Static and Dynamic Carbonation of Lightweight Concrete Masonry Units

Hilal El-Hassan

Department of Civil Engineering and Applied Mechanics McGill University, Montreal

December 2012

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Doctor of Philosophy in Engineering

© Hilal El-Hassan, 2012 All Rights Reserved

ABSTRACT Static and dynamic carbonation curing at early age was developed for ordinary Portland cement (OPC) and Portland limestone cement (PLC) concrete masonry units (CMU) production. It is intended to replace conventional steam curing, improve the CMU performance, reduce energy consumption, and permanently sequester carbon dioxide in concrete. Concrete slabs representing the face shell of a 20-cm CMU as well as full sized CMU were used throughout the carbonation process. In static carbonation, it was found that initial air curing was vital to maximize carbonation reaction. After a procedure of casting, air curing, carbonation curing, and water compensation in subsequent hydration, carbonated CMUs had shown equivalent strength to steamed CMU but much better resistance to freeze-thaw damage. Carbonate-reinforced cement matrix played a critical role in improving freeze thaw resistance. In dynamic carbonation, the initial air curing was combined with carbonation with controlled relative humidity. The production cycle was significantly reduced to avoid initial air curing. The process proved to be a valid replacement of the static system in terms of CO2 uptake and compressive strength. While both OPC and PLC concretes displayed the hydration and carbonation products, only OPC concrete demonstrated an intermix of these products in the form of calcium silicate hydrocarbonate and a phase transformation of poorly crystalline aragonite and vaterite into well crystalline calcite. Based on 24% CO2 uptake, the CMU production in US and Canada is capable of sequestering 2 million tons CO2 per year. It is equivalent to 2.5% carbon emission reduction for US and Canada cement industry.

– i – RÉSUMÉ La carbonatation par méthode statique et dynamique a été développée pour la cure rapide de blocs de bétons composés à partir de ciment Portland ordinaire ainsi que de ciment Portland à base de calcaire. Cette approche vise à remplacer le procédé traditionnel de cure de blocs de bétons par étuvage afin d’améliorer leur performance, réduire la consommation d’énergie, et séquestrer le dioxyde de carbone de manière indéfinie. La façade extérieure d’un bloc de béton de 20-cm, représenté par une dalle de béton, ainsi qu’un bloc de pleine taille, ont été utilisés durant le procédé de carbonatation. Les résultats indiquent qu’il est essentiel de curer les blocs par air contrôlé avant d’employer la carbonatation statique. Suivant la procédure de moulage, cure à l’air, carbonatation, et compensation de l’eau à travers hydratation suivie, les blocs de bétons carbonatés ont témoigné une résistance comparable à celle de blocs durcis à la vapeur, cependant une résistance supérieure aux dégâts de gel-dégel. La microstructure du ciment carbonaté-renforcé a joué un rôle crucial dans l’amélioration de la résistance du gel- dégel. Dans la carbonatation dynamique, le durcissement initial par étuvage a été combiné avec carbonatation sous une humidité relative contrôlée. Afin de réduire le cycle de production, le durcissement initial par étuvage a été éliminé. La carbonatation dynamique s’est avérée être un remplacement valable du système statique en termes d’absorption de CO2 et résistance à la compression. Bien que le ciment Portland ordinaire ainsi que le ciment Portland à base de calcaire ont confirmé des produits d’hydratation et de carbonatation, seul le ciment Portland a fait preuve de la capacité de ses produits d’hydratation et de carbonatation de se mélanger sous la forme de calcium hydrocarbonate silicate . De plus, de l'aragonite mal cristallisé et de la vatérite ont subi une transformation de phase dérivant en calcite cristalline. Basé sur une capacité d’absorption de CO2 de 24%, la production de blocs de bétons aux États-Unis et au

Canada a le potentiel annuel de séquestrer 2 millions de tonnes de CO2. Ce fait signifie que la réduction d’émissions de dioxyde de carbone de ces deux pays dans l’industrie de ciment est égale à 2.5%.

– ii – ACKNOWLEDGEMENTS Foremost, I would like to thank my supervisor, Professor Yixin Shao, for his patient guidance, motivation, enthusiasm, and inspiration. I sincerely appreciate all his contributions of time and ideas to make my PhD experience productive and stimulating. The expanded slag aggregates and cement have been provided from Lafarge Canada. I would like to especially thank the President of Boehmer block, Mr. Paul Hargest, whose support and interest in our work motivated us to prosper for perfection. Microstructure study was carried out with the technical support of the Department of Mining and Materials Engineering, specifically Helen Campbell, Line Mongeon, and Pierre Hudon. The full understanding of lab equipment and experiments was possible through the aid of Ron Sheppard, Bill Cook, Marek Przykorski, and John Bartczak. Lastly, I would like to thank my fiancé, Nihale, for her patience, support, and understanding. The constant encouragement of my parents is what drove me to thrive for excellence throughout these years. Special thanks to my friend, Zaid Ghouleh, whose extensive help in the microstructure analysis was of great value. I would also like to thank my friend, Lana Tayyara, for translating the abstract to French.

– iii – CONTRIBUTIONS OF AUTHORS The manuscripts included in this thesis and Chapters 3 to 8 will be or have been submitted for publication in scientific journals. The author was responsible for conducting the research, analyzing the data, and preparing the manuscripts. The author’s supervisor, Dr. Yixin Shao, provided general guidance and editorial revisions throughout the entire process. Also, Mr. Zaid Ghouleh contributed to the presented research particularly in microstructural analysis.

– iv – TABLE OF CONTENTS Abstract ...... i Résumé ...... ii Acknowledgements ...... iii Contributions of Authors ...... iv List of Tables ...... xi List of Figures ...... xiii

Chapter 1 - Introduction ...... 1 Overview ...... 1 Steam curing ...... 1 Carbonation curing ...... 2

CO2 utilization in concrete masonry units production...... 2 Carbon dioxide emission ...... 3 Research Objectives ...... 3 Thesis Structure...... 5 References ...... 6 Figures ...... 8

Chapter 2 - Literature Review ...... 9 Carbonation Curing of Concrete Masonry Units ...... 9 Weathering carbonation ...... 9 Early-age carbonation ...... 10 Carbonation products ...... 11 Other carbonation applications ...... 12 Factors influencing carbonation ...... 13

Quantification of CO2 Uptake ...... 14 Mass gain method ...... 14 Mass curve method ...... 15 Thermogravimetry (TG) method ...... 15 Constant temperature pyrolysis method ...... 16

– v – Chemical analysis by hydrochloric acid titration method ...... 16 Strength Development...... 17 Early strength ...... 17 Late strength ...... 17 Durability ...... 18 Energy Consumption...... 19 Portland Limestone Cement (PLC) in Concrete ...... 19 References ...... 20

Chapter 3 - Static Carbonation of Lightweight OPC Concrete Masonry Units ..... 28 Preface ...... 28 Introduction ...... 30 Experimental Investigations ...... 31 Concrete sample preparation ...... 31 Curing procedures ...... 32 Internal relative humidity measurement ...... 33

CO2 uptake estimation ...... 33 Performance evaluation ...... 34 Experimental Results and Discussion ...... 35 Effect of initial curing on internal relative humidity ...... 35 Effect of initial curing on carbonation reaction ...... 36 Compressive strength ...... 39 pH measurement ...... 40 Carbonation curing of 20-cm CMU...... 40 Carbon storage in blocks and bricks ...... 42 The network operation and cost estimate ...... 42 Conclusions ...... 43 Acknowledgements ...... 44 References ...... 45 Tables ...... 47 Figures ...... 52

– vi – Chapter 4 – Reaction Products of Lightweight OPC Concrete Masonry Units by Static Carbonation ...... 62 Preface ...... 62 Introduction ...... 63 Experimental Investigations ...... 65 Concrete sample preparation ...... 65

Measurement of CO2 in concrete...... 65 Performance evaluation ...... 67 Method of determining cement content (CC) in powder samples ...... 67 Experimental Results and Discussion ...... 69 Carbonation behavior ...... 69 Compressive strength ...... 70 Phase analysis ...... 71 TG/DTG analysis ...... 71 SEM analysis ...... 75 Conclusions ...... 76 Acknowledgements ...... 77 References ...... 77 Tables ...... 80 Figures ...... 82

Chapter 5 - Static Carbonation of Lightweight PLC Concrete Masonry Units...... 89 Preface ...... 89 Introduction ...... 90 Experimental Investigations ...... 91 PLC concrete sample preparation ...... 91 Curing procedures ...... 92 Internal relative humidity measurement ...... 93

CO2 uptake estimation ...... 94 Performance evaluation ...... 95 Experimental Results and Discussion ...... 95

– vii – Effect of initial curing on water content ...... 95 Effect of initial curing on carbonation reaction ...... 97 Compressive strength ...... 99 pH measurement ...... 100 Static carbonation curing of 20-cm PLC CMU ...... 101 Conclusions ...... 102 Acknowledgements ...... 103 References ...... 103 Tables ...... 106 Figures ...... 109

Chapter 6 - Reaction Products of Lightweight PLC Concrete Masonry Units by Static Carbonation ...... 116 Preface ...... 116 Introduction ...... 117 Experimental Investigations ...... 118 PLC concrete sample preparation ...... 118 Method of determining cement content (CC) in powder samples ...... 119

Measurement of CO2 in PLC concrete ...... 120 Performance evaluation ...... 121 Experimental Results and Discussion ...... 122

Cement content and CO2 content in concrete powder ...... 122 Compressive strength ...... 123 Phase analysis ...... 123 TG/DTG analysis ...... 124 SEM analysis ...... 127 Comparison with ordinary Portland cement ...... 128 Conclusions ...... 129 Acknowledgements ...... 130 References ...... 130 Tables ...... 133

– viii – Figures ...... 135

Chapter 7 - Dynamic Carbonation of Lightweight OPC Concrete Masonry Units...... 143 Preface ...... 143 Introduction ...... 144 Experimental Investigations ...... 145 Concrete sample preparation ...... 145 Curing procedures ...... 146

CO2 uptake estimation ...... 147 Method of determining cement content (CC) in powder samples ...... 148 Performance evaluation ...... 149 Experimental Results and Discussion ...... 149 Dynamic carbonation behavior ...... 149 Compressive strength ...... 151 Relative humidity control ...... 152 Effect of carbonation pressure ...... 153 Phase analysis ...... 153 TG/DTG analysis ...... 154 SEM analysis ...... 158 Dynamic carbonation of 20-cm CMU ...... 159 Conclusions ...... 160 Acknowledgements ...... 162 References ...... 162 Tables ...... 165 Figures ...... 168

Chapter 8 - Performance Comparison of OPC and PLC Concretes Subject to Early Carbonation ...... 179 Preface ...... 179 Introduction ...... 180

– ix – Experimental Investigations ...... 181 Concrete sample preparation and curing ...... 181

CO2 uptake estimation ...... 181 Performance evaluation ...... 181 Experimental Results and Discussion ...... 183 Carbonation degree ...... 183 Compressive strength ...... 184 Water absorption...... 185 Freeze-thaw resistance ...... 186 Phase analysis ...... 187 TG/DTG analysis ...... 187 SEM analysis ...... 189 Carbonation curing model ...... 190 Conclusions ...... 192 Acknowledgements ...... 193 References ...... 193 Tables ...... 195 Figures ...... 196

Chapter 9 - Conclusions ...... 206 Conclusions ...... 206 Suggestions for Future Work ...... 211 Statement of Originality ...... 213

– x – LIST OF TABLES

Table 3.1: CO2 uptake and compressive strength (MPa) of different aggregate concretes ...... 47 Table 3.2: Mixture proportion ...... 47 Table 3.3: Curing procedures ...... 48 Table 3.4: Thermal decomposition analysis after 28 days ...... 49

Table 3.5: Comparison of CO2 uptake by three methods ...... 50 Table 3.6: Compressive strength of concrete with water spray ...... 50 Table 3.7: pH of surface sample at 28 days ...... 51

Table 3.8: CO2 uptake in 20-cm CMU samples ...... 51 Table 3.9: Energy consumption and cost estimate ...... 51 Table 4.1: Mixture proportion of slab samples ...... 80 Table 4.2: Chemical composition of the cement ...... 80

Table 4.3: CO2 content and uptake from concrete samples ...... 80 Table 4.4: Cement content (CC) in powder samples ...... 80

Table 4.5: CO2 content and uptake from powder samples ...... 81 Table 4.6: Total calcium hydroxide and calcium carbonate content (%) ...... 81 Table 4.7: Dehydration mass and decarbonation mass ...... 81 Table 5.1: Chemical composition of PLC ...... 106 Table 5.2: Mixture proportion of PLC concrete ...... 106 Table 5.3: Curing conditions of PLC concrete ...... 106

Table 5.4: Comparison of CO2 uptake by three methods ...... 107 Table 5.5: pH measurement after 28 days ...... 107

Table 5.6: CO2 uptake of PLC concrete blocks ...... 108 Table 6.1: Chemical composition of PLC ...... 133 Table 6.2: Mixture proportion of PLC concrete ...... 133 Table 6.3: Cement content (CC) in powder samples ...... 133

Table 6.4: CO2 content in powder samples ...... 133 Table 6.5: Percentage of produced poorly and well crystalline calcium carbonates .... 133 Table 6.6: Total calcium hydroxide and calcium carbonate content (%) ...... 134 Table 6.7: Dehydration mass and decarbonation mass of PLC concrete ...... 134

– xi – Table 7.1: Mixture proportion of concretes ...... 165 Table 7.2: Curing conditions ...... 165 Table 7.3: Water content in concretes ...... 165

Table 7.4: CO2 content (%) and CO2 uptake (%) by thermal analysis ...... 166 Table 7.5: Cement content (CC) in concrete powder samples ...... 166 Table 7.6: Effect of gas pressure on dynamic carbonation ...... 167 Table 7.7: Percentage of poorly and well crystalline carbonates ...... 167 Table 7.8: Calcium hydroxide and calcium carbonate content after 1 and 28 days ..... 167 Table 7.9: Dehydration and decarbonation mass ...... 167

Table 7.10: CO2 content and uptake of concrete blocks ...... 167 Table 8.1: Chemical composition of OPC and PLC concretes ...... 195

Table 8.2: CO2 uptake and content (%) of OPC and PLC concretes ...... 195 Table 8.3: Percentage of produced poorly and well crystalline carbonated in OPC and PLC concretes ...... 195

– xii – LIST OF FIGURES Fig. 3.1: Sieve analysis of different aggregates ...... 52 Fig. 3.2: Sieve analysis of expanded shale ...... 52 Fig. 3.3: Sieve analysis of expanded slag aggregates ...... 53 Fig. 3.4: Schematic of carbonation setup ...... 53 Fig. 3.5: Schematic of measuring internal relative humidity ...... 54 Fig. 3.6: Water loss due to initial curing ...... 54 Fig. 3.7: Internal relative humidity during initial curing ...... 55

Fig. 3.8: Water loss due to carbonation and CO2 uptake ...... 55 Fig. 3.9: Mass curves of 4-hour carbonation ...... 56 Fig. 3.10: Mass curve of 4-day carbonation ...... 56 Fig. 3.11: Compressive strength after 4 hours ...... 57 Fig. 3.12: Compressive strength after 1 day ...... 57 Fig. 3.13: Compressive strength with varied carbonation duration ...... 58 Fig. 3.14: Effect of initial curing on compressive strength ...... 58 Fig. 3.15: Effect of carbonation time on compressive strength ...... 59 Fig. 3.16: Standard CMU made in lab ...... 59 Fig. 3.17: Compressive strength of CMU ...... 60 Fig. 3.18: XRD patterns of CMU blocks ...... 60 Fig. 3.19: Carbonated concrete (18a+4c) after 28 days sprayed with Phenolphthalein . 61 Fig. 3.20: Hydrated concrete (0a) after 28 days sprayed with Phenolphthalein ...... 61 Fig. 4.1: Schematic of carbonation setup ...... 82 Fig. 4.2: Compressive strength of concretes at 1-day and 28-day ...... 82 Fig. 4.3: XRD patterns of concretes (1-day) ...... 83 Fig. 4.4: XRD patterns of concretes (28-day) ...... 83 Fig. 4.5: DTG curves of concretes of 1-day ...... 84 Fig. 4.6: Mass loss due to dehydration and decarbonation in concretes (1-day) ...... 84 Fig. 4.7: DTG curves of concretes of 28-day ...... 85 Fig. 4.8: Mass loss due to dehydration and decarbonation in concretes (28-day) ...... 85 Fig. 4.9: SEM micrograph of hydrated reference concrete (0a) after 28 days ...... 86 Fig. 4.10: SEM micrograph of carbonated concrete (0a+4c) after 28 days ...... 86

– xiii – Fig. 4.11: SEM micrograph of carbonated concrete (18a+4c) after 1 day ...... 87 Fig. 4.12: SEM micrograph of carbonated concrete (18a+4c) after 28 days ...... 87 Fig. 4.13: SEM micrograph of carbonated concrete (18a+4c+sp) after 28 days ...... 88 Fig. 4.14: SEM micrograph of carbonated concrete (18a+4c+sp) after 28 days ...... 88 Fig. 5.1: Schematic of carbonation setup ...... 109 Fig. 5.2: Schematic of measuring internal relative humidity ...... 109 Fig. 5.3: Water loss during initial curing ...... 110 Fig. 5.4: Relative humidity inside PLC sample during initial curing ...... 110

Fig. 5.5: Water loss due to carbonation and CO2 uptake (Eq. 5.1) ...... 111 Fig. 5.6: Mass curves of carbonated PLC concretes ...... 111 Fig. 5.7: Thermal decomposition analysis after 28 days ...... 112 Fig. 5.8: Compressive strength after 4 hours ...... 112 Fig. 5.9: Compressive strength after 1 day ...... 113 Fig. 5.10: Effect of initial curing on 1-day compressive strength...... 113 Fig. 5.11: Compressive strength after 28 days ...... 114 Fig. 5.12: Effect of initial curing on 28-day compressive strength ...... 114 Fig. 5.13: Full-size 20-cm PLC concrete block ...... 115 Fig. 5.14: Compressive strength of PLC concrete blocks ...... 115 Fig. 6.1: Schematic of carbonation setup ...... 135 Fig. 6.2: Compressive strength of PLC concretes ...... 135 Fig. 6.3: XRD patterns of PLC concretes after 1 day ...... 136 Fig. 6.4: XRD patterns of PLC concretes after 28 days ...... 136 Fig. 6.5: DTG curves of 1-day PLC concretes ...... 137 Fig. 6.6: Mass loss due to dehydration and decarbonation in 1-day PLC concretes .... 137 Fig. 6.7: DTG curves of 28-day PLC concretes ...... 138 Fig. 6.8: Mass loss due to dehydration and decarbonation in 28-day PLC concretes .. 138 Fig. 6.9: SEM micrograph of PLC hydrated reference (0a) after 1 day ...... 139 Fig. 6.10: SEM micrograph of PLC carbonated concrete (18a+4c) after 1 day ...... 139 Fig. 6.11: SEM micrograph of PLC carbonated concrete (18a+4c) after 28 days ...... 140 Fig. 6.12: SEM micrograph of sprayed PLC carbonated concrete (18a+4c+sp) after 1 day ...... 140

– xiv – Fig. 6.13: SEM micrograph of sprayed PLC carbonated concrete (18a+4c+sp) after 28 days ...... 141 Fig. 6.14: SEM micrograph of sprayed PLC carbonated concrete (18a+4c+sp) after 28 days ...... 141 Fig. 6.15: SEM micrograph of PLC powder ...... 142 Fig. 7.1: Schematic of carbonation setup ...... 168 Fig. 7.2: Six-hour compressive strength of slab samples ...... 168 Fig. 7.3: Compressive strength of slab samples ...... 169 Fig. 7.4: Relative humidity and temperature during 4-hour carbonation ...... 169 Fig. 7.5: XRD patterns of 1-day slab samples ...... 170 Fig. 7.6: XRD patterns of 28-day slab samples ...... 170 Fig. 7.7: Comparison of DTG curves of 1-day slab samples ...... 171 Fig. 7.8: Mass loss due to dehydration and decarbonation in 1-day concretes ...... 171 Fig. 7.9: Comparison of DTG curves of 28-day slab samples ...... 172 Fig. 7.10: Mass loss due to dehydration and decarbonation in 28-day concretes ...... 172 Fig. 7.11: SEM micrograph of carbonated concrete (2a+4c) after 28 days ...... 173 Fig. 7.12: SEM micrograph of carbonated concrete (2a+4c) after 28 days ...... 173 Fig. 7.13: SEM micrograph of carbonated concrete (2a+4c+sp) after 1 day ...... 174 Fig. 7.14: SEM micrograph of carbonated concrete (2a+4c+sp) after 28 days ...... 174 Fig. 7.15: SEM micrograph of carbonated concrete (2a+18c) after 1 day ...... 175 Fig. 7.16: SEM micrograph of carbonated concrete (2a+18c) after 28 days ...... 175 Fig. 7.17: SEM micrograph of steam concrete (2a+4s) after 28 days ...... 176 Fig. 7.18: SEM micrograph of steam concrete (2a+4s) after 28 days ...... 176 Fig. 7.19: SEM micrograph of hydrated concrete (0a) after 28 days ...... 177 Fig. 7.20: Full-size 20-cm concrete block ...... 177 Fig. 7.21: Compressive strength of 20-cm concrete blocks ...... 178 Fig. 8.1: Schematic of freeze-thaw test ...... 196 Fig. 8.2: Compressive strength of OPC and PLC concretes ...... 196 Fig. 8.3: Water absorption of 28-day OPC and PLC concretes ...... 197 Fig. 8.4: Freeze-thaw resistance of 28-day OPC and PLC concretes ...... 197 Fig. 8.5: Hydrated OPC concrete after 6 freeze-thaw cycles (0a-Ref) ...... 198

– xv – Fig. 8.6: Hydrated PLC concrete after 6 freeze-thaw cycles (0a-Ref) ...... 198 Fig. 8.7: Commercial concrete block after 6 freeze-thaw cycles ...... 199 Fig. 8.8: Steamed OPC concrete after 6 freeze-thaw cycles (2a+4s) ...... 199 Fig. 8.9: Carbonated OPC concrete after 10 freeze-thaw cycles (18a+4c) ...... 200 Fig. 8.10: Carbonated OPC concrete with water compensation after 10 freeze-thaw cycles (18a+4c+sp) ...... 200 Fig. 8.11: Carbonated PLC concrete after 10 freeze-thaw cycles (18a+4c) ...... 201 Fig. 8.12: Carbonated PLC concrete with water compensation after 10 freeze-thaw cycles (18a+4c+sp) ...... 201 Fig. 8.13: XRD patterns of 1-day OPC and PLC concretes ...... 202 Fig. 8.14: XRD patterns of 28-day OPC and PLC concretes ...... 202 Fig. 8.15: DTG curves of 1-day OPC and PLC concretes ...... 203 Fig. 8.16: DTG curves of 28-day OPC and PLC concretes ...... 203 Fig. 8.17: SEM micrograph of OPC and PLC hydrated reference (0a) after 1 day ...... 204 Fig. 8.18: SEM micrograph of OPC and PLC hydrated reference (0a) after 28 days .. 204 Fig. 8.19: SEM micrograph of carbonated OPC and PLC concrete (18a+4c) after 1 day ...... 204 Fig. 8.20: SEM micrograph of carbonated OPC and PLC concrete (18a+4c+sp) after 28 days ...... 205 Fig. 8.21: Carbonation curing model: (a) After casting; (b) After initial air curing; (c) After carbonation; (d) After water compensation and subsequent hydration ...... 205

– xvi – Chapter 1

–

INTRODUCTION

OVERVIEW Concrete masonry units (CMU) have been utilized in the building construction industry since the 1950s. Whether as load bearing or non-load bearing walls, the North American market for concrete blocks and bricks is projected to increase to 4.3 billion units per year in 2014 (Freedonia 2010). In comparison to cast-in-place concrete, masonry block structures not only exhibit superior performance due to precast quality, but also represent a low environmental impact construction system. The increased demand for masonry units can be explained by comparing one square meter of solid concrete wall to a masonry unit wall. The latter requires 48% less material due to internal cavities, consumes 65% less cement for equivalent strength (10% compared to 15% in regular concrete), reduces CO2 emission by 65%, and enhances thermal insulation efficiency by 70%.

Steam curing The North American concrete block is normally steam cured. Autoclaving or high-pressure steam curing is utilized in several block plants, but the more dominant curing method is atmospheric steam curing. The 24-hour cycle is carried out in a steam kiln at a relative humidity of approximately 95% and temperature of 70ºC. This energy- intensive process consumes 0.59 GJ and 0.71 GJ of energy for 1 cubic meter in block form for atmospheric and high-pressure steam curing, respectively (Kawai 2005). This procedure ensures high early strength for transport and an accelerated production cycle for maximum plant efficiency.

– 1 – Carbonation curing Carbonation curing was systematically studied in the 1970s to replace the energy- intensive steam curing (Berger 1972; Klemm 1972; Young 1974; Sorochkin 1975; Bukowski 1979; Goodbrake 1979a; Goodbrake 1979b). This alternative curing method uses high purity carbon dioxide to accelerate curing. However, there has been no large- scale commercial application since then, possibly due to the high cost of CO2 gas production, which is difficult to justify by the accelerated production. Even so, this situation may change in the near future. As regulations requiring reductions in CO2 emissions are developed, large quantities of high purity, low cost carbon dioxide could soon be available. The high purity CO2 can be recovered from cement kiln flue gas. The energy required for CO2 recovery and compression is estimated to be 0.89 GJ/ton of CO2

(=247 kWh/tCO2) (Halmann 1999). Assuming masonry concrete has a density of 2400 3 kg/m , a cement content of 10% by total mass and a CO2 uptake capacity of 10% by cement mass, 24 kg of CO2 gas is required for one cubic meter of masonry concrete to perform carbonation curing (Shao et al. 2006). The energy consumption in recovering 3 that amount of CO2 is approximately about 0.02 GJ/m . If CO2 uptake capacity of concrete can be enhanced to its theoretical limit, 50% of cement mass (Steinour 1956), the energy consumption will be increased in the recovery process to 0.1 GJ per one m3 concrete. In comparison to the energy used in atmospheric steam curing (0.59 GJ per m3 concrete), the total energy required by the carbonation process can still be significantly lower, even if the energy for CO2 transport is also considered. This energy efficiency will motivate concrete block producers to consider this economically beneficial industry process.

CO2 utilization in concrete masonry units production Concrete blocks are mass-produced, porous in nature, and require accelerated curing. These factors make CMU ideal candidate products for the carbonation process and CO2 utilization. Early research had found that carbonation treatment of steamed cured CMU could reduce service shrinkage and its related cracking in masonry structures (Shideler 1955; Toennies 1963). Reaction of cement with carbon dioxide at early age is a

CO2 sequestration process. One 20-cm block, with a mass of 18 kg containing 10%

– 2 – cement, could sequester at least 0.18 kg of CO2 (Shao 2006). With a 10% CO2 uptake capacity, the annual production of 4.3 billion bricks and blocks can sequester 1.278 million tons of CO2 gas per year. If the CO2 uptake capacity can be doubled, the carbon sequestration in concrete blocks can reach 2.5 million tons per year.

Carbon dioxide emission Fig. 1.1 is an estimation of the carbon dioxide emission up to the year 2030

(Mongabay 2009). The cement industry contributes around 5% of global CO2 emission, though it is making efforts to reduce CO2 emission. One of the sustainable developments is the use of blended cements to replace traditional Portland cement to reduce the clinker demand and its associated emission. Ground granulated blast furnace (GGBF) slag has been used in making slag cement for more than thirty years. Portland limestone cement (PLC) with replacement of 4-14% cement clinker with ground limestone is a relatively new product of blended cement. If concrete blocks can be made with Portland limestone cement, and carbonation curing can still be effectively utilized, the produced blocks are truly green products, and the process will show environmental, economical, and technical benefits.

RESEARCH OBJECTIVES While past research has targeted rapid strength gain, reduction in service shrinkage, and acceleration in setting using prolonged carbonation curing, the proposed research will develop an early-age carbonation technology to maximize CO2 uptake in concrete masonry units. This will be achieved by optimizing process parameters and curing systems in a 24-hour timeframe while maintaining comparable compressive strength and durability in comparison to steam-cured and hydrated counterparts. Early-age carbonation refers to carbonation within 24 hours after cast forming. It can be an immediate carbonation after casting or carbonation after initial hydration for a few hours. While the former reaction takes place between calcium silicates and carbon dioxide when concrete is fresh, the latter occurs between hydration products and carbon dioxide after concrete hardens in initial curing. The different process window requires different preconditioning of the CMU. Both methods are suited for large-scale

– 3 – production. Unlike previous work on early carbonation which was focused more on cement pastes and mortars, the current work will employ lightweight concrete mixes with expanded slag aggregates to simulate the real concrete products and promote full-scale applications. Static and dynamic systems are studied in search for the optimum method that satisfies the main objective of this research. Both ordinary Portland cement (OPC) and Portland limestone cement (PLC) concretes will undergo static carbonation. It is performed after controlled initial air curing in an environmental chamber, while dynamic carbonation is carried out after 2-hour open air curing for OPC concrete to demonstrate its feasibility. In the dynamic system, diffusion of carbon dioxide into concrete is accomplished by circulating the CO2 at low pressure. The proposed tasks will investigate:

 The effect of controlled initial hydration on CO2 uptake and compressive strength.  The effect of early-age static carbonation on the concrete microstructure.  The effect of replacing OPC by PLC in the mix design to produce greener concrete.  The production of 20-cm concrete blocks to validate the results obtained from slab samples.

 The effect of dynamic carbonation on CO2 uptake, compressive strength, and the concrete microstructure for fresh concrete process.  A method to determine cement content in concrete powder for quantitative analysis of hydration and carbonation products using thermogravimetric analysis (TGA).

 A comparison of OPC and PLC concrete in terms of CO2 uptake, compressive strength, and microstructural changes.  The effect of early-age carbonation on the durability of concrete.

– 4 – THESIS STRUCTURE This thesis follows a manuscript-based structure. A preface is included in each chapter to introduce the objectives, the background, and the connection to other chapters. Chapter 1 serves as an introduction to the research along with the objectives and background needs. Chapter 2 presents the literature review and emphasizes the past work done in the domain of carbonation. Chapters 3, 4, 5, 6, and 7 are on static carbonation of OPC and PLC concretes. Each chapter consists of an introduction, literature review, and a conclusion, and therefore, repetition is, in some cases, inevitable. Chapter 3 highlights the extensive study on the effect of initial hydration on the carbonation degree and compressive strength of OPC concrete. Static carbonation is the curing system employed. More than 200 samples were cast in this part to select the most appropriate batch for further microstructure examination. Full-size (20-cm) OPC concrete blocks were also produced to confirm the possibility of scale-up. Chapter 4 investigates the effect of static carbonation on the microstructure of OPC concrete in comparison to hydration using optimized batches from Chapter 3. TG/DTG, SEM, XRD, and chemical analysis by hydrochloric acid (HCl) titration are employed for this study. In order to produce greener concrete blocks with the incorporation of less cement, OPC is replaced by PLC. Chapter

5 and 6 discuss the effect of this replacement on CO2 uptake, compressive strength, and concrete microstructure. Controlled initial air curing utilized through Chapters 3 to 6 is a time-consuming process. It takes 4-18 hours for preconditioning. In an attempt to reduce the total process time during carbonation, a dynamic system is discussed in Chapter 7. The process initially cures concrete for 2 hours in open air, but uses a pump/vacuum system to circulate the CO2 during carbonation. This reduced the process cycle from 1 day to 6 hours and allowed simultaneous carbonation and drying. The effect of dynamic carbonation on microstructure was also investigated. A complete comparison of OPC and

PLC concretes in terms of CO2 uptake, compressive strength, and microstructural changes is presented in Chapter 8. The conclusions are summarized in Chapter 9. Suggestions for future work are also presented.

– 5 – REFERENCES Berger, R. L., Young, J. F., and Leung, K. (1972). "Accelerated curing of cementitious systems by carbon dioxide - Hydraulic calcium silicates and aluminates - Part II." Cement and Concrete Research 2: 647-652.

Bukowski, J. M., and Berger, R. L. (1979). "Reactivity and strength development of CO2 activated non-hydraulic calcium silicates." Cement and Concrete Research 9(1): 57-68.

Freedonia, G. (2010). "Brick and block demand to reach 12.4 billion units, $8 billion by 2014." Journal of Concrete Products (http://www.freedoniagroup.com/DocumentDetails.aspx?DocumentId=506289 ).

Goodbrake, C. J., Young, J. F., and Berger, R. L. (1979a). "Reaction of beta-dicalcium silicate and tricalcium silicate with carbon dioxide and water." Journal of the American Ceramic Society 62(3-4): 168-171.

Goodbrake, C. J., Young, J. F., and Berger, R. L. (1979b). "Reaction of hydraulic calcium silicates with carbon dioxide and water." Journal of the American Ceramic Society 62(9-10): 488-491.

Halmann, M. (1999). "Greenhouse gas carbon dioxide mitigation: Science and Technology." Boca Raton, Fla, Lewis Publishers.

Kawai, K., and Sugiyama, T. (2005). "Inventory data and case studies for environmental performance evaluation of concrete structures." Journal of Advanced Concrete Technology: 435-456.

Klemm, W. A., and Berger, R. L. (1972). "Accelerated curing of cementitious systems by carbon dioxide - Part I - Portland Cement." Cement and Concrete Research 2: 567-576.

Mongabay (2009). "Carbon Emission Charts by Country 1990-2030." (http://rainforests.mongabay.com/09-carbon_emissions.htm).

Shao, Y., Mirza, M. S., and Wu, X. (2006). "CO2 sequestration using calcium-silicate concrete." Canadian Journal of Civil Engineering 33: 776-784.

Shideler, J. J. (1955). "Investigation of the moisture-volume stability of concrete masonry units." Portland Cement Association Research and Development Laboratories Bulletin D3.

Sorochkin, M. A., Shchurov, A. F., and Safonov, I. A. (1975). "Study of The Possibility of Using Carbon Dioxide for Accelerating the Hardening of Products Made From Portland Cement." Journal of Applied Chemistry of the USSR 48(6): 1211-1217.

– 6 – Steinour, H. (1956). "Some Effects of Carbon Dioxide on Mortar and Concrete." American Concrete Institute Journal 30: 905-907.

Toennies, H. T., and Shideler, J. J. (1963). "Plant drying and carbonation of concrete block - NCMA-PCA cooperative program." American Concrete Institute Journal 60(33): 617-632.

Young, J. F., Berger, R. L., Breese, J. (1974). "Accelerated Curing of Compacted Calcium Silicate Mortars on Exposure to CO2." Journal of the American Ceramic Society 57(9): 394-397.

– 7 – FIGURES

Fig. 1.1: World carbon dioxide emissions by country

– 8 – Chapter 2

–

LITERATURE REVIEW

CARBONATION CURING OF CONCRETE MASONRY UNITS The effect of static carbonation on concrete masonry units has been studied in the 50s and 60s. Initially, carbonation was intended to reduce the service shrinkage of concrete blocks by drying and preshrinking at early stages of hydration (Leber 1956; Polisner 1956; Toennies 1960; Toennies 1963; Freedman 1969). Concrete masonry units were initially preset at 50% relative humidity and 25ºC, steam cured for 16 hours at 71-

77ºC, and carbonated by flue gas (10%), pure CO2 (99.5%), or a combination of the two at 37-93ºC and 45-70% RH (Toennies 1960). Other concrete block carbonation schemes were a combination of 6-24 hours of steam curing at 65-121ºC and RH of 14-25% and carbonation with flue gas of 10% concentration (Toennies 1963). The resulting drying shrinkage was reduced by 42-43% due to carbon dioxide exposure (Toennies 1960;

Toennies 1963). The resulting CO2 uptake reached 27% after 14 days of air curing at 50% relative humidity and 25ºC and 4 days of carbonation at 53% relative humidity (Shideler 1955). The 28-day compressive strength of the carbonated sample was found to be higher than the hydrated reference (Shideler 1955). The static and dynamic carbonation of freshly cast concrete blocks in a 24-hour timeframe have not yet been investigated.

Weathering carbonation Carbonation curing is different from weathering carbonation in that the former takes place in a short period of time after casting, while the latter is a slow process afflicting mature concrete in service (Neville 1996). The reaction involves atmospheric

CO2 (partial pressure = 0.038 atm) and concrete hydration products (Tans 2008). It has a detrimental effect on the hydrated, hardened concrete. Past research studied the effect of atmospheric weathering carbonation on the physical and transport properties of concrete

– 9 – (Papadakis 1991). The primary consequence of atmospheric carbonation is shrinkage, which can lead to unstable dimensional change and cracking in restraint (Powers 1962). Another significant deterioration mechanism resulting from weathering carbonation is the corrosion of steel reinforcement (Parrott 1996; Bertolini 2006). Calcium hydroxide (CH) is one of the concrete hydration products. Its basic structure produces a passivation layer for reinforcing steel and protects it from corrosion.

However, long term exposure to low concentration CO2 consumes the CH according to Eq. 2.1, leading to a drop in pH of pore solution below the threshold value, and ultimately the corrosion of the steel reinforcement (Moorehead 1986; Richardson 1988; Parrott 1996; Bertolini 2006; Mehta 2006):

Ca(OH)2 + CO2 → CaCO3 + H2O (2.1)

Another hydration product, C-S-H, reacts with CO2 (Eq. 2.2) and becomes progressively decalcified upon extended exposure leading to the eventual formation of silica gel (Young 1974; Goodbrake 1979b). xCaO·SiO2·yH2O + (x-x’)CO2 → x’CaO·SiO2·y’H2O + (x-x’)CaCO3 + (y-y’)H2O (2.2) Gypsum is present in cement clinker to defer flash set. The hydration of gypsum in concrete produces ettringite, which in turn carbonates (Eq. 2.3) and produces calcite, gypsum, and aluminate gel as (Nishikawa 1992; Grounds 1998):

6CaO·Al2O3·3SO3·32H2O + 3CO2 →

3CaCO3 + 3CaO·SO3·2H2O + Al2O3·xH2O + (26-x)H2O (2.3)

The monosulphoaluminate, or AFm, phase produced due to hydration of tricalcium aluminate will react with carbon dioxide according to Eq. 2.4 (Venhuis 2001):

4CaO·Al2O3·SO3·12H2O + 3CO2 →

3CaCO3 + CaO·SO3·2H2O + Al2O3·3H2O + 7H2O (2.4)

Early-age carbonation Unlike weathering carbonation, early-age carbonation is when fresh or partially hydrated concrete is deliberately exposed to carbon dioxide to accelerate early strength and improve mechanical properties (Berger 1972; Young 1974; Bukowski 1979; Goodbrake 1979b). This early exposure to carbon dioxide would contribute to the rapid

– 10 – strength gain through the following reactions (Young 1974; Goodbrake 1979b):

C3S+(3-x)CO2+yH2O→CxSHy+(3-x)CaCO3 (2.5)

C2S+(2-x)CO2+yH2O→CxSHy+(2-x)CaCO3 (2.6)

The aqueous reaction of CO2 with C2S and C3S generates C-S-H and calcium carbonates without producing calcium hydroxide as seen in Eq. 2.5 and 2.6 (Berger 1972; Young 1974). The initially formed C-S-H is suggested to be similar in microstructure to that generated by normal hydration and is further carbonated according to Eq. 2.2 to produce silica gel and calcium carbonates (Goodbrake 1979b). The CO2-induced binding effect offers high early strength within a few minutes to a few hours (Young 1974).

At later stages of hydration, concrete consists of C2S, C3S, C-S-H, and CH concurrently. It is widely accepted that once this concrete is exposed to CO2, calcium hydroxide and C-S-H can be carbonated simultaneously (Groves 1991; Matsusato 1992; Thiery 2007; Castellote 2009). However, some research isolated these hydration products to study the difference in carbonation rate when exposed to CO2 (Groves 1991; Thiery 2007). Observations clarified that the carbonation rate of calcium hydroxide is initially higher than that of C-S-H, but the former’s rate slows down due to the formation of carbonate microcrystals at the surface of the calcium hydroxide crystal. In contrast, the carbonation rate of C-S-H has been shown to be constant (Groves 1991; Thiery 2007). Excessive carbonation of hydrated concrete may result in reduction in the Ca/Si ratio in calcium silicate hydrates. This progressive decalcification eventually leads to the formation of silica gel and calcite, while leaving some residual uncarbonated calcium hydroxide (Slegers 1976; Groves 1991; Richardson 1993; Castellote 2009).

Carbonation products The exothermic reactions release 74, 347, and 184 kJ/mol due to the carbonation of CH, C3S, and β-C2S respectively (Goodbrake 1979a; Moorehead 1986). The products of carbonation are normally identified as calcite although calcium carbonate exists in three different polymorphs (in order of decreasing stability): well crystalline calcite and poorly crystalline aragonite and vaterite (Goto 1995; Lange 1996). Observations reveal that aragonite tends to form as the primary product from the carbonation of β-C2S (Goto 1995), or when the material dries out during carbonation (Goodbrake 1979b). It has been

– 11 – noted that these unstable polymorphs (aragonite and vaterite) eventually recrystallize at room temperature to form the stable calcite polymorph (Sawada 1997).

Other chemical reactions involving C3S, β-C2S, and γ-C2S with CO2 suggest the formation of basic calcium carbonates (BCC: 2CaCO3·Ca(OH)2·1.5H2O) and amorphous basic calcium carbonate (ABCC: Ca3(OH)6x(CO3)3-3x·3yH2O) (Goto 1995). In addition, magnesia-based cements or slag with high MgO content carbonate to produce magnesite,

MgCO3 (Pearce 2002). Another carbonation product is C-S-H gel, which is found intermixed with calcium carbonate (Berger 1972). A microstructural study performed on β-C2S and C3S exposed to 100% CO2 at 1 bar resulted in the formation of more calcium carbonate than that expected from the simple stoichiometric balancing of the reactions (Eqs. 2.5-2.6). Calcium carbonates bound within the C-S-H matrix had not been accounted for

(Goodbrake 1979a). This coexistence of C-S-H and CaCO3 has been introduced in past research as calcium silicate hydrocarbonate (Berger 1972; Ramachandran 1986; Goto 1995). SEM observations discussed the presence of amorphous, non-crystalline calcium carbonates surrounding cement particles in near pure form (Shtepenko 2006; Villain 2006). This form of carbonation product was theorized to have binding properties (Liu 2001). Some studies identified crystalline forms of carbonates near the surface of cement particles (Monkman 2006), while others discovered micro crystals growing in cement pores (Bertos 2004).

Other carbonation applications Carbonation has been investigated in many applications for over 50 years. Hydrated lime, used as an adhesive filler between bricks and stones in construction of buildings, was carbonated to form a durable carbonate matrix (Moorehead 1986). Masonry units were also carbonated to reduce service shrinkage and improve compressive strength (Shideler 1955). Supercritical CO2 exposure improved the performance of glass fiber reinforcement due to the reduction of calcium hydroxide content and pH of cured concrete (Jones Jr. 1997). Such carbonation has also been employed to create a dense cementitious product with polymeric fiber reinforcement (Knopf 2002). Cement and asbestos fiber blocks were cured with a carbon dioxide/steam

– 12 – mixture at elevated temperature (up to 150ºC) and pressure (up to 175 psi or 1206 kPa) (Staley 1950). Concrete pipes with normal-weight granite aggregates were cured employing a combined sequence of initial hydration and subsequent carbonation (Rostami 2011). Pretreatment of cement, asbestos fibers, and silica blocks with dilute

CO2 (20%) at pressures up to 30 psi (or 206 kPa) were carried out to achieve initial set and permit safe handling before autoclaving (Schulze 1967). In addition, a curing regimen incorporated carbon dioxide with combustion emissions into the early stages of steam curing of concrete (Soroushian 1999). Moreover, the carbonation of cement in wood reinforced concrete solved the compatibility problem due to accelerated hardening (Schmidt 1988; Hermawan 2002; Soroushian 2003). Carbonation ensured the feasibility of waste stabilization in concrete (Bertos 2004), whether at supercritical (Venhuis 2003) or vacuumed conditions (Venhuis 2001). Stabilization has been achieved through both physical immobilization and chemical bounding of metals (Lange 1996; Walton 1997). While static carbonation was the dominant method for carbonation curing, an alternate dynamic system was introduced in the carbonation of compact calcium silicate mortars (Young 1974). The results of the pseudo-dynamic system using flush-through could reach similar carbonation degree as the static system. However, the flush-through system didn’t require a pressurized chamber.

Factors influencing carbonation Carbonation was initially studied to offset the deterioration mechanism of weathering carbonation on hydrated concrete. In simulated carbonation reactions, many factors influencing carbonation were addressed including the nature of the reacting material, the physical characteristics of the solid, and the reaction’s exposure conditions

(Lange 1996). The reactivity of a material to CO2 is crucial. If the material is non-reactive to CO2, no carbonation will take place. Also, carbonation is affected by the penetration rate of the gas into the material. Excessive surface free water may hinder the reaction by blocking the pores, while insufficient amounts will prevent the diffusivity of CO2. Porous samples exhibit higher penetration of gas in capillaries and greatly improve carbonation.

Many other factors affect the reactivity such as pressure and CO2 concentration, where respective increases can enhance reaction rate to a certain extent (Moorehead 1986). On

– 13 – the other hand, an increase in temperature or specimen thickness has been noted to decrease the rate of the reaction (Verbeck 1958; Moorehead 1986). The relative humidity during carbonation and the moisture content of the samples are debated as the most crucial factors to maximize carbonation efficiency (Shideler 1955; Verbeck 1958; Kroone 1959; Toennies 1960; Toennies 1963). Optimum carbonation extent has been recorded for relative humidity and moisture content between 50-60%. At very low (11%) and very high (100%) relative humidity, the carbonation degree was limited (Toennies 1960). A favorable relative humidity between 15-35% led to maximum shrinkage reduction (Toennies 1963). In other work, the relative humidity was chosen between 50-55% for the carbonation of concrete masonry units and calcium silicate mortars (Shideler 1955; Young 1974).

QUANTIFICATION OF CO2 UPTAKE

CO2 uptake can be quantified by the mass gain method, mass curve method, thermogravimetry (TG) method, constant temperature pyrolysis method, and chemical analysis by hydrochloric acid (HCl) titration method.

Mass gain method The mass gain method compares the masses of the specimens before and after carbonation curing. Carbonation is an exothermic reaction that usually leads to the release of free water. For this reason, to accurately quantify the mass gain arising from the CO2 uptake, water lost and condensed on the interior of the chamber should be incorporated into the formula, as follows (Monkman 2006):

Mass gain (%) = (2.7)

Eq. 2.7 indicates that the increase in mass, designated by Final mass – Initial mass, is calculated due to CO2 uptake. The mass of water condensed on the wall of the chamber, Mass of water loss, along with the water included in the mass of the samples represent the initial amount of mixing water. By dividing the numerator by the mass of binder

(cement), the CO2 uptake can be calculated based on cement mass instead of concrete mass.

– 14 – Mass curve method The concrete samples are placed in a sealed chamber during carbonation. By placing the entire carbonation setup onto a digital balance, the real-time mass curve can be recorded with respect to time. The resulting mass curve is initially produced by zeroing the chamber after vacuuming, and the increase in mass is due to the precipitation of carbonates. At the end of the process, CO2 gas is released, and the residual mass, M, is measured. A second residual mass, m, is obtained by calibrating the system with CO2- insensitive styrofoam samples of the same volume. The difference between M and m represented the CO2 uptake by concrete. Eq. 2.8 is the basis for the mass curve measurement. M m CO2 uptake (%) = (2.8) Mass of cement

Thermogravimetry (TG) method Thermogravimetry has been employed for the quantification of carbonates in several investigations. The main interest in this approach is to detect the decomposition of carbonates. Free water is typically released at temperatures below 105ºC, while bound water in hydration products can reach up to 400ºC (Johnson 2000). The interpretations of the mass loss data vary, but from a general perspective, the mass loss between 600 °C and

800°C can be regarded as the loss of CO2 from decarbonation of calcite (Matsushita

2000). In the case of decomposition of magnesite, MgCO3, the major mass loss is between 300ºC and 350ºC (Knopf 2002). Carbonates are usually classified according to crystallinity. Poorly crystalline carbonate phases decompose between 200°C and 600°C, and decarbonation of well-formed calcite, vaterite and aragonite crystals occurs at higher temperatures (Goto 1995). Another approach has suggested that carbon dioxide mass loss can be divided into three types (Kroone 1959): Unstable carbon dioxide (from calcite intimately associated with silica) is lost below 500°C, stable carbon dioxide (from calcite sufficiently separated from silica to behave as pure calcite) evolves between 500°C and 700°C, while carbon dioxide associated with alkali metal carbonates is released above 700°C. Other research has associated the mass loss between 325ºC and 600ºC with the decomposition of small, poorly crystalline calcium carbonates, and discussed the

– 15 – possibility of an overlap with combined water (which can decompose at temperatures up to 400ºC) (Bukowski 1979). In order to identify the onset and end temperatures for the mass loss reactions in the TG curves, it has been suggested that the DTG or derivative thermogravimetric curve can be used (Kneller 1994). With such variance in identification of chemical compounds at different temperatures, it is crucial to follow a certain mass loss convention based on temperature ranges. A recent study by Ramachandran indicated that evaporable water and poorly bound water in C-S-H decomposed at temperatures up to 105ºC, water bound with hydration products, C-S-H and C-A-H, was released between 105ºC-420ºC, and calcium hydroxide decomposed between 420C-470C (Ramachandran 2001). As for the carbonation products, decarbonation of poorly crystalline calcium carbonates attributed to aragonite and vaterite occurred between 470C-720C, and decarbonation of well crystalline calcite between 720C-950C (Ramachandran 2001; Li 2003).

Constant temperature pyrolysis method Constant temperature pyrolysis has also been used to identify and quantify the carbonates formed due to the carbonation reaction. Mass loss below 105°C is attributed to evaporable water, between 105°C and 350°C to combined water, and 350°C to 1000°C to the decomposition of carbonates, with the latter representing the CO2 uptake (Goodbrake 1979a). However, considering the before-mentioned findings regarding thermal decompositions, this approach is relatively inaccurate. It is therefore suggested that the temperature range between 350°C and 470°C be representative of the decomposition of calcium hydroxide, and that between 470°C and 950°C be of the decarbonation of calcium carbonates. Accordingly, the CO2 content can be calculated based on the following equation:

CO2 content (%) = (2.9)

Chemical analysis by hydrochloric acid titration method

The CO2 content can be quantified by measuring the gas produced from an acidified sample. The process involves the coulometric titration of the sample in

– 16 – hydrochloric acid (HCl) and capturing the liberated CO2 gas in an absorption bulb (Huffman 1977). Other similar techniques titrate using barium hydroxide solution

(Schollenberger 1958). The volume of liberated CO2 can be measured by heating the material in the presence of potassium sulfate and sulfuric acid (Bessey 1939). The CO2 evolved from the reaction of hydrochloric acid and carbonates can be quantified by means of an inverted and submerged burette (Huffman 1977; Pile 1998). The reaction (Eq. 2.10) produces calcium chloride solution and carbon dioxide gas.

CaCO3(s) + 2HCl(aq)  CaCl2(aq) + CO2(g) + H2O(l) (2.10)

STRENGTH DEVELOPMENT Early strength Carbonation is defined as an accelerated curing mechanism (Young 1974). Early- age carbonation is proved to be beneficial, unlike weathering carbonation. Five-minute carbonation of a cement mortar was found to achieve equivalent compressive strength to a 24-hour hydrated counterpart (Klemm 1972; Young 1974). In other cases, similar mortars were carbonated for 30 minutes and achieved the ASTM 28-day required strength for Portland cement (Wagh 1995). C3S and β-C2S mortars carbonated for 81 minutes had more than 3 times the strength of a comparable mortar hydrated for 24 hours (Young 1974). Unlike the previously mentioned cases of atmospheric and pressurized carbonation, vacuum carbonation allows a higher water-to-binder (w/b) ratio by lowering the amount of free water and enhancing CO2 penetration (Venhuis 2001). This process of vacuumed carbonation found that the strength of a mortar subjected to 38 minutes of carbonation was higher than that of mortars hydrated for 7 and 28 days in air (Hannawayya 1984). Furthermore, 45% of the 28-day compressive strength of vacuum- dewatered lightweight aggregate concrete (LWAC) was achieved after 15 minutes of carbonation (Malinowski 1983).

Late strength Past research discussed the effect of carbonation on subsequent hydration. Early- age carbonation did not hinder subsequent hydration and produced higher compressive strength up to 28 days (Sorochkin 1975). In other cases, 3-day hydrated samples

– 17 – exhibited 50% less strength than carbonated samples hydrated for the same duration (Klemm 1972). Prolonged carbonation durations of mature concrete have shown to improve the ultimate compressive strength (Chang 2003; Jerga 2004; El-Turki 2009). A continuous strength gain is attained in an extensive 52-week exposure of hydrated concrete to 10% concentrated CO2 in 60% relative humidity (Matsusato 1992). Another research of prolonged carbonation of 28-day moist-cured concrete has shown an increase in strength (Chang 2006). Furthermore, air-cured concrete (at 20ºC and 50% relative humidity) was carbonated for 40 days at 65% relative humidity and led to an increase in compressive strength (El-Turki 2009). Microstructural studies, including SEM, suggested that carbonation of hydration products in normal concrete could promote the fast nucleation of calcite crystals, forming a continuous phase in which silica gel is left as an impurity, leading to an increase in strength. On the other hand, some other research discussed that the calcite formed in carbonated porous concrete is possibly different from normal concrete, wherein no continuous phase of calcite is formed. The strength consequently is decreased, as silica gel exhibits poor binding properties and calcite crystals behave similarly to unbound sand grains (Sauman 1971; Thaulow 2001).

DURABILITY Durability of carbonated concrete was addressed in several matters. Long-term atmospheric carbonation deposits calcium carbonates in the capillaries, and lowers sorptivity due to reduction in the porosity (Dias 2000). Additional research observed that prolonged exposure of hydrated concrete to CO2 reduces porosity (Song 2007). Furthermore, the tendency to develop undesired efflorescence is reduced due to carbonation (Weber 1941). Steam-cured concrete blocks have been carbonated to improve their resistance to atmospheric carbonation shrinkage (Toennies 1963). Similarly, carbonation of precast concrete offers a preshrinking curing phenomenon that could reduce service shrinkage (Leber 1956; Polisner 1956; Toennies 1960; Freedman 1969). The effect of carbonation on other durability aspects, including freeze-thaw resistance, sulfate attack, surface resistivity, has been only recently studied. It has been found that carbonation increases the resistance to the previously mentioned durability issues in comparison to steam-cured and hydrated concrete pipes (Rostami 2011).

– 18 – ENERGY CONSUMPTION Carbonation is proposed as an alternative to steam curing in order to minimize the cost of the curing process. Previous research on steam curing showed that atmospheric pressure steam curing consumes 0.59 GJ of energy and autoclaving consumes 0.71 GJ of energy per 1 m3 of concrete in block form (Kawai 2005). On the other hand, carbonation requires 0.89 GJ/ton of CO2 to recover and compress the gas for use (Halmann 1999). Assuming masonry concrete has a density of 2400 kg/m3, a cement content of 10% by total mass, and a CO2 uptake capacity of 10% by cement mass (Shao 2006), 24 kg of CO2 gas is required for one cubic meter of masonry concrete to perform carbonation curing.

The energy consumption in recovering that amount of CO2 is approximately about 0.02 3 GJ/m . If CO2 uptake capacity of concrete can be enhanced to their theoretical limit, 50% of cement mass (Steinour 1956), the energy consumption will be increased in recovery 3 process to 0.1 GJ per one m concrete. Even if the energy for CO2 transport is considered, the total energy required by the carbonation process can still be significantly lower than atmospheric steam curing.

PORTLAND LIMESTONE CEMENT (PLC) IN CONCRETE Currently, PLC is gaining more widespread use throughout the concrete industry as it offers technical, economical, and environmental benefits (Detwiler 1996; Tsivilis 2002; Bonavetti 2003). The addition of limestone through the use of PLC may in some cases alter the hydration of concrete. The presence of calcium carbonate has shown to accelerate the hydration of C3S and C2S (Pera 1999). It has also been shown that as a result of calcium carbonate consumption, the hydrated carbosilicates and carboaluminates might be produced along with conventional hydration products (Pera 1999). Moreover, limestone addition has led to the increase of early strength, the reduction of water demand, the control of bleeding in concrete with low cement content, and the improvements of concrete workability (Moir 1997; Tsivilis 1998). Among the economic benefits are the possibility to obtain cement with a strength development similar to that of Portland cement at low production and investment costs per ton of cement (Baron 1987).

– 19 – As limestone is softer than clinker, it will achieve a finer particle size when interground, producing an improved particle size distribution and improving particle packing (Tsivilis 2002). Also, the fine limestone particles act as nucleation sites, thus increasing the rate of hydration of the calcium silicates at early ages, leading to an improved distribution of hydrates (Soroka 1977; Bonavetti 2003). It has been found that the early strength was improved with 15-30% limestone addition, but was slightly decreased with 30-50% addition. However, the 28-day compressive strength was reduced by 8-19% with any limestone addition (Pera 1999; Bonavetti 2003). The carbonation of PLC is greatly affected by the amount of limestone added to the cement clinker. With over 19% limestone addition, it was reported that PLC concrete was more vulnerable to weathering carbonation (Parrott 1996). CaCO3 present as limestone reacts chemically with aluminate phases to form carboaluminates (Matschei

2007). Ettringite also forms due to the presence of CaCO3, while C-S-H incorporates significant amounts of calcium carbonates into its structure forming calcium silicocarbonate hydrates (Ramachandran 1986). However, no work has been done on the carbonation of PLC concrete within a 24-hour timeframe.

REFERENCES Baron, J., and Douvre, C. (1987). "Technical and economical aspects of the use of limestone filler additions in cement." World Cement 18(3): 100-104.

Berger, R. L., Young, J. F., and Leung, K. (1972). "Accelerated curing of cementitious systems by carbon dioxide - Hydraulic calcium silicates and aluminates - Part II." Cement and Concrete Research 2: 647-652.

Bertolini, L., Elsener, B., Pedeferri, P., and Polder, R. B. (2006). "Carbonation-Induced Corrosion." In Corrosion of Steel in Concrete: 79-90.

Bertos, M. F., Simons, S. J. R., and Hills, C. D. (2004). "A review of accelerated carbonation technology in the treatment of cement-based materials and sequestration of CO2." Journal of Hazardous Materials B112: 193-205.

Bessey, G. E. (1939). "Gas-volume Method for Determination of Carbonate in Calcareous Materials." Society of Chemical Industry -- Journal -- Transactions and Communications 58(5): 178-180.

– 20 – Bonavetti, V., Donza, H., Menendez, G., Cabrera, O., Irassar, E. F. (2003). "Limestone filler cement in low w/c concrete: A rational use of energy." Cement and Concrete Research 33(6): 865-871.

Bukowski, J. M., and Berger, R. L. (1979). "Reactivity and strength development of CO2 activated non-hydraulic calcium silicates." Cement and Concrete Research 9(1): 57-68.

Castellote, M., and Fernandez, L. (2009). "Chemical changes and phase analysis of OPC pastes carbonated at different CO2 concentrations." Materials and Structures 42(4): 515-525.

Chang, C., and Chen, J. (2006). "The experimental investigation of concrete carbonation depth." Cement and Concrete Research 36: 1760-1767.

Chang, C., and Yeih, W. C. (2003). "Mechanical properties of carbonated concrete." Journal of the Chinese Institute of Engineers 26(4): 513-522.

Detwiler, R. J., Tennis P. D. (1996). "The use of limestone in Portland cement: a state-of- the-art review." Portland Cement Association.

Dias, W. P. S. (2000). "Reduction of concrete sorptivity with age through carbonation." Cement and Concrete Research 30(8): 1255-1261.

El-Turki, A. B. R. (2009). The effect of carbonation on the mechanical and microstructural properties of lime and cement. Proceedings of the 11th International Conference on Non-conventional Materials and Technologies, Bath, UK, NOCMAT.

Freedman, S. (1969). "Carbonation Treatment of Concrete Masonry Units." Modern Concrete 33(5): 33-36.

Goodbrake, C. J., Young, J. F., and Berger, R. L. (1979a). "Reaction of beta-dicalcium silicate and tricalcium silicate with carbon dioxide and water." Journal of the American Ceramic Society 62(3-4): 168-171.

Goodbrake, C. J., Young, J. F., and Berger, R. L. (1979b). "Reaction of hydraulic calcium silicates with carbon dioxide and water." Journal of the American Ceramic Society 62(9-10): 488-491.

Goto, S., Suenaga, K., and Kado, T. (1995). "Calcium silicate carbonation products." Journal of the American Ceramic Society 78(11): 2867-2872.

Grounds, T., Midgley, H. G., and Novell, D. V. (1998). "Carbonation of Ettringite by Atmospheric Carbon Dioxide." Thermochimica Acta 135: 347-352.

– 21 – Groves, G. W., and Brough, A. (1991). "Progressive Changes in the Structure of Hardened C3S Cement Pastes Due to Carbonation." Journal of the American Ceramic Society 74(11): 2891-2896.

Halmann, M. (1999). "Greenhouse gas carbon dioxide mitigation: Science and Technology." Boca Raton, Fla, Lewis Publishers.

Hannawayya, F. (1984). "Microstructural Study of Accelerated Vacuum Curing of Cement Mortar with Carbon Dioxide - Part 1." World Cement 15(9): 326-334.

Hermawan, D., Hata, T., Kawai, S., Nagadomi, W., and Kuroki, Y. (2002). "Effect of Carbon Dioxide–Air Concentration in the Rapid Curing Process on the Properties of Cement bonded Particleboard." Journal of Wood Science 48(3): 179-184.

Huffman, E. W. D. (1977). "Performance of a new automatic carbon dioxide coulometer." Microchemical Journal 22: 567-573.

Jerga, J. (2004). "Physico-mechanical properties of carbonated concrete." Construction and Building Materials 18(9): 645-652.

Johnson, D. (2000). "Accelerated carbonation of waste calcium silicate materials." Society of Chemical Industry 10.

Johnson, D. C., MacLeod, C. L., and Hills, C. D. (2003). Accelerated Carbonation of Stainless Steel Slag. In Proceedings of the 5th International Conference on the Environmental and Technical Implications of Construction with Alternative Materials, Spain, ISCOWA.

Jones Jr., R. H. (1997). Cement Treated with High-pressure CO2. United States Patent. United States of America. 5: 562.

Kawai, K., and Sugiyama, T. (2005). "Inventory data and case studies for environmental performance evaluation of concrete structures." Journal of Advanced Concrete Technology: 435-456.

Klemm, W. A., and Berger, R. L. (1972). "Accelerated curing of cementitious systems by carbon dioxide - Part I - Portland Cement." Cement and Concrete Research 2: 567-576.

Kneller, W. A., Gupta, J., Borkowski, M. L., and Dollimore, D. (1994). "Determination of Original Free Lime Content of Weathered Iron and Steel Slags by Thermogravimetric Analysis." Transportation Research Record(1434): 17-22.

Knopf, F. C., and Dooley, K. M. (2002). Pressure-assisted Molding and Carbonation of Cementitious Materials. United States Patent. United States of America. 6: 174.

– 22 – Kroone, B., and Blakey, F. A. (1959). "Reaction Between Carbon Dioxide Gas and Mortar." American Concrete Institute Journal 31(6): 497-510.

Lange, L. C., Hills, C. D., and Poole, A. B. (1996). "The effect of accelerated carbonation on the properties of cement-solidified waste forms." Waste Management 16(8): 757-763.

Leber, I., Blakey, F. A. (1956). "Some Effects of Carbon Dioxide on Mortars and Concrete." American Concrete Institute Journal 28(3): 295-308.

Li, D., Shi, H., Deng, J., and Xu, Y. (2003). "Study on the thermal decomposition kinetics of nano-sized calcium carbonates." Journal of Zhejiang University 4(3): 363-368.

Liu, L., Ha, J., Hashida, T., and Teramura, S. (2001). "Development of a CO2 Solidification Method for Recycling Autoclaved Lightweight Concrete Waste." Journal of Materials Science Letters 20(19): 1791-1794.

Malinowski, R., and Rodhe, M. (1983). "Vacuum Carbonation of Lightweight Aggregate Concrete." Nordic Concrete Research(2): 166-182.

Matschei, T., Lothenbach, B., and Glasser, F. P. (2007). "The role of calcium carbonate in cement hydration." Cement and Concrete Research 37(4): 551-558.

Matsusato, H., and Ogawa, K. (1992). Studies on the carbonation of hydrated cement and its effect on microstructure and strength. Ninth International Congress on the Chemistry of Cement. India, National Council for Cement and Building Materials.

Matsushita, F., Aono, Y., and Shibata, S. (2000). "Carbonation degree of autoclaved aerated concrete." Cement and Concrete Research 30: 1741-1745.

Mehta, P. K., and Monteiro, J. P. (2006). Concrete Microstructure, Properties and Material. New York, McGraw Hill.

Moir, G., and Kelham, S. (1997). "Developments in manufacture and use of Portland limestone cement." American Concrete Institute Journal 172: 797-819.

Monkman, S., and Shao, Y. (2006). "Assessing the carbonation behavior of cementitious materials." Journal of Materials in Civil Engineering November/December: 768- 776.

Moorehead, D. R. (1986). "Cementation by the carbonation of hydrated lime." Cement and Concrete Research 16: 700-708.

Neville, A. M. (1996). Properties of Concrete. New York, Wiley.

– 23 –

Nishikawa, T., Suzuki, K., Ito, S., Sato, K., and Takebe, T. (1992). "Decomposition of Synthesized Ettringite by Carbonation." Cement and Concrete Research 22(1): 6- 14.

Papadakis, V. G., Vayenas, C. G., and Fardis, M. N. (1991). "Fundamental Modeling and Experimental Investigation of Concrete Carbonation." American Concrete Institute Materials Journal 88(4): 363-373.

Parrott, L. J. (1996). "Some effects of cement and curing upon carbonation and reinforcement corrosion in concrete." Materials and Structures 29: 164-173.

Pearce, F. (2002). "Green Foundations: It's Time to Make the Concrete Jungle Emulate the Real Thing." New Scientist 175(2351): 38-40.

Pera, J., Husson, S., Guilhot, B. (1999). "Influence of finely ground limestone on cement hydration." Cement and Concrete Composites 21(2): 99-105.

Pile, D. L., Benjamin, A. S., Lackner, K. S., Wendt, C. H., and Butt, D. P. (1998). "A Precise Method for Determining the CO2 Content of Carbonate Materials." Journal of Chemical Education 75(12): 1610-1614.

Polisner, L. (1956). "Block Curing with CO2." Concrete 64(12): 23-24.

Powers, T. C. (1962). "Hypothesis on Carbonation Shrinkage." Portland Cement Association Research and Development Laboratories 4(2): 40-50.

Ramachandran, V. S., and Beaudoin, J. J. (2001). Handbook of Analytical Techniques in Concrete Science and Technology, William Andre Publishing/Noyes

Ramachandran, V. S., and Zhang, C. (1986). "Influence of CaCO3 on hydration and microstructural characteristics of tricalcium silicate." Cemento 83(3): 129-152.

Richardson, I. G., and Grooves, G. W. (1993). "The incorporation of minor and trace elements into calcium silicate hydrate gel in hardened cement pastes." Cement and Concrete Research 23: 131.

Richardson, M. G. (1988). "Carbonation of Reinforced Concrete: Its Causes and Management." CITIS Ltd.: 205.

Rostami, V., Shao, Y., and Boyd, A. J. (2011). "Durability of concrete pipes subjected to combined steam and carbonation curing." Construction and Building Materials 25: 3345-3355.

– 24 – Sauman, Z. (1971). "Carbonization of porous concrete and its main binding components." Cement and Concrete Research 1(6): 645-662.

Sawada, K. (1997). "The mechanism of crystallization and transformation of calcium carbonates." Pure and Appl. Chem. 69(5): 921-928.

Schmidt, E. (1988). Process for Accelerating Hardening of Cement with Fibre-reinforced and Cement-bound Plates. 4,746,481.

Schollenberger, C. J., and Whittaker, C. W. (1958). "Improved Determination of Carbonates in Soils." Soil Science 85: 10-13.

Schulze, H. C. (1967). Method of Curing Articles Manufactured from Cement and Asbestos. United States Patent. United States of America. 3,356,779.

Shao, Y., Mirza, M. S., and Wu, X. (2006). "CO2 sequestration using calcium-silicate concrete." Canadian Journal of Civil Engineering 33: 776-784.

Shideler, J. J. (1955). "Investigation of the moisture-volume stability of concrete masonry units." Portland Cement Association Research and Development Laboratories Bulletin D3.

Shtepenko, O., Hills, C., and Brough, A. (2006). "The effect of carbon dioxide on beta- dicalcium silicate and Portland cement." Chemical Engineering Journal 118: 107- 118.

Slegers, P. A., and Rouxhet P. G. (1976). "Carbonation of the hydration products of tricalcium silicate." Cement and Concrete Research 6(3): 381-388.

Song, H. W., and Kwon, S. J. (2007). "Permeability characteristics of carbonated concrete considering capillary pore structure." Cement and Concrete Research 37(6): 909-915.

Sorochkin, M. A., Shchurov, A. F., and Safonov, I. A. (1975). "Study of The Possibility of Using Carbon Dioxide for Accelerating the Hardening of Products Made From Portland Cement." Journal of Applied Chemistry of the USSR 48(6): 1211-1217.

Soroka, I., and Setter, N. (1977). "The effect of fillers on strength of cement mortars." Cement and Concrete Research 7(4): 449-456.

Soroushian, P., and Hsu, J. W. (1999). Accelerated Curing of Cement-based Materials. United States Patent. United States of America. 5,935,317.

Soroushian, P., Won, J. P., Chowdhury, H., and Nossoni, A. (2003). "Development of Accelerated Processing Techniques for Cement-bonded Wood Particleboard." Cement and Concrete Composites 25(7): 721-727.

– 25 –

Staley, R. W. (1950). Method of Curing Cement-Type Cold Molding Compositions. United States Patent. United States of American. 2,496,895.

Steinour, H. (1956). "Some Effects of Carbon Dioxide on Mortar and Concrete." American Concrete Institute Journal 30: 905-907.

Tans, P. (2008). "Trends in Atmospheric Carbon Dioxide."

Thaulow, N. L. R. J., Wagner, K., and Sahu, S. (2001). "Effect of calcium hydroxide content on the form, extent, and significance of carbonation Materials Science of Concrete Special Volume: Calcium Hydroxide in Concrete." USA.

Thiery, M., and Villain, G. (2007). "Investigation of the carbonation front shape on cementitious materials: Effects of the chemical kinetics." Cement and Concrete Research 37(7): 1047-1058.

Toennies, H. T. (1960). "Artificial carbonation of concrete masonry untis." American Concrete Institute Journal(56-42): 737-755.

Toennies, H. T., and Shideler, J. J. (1963). "Plant drying and carbonation of concrete block - NCMA-PCA cooperative program." American Concrete Institute Journal 60(33): 617-632.

Tsivilis, S., Chaniotakis, E., Kakali, G., and Batis, G. (2002). "An analysis of the properties of Portland limestone cements and concretes." Cement and Concrete Composites 24(3): 371-378.

Tsivilis, S., Kakali, G.,Chaniotakis, E., and Souvaridou, A. (1998). "A study of the hydration of Portland limestone cement by means of TG." Journal of Thermal Analysis 52: 863-870.

Venhuis, M. A., and Reardon, E. J. (2001). "Vacuum method for carbonation of cementitious waste forms." Environmental Science and Technology 35(20): 4120- 4125.

Venhuis, M. A., and Reardon, E. J. (2003). "Carbonation of Cementitious Wasteforms Under Supercritical and High Pressure Subcritical Conditions." Environmental Technology 24(7): 877-887.

Verbeck, G. J. (1958). "Carbonation of Hydrated Portland Cement." American Society for Testing Materials Bulletin 87(205): 16-36.

Villain, G., and Platret, G. (2006). "Two experimental methods to determine carbonation profiles in concrete." American Concrete Institute Materials Journal 103(4): 265- 271.

– 26 –

Wagh, A. S., Singh, D., and Knox, L. J. (1995). "Lab studying greenhouse effect on concrete setting." Concrete International 17(4): 41.

Walton, J. C., Bin-Shafique, S., Smith, R. W., Gutierrez, N., and Tarquin, A. (1997). "Role of Carbonation in Transient Leaching of Cementitious Wasteforms." Environmental Science and Technology 31(8): 2345-2349.

Weber, J., and Matthei, R. (1941). "Find carbon dioxide gas under pressure an efficient curign agent for cast stone." Concrete (Chicago) 41(7): 31-34.

Young, J. F., Berger, R. L., Breese, J. (1974). "Accelerated Curing of Compacted Calcium Silicate Mortars on Exposure to CO2." Journal of the American Ceramic Society 57(9): 394-397.

– 27 – Chapter 3

—

STATIC CARBONATION OF LIGHTWEIGHT OPC CONCRETE MASONRY

UNITS

PREFACE Weathering carbonation has been proven to be detrimental to concrete. On the other hand, early-age carbonation has shown to present economical, technical, and environmental benefits. However, the carbonation reaction of freshly cast concrete is hindered by surface saturation due to the presence of free water. Initial air curing served as a presetting procedure, which enhanced the efficiency of the reaction due to release of surface free water and production of capillaries. The research in this chapter aims to produce the optimum curing conditions to replace steam curing for concrete masonry units. The research was initiated by extending initial air curing and carbonation times to 14 and 4 days, respectively, to maximize the degree of carbonation. Three different types of aggregates were investigated: granite with riversand, expanded shale with riversand, and expanded slag. While the first two aggregates required a combination to create a well-graded product, the latter was used as received. The sieve analysis of each aggregate is presented in Fig. 3.1 and 3.2. In order to choose the most efficient aggregate for carbonation, the proposed initial carbonation scheme was applied, and the CO2 uptake and compressive strength were measured. The results illustrated in Table 3.1 show that carbonation of expanded slag came up with the highest uptake and compressive strength. Due to 14 days of air curing, most of the water was lost and the hydrated sample could not produce comparable compressive strength. With an industrial timeframe of 24 hours, the duration of each process was shortened to meet the time restrictions. The proposed curing scheme has three stages: initial air curing (up to 18 hours), short term carbonation (up to 96 hours), and subsequent hydration. Each step plays a unique role to enhance the early and late performance of concrete. The initial air curing

– 28 – reduces free capillary water content at the surface by maintaining equilibrium with the drier atmosphere. The second step, carbonation, modifies the concrete properties and microstructure by binding CO2 to the cement matrix. The third step, which is subsequent hydration, follows carbonation since the short-term carbonation curing leaves the opportunity for subsequent hydration to obtain enhanced performance of the concrete comparable to its counterpart cured by normal hydration curing. The proposed curing regime is characterized in terms of carbonation degree and compressive strength. To ensure the validity of the curing scheme, 20-cm lightweight concrete masonry units are cast and cured. The CO2 uptake and compressive strength are also measured, and the carbonation products are characterized by XRD analysis. The chapter presented here has been accepted for publishing with minor change in the American Concrete Institute (ACI) Materials Journal.

– 29 – INTRODUCTION Concrete masonry units (CMU) have been widely used in building construction as load bearing and non-load bearing walls. The North American market for concrete blocks and bricks is projected to increase to 4.3 billion units per year in 2014 (Freedonia 2010). In comparison to cast-in-place concrete, masonry block structures not only exhibit superior performance due to precast quality, but also represent a low environmental impact construction system. With reference to one square meter of solid concrete wall, a masonry partitioning wall using 20-cm CMU requires 48% less material due to internal cavities, leading to 65% less cement and 65% less CO2 emission. CMUs produced in North America are typically steam cured. While steam curing accelerates strength gain and shortens the production cycles, the process itself is energy intensive. It is estimated, for one cubic meter of concrete in block form, atmospheric pressure steam curing consumes 0.59 GJ and autoclave curing consumes 0.71 GJ (Kawai 2005). The alternative curing method for CMU production is carbonation curing which uses high purity carbon dioxide (99.5% of CO2) or low purity flue gas (14% of CO2) for accelerated hydration and durability improvement. Carbonation treatment of concrete masonry units was studied in the 60’s. The purpose of the investigation at that time was to “preshrink” CMUs in order to reduce weathering carbonation induced cracking in masonry structures (Toennies 1960; Toennies 1963). CMUs were first steam cured at atmospheric pressure for 5 to 18 hours and then treated with hot flue gas for 3 to 36 hours. It was found that service shrinkage of CMUs exposed to weathering carbonation could be reduced by 43% with hot flue gas treatment for 24 hours at a temperature of 93oC and a relative humidity of 25% (Toennies 1963). Recent work on the carbonation curing of lightweight concrete products showed that pre-conditioning in a dry environment led to a higher carbonation degree but less hydration degree than pre- conditioning in a moist environment (Shi 2008). Carbonation curing can also be used to accelerate hydration, a process similar to steam curing. The strength of cement paste cured for 5 minutes by high purity carbon dioxide (99.5%) could reach a strength equivalent to normal curing for 24 hours (Young 1974). However, carbonation curing has never been adopted in large-scale production. This was possibly because flue gas

– 30 – carbonation was not effective in hydration acceleration and pure gas carbonation was expensive. The latter situation may change in the near future. Large quantities of high purity, low cost carbon dioxide could soon be available as regulations requiring reductions in CO2 emissions are developed. In this case, pure gas carbonation can simultaneously accelerate strength development, stabilize the dimension, and enhance the durability. By reducing the hydroxyl ion and precipitating CaCO3 on the surface layer, carbonation curing could improve the concrete resistance to sulfate attack, freeze-thaw cycling, and acid attack (Rostami 2011). Since carbonation is a CO2 uptake process (Shao

2006), recovered cement kiln CO2 can be recycled into concrete products and contribute to carbon emission reduction. The purpose of this research is to develop a carbonation curing process that can be used to replace steam curing for CMU production. High purity CO2 (99.5%) will be used to simulate the recovered cement kiln flue gas carbon dioxide. The goal is to shorten the carbonation duration to 2 to 4 hours with the help of initial curing ranging from 0 to 18 hours. The effect of initial curing on degree of carbonation reaction is evaluated to promote maximum possible CO2 uptake in CMU. The CO2 uptake is estimated using mass gain, mass curve, and thermal analysis. Early age and 28-day performance of carbonated lightweight CMU are examined in terms of CO2 uptake and strength gain.

EXPERIMENTAL INVESTIGATIONS Concrete sample preparation The carbonation curing parameters will be studied using rectangular concrete slab samples of 38 mm thickness to simulate the typical web or face shell of a hollow concrete masonry unit (CMU). The optimized process parameters are then applied to carbonation curing of 20-cm CMU. The samples are prepared according to the commercial CMU mix design with lightweight expanded slag aggregates. Fig. 3.3 shows the sieve analysis of slag aggregates based on particle size distribution. The as-received aggregates are nearly saturated with a water content of 5% and a maximum absorption capacity of 9%. Table 3.2 summarizes the mixture proportion of both slab and CMU samples. By mass, it includes a water-to-cement ratio of 0.4, an aggregate-to-cement ratio of 6.23, and cement content of 13% of the total mass. Each rectangular slab sample weighs approximately 680

– 31 – grams and each CMU block weighs 15 kg with a density of 1839 kg/m3 each. For rectangular slabs, raw materials were mixed in a pan mixer, cast into 127 x 76 x 38 mm mold and compact formed using a vibrating hammer to simulate the industry production of CMU. Because of the dry mix, concrete was demolded right after casting for initial curing. For 20-cm CMU blocks, materials were mixed in a mechanical drum mixer and compact formed by a manual block machine. They were typical 20-cm CMU blocks with the thickness of the web or face shell ranging from 25 to 33 mm. The density was calculated based on ASTM C140 (ASTM 2004b).

Curing procedures Curing procedures of different batches using rectangular slab samples are summarized in Table 3.3. For comparison, Batches 1-4 are steam cured and Batches 5-15 are carbonation cured. One set of slab batches were also prepared for normal hydration in sealed plastic bags to serve as controls. Steam curing took place in a steam cooker for a period of 4 hours with maximum temperature of 80C and relative humidity of 95%. Initial curing of 0, 4, 6, and 8 hours at 22±1ºC and relative humidity of 80% was applied prior to steam. The carbonation curing setup is shown in Fig. 3.4. Initial curing was performed on fresh concrete for 0, 4, 6, 8, 18, 336 hours, respectively, in an environmental chamber at a relative humidity of 50% and a temperature of 25C. The purpose of initial curing was to reduce the free water on the surface and allow diffusion of carbon dioxide. Initial curing of 0 hour was actually immediate carbonation of fresh concrete and served as a reference. Initial curing of 18 hours was to simulate an overnight curing and was likely the longest preset that can be accepted by commercial production. Initial curing of 336 hours (14 days) and carbonation for 96 hours (4 days) were designed to repeat Shideler’s tests in 1955 to understand the condition for high carbonation degree (Shideler 1955). Concrete slab samples after initial curing were placed in a sealed chamber in Fig. 3.4, which was then vacuumed to about 0.7 bars below atmospheric and filled with carbon dioxide gas to a pressure of 1 bar. The chamber was placed on a digital balance to obtain the mass curve of concrete during carbonation. The carbonation duration varied from 2 to 4 hours. A period of 96 hours was also investigated to study the effect of extreme exposure time. The effect of initial curing and carbonation curing were

– 32 – evaluated based on water loss, CO2 uptake, and compressive strength. To compensate for the water loss during initial curing and carbonation curing, water spray was applied to Batches 6, 8 and 12 to restore the original water content and examine its effect on subsequent hydration after carbonation. The temperature, relative humidity, pressure, samples’ initial and final mass, and mass of water condensed on the wall of the chamber were recorded. The best combination of initial curing and carbonation curing from slab tests was selected for 20-cm CMU production. Control concrete as reference to carbonation underwent the same initial curing of 0, 4, 6, 8, and 18 hours in an environmental chamber at a 50% RH and a 25C for each hydration control batch.

Internal relative humidity measurement Initial air curing was employed to remove the surface free water and produce capillaries to facilitate the penetration of carbon dioxide gas. To monitor distribution of internal relative humidity of concrete during the period of initial curing, a Sensiron EK- H4 relative humidity meter was employed. Three humidity probes were inserted in a slab sample respectively at 5 mm from top (T), 5 mm from bottom (B), and in the middle (M). Fig. 3.5 presents a schematic of the slab samples with 3 probes inserted at each depth.

CO2 uptake estimation In order to measure the degree of carbonation, three methods were utilized for the estimation of CO2 uptake: mass gain, mass curve, and thermal decomposition analysis.

Mass gain method estimates CO2 uptake in concrete by comparing the mass of samples before and after carbonation (Eq. 3.1). Carbonation-induced water loss was collected by absorbent paper and added to the final mass. By treating the system as a closed system, it was imperative to include the evaporated water, which was initially inside the samples prior to carbonation.

CO2 uptake (%) = (3.1)

Mass curve method was executed by placing the carbonation setup on a digital balance, which was zeroed after vacuuming the chamber. A mass curve was recorded as mass versus time until the end of the process at which CO2 was released and the residual

– 33 – mass, M, was measured. The system was calibrated by repeating the tests using CO2- insensitive styrofoam samples of the same volume to obtain second residual mass, m. The difference between M and m represented the CO2 uptake by concrete (Eq. 3.2). Data collected by mass gain and mass curve methods are two simultaneous measurements from the same process and therefore should be comparable. They are independent from any CO2 content existing before carbonation.

CO2 uptake (%) = (3.2)

Thermal decomposition analysis was also performed to estimate the amount of carbonates in concrete. Instead of using a classical thermogravimetry device, which allows only a few micrograms of powder, a furnace of maximum temperature of 1100ºC was employed to test large concrete samples with mass range of 35-70 g for each. Separation of paste from concrete was thus avoided. Concrete samples were heated up to 105C, 470C, and 950C to quantitatively measure the evaporable water, bound water in hydration products, and carbon dioxide in carbonates respectively (Goodbrake 1979b; o Johnson 2000). The mass loss between 470 C and 950ºC represents CO2 content and includes pre-existing and carbonation-induced carbonates (Eq. 3.3):

CO2 content (%) = (3.3)

Performance evaluation Each slab batch made for carbonation was compared to two control batches: normal hydration control and steam control. Both controls underwent the same initial curing process after which samples of normal hydration control were placed in a sealed plastic bag and samples of steam control were steam-cured followed by sealed hydration to the test age. The compressive strength was measured at 1 and 28 days after casting for all three curing schemes following ASTM standards C140 (ASTM 2004b). Three samples per batch were tested and averaged. For 20-cm CMU, compressive strength was compared between carbonated and hydrated concrete at 1 and 28 days. XRD analysis of CMU was performed. Powder samples were obtained by drilling the surface of fractured CMU samples to a depth not more than 3 mm and then sieved to pass a 125 m sieve. The powder was subjected to X-Ray Diffraction (XRD) analysis by

– 34 – means of a Philips PW1710 Powder Diffractometer (Cu, K radiation, X'celerator proportional detector, scan interval 10-100, 0.02, and 0.5 seconds per step) to identify carbonation and hydration products.

In order to ensure permanent storage of CO2 in concrete and to realize the change in alkalinity of carbonated concrete, phenolphthalein was sprayed on a freshly cut surface of concrete. A purple color indicated basic regions, while an acidic region remained colorless. The solution was made with 5 g phenolphthalein powder, 100 ml of water, and 100 ml of acetone solution. Samples’ pH was measured using a pH meter (Extech PH110 model), which has a flat surface electrode at its tip. An absorptive paper of 10x10 mm was placed on the concrete surfaces and distilled water of 100µl was dropped on it for 15 minutes. An additional 100µl of distilled water was added if the water was absorbed. To ensure an average, the pH readings of three random spots on the cut surface were recorded.

EXPERIMENTAL RESULTS AND DISCUSSION Effect of initial curing on internal relative humidity Initial curing in a controlled environment (25ºC and 50% RH) was introduced in the carbonation process to justify the water content in concrete and its related internal relative humidity. To explore the possibility of application of carbonation curing to CMU production, the entire curing process cannot exceed 24 hours in comparison to current steam curing practice. Therefore, the process window of initial curing is limited to 0, 4, 6, 8, and 18 hours at 25ºC and 50% RH. Shideler’s experiment was repeated with initial curing of 14 days at the same environmental conditions to examine the maximum possible reaction efficiency (Shideler 1955). The water loss curve due to initial curing up to 14 days is plotted in Fig. 3.6. Percent water loss is a ratio of mass loss during initial curing in a specified time over total initial water mass. The total initial water mass is the sum of the mixing water (water-to-cement ratio of 0.4) and the water in wet expanded slag aggregates (5% of the total slag mass). The mass loss during initial curing was recorded by a digital balance over a period of 14 days at 25ºC and 50% RH. It was apparent that water loss was proportional to the duration of initial curing within 24 hours. The most significant loss occurred in first 4 hours, reaching 32%. Water loss due to 6 and

– 35 – 8 hours initial curing was basically no different from 4 hours. However, curing of 18 hours increases the water loss to 51%, which was very close to that by 24 hours. Eighteen hours seemed to be the maximum possible time for initial curing and could be executed through an overnight shift. At 14 days, the water loss reached 81%. The loss of water reduces the water content in concrete, and is expected to reduce the internal relative humidity for better carbonation efficiency. Fig. 3.7 shows the corresponding relative humidity measurement up to 14 days at the same initial curing conditions. Although RH loss was more rapid on the surface than at the bottom, they were proportional and eventually reached equilibrium with the environment of 50% RH in about 14 days. It was interesting to notice that at 24 hours, the RH was 85% near the top surface, 87% at the center and 93% at the bottom. They were much higher than the environmental RH of 50%, which was considered as an ideal condition for carbonation. It is, therefore, conclusive that within 24 hours initial curing, it is not possible to achieve optimal relative humidity of 50%. Comparison of Fig. 3.7 with Fig. 3.6 implies that water loss of 53% at 24 hours does not considerably decrease the relative humidity. Moisture movement from inside out occurs during initial curing to reach equilibrium.

Effect of initial curing on carbonation reaction

Degree of carbonation is characterized by CO2 uptake. Three methods are used to quantify CO2 uptake in carbonated concrete. Fig. 3.8 shows water loss collected from carbonation process and the CO2 uptake by mass gain method (Eq. 3.1). Two groups of data are presented in Fig. 3.8. First group includes the first 5 batches with constant carbonation time of 4 hours and varied initial curing of 0, 4, 6, 8 and 18 hours to study the effect of initial curing. The second group involves the last three batches with constant initial curing of 18 hours and varied carbonation duration of 2, 4, and 96 hours to investigate the effect of carbonation time. Percent water loss due to exothermic carbonation curing is defined as a ratio of water collected in chamber after carbonation over total initial water mass. In the first group of 4-hour carbonation, immediate carbonation with no initial curing resulted in a CO2 uptake of 7.5% with a carbonation water loss of 2.2%. It was indicative of a low degree of reaction. After 4, 6, and 8 hour initial air curing, water loss

– 36 – due to initial curing was of close value of 32-33%. Nevertheless, their CO2 uptake was different at 21.3, 22.8, and 23.5% with carbonation water loss at 8.2, 9.6, and 10.2%, respectively. Obviously, initial curing reduced free water, making room for gas diffusion and calcium carbonate precipitation. Initial curing of 18 hours removed 51% free water and promoted CO2 uptake to 24.2% with a carbonation water loss of 5.9%. Prolonged initial curing is not directly beneficial to reaction efficiency. In the second group, initial curing was fixed at 18 hours. When 2-hour carbonation was compared with 4-hour carbonation, longer carbonation evaporated more free water and promoted higher CO2 uptake. In prolonged carbonation, CO2 uptake by 96 hours was increased by 44% in comparison to that by 4 hours. However, water evaporation was reduced possibly due to the re-absorption of water during the 4-day process. A CO2 uptake of 34.8% in 4-day carbonation represented a degree of carbonation of nearly 70%, if full carbonation is considered as 100% (Steinour 1956).

Obviously, longer carbonation time could promote CO2 uptake and enhance the carbon storage capacity of concrete.

To verify the CO2 uptake by mass gain method (Eq. 3.1), mass curves were obtained. Five batches were compared in Fig. 3.9 with initial curing of 0, 4, 6, 8 and 18 hours. During the 4-hour carbonation process, 60-70% of the reaction occurred in the first hour and 80-90% in the second hour. Extensive carbonation of 96 hours was also studied. Its mass curve is shown in Fig. 3.10. Three points are taken in comparison at 4, 24, and 96 hours. After only 4 hours, the uptake already reached 68% of the total 4-day uptake, and 86% after 1 day. The third method to quantify the calcium carbonates in concrete is thermal decomposition analysis. Table 3.4 summarizes the results of the concrete batches after 28 days in different curing conditions. Mass loss was categorized into three components: evaporable water (up to 105ºC), combined water (between 105ºC and 470ºC) and carbon dioxide (between 470ºC and 950ºC). Higher bound water indicates more hydration products. It is noted that zero initial curing has the lowest uptake but highest bound water. Other samples, as 18a + 4c, have higher uptake, which is associated with high carbonates, but low bound water content. It can be concluded that the hydration and carbonation products cannot coexist in high quantities in one sample, or in other words, hydration

– 37 – products can be consumed in the carbonation reaction to produce more carbonates. However, once the water spray technique is applied to the batch with 18-hour initial curing followed by 4-hour carbonation curing, both hydration and carbonation products appear to be high. Apparently, water compensation after initial curing followed by carbonation curing is effective and beneficial for subsequent hydration and overall performance. Table 3.4 also compares bound water content between carbonated and steam cured concretes. After the same initial curing, the amount of hydration products is similar in carbonated and steam cured concretes, suggesting carbonation can technically replace steam to accelerate hydration. Carbon dioxide content in the last column of Table 3.4 is calculated based on total concrete mass used in thermal analysis. If the value is divided by 13%, the cement content, CO2 content is converted to a cement basis and comparable to the other two methods. In order to obtain the CO2 uptake, the CO2 content of the hydrated reference can be subtracted from the respective carbonated concrete.

Table 3.5 shows the CO2 uptake of all batches using three methods: mass gain, mass curve, and thermal analysis. For each batch, mass gain is average of three samples and mass curve recorded mass increase of entire batch. These two independent measurements are taken from the same setup at the same time, and thus, are directly comparable. CO2 uptake by thermal analysis is performed on fractured concrete pieces and serves as a third check. Since all carbonates, pre-existing and carbonation-induced, are thermally decomposed, it is necessary to deduct the initial pre-existing carbonates in the cement (1-2%). Thus, the thermal analysis results of Table 3.5 represent the CO2 uptake only due to the carbonation reaction. Carbonation of 4 hours is capable of storing

9.1% CO2 without initial curing and 23.5% CO2 as average with initial curing between 4 and 18 hours. CO2 uptake reached 22.0% by 2-hour carbonation and 35.3% by 96-hour carbonation after 18-hour initial curing. The average CO2 uptake is also shown in the last column of Table 3.5. In comparison to CO2 uptake of 6% in granite aggregate concrete after 18-hour initial curing (Rostami 2011), degree of carbonation achieved in this study is considered as very high. This is possibly attributed to the porous nature of CMU and the use of porous lightweight slag aggregates.

– 38 – Compressive strength The accelerated strength gain of concrete 4 hours after casting without initial curing is shown in Fig. 3.11. In comparison to hydration reference of 1.9 MPa, carbonation strength reached 5.6 MPa and steam strength 6.3 MPa. Early strength was improved by nearly three times due to accelerated curing. It seemed steam was slightly more effective. Twenty-four-hour strength was used to evaluate the effect of initial curing. They are compared in Fig. 3.12. The batches 0a, 4a, 6a and 8a represent initial air curing time followed by 4-hour accelerated curing and subsequent hydration to 24 hours. For instance, the 24-hour strength of carbonated concrete 8a+4c was tested after initial curing of 8 hours, carbonation of 4 hours, and subsequent hydration in a sealed plastic bag for 12 hours. Hydration reference for each batch also underwent different initial curing before sealed for normal curing. When comparing the results of all batches, hydration control without initial curing turned out to have the highest one-day strength because there was no water loss during the curing. One-day strengths of carbonated and steamed concretes without initial curing were of same value. With same initial curing time, carbonated samples were comparable to steamed although the RH of initial curing was different. It was 80% RH in steam batches and 50% RH in carbonation batches. It appeared that 4- hour carbonation without initial curing yielded highest carbonation strength in one day. Fig. 3.13 shows the effect of carbonation duration with fixed initial curing of 18 hours. The first two batches (18a+4c and 18a+2c) were tested one day after casting. It was obvious that longer carbonation time produced more strength. The last batch (18a+96c) was tested nearly 5 days after casting. Its high strength was attributed to both higher carbonation reaction and longer curing time. The corresponding carbonation degree was 70%. In case of the 28-day compressive strength in Fig. 3.14 and 3.15, the highest 4- hour carbonation strengths reached 15.5 MPa and 15.3 MPa with 4-hour and 18-hour initial curing respectively. The best hydration strength was 18.2 and 17.9 MPa for the reference and steam cured samples respectively. Obviously 4-hour carbonation produced lower late strength than hydration reference. The subsequent hydration was affected by water loss in early curing. Although extensive carbonation of 4-day produced comparable

– 39 – 28-day strength to the reference, it was not practical in commercial production. In an attempt to improve the compressive strength of short-term carbonated concrete, water spray after carbonation was adopted to restore the lost water during early curing. Water was slowly added until surface saturation. The process could last a few days until the lost water were all compensated. The results are summarized in Table 3.6. The best reference is the hydration without initial curing. Water compensation increased the compressive strength by 20-25% in comparison to carbonation without spray. This increase in strength is associated with the high bound water and carbon content presented in Table 3.4. The 28-day strength of 18a+4c after spraying was 17.9 MPa, which was comparable to the hydration strength (0a) of 18.2 MPa, and steam strength (4a+4s) of 17.9 MPa. The results show that carbonation followed by water compensation is beneficial to compressive strength. It is clear that early carbonation does not hinder subsequent hydration. It is the water loss during initial curing and carbonation that reduces hydration degree. With water compensation through spray, it is possible to make carbonation strength comparable to the best hydration reference. pH measurement Table 3.7 shows the pH values of carbonated, steam cured, and control samples. It is known that carbonation usually reduces the pH of concrete. This phenomenon can be noticed in comparing the carbonated and control samples. In all cases, the former’s pHs are in the range of 9-11, while the latter’s are in the range of 11-12. Also, the pHs of the steam-cured samples are similar to that of the control, as they were not exposed to carbon dioxide gas. The pH reading of hydration reference was lower than 12.5-13.0. It was due to the extraction method used in this study. Because of porous nature of block samples, it was difficult to produce representative pore solution by extraction. Therefore, the percent reduction is more relevant. pH of concrete after 2 to 4 hr carbonation was reduced by 9.0 – 19.2%.

Carbonation curing of 20-cm CMU Process parameters are selected for 20-cm CMU production. To optimize strength gain and CO2 uptake, it appears 18-hour initial curing followed by 4-hour carbonation

– 40 – and water compensation can produce a strength comparable to hydration reference and reach a CO2 uptake of 24% based on cement. Its carbonation degree is about 48%. Although 4, 6 and 8-hour initial curing can also be considered, it is difficult to implement based on a work shift. Eighteen-hour initial curing can be accomplished through overnight operation. The carbon storage capacity of concrete produced by this process is excellent. Standard 20-cm lightweight CMU block was made in the laboratory (Fig. 3.16) and carbonated in a setup similar to that shown in Fig. 3.4. The fresh blocks underwent an initial curing of 18 hours at 50% RH and 25ºC. Water loss due to initial curing was about 50%, very close to that observed in slab tests. The blocks were then carbonated at 0.1

MPa gas pressure for 4 hours. Water loss due to carbonation was about 5%. CO2 uptake by three block samples, at the age of 1 day, is presented in Table 3.8. The three methods were used in CO2 uptake estimation. In comparison, the three blocks demonstrated an uptake in the range of 23-25%. These results agreed with that of the slab samples, whose uptake was approximately 24%. The compressive strength of 1 day and 28 days CMU is presented in Fig. 3.17. Prior to testing, the blocks were capped with a white capping cement to ensure uniform load distribution on the block. One-day strength was tested after 18 hours of initial curing, 4 hours of carbonation and 2 hours of subsequent hydration. The carbonated CMU blocks exhibited almost twice the strength of the hydration reference in one day.

This was associated to the formation of carbonates at a 24% CO2 uptake. However, after 28 days, the hydration strength was slightly higher than carbonation strength in subsequent hydration because of the loss of water in initial curing in carbonation process. Similar to slab samples, water compensation through surface spray was also introduced after carbonation to restore the lost water. Thus, the 28-day compressive strength was comparable to the hydrated reference. CMU can be produced by carbonation process to replace steam and achieve high early strength, equivalent late strength, and superior CO2 uptake capacity. After 28 days, the carbonated and hydrated block samples are analyzed using XRD, and the results are shown in Fig. 3.18. Calcite and tricalcium silicate are the dominant compounds found in the carbonated sample, with the latter overlapping the

– 41 – former in most peaks. Dicalcium silicate was found but in weak counts and overlapped tricalcium silicate in several peaks. On the other hand, the XRD pattern identified calcium hydroxide, dicalcium silicate, and tricalcium silicate in the hydrated reference. A unique calcium hydroxide peak was detected at 18 2θ. It is clear calcium hydroxide was eliminated in carbonated concrete. Phenolphthalein was sprayed on a freshly cut surface of concrete to identify the carbonated and hydrated regions. Fig. 3.19 and 3.20 show carbonated and hydrated concretes respectively. After 28 days, the carbonated sample preserved a colorless surface with permanent carbon sequestration. On the other hand, a purple color in the hydrated sample indicated no carbonation reaction.

Carbon storage in blocks and bricks

Reaction of cement with carbon dioxide at early age is a CO2 sequestration process. If one 20-cm block with a mass of 15 kg contains 13% cement could have a CO2 uptake of 24% based on cement, one block has the capacity to store 0.47 kg of CO2 in a thermodynamically stable calcium carbonate form. Assuming every block or brick has the same carbon storage capacity, the projected annual production of 4.3 billion units in

US and Canada market can thus consume 2 million tons of CO2 per year. The capacity for carbon capture and storage (CCS) in geologic formation is approximately 1 million ton per year per site (IPCC 2005). CO2 utilization in concrete blocks and bricks production is equivalent to carbon sequestration in two geologic formation sites. The cement annual production in the United States and Canada is about 100 million tons with CO2 emission of 80 million tons. If all block and brick plants adopt carbonation curing with the same

CO2 uptake rate, CO2 utilization in their production lines alone could reduce carbon emission by 2.5% for cement industry.

The network operation and cost estimate

To implement CO2 utilization at the vicinity of carbon sources, a network needs to be established. It shall include carbon capture, compression, transport, storage, and utilization. The network will be operational based on the assumption that large quantities of high purity and low cost carbon dioxide will be produced as regulations requiring

– 42 – reductions in CO2 emissions are developed and will be otherwise transported to remote area for underground geologic storage. For CO2 utilization in block production, the pressurized high purity CO2 is instead transported to block plants. The cost estimate was made by comparing carbonation curing with steam curing in Table 3.9. It is assumed that the cost estimate starts when the pressurized gas arrives at block plants. The static carbonation process does not require additional energy except the initial air curing in environmental chamber at 50% RH and ambient temperature. For laboratory setting used in this preliminary study, the environmental chamber consumes 2.25 kW, but can fit up to 160 slab samples with a total concrete volume of 0.058 m3. Accordingly, per cubic meter of concrete, initial curing by 4-hour and 18-hour consumes 155 and 696 kWh and cost $9.27 and $41.74 respectively, assuming 1 kWh of electricity costs $0.06. Energy consumption in atmospheric steam curing is well studied. It is approximately 164 kWh/m3 concrete (Kawai 2005) and costs $9.84/m3 concrete. With reference to steam, the cost by 4-hour initial curing is comparable but the cost by 18-hour initial curing is too high. In order to reduce the cost without compromising the superior performance, the initial curing duration must be modified while maintaining more than 50% water loss. It appears that the water lost during initial curing enhances the carbonation degree and maintains the compressive strength if followed by water spray. Technically, it is not necessary, and even impossible, to have 18 hours initial curing in a controlled environmental chamber in industry production. There is room to reduce the energy consumption in carbonation by using different initial curing and eliminating vacuum. While the industrial steam curing mechanism is optimized for mass production, the cost of the network operation for carbonation should also be analyzed and optimized to make the system economically feasible and sustainable.

CONCLUSIONS

1. Initial curing is beneficial for CO2 uptake in an early carbonation process of CMUs for an accelerated curing and carbon dioxide storage. However, initial curing could be detrimental to late strength development because of water loss. Therefore, initial curing shall be minimized to secure the performance and reduce

– 43 – the process cost. With water compensation immediately after carbonation, high

early strength, equivalent late strength, and superior CO2 uptake can be achieved. 2. For early carbonation targeted in the first 24 hours after casting, it is not possible to reduce the internal relative humidity to an ideal level of 50-60%. No matter what drying process is used, internal RH remains higher than 80% within 24 hours. Water content is a better parameter to justify the condition for carbonation.

3. CO2 uptake capacity by CMU is dependent on initial curing. Taking cement

binder in CMU as reference, the CO2 uptake in 4-hour carbonation treatment reached approximately 8.5% by zero initial curing, 22% by 4 to 8 hours initial curing, and 24% by 18 hours initial curing. Longer carbonation time of 96 hours

could promote CO2 uptake to 35%. It is corresponding to a degree of carbonation of 70%. 4. The process parameters obtained from slab samples are successfully applied to

20-cm CMU production in laboratory. Both CO2 uptake and strength gain in CMU are of close value to slab tests. Therefore, laboratory work can be scaled up to commercial production. XRD patterns illustrate that the carbonation products

are stable calcite with the consumption of C2S, C3S, CH, and possibly CSH. 5. Annual cement production in USA and Canada is 100 million tons and its

corresponding CO2 emission is about 80 million tons. If all CMUs produced in these two countries are cured by carbonation for accelerated hydration and carbon

storage, the CMU production could consume 2 million tons of CO2 every year based on an uptake rate of 24%. The corresponding emission reduction for cement industry reaches 2.5% by CMU production alone. 6. Low energy and low cost initial curing methods shall be developed to make carbonation curing economically competitive to steam process.

ACKNOWLEDGEMENTS The authors gratefully acknowledge the financial support by the Natural Science and Engineering Research Council (NSERC) of Canada and Canadian Concrete Masonry Producers Association (CCMPA), and the supply of expanded slag aggregates by Lafarge Canada.

– 44 –

REFERENCES ASTM (2004b). "Standard Test Methods for Sampling and Testing Concrete Masonry Units and Related Units." ASTM International C140.

Freedonia, G. (2010). "Brick and block demand to reach 12.4 billion units, $8 billion by 2014." Journal of Concrete Products (http://www.freedoniagroup.com/DocumentDetails.aspx?DocumentId=506289 ).

Goodbrake, C. J., Young, J. F., and Berger, R. L. (1979b). "Reaction of hydraulic calcium silicates with carbon dioxide and water." Journal of the American Ceramic Society 62(9-10): 488-491.

IPCC (2005). "IPCC Special Report on Carbon Dioxide Capture and Storage."

Johnson, D. (2000). "Accelerated carbonation of waste calcium silicate materials." Society of Chemical Industry 10.

Kawai, K., and Sugiyama, T. (2005). "Inventory data and case studies for environmental performance evaluation of concrete structures." Journal of Advanced Concrete Technology: 435-456.

Rostami, V., Shao, Y., and Boyd, A. J. (2011). "Durability of concrete pipes subjected to combined steam and carbonation curing." Construction and Building Materials 25: 3345-3355.

Shao, Y., Mirza, M. S., and Wu, X. (2006). "CO2 sequestration using calcium-silicate concrete." Canadian Journal of Civil Engineering 33: 776-784.

Shi, C., and Wu, Y. (2008). "Studies on some factors affecting CO2 curing of lightweight concrete products." Resources, Conservation, and Recycling(52): 1087-1092.

Shideler, J. J. (1955). "Investigation of the moisture-volume stability of concrete masonry units." Portland Cement Association Research and Development Laboratories Bulletin D3.

Steinour, H. (1956). "Some Effects of Carbon Dioxide on Mortar and Concrete." American Concrete Institute Journal 30: 905-907.

Toennies, H. T. (1960). "Artificial carbonation of concrete masonry untis." American Concrete Institute Journal(56-42): 737-755.

Toennies, H. T., and Shideler, J. J. (1963). "Plant drying and carbonation of concrete block - NCMA-PCA cooperative program." American Concrete Institute Journal 60(33): 617-632.

– 45 –

Young, J. F., Berger, R. L., Breese, J. (1974). "Accelerated Curing of Compacted Calcium Silicate Mortars on Exposure to CO2." Journal of the American Ceramic Society 57(9): 394-397.

– 46 – TABLES

Table 3.1: CO2 uptake and compressive strength (MPa) of different aggregate concretes

CO2 Carbonated Strength Reference Strength Label Aggregate Description Uptake (%) 18-day 28-day 18-day 28-day GR Granite + Riversand 33.15 10.47 11.71 7.74 7.85 ER Expanded Shale + Riversand 31.36 19.45 21.51 10.94 12.22 ES Expanded Slag 39.43 22.82 26.40 14.38 17.19

Table 3.2: Mixture proportion Slab CMU Mass Percent (g) (g) (kg/m3) (%) Ordinary Portland Cement 88 1967 241 13 Water 35 787 96 5 SSD Expanded Slag 554 12251 1502 82 Concrete Sample 677 15004 1839 100

– 47 – Table 3.3: Curing procedures

Subsequent

Initial curing Steam curing Carbonation curing Hydration

RH T t RH T t t Water t

Batch # Condition (%) (°C) (hours) (%) (°C) (hours) (hours) Spray (g) (days)

1 0a + 4s - - 0 95±5 75±5 4 - - 28

2 4a + 4s 80±5 22±1 4 95±5 75± 5 4 - - 28

3 6a + 4s 80±5 22±1 6 95±5 75±5 4 - - 28

4 8a + 4s 80±5 22±1 8 95±5 75±5 4 - - 28

5 0a + 4c - - 0 - - - 4 - 28

6 0a + 4cw - - 0 - - - 4 1±0.2 28

7 4a + 4c 50±1 25±0.2 4 - - - 4 - 28

8 4a + 4cw 50±1 25±0.2 4 - - - 4 17±2 28

9 6a + 4c 50±1 25±0.2 6 - - - 4 - 28

10 8a + 4c 50±1 25±0.2 8 - - - 4 - 28

11 18a + 4c 50±1 25±0.2 18 - - - 4 - 28

12 18a + 4cw 50±1 25±0.2 4 - - - 4 29±2 28

13 18a + 2c 50±1 25±0.2 18 - - - 2 - 28

14 18a + 96c 50±1 25±0.2 18 - - - 96 - 28

15 336a+96c 50±1 25±0.2 336 - - - 96 - 28 a – Initial air curing; s – Steam curing; c – Carbonation; RH – Relative humidity; T – Temperature; t – Time; W – Water sprayed after carbonation.

– 48 – Table 3.4: Thermal decomposition analysis after 28 days Initial Mass Evaporated Water1 Combined Water2 Carbon Dioxide3 Sample (g) Mass (g) %4 Mass (g) %4 Mass (g) %4 0a + 4c 47.25 0.90 1.90 1.76 3.72 0.68 1.44 0a + 4s 48.26 0.72 1.49 1.86 3.85 0.12 0.25 0a 38.84 1.10 2.83 2.55 6.57 0.01 0.25 4a + 4c 76.45 1.09 1.43 2.35 3.07 2.37 3.10 4a + 4s 35.58 0.52 1.46 1.12 3.15 0.09 0.25 4a 45.91 0.91 1.98 1.95 4.25 0.12 0.26 6a + 4c 44.33 0.49 1.11 1.00 2.26 1.46 3.30 6a + 4s 38.76 0.73 1.88 0.95 2.45 0.10 0.26 6a 34.92 0.98 2.81 1.52 4.35 0.09 0.26 8a + 4c 47.26 0.42 0.89 1.82 3.85 1.58 3.34 8a + 4s 55.4 0.66 1.19 1.62 2.92 0.14 0.25 8a 47.28 1.16 2.45 1.79 3.79 0.12 0.25 18a + 4c 65.15 0.67 1.03 1.26 1.93 2.32 3.56 18a + 4c5 71.14 0.7 0.98 1.39 1.95 2.56 3.60 18a + 2c 50.69 0.47 0.93 2.08 4.10 1.59 3.13 18a + 96c 40.57 0.46 1.13 1.05 2.59 1.98 4.88 18a 43.61 0.50 1.15 1.42 3.26 0.11 0.25 1 – Difference in mass between room temperature and 105°C 2 – Difference in mass between 105°C and 470°C 3 – Difference in mass between 470°C and 950°C 4 – Percentage of initial sample mass 5 – Sprayed after carbonation

– 49 – Table 3.5: Comparison of CO2 uptake by three methods

CO2 Uptake (% Cement Mass) Average Mass Thermal due to Condition Mass Gain Curve analysis carbonation 0a - - 0.0 0.0 0a + 4c 7.5 ± 1.8 9.0 9.2 8.6 4a + 4c 21.3 ± 1.4 22.1 21.9 21.8 6a + 4c 22.8 ± 1.8 23.6 23.5 23.3 8a + 4c 23.5 ± 1.6 24.0 23.8 23.8 18a + 4c 24.2 ± 0.4 24.4 25.5 24.7 18a + 2c 20.4 ± 0.4 21.7 22.2 21.4 18a + 96c 34.9 ± 1.6 35.2 35.6 35.2

Table 3.6: Compressive strength of concrete with water spray

Batch After 1 Day (MPa) After 7 Days (MPa) After 28 Days (MPa) 0a + 4c 8.6 ± 0.7 11.2 ± 1.0 15.1 ± 1.5 0a + 4cW 10.0 ± 0.7 12.6 ± 1.3 15.9 ± 1.7 0a 8.8 ± 1.0 16.4 ± 1.5 18.2 ± 1.3 4a + 4c 7.9 ± 0.6 10.1 ± 1.0 15.5 ± 1.6 4a + 4cW 8.1 ± 0.6 12.1 ± 1.1 15.8 ± 1.3 4a 4.4 ± 0.3 11.4 ± 0.9 13.2 ± 1.4 18a + 4c 9.9 ± 0.6 11.1 ± 0.9 15.3 ± 1.2 18a + 4cW 10.5 ± 1.2 14.6 ± 1.2 17.9 ± 1.4 18a 5.2 ± 0.5 12.0 ± 1.1 14.4 ± 1.2 W – Water sprayed after carbonation

– 50 – Table 3.7: pH of surface sample at 28 days % Reduction Batch pH @ 28 days by carbonation 0a + 4c 10.8 ± 1.1 9.0 0a + 4s 11.5 ± 1.2 - 0a 11.8 ± 1.2 - 4a + 4c 10.5 ± 0.8 11.0 4a + 4s 11.7 ± 1.0 - 4a 11.8 ± 1.0 - 6a + 4c 10.1 ± 0.8 13.3 6a + 4s 11.6 ± 1.2 - 6a 11.7 ± 1.1 - 8a + 4c 9.9 ± 1.0 11.7 8a + 4s 11.7 ± 1.0 - 8a 11.2 ± 0.9 - 18a + 4c 9.2 ± 1.1 18.1 18a + 2c 10.1 ± 1.0 10.1 18a + 96c 9.1 ± 0.8 19.2 18a 11.3 ± 1.0 -

Table 3.8: CO2 uptake in 20-cm CMU samples

CO2 Uptake (%) Mass Mass Thermal CMU # Gain Curve analysis 1 23.8 24.0 23.5 2 23.1 23.6 24.3 3 23.6 23.8 23.9

Table 3.9: Energy consumption and cost estimate Power Time Power Volume Power Cost Process (kW) (hr) (kWh) (m3) (kWh/m3) ($/m3) Carbonation 2.25 4 9 0.058 154.58 9.27 Carbonation 2.25 18 40.5 0.058 695.61 41.74 Steam Curing2 - - - 1.00 164.00 9.84

– 51 – FIGURES

120

100

80

60

% Passing % Granite 40 Riversand

20 Expanded Slag

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 Sieve Size (mm) Fig. 3.1: Sieve analysis of different aggregates

120

100

80

60 % Passing % 40

20

0 0 5 10 15 20 25 30 Sieve Size (mm) Fig. 3.2: Sieve analysis of expanded shale

– 52 – 120

100

80

60 % Passing % 40

20 Maximum absorption capacity = 9%

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 Sieve Size (mm)

Fig. 3.3: Sieve analysis of expanded slag aggregates

Fig. 3.4: Schematic of carbonation setup

– 53 –

Fig. 3.5: Schematic of measuring internal relative humidity

90

80 81 70 24 hours 60 14 days 53 50 18 hours 51 40 33 33 8 hours 30 32 6 hours Initial Curing Conditions: 20 4 hours T = 25°C Initial Curing Initial Water (%) Loss RH = 50% 10

0 0 24 48 72 96 120 144 168 192 216 240 264 288 312 336 360 Time (Hours)

Fig. 3.6: Water loss due to initial curing

– 54 – 100 Bottom 95 Middle 90 Top 85 80 75 70

65 Relative Humidity Humidity Relative(%) 60 55 50 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Time (Days)

Fig. 3.7: Internal relative humidity during initial curing

40 Water Loss w.r.t Initial Water Mass 34.8 35 Carbon Uptake w.r.t Initial Cement Mass 30 24.2 25 22.8 23.5 21.3 20.4 20

Percent (%) Percent 15 9.6 10.2 10 7.5 8.2 5.9 5.3 4.2 5 2.2

0 0a + 4c 4a + 4c 6a + 4c 8a + 4c 18a + 4c 18a + 2c 18a + 96c Batch

Fig. 3.8: Water loss due to carbonation and CO2 uptake

– 55 – 30 18a+4c 8a+4c 25 6a+4c 24.36 23.98 4a+4c 23.61 0a+4c 20 22.11

15

Uptake Uptake (%)

2

10 % CO % 8.99

5

0 0 60 120 180 240 300 Carbonation Time (minutes)

Fig. 3.9: Mass curves of 4-hour carbonation

40 18a + 96c 35 35 30

30

25 24

20

Uptake Uptake (%)

2 15 CO 10

5

0 0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102 Carbonation Time (Hours)

Fig. 3.10: Mass curve of 4-day carbonation

– 56 – 9 4-hour hydration 8 4-hour carbonation 7 4-hour steam 6.3 6 5.6

5

4

3 1.9

2 Compressive Strength (MPa)

1

0 0a

Fig. 3.11: Compressive strength after 4 hours

14 4 hr carb + hydration to 24 hr hydration to 24 hr (ref) 12 4 hr steam + hydration to 24 hr 9.5 10 8.6 8.8 8.7 7.9 7.2 8 7.0 6.3 5.8 6 5.0 5.3 4.4

4

DayCompressive Strength (MPa) -

2 One

0 0a 4a 6a 8a Batch

Fig. 3.12: Compressive strength after 1 day

– 57 – 16 Carbonated 13.7 14 Reference 12 9.9 10

8 6.6 6.7 6 5.2 4.5

4

DayCompressive Strength (MPa) -

2 One

0 18a + 4cu 18a + 2cu 18a + 96cu Batch

Fig. 3.13: Compressive strength with varied carbonation duration

25 4 hr carbonation + hydration to 28 days hydration to 28 days 4 hr steam + hydration to 28 days 20 18.2 17.9 16.7 15.1 15.5 15 13.2 12.9 12.9 12.2 10.4 10 9.2 7.6

Day CompressiveDay Strength (MPa) 5

- 28

0 0a 4a 6a 8a Batch

Fig. 3.14: Effect of initial curing on compressive strength

– 58 – 20 17.7 18 Carbonated 15.3 Reference 16 14.4 14.4 14 12.8 12.6 12 10 8 6

Day CompressiveDay Strength (MPa) 4 -

28 2 0 18a + 4cu 18a + 2cu 18a + 96cu Batch

Fig. 3.15: Effect of carbonation time on compressive strength

Fig. 3.16: Standard CMU made in lab

– 59 – 20 1-day 28-day 14.1 15 13.9 12.7

9.2 10 7.8

4.3

5 Compressive Strength (MPa) CompressiveStrength

0 18a+4c 0a (Ref) 18a+4c+sp Sample Fig. 3.17: Compressive strength of CMU

Fig. 3.18: XRD patterns of CMU blocks

– 60 –

Fig. 3.19: Carbonated concrete (18a+4c) after 28 days sprayed with Phenolphthalein

Fig. 3.20: Hydrated concrete (0a) after 28 days sprayed with Phenolphthalein

– 61 – Chapter 4

—

REACTION PRODUCTS OF LIGHTWEIGHT OPC CONCRETE MASONRY

UNITS BY STATIC CARBONATION

PREFACE Among the many different batches cast in the previous chapter, four were selected based on superior CO2 uptake and compressive strength. The microstructure of the four batches were investigated thoroughly: the best hydrated (0a), freshly carbonated (0a+4c), air cured and carbonated (18a+4c), and sprayed after carbonation (18a+4c+sp). They were chosen to study the effect of air curing on the carbonation reaction, the effect of spraying on the concrete microstructure and subsequent hydration, and compare carbonated batches to their hydrated counterparts. The curing regimes employed in this part of the research are the same as those used in the previous chapter. Carbonation was performed on concretes either immediately after casting or after 18-hour air curing. The spraying technique is introduced to compensate for the water lost during air curing and carbonation to facilitate subsequent hydration. Two experiments utilized in this chapter require the use of a powder representing the concrete sample. However, separation of aggregates from paste is a near impossible task. A process involving initial air curing, carbonation curing, water compensation, and subsequent hydration was developed to maximize the degree of carbonation and hydration. The compressive strength is monitored after 1 and 28 days in order to study the evolution of strength through subsequent hydration. Reaction products of carbonation–cured concretes at early and late age were characterized using thermogravimetrical (TG) analysis, X-ray diffraction analysis, and scanning electron microscopy. The chapter presented here has been accepted for publishing with minor change in the American Society of Civil Engineers (ASCE) Journal of Materials in Civil Engineering.

– 62 – INTRODUCTION Carbonation of concrete at early age is an accelerated curing process (Young 1974). It can be carried out immediately after concrete casting or after initial air curing. The process is best suited for precast concrete production. It was found that immediate carbonation of 5 minutes on fresh cement paste compact could produce a strength equivalent to that by normal curing of 24 hours (Klemm 1972; Young 1974). In another study, 30-minute immediate carbonation on cement compacts led to a strength comparable to 28-day hydrated reference (Wagh 1995). Carbonation after initial air curing was also examined. Although the strength increase was not significant by carbonation after overnight air curing, durability of concrete could be considerably improved (Rostami 2011). Carbonation was also employed to treat concrete blocks after steam curing to reduce shrinkage cracking of masonry structure during the service (Toennies 1960; Toennies 1963). In addition, carbonation curing was capable of recycling carbon dioxide recovered from cement kilns

(Shao 2006). The process offers an economic means of CO2 utilization in the vicinity of

CO2 sources to avoid distant disposal for carbon emission reduction. A recent study showed that if a sufficient amount of free water could be removed from concrete before carbonation treatment, it was possible to reach a CO2 uptake of 23% in terms of cement mass with a concrete strength equivalent to the hydrated references (El-Hassan 2012a). The corresponding reaction efficiency was 46% if full carbonation was considered as 100%, significantly higher than the efficiency of 12-20% reported in previous studies (Young 1974; Rostami 2011). At such a high uptake rate, one typical 20-cm masonry block could consume 0.33 kg of carbon dioxide. A concrete with such high CO2 content is no longer a normal concrete by conventional hydration curing. It is, therefore, necessary to carry out a study to understand the reaction products from this early-age carbonation process. The previous studies on carbonation reaction products were focused on either immediate carbonation or weathering carbonation. For immediate carbonation curing, most studies used calcium silicates or cement pastes as materials to avoid the influence of aggregates on powder analysis. Immediate carbonation took place when dissolved carbon dioxide reacted with calcium ions in the pore solution or on surface of cement particles.

– 63 – The final structure was a hybrid of calcium silicate hydrate and calcium carbonate (Klemm 1972; Young 1974; Goto 1995). XRD analysis was commonly utilized. It was found that the carbonation of calcium silicates led to the formation of calcite, aragonite, and vaterite with stable calcite being the dominant final product (Sawada 1997). However, in the case of incomplete carbonation, unreacted calcium silicates also hydrated into calcium hydroxide (Rostami 2012). Additionally, TG/DTG was employed to investigate carbonation. The mass loss detected at different temperature ranges could be used to estimate the amount of hydration and carbonation products (Goodbrake 1979b; Goto 1995; Shtepenko 2006). Morphological changes due to carbonation were studied using SEM. The findings of the past work were inconsistent. While some identified crystal near the surface of cement particles (Monkman 2006), other studies showed amorphous and non-crystalline forms (Shtepenko 2006). Reaction products in weathering carbonation were created from reactions between carbon dioxide and hydration products, leading to decomposition of C-S-H and production of calcium carbonates and silica gel (Hyvert 2010). Weathering carbonation is not a desired reaction for concrete. The purpose of this paper is to examine the reaction products of lightweight concrete cured by 4-hour carbonation after an 18-hour initial air curing. This early carbonation curing within 24 hours is different from immediate carbonation or weathering carbonation in that the process window allows the removal of free water for maximized CO2 uptake and the addition of curing water after 4-hour carbonation for maximized hydration. CO2 uptake and strength gain are quantified. Carbonation products are characterized by X-ray Diffraction (XRD), thermogravimetry analysis (TG), and scanning electron microscopy (SEM). Most of the previous research on reaction products were conducted using cement or pure calcium silicates in microstructure analysis (Young 1974; Goto 1995). This study is focused on lightweight concrete with expanded slag aggregates. A method is developed to quantify the cement content in a concrete powder sample, which is prepared for TG analysis.

– 64 – EXPERIMENTAL INVESTIGATIONS Concrete sample preparation Rectangular concrete samples 127 mm long, 76 mm wide, and 38 mm thick were prepared to simulate a typical web of a 20-cm concrete masonry unit (CMU). The mixture proportion is shown in Table 4.1. ASTM Type I ordinary Portland cement and saturated surface dry (SSD) expanded slag aggregates with a dry bulk density of 951 kg/m3 were used. The composition of OPC is summarized in Table 4.2. The water content in the as-received slag aggregates was 5% by mass with a maximum absorption capacity of 9%. According to the sieve analysis, the well-graded aggregates ranged in size between 0.2 and 6 mm (El-Hassan 2012a). Based on the water added for mixing, the water to cement ratio (w/c) was 0.4. If water in aggregates was also considered as part of total water in concrete, w/c reached 0.71. The raw materials were mixed in a pan mixer and samples were compact formed using a vibrating hammer to simulate the industry production of CMU. Concrete was then demolded right after casting. Four curing schemes were investigated. Sealed hydration in a plastic bag served as reference (0a). Immediate carbonation of four hours followed by a subsequent hydration was also examined (0a+4c). Initial air curing of 18 hours in an environmental chamber of 50% relative humidity (RH) and 25ºC was applied to batches (18a+4c) before 4-hour carbonation. To compensate for water loss during initial air curing and carbonation curing, water spray was devised immediately after carbonation to restore all lost water (18a+4c+sp). This was done slowly over the next two weeks. The detailed procedure is described in Chapter 3. The subsequent hydration was carried out in a sealed bag at a relative humidity of 80±5% and room temperature (24±1ºC) until 24 hours and 28 days for strength tests. The fractured samples were preserved in an acetone solution to stop hydration. Acetone exchange with water to stop hydration was reported as the method to least affect the microstructure (Collier 2008). Prior to analysis, the samples were pre- dried overnight at 60oC.

Measurement of CO2 uptake in concrete Fig. 4.1 shows the setup of early carbonation curing. Concrete samples with or without initial air curing were placed in a carbonation chamber, which was vacuumed to

– 65 – about 0.7 bars within a few seconds, and then filled with carbon dioxide gas to an absolute pressure of 1 bar. The very short period of vacuum should not change the porosity of concrete. During 4-hour carbonation, the pressure was kept constant so that the consumed CO2 could be replenished. The relative humidity of the gas reached nearly 85% at the end of the 4-hour carbonation.

Two methods were used to quantify CO2 uptake in concrete with no need of paste separation: mass gain method and furnace decomposition method. Three samples for each batch were tested for average uptake. Mass gain method compares mass of concrete before and after carbonation (Eq. 4.1). Carbonation-induced water loss due to exothermic reaction was collected by absorbent paper and added to the final mass. By treating the system as a closed system, it was imperative to include the lost water, which was initially part of the concrete mass prior to carbonation. Since not all evaporated water, especially the vapor, can be collected, mass gain method underestimated the uptake and is, therefore, more conservative. CO2 uptake so measured is an average of the entire batch and independent of carbonates which may exist before carbonation curing.

CO2 uptake (%) = (4.1)

Furnace decomposition analysis was performed to estimate CO2 content in concrete as a comparison. Three samples per batch were tested for average. Instead of using classical thermogravimetric analysis, which allows only micrograms of powder, a furnace of maximum temperature of 1100ºC was employed to ignite large concrete samples with a mass ranging from 35 g to 70 g for each. In this case the need to separate paste from concrete was avoided. The concrete chunk was selected in such a way that it contained the entire thickness to provide an average. Concrete chunk samples were heated up to 105C, 470C, and 950C to quantitatively measure the corresponding evaporable water, bound water in hydration products, and carbon dioxide in carbonation products respectively. The temperature range convention was set up based on the DTG curves. The mass at each temperature was recorded. CO2 content can be calculated based on Eq. 4.2.

CO2 content (%) = (4.2)

– 66 – Mass of cement in Eqs. 4.1 and 4.2 can be estimated by concrete mass multiplied by 0.13, which is the cement content in concrete.

Performance evaluation The carbonated concretes were compared to a hydrated concrete reference with no initial air curing. The compressive strength was tested after 1 day and 28 days subsequent hydration following ASTM standards (ASTM 2004b). Three rectangular specimens for each batch were tested and averaged with a compressive area of 127 mm x 38 mm. X-Ray Diffraction analysis was performed on powder samples. Two specimens per batch were used to confirm the obtained results. The XRD patterns were obtained after 1 and 28 days compressive strength tests using a Philips PW1710 Powder Diffractometer (Cu, K radiation, X'celerator proportional detector, scan interval 10- 100, 0.02, and 0.5 seconds per step) in order to identify phases such as calcium hydroxide, calcium silicates, and calcium carbonates in carbonated and hydrated concretes. The same powder samples were further subjected to thermogravimetry analysis by employing a thermal analyzer (NETZSCH, TG 449 F3 Jupiter) with a resolution of 0.01 mg. Similarly, two samples were analyzed for one batch. The TG and DTG curves were thus obtained in terms of mass loss between 25C and 950C at a heating rate of 10C/min. Scanning electron microscopy (SEM) analysis was performed on the four concretes after 1 and 28 days. Two specimens per batch were analyzed. The samples from fractured surface of tested concrete were coated with a gold-palladium layer to ensure conductivity. High-vacuum scanning electron microscopy in backscattered electron (BSE) mode (Hitachi S4000) was employed to conduct morphological analysis.

Method of determining cement content (CC) in powder samples To quantify carbonation reaction products using TG analysis, the use of powder samples is necessary. A procedure was developed to estimate the cement content in powder samples that were collected from a 5-mm thick surface layer of concrete. The

– 67 – samples used in this experiment were different from those used in furnace decomposition analysis. To obtain an average, two samples per batch were chosen: 1) A 5-mm thick concrete surface layer was cut off by saw after mechanical testing. Part of the surface concrete was crushed by hammer and then sieved to pass 125- µm to remove large size aggregates. This powder is still a mix of cement and aggregate and will be used for analysis. TG test was then performed on the

powder of about 120 mg to obtain CO2 loss between 470ºC and 950ºC from

powder mass of Mp. This CO2 loss is referred to as (CO2)p. 2) To quantify the cement content in this powder sample, the other part of the surface concrete chunk of about 10-12 g was ignited in a furnace between 470ºC

and 950ºC for CO2 content, (CO2)c. This concrete mass is referred to Mc.

3) Assuming that on the 5-mm thick surface layer of the same concrete, the CO2 content shall be the same whether calculated from powder sample or concrete chunk, that is:

CO2 content (%) = = (4.3)

Since cement content in concrete is known as 0.13, the cement content (CC) in powder sample can be estimated by Eq. 4.3. This correction shall be carried out for each sample. The cement content in powder is needed to allow the expression of percent hydration and carbonation products in terms of cement using TG/DTG analysis.

CO2 content of powder samples was also determined by coulometric titration in a hydrochloric acid (HCl) solution (Huffman 1977). Two samples were used to obtain an average. The powder sample was immersed in the acid solution in a pipette which was sealed from the opening by a cork. A thin plastic tube was connected at one end to the inside of the pipette and at the other end to a coulometer. The reaction between hydrochloric acid (HCl) and carbon-containing compound such as calcium carbonate led to the release of carbon dioxide gas. The coulometer measured the amount of carbon released, and ultimately the carbon dioxide by stoichiometric proportions.

– 68 – EXPERIMENTAL RESULTS AND DISCUSSION Carbonation behavior

The degree of carbonation is characterized by CO2 uptake. Table 4.3 compares

CO2 content of concretes using two different methods. The CO2 content determined by mass gain method is in fact the CO2 uptake due to carbonation. It was measured instantly after the carbonation process. Fresh concrete (0a+4c) was carbonated for 4 hours and resulted in an uptake of 7.5%. Obviously the presence of surface free water obstructed the diffusion of carbon dioxide and prevented carbonation to the core. In the case of the air cured samples (18a+4c), CO2 uptakes reached 24-25%. It was the initial air curing that created capillary space for carbonation products to precipitate. The second method presented in Table 4.3 is furnace analysis using samples each with a mass range of 35-70 g. Mass loss due to decarbonation between 470ºC and 950ºC is considered as CO2 content. Unlike mass gain, thermal analysis detects both originally- existed and carbonation-induced carbonates in concrete. The original CO2 content was determined from hydration reference. It was 1.7-1.9%, indicating the presence of a small amount of carbonates in the as-received cement. In carbonated concretes, CO2 content reached 8.9-11.0% without air curing and 24.0-27.7% with initial air curing. The results were consistent between 1 and 28 days, suggesting that carbonation is a permanent sequestration process. To obtain CO2 uptake due to carbonation by thermal method, CO2 content detected in the reference should be subtracted from that of carbonated concretes.

The CO2 uptakes presented in Table 4.3 are thus the average of mass gain method and furnace analysis method. The results showed that 18-hour air curing had enhanced CO2 uptake by three times in comparison to immediate carbonation. Thermogravimetry analysis (TG/DTG) of powder samples was also conducted to quantify reaction products. To do so, cement content in powder sample should be first determined. It was accomplished by combining TG analysis with furnace analysis using Eq. 4.3. The sample mass was 120 mg each in TG tests and 10-12 g each in furnace test. The cement contents (CC) for powder samples are presented in Table 4.4. Unlike the cement content in concrete chunk which is a constant of 13%, the cement content in powder samples is higher, ranging between 0.40-0.75. It is because of the removal of large pieces of aggregates over 125 µm. CO2 content determined by TG method is

– 69 – compared with HCl method in Table 4.5. CO2 uptakes were calculated by subtracting the original CO2 content in hydration reference from total. Their CO2 uptake averages obtained from powder samples (Table 4.5) were comparable to that obtained from concrete samples (Table 4.3). In Table 4.5, the percent cement content (CC) was used in calculating the mass of dry cement in powder samples so that the comparison was on the same dry cement base. Therefore, the furnace results in Table 4.3 and TG results in Table 4.5 provided an additional comparison. The difference of the two tests was in that furnace tests in Table 4.3 employed concrete sample mass, 35-70 g each, representing a through- thickness average; while TG tests in Table 4.5 used much smaller mass, 120 mg each, representing 5-mm surface layer. It was clear that the surface layer contained higher CO2 content than the through-thickness average in immediate carbonation (0a+4c). With initial air curing (18a+4c), the difference between surface layer and through-thickness average was much smaller, owing to a higher degree of reaction.

Compressive strength Fig. 4.2 compares the compressive strength of four concretes at 24 hours and 28 days after casting. The 24-hour duration represents the initial air curing followed by 4- hour carbonation and subsequent hydration to 24 hours. For instance, sample 18a+4c underwent 18 hours of air curing followed by 4 hours of carbonation and 2 hours of subsequent hydration. While immediate carbonation (0a+4c) did not produce higher strength than the hydration reference after 24 hours, the initial air curing made carbonated concrete stronger than its hydrated counterpart. In a period of two hours of subsequent hydration, only a small portion of lost water was possibly compensated by surface spray. Nevertheless, carbonated concrete with water spray (18a+4c+sp) had already exhibited a strength 6.5% higher than concrete without water compensation. After 28 days, the hydrated reference exhibited higher compressive strength than the carbonated (0a+4c and 18a+4c) owing to the water loss in initial curing. Therefore, water spray after carbonation was essential to restore the lost water. Consequently, a strength increase of 17% was observed after water compensation. The results indicate that carbonation followed by a water compensation is beneficial. Early carbonation does not hinder subsequent hydration. It is the water loss during early air curing and carbonation

– 70 – curing that reduces degree of hydration. With water compensation through surface spray, it is possible to make carbonation strength comparable to hydration counterpart.

Phase analysis XRD patterns of concretes at 1-day are plotted in Fig. 4.3. In carbonated concretes, the peaks of calcium carbonates were strong and dominant. In comparison to the hydration reference at age of 1 day (0a), immediate carbonation followed by hydration (0a+4c) produced both carbonation and hydration products. A calcium hydroxide peak at 18 2θ was distinctive. The formation of calcium hydroxide was believed to have occurred during the subsequent hydration after carbonation. On the other hand, carbonation after 18-hour air curing (18a+4c) eliminated calcium hydroxide and partially consumed anhydrous calcium silicates, leading to a CaCO3 dominated diffraction pattern. The consumption of calcium silicate phases was also evident. Fig. 4.4 shows XRD patterns of concretes at 28 days after subsequent hydration. A detailed comparison of the carbonated concretes between 1 and 28 days revealed that the original vaterite and aragonite peaks at 23o and 36o 2θ respectively existed at 1 day were significantly reduced after 28 days. This was possibly related to the transformation of the unstable, poorly crystalline CaCO3, such as aragonite and vaterite, to more stable, well crystalline polymorph, such as calcite (Sawada 1997). The hydrated reference did not differ between 1 and 28 days with slightly higher peaks of calcium hydroxide. A slight increase in baseline between 25º and 35º 2θ was also an indication of the presence of C-S-H. It was noted that the intensity of the peak at 29.4º 2θ, which was calcite overlapped by C3S, was stronger in concrete with 18-hour air curing than immediate carbonation. It was consistent with the CO2 uptake.

TG/DTG analysis DTG curves of four concretes after 1 day curing are shown in Fig. 4.5. According to Ramachandran and Beaudoin (2001), the mass loss along TG curves can be classified into 6 categories: Below 105ºC, the mass loss is associated with the evaporable water and possibly the poorly formed C-S-H. The water loss between 105C and 200C represents the bound water in low temperature C-S-H and ettringite, while the loss in the range of

– 71 – 200C-420C is associated to the bound water in well-formed hydration products including C-S-H and C-A-H. Bound water in calcium hydroxide is decomposed between

420ºC-470ºC. While CO2 in poorly crystalline calcium carbonates is decomposed between 470C-720C, that in well crystalline calcium carbonates lies in the range of 720C-950C (Ramachandran 2001; Li 2003). The DTG curves in Fig. 4.5 are thus compared. In the hydration reference (0a) and the freshly carbonated (0a+4c) concretes, an endotherm at 450C indicates dehydration of Ca(OH)2 (Chang 2006). This is confirmed by the Ca(OH)2 peaks in the XRD patterns. After 470C, the carbonates formed in 0a+4c were decomposed in a broad range in the poorly crystalline range. On the other hand, concretes with initial air curing (18a+4c and 18a+4c+sp) have shown two distinct endothermic peaks, at 689C and 770C. The mass loss between 470-720ºC suggests decomposition of poorly crystalline calcium carbonate, designated by aragonite and vaterite (Ramachandran 2001; Li 2003). The carbonates with better crystallinity decompose above 720C and are associated with the stable polymorph of calcite (Ramachandran 2001). This is confirmed with the earlier work that the carbonation of C-

S-H would produce poorly crystalline CaCO3 such as aragonite, and vaterite (Villain 2006). Calcium carbonates generated by carbonation of calcium hydroxide are more crystalline and decomposed at temperature higher than 720ºC (Moorehead 1986).

Carbonation products from C3S and C2S reaction may be decomposed either at low or high temperature, dependent on the duration of carbonation (Goto 1995). It is noted that concretes with air curing (18a+4c and 18a+4c+sp) show a broad exotherm prior to second endotherm. This exotherm is thought to be due to the energy release during a crystalline growth of calcite accompanied by a reduction of surface energy (Moorehead 1986). Based on TG/DTG analysis, it is possible to quantify the hydration and carbonation products. Water loss due to dehydration between 105ºC-420ºC is associated with hydration products such as C-S-H and C-A-H. Water loss due to dehydration in the range of 420ºC-470ºC represents calcium hydroxide. The carbonates decomposed between 470ºC-720ºC and 720ºC-950ºC are corresponding to poorly crystalline (aragonite and vaterite) and well crystalline (calcite) polymorphs respectively. The results are shown in Fig. 4.6. With only a 4-hour carbonation scheme followed by 20 hours of hydration, the freshly carbonated concrete (0a+4c) exhibited the highest quantity of

– 72 – hydration products. The carbonated concretes with initial air curing showed less hydration products because some of them were consumed by the successive 4-hour carbonation. The water spray seemed to be effective in promoting subsequent hydration after carbonation. Concrete with water compensation (18a+4c+sp) had seen an increase in hydration by 1.8% if compared to concrete with no water spray (18a+4c). Relatively, the hydrated reference displayed less hydration products in 24 hours which confirmed early carbonation was an accelerated curing. CH formation is another indicator of the degree of hydration. The sprayed and freshly carbonated concrete showed the highest amount of CH, followed by the hydrated reference. Concretes with initial air curing (18a+4c) contained much less CH content due to carbonation after 18-hour hydration. It is possible that, even with 1.64% loss in the range between 420-470ºC, CH is not detected by XRD because of the presence of amorphous forms of CH (Berger 1973). It was noted that the sequence of hydration and carbonation could have great effect on reaction products. CO2 loss due to decarbonation is also shown in Fig. 4.6. The reference showed some loss possibly due to decarbonation of carbonates that were present in as-received cement. For carbonation-induced calcium carbonates, 70-75% of the CO2 loss occurred below 720C. The poorly crystalline calcium carbonates were clearly dominant in carbonated concrete at age of one day. The initial air curing of 18 hours allowed concretes to gain more carbonation reaction products and thus more strength. DTG curves of concretes after 28 days subsequent hydration are plotted in Fig. 4.7. The absence of endothermic peaks between 110C-120C was indicative of an increase in stability of hydration products after 28 days. It was noted that, after 28 days of hydration, the endotherm at 450C appeared in all concretes, which signified the formation of CH. However, it was not evidenced by XRD peaks possibly due to the formation of amorphous calcium hydroxide (Berger 1973). In addition, the exotherm was no longer present after 28 days, and the decomposition temperature of carbonates was increased by 15C to 25C. This shift in decomposition temperature implied a better crystallinity of the carbonates due to the transformation of the poorly crystalline, unstable

CaCO3, such as aragonite and vaterite, to the more crystalline, stable calcite. Fig. 4.8 summarizes hydration and carbonation products of concretes at 28 days. While the sealed hydrated reference produced significantly more hydration products in

– 73 – 28-day subsequent hydration, carbonated concretes experienced only a slight increase which was not proportional to the strength gain in Fig. 4.2. Water compensation played a critical role. It promoted hydration products in carbonated concrete (18a+4c+sp) and thus the strength. It was also evident that subsequent hydration could occur after carbonation. The highest CH loss was seen in sprayed and freshly carbonated concrete followed by the reference. Anhydrous cement and adequate water supply ensured a suitable environment for the development of calcium hydroxide. Fig. 4.8 also displays the CO2 loss at two temperature ranges. After 28 days, the CO2 loss in carbonated concretes with air curing (18a+4c) due to decarbonation of poorly crystalline carbonates was reduced to 52-60%. In other words, more well crystalline phases such as calcite formed during subsequent hydration. In the case of the freshly carbonated concrete (0a+4c), the poorly crystalline carbonates was reduced to 13% of the total carbonates after 28 days. Comparison of Fig. 4.8 with Fig. 4.6 suggested that phase transformation did occur in subsequent hydration.

Table 4.6 summarizes CH and CaCO3 content in terms of cement. In order to obtain an average, 2 samples were used per batch. While the mass loss between 420C- 470C is associated with the water in CH, that between 470C-950C is coupled with

CO2 in calcium carbonate of all polymorphs. Accordingly, the amount of calcium hydroxide produced during 1 day and 28 days reached 13% in carbonated concrete with water compensation (18a+4c+sp), hydrated reference, and fresh carbonated concrete (0a+4c). The presence of water played a critical role in the development of CH after carbonation. The CaCO3 content in carbonated concrete in Table 4.6 included that which originally existed and that carbonation-produced. It was 28-30% in immediate carbonation and 59-67% in carbonation after 18-hour air curing. While in immediate carbonation (0a+4c), carbon dioxide reacted with C2S and C3S creating C-S-H and

CaCO3, carbonation after air curing (18a+4c) allowed the reaction to occur not only to anhydrous but to hydration products, mainly CH and C-S-H as well (Young 1974; Goodbrake 1979b). However, considering the significant amounts of solids produced in carbonation, the compressive strength is not enhanced proportionally. Carbonated concrete contains more CH and CaCO3 as carbonation products that fill up the voids produced by air curing, but the binding ability of carbonation products seems not to be greater. Therefore, the strength of carbonated concretes is not considerably increased.

– 74 – Nevertheless, the durability of concrete can be significantly enhanced due to the reduction of calcium hydroxide and the precipitation of calcium carbonates on surface (Rostami 2011).

SEM analysis SEM tests were performed to study the effect of carbonation on morphology of concrete. Microstructure of typical hydrated concrete is shown in Fig. 4.9. Ettringite needles were clearly seen after 28 days. In case of the freshly carbonated concrete (0a+4c), Fig. 4.10 shows the aggregate-cement interface. Ettringite needles were also detected in a grainy shape and embedded in C-S-H. Even with a low degree of carbonation, the microstructure was much denser in carbonated concrete than the hydrated reference. It was difficult to distinguish C-S-H and CaCO3 through their morphology. Fig. 4.11 and 4.12 show the micrographs of highly carbonated concrete (18a+4c) at the age of 1 and 28 days respectively. Pebble and flake-like grains of 1-2 m in size were identified on cement particles. The effect of water compensation on the microstructure of concrete after 28 days subsequent hydration is shown in Fig. 4.13 and 4.14. Two different areas in the concrete sample are considered. The SEM micrographs show that the microstructure is typical of amorphous. Water compensation produced concrete (18a+4c+sp) with more C-S-H in subsequent hydration.

It is apparent that the amorphous phase that intermingles C-S-H with CaCO3 dominates the microstructure of carbonated concretes with initial air curing. Effort has been made to identify this phase. Table 4.7 shows the mass lost between 105C-200C, poorly hydrated C-S-H, (DH1), 200ºC-420ºC, well hydrated C-S-H/C-A-H, (DH2), 470-

720ºC, poorly crystalline CaCO3, (DC1), and 720-950ºC, well crystalline CaCO3, (DC2). In the hydrated reference (0a), 28 days of hydration lead to an increase in DH1 and DH2, which leads to an increase in compressive strength. The carbonated samples exhibit a different behavior. During the 28-day hydration period, at room temperature, poorly crystalline CaCO3 polymorphs tend to recrystallize and transform into crystalline calcite. It is evidenced by the decrease in DC1 and increase in DC2. The same is seen in subsequent C-S-H formation. There is a transformation from the poorly hydrated to the

– 75 – well hydrated. DH1/DC1 ratios are shown in Table 4.7. They represent the ratios of poorly hydrated C-S-H to poorly crystalline CaCO3. It is noted that these ratios are almost constant for the concretes with air curing during hydration. They are about 0.35 for air cured carbonated concrete and 0.38 for that with water compensation. Such a constancy can only mean that these poorly crystalline carbonates and poorly hydrated C-S-H are that of one component. This component is possibly amorphous calcium silicate hydrocarbonate. This was also observed in a study on carbonation of β-C2S system (Goto 1995). Among the air cured carbonated concretes, it is observed that the higher the sum of the mass loss between 105-200ºC and 470-720ºC, the higher the compressive strength.

CONCLUSIONS The microstructure of carbonation-cured lightweight concretes has been studied. The reaction products due to carbonation are identified and quantified. Based on the findings of this study, the following conclusions may be drawn:

1. With the help of initial air curing, a CO2 uptake of about 23% in terms of cement was achieved in a lightweight concrete with expanded slag aggregates. It was corresponding to a carbonation degree of 46%. Water compensation after carbonation seemed to be critical to make carbonated concrete comparable to hydration references. The process was best suited to precast products without steel reinforcement and can be utilized to improve durability and sequester carbon dioxide for emission reduction. 2. Carbonation reaction products in concrete were characterized. A method was developed to determine the cement content in concrete powder sample through thermal analysis. It would allow hydration and carbonation products of concrete to be expressed in terms of cement content. 3. The carbonation-induced carbonates could be divided into poorly and well crystalline polymorphs. It was found that the poorly crystalline phases decreased from 70-75% to 50-60% after 28 days subsequent hydration. It was apparent that phase transformation occured during hydration. The dominant polymorph of calcium carbonate was calcite, which was more stable.

– 76 – 4. The coexistence of carbonation and hydration products was observed. The ratio of low temperature hydrates to low temperature carbonates (DH1/DC1) was indicative that these two phases were of one component. The constancy of this ratio suggested the possible presence of an amorphous calcium silicate hydrocarbonate. Although more reaction solids were produced in carbonated concretes, the binding ability of calcium silicate hydrocarbonates didn’t seem to be stronger than that of calcium silicate hydrates. However, more solids might imply a better resistance to permeation. The complete analysis of TG/DTG, SEM, and XRD indicated that the carbonation curing consumed calcium hydroxide, calcium silicate hydrates, and calcium silicates to produce calcium carbonates in different polymorphs and calcium silicate hydrocarbonates.

ACKNOWLEDGEMENTS The authors gratefully acknowledge the support by the Natural Science and Engineering Research Council (NSERC) of Canada, Canadian Concrete Masonry Producers Association (CCMPA), and Lafarge Canada.

REFERENCES ASTM (2004b). "Standard Test Methods for Sampling and Testing Concrete Masonry Units and Related Units." ASTM International C140.

Berger, R. L., and McGregor, J. D. (1973). "Effect of temperature and water-solid ratio on growth of Ca(OH)2 crystals formed during hydration of Ca3SiO5." Journal of American Ceramic Society 56(2): 73-79.

Chang, C., and Chen, J. (2006). "The experimental investigation of concrete carbonation depth." Cement and Concrete Research 36: 1760-1767.

Collier, N. C., Sharp, J. H., Milestone, N. B., Hill, J., and Godfrey, I. H. (2008). "The influence of water removal techniques on the composition and microstructure of hardened cement pastes." Cement and Concrete Research 38: 737-744.

El-Hassan, H., Shao, Y., and Ghouleh, Z. (2012a). "Effect of initial curing on carbonation of lightweight concrete masonry units." American Concrete Institute Journal.

– 77 – Goodbrake, C. J., Young, J. F., and Berger, R. L. (1979b). "Reaction of hydraulic calcium silicates with carbon dioxide and water." Journal of the American Ceramic Society 62(9-10): 488-491.

Goto, S., Suenaga, K., and Kado, T. (1995). "Calcium silicate carbonation products." Journal of the American Ceramic Society 78(11): 2867-2872.

Huffman, E. W. D. (1977). "Performance of a new automatic carbon dioxide coulometer." Microchemical Journal 22: 567-573.

Hyvert, N., Sellier, A., and Duprat, F. (2010). "Dependency of C-S-H carbonation rate on CO2 pressure to explain transition from accelerated tests to natural carbonation." Cement and Concrete Research 40: 1582-1589.

Klemm, W. A., and Berger, R. L. (1972). "Accelerated curing of cementitious systems by carbon dioxide - Part I - Portland Cement." Cement and Concrete Research 2: 567-576.

Li, D., Shi, H., Deng, J., and Xu, Y. (2003). "Study on the thermal decomposition kinetics of nano-sized calcium carbonates." Journal of Zhejiang University 4(3): 363-368.

Monkman, S., and Shao, Y. (2006). "Assessing the carbonation behavior of cementitious materials." Journal of Materials in Civil Engineering November/December: 768- 776.

Moorehead, D. R. (1986). "Cementation by the carbonation of hydrated lime." Cement and Concrete Research 16: 700-708.

Ramachandran, V. S., and Beaudoin, J. J. (2001). Handbook of Analytical Techniques in Concrete Science and Technology, William Andre Publishing/Noyes

Rostami, V., Shao, Y., and Boyd, A. J. (2011). "Durability of concrete pipes subjected to combined steam and carbonation curing." Construction and Building Materials 25: 3345-3355.

Rostami, V., Shao, Y., and Boyd, A. J. (2012). "Microstructure of cement paste subject to early carbonation curing." Cement and Concrete Research 42(1): 186-193.

Sawada, K. (1997). "The mechanism of crystallization and transformation of calcium carbonates." Pure and Appl. Chem. 69(5): 921-928.

Shao, Y., Mirza, M. S., and Wu, X. (2006). "CO2 sequestration using calcium-silicate concrete." Canadian Journal of Civil Engineering 33: 776-784.

– 78 – Shtepenko, O., Hills, C., and Brough, A. (2006). "The effect of carbon dioxide on beta- dicalcium silicate and Portland cement." Chemical Engineering Journal 118: 107- 118.

Toennies, H. T. (1960). "Artificial carbonation of concrete masonry untis." American Concrete Institute Journal(56-42): 737-755.

Toennies, H. T., and Shideler, J. J. (1963). "Plant drying and carbonation of concrete block - NCMA-PCA cooperative program." American Concrete Institute Journal 60(33): 617-632.

Villain, G., and Platret, G. (2006). "Two experimental methods to determine carbonation profiles in concrete." American Concrete Institute Materials Journal 103(4): 265- 271.

Wagh, A. S., Singh, D., and Knox, L. J. (1995). "Lab studying greenhouse effect on concrete setting." Concrete International 17(4): 41.

Young, J. F., Berger, R. L., Breese, J. (1974). "Accelerated Curing of Compacted Calcium Silicate Mortars on Exposure to CO2." Journal of the American Ceramic Society 57(9): 394-397.

– 79 – TABLES Table 4.1: Mixture proportion of slab samples

Constituent Mass (g) Mass (kg/m3) Percent (%) Ordinary Portland Cement 88 241 13 Water 35 96 5 SSD Expanded Slag 554 1502 82 Concrete Sample 677 1839 100

Table 4.2: Chemical composition of the cement

Constituent (%)

CaO SiO2 Al2O3 Fe2O3 MgO Na2O K2O SO3 CO2 63.1 19.8 4.9 2.0 2.0 0.9 - 3.8 1.72

Table 4.3: CO2 content and uptake from concrete samples Furnace method (%) Mass gain Average CO2 uptake Sample method (%) 1 day 28 days due to carbonation (%) 0a (Ref) 0.0 1.7 ± 0.2 1.9 ± 0.4 0.0 0a + 4c 7.5 ± 1.8 8.9 ± 0.6 11.1 ± 0.7 8.0 18a + 4c 24.2 ± 0.4 24.0 ± 1.2 27.4 ± 1.3 24.0 18a + 4c + sp 25.7 ± 1.2 24.1 ± 1.2 27.7 ± 1.3 24.6

Table 4.4: Cement content (CC) in powder samples

Test Furnace Uptake (%) TG/DTG Uptake (%) Sample Date Mc (g) (ΔCO2)C (g) Mp (mg) (ΔCO2)P (mg) CC 0a (Ref) 1 day 12.18 0.06 121.6 3.53 0.75 0a + 4c 1 day 10.25 0.17 119.5 6.57 0.44 18a + 4c 1 day 11.02 0.37 120.2 13.12 0.42 18a + 4c + sp 1 day 10.76 0.39 120.9 14.63 0.44 0a (Ref) 28 day 11.24 0.06 119.2 2.27 0.45 0a + 4c 28 day 12.45 0.21 119.9 6.93 0.45 18a + 4c 28 day 10.78 0.38 120.4 12.90 0.40 18a + 4c + sp 28 day 10.31 0.39 120.8 14.66 0.42

– 80 – Table 4.5: CO2 content and uptake from powder samples

CO2 Content (%) CO2 Uptake (%) Sample TG 1-day TG 28-day HCl 28-day Average 0a (Ref) 3.9 ± 0.5 4.3 ± 0.4 4.9 ± 0.5 0.00 0a + 4c 12.5 ± 0.9 12.9 ± 0.8 13.6 ± 0.9 8.65 18a + 4c 26.0 ± 1.2 26.8 ± 1.3 27.3 ± 1.3 22.33 18a + 4c + sp 27.5 ± 1.3 28.9 ± 1.4 27.9 ± 1.4 23.41

Table 4.6: Total calcium hydroxide and calcium carbonate content (%)

Sample # 0a 0a + 4c 18a + 4c 18a + 4c + sp 1-day 5.54 8.50 4.11 6.73 Ca(OH)2 by TG 28-day 13.18 13.71 11.19 13.71 1-day 8.86 28.41 59.09 62.50 CaCO3 by TG 28-day 9.77 29.32 60.91 65.68 1-day 14.40 36.91 63.20 69.23 Ca(OH)2 + CaCO3 28-day 22.95 43.03 72.10 79.39

Table 4.7: Dehydration mass and decarbonation mass

DH1a (%) DH2b (%) DC1c (%) DC2d (%) DH1/DC1

Sample 1-day 28-day 1-day 28-day 1-day 28-day 1-day 28-day 1-day 28-day

0a (Ref) 4.81 6.39 2.51 6.39 0.00 0.00 0.00 0.00 n/a n/a 0a + 4c 7.99 5.31 3.45 5.98 5.72 1.10 2.91 7.51 1.40 4.82 18a + 4c 5.60 3.95 3.55 5.92 16.05 11.67 6.05 10.83 0.35 0.34

18a + 4c + sp 6.52 5.69 4.50 7.12 17.60 14.37 6.06 9.20 0.37 0.40 DH1 (105-200ºC), poorly hydrated C-S-H DH2 (200-420ºC), well hydrated C-S-H and C-A-H DC1 (470-720ºC), poorly crystalline CaCO3 DC2 (720-950ºC), well crystalline CaCO3

– 81 – FIGURES

Fig. 4.1: Schematic of carbonation setup

22 1-Day Strength 20 18.2 28-Day Strength 17.9 18 16 15.1 15.3 14 12 10.5 9.9 10 8.8 8.6 8 6

4 Compressive Strength (MPa) CompressiveStrength 2 0 0a 0a + 4c 18a + 4c 18a + 4c + sp Samples Fig. 4.2: Compressive strength of concretes at 1-day and 28-day

– 82 –

Fig. 4.3: XRD patterns of concretes (1-day)

Fig. 4.4: XRD patterns of concretes (28-day)

– 83 –

Fig 4.5: DTG curves of concretes of 1-day

32 105C

16 11.45

12 11.02

Mass Loss(%) Mass

9.15

9.04

7.32 6.60

8 6.59 3.46

4 3.33

2.07

1.64

1.35

1.00 0.54 0 0a 0a + 4c 18a + 4c 18a + 4c + sp Sample # Fig. 4.6: Mass loss due to dehydration and decarbonation in concretes (1-day)

– 84 –

Fig. 4.7: DTG curves of concretes of 28-day

28 105C

16 15.34

12.81

12.78 11.41

12 11.29

9.87

9.78

Mass Loss (%) LossMass 8.09

8

4.77

3.67

3.33 3.33 3.21

4 2.72 0.58 0 0a 0a + 4c 18a + 4c 18a + 4c + sp Sample # Fig. 4.8: Mass loss due to dehydration and decarbonation in concretes (28-day)

– 85 –

Fig. 4.9: SEM micrograph of hydrated reference concrete (0a) after 28 days

Fig. 4.10: SEM micrograph of carbonated concrete (0a+4c) after 28 days

– 86 –

Fig. 4.11: SEM micrograph of carbonated concrete (18a+4c) after 1 day

Fig. 4.12: SEM micrograph of carbonated concrete (18a+4c) after 28 days

– 87 –

Fig. 4.13: SEM micrograph of carbonated concrete (18a+4c+sp) after 28 days

Fig. 4.14: SEM micrograph of carbonated concrete (18a+4c+sp) after 28 days

– 88 – Chapter 5

—

STATIC CARBONATION OF LIGHTWEIGHT PLC CONCRETE MASONRY

UNITS

PREFACE Previous chapters presented a detailed study on carbonation curing of lightweight concrete masonry units with Ordinary Portland Cement (OPC). The feasibility of applying the developed carbonation curing scheme to concrete masonry units with Portland limestone cement to replace the traditional steam curing technique was examined in this chapter. Portland Limestone Cement (PLC) is produced with a partial replacement of cement clinker with limestone. The resulting cement offers the production of environmentally friendly concrete by reducing CO2 emissions with less cement used. It was imperative to analyze the carbonation behavior of PLC, which can make PLC concrete greener by having a higher net gain in carbon emission reduction. The chapter aims at investigating the effect of initial air curing and carbonation curing on performance of PLC concrete masonry units and the possibility of replacing OPC with PLC for dry mix concrete. A similar early carbonation procedure to that utilized in

Chapter 3 is employed. CO2 uptake, compressive strength, and pH values are characterized to assess the performance of the carbonated PLC concrete in comparison to the hydrated and steamed counterparts.

– 89 – INTRODUCTION Previous chapters have investigated the carbonation reaction of Ordinary Portland Cement (OPC) concrete in a carbon dioxide curing process. While OPC has been the binder in the majority of concrete structural components, Portland Limestone Cement (PLC) was introduced as a greener substitute by replacing 15% cement clinker with limestone. It has shown technical and economical benefits. The use of PLC concrete with 10-35% limestone was reported to be able to increase early strength, reduce water demand, reduce bleeding in concrete with low cement content, and improve workability (Moir 1997; Tsivilis 1998). It is also possible to obtain a cement including up to 30% limestone with an equivalent strength to that of Portland cement at lower production costs per ton of cement (Baron 1987). Portland Limestone Cement (PLC) is produced by intergrinding Portland cement clinker, limestone, and calcium sulfate. Although limestone is a thermodynamically stable compound, when added to cement, it may affect hydration, carbonation, and compressive strength of concrete. It is first proposed that the fine limestone particles act as nucleation sites, thus increasing the rate of hydration of the calcium silicates and calcium hydroxide at early ages and possibly improving the distribution of hydrates

(Soroka 1977; Bonavetti 2003). Second, CaCO3 will react chemically with aluminate phases to form carboaluminate phases (Matschei 2007). Finally, because limestone is softer than clinker, it will achieve a finer particle size when interground, producing an improved particle size distribution and improving particle packing (Tsivilis 2002). Past research investigated the hydration of PLC during 28 days up to 18 months. The studies mostly focused on the relationship between compressive strength and the amount of added limestone. Limestone additions of up to 5% increased the early-age strength as a combined result of improved particle packing (Sprung 1991), increase in the rate of cement hydration (Vuk 2001; Bonavetti 2003), and early production of calcium carboaluminate (Voglis 2005). Even with 5% limestone blended with cement opposed to being interground, the strength was relatively unaffected (Livesey 1991; Schmidt 1992; Matthews 1994). A 12-month study by Hawkins reported a strength decrease with addition of limestone (0-8%) (Hawkins 1986).

– 90 – Limestone additions above 15% had a more severe effect on the compressive strength. Cement blended with 15% limestone or more is labeled as Portland limestone cement (CSA 2008). Limestone additions of 10-50% acted as a diluent and either improved the early strength or did not affect it, but a strength reduction of 8-19% was noticed after 28 days (Pera 1999; Bonavetti 2003). Others have studied replacements of more than 25% limestone, which resulted in much more significant strength reductions (Matthews 1994; Pera 1999; Dhir 2007). Comparable compressive strengths were only observed when PLC with 15% limestone was ground to 511 m2/kg as compared to 303 m2/kg in OPC (Voglis 2005). With higher limestone replacements, compressive strength was proportionally diminished. The current investigation aims to understand the processes of carbonation and hydration when PLC is used as a binding material. Currently, CSA A3000 allows the blending of up to 15% limestone in General Use Type 1 Portland cement, while European standard EN 197-1 identifies two types of cements containing 6-20% and 21-35% respectively (ECS 2000; CSA 2008). Major North American cement producers, such as Lafarge North America and Holcim Canada, have already been blending 3-5% limestone in their OPC and up to 15% in PLC. The static carbonation procedure used in Chapter 3 with OPC concrete will be followed to examine the effect of initial air curing on CO2 uptake and strength gain at 1 and 28 days. The pH of concrete so produced is also determined after 28 days. The purpose of the study is to investigate if PLC can be used in masonry block production using carbonation curing treatment. Therefore, 20-cm PLC concrete blocks will be fabricated and cured by carbonation. There is no similar work reported. The challenge is, with reduced reactive Portland cement content, if PLC concretes could still show comparable carbonation behavior and physical performance in comparison with OPC counterparts.

EXPERIMENTAL INVESTIGATIONS PLC concrete sample preparation Slab samples representing the 38-mm web of a 20-cm lightweigh PLC concrete masonry unit were used to study the effect of process parameters. The mixture proportion of commercial CMU was used with Portland limestone cement (PLC) as binder. PLC

– 91 – used in this project is a commercial product manufactured and marketed by Holcim Canada. About 13-15% limestone is interground with OPC clinker. The chemical composition of PLC is shown in Table 5.1. CO2 content in as-received PLC is about 6.97% which is equivalent to a limestone content of 15.84%. The cement is considered as a high limestone cement and is, therefore, labeled as Portland limestone cement. To make early strength comparable to OPC, PLC is ground finer. The fineness of PLC is 500 m2/kg in comparison to 390 m2/kg in OPC. Expanded slag aggregates were used, ranging in size from 0.2 to 6 mm with a dry bulk density of 951 kg/m3 and an internal water content of 5% (El-Hassan 2012a). Table 5.2 summarizes the mix design of the laboratory slab and CMU samples. By mass, it included a water-to-cement ratio of 0.40, an aggregate-to-cement ratio of 6.23, and cement content of 0.13. With water in the aggregates considered as part of the hydration water, the water to cement ratio reached 0.71. For the slabs, the raw materials were mixed in a pan mixer, cast into 127 x 76 x 38 mm mold, and then compact formed using a vibrating hammer to simulate the industry production of CMU. PLC concrete was demolded right after casting for initial curing due to the dry mix. For 20-cm CMU blocks, materials were mixed in a mechanical drum mixer and compact formed by a manual block machine. They were typical 20-cm lightweight CMU blocks with web thickness ranging from 25 to 33 mm and a density of 1839 kg/m3.

Curing procedures Table 5.3 summarizes the different curing procedures. Steam cured batches 1-4, hydrated batches 5-9, and carbonation cured batches 10-16 were prepared to compare carbonation with steam curing and hydration. Steam curing took place in a steam cooker for a period of 4 hours with maximum temperature of 80C and relative humidity of 95%. Initial curing periods of 0, 4, 6, and 8 hours were applied prior to steam for comparison with carbonation. Hydration was carried out in similar initial curing conditions but specimens were then placed in a sealed bag for subsequent hydration. Carbonation curing required a preconditioning. It was accomplished by initial air curing. It was performed on fresh PLC concrete in an environmental chamber at a relative humidity of 50% and a temperature of 25C. The duration of initial air curing was varied among 0, 4, 6, 8, and

– 92 – 18 hours to study its effect on the physical properties of PLC concrete. The purpose of initial curing was to remove surface free water and produce enough capillaries to ease the penetration of CO2 into the sample. Immediate carbonation of fresh PLC concrete served as reference and was performed with no air curing (0a+4c). Initial air curing of 18 hours was to simulate an overnight curing and was likely the longest presetting duration accepted by commercial production. Fig. 5.1 shows the static carbonation curing setup. PLC concrete samples after initial air curing were placed in a sealed chamber which was then vacuumed to about 0.7 bars and filled with carbon dioxide gas to an absolute pressure of 1 bar. The chamber was placed on a digital balance to obtain the mass curve of PLC concrete during carbonation. The carbonation duration varied from 2 to 4 hours.

Water loss, CO2 uptake, and compressive strength provided a basis to evaluate the effect of initial curing and carbonation curing. To compensate for the water loss during the curing process, water spray was applied to Batch 15 (18a+4c+sp) to restore the original water content and examine its effect on subsequent hydration after carbonation. The spray was stopped when the surface was saturated. It took a few days for all lost water to be compensated (El-Hassan 2012a). The subsequent hydration was carried out in a sealed bag at a relative humidity of 80±5% and room temperature (24±1ºC) until 24 hours and 28 days for strength tests. The best combination of initial curing and carbonation curing from slab tests was selected for 20-cm block production.

Internal relative humidity measurement As in Chapter 3, initial air curing was employed to remove the surface free water and produce capillaries to facilitate the penetration of carbon dioxide gas. To monitor distribution of internal relative humidity of PLC concrete during the period of initial curing, a Sensiron EK-H4 relative humidity meter was employed. Three humidity probes were placed in a concrete slab respectively at 5 mm from top (T), 5 mm from bottom (B), and in the middle (M). Fig. 5.2 presents a schematic of the slab samples with one probe inserted at each depth.

– 93 – CO2 uptake estimation In order to measure the degree of carbonation, three methods were utilized for the estimation of CO2 uptake: mass gain, mass curve, and thermal decomposition (furnace) analysis.

Mass gain method estimated CO2 uptake in PLC concrete by calculating the mass change of samples before and after carbonation (Eq. 5.1). Carbonation-induced water loss was collected by absorbent paper and added to the final mass. By treating the system as a closed system, it was imperative to include the evaporated water, which was initially inside the samples prior to carbonation. Since not all evaporated water, especially the vapor, can be collected, mass gain method underestimated the uptake and was therefore conservative.

CO2 uptake (%) = (5.1)

Mass curve method was executed by placing the carbonation setup on a computer- connected digital balance, which was zeroed after vacuuming the chamber. A mass curve was recorded as mass versus time until the end of process at which time CO2 was released and the residual mass, M, was measured. The system was calibrated by repeating the tests using CO2-insensitive styrofoam samples of the same volume to obtain a second residual mass, m. The difference between M and m represented the CO2 uptake by concrete mass. Eq. 5.2 is the basis for the mass curve measurement. Data collected by mass gain and mass curve methods can be compared because they are obtained from two simultaneous measurements from the same process. M  m CO2 uptake (%) = (5.2) Mass of cement Thermal decomposition (furnace) analysis was also performed to estimate the amount of carbonates in PLC concrete. Instead of using a thermogravimetry device, which only allows a few micrograms of powder, a furnace of maximum temperature of 1100ºC was employed to test through-thickness concrete samples with mass range of 35- 70 g. Separation of paste from concrete was thus avoided. PLC concrete samples were ignited at 105C, 470C, and 950C to quantitatively measure the evaporable water, bound water in hydration products, and carbon dioxide in carbonates respectively. The temperature range convention was established based on the DTG curves that associated

– 94 – temperature ranges 105ºC-470ºC and 470ºC-950ºC to hydration and carbonation products respectively. The mass at each temperature was recorded. CO2 content was calculated based on Eq. 5.3 (Johnson 2000).

CO2 content (%) = (5.3)

Performance evaluation Each batch cast for carbonation was compared to 2 control batches: normal hydration control and steam control. Although both controls followed the same initial hydration process, the former was then placed in a sealed plastic bag, while the latter was steam cured afterwards. The compressive strength was measured after 1 and 28 days for carbonated, steam cured, and hydration control samples following ASTM C140 (ASTM 2004b). Three samples per batch were tested and averaged. The fractured samples were preserved in an acetone solution to stop hydration. Acetone exchange with water to stop hydration was reported as the least damaging method to preserve the microstructure (Collier 2008). Prior to analysis, the samples were pre-dried overnight at 60oC. A simple and convenient method was employed for measuring the pH of pore solution of concretes (Heng 2004). An absorptive paper of 10×10 mm size was placed on the concrete surface and 100 μl of distilled water was added to it. After a 15-minute extraction, the hydroxyl ions came to equilibrium and an Extech PH110 pH meter, which had a flat sensor at its tip, was used to measure the pH of the cut layers of the samples. To ensure consistency of results, the pH readings of three random spots on the cut surface were averaged and reported.

EXPERIMENTAL RESULTS AND DISCUSSION Effect of initial air curing on water content Initial air curing was performed in an environmental chamber at 25ºC and 50% RH on fresh concrete immediately after casting and demoulding. The moisture content in concrete decreased gradually as it tried to reach equilibrium with the controlled relative humidity of the environmental chamber. While the top and side surfaces were exposed to

– 95 – the drying environment, the bottom face was much less exposed as being supported by a bottom plate. To explore the possibility of applying carbonation curing to PLC CMU production, the entire curing process should not exceed 24 hours in comparison to current steam curing practice. The process window of initial curing was thus also limited. Therefore, initial air curing of 0, 4, 6, 8, and 18 hours was selected in this study. The water loss due to initial air curing was calculated based on Eq. 5.4 and plotted in Fig. 5.3.

Water Loss (%) = (5.4)

In Eq. 5.4, the total initial water mass includes the mixing water needed to obtain a water- to-cement ratio of 0.4 and the pre-existing water in the expanded slag aggregates (5% of the total slag mass). The water lost during initial curing was recorded by placing the samples inside the environmental chamber on top of a digital balance for a period of 14 days. Mass was recorded as a function of time. It was apparent that water loss was proportional to the duration of initial curing within 24 hours. The most significant loss occurred in first 4 hours, reaching 37% with slight difference than the 6-hour and 8-hour readings. However, curing of 18 hours increased the water loss to 51%, which was very close to that cured by 24 hours. Eighteen hours seemed to be the maximum possible time for initial curing and could be executed through an overnight shift. At 14 days, the water loss reached 74%. Similar results were obtained during initial air curing of OPC concrete samples with 32%, 51%, and 81% after 4 hours, 18 hours, and 14 days. Significant water loss reduced water content in PLC concrete, and was expected to reduce the internal relative humidity for better carbonation efficiency. Fig. 5.4 shows the corresponding relative humidity measurement up to 14 days at the same initial curing conditions. Probe T measured RH of concrete 5 mm from the top surface, probe C at the center, and probe B 5 mm from the bottom surface as seen in the schematic of Fig. 5.2. Comparison of Fig. 5.4 with Fig. 5.3 shows that water loss of 51% at 24 hours does not considerably decrease the relative humidity. It is indicative that moisture movement inside PLC concrete occurs during the initial air curing. Nevertheless, free water removal was still crucial in making room for CO2 gas diffusion and calcium carbonate precipitation.

– 96 – Effect of initial curing on carbonation reaction

CO2 uptake describes the degree of carbonation in PLC concrete. One of two possible carbonation reactions may occur: The presence of limestone particles may induce a nucleation effect and promote the growth of more calcium carbonates in the PLC concrete sample. On the other hand, less amount of carbonation reactants present in the cement might lead to a decrease in uptake. Fig. 5.5 shows water loss due to exothermic carbonation reaction and the CO2 uptake by mass gain method (Eq. 5.1). The data was presented in two groups. First group included the first 5 batches with constant carbonation time of 4 hours and varied initial curing of 0, 4, 6, 8 and 18 hours to study the effect of initial air curing. The second group involved the last two batches with constant initial curing of 18 hours and varied carbonation duration of 2 and 4 hours to investigate the effect of carbonation time. Water loss due to carbonation curing is calculated based on Eq. 5.5:

Water Loss (%) = (5.5)

For the first group of 4-hour carbonation, immediate carbonation with no initial curing resulted in a CO2 uptake of 6.1% with a carbonation water loss of 0.5%. It was indicative of a low degree of reaction. After 4, 6, and 8 hour initial air curing, water loss due to initial curing was in the range of 37-39%. Nevertheless, their CO2 uptakes were different at 16.1, 17.4, and 18.3% with carbonation water loss at 4.9, 5.6, and 6.5% respectively. For initial curing of 18 hours, water loss in initial curing reached 51%, leading to a much lower water loss due to carbonation. Its CO2 uptake, nonetheless, was not much higher. It was conclusive that initial air curing reduced free water and made room for gas diffusion and calcium carbonate precipitation. On the other hand, OPC concrete carbonation showed slightly higher degree of reactivity in Chapter 3 with uptake reaching 24% with similar initial air curing conditions. In the second group, where initial curing was fixed at 18 hours, 4-hour carbonated sample resulted in a CO2 uptake of 18.3% and carbonation water loss of 4.2%. On the other hand, the 2-hour carbonated sample reached an uptake of 17% associated with 3.3% water loss during carbonation. Obviously, longer carbonation duration resulted in a higher carbonation degree and more water release due to the exothermic reaction.

– 97 – To verify the CO2 uptake by mass gain method (Eq. 5.1), mass curves were obtained. Five batches were compared in Fig. 5.6 with initial air curing of 4, 6, 8 or 18 hours and carbonation curing of 2 or 4 hours. During the 4-hour carbonation process, 60- 70% of the reaction occurred in the first hour and 80-90% in the second hour. The carbonation reaction efficiency decreased progressively with time due to moisture saturation on concrete surfaces and inside the chamber. Thermal decomposition of concrete pieces after strength tests was also used to quantify the amount of calcium carbonates in PLC concrete. The results of twelve batches after 28 days in different curing conditions are summarized in Fig. 5.7. Mass loss was categorized into three components: evaporable water (up to 105ºC), bound water in hydration products (between 105ºC and 470ºC) and carbon dioxide (between 470ºC and 950ºC). It was noted that 4-hour carbonated concrete without initial curing had the lowest uptake but high bound water. Higher bound water indicated more hydration products. Other samples, as 18a+4c, had higher uptake, which was associated with high carbonates, and low bound water content. It was concluded that the hydration products and carbonation products coexisted in one sample. Water compensation through spraying was used to promote subsequent hydration after carbonation. Comparing the hydrated samples (0a, 4a, 6a, 8a, and 18a) indicated that initial air curing reduced the amount of bound water. This was due to water loss during air curing which was crucial for subsequent hydration. As for the evaporated water, more CO2 uptake resulted in less evaporable water, as the exothermic carbonation reaction released free water. Carbon dioxide content in Fig. 5.7 was calculated based on total concrete mass used in furnace analysis. CO2 uptake was estimated based on cement mass by dividing the value by the cement content (13%). Thus, the resulting value was compared to the other two methods.

Table 5.4 shows the CO2 uptake of all batches using three methods: mass gain, mass curve, and furnace analysis. For each batch, mass gain was average of at least three samples and mass curve recorded mass increase of an entire batch. These two independent measurements were directly comparable as they were taken from the same process simultaneously. CO2 content by furnace analysis was performed on fractured PLC concrete chunks and served as a third check. Carbonation of 4 hours was capable of storing 6% CO2 without initial curing and 17% CO2 as average with initial curing

– 98 – between 4 and 18 hours. CO2 uptake reached 17% by 2-hour carbonation and 18.3% by 4-hour carbonation after 18-hour initial curing. The degree of carbonation was considered lower than that in OPC concrete samples in Chapter 3 with an uptake for OPC reaching 23-24%. This is likely attributed to the replacement of cement by 15% limestone. The comparison of thermal decomposition analysis with the other two mass gain methods showed an additional 6-7% CO2 content. This was due to the presence of 13% added limestone, which decomposed alongside the produced carbonates when heated to

950ºC. This original CO2 content from added limestone was subtracted from total CO2 content of carbonated concretes to gain the absolute value due solely to carbonation reaction. The average CO2 uptake by three methods is also presented in Table 5.4.

Compressive strength The effect of immediate carbonation and steam curing on fresh PLC concrete compressive strength was studied through Fig. 5.8. After 4 hours of hydration, the reference (0a) did not present much strength and reached 1.6 MPa. On the other hand, while carbonation (0a+4c) enhanced the compressive strength by 275% to 4.4 MPa, the steam cured sample (0a+4s) was slightly more effective and exceeded 5 MPa. The compressive strength was measured at 1 and 28 days after casting. Fig. 5.9 shows the strength variation among carbonated, steam, and hydrated specimens. The batches 0a, 4a, 6a, and 8a represent initial air curing time followed by 4-hour carbonation curing and subsequent hydration to 24 hours. For instance, the 24-hour strength of carbonated PLC concrete 8a+4c was tested after initial air curing of 8 hours, carbonation of 4 hours, and subsequent hydration in a sealed plastic bag for 12 hours. Hydration reference for each batch also underwent different initial air curing before sealed for normal curing. When comparing the results of hydrated batches with different initial air curing, the batch without initial air curing (0a) turned out to have the highest one-day strength (7 MPa) because there was no water loss during the curing. Carbonated samples with the same curing time showed better compressive strength than the hydrated control and comparable to the steam control. However, the steam samples that have been air cured prior to steam curing show a decrease in strength due to water release during air curing.

– 99 – Concretes with 18-hour air curing and 4-hour carbonation were tested for compressive strength and plotted in Fig. 5.10. Air curing was proven beneficial to enhance CO2 uptake while maintaining a comparable compressive strength. However, when the sample was not carbonated (18a), the void produced by water release was not filled with carbonation products and hydration was incomplete, which explained the lag in compressive strength. In comparison to 4a, 6a, and 8a, hydrated concrete 18a showed higher compressive strength. With longer air curing, the latter concrete hardened faster than the remaining samples, resulting in a higher 24-hour strength. Water compensation after carbonation (18a+4c+sp) had enhanced strength by 30% due to the formation of hydration products. Fig. 5.11 and 5.12 present the compressive strength after 28 days. The highest compressive strength of 4-hour carbonated concrete reached 14.3 MPa with water compensation (18a+4c+sp). It was directly comparable to the best hydration reference of 14.1 MPa with no air curing (0a) and better than steam strength. However, unlike at 1 day, the compressive strength of 18a could not compare to the remaining hydrated concretes due to more water loss during initial air curing, which is essential for subsequent hydration. The 28-day compressive strengths of hydrated and carbonated OPC concretes were also obtained. They were 18.2 MPa for hydration reference (0a) and 17.9 MPa for carbonation with water compensation (18a+4c-sp). In comparison to OPC concrete subjected to same process, PLC concrete exhibited a strength 20% lower owing to a lower degree of hydration and carbonation reaction. pH measurement The pH values of carbonated, steam cured, and hydration control samples are shown in Table 5.5. Carbonation is known to reduce pH of concrete pore solution. This phenomenon can be noticed in comparing the carbonated and control samples. In all cases, pH values of all carbonated concretes were in the range of 9-10.5, while pH values of hydrated references were greater than 12. Without exposure to carbon dioxide gas, the steam-cured samples presented pHs similar to that of the hydrated control. The pH reading of hydration reference was lower than 13.0. It was due to the non-destructive test

– 100 – method used in this study. It was difficult to produce representative pore solution by extraction because of porous nature of block samples. The percent reduction is therefore more relevant. The pH of PLC concrete after 2 to 4 hr carbonation was reduced by 16.5- 26.8%. The pH obtained in Table 5.5 is an average value of the whole sample, considering chunks from the surface and the core. Studies carried out in Chapter 3 described the loss in alkalinity in concrete due to carbonation. The pH of freshly carbonated sample 0a+4c was 10.5 compared to 12.2 and 12.6 for the steam and hydrated counterparts respectively. This pattern was visible throughout the batches with 4, 6, and 8 hours of air curing. However, higher CO2 uptake specimens, associated with 18 hours of air curing and 4 hours to 4 days of carbonation, showed an even larger decrease in pH to 9.5 compared to 12.5 for the 18-hour hydrated reference. This phenomenon was due to a higher degree of carbonation reaction, which consumed calcium hydroxide, the alkaline compound in concrete, and became less alkaline.

Static carbonation of 20-cm PLC CMU Static carbonation of 20-cm full-size PLC CMU was also carried out using the optimized process parameters. The procedure included 18-hour initial air curing, 4-hour carbonation, and water compensation (18a+4c+sp). Hydration reference (0a) was also tested at the same time. Standard 20-cm CMU blocks were made in the laboratory using a manual block machine (Fig. 5.13) and carbonated in a setup similar to that shown in Fig. 5.1. The hydrated reference was kept in a fog room with 100% relative humidity until strength test. The carbonated sample was air cured for 18 hours in an environmental chamber at 50% relative humidity and 25ºC. While the water loss due to initial curing ranged between 45-50%, similar to that of slab samples, water lost during carbonation was only 5%.

The CO2 uptake was measured using mass gain formula after carbonation and is presented in Table 5.6. Based on cement content, the 4-hour static carbonation curing blocks reached an uptake of 17%, which was close to carbonated slabs ranging between 17-18%. Their carbonation degree was about 34%. The carbon storage capacity of PLC concrete produced by this process was excellent.

– 101 – The compressive strength was measured after 1 day and 28 days. To ensure load distribution on the block, the blocks were capped with a white capping cement prior testing. Fig. 5.14 shows the obtained strengths. One-day strength was, in the 4-hour carbonated sample (18a+4c+sp), obtained after 18 hours of initial curing, 4 hours of static carbonation and 2 hours of subsequent hydration. The carbonated CMU blocks were 20% higher than the strength of the hydrated reference in one day. This was associated to the formation of carbonates at a 17% CO2 uptake. During curing, 50% of the water was lost, but was restored after carbonation by water spray. The final strength of water compensated 20-cm CMU blocks was nearly equivalent to the hydration reference. Therefore, CMU with PLC as binder can be produced by carbonation process to replace steam and normal hydration and achieve high early strength, equivalent late strength, superior CO2 uptake capacity, and reduce carbon emissions during cement production. The so-produced blocks are truly green products and the process should show environmental, economical, and technical benefits (Tsivilis 2002).

CONCLUSIONS The carbonation behavior of PLC concrete with 15% limestone powder was studied. The CO2 uptake, compressive strength, and water loss were compared among steam-cured, hydrated, and carbonated PLC concretes. Based on the findings of this study, the following conclusions may be drawn:

1. Initial curing had a considerable effect on the CO2 uptake. Freshly cast samples were carbonated for 4 hours and reached an uptake of 6% by cement mass, while 18-hour air cured samples followed by 4-hour carbonation obtained 18% uptake. The air curing process is vital for maximizing uptake as it provides a passage for

CO2 to penetrate the sample by removing the capillary water. 2. Initial curing affected compressive strength. The compressive strength was reduced by 5-10% due to the loss of water during air curing and carbonation, as subsequent hydration is hindered without sufficient amount of water. A wetting technique was suggested to replace the lost water and improved the compressive strength of the carbonated samples by 10-30%.

– 102 – 3. A comparison between OPC and PLC compressive strengths showed a 20-30% decrease in 1 and 28-day strength in PLC. Substitution of cement clinker, which represents the reactants in the hydration reaction, produced concrete with less hydration products. It is also proposed that the crushed limestone absorbed part of the mixing water that is required for subsequent hydration. 4. The pH values of carbonated PLC concretes are reduced due to high degree of carbonation reaction. This reduction could be detrimental to reinforced concrete due to the consumption of the alkaline CH phase that protects the reinforcement from corrosion. However, it does not affect the service performance of concrete masonry unit applications.

5. PLC concretes have shown lower CO2 uptake and compressive strength in comparison to OPC concrete in masonry unit applications with both hydration and carbonation curing. This is possibly attributed to 15% replacement of cement by limestone and porous nature of concrete masonry unit products. The higher Blaine fineness of PLC does not help improve the strength in CMU application. 6. The use of Portland limestone cement in place of ordinary Portland cement has significant economical and environmental benefits. It reduces carbon emission

and energy consumption in cement production and promotes net gain in CO2 utilization and sequestration. To improve the strength of PLC concrete, a ternary material system including silica fume could be beneficial.

ACKNOWLEDGEMENTS The authors gratefully acknowledge the support by the Natural Science and Engineering Research Council (NSERC) of Canada, Canadian Concrete Masonry Producers Association (CCMPA), and Lafarge Canada.

REFERENCES ASTM (2004b). "Standard Test Methods for Sampling and Testing Concrete Masonry Units and Related Units." ASTM International C140.

Baron, J., and Douvre, C. (1987). "Technical and economical aspects of the use of limestone filler additions in cement." World Cement 18(3): 100-104.

– 103 –

Bonavetti, V., Donza, H., Menendez, G., Cabrera, O., Irassar, E. F. (2003). "Limestone filler cement in low w/c concrete: A rational use of energy." Cement and Concrete Research 33(6): 865-871.

Collier, N. C., Sharp, J. H., Milestone, N. B., Hill, J., and Godfrey, I. H. (2008). "The influence of water removal techniques on the composition and microstructure of hardened cement pastes." Cement and Concrete Research 38: 737-744.

CSA (2008). Cementitious materials compendium, CSA. CSA-A3000.

Dhir, R. K., Limbachiya, M. C., McCarthy, M. J., and Chaipanich, A. (2007). "Evaluation of portland limestone cements for use in concrete construction." Materials and Structures 40: 459-473.

ECS (2000). Cement: composition, specifications, and conformity criteria, Part 1:Common cements. EN 197-1, European Committee for Standardization. EN/TC51/WG 6.

El-Hassan, H., Shao, Y., and Ghouleh, Z. (2012a). "Effect of initial curing on carbonation of lightweight concrete masonry units." American Concrete Institute Journal.

Hawkins, P. (1986). "Personal Communication to R. E. Gebhardt."

Heng, M., and Murata, K. (2004). "Aging of concrete buildings and determining the pH value on the surface of concrete by using a handy semi-conductive pH-meter": 1087-1090.

Johnson, D. (2000). "Accelerated carbonation of waste calcium silicate materials." Society of Chemical Industry 10.

Livesey, P. (1991). "Performance of limestone-filled cements." Blended Cements in Construction: 1-15.

Matschei, T., Lothenbach, B., and Glasser, F. P. (2007). "The role of calcium carbonate in cement hydration." Cement and Concrete Research 37(4): 551-558.

Matthews, J. D. (1994). "Performance of limestone filler cement concrete." In Euro- Cements –Impact of ENV 197 on Concrete Construction: 113-147.

Moir, G., and Kelham, S. (1997). "Developments in manufacture and use of Portland limestone cement." American Concrete Institute Journal 172: 797-819.

Pera, J., Husson, S., Guilhot, B. (1999). "Influence of finely ground limestone on cement hydration." Cement and Concrete Composites 21(2): 99-105.

– 104 – Schmidt, M. (1992). "Cement with interground additives - Capabilities and environmental relief, Part 1." Zement-Kalk-Gips 45(4): 87-92.

Soroka, I., and Setter, N. (1977). "The effect of fillers on strength of cement mortars." Cement and Concrete Research 7(4): 449-456.

Sprung, V. S., and Siebel, E. (1991). "Assessment of the suitability of limestone for producing portland limestone cement." ZKG International 3: 43-48.

Tsivilis, S., Chaniotakis, E., Kakali, G., and Batis, G. (2002). "An analysis of the properties of Portland limestone cements and concretes." Cement and Concrete Composites 24(3): 371-378.

Tsivilis, S., Kakali, G.,Chaniotakis, E., and Souvaridou, A. (1998). "A study of the hydration of Portland limestone cement by means of TG." Journal of Thermal Analysis 52: 863-870.

Voglis, N., Kakali, G., Chaniotakis, E., and Tsivilis, S. (2005). "Portland-limestone cements. Their properties and hydration compared to those of other composite cements." Cement and Concrete Composites 27: 191-196.

Vuk, T., Tinta, V., Gabrovšek, R. and Kauĉiĉ, V. (2001). "The effects of limestone addition, clinker type and fineness on properties of Portland cement." Cement and Concrete Research 31(1): 135-139.

– 105 – TABLES Table 5.1: Chemical composition of PLC

Constituent (%)

CaO SiO2 Al2O3 Fe2O3 MgO Na2O K2O SO3 CO2 59.8 21.6 4.5 2.7 1.7 0.8 - 3.2 6.97

Table 5.2: Mixture proportion of PLC concrete

Slab (g) CMU (g) Mass (kg/m3) Percent (%) Portland Limestone Cement 88 1967 241 13 Water 35 787 96 5 SSD Expanded Slag 554 12251 1502 82 Concrete Sample 677 15004 1839 100

Table 5.3: Curing conditions of PLC concrete

Subsequent Initial air curing Steam curing Carbonation curing Hydration

RH T t RH T t t Water t

Batch # Condition (%) (°C) (hours) (%) (°C) (hours) (hours) Spray (g) (days)

1 0a + 4s - - 0 95±5 75±5 4 - - 28 2 4a + 4s 80±5 22±1 4 95±5 75± 5 4 - - 28 3 6a + 4s 80±5 22±1 6 95±5 75±5 4 - - 28 4 8a + 4s 80±5 22±1 8 95±5 75±5 4 - - 28 5 0a - - 0 - - - - - 28 6 4a 80±5 22±1 4 - - - - - 28 7 6a 80±5 22±1 6 - - - - - 28 8 8a 80±5 22±1 8 - - - - - 28 9 18a 80±5 22±1 18 - - - 28

10 0a + 4c - - 0 - - - 4 - 28 11 4a + 4c 50±1 25±0.2 4 - - - 4 - 28 12 6a + 4c 50±1 25±0.2 6 - - - 4 - 28 13 8a + 4c 50±1 25±0.2 8 - - - 4 - 28 14 18a + 4c 50±1 25±0.2 18 - - - 4 - 28 15 18a + 4c + sp 50±1 25±0.2 4 - - - 4 33±2 28

16 18a + 2c 50±1 25±0.2 18 - - - 2 - 28 a - Initial air curing; s - Steam curing; c - Carbonation; RH - Relative humidity; T - Temperature; t - Time

– 106 – Table 5.4: Comparison of CO2 uptake by three methods Average CO CO Uptake (%) CO Content (%) 2 2 2 uptake due to Condition Mass Gain Mass Curve Furnace Analysis carbonation (%) 0a (Ref) n/a n/a 6.52 ± 0.5 0 0a + 4c 6.08 ± 1.4 5.98 13.43 ± 0.6 6.32 4a + 4c 16.14 ± 1.8 15.82 22.57 ± 1.2 16.00 6a + 4c 17.38 ± 1.7 16.91 24.26 ± 1.3 17.34 8a + 4c 18.32 ± 2.0 17.99 25.11 ± 1.2 18.30 18a + 4c 18.34 ± 1.6 18.23 25.31 ± 1.1 18.45 18a + 4c + sp 18.29 ± 1.7 18.09 25.85 ± 1.3 18.57 18a + 2c 16.97 ± 1.6 16.87 24.25 ± 1.5 17.19

Table 5.5: pH measurement after 28 days % Reduction by Sample pH @ 28 days Carbonation 0a + 4c 10.54 ± 1.18 16.5 0a + 4s 12.18 ± 1.22 - 0a (Ref) 12.62 ± 1.06 - 4a + 4c 10.49 ± 0.78 17.7 4a + 4s 12.61 ± 1.01 - 4a 12.75 ± 0.95 - 6a + 4c 10.13 ± 0.87 20.2 6a + 4s 12.52 ± 1.45 - 6a 12.69 ± 1.23 - 8a + 4c 9.97 ± 1.12 20.6 8a + 4s 12.62 ± 1.67 - 8a 12.56 ± 0.91 - 18a + 4c 9.52 ± 1.04 23.9 18a + 2c 10.12 ± 0.87 19.1 18a 12.51 ± 1.12 -

– 107 – Table 5.6: CO2 uptake of PLC concrete blocks

CO2 uptake (%) Sample Block #1 Block #2 Block #3 Average 18a + 4c 16.65 17.98 17.54 17.39 ± 0.68 18a + 4c + sp 17.16 17.89 16.92 17.32 ± 0.51

– 108 – FIGURES

Fig. 5.1: Schematic of carbonation setup

Fig. 5.2: Schematic of measuring internal relative humidity

– 109 –

Fig. 5.3: Water loss during initial curing

Fig. 5.4: Relative humidity inside PLC sample during initial curing

– 110 – 25 Water Loss w.r.t Initial Water Mass

Carbon Uptake w.r.t Initial Cement Mass 20 18.3 18.3 17.4 17.0 16.1 15

10 Percent (%) Percent 6.5 6.1 5.6 4.9 4.2 5 3.3

0.5 0 0a + 4c 4a + 4c 6a + 4c 8a + 4c 18a + 4c 18a + 2c Batch

Fig. 5.5: Water loss due to carbonation and CO2 uptake (Eq. 5.1)

20 18.23 18 16.87 17.99 16.91 16 15.82 14 12

10 Uptake (%) Uptake 2 8 4a+4c 6 6a+4c % CO % 8a+4c 4 18a+4c 2 18a+2c 0 0 60 120 180 240 300 Carbonation Time (minutes) Fig. 5.6: Mass curves of carbonated PLC concretes

– 111 – 18a 0.89 5.46 7.41 18a + 2c 3.15 2.92 10.92 18a + 4c + sp 3.36 3.18 16.32 18a + 4c 3.29 2.42 12.86 8a 0.90 5.79 7.43 Evaporable Water 8a + 4c 3.26 2.80 10.64 Bound Water

Sample 6a 0.89 5.81 7.52 Carbon Dioxide 6a + 4c 3.15 3.84 10.35 4a 0.88 6.09 13.56 4a + 4c 2.93 3.90 9.48 0a (Ref) 0.85 9.08 17.74 7.77 14.32 0a + 4c 1.75

0 5 10 15 20 25 30 Mass loss (%)

Fig. 5.7: Thermal decomposition analysis after 28 days

9 4-hour hydration 8 4-hour carbonation 7 4-hour steam 6 5.2 5 4.4 4 3

2 1.6 Compressive Strength (MPa) CompressiveStrength 1 0 0a

Fig. 5.8: Compressive strength after 4 hours

– 112 –

16 4-hr carbonation+hydration till 1 day 14 1-day hydration 12 4 hr steam+hydration till 1 day

10 9.0 8.7 8.5 8.4 8.7 8.7 8.7 8.7

8

7.0

6.2

6.1 6.0 6

4 Day Compressive Strength (MPa) Strength Day Compressive

- 2 One 0 0a 4a 6a 8a Batch Fig. 5.9: Compressive strength after 1 day

12 Carbonated 10.2 10 Hydrated Reference 7.9 8 7.5 7.0 6.8 6.8

6

4

Day Compressive Strength (MPa) Strength DayCompressive 2 -

One 0 18a+4c 18a+2c 18a+4c+sp Batch Fig. 5.10: Effect of initial curing on 1-day compressive strength

– 113 – 22 4-hr carbonation+hydration till 28 days 20 1-day hydration 18 4 hr steam+hydration till 28 days

16 14.1

13.8

13.4

13.1 13.1

14 12.4

12.8

12.6

12.3

12.5

12.4 12.2 12 10 8 6

4

Day Compressive Strength (MPa) Strength DayCompressive -

28 2 0 0a 4a 6a 8a Batch Fig. 5.11: Compressive strength after 28 days

20 Carbonated 18 Hydrated Reference 16 14.3 14 12.8 11.8 11.5 12 11.4 11.4 10 8 6

4 Day Compressive Strength (MPa) Strength DayCompressive

- 2 28 0 18a+4c 18a+2c 18a+4c+sp Batch Fig. 5.12: Effect of initial curing on 28-day compressive strength

– 114 –

Fig. 5.13: Full-size 20-cm PLC concrete block

20 1-day 28-day 15 13.2 12.5 11.5

10 7.2 6.5 5.8

5 Compressive Strength (MPa) Compressive Strength

0 0a (Ref) 18a+4c 18a+4c+sp Batch Fig. 5.14: Compressive strength of PLC concrete blocks

– 115 – Chapter 6

—

REACTION PRODUCTS OF LIGHTWEIGHT PLC CONCRETE MASONRY UNITS

BY STATIC CARBONATION

PREFACE The physical performance and carbonation behavior of PLC concrete with 15% interground limestone was investigated in Chapter 5. A comparison of carbonated PLC concrete with hydrated counterparts showed that carbonation could produce higher early strength at 1-day and comparable late strength at 28 days if water compensation by surface spray was implemented immediately after carbonation. It was found that initial air curing seemed to be necessary to promote high carbonation degree. Of the initial curing of 4, 6, 8, and 18 hours, the best combination appeared to be a procedure by air curing of 18 hours, followed by carbonation curing of 4 hours, and immediate water spray for compensation. The best reference is the sealed hydration (0a) with no water loss during the entire curing process. Concretes underwent 4-hour carbonation preceded by 18-hour air curing (18a+4c) could uptake 18% CO2 based on cement content. It was corresponding to a 39% reaction efficiency. Water compensation through surface spray restored the lost water and enhanced the 28-day compressive strength to be comparable to the hydration reference. In this chapter, microstructure changes of three PLC concretes: hydration reference, carbonated concrete without water compensation and carbonated concrete with water compensation, were analyzed using XRD, TG/DTG, and SEM.

– 116 – INTRODUCTION Portland Limestone Cement (PLC) is produced by intergrinding Portland cement clinker, thermodynamically stable limestone, and calcium sulfate. The effect of initial air curing on the carbonation reaction and physical properties of PLC concrete was investigated in Chapter 5. In comparison to Chapter 3 which was involved with OPC concrete, PLC appeared to be less reactive in terms of carbonation efficiency and compressive strength development. It is, therefore, crucial to understand how the use of PLC instead of OPC alters the performance and carbonation reactivity of concrete masonry units. When added to cement, limestone may affect hydration, carbonation, and compressive strength of concrete. It was first proposed that the fine limestone particles act as nucleation sites, thus increasing the rate of hydration of the calcium silicates and calcium hydroxide at early ages and possibly improving the distribution of hydrates (Soroka 1977; Bonavetti 2003). Ramachandran (1986) has reported that with an increase in the fineness and amount of CaCO3, the hydration rate of C3S also increases. In addition, CaCO3 will react chemically with aluminate phases to form carboaluminate phases (Matschei 2007). Finally, because limestone is softer than clinker, it will achieve a finer particle size when interground, producing an improved particle size distribution and improving particle packing (Tsivilis 2002). XRD and SEM analysis have shown that the depth of carbonation increased during 18-month drying (at 20ºC and 60% relative humidity) when more than 19% of Portland cement clinker was replaced with limestone. However, with less than 15% replacement, weathering carbonation degree was not affected (Parrott 1996). Hydration products, as ettringite, formation have also been accelerated by the presence of CaCO3, while C-S-H incorporated significant amount of calcium carbonates into its structure and formed calcium silicocarbonated hydrates (Ramachandran 1986). The effect of limestone addition of up to 35% on the durability of concrete has been studied. With 15% or below addition, the limestone can positively affect water permeability and sorptivity of the concrete due to improved cement particle size distribution (Tsivilis 2003). The physical performance and carbonation behavior of PLC concrete with 15% interground limestone was investigated in Chapter 5. A comparison of carbonated PLC

– 117 – concrete with hydrated counterparts showed that carbonation could produce higher early strength at 1-day and comparable late strength at 28 days if water compensation by surface spray was implemented immediately after carbonation. It was found that initial air curing seemed to be necessary to promote high carbonation degree. Of the initial curing of 4, 6, 8 and 18 hours, the best combination appeared to be a procedure by air curing of 18 hours, followed by carbonation curing of 4 hours and immediate water spray for compensation. On the other hand, the best reference is the sealed hydration (0a) with no water loss during entire curing. Concretes that underwent 4-hour carbonation preceded by

18-hour air curing (18a+4c) had shown 18% CO2 uptake. Water compensation through surface spray restored the lost water and enhanced the 28-day compressive strength to be comparable to the hydration reference. In order to investigate the microstructure and characterize the reaction products of PLC concrete, the process parameters will be a basis for selecting the optimum batches. Static carbonation setup developed in Chapter 3-4 will be used to perform carbonation curing of PLC concrete (El-Hassan 2012b). Three batches out of those in Chapter 5 were selected (0a, 18a+4c, 18a+4c+sp). The reaction products of three PLC concretes, hydration reference, carbonated concrete without water compensation and carbonated concrete with water compensation, are characterized by XRD, SEM, and TG/DTG. It is vital to investigate whether PLC can replace OPC in manufacturing concrete masonry units.

EXPERIMENTAL INVESTIGATIONS PLC concrete sample preparation To simulate a typical web or face shell of a 20-cm lightweight PLC concrete masonry units, rectangular concrete samples 127 mm long, 76 mm wide, and 38 mm thick were prepared. PLC used in this project was a commercial product manufactured and marketed by Holcim Canada. About 13-15% limestone is interground with OPC clinker. The chemical composition of PLC is shown in Table 6.1. CO2 content in as- received PLC is about 6.97%, which is equivalent to a limestone content of 15.84%. The cement is considered as a high limestone cement and is, therefore, labeled as Portland

– 118 – limestone cement. To make early strength comparable to OPC, PLC is ground finer. The Blaine fineness of PLC is 500 m2/kg in comparison to 390 m2/kg in OPC. The as- received lightweight slag aggregates had a dry bulk density of 951 kg/m3 with a water content of 5% by mass at saturated surface dry (SSD). According to the sieve analysis in Chapter 3, the well-graded aggregates ranged in size between 0.2 and 6 mm (El-Hassan 2012a). The mixture proportion is shown in Table 6.2. The water to cement ratio (w/c) incorporated into the mix reached 0.71 including the added water (w/c = 0.4) and the water in the aggregates. The raw materials were mixed in a pan mixer and samples were compact formed using a vibrating hammer to simulate the industry production of CMU. PLC concrete was then demolded right after casting. Three curing schemes were investigated. Sealed hydration in a plastic bag served as reference (0a). Initial air curing of 18 hours in an environmental chamber of 50% relative humidity (RH) and 25ºC was applied to batch 18a+4c before carbonation. After initial air curing, 4-hour carbonation curing was performed. A schematic of the carbonation setup is shown in Fig. 6.1. Water compensation through spray was devised immediately after carbonation to restore all the lost water (18a+4c+sp). The spray was stopped when the surface was saturated. It took a few days for all lost water to be compensated (El-Hassan 2012a). The subsequent hydration was carried out in a sealed bag at a relative humidity of 80±5% and room temperature (24±1ºC) until 24 hours and 28 days for compressive strength tests. The fractured concrete samples were preserved in an acetone solution to stop hydration for microstructure analysis. Acetone exchange with water to stop hydration was reported as the least damaging method to preserve the microstructure (Collier 2008). Prior to analysis, the samples were pre-dried overnight at 60oC.

Method of determining cement content (CC) in powder samples To conduct microstructure analysis of concrete using TG, powder samples were required. Powder samples were obtained by hammer crushing of the concrete surface layer of 5-mm thick and then sieving to pass a 125-µm sieve to remove coarse aggregates. No attempts were made to separate the fine aggregates from the cement. Therefore, powder samples were a mix of paste and fine aggregates. To quantify carbonation and hydration products, cement content in powder samples has to be

– 119 – determined. A procedure, similar to that employed in Chapter 4, was developed to collect powder samples within 5-mm thick surface of PLC concrete and estimate the cement content in powder: 1) A 5-mm thick concrete surface layer was cut off by a saw after mechanical testing. Part of the surface concrete was crushed by hammer and then sieved to pass 125-µm to remove large size aggregates. This is the powder that will be used for analysis. TG test was then performed on the powder of about 120 mg to obtain

CO2 loss, (CO2)p, between 470ºC and 950ºC from powder mass of Mp. 2) To quantify the cement content in this powder sample, the other part of surface layer of the same concrete of about 10-12 g was ignited in a furnace between

470ºC and 950ºC for CO2 content, (CO2)c, out of a concrete mass of Mc. 3) Assuming that on the surface layer of 5-mm thick from the same concrete, the

CO2 content will be the same whether calculated from powder sample or concrete chunk, that is:

CO2 content (%) = = (6.1)

Since cement content in PLC concrete is known as 0.13, the cement content (CC) in powder sample can be determined by Eq. 6.1. This correction shall be carried out for each sample and will allow the expression of reaction products in terms of cement content.

Measurement of CO2 uptake in PLC concrete

In Chapter 5, CO2 uptake by PLC concrete through carbonation curing was quantified by three methods: mass gain method, mass curve method, and furnace decomposition method. Measurements were performed using fractured concrete samples of entire thickness to obtain an average. To assure the powder analysis by TG represents the properties of PLC concrete, CO2 content in powder samples was measured using TG method and compared with that in concrete by independent furnace decomposition method. A furnace of maximum temperature of 1100ºC was employed to test large concrete samples with mass range of 35-70 g. Separation of paste from concrete was thus avoided. PLC concrete samples were ignited up to 105C, 470C, and 950C to

– 120 – quantitatively measure the evaporable water, bound water in hydration products, and carbon dioxide in carbonates respectively. The temperature range convention was established based on the DTG curves that associated temperature ranges 105-420ºC and 470ºC-950ºC to hydration and carbonation products respectively. The mass at each temperature was recorded. The CO2 content was calculated based on Eq. 6.2.

CO2 content (%) = (6.2)

Thermogravimetry (TG) analysis was used to estimate the CO2 uptake in powder samples of 120 mg each. The powder was subjected to thermogravimetry analysis by employing a thermal analyzer (NETZSCH, TG 449 F3 Jupiter) with a resolution of 0.01 mg. The TG/DTG curves were thus obtained by plotting the mass loss between 25C and

950C at a heating rate of 10C/min. The mass loss between 470ºC-950ºC represents CO2 content in concrete. The entire TG curve will be used to quantify hydration and carbonation products.

As comparison, CO2 content of powder samples was also determined by coulometric titration in a hydrochloric acid (HCl) solution (Huffman 1977). Two samples were used to obtain an average. The powder sample was immersed in the acid solution in a pipette which was sealed from the opening by a cork. A thin plastic tube was connected at one end to the inside of the pipette and at the other end to a coulometer. The reaction between hydrochloric acid (HCl) and carbon-containing compound such as calcium carbonate led to the release of carbon dioxide gas. The coulometer measured the amount of carbon released, and ultimately the carbon dioxide by stoichiometric proportions.

Performance evaluation The carbonated PLC concretes were compared to a hydrated concrete reference with no initial air curing. The compressive strength was tested after 1 day and 28 days subsequent hydration following ASTM C140 (ASTM 2004b). Three rectangular specimens for each batch were tested and averaged with a compressive area of 127 mm x 38 mm. The same concrete powder as used in TG/DTG tests was analyzed under XRD. The powder was subjected to XRD analysis by means of a Philips PW1710 Powder

– 121 – Diffractometer (Cu, K radiation, X'celerator proportional detector, scan interval 10- 100, 0.02, and 0.5 seconds per step) in order to relatively quantify the amount of calcium hydroxide, calcium silicates, and calcium carbonates in carbonated and hydrated samples. A Hitachi S4000 with an EDAX Phoenix EDS microanalysis system was employed to perform high-vacuum SEM morphology analysis of fractured concrete in a backscattered electron (BSE) mode. The air-exposed surface of a slab sample was chiseled to obtain gravel-size (5-10 mm) granules, and then a thin gold palladium coating was applied to ensure conductivity during analysis. The micrographs were used to identify the hydration and carbonation products in PLC concrete.

EXPERIMENTAL RESULTS AND DISCUSSION

Cement content and CO2 content in concrete powder Cement content of concrete powder was determined using Eq 6.1. All reaction products will be expressed in terms of dry cement content. The cement content (CC) of different concrete batches as correction factors are presented in Table 6.3. Their CO2 contents based on dry cement are summarized in Table 6.4. The TG analysis of the hydrated reference detected approximately 7% of pre-existing carbonates, which were generally in good agreement with furnace analysis. CO2 content determined by TG method is compared with HCl titration and furnace analysis methods in Table 6.4. While TG and furnace analysis are performed based on a temperature range of 470ºC-950ºC, HCl titration acts as a calibration that is independent of temperature range selection. The obtained carbonates in the hydration reference are to be deducted from those of the carbonated samples to obtain the CO2 uptake. It can be noted that 28 days after casting, the samples recorded similar CO2 uptake as to 1 day, confirming permanent sequestration of CO2. In Table 6.3, the percent cement content (CC) was used in calculating the mass of dry cement in powder samples so that the comparison was on the same dry cement base. Therefore, the furnace results in Table 6.4 and Table 6.3 provided an additional comparison. The difference of the two tests was in that furnace tests in Table 6.4 employed concrete sample mass, 35-70 g each, representing a through-thickness average; while that in Table 6.3 used much smaller mass, 10 g each, representing 5-mm surface

– 122 – layer. The absolute CO2 uptake calculated using concrete part of Eq. 6.1 appeared to be very close to that by furnace analysis in Table 6.4. It was obvious that with initial air curing (18a+4c), the difference between surface layer and through-thickness average was relatively insignificant, leading to a higher degree through-thickness reaction. With the help of cement content, the powder analysis determined the reaction products of concrete.

Compressive strength Fig. 6.2 shows the 1-day and 28-day compressive strength of the three concretes. The 24-hour duration represents the initial air curing followed by 4-hour carbonation and subsequent hydration to 24 hours. For instance, sample 18a+4c underwent 18 hours of air curing followed by 4 hours of carbonation and 2 hours of subsequent hydration. At 1-day, it was obvious that carbonation was an accelerated hydration process with 13% higher strength in the carbonated concrete with respect to its hydrated counterpart. Within 2 hours of subsequent hydration, water compensation improved the compressive strength by 29% in comparison to the non-sprayed carbonated concrete, even though only part of the lost water was restored. After 28 days, the hydrated reference exhibited higher compressive strength than the carbonated (18a+4c) owing to the water loss in early curing. Therefore, water spray after carbonation was essential to restore the lost water. Accordingly, a strength increase of 12% was recorded after water compensation. The results indicate that carbonation followed by a water compensation is beneficial. Early carbonation does not hinder subsequent hydration. It is the water loss during early air curing and carbonation curing that reduces degree of hydration. With water compensation through surface spray, it is possible to make carbonation strength comparable to hydration counterpart.

Phase analysis The 1-day XRD patterns for carbonated and hydrated PLC concretes are plotted in Fig. 6.3. Calcium carbonate peaks including calcite, vaterite, and aragonite are dominant and strong throughout carbonated samples. In comparison to the hydrated reference (0a), higher carbonate peaks at 29.4º, 32º, and 43º 2θ are present in the carbonated batches. An intermix between hydration and carbonation products is noticed in the carbonated

– 123 – samples. Calcium hydroxide is distinctively identified at 18º 2θ in hydrated samples. However, with absence of this peak in the carbonated PLC concrete, it is conclusive that calcium hydroxide is consumed. In addition, anhydrous calcium silicates are consumed, evidenced by the decrease in peak intensity at 41.5º 2θ, which indicate the existence of

C3S. Ettringite has also been consumed. The lack of peaks at 52.5º 2θ after carbonation provide evidence of this phenomenon. Fig. 6.4 shows the XRD patterns of the same PLC concretes after 28 days subsequent hydration. In comparison to the patterns obtained after 1 day, the amount of vaterite and aragonite have not decreased significantly as was the case with OPC samples in Chapter 4 (El-Hassan 2012b). The hydrated reference does not differ between 1 and 28 days with higher peaks of calcium hydroxide as the sample is hydrated for 27 more days. Calcite peaks are identified at 26.8º and 29.4º 2θ in the hydrated reference, which are associated to the added limestone in the cement. While aragonite is suggested to form when the sample is allowed to dry out as in the case of 18a+4c (Goodbrake 1979b), vaterite forms due to the rapid carbonation reaction and high heat release (Bukowski 1979; Goto 1995).

TG/DTG analysis Fig. 6.5 is a plot of the DTG curves of the investigated batches at 1 day. According to Ramachandran and Beaudoin (2001), the mass loss along TG curves can be classified into 6 categories: Below 105ºC, the mass loss is associated with the evaporable water and possibly the poorly formed C-S-H. The water loss between 105C and 200C represents the bound water in low temperature C-S-H and ettringite, while the loss in the range of 200C-420C is associated to the bound water in well-formed hydration products including C-S-H and C-A-H. Bound water in calcium hydroxide is decomposed between

420ºC-470ºC. While CO2 in poorly crystalline calcium carbonates designated by aragonite and vaterite are decomposed between 470C-720C, that in well crystalline calcite lies in the range of 720C-950C (Ramachandran 2001; Li 2003). The DTG curves of PLC concretes were compared: Drops between 105C-420C were common among all samples, while those between 420C-470C were only present in the hydrated reference. Previous work suggested that this last loss was due to the dehydration of

– 124 – Ca(OH)2 (Chang 2006). This was evidenced by the presence of Ca(OH)2 peaks in the XRD patterns. After 470C, the carbonates formed in carbonated PLC concretes were decomposed in a narrow drop in the well crystalline range. Calcium carbonates generated by carbonation of calcium hydroxide were better crystalline and decomposed at temperature higher than 720ºC (Moorehead 1986). The reference hydration concrete also showed a similar, but less intense CaCO3 peak, due to added limestone. Fig. 6.6 shows detailed mass loss based on TG/DTG analysis. It can be noted, using the previously mentioned mass loss distribution at each temperature range, that the hydrated reference had the lowest quantity of C-S-H and C-A-H and exceeded by the carbonated counterparts. PLC concrete consisted of anhydrous calcium silicates prior to carbonation. Due to 4-hour carbonation, the calcium silicates were consumed to form C- S-H, as carbonation is an accelerated hydration process (Young 1974). This explains the increase in C-S-H and C-A-H contents after carbonation. As for the sprayed carbonated batch (18a+4c+sp), the 2 hours of subsequent hydration allowed further formation of C- S-H, which was detected by comparing to the non-sprayed sample (18a+4c). CH formation was a direct indication of the degree of hydration of any sample. The hydrated reference formed the highest amount of CH associated with no water loss during 1-day curing. The sprayed carbonated sample (18a+4c+sp) showed a comparable amount followed by its drier counterpart (18a+4c). It is possible that, even with 0.8% loss in the range between 420-470ºC, CH was not detected by XRD because of the presence of amorphous forms of CH (Berger 1973). A sufficient water supply promoted the hydration of unreacted calcium silicates in the carbonated PLC concrete. Beyond 470C, the hydrated reference showed some loss due to decarbonation of pre-existing carbonates.

The respective CO2 loss of carbonated concretes represented the CO2 content, and included the added limestone and produced carbonates. The mass loss between 470- 720ºC, representing the carbonates produced from carbonation of C-S-H, showed the least thermal stability, suggesting poorly crystalline calcium carbonate designated by aragonite and vaterite (Ramachandran 2001; Li 2003; Villain 2006). The carbonates with better crystallinity, such as calcite, required higher temperatures to decompose and were lost above 720C (Ramachandran 2001).

– 125 – Fig. 6.7 is a plot of the DTG curves after 28 days subsequent hydration. The absence of the endotherm in the PLC concretes between 120-130ºC was indicative of an increase in stability of hydration products after 28 days. The mass loss at 107ºC in the hydrated reference was a possible evidence of the formation of ettringite, which was also detected in the XRD patterns. While the hydration reference concrete showed a noticeable drop between 420-470C, the carbonated concretes did not show similar results. This endotherm was associated with the decomposition of CH, which was produced during the 28-day subsequent hydration. Unlike Chapter 4, which identified a peak shift due to transformation of poorly crystalline aragonite and vaterite into well crystalline calcite during subsequent hydration of OPC, this was not the case in PLC concrete (El-Hassan 2012b). Fig. 6.8 summarizes hydration and carbonation products of PLC concretes at 28 days. The major change between 1 and 28 days was the hydrated reference concrete, in which increase in mass loss between 105-470ºC was indicative of increase in hydration products. Sealed curing conditions in hydration reference prevented evaporation of water, which was essential for subsequent hydration. Two carbonated concretes exhibited similar behavior from 1 to 28 days, but the losses were higher due to hydration. The increase of C-S-H and C-A-H in hydrated reference concrete led to increase in compressive strength as seen in Fig 6.2. In the carbonated samples, the amounts of C-S-H and C-A-H did not differ between 1 and 28 days. It was suggestive that carbonation of PLC may slightly hinder the hydration of silicates and aluminates, thus affecting the compressive strength. In order to study the effect of subsequent hydration on the carbonates stability, the percentage of poorly crystalline and well crystalline carbonates were calculated. Table 6.5 summarizes the poorly crystalline (470-720ºC) and well crystalline (720-950ºC) proportions of carbonation-induced carbonates in each batch. After 1 day, 50-56% of the total 4-hour reaction carbonates were poorly crystalline, while 44-50% were well crystalline. After 28 days, similar results were reported. It was conclusive, evidenced by the DTG curves, that there was no transformation of unstable carbonates (aragonite and vaterite) into more stable polymorphs (calcite) in PLC concretes.

– 126 – Table 6.6 shows the CH and CaCO3 content in terms of cement. The mass loss between 420C-470C was associated to the water in CH, while that between 470C-

950C was attributed to the CO2 in calcium carbonate of all polymorphs. Accordingly, the amount of calcium hydroxide produced during 1 day and 28 days reached maximum in the hydrated reference. Water presence played a critical role in the development of CH. However, carbonated concrete lost a significant amount of CH even though the DTG curves showed no endotherm. The CaCO3 content in carbonated concrete in Table 6.6 included the originally existed and the carbonation-produced. It was 15-16% in the hydrated reference and 59-62% in carbonation after 18-hour air curing. Carbonation after air curing (18a+4c and 18a+4c+sp) allowed the reaction to occur not only to anhydrous silicates but to hydration products, mainly CH and C-S-H as well (Young 1974;

Goodbrake 1979b). The total solids consisting of CH and CaCO3 are shown in Table 6.6.

Even though, the carbonated concrete had four times as much solids from CH and CaCO3 than the hydrated reference, it did not contribute to the compressive strength as seen in Fig. 6.2. The carbonated concrete (18a+4c) reached a strength of 12.8 MPa after 28 days with 65% solids from CH and CaCO3, while the hydrated concrete (0a) achieved 14.1 MPa with only 20-25% solids. It appears that the hydration products had a more significant effect on the strength.

SEM analysis Fig. 6.9 shows the microstructure of a hydrated reference after 1 day. Ettringite are incorporated alongside C-S-H. This was evidenced by the XRD patterns of the hydrated reference. Previous research resulted in similar observations (Ramachandran 1986). The compressive strength obtained after 1 day was comparable to the carbonated concrete due to the dense structure of the hydrated PLC concrete. The microstructure of carbonated sample (18a+4c) after 1 and 28 days are presented in Fig. 6.10 and 6.11. It was observed that crystals have formed around ball- like shapes with dimensions on the order of 5 m and 1 m respectively. Such a porous microstructure after 28 days explains the lower compressive strength of the carbonated concrete, as more voids were present between the crystals. Water compensation after carbonation (18a+4c+sp) aimed to maximize CO2 uptake while maintaining a high

– 127 – compressive strength. Fig. 6.12, 6.13, and 6.14 show a relatively denser microstructure after 1 and 28 subsequent hydration with the formation of sharp closely packed crystals. As a result, the compressive strength exceeds that of the hydrated and other carbonated concrete at both durations. Examination of powder PLC in Fig. 6.15 showed that 10 m particles represent the ground limestone added to the cement in production phase. Larger irregular shapes (15-20 m) were also inspected and identified as a form of anhydrous calcium silicates, either C3S or C2S. Table 6.7 presents the results of a study on the possibility of the formation of calcium silicate hydrocarbonate. The mass loss distribution was classified: mass lost between 105C-200C as poorly hydrated C-S-H, (DH1), 200ºC-420ºC as well hydrated

C-S-H/C-A-H, (DH2), 470-720ºC as poorly crystalline CaCO3, (DC1), and 720-950ºC as well crystalline CaCO3, (DC2). In the hydrated reference (0a), 28 days of hydration led to an increase in DH1 and DH2, which contributed to an increase in compressive strength. The carbonated samples, however, exhibited a far more different phenomenon. While the hydration products (DH1 and DH2) showed a slight increase in percentage, the carbonates, whether poorly or well crystalline (DC1 or DC2) showed insignificant change. Due to early formed crystalline phases in carbonated PLC concrete, the carbonates did not become more crystalline during subsequent hydration. The obtained ratio of DH1/DC1 did not remain constant as in the case of OPC (El-Hassan 2012b). This showed that the reaction products were not correlated in any form and calcium silicate hydrocarbonate was not likely a well-defined reaction product in the carbonation of PLC concrete. However, the intermingled hydrates and carbonates could still serve as the binding matrix in carbonated PLC concrete.

Comparison with ordinary Portland cement

The comparison of OPC and PLC carbonated concretes includes CO2 uptake, strength development and microstructure changes. OPC uptake exceeded that of PLC by up to 5% for 18a+4c. The 15% cement substitution with limestone decreased the number of carbonation reactants, including C3S and C2S, and ultimately decreased the rate of reaction. Past research on the effect of 18-month drying (20ºC and 60% RH) on different

– 128 – cements had discussed that limestone addition was more beneficial for hydration rather than carbonation, and that at least 19% replacement would affect carbonation (Parrott 1996). The XRD patterns after 1 day for both types of carbonated PLC concrete identified vaterite, aragonite, and calcite as carbonation products. However, after 28 days, while OPC showed less amounts of aragonite and vaterite and more calcite, there was little to no change in PLC specimens. TG/DTG also showed that, in both OPC and PLC concretes ,C-S-H and C-A-H did not greatly develop after 28 days of hydration. On the other hand, the recorded mass loss at 420-470ºC, which was associated to CH decomposition, increased from 1 to 28 days proving that carbonation did not hinder the formation of CH. In addition, a pattern of transformation of carbonates to the more crystalline polymorphs with higher 28-day decomposition temperatures was found in OPC but not in PLC. It was conclusive that carbonates produced in OPC carbonation are of better crystallinity after 28 days. The SEM images presented an amorphous form of carbonates throughout the carbonated OPC concrete, while the PLC showed more crystalline microstructure. Further inspection and analysis of the formation of these micro crystals has led to the conclusion that the addition of fine limestone may have contributed to the growth of the crystals.

CONCLUSIONS The carbonation curing of PLC concrete was studied. The microstructure of carbonated PLC concrete was examined. Based on the findings of this study, the following conclusions may be drawn:

1. In the presence of initial air curing, a CO2 uptake of about 19% in terms of cement was achieved in a lightweight PLC concrete with expanded slag aggregates. It was corresponding to a carbonation degree of 38%. Compared to 23% in the carbonation of OPC concrete, a reaction efficiency of 46%, PLC showed lower degree of reactivity due to less amounts of reactants. 2. Carbonation consumed calcium hydroxide, calcium silicates, and ettringite to produce calcium carbonates in its three polymorphs. The XRD peaks show

– 129 – evidence of this phenomenon. Even though a mass loss between 420-470ºC was detected in the carbonated PLC concrete, it was not identified by XRD possibly due to its amorphous nature. 3. The carbonation-induced calcium carbonates are distinguished between poorly and well crystalline. For 18-hour air cured and 4-hour carbonated PLC concrete, 47-50% of the total carbonation-induced carbonates are poorly crystalline throughout the 28 days, while 50-53% remain well crystalline. Water compensated concrete resulted in 52-56% poorly crystalline carbonates with 44- 48% well crystalline during 28 days. It was conclusive that no phase transformation of carbonates exist in PLC concrete during hydration. 4. The coexistence of carbonation and hydration products was observed. The ratio of low temperature hydrates to low temperature carbonates (DH1/DC1) showed that the reaction products do not form a phase of calcium silicate hydrocarbonate. With crystalline carbonates forming early in the curing process, no phase

transformation of CaCO3 is noticed during subsequent hydration. The final binding matrix is a hybrid of hydrates and carbonates.

ACKNOWLEDGEMENTS The authors gratefully acknowledge the support by the Natural Science and Engineering Research Council (NSERC) of Canada, Canadian Concrete Masonry Producers Association (CCMPA), and Lafarge Canada.

REFERENCES ASTM (2004b). "Standard Test Methods for Sampling and Testing Concrete Masonry Units and Related Units." ASTM International C140.

Berger, R. L., and McGregor, J. D. (1973). "Effect of temperature and water-solid ratio on growth of Ca(OH)2 crystals formed during hydration of Ca3SiO5." Journal of American Ceramic Society 56(2): 73-79.

Bonavetti, V., Donza, H., Menendez, G., Cabrera, O., Irassar, E. F. (2003). "Limestone filler cement in low w/c concrete: A rational use of energy." Cement and Concrete Research 33(6): 865-871.

– 130 – Bukowski, J. M., and Berger, R. L. (1979). "Reactivity and strength development of CO2 activated non-hydraulic calcium silicates." Cement and Concrete Research 9(1): 57-68.

Chang, C., and Chen, J. (2006). "The experimental investigation of concrete carbonation depth." Cement and Concrete Research 36: 1760-1767.

Collier, N. C., Sharp, J. H., Milestone, N. B., Hill, J., and Godfrey, I. H. (2008). "The influence of water removal techniques on the composition and microstructure of hardened cement pastes." Cement and Concrete Research 38: 737-744.

El-Hassan, H., Shao, Y., and Ghouleh, Z. (2012a). "Effect of initial curing on carbonation of lightweight concrete masonry units." American Concrete Institute Journal.

El-Hassan, H., Shao, Y., and Ghouleh, Z. (2012b). "Reaction Products in Carbonation Cured Lightweight Concrete." ASCE Materials Journal.

Goodbrake, C. J., Young, J. F., and Berger, R. L. (1979b). "Reaction of hydraulic calcium silicates with carbon dioxide and water." Journal of the American Ceramic Society 62(9-10): 488-491.

Goto, S., Suenaga, K., and Kado, T. (1995). "Calcium silicate carbonation products." Journal of the American Ceramic Society 78(11): 2867-2872.

Huffman, E. W. D. (1977). "Performance of a new automatic carbon dioxide coulometer." Microchemical Journal 22: 567-573.

Li, D., Shi, H., Deng, J., and Xu, Y. (2003). "Study on the thermal decomposition kinetics of nano-sized calcium carbonates." Journal of Zhejiang University 4(3): 363-368.

Matschei, T., Lothenbach, B., and Glasser, F. P. (2007). "The role of calcium carbonate in cement hydration." Cement and Concrete Research 37(4): 551-558.

Moorehead, D. R. (1986). "Cementation by the carbonation of hydrated lime." Cement and Concrete Research 16: 700-708.

Parrott, L. J. (1996). "Some effects of cement and curing upon carbonation and reinforcement corrosion in concrete." Materials and Structures 29: 164-173.

Ramachandran, V. S., and Beaudoin, J. J. (2001). Handbook of Analytical Techniques in Concrete Science and Technology, William Andre Publishing/Noyes

Ramachandran, V. S., and Zhang, C. (1986). "Influence of CaCO3 on hydration and microstructural characteristics of tricalcium silicate." Cemento 83(3): 129-152.

– 131 – Soroka, I., and Setter, N. (1977). "The effect of fillers on strength of cement mortars." Cement and Concrete Research 7(4): 449-456.

Tsivilis, S., Chaniotakis, E., Kakali, G., and Batis, G. (2002). "An analysis of the properties of Portland limestone cements and concretes." Cement and Concrete Composites 24(3): 371-378.

Tsivilis, S., Tsantilas, J. (2003). "The permeability of Portland limestone cement concrete." Cement and Concrete Research 33(9): 1465-1471.

Villain, G., and Platret, G. (2006). "Two experimental methods to determine carbonation profiles in concrete." American Concrete Institute Materials Journal 103(4): 265- 271.

Young, J. F., Berger, R. L., Breese, J. (1974). "Accelerated Curing of Compacted Calcium Silicate Mortars on Exposure to CO2." Journal of the American Ceramic Society 57(9): 394-397.

– 132 – TABLES Table 6.1: Chemical composition of PLC

Constituent (%)

CaO SiO2 Al2O3 Fe2O3 MgO Na2O K2O SO3 CO2 59.8 21.6 4.5 2.7 1.7 0.8 - 3.2 6.97

Table 6.2: Mixture proportion of PLC concrete Slab CMU Mass Percent (g) (g) (kg/m3) (%) Portland Limestone Cement 88 1967 241 13 Water 35 787 96 5 SSD Expanded Slag 554 12251 1502 82 Concrete Sample 677 15004 1839 100

Table 6.3: Cement content (CC) in powder samples

Test Furnace Uptake (%) TG/DTG Uptake (%) Sample Date Mc (g) (ΔCO2)C (g) Mp (mg) (ΔCO2)P (mg) CC 0a - Ref 1 day 9.21 0.08 120.6 3.14 0.38 18a + 4c 1 day 10.54 0.37 119.9 8.32 0.26 18a + 4c + sp 1 day 11.05 0.39 121.0 8.08 0.25 0a - Ref 28 day 10.76 0.09 120.2 1.92 0.24 18a + 4c 28 day 10.32 0.35 121.5 7.20 0.23 18a + 4c + sp 28 day 9.89 0.34 120.4 9.02 0.28

Table 6.4: CO2 content in powder samples

CO Content (%) 2 Average CO2 Sample TG 1-day TG 28-day HCl 28-day Furnace 28-day Uptake (%) 0a (Ref) 6.91 6.62 7.01 6.52 0 18a + 4c 26.71 26.06 25.98 25.31 19.25 18a + 4c + sp 26.92 26.54 26.55 25.85 19.70

Table 6.5: Percentage of produced poorly and well crystalline calcium carbonates

Poorly Crystalline Well Crystalline Sample 1-day 28-day 1-day 28-day 18a+4c 50.5 47.4 49.5 52.6 18a + 4c + sp 56.1 52.7 43.9 47.3

– 133 – Table 6.6: Total calcium hydroxide and calcium carbonate content (%)

Sample 0a 18a + 4c 18a + 4c + sp 1-day 3.82 3.33 3.64 Ca(OH)2 by TG 28-day 10.21 7.59 8.02 1-day 15.70 60.70 61.18 CaCO3 by TG 28-day 15.05 59.23 60.32 1-day 19.53 64.03 64.83 Ca(OH)2 + CaCO3 28-day 25.25 66.82 68.34

Table 6.7: Dehydration mass and decarbonation mass of PLC concrete

DH1a (%) DH2b (%) DC1c (%) DC2d (%) DH1/DC1

Sample 1-day 28-day 1-day 28-day 1-day 28-day 1-day 28-day 1-day 28-day

0a (Ref) 3.40 4.88 3.19 3.43 0.00 0.00 0.00 0.00 n/a n/a 18a + 4c 3.70 3.96 4.47 4.57 9.99 9.22 9.81 10.22 0.370 0.429

18a + 4c + sp 4.19 4.54 5.48 5.25 11.23 10.49 8.78 9.43 0.373 0.432 DH1 (105-200ºC), poorly hydrated C-S-H DH2 (200-420ºC), well hydrated C-S-H and C-A-H DC1 (470-720ºC), poorly crystalline CaCO3 DC2 (720-950ºC), well crystalline CaCO3

– 134 – FIGURES

Fig. 6.1: Schematic of carbonation setup

20 1-day 28-day 14.3 15 14.1 12.8

10.2 10 7.9 7.0

5 Compressive Strength (MPa) Compressive Strength

0 0a (Ref) 18a+4c 18a+4c+sp Batch Fig. 6.2: Compressive strength of PLC concretes

– 135 –

Fig. 6.3: XRD patterns of PLC concretes after 1 day

Fig. 6.4: XRD patterns of PLC concretes after 28 days

– 136 –

Fig. 6.5: DTG curves of 1-day PLC concretes

28 105C

16 15.27

14.03 12.68

12 11.65

9.67

Mass Loss(%)Mass 8.17

8 6.59 4.04

4 2.87

0.93

0.89 0.81 0 0a 18a + 4c 18a + 4c + sp Sample # Fig. 6.6: Mass loss due to dehydration and decarbonation in 1-day PLC concretes

– 137 –

Fig. 6.7: DTG curves of 28-day PLC concretes

32 105C

16 14.47

13.19 12.87

12 12.07

9.79

Mass Loss(%) Mass

8.53 8.32

8

3.97 2.65

4 2.48

1.95 1.85

0 0a 18a + 4c 18a + 4c + sp Sample # Fig. 6.8: Mass loss due to dehydration and decarbonation in 28-day PLC concretes

– 138 –

Fig. 6.9: SEM micrograph of PLC hydrated reference (0a) after 1 day

Fig. 6.10: SEM micrograph of PLC carbonated concrete (18a+4c) after 1 day

– 139 –

Fig. 6.11: SEM micrograph of PLC carbonated concrete (18a+4c) after 28 days

Fig. 6.12: SEM micrograph of sprayed PLC carbonated concrete (18a+4c+sp) after 1 day

– 140 –

Fig. 6.13: SEM micrograph of sprayed PLC carbonated concrete (18a+4c+sp) after 28 days

Fig. 6.14: SEM micrograph of sprayed PLC carbonated concrete (18a+4c+sp) after 28 days

– 141 –

Fig. 6.15: SEM micrograph of PLC powder

– 142 – Chapter 7

—

DYNAMIC CARBONATION OF LIGHTWEIGHT OPC CONCRETE MASONRY

UNITS

PREFACE Previous chapters provided a detailed performance evaluation and microstructure analysis of carbonated OPC and PLC concrete employing a static carbonation system in comparison to hydrated and steam-cured counterparts. The results showed that CO2 uptake could be maximized by removing evaporable water and producing capillaries for carbon dioxide to penetrate and carbonation products to precipitate. However, with more than half of the water lost during curing, the strength of the carbonated concrete could not compare to the hydrated reference, and thus, water compensation through spray was devised to restore the lost water. The procedure of initial air curing, carbonation curing, and water compensation was required to maximize carbonation reactivity and strength gain. In this regard, a carbonation system was introduced to combine these requirements in a single cycle to be designated as the dynamic system. Based on the results of the previous chapters, OPC is used owing to a higher degree of carbonation and compressive strength. Steam and hydrated control batches were cast to compare the microstructure and properties at 1 day and 28 days after casting. Carbonation products were characterized using thermogravimetrical analysis, X-ray diffraction analysis, and scanning electron microscopy.

– 143 – INTRODUCTION The effect of early-age static carbonation on the properties and microstructure of OPC and PLC concrete masonry units was investigated in Chapters 3-6. One of the main objectives was to examine the effect of initial air curing on water loss, compressive strength, CO2 uptake, and microstructure changes. The findings showed that CO2 uptake could be maximized by removing evaporable water and producing capillaries for carbon dioxide to penetrate and carbonation products to precipitate. Because of the significant water loss due to the initial air curing at low relative humidity, the strength of carbonated concrete was lower than that of hydrated references. Therefore, surface water spray procedure immediately after carbonation was devised to promote subsequent hydration. It was recommended that a procedure, including initial air curing, carbonation curing, and water compensation, be followed to achieve maximum carbonation reactivity and strength gain (El-Hassan 2012a; El-Hassan 2012b). To achieve a CO2 uptake of 23% based on cement content, static carbonation system requires a total duration of 22 hours including 18 hours of initial air curing and 4 hours of carbonation. It would be economically beneficial if the two-step process, initial curing and carbonation curing, could be combined into one so that the total process duration of 22 hours could be shortened to a few hours and the production could be completed in one work shift of 8 hours. This process is considered as immediate carbonation after casting. Immediate carbonation is a chemical reaction between carbon dioxide and dicalcium silicate and tricalcium silicate, with the governing reactions shown by Eq. 7.1- 7.2. The carbonation reaction products are a hybrid of calcium-silicate-hydrate (C-S-H) and calcium carbonate (CaCO3). High early strength can be obtained within a few minutes to a few hours (Young 1974; Goodbrake 1979b).

C3S+(3-x)CO2+yH2O→CxSHy+(3-x)CaCO3 (7.1)

C2S+(2-x)CO2+yH2O→CxSHy+(2-x)CaCO3 (7.2) Most of the previous research on immediate carbonation after casting was carried out with fresh calcium silicate or Portland cement pastes and mortars with low water to cement ratio in a range of 0.15-0.36 and static compact forming (Young 1974; Monkman 2006). The water content before and during carbonation was found to be of utmost importance (Young et al. 1974). While insufficient water could not provide the aqueous

– 144 – environment required for ions dissolution, excess water could limit the reaction by blocking CO2 gas diffusion (Young 1974; Goodbrake 1979b). Because of this difficulty, the CO2 uptake of fresh cement paste or mortar was limited to 10-12% on cement basis (Berger 1972). Carbonation of fresh concrete was also investigated. Several patents were issued (Jones Jr. 1997; Soroushian 1999). The processes were developed mainly for rapid strength gain. No CO2 uptake and carbonation efficiency were reported. The purpose of this chapter is to examine if a dynamic carbonation system can be developed to combine initial air curing and carbonation curing to reduce the time required for the process and gain carbonation efficiency comparable to static systems. It is hoped that dynamic carbonation could remove the free water and inject carbon dioxide simultaneously. The gas circulation can be used to control relative humidity in the chamber and enhance carbonation efficiency. It is different from static carbonation in that the process window allows drying and carbonation to occur at the same time for maximized carbonation reaction. Because of the water loss during the dynamic carbonation, water compensation through surface water spray is still necessary for maximized subsequent hydration. A pseudo-dynamic system using flush-through was studied in earlier work (Young 1974). The results showed that flush-through system could reach similar carbonation degree as the static system. However, the flush-through system didn’t require a pressurized chamber. The proposed system in this chapter is different from flush-through concept in that it runs on a closed-loop circulation with in- situ drying capacity and varied pressure. The carbonation degree will be measured using thermal analysis. SEM and XRD analysis will also be performed to characterize and identify the location, morphology, and stability of the carbonation products developed in concrete.

EXPERIMENTAL INVESTIGATIONS Concrete sample preparation The controlled dynamic carbonation process was investigated using concrete slab samples 127 mm long, 76 mm wide, and 38 mm thick representing a typical web of a 20- cm concrete masonry unit (CMU). The samples were prepared according to the commercial block mix design using general use ordinary Portland cement (OPC) and

– 145 – lightweight expanded slag aggregates with saturated surface dry (SSD). The chemical composition of OPC is shown in Table 4.2 in Chapter 4. The dry bulk density of expanded slag aggregates was 951 kg/m3 and the water absorption at saturated surface dry was 5%. The well-graded aggregates had a size distribution between 0.2 and 6 mm according to the sieve analysis reported in Chapter 3 (El-Hassan 2012a). Table 7.1 summarizes the mixture proportion of the slab samples and concrete blocks. By mass, it includes a mixing water-to-cement ratio of 0.4, aggregate-to-cement ratio of 6.23, and cement content of 0.13. The total water-to-cement ratio in concrete reached 0.71 considering also the amount of water in aggregates (5%). Each rectangular concrete slab sample and 20-cm CMU block weighs approximately 680 grams and 15 kg respectively, with a density of 1839 kg/m3. The raw materials of slab samples were mixed in a pan mixer, and compact formed using a vibrating hammer to simulate the industry production of CMU. Because of the dry mix, concrete was demolded right after casting for initial curing. For 20-cm CMU blocks, materials were mixed in a mechanical drum mixer and compact formed by a manual block machine. They were typical 20-cm CMU blocks with face shell or web thickness ranging from 25 to 33 mm.

Curing procedures Table 7.2 summarizes the different curing procedures of the batches using slab samples. The first and second batches are normally hydrated and steam cured. They serve as a control to dynamic carbonation. The other 3 batches are carbonated in the dynamic system. Steam curing took place in a steam cooker for a period of 4 hours with maximum temperature of 80C and relative humidity (RH) of 95%. Initial sitting time of 2 hours at 24±1oC and relative humidity of 50±10% was applied prior to steam. Dynamic carbonation curing was carried out in a setup shown in Fig. 7.1. The samples were also initially placed for 2 hours in the room temperature 24C and RH of 50±10%. The initial sitting time could also reduce the surface free water and allow diffusion of carbon dioxide. Setting of 2 hours was to simulate the same time a block plant requires before steam curing or autoclaving a batch of concrete blocks. After initial sitting, the samples were placed in the sealed chamber in Fig. 7.1, which was then filled with carbon dioxide to an absolute pressure of 0.5 or 1 bar. A constant gas flow was maintained using a

– 146 – circulation pump. A desiccator was attached to the loop with a valve so that the system has the capacity to dry the gas if relative humidity of the gas was high. The carbonation duration investigated was 4 and 18 hours. The prolonged 18-hour carbonation represented the overnight operation. Water spray was also devised after 2-hour open-air curing and 4- hour dynamic carbonation to promote subsequent hydration in batch 2a+4c. In static carbonation, it was shown that surface water removal can improve the

CO2 uptake. During carbonation, much free water migrates from the core to the surface. This causes surface saturation and hinders the carbonation reaction. By maintaining a relatively dry surface throughout curing, the degree of carbonation can be maximized. Only at an optimum relative humidity can this mechanism be applied. In order to monitor the relative humidity while curing, an RH probe is installed inside the chamber as shown in Fig. 7.1. Water evaporation due to exothermic carbonation reaction increases the RH in chamber. The released water is thus removed by the desiccant material, and the relative humidity is maintained at desired level. This proposed curing system results in drying the sample while carbonating it.

CO2 uptake estimation Thermal analysis was utilized to estimate the degree of carbonation. The mass gain method, which was developed for static carbonation, was proven inaccurate in the dynamic system since water absorbed by desiccant material could not be measured precisely and significant amount of water vapor was also lost in circulation.

Two thermal analysis methods were employed in CO2 uptake estimate: furnace method and thermogravimetry (TG) method. A furnace of maximum temperature of

1100ºC was used to determine CO2 content in concrete chucks of 35-70g each without the need to separate the paste from the aggregates. The samples were chosen in such a way that the entire thickness of concrete was retained to provide an average throughout the thickness. Concrete chunk specimens were heated up to 105C, 470C, and 950C to quantitatively measure the evaporable water, bound water in hydration products, and carbon dioxide in carbonates respectively. The mass at each temperature was recorded.

The CO2 content in concrete can be calculated based on Eq. 7.3 (Johnson 2000).

– 147 –

CO2 content (%) = (7.3)

Thermogravimetry tests were performed with powder samples which were obtained by crushing the 5-mm surface layer of fractured concrete samples and then sieving to pass a 125-µm sieve. The powder was subjected to thermogravimetry analysis by employing a thermal analyzer (NETZSCH, TG 449 F3 Jupiter) with a resolution of 0.01 mg. The TG/DTG curves were thus obtained by plotting the mass loss between 25C and 950C at a heating rate of 10C/min. No attempts were made to separate the fine aggregates from the cement. Therefore, a method was developed to account for cement content in the powder samples.

Method of determining cement content (CC) in powder samples Powder samples are necessary for TG analysis to quantify carbonation and hydration products. A procedure was developed to collect powder samples within 5-mm thick surface of concrete and estimate the cement content in powder: 1) A 5-mm thick concrete surface layer was cut off by a saw after mechanical testing. Part of the surface concrete was crushed by a hammer and then sieved to pass 125-µm to remove large size aggregates. This powder is a mix of cement and aggregate and will be used for TG analysis. TG test was then performed on the

powder of about 120 mg to obtain CO2 loss between 470 and 950ºC from powder

mass of Mp. This CO2 loss is referred to as (CO2)p. 2) To quantify the cement content in this powder sample, the other part of surface layer of the same concrete of about 10-12 g was ignited in a furnace between 470

and 950ºC for CO2 content, (CO2)C, out of a concrete mass of Mc. 3) Assuming that on the surface layer of 5-mm thick from the same concrete, the

CO2 content will be the same whether calculated from powder sample or concrete chunk, that is:

CO2 content (%) = = (7.4)

– 148 – Since cement content in concrete is known as 0.13, the cement content (CC) in powder sample can be determined by Eq. 7.4. This correction shall be carried out for each sample and will allow the expression of reaction products in terms of cement content.

Performance evaluation In order to study the effect of carbonation, each batch produced for dynamic carbonation was compared to 2 control batches: normal hydration control and steam control. While the steam control underwent the same initial curing process after which they were steam-cured followed by sealed hydration to the test age, the hydrated control was placed in a sealed plastic bag after casting. The compressive strength was measured after 1 and 28 days subsequent hydration following ASTM C140 (ASTM 2004b). Acetone exchange with water was used to stop hydration before XRD and TG analysis. It was reported that acetone exchange was the least damaging method to preserve the microstructure (Collier 2008). Prior to analysis, the samples were pre-dried overnight at 60oC.

Powder samples similar to those used in the TG/DTG CO2 uptake estimation method were utilized for X-Ray Diffraction analysis. The XRD patterns were obtained at 1 and 28 days after casting by means of a Philips PW1710 Powder Diffractometer (Cu, K radiation, X'celerator proportional detector, scan interval 10-100, 0.02, and 0.5 seconds per step) in order to relatively quantify the amount of calcium hydroxide, calcium silicates, and calcium carbonates in carbonated and control batches. One and 28 days after casting, SEM was performed on the five batches. A gold/palladium coating was applied to the concrete samples collected from fracture surface to ensure conductivity during the experiment. The coated specimens were then analyzed using high-vacuum scanning electron microscopy in backscattered electron (BSE) mode employing a Hitachi S4000 microanalysis system.

EXPERIMENTAL RESULTS AND DISCUSSION Dynamic carbonation behavior Water content is an indication of carbonation behavior. Table 7.3 summarizes the water content change during the proposed curing procedure. In two hours of sitting,

– 149 – samples showed a water loss of 8-9%. After 4 hours of dynamic carbonation, carbonated samples (2a+4c and 2a+4c+sp) lost an additional 29-30%. In comparison, during 4-hour static carbonation, the water loss was only 3-6% due to a water-saturated atmosphere in the chamber (El-Hassan 2012a). This was also attributed to the significant water loss during initial air curing in the static carbonation system. However, in the dynamic carbonation system, the relative humidity was maintained at 53% using a desiccator, leading to a simultaneous drying and carbonation reaction. Water spray was performed on 4-hour carbonated concrete (2a+4c+sp) to compensate for the water loss and ensure a comparable subsequent hydration. The final water content after compensation was 100%. Prolonged dynamic carbonation of 18 hours was also investigated. Water loss reached as much as 46%. The dynamic system had proven effective in free water removal at early age.

CO2 uptake is the measurement of degree of carbonation and determined by furnace decomposition at 470-950ºC. CO2 uptakes were calculated by subtracting the original CO2 content in hydration reference from total CO2 content in carbonated concretes. Their CO2 uptake averages obtained from 1 day and 28 days are presented in Table 7.4. Two independent furnace tests are compared: one using large concrete chunk each of 35-70 g mass, representing a through-thickness average; another using smaller mass, 10-12 g each, representing 5-mm surface layer. It was clear that the surface layer contained higher CO2 content than the through-thickness average in 4-hour carbonation (2a+4c). With 18-hour carbonation (2a+18c), the difference between surface layer and through-thickness average was much smaller, indicating a higher degree of through- thickness reaction. Four-hour dynamic carbonation of fresh concrete reached a CO2 uptake of 12-14%. The prolonged reaction in the same system could increase the uptake rate to 20-22%. Both were based on dry cement content in concrete. The ability to dry the sample and remove surface water while carbonating resulted in notable carbonation reactivity. Static carbonation system could only reach an uptake of 8% with no initial air curing. The hydrated reference was sealed in a bag for 1 and 28 days. The associated 4% carbon dioxide content was possibly an indication of the presence of limestone in the as- received cement. Similar results were seen in the steam control. A comparison between 1

– 150 – day and 28 days showed a noticeable constancy in the results of each batch, suggesting that carbonation is a permanent sequestration process. To perform quantitative TG analysis, cement content in concrete powder samples shall be determined. It is accomplished by comparing the CO2 content in concrete chunk and in concrete powder, both of which were collected from 5-mm surface layer. In the calculation of cement content (CC) using Eq. 7.4, thermal analysis of concrete pieces of 10-12 g each was performed in a furnace, while thermogravimetry analysis (TG/DTG) was conducted with 120 mg of concrete powder for each. The cement content (CC) in the concrete powder was thus estimated using Eq. 7.4 and the results are presented in Table 7.5. The percent cement content of powder samples, ranging from 50-80%, will be used for quantitative TG analysis of reaction products.

Compressive strength The accelerated strength gain of concrete at the age of 6 hours after casting is shown in Fig. 7.2. In comparison to hydration reference of 2.2 MPa, carbonation strength reached 6.0 MPa, and steam strength 6.7 MPa. The carbonated sample reached 90% of the strength of the steam-cured at the same duration and more than 150% that of the reference. Early strength was significantly improved by accelerated curing. It seemed steam was slightly more effective. After 1 day and 28 days, the compressive strengths were tested and presented in Fig. 7.3. The 24-hour strength was used to evaluate the effect of carbonation duration. This compressive strength was measured after the complete curing process including initial sitting time, carbonation curing or steam curing, and subsequent hydration up to 24 hours. For instance, sample 2a+4c was tested after 2 hours of sitting, 4 hours of carbonation, and 18 hours of subsequent hydration. The hydration reference, however, was kept in a sealed plastic bag throughout the process. In comparison, 4-hour carbonation followed by water compensation (2a+4c+sp) resulted in the highest compressive strength due to the presence of sufficient water for subsequent hydration during the remaining 18 hours. In comparison to the concrete without compensation (2a+4c), the addition of curing water promoted a 24% increase in strength within 18 hours. Prolonged carbonation associated with high CO2 uptake resulted in a strength

– 151 – higher than both steam and hydrated references. The high carbon dioxide content would have filled the voids of evaporated water and increased the compressive strength. Both hydrated and steamed reference concretes showed similar compressive strength after 1 day. Carbonated concrete without water compensation (2a+4c) was the weakest owing to its insignificant water content in subsequent hydration. Compressive strengths of the 28-day concretes are also shown in Fig. 7.3. The hydrated reference concrete was stronger than steamed reference, with the former being cured in a sealed condition while the latter with 8% water loss in first two hours sitting time. It was observed that concrete by steam curing could have higher early strength but lower late strength. Because of the water loss in 2-hour initial sitting, concretes with less curing water (2a+4s and 2a+4c) did not produce enough hydration products during subsequent hydration in comparison to their hydrated counterpart. On the other hand, spraying the concrete after carbonation (2a+4c+sp) promoted the formation of hydration products and ultimately improved the compressive strength by 12% compared to the non- sprayed concrete (2a+4c). Prolonged carbonation of 18 hours had a superior compressive strength with carbonation products occupying the majority of the microstructure even with only 54% remaining water without compensation. Although 18-hour carbonation process is not practical in industry production, the result shows longer carbonation in dynamic system can enhance carbonation reactivity. Instead 4-hour carbonation is more practically feasible. Its 28-day strength after water compensation was comparable to hydration reference and very close to that by 18-hour carbonation treatment. The results show that dynamic carbonation followed by water compensation is beneficial to compressive strength and carbonation reaction. It is clear that early carbonation does not hinder subsequent hydration. It is the water loss during carbonation that reduces hydration degree. With water compensation through spray, it is possible to make carbonation strength comparable to the best hydration reference.

Relative humidity control During 4-hour carbonation, the relative humidity was monitored. As carbonation is an exothermic reaction, it causes the evaporation of free water, and increases the relative humidity inside reaction chamber to nearly 100%. As water-saturated air prevents

– 152 – further evaporation, the concrete surface pores clog and prevent the gas from penetrating the core. The proposed dynamic carbonation regime dries and carbonates the sample simultaneously. Fig. 7.4 shows the effect of controlling relative humidity on carbonation. The chamber air temperature is considered an indication of carbonation. Due to the heat release during the reaction, the air temperature inside the chamber increases, i.e. the more the carbonation reaction, the higher the temperature. It can be noticed that the temperature increases rapidly during the first 50 minutes, where carbonation is at its highest efficiency. For the remaining 190 minutes, it increases but at a slower rate. The relative humidity was controlled at 53% based on previous investigations (Shideler 1955; Verbeck 1958). Alternately, when the relative humidity was not controlled, the temperature remained constant if not slightly decreased during the last 190 minutes. At the beginning of the test, the drop in the RH curve was due to passing of the gas initially in the desiccant material to lower the humidity to the desirable percentage. The temperature also decreased as the colder desiccant material cooled the gas down.

Effect of carbonation pressure

The effect of carbonation pressure on CO2 uptake was investigated. Instead of carbonating at 1 bar, the pressure was reduced to 0.5 bars. The same dynamic carbonation system was employed and the CO2 uptake was measured using furnace analysis in Table

7.6. Even at a lower carbonation pressure, the CO2 uptake was not affected. It is noted that due to the circulation pump, the flow was maintained passing through the concrete samples. This resulted in similar uptakes between 0.5 and 1 bar. The compressive strengths are also shown in Table 7.6. Carbonating at 0.5 bars did not significantly change the compressive strength. Thus, the dynamic carbonation system proved to be effective even at a lower gas pressure.

Phase analysis XRD patterns of concretes at age of 1 day are plotted in Fig. 7.5. The carbonated samples showed strong and dominant calcium carbonate peaks. Except the 18-hour carbonated concrete, a distinctive calcium hydroxide peak is identified at 18º in all other concretes including 4-hour carbonation. With 40% of water loss during 4-hour

– 153 – carbonation, carbonated concrete (2a+4c) could not produce as much CH as the steam, or hydrated reference concretes. Water compensation makes carbonated concrete (2a+4c+sp) comparable to hydration reference. Calcium hydroxide is believed to have been formed during subsequent hydration. Ettringite was also detected at 8 and 52 2θ. It usually tends to form due to the reaction of the aluminates and gypsum in the concrete during subsequent hydration. The other hydration product, C-S-H, did not display strong diffraction lines due to its amorphous nature. It is revealed by a slight increase in the baseline between 25º and 35º 2θ. A calcium carbonate dominated diffraction pattern is visible in the carbonated concretes with complete elimination of CH in 18 hours carbonation and significant consumption of anhydrous calcium silicates. It was noted that the intensity of the peak at 29.4º 2θ, which was calcite overlapped by C3S, was stronger in concrete with 18-hour carbonation than 4-hour carbonation. It was consistent with the

CO2 uptake in Table 7.4. Fig. 7.6 shows the XRD patterns of concretes 28 days after subsequent hydration. A slight increase in calcium hydroxide peak is observed due to subsequent hydration. In comparison between 1 day and 28 days, original aragonite and vaterite peaks at 23º, 36º, and 43º 2θ in the 18-hour carbonated sample have decreased in intensity. This is possibly due to the transformation of the unstable, poorly crystalline CaCO3, such as aragonite and vaterite, to more stable, well crystalline polymorph, such as calcite (Sawada 1997). However, this phenomenon is not easily detected in the remaining carbonated batches, which may indicate little or no transformation. A slight increase in baseline between 25º and 35º 2θ was also an indication of the presence of C-S-H. It was noted that the intensity of the peak at 41.3º 2θ, which was pure C3S, was stronger in hydrated concrete than 4- hour and 18-hour carbonation. It is an indication that C3S is consumed during carbonation and partially eliminated with prolonged carbonation.

TG/DTG analysis The DTG curves of 5 concretes at 1 day after casting are shown in Fig. 7.7. According to Ramachandran and Beaudoin (2001), the mass loss along TG curves can be classified into 6 categories: Below 105ºC, the mass loss is associated with the evaporable water and possibly the poorly formed C-S-H. The water loss between 105C and 200C

– 154 – represents the bound water in low temperature C-S-H and ettringite, while the loss in the range of 200C-420C is associated to the bound water in well-formed hydration products including C-S-H and C-A-H. CO2 in poorly crystalline calcium carbonates is decomposed between 470C-720C, while that in well crystalline calcium carbonates lies in the range of 720C-950C (Ramachandran 2001; Li 2003). This classification of reaction products based on decomposition temperature is used in quantitative analysis for comparison. The mass losses detected between 105ºC-420C are associated to hydration products (C-S-H and C-A-H). However, in between 420-470C, the steam-cured, and hydrated reference concretes show an endotherm, indicating a mass loss associated to the dehydration of

Ca(OH)2 (Chang 2006). The results are consistent with the Ca(OH)2 peaks in the XRD patterns. The majority of the produced carbonates are lost in the range of 470C to 750C. The carbonates produced from carbonation of hydration products, represented by a mass loss between 470-720ºC, showed the least thermal stability, suggesting poorly crystalline calcium carbonate (Villain 2006). This confirms the earlier work that the carbonation of hydrated cement would produce poorly crystalline calcite, aragonite, and vaterite (Cole 1960; Bertos 2004). Calcium carbonates generated by carbonation of calcium hydroxide are better crystalline and decomposed at temperatures higher than

720ºC (Moorehead 1986). Carbonation products from C3S and C2S reaction may be decomposed either at low temperature or high temperature, dependent on the duration of carbonation (Goto 1995). Fig. 7.8 presents the mass loss due to dehydration and decarbonation in 1-day concretes. The steamed concrete shows the highest C-S-H and C-A-H content followed by concrete of 18-hour carbonation (2a+18c) and the hydration reference. Steam curing and prolonged carbonation accelerate hydration and produce more hydration products. The hydrated reference shows comparable hydration in 1 day, which explains the corresponding compressive strength. The 4-hour carbonated concrete hydrates in less proportion since part of the hydrated cement was carbonated. Water compensation by spraying, on the other hand, promotes hydration development during the 18 hours of subsequent hydration. It shows a slight increase in C-S-H and C-A-H content. Ca(OH)2 formation is another indication of degree of hydration. Concrete after carbonation and water compensation during the curing procedure promoted further subsequent hydration

– 155 – and produced comparable Ca(OH)2 to the steam sample. Carbonation for 4 hours does not consume all the calcium silicates, and the remaining are converted into Ca(OH)2 with the availability of water. The hydrated reference could not produce sufficient Ca(OH)2 in 1- day hydration. Concerning the carbonates, which are presented in Fig. 7.8 as well, both hydrated and steamed controls show around 4% mass loss due to pre-existing limestone. Table 7.7 shows the percentage of poorly and well crystalline carbonates produced during carbonation. Whether 4 or 18-hour carbonation, 51-53% of the produced carbonates are well crystalline and 47-49% are poorly crystalline. A near-equal distribution of carbonates between the two phases is clearly obtained after 1 day subsequent hydration. Fig. 7.9 plots the DTG curves after 28 days of subsequent hydration. The curves of the 5 different batches are compared: Other than sample 2a+4s, all batches have similar patterns between 105C-420C. In the case of 2a+4s, the endotherm at 119C is correlated with the decomposition of ettringite which is evidenced by the XRD analysis. The endotherm between 420C-470C present in all batches except 2a+18c indicates the decomposition of Ca(OH)2. This is also supported by the peaks at 18º 2θ in the XRD patterns. After 470C, the carbonates formed in the carbonated samples are lost at the same temperature range, 780C-800C. However, concrete by 2a+18c exhibits another drop at 705C, possibly due to the formation of carbonates with poor crystallinity. The carbonates decomposition temperatures increased by 90C to 100C in the carbonated specimens after 28 days. This shift in decomposition temperature indicates a better crystallinity of the carbonates due to the transformation of the poorly crystalline, unstable phases of aragonite and vaterite to the more crystalline, stable calcite as suggested by the XRD (Ramachandran 2001; Li 2003). It should be noted that sample 2a+18c showed a broad exotherm just prior to decomposition. This exotherm is thought to be due to the energy release during a crystalline growth of calcite accompanied by a reduction of surface energy (Moorehead 1986). After 28 days, the detailed mass loss at each temperature range is investigated as shown in Fig. 7.10. The major increase in mass loss was seen in the samples with enough water to undergo subsequent hydration. While the steam-cured sample preserved the C-S- H and C-A-H formed after 1 day, the sprayed sample formed more hydration products during the 27 days. With sufficient water present due to sealed conditions, the hydrated

– 156 – control showed a high mass loss due to decomposition of Ca(OH)2. Similarly, due to spraying, the concrete sample (2a+4c+sp) resulted in a comparable mass loss. Water availability is crucial for Ca(OH)2 development. With a mass loss of 1.43% between

420ºC and 470ºC, the sprayed carbonated sample (2a+4c+sp) shows sign of Ca(OH)2 formation. However, it may have not been identified in XRD possibly due to the presence of an amorphous form of calcium hydroxide (Berger 1973). The mass loss due to decarbonation is also summarized in Fig. 7.8. It is noted that the mass loss between 470- 720ºC and 720-950ºC have changed after 28 days. According to Table 7.7, after 28 days, 65-68% of the produced carbonates in 4-hour carbonation are in the crystalline form, while reaching 76% in 18-hour carbonation. The increase in amount of well crystalline phases is due to the presence of more stable forms of calcium carbonate. A detailed comparison of 4-hour to 18-hour carbonation showed that the carbonates produced after 4 hours had a different morphology than those during the 4 hours. If all additional carbonates produced in the following 14 hours had the stability as those in the first 4 hours, then after 28 days, the ratio of poorly to well crystalline should be the same as that of the 4-hour carbonated samples. However, as indicated in Table 7.7, a further 11% of carbonates have transformed into crystalline. This signified that more than 70% of the 14-hour carbonates have recrystallized, which in turn shows that these carbonates were likely less stable than those produced during the first 4 hours. It is possible that prolonged carbonation (2a+18c) promoted more polymorph transformation during subsequent hydration than 4-hour carbonated concrete (2a+4c).

Table 7.8 summarizes the produced Ca(OH)2 and CaCO3 from the losses between 420C-470C and 470C-950C using stoichiometric proportions. Accordingly, the amount of calcium hydroxide produced during 1 and 28 days was maximized by a process of carbonation and water spray (2a+4c+sp). The CaCO3 content in carbonated concrete included the originally existed and the carbonation-produced. It was 40-46% and 58-65% in 4-hour and 18-hour carbonation respectively. It was noted that in both carbonation schemes, carbon dioxide reacted with C2S and C3S to produce C-S-H and

CaCO3, and with hydration products, mainly Ca(OH)2 and C-S-H to generate CaCO3 and silica gel (Young 1974; Goodbrake 1979b). With higher percent of solids from Ca(OH)2 and CaCO3 in carbonated concrete, the compressive strength was not increased

– 157 – proportionally. Nevertheless, the durability of concrete can be significantly enhanced due to the reduction of calcium hydroxide and the precipitation of calcium carbonates on surface (Rostami 2011).

SEM analysis In order to investigate the microstructure changes and morphology of concrete during carbonation and subsequent hydration, SEM test was performed after 1 and 28 days. Fig. 7.11 and 7.12 show 28-day micrographs of 4-hour carbonated concrete with different magnifications. The microstructure is typical of intermix of carbonation and hydration products. Similar results have been obtained in the identification of carbonation products (Shtepenko 2006). This phenomenon is repeatedly found throughout the sample. Fig. 7.13 displays a flakey image of sprayed 4-hour carbonated sample (2a+4c+sp) after 28 days. The purpose of choosing 28 days is to study the effect of water spraying on the microstructure after subsequent hydration. Accordingly, the SEM micrograph shows a denser microstructure due to the formation of hydration along with carbonation products. Fig. 7.14, on the other hand, shows another location in the same concrete with ettringite occupying the majority of the micrograph. This confirms that ettringite may form during subsequent hydration and is evidenced the XRD patterns. Prolonged carbonation is investigated in Fig. 7.15. The aggregate-concrete interface is investigated. A dense amorphous pattern suggests an intermix of CSH with calcium carbonates. Additionally, the highly carbonated sample (2a+18c) shows a dense granular, pebble-like microstructure in Fig. 7.16, which can be associated with the high

CO2 content. Fig. 7.17 and 7.18 display the effect of steam curing on a concrete specimen after 1 and 28 days respectively. Ettringite needles are distributed throughout the sample whether at 1 or 28 days, which are also verified by the XRD analysis. On the other hand, the hydrated sample shows a calcium hydroxide hexagon crystal surrounded by ettringite and C-S-H as seen in Fig. 7.19. The detected compounds are hydration products that result from the 28-day subsequent hydration. Such CH and C-S-H coexistence is supported by previous research results (Richardson 1993).

It is evident that the amorphous phase that intermingles C-S-H with CaCO3 dominates the microstructure of carbonated concrete. Effort has been made to identify

– 158 – this phase. Table 7.9 shows the mass lost between 105C-200C as poorly hydrated C-S- H, (DH1), 200ºC-420ºC as well hydrated C-S-H/C-A-H, (DH2), 470-720ºC as poorly crystalline CaCO3, (DC1), and 720-950ºC as well crystalline CaCO3, (DC2). Both concretes (2a+4c and 2a+18c) underwent the same process except the carbonation durations. Such a comparison aims at studying the effect of longer carbonation time on the amount of hydration and carbonation products. By 1 day, poorly hydrated phases (DH1) reach 3.58% and 4.81% with 4 and 18 hours of carbonation respectively. It is evident that longer carbonation accelerates hydration. After 28 days, the mass loss differs by little possibly due to increase in stability of hydration products. The low temperature carbonates, DC1, also decrease between 1 and 28 days due to transformation of poorly crystalline aragonite and vaterite to well crystalline calcite (Sawada 1997). The ratio of DH1/DC1 was calculated. It was almost constant at 0.47 and 0.85 for the two concretes at 1 and 28 days, respectively. Such a constancy can only mean that poorly hydrated CSH and low temperature carbonates are that of one component. This component is amorphous calcium silicate hydrocarbonate. This was also observed in a study on carbonation of β-

C2S system (Goto 1995). In case of the sprayed sample, the additional 18 hours of subsequent hydration up to 1 day increased the amount of hydration products, leading to different DH1/DC1 ratios of 0.74 and 1.43.

Dynamic carbonation of 20-cm CMU Full-size 20-cm CMUs were produced based on process parameters. Four batches were cast and cured with sealed hydration and dynamic carbonation. The hydrated reference is kept in a fog room with 100% relative humidity until strength testing. The carbonated batches are distinguished based on curing durations: 4-hour carbonated (2a+4c), 4-hour carbonated with water compensation (2a+4c+sp), and 18-hour carbonated (2a+18c). In slab samples, based on cement content, while the 4-hour dynamic carbonation curing reached an uptake of 13%, 18-hour curing achieved 20%. Their carbonation degree was about 24% and 40% respectively. The hydrated reference reached a compressive strength comparable to the slab samples used in earlier investigations. Only 4-hour dynamic carbonation with water spray and 18-hour dynamic carbonated concretes attained a comparable compressive strength with respect to their

– 159 – hydrated counterpart. The carbon storage capacity of concrete produced by this process was excellent. Standard 20-cm CMU block was made in the laboratory (Fig. 7.20) and carbonated in a setup similar to that shown in Fig. 7.1. The fresh blocks underwent an initial open-air sitting of 2 hours at 50±10% RH and 24±1ºC. Water loss due to initial sitting was about 7-9%, very close to that observed in slab tests. The blocks were then carbonated in the dynamic system at 0.1 MPa absolute gas pressure for 4 and 18 hours.

Water loss during dynamic carbonation was about 30-35%. CO2 uptake by three block samples, at the age of 28 days, is presented in Table 7.10. Furnace analysis was used in

CO2 content estimation. In comparison, the three blocks demonstrated CO2 uptake capacity in the order of 12.5%, which was close to 13% from slab tests. The compressive strength of CMU was measured after 1 day and 28 days. Prior to testing, the blocks were capped with a white capping cement to ensure load distribution on the block. Fig. 7.21 presents the recorded strengths. One-day strength was, in the 4- hour carbonated sample (2a+4c), tested after 2 hours of sitting, 4 hours of dynamic carbonation and 18 hours of subsequent hydration. The dynamic carbonated CMU blocks were comparable to the strength of the hydrated reference in one day. This was associated to the formation of carbonates at 12-20% CO2 uptake. However, after 28 days, the hydration strength exceeded that of carbonation in subsequent hydration because of the loss of water in dynamic carbonation. Water spray restored the lost water during curing. The final strength of water compensated 20-cm CMU blocks had exhibited equivalent strength to hydration reference. Therefore, CMU can be produced by dynamic carbonation process to replace steam curing and static carbonation and achieve high early strength, equivalent late strength, and superior CO2 uptake capacity in a short period of process window that is acceptable by large scale production.

CONCLUSIONS The carbonation capacity, strength development, and microstructure change of dynamic carbonation-cured lightweight concrete masonry units were studied. The reaction products due to carbonation are identified and quantified. Based on the findings of this study, the following conclusions may be drawn:

– 160 – 1. The dynamic gas flow allows the penetration of CO2 into the concrete and further promotes the degree of carbonation. Within 6 and 20 hours after casting, the

carbonated concretes have exhibited a CO2 uptake of 12% and 20% in terms of cement, leading to a carbonation degree of 24% and 40% respectively. The dynamic system can control RH level to prevent water accumulation on the

surface and provide a route of capillaries for CO2 to infiltrate. Water compensation through water spray after carbonation is a critical part of the curing scheme, which makes carbonated concrete comparable to hydration references. 2. Dynamic carbonation effectively combines drying and carbonation into one process that shortens curing and can be possibly done in a 8-hour work shift. Carbonation of 4 hours produces a concrete with more carbonation products on surface while carbonation of 18 hours makes a uniform carbonation through the thickness. Longer carbonation time promotes a higher degree of reaction. 3. The gas pressure has no apparent effect on carbonation degree in the dynamic system. Carbonation at 0.5 bar yields similar results as carbonation at 1 bar. Therefore, it is possible to employ the dynamic circulation system in a sealed chamber, which is not necessary to be pressure resistant. 4. The carbonation-induced carbonates are characterized by poorly and well crystalline polymorphs. It is found that the poorly crystalline phases in 4-hour carbonated samples dropped from 47-48% to 32-35% after 28 days subsequent hydration. On the other hand, the 18-hour counterpart decreased from 48 to 24%. It was apparent that the carbonates produced after 4 hours had less stability than those created during the first 4 hours. The phase transformation during hydration could occur evidenced by XRD patterns and shift in the DTG curves. The dominant polymorph of calcium carbonate was calcite, the most stable polymorph. 5. The coexistence of carbonation and hydration products has been observed using TG/DTG and SEM. The ratio of low temperature hydrates to low temperature carbonates (DH1/DC1) is indicative that these two amorphous phases are of one component. The constancy of this ratio suggests the possible presence of an amorphous calcium silicate hydrocarbonate. The complete analysis of TG/DTG,

– 161 – SEM, and XRD indicates that the carbonation curing partially consumes calcium silicates and calcium silicate hydrates, and produces calcium carbonates in different polymorphs and calcium silicate hydrocarbonates.

ACKNOWLEDGEMENTS The authors gratefully acknowledge the support by the Natural Science and Engineering Research Council (NSERC) of Canada, Canadian Concrete Masonry Producers Association (CCMPA), and Lafarge Canada.

REFERENCES ASTM (2004b). "Standard Test Methods for Sampling and Testing Concrete Masonry Units and Related Units." ASTM International C140.

Berger, R. L., and McGregor, J. D. (1973). "Effect of temperature and water-solid ratio on growth of Ca(OH)2 crystals formed during hydration of Ca3SiO5." Journal of American Ceramic Society 56(2): 73-79.

Berger, R. L., and Young, J. F. (1972). "Acceleration of Hydration of Calcium Silicates by Carbon-Dioxide Treatment." Nature-Physical Science 240(97): 16-&.

Bertos, M. F., Simons, S. J. R., and Hills, C. D. (2004). "A review of accelerated carbonation technology in the treatment of cement-based materials and sequestration of CO2." Journal of Hazardous Materials B112: 193-205.

Chang, C., and Chen, J. (2006). "The experimental investigation of concrete carbonation depth." Cement and Concrete Research 36: 1760-1767.

Cole, W. F., and Kroone, B. (1960). "Carbonate minerals in hydrated Portland cement." Cement and Concrete Research 31(12): 1275-1295.

Collier, N. C., Sharp, J. H., Milestone, N. B., Hill, J., and Godfrey, I. H. (2008). "The influence of water removal techniques on the composition and microstructure of hardened cement pastes." Cement and Concrete Research 38: 737-744.

El-Hassan, H., Shao, Y., and Ghouleh, Z. (2012a). "Effect of initial curing on carbonation of lightweight concrete masonry units." American Concrete Institute Journal.

El-Hassan, H., Shao, Y., and Ghouleh, Z. (2012b). "Reaction Products in Carbonation Cured Lightweight Concrete." ASCE Materials Journal.

– 162 – Goodbrake, C. J., Young, J. F., and Berger, R. L. (1979b). "Reaction of hydraulic calcium silicates with carbon dioxide and water." Journal of the American Ceramic Society 62(9-10): 488-491.

Goto, S., Suenaga, K., and Kado, T. (1995). "Calcium silicate carbonation products." Journal of the American Ceramic Society 78(11): 2867-2872.

Johnson, D. (2000). "Accelerated carbonation of waste calcium silicate materials." Society of Chemical Industry 10.

Jones Jr., R. H. (1997). Cement Treated with High-pressure CO2. United States Patent. United States of America. 5: 562.

Li, D., Shi, H., Deng, J., and Xu, Y. (2003). "Study on the thermal decomposition kinetics of nano-sized calcium carbonates." Journal of Zhejiang University 4(3): 363-368.

Monkman, S., and Shao, Y. (2006). "Assessing the carbonation behavior of cementitious materials." Journal of Materials in Civil Engineering November/December: 768- 776.

Moorehead, D. R. (1986). "Cementation by the carbonation of hydrated lime." Cement and Concrete Research 16: 700-708.

Ramachandran, V. S., and Beaudoin, J. J. (2001). Handbook of Analytical Techniques in Concrete Science and Technology, William Andre Publishing/Noyes

Richardson, I. G., and Grooves, G. W. (1993). "The incorporation of minor and trace elements into calcium silicate hydrate gel in hardened cement pastes." Cement and Concrete Research 23: 131.

Rostami, V., Shao, Y., and Boyd, A. J. (2011). "Durability of concrete pipes subjected to combined steam and carbonation curing." Construction and Building Materials 25: 3345-3355.

Sawada, K. (1997). "The mechanism of crystallization and transformation of calcium carbonates." Pure and Appl. Chem. 69(5): 921-928.

Shideler, J. J. (1955). "Investigation of the moisture-volume stability of concrete masonry units." Portland Cement Association Research and Development Laboratories Bulletin D3.

Shtepenko, O., Hills, C., and Brough, A. (2006). "The effect of carbon dioxide on beta- dicalcium silicate and Portland cement." Chemical Engineering Journal 118: 107- 118.

– 163 –

Soroushian, P., and Hsu, J. W. (1999). Accelerated Curing of Cement-based Materials. United States Patent. United States of America. 5,935,317.

Verbeck, G. J. (1958). "Carbonation of Hydrated Portland Cement." American Society for Testing Materials Bulletin 87(205): 16-36.

Villain, G., and Platret, G. (2006). "Two experimental methods to determine carbonation profiles in concrete." American Concrete Institute Materials Journal 103(4): 265- 271.

Young, J. F., Berger, R. L., Breese, J. (1974). "Accelerated Curing of Compacted Calcium Silicate Mortars on Exposure to CO2." Journal of the American Ceramic Society 57(9): 394-397.

– 164 – TABLES Table 7.2: Mixture proportion of concrete Slab CMU Mass Percent (g) (g) (kg/m3) (%) Ordinary Portland Cement 88 1967 241 13 Water 35 787 96 5 SSD Expanded Slag 554 12251 1502 82 Concrete Sample 677 15004 1839 100

Table 7.2: Curing conditions

Subsequent Initial Sitting Steam Curing Carbonation Curing Hydration

RH T t RH T t RH t Water t Sample (%) (°C) (hours) (%) (°C) (hours) (%) (hours) Sprayed (g) (days)

0a (Ref) n/a n/a n/a n/a n/a n/a n/a n/a n/a 28

2a + 4s 50 ± 10 24 ± 1 2 95 ± 5 75 ± 5 4 n/a n/a n/a 28

2a + 4c 50 ± 10 24 ± 1 2 n/a n/a n/a 55 ± 5 4 n/a 28 2a + 4c + sp 50 ± 10 24 ± 1 2 n/a n/a n/a 55 ± 5 4 25 28

2a + 18c 50 ± 10 24 ± 1 2 n/a n/a n/a 55 ± 5 18 n/a 28 a – initial sitting; s – steam curing; c – carbonation; RH – relative humidity; T – temperature; t – time; sp – sprayed water

Table 7.3: Water content in concretes Water content (%) After After 2-hr After After Sample forming sitting carbonation water spray Final 0a (Ref) 100 n/a n/a n/a 100 2a + 4s 100 92 n/a n/a 92 2a + 4c 100 91 62 n/a 62 2a + 4c + sp 100 91 61 100 100 2a + 18c 100 91 54 n/a 54

– 165 – Table 7.4: CO2 content (%) and CO2 uptake (%) by thermal analysis

CO2 content (through Average CO2 content Average thickness concrete) (surface concrete) CO2 CO2 Sample 1-day 28-day uptake 1 day 28-day uptake 0a (Ref) 4.46 ± 0.35 5.16 ± 0.43 0 4.62 5.83 0 2a + 4s 5.01 ± 0.41 3.9 ± 0.65 0 3.47 4.64 0 2a + 4c 17.16 ± 0.88 18.3 ± 0.91 12.92 20.47 18.35 14.19 2a + 4c + sp 17.45 ± 0.65 18.47 ± 0.87 13.15 18.87 19.52 13.97 2a + 18c 25.35 ± 0.59 24.39± 0.64 20.06 25.69 28.33 21.79

Table 7.5: Cement content (CC) in concrete powder samples

Test Furnace Uptake (%) TG/DTG Uptake (%) Sample Date Mc (g) (ΔCO2)C (g) Mp (mg) (ΔCO2)P (mg) CC 0a (Ref) 1-day 11.28 0.07 120.5 3.90 0.70 2a + 4s 1-day 10.12 0.05 120.6 2.93 0.70 2a + 4c 1-day 10.55 0.28 119.3 12.94 0.53 2a + 4c + sp 1-day 11.07 0.27 120.9 11.41 0.50 2a + 18c 1-day 12.06 0.40 119.8 16.01 0.52 0a (Ref) 28-day 12.22 0.09 120.4 5.47 0.78 2a + 4s 28-day 11.47 0.07 121.2 4.61 0.82 2a + 4c 28-day 11.56 0.28 120.6 10.62 0.48 2a + 4c + sp 28-day 10.21 0.26 120.2 11.50 0.49 2a + 18c 28-day 10.77 0.40 119.6 17.62 0.52

Table 7.6: Effect of gas pressure on dynamic carbonation 0.5-bar 1-bar 28-day 28-day CO2 Compressive CO2 Compressive Sample Content (%) Strength (MPa) Content (%) Strength (MPa) 2a + 4c 16.29 ± 1.02 15.23 ± 1.03 17.16 ± 0.88 15.85 ± 0.89 2a + 4c + sp 16.91 ± 0.97 16.98 ± 0.98 17.45 ± 0.65 17.83 ± 1.09 2a + 18c 24.51 ± 0.75 17.39 ± 1.32 25.35 ± 0.59 18.16 ± 1.25

– 166 – Table 7.7: Percentage of poorly and well crystalline carbonates

Poorly Crystalline Well Crystalline Sample # 1-day 28-day 1-day 28-day 2a + 4c 48.4 32.8 51.6 67.2 2a + 4c + sp 47.6 34.7 52.4 65.3 2a + 18c 47.9 23.6 52.1 76.4

Table 7.8: Calcium hydroxide and calcium carbonate content after 1 and 28 days Phase content (%) 0a (Ref) 2a + 4s 2a + 4c 2a + 4c + sp 2a + 18c 1-day 5.26 6.41 3.49 5.02 2.26 Ca(OH)2 by TG 28-day 13.40 6.78 4.11 5.88 2.34 1-day 10.50 7.89 46.52 42.89 58.40 CaCO3 by TG 28-day 13.25 10.55 41.70 44.37 64.39 1-day 15.76 14.30 50.01 47.90 60.66 Ca(OH)2 + CaCO3 28-day 26.65 17.33 45.81 50.25 66.73

Table 7.9: Dehydration and decarbonation mass

DH1 (%) DH2 (%) DC1 (%) DC2 (%) DH1/DC1

Sample 1 day 28 day 1 day 28 day 1 day 28 day 1 day 28 day 1 day 28 day

0a (Ref) 4.88 5.13 4.06 4.63 0.00 0.00 0.00 0.00 n/a n/a 2a+4s 8.57 7.03 2.86 5.00 0.00 0.00 0.00 0.00 n/a n/a 2a + 4c 3.58 3.50 4.72 5.21 7.67 4.10 8.17 8.42 0.47 0.85 2a + 4c + sp 5.04 6.80 3.60 4.12 6.79 4.76 7.46 8.94 0.74 1.43

2a + 18c 4.81 4.56 5.00 5.65 10.10 5.32 10.97 17.18 0.48 0.86 DH1 (105-200ºC), poorly hydrated C-S-H DH2 (200-420ºC), well hydrated C-S-H and C-A-H DC1 (470-720ºC), poorly crystalline CaCO3 DC2 (720-950ºC), well crystalline CaCO3

Table 7.10: CO2 content and uptake of concrete blocks

28-day CO2 Content (%) Average CO2 Sample Block #1 Block #2 Block #3 Average Uptake (%) 0a (Ref) 5.21 5.64 5.55 5.47 ± 0.23 0 2a + 4c 17.91 18.09 18.58 18.19 ± 0.35 12.73 2a + 4c + sp 17.45 18.42 18.02 17.96 ± 0.49 12.50 2a + 18c 25.20 25.39 24.35 24.98 ± 0.55 19.51

– 167 – FIGURES

Fig. 7.1: Schematic of carbonation setup

9 8 6-Hour Strength 6.7 7 6.0 6 5 4

3 2.2 2

Compressive Strength (MPa) CompressiveStrength 1 0 2a+4c 2a+4s 0a - Ref Sample

Fig. 7.2: Six-hour compressive strength of slab samples

– 168 – 25 1-day 28-day 20 18.2 17.8 18.2 15.9 15.2 15

9.7 10 9.3 8.8 7.9 8.3

5 Compressive Strength (MPa) CompressiveStrength

0 2a+18c 2a + 4c 2a + 4s 2a+4c+sp 0a - Ref Sample

Fig. 7.3: Compressive strength of slab samples

Fig. 7.4: Relative humidity and temperature during 4-hour carbonation

– 169 –

Fig. 7.5: XRD patterns of 1-day slab samples

Fig. 7.6: XRD patterns of 28-day slab samples

– 170 –

Fig. 7.7: Comparison of DTG curves of 1-day slab samples

24 105C

16

13.27

12.42 11.43

12 10.84

9.96

9.81

9.62

8.94

8.91

8.64 8.30

Mass Loss (%) Loss Mass 8 3.17

4 2.72

1.56

1.45

1.28

1.22

0.85

0.75 0.55 0 0a 2a + 4s 2a + 4c 2a + 4c + sp 2a + 18c Sample # Fig. 7.8: Mass loss due to dehydration and decarbonation in 1-day concretes

– 171 –

Fig. 7.9: Comparison of DTG curves of 28-day slab samples

24 105C

470C

16 12.03

12 10.92

10.31

10.28

10.21

9.81

9.76

9.72

9.29 9.06

Mass Loss (%) Loss Mass 8

4.96 3.96

4 3.26

1.65

1.43

1.00

0.87

0.68 0.57 0 0a 2a + 4s 2a + 4c 2a + 4c + sp 2a + 18c Sample #

Fig. 7.10: Mass loss due to dehydration and decarbonation in 28-day concretes

– 172 –

Fig. 7.11: SEM micrograph of carbonated concrete (2a+4c) after 28 days

Fig. 7.12: SEM micrograph of carbonated concrete (2a+4c) after 28 days

– 173 –

Fig. 7.13: SEM micrograph of carbonated concrete (2a+4c+sp) after 1 day

Fig. 7.14: SEM micrograph of carbonated concrete (2a+4c+sp) after 28 days

– 174 –

Fig. 7.15: SEM micrograph of carbonated concrete (2a+18c) after 1 day

Fig. 7.16: SEM micrograph of carbonated concrete (2a+18c) after 28 days

– 175 –

Fig. 7.17: SEM micrograph of steam concrete (2a+4s) after 28 days

Fig. 7.18: SEM micrograph of steam concrete (2a+4s) after 28 days

– 176 –

Fig. 7.19: SEM micrograph of hydrated concrete (0a) after 28 days

Fig. 7.20: Full-size 20-cm concrete block

– 177 – 20 1-day 28-day 14.9 15.0 15 14.1 12.4

10 9.2 7.8 7.1

5 4.3 Compressive Strength (MPa) CompressiveStrength

0 0a (Ref) 2a + 4c 2a + 4c + sp 2a + 18c Sample

Fig. 7.21: Compressive strength of 20-cm concrete blocks

– 178 – Chapter 8

—

PERFORMANCE COMPARISON OF OPC AND PLC CONCRETES SUBJECT TO

EARLY CARBONATION

PREFACE In the previous chapters, the carbonation behaviour of concrete masonry units made of Ordinary Portland Cement (OPC) and Portland Limestone Cement (PLC) was studied. While PLC concrete consumes less clinker and produces environmentally friendly concrete, the carbonation reactivity and strength gain are lower in PLC than in OPC concrete produced and cured at the same conditions. This chapter summarizes a comparison of carbonated OPC and PLC concrete in terms of strength gain, CO2 uptake capacity, and durability performance. The CO2 uptake capacity of the two concretes was compared using mass gain, mass curve, furnace analysis, hydrochloric acid titration, and thermogravimetry methods. The durability was characterized by water absorption and freeze-thaw resistance. Microstructural analysis was studied through X-Ray diffraction (XRD), thermogravimetry (TG), and scanning electron microscopy (SEM).

– 179 – INTRODUCTION Previous work in Chapters 3 through 6 have demonstrated the use of ordinary

Portland cement (OPC) and Portland limestone cement (PLC) as binder and CO2 reactant in concrete masonry unit production by early carbonation curing. While OPC is the most commonly used conventional cement, PLC is slowly getting accepted as an environmentally friendly and economically competitive alternative. Based on CSA A3000-08 (CSA 2008), limestone addition up to 5% is allowed with no label. Therefore, the OPC used in this thesis is actually the low limestone cement (3-5%) in comparison to high limestone cement with 13-15% limestone and clearly labeled as PLC. The use of limestone is an important sustainable development by the cement industry. It was vital to understand the differences of using either cement in the carbonation curing process of concrete masonry units by evaluating their performance and analyzing the microstructure changes. This chapter presents a study on the comparison of early carbonation behavior of OPC and PLC concrete subject to early carbonation. The carbonation curing consists of a 4-hour carbonation period after an initial hydration of 18 hours. Water compensation through spray has been incorporated to provide subsequent hydration after carbonation.

The comparison presented in this work highlights the differences in CO2 uptake, compressive strength, and microstructure changes characterized by X-Ray diffraction (XRD), thermogravimetry (TG), and scanning electron microscopy (SEM). Based on ASTM C140 and C1262, concrete masonry units shall pass the following tests and satisfy the minimum requirement: 28-day strength, density, water absorption, and resistance to freeze-thaw (ASTM 2004a; ASTM 2004b). Durability is always the concern for concrete products in service. Early carbonation has proven effective in improving durability performance by eliminating or reducing calcium hydroxide. It was found that carbonation enhances the resistance to freeze-thaw and sulfate attack in normal weight concrete (Rostami 2011). Other work showed that carbonation of concrete, either preceded by initial air curing or steam curing, offered a preshrinking curing scheme that could reduce service shrinkage (Shideler 1955; Toennies 1960; Toennies 1963). This chapter will compare OPC with PLC concretes, both hydrated and carbonated, in terms of their strength gain, carbonation reactivity, water absorption, freeze-thaw response, and microstructural development. A carbonation

– 180 – model is introduced to explain the effect of early carbonation on durability behavior.

EXPERIMENTAL INVESTIGATIONS Concrete sample preparation and curing The carbonation curing of concrete masonry units (CMU) were studied using 127 x 76 x 38 mm rectangular concrete slabs to simulate the typical web or face shell of a hollow 20-cm CMU. The chemical composition of both cements is shown in Table 8.1.

The CO2 content is 1.72% in OPC and 6.97% in PLC. It is corresponding to limestone addition of 3.9% in OPC and 15.8% in PLC. OPC concrete slabs were cast and cured according to section "Concrete sample preparation" in Chapter 3, and process of PLC concrete slabs was described in section "PLC concrete sample preparation" in Chapter 5.

CO2 uptake estimation

CO2 uptake was estimated utilizing five different techniques: mass gain, mass curve, furnace analysis, TG, and HCl titration. The measurement of CO2 uptake of OPC concrete subject to early carbonation was described in Chapters 3 and 4, section "CO2 uptake estimation", while that of PLC concrete was given in Chapters 5 and 6, section

"CO2 uptake estimation".

Performance evaluation Carbonated concretes were compared to a normal hydration control. Carbonation curing was carried out after 18-hour initial curing, while hydration reference concretes were placed in a sealed plastic bag after casting. The compressive strength was measured after 1 and 28 days for carbonated and hydration control samples following ASTM C140 (ASTM 2004b). Three samples per batch were tested and averaged. The fractured samples were preserved in an acetone solution to stop hydration. Acetone exchange with water to stop hydration was reported as the least damaging method to preserve the microstructure (Collier 2008). Prior to analysis, the samples were pre-dried overnight at 60oC. A different set of batches was cast and cured for 28 days to test for water absorption. This test was performed according to ASTM C140 (ASTM 2004b), which

– 181 – was developed for concrete masonry units. The procedure stated that the 28-day subsequent hydration concretes were first oven-dried at 60C for 3 days or until a constant mass was obtained. They were then immersed in tap water and weighed at SSD every 24 hours to check the increase in mass. The saturation stage was reached when the mass increase was less than 0.5% of the heavier mass. Thus, the water absorption due to immersion was acquired following Eq. 8.1:

– Water absorption after immersion, % = (8.1)

Slab samples were cured for 28 days before undergoing cycles of freezing and thawing according to ASTM C1262 (ASTM 2004a). The accelerated test is the standard test method for evaluating the freeze-thaw durability of manufactured concrete masonry units. It specified that the samples be placed in either water or saline solution until fully saturated. For this experiment, concretes were placed in saline solution after a sequence of curing, drying, and saline solution submersion. Prior cycling test, the concrete were hydrated for 28 days, air dried at 25ºC and 60% relative humidity, and placed in saline solution for 24 hours. Freeze-thaw cycling was commenced. A full freeze-thaw cycle consisted of 16 hours of freezing at -18±5C and 8 hours of thawing at 24±5C. Fig. 8.1 shows a schematic of the freeze-thaw test. The water thickness surrounding the concrete is to be in a range of 3 to 38 mm from all sides with a 3-mm spacing from the bottom. At any point during the test, the specimens were not to be fully submerged. The lid was placed on top of the container to prevent evaporation during any stage of the experiment. The standard requires after 5, 10, 15, and 20 cycles, the samples be removed from the containers, and using a squeeze bottle and surface rubbing, any loose parts be cleaned. The loss of mass was represented by the residue that was then dried in an oven overnight at 105C. The loss of mass was an indication of the resistance to freeze-thaw cycles. The experiment was terminated if the mass loss exceeded 15% or 20 cycles had been reached, whichever came first. The following concretes were tested for freeze-thaw resistance: two carbonated OPC concretes (18a+4c and 18a+4c+sp), two carbonated PLC concretes (18a+4c and 18a+4c+sp), three hydrated reference concretes (hydrated OPC, hydrated PLC and hydrated commercial block), and one steam-cured OPC concrete (2a+4s from Chapter 7). For each batch, two samples were tested for average. The commercial

– 182 – concrete block was purchased and tested to serve as second hydration reference. The 20- cm commercial normal weight concrete block had shown a compressive strength of 24.8 MPa and a mass of 17-18 kg for each. Two samples were made by cutting the web to obtain a dimension of 127 x 76 x 38 mm. The normal weight concrete samples from commercial blocks were stronger and heavier than the lightweight concrete developed in this thesis. One and 28 days after casting, SEM analysis was performed on the six concretes: three OPC (hydrated reference, carbonated, and carbonated with water compensation) and three PLC (hydrated reference, carbonated, and carbonated with water compensation). Gravel-size granules were chiseled from the surface of a concrete slab. The exposed fracture surface was coated by a gold/palladium coating to ensure conductivity during the experiment. The coated specimens were then analyzed using high-vacuum scanning electron microscopy employing a Hitachi S4000 in backscattered electron (BSE) mode with an EDAX Phoenix EDS microanalysis system. The micrographs were used to study the morphology of carbonated and hydrated concretes. Powder samples collected from concrete were subject to thermogravimetry (TG) analysis by employing a thermal analyzer (NETZSCH, TG 449 F3 Jupiter) with a resolution of 0.01 mg. Similarly, two samples were analyzed per batch. The TG and DTG curves were thus obtained in terms of mass loss between 25C and 950C at a heating rate of 10C/min. The XRD patterns were obtained at 1 and 28 days after casting by means of a Philips PW1710 Powder Diffractometer (Cu, K radiation, scan interval 10-100, 0.02, and 0.5 seconds per step). Both OPC and PLC concrete powder analyses were performed to qualitatively and quantitatively determine the amount of calcium hydroxide, calcium silicates, and calcium carbonates in carbonated and control concretes for comparison.

EXPERIMENTAL RESULTS AND DISCUSSION Carbonation degree

The carbonation degree of concrete was estimated by the CO2 uptake. Table 8.2 summarizes the CO2 uptake and content using 5 different methods. While CO2 uptake by mass gain and mass curve methods are absolute, the thermal analysis by furnace method

– 183 – and TG method as well as chemical analysis by HCl titration yield CO2 content and are relative. The former is not affected by initial CO2 content, while the latter has to subtract the CO2 content of hydrated reference from carbonated concretes to acquire the CO2 uptake due to carbonation reaction. The results by 5 methods are compared and averaged.

It was shown that OPC concretes yielded approximately 30% more CO2 uptake than PLC concretes. This comparison is based on dry cement. Since limestone is non-reactive with carbon dioxide, the carbonation degree will be increased for PLC concrete if the comparison is referred only to reactive component in cement, i.e. the Portland cement. In

PLC, the limestone content was 15.8% based on 6.97% CO2 content in the PLC composition. CO2 uptake by PLC concrete is increased to 22.6% in PLC concrete if the uptake of 19.0% is divided by 0.842. Therefore, if only reactive Portland cement is considered as a base for comparison, PLC has shown comparable carbonation degree to

OPC concrete. The CO2 uptake was 22.6% in PLC versus 23.5% in OPC. It was reported that PLC concrete was more vulnerable to weathering carbonation when the limestone addition was over 19% (Parrott 1996). This phenomenon was not observed in early carbonation, even though the Portland cement in PLC blend was ground finer than in

OPC. PLC was less reactive with CO2 in early carbonation curing for concrete masonry unit application. The effect of cement particle size on carbonation degree was not apparent.

Compressive strength The compressive strength after 1 and 28 days are investigated in Fig. 8.2. At 1 day, the carbonated OPC (18a+4c and 18a+4c+sp) gained 20-25% more strength than the carbonated PLC. In addition, the hydrated OPC had seen a 26% increase in strength in comparison to hydrated PLC. After 28 days, the strength of carbonated OPC concretes was still 20-25% higher, while that of hydrated OPC was 29% higher than its PLC counterpart. In general, even though PLC is finer than OPC, the strength of PLC concrete was lower than OPC. It is the Portland cement content, not the particle size, that dominates the final strength in both hydration and carbonation. It should be noted that lightweight concrete masonry units are considered as special concrete. Unlike conventional concrete with higher cement content and denser packing, the investigated

– 184 – concrete contains low cement content (13%), lightweight aggregate, and is porous in structure due to dry mix and compact forming. It is conclusive that, with less cement content (13%) in lightweight porous concrete blocks in comparison to regular concrete (15-20%), and an additional 15% replacement of cement by limestone, PLC did not promote more carbonation and hydration, as well as their related compressive strength gain. Previous studies concerning the compressive strength of PLC showed that the limestone addition of 15-30% improved the strength, but a slight decrease was noticed with 30-50% addition. However, in both cases, a strength reduction by 8-19% was noticed after 28 days (Pera 1999; Bonavetti 2003). The compressive strength of carbonated OPC and PLC was also compared. With higher CO2 uptake, the carbonated OPC concretes resulted in a higher compressive strength in comparison to PLC. A further microstructure study may explain the reason of different strengths.

Water absorption The water absorption of OPC and PLC concretes is compared in Fig. 8.3. Preconditioning at 60ºC for 3 days or until a constant mass was obtained caused the evaporation of part of the free water but did not affect the bound water in hydration products. According to Eq. 8.1, the oven-dried concrete mass was the basis for water absorption study. Thus, at initial oven-dried state, the permeable voids included the capillaries in the cement paste and the available cavity space owing to the nature of the porous concrete masonry units. In case of the hydrated references, the water absorption was similar for both OPC and PLC concretes. However, the 4-hour carbonated OPC concrete (18a+4c) exhibited a relatively lower water absorption than its hydrated counterpart, which is associated with the precipitation of carbonates. Water compensation through water spray also reduced the studied parameter, with OPC having the superior performance among the two concretes. It was noted that unlike carbonated OPC (18a+4c), in which the reaction lowered water absorption, PLC carbonation (18a+4c) increased the water absorption. It seemed, however, that the simultaneous incorporation of hydration and carbonation products in PLC concrete through water compensation led to somewhat improved results. It is conclusive that carbonation relatively reduced water

– 185 – absorption and a further reduction was noticed with water compensated carbonated concrete.

Freeze-thaw resistance Freeze-thaw resistance in concretes was measured through mass loss during freeze-thaw cycling. It is presented in Fig 8.4 as mass loss against cycles. ASTM C1262 requires the experiment be terminated at 20 cycles or when total mass loss exceeded 15%. It was interesting to notice that OPC and PLC concretes responded to freeze thaw cycles in a similar manner, but carbonation and hydration showed quite a difference. The hydrated reference lightweight concrete in both OPC and PLC cases lost more than 15% of mass in just 3 cycles. The steam-cured OPC concrete was the worst in freeze thaw resistance. Its mass loss was over 25% at 3 cycles. Even the commercial normal weight concrete block, which was much stronger than lightweight concrete, also failed in about 3 cycles. Total disintegration took place after 6 cycles. Figs. 8.5, 8.6, and 8.7 show the remaining mass of three hydration references after 6 cycles. Only 20% mass was retained. The freeze-thaw damage to hydrated concretes was extremely severe. The steam-cured concrete (2a+4s) shown in Fig. 8.8 was totally disintegrated after 6 cycles and turned into a sandy mush with no solids. Nevertheless, the carbonated concretes had demonstrated much better resistance. Carbonated OPC concretes (18a+4c and 18a+4c+sp) had lost 15% mass after 10 cycles. Their remaining specimens are shown in Fig. 8.9 and 8.10 respectively. The degree of damage in carbonated lightweight concrete was much less than that in hydration references and steam-cured concrete. Carbonated PLC concretes (18a+4c and 18a+4c+sp) had seen a similar 15% mass loss after 10 cycles. Their respective specimens after 10 cycles are shown in Fig. 8.11 and 8.12. It was clear that carbonated concretes were much more resistant to freeze-thaw damage in comparison to hydrated and steam-cured concrete whether the concretes were OPC or PLC, or whether they were lightweight low-strength concrete or normal weight higher strength concrete. It is likely that the calcium carbonates produced during carbonation reaction have altered the microstructure of the binding phase and played a significant role in improving freeze- thaw resistance.

– 186 – Phase analysis Microstructures of hydrated and carbonated concretes are analyzed and compared using XRD, TG, and SEM. The XRD patterns after 1 day for both concretes are compared in Fig. 8.13. Hydrated OPC and PLC concrete showed different patterns. While both concretes showed C2S and C3S phases with Ca(OH)2, it was only hydrated PLC that showed peaks of vaterite, which can be attributed to the added limestone (15.84%). Ettringite was detected at 52º 2θ, but the more distinctive peak at 9º 2θ appeared to be much smaller with respect to the other peaks. In addition, a slight increase in baseline of hydrated OPC between 25º and 35º 2θ was also an indication of the presence of C-S-H. On the other hand, carbonation resulted in more vaterite and aragonite peaks in PLC than in OPC. In both cases, carbonation consumed Ca(OH)2 and ettringite. Fig. 8.14 displays the XRD patterns after 28 days of OPC and PLC concretes. The comparison between hydrated concretes showed similar results to those obtained after 1 day. While vaterite and aragonite phases decreased during subsequent hydration in carbonated OPC concrete, they were maintained in PLC counterpart.

TG/DTG analysis Phase analysis was further quantified by TG analysis. DTG curves of 1-day concretes are compared in Fig. 8.15. The hydrated OPC and PLC showed similar peaks of

CSH, CH, and CaCO3 with a more identifiable endotherm of CH in OPC. In the carbonated concrete (18a+4c), the addition of limestone in PLC appeared to produce only one endothermic peak at 746ºC, while OPC showed two at 689ºC and 770ºC. Decomposition temperatures in the range of 470-720ºC were associated with poorly crystalline carbonates designated by aragonite and vaterite (Ramachandran 2001; Li 2003). The well crystalline carbonates attributed to calcite decompose in a higher temperature range of 720-950ºC (Ramachandran 2001). The endotherms were indicative of possible production of poorly crystalline and well crystalline calcium carbonate polymorphs in OPC. Water compensation through water spray did not change the results with respect to the non-compensated concretes. A similar phenomenon of double endothermic peaks in OPC concrete with water compensation was observed.

– 187 – After 28 days of subsequent hydration, the resulting DTG curves are presented in Fig. 8.16. Hydration products C-S-H and C-A-H did not show significant increase after 28 days of hydration. Calcium hydroxide started to develop after carbonation in OPC. This was evidenced by the peaks in the sprayed carbonated OPC concrete, which were not seen at 1 day. It is possible that these peaks were not identified in the XRD patterns due to the amorphous nature of some CH (Berger 1973). A transformation of carbonates from poorly crystalline (aragonite and vaterite) to well crystalline (calcite) was observed in OPC concrete with double peaks becoming a single peak. The decomposition of carbonates in OPC took place at 781-793ºC, approximately 30-40ºC higher than that of PLC. Similar transformation was noticed in previous studies (Sawada 1997). However, this transformation was not observed in PLC. The endothermic peaks at 1 day and 28 days in PLC concrete were kept the same. There was no phase transformation during subsequent hydration in PLC concrete. It is apparent that carbonated OPC concrete has experienced phase transformation of carbonates, producing carbonates with better crystallinity after 28 days subsequent hydration. Table 8.3 summarizes percentage of poorly and well crystalline carbonates produced during carbonation and subsequent hydration. It was noted that in carbonated OPC concrete, during 28-day subsequent hydration, carbonates had a tendency to transform into the more crystalline polymorph, calcite, with an increase from 25-28% to 41-49%. However, this is not the case in PLC. The percentage of well crystalline carbonates remains constant in the range of 47-52%. This indicates more amounts of well crystalline carbonates in PLC carbonated concrete. The effect of crystallinity on compressive strength is investigated by comparing poorly crystalline phase percentage in Table 8.3 with compressive strength in Fig. 8.2. The comparison was carried out at 1 day and 28 days separately due to the formation of hydration products between the two tests, which has a notable effect on the strength. A general comparison of 1-day OPC and PLC showed that more poorly crystalline carbonates in OPC concrete were associated with higher compressive strength. The higher strength of carbonated OPC produced during 28- day subsequent hydration was associated with the presence of more poorly crystalline carbonates. Therefore, it is likely that poorly crystalline carbonates yield a higher strength than polymorphs with higher crystallinity.

– 188 – SEM analysis The microstructure of OPC and PLC concretes was compared using SEM. The micrographs of hydrated reference of both concretes at age of 1 day are shown in Fig. 8.17. Ettringite needles were developed in similar form in OPC and PLC hydrated concretes with a C-S-H dominant amorphous microstructure. Similar micrographs of 28- day concrete are displayed in Fig. 8.18. Carbonation curing was carried out on OPC and PLC concretes (18a+4c). The micrograph comparison is shown in Fig. 8.19 after 1 day curing. While PLC showed crystalline forms ranging between 2 and 5 µm, the OPC carbonated sample displayed an amorphous microstructure. With mass loss over 720ºC, it was obvious that crystalline carbonates existed in the carbonated OPC concrete, but were not shown in SEM micrograph possibly due to their smaller size (nanometer). With no crystals identified in the hydrated PLC reference, it was conclusive that the newly formed crystals in the carbonated sample were due to carbonation and not initially present in the cement. In comparison to the needle-shaped ettringite that apparently formed from a single origin, the crystals seen in carbonated concrete formed in a random pattern scattered in a large area. Water compensation through spray affected the microstructure development of the carbonated concretes. Fig. 8.20 illustrates the changes observed in carbonated OPC and PLC concretes after 28 days subsequent hydration. Similar to 4-hour carbonated concrete after 1 day, PLC concrete after 28-day hydration with water compensation exhibited carbonate crystals with a rather porous microstructure. On the other hand, OPC concretes were more poorly crystalline, making contribution to higher compressive strength. The limestone particles ranging in size between 9-11 μm (Fig. 6.15) may serve as nuclei for calcium carbonate growth. The resulting CO2 uptake of PLC concrete is 22% based on Portland cement content. The size of carbonates in PLC concrete is in the range of 2-5 μm. They are well crystalline, formed right after carbonation and remain the same polymorph and size in subsequent hydration. Nevertheless, this network-type distribution of carbonate crystals results in a more porous microstructure. On the contrary, carbonated OPC concrete has shown lower amount of crystalline phases after 28 days subsequent hydration along with smaller sized crystalline carbonates in general. Even though the CO2

– 189 – uptake was similar, because of the much smaller carbonate particles with the dominant amorphous microstructure, carbonated OPC concrete has gained more strength.

Carbonation curing model To understand the improved freeze-thaw resistance of carbonated concrete, an effort was made to model the carbonation curing process and the microstructure of the concrete so produced. The schematic of the model is shown in Fig. 8.21. The process includes initial air curing, carbonation curing, and subsequent hydration (18a+4c+sp). Fig. 8.21 (a) shows the close-up of one typical unit in concrete, cement, aggregate, and aggregate-cement interface, near the surface of concrete slab where carbon dioxide diffusion takes place. Immediately after casting, the capillary pores are filled up with water and the interface between aggregate and cement paste is covered by a thin water film. After 18-hour initial air curing at 50% RH and 25ºC, as shown in Fig. 8.21 (b), cement is hydrated, 51% of total initial water (including water in lightweight slag aggregates) is evaporated, and capillary pore space are created. If only mixing water is considered (w/c=0.4), 91% mixing water has been removed. A thin water film is expected to remain on the surface of capillary pores and aggregate. The empty capillary pores create space for gas to penetrate and the carbonates to precipitate. Upon carbonation (Fig.

8.21 (c)), CO2 gas diffuses from surface through the porous path and dissolves in the water to produce carbonic acid, which in turn reacts with the calcium ions in water to form calcium carbonates. In comparison to the water film thickness, calcium carbonate forms a thicker film due to higher quantities of precipitation (20-25% by mass of cement). A similar phenomenon is observed at the aggregate surface. It is possible that the capillary space created during initial air curing may be filled due to calcium carbonate precipitation. It is most likely to happen near surface. With water loss of 91% by mass of 3 mixing water, 32 cm of pore space are created. However, a 24% CO2 uptake can only occupy 17 cm3, leaving approximately half empty. After 28-day subsequent hydration with or without water compensation in carbonated concrete (18a+4c+sp or 18a+4c), hydration continues and hydration products start to fill capillary pores as depicted in Fig. 8.21 (d). The final reaction product is an intermingled carbonation and hydration

– 190 – products. The remaining empty pore space can serve as free space for freezing water to move in to avoid frost expansion damage. The freeze-thaw behavior of concretes was explained using carbonation model. Hydrated reference and steam-cured concrete form a uniform cement matrix of hydration products, which provides exceptional compressive strength, but is not resistant to freeze- thaw expansive stresses. Both laboratory and commercial concretes are not air-entrained. On the other hand, carbonated concretes have demonstrated a homogenous microstructure with intermingled carbonation and hydration products. The carbonate-modified C-S-H structure works as a particulate reinforced C-S-H composite. This reinforcement and densification happen more on the surface than in the core, resulting in more resistance to permeation. The water loss in initial air curing creates more free space and might not be fully filled up by carbonates and hydrates. The resulting concretes will have capillary pores coated by calcium carbonate, which will serve as carbonate-coated entrained air, and is attributed to the excellent freeze-thaw resistance. In case of hydrated concrete, including steamed and commercial concretes, the lack of entrained air in paste makes concrete weak in freeze-thaw resistance. The particulate carbonate-reinforced CSH has a critical effect on the concrete and is seen more on the durability than on the strength. It is interesting to point out that the absorption calculated based on ASTM C140 are relatively of close values for both carbonated and hydrated concretes (Figs 8.3). This is suggestive that the freeze-thaw performance of the concrete is not dominated by voids in porous masonry concrete, but by the microstructure of binding matrix. ASTM C1262 for freeze- thaw resistance is a severe test method. Concretes after carbonation can perform much better in such a severe condition. It is conclusive that while hydrated concrete without entrained air provides high compressive strength, it lacks in resistance to freeze-thaw. Steam curing produced early strength due to accelerated hydration, but also produced much pore space, which reduced the 28-day strength and the resistance to freeze-thaw. Carbonation with water compensation, however, has shown to improve freeze-thaw resistance while maintaining a comparable strength. It appears that pore space produced by initial air curing and coated by carbonation products resulted in the highest resistance to freeze-thaw attack.

– 191 – CONCLUSIONS From the comparison of physical properties and microstructure of OPC and PLC concretes, the following conclusions can be drawn: 1. Hydrated OPC concrete has shown a compressive strength 27% higher than hydrated PLC concrete, even though PLC is ground finer than OPC. With 15% clinker replaced by limestone, PLC experienced a lower degree of hydration, which was associated with a less binding phase. Compressive strength of carbonated OPC concrete also exceeded that of PLC by 20-25%. The addition of limestone particles to the initial cement mix in PLC yielded the formation of crystalline carbonates, which may also be attributed to the strength reduction. It appeared that more poorly crystalline carbonates and smaller sized crystalline carbonates produced by carbonation of OPC led to a higher compressive strength. 2. Concrete durability through freeze-thaw resistance was measured for hydrated and carbonated concretes. There was no considerable difference between the two types of concrete as far as the frost resistance is concerned. However, for both concretes (PLC and OPC), the carbonated specimens lost 15% of initial mass after 10 cycles, while hydrated references and steam-cured concrete reached that threshold limit in 3-4 cycles. Carbonated concrete is much more resistant to freeze-thaw damage. The capillary space created by water loss during the carbonation reaction and carbonation products precipitation on the surface of those capillaries enhanced the freeze-thaw resistance by 300%. ASTM C1262 for freeze-thaw resistance in saline solution is a very severe test method for concrete masonry units. 3. Carbonation of OPC and PLC concrete consumed calcium silicates and calcium hydroxide. Ettringite was significantly reduced after carbonation. Aragonite, vaterite, and calcite were the carbonation products after 1 day. In subsequent 28- day hydration, phase transformation was observed in OPC, where poorly crystalline carbonates designated by aragonite and vaterite were transformed into well crystalline calcite. Nevertheless, this transformation was not seen in PLC concrete.

– 192 – 4. The carbonation model showed that the proposed curing regimen consisting of initial air curing, carbonation, and subsequent hydration with water compensation could provide the concrete with a carbonate-reinforced cement matrix. The freeze- thaw resistance was improved with comparable compressive strength, owing to the formation of a densified homogenous microstructure of carbonation and hydration products. Initial air curing is necessary if high carbonation reaction efficiency is desired. The created free space which is not filled by reaction products may be beneficial to freeze-thaw resistance as entrained air if coated with a carbonation product layer. It has shown no negative effect on strength. However, in steam-cured concrete, these uncoated voids have shown to reduce both strength and freeze-thaw resistance.

ACKNOWLEDGEMENTS The authors gratefully acknowledge the support by the Natural Science and Engineering Research Council (NSERC) of Canada, Canadian Concrete Masonry Producers Association (CCMPA), and Lafarge Canada.

REFERENCES ASTM (2004a). "Standard Test Method for Evaulating the Freeze-Thaw Durability of Dry-Cast Segmental Retaining Wall Units and Related Concrete Units." ASTM International.

ASTM (2004b). "Standard Test Methods for Sampling and Testing Concrete Masonry Units and Related Units." ASTM International.

Berger, R. L., and McGregor, J. D. (1973). "Effect of temperature and water-solid ratio on growth of Ca(OH)2 crystals formed during hydration of Ca3SiO5." Journal of American Ceramic Society 56(2): 73-79.

Bonavetti, V., Donza, H., Menendez, G., Cabrera, O., Irassar, E. F. (2003). "Limestone filler cement in low w/c concrete: A rational use of energy." Cement and Concrete Research 33(6): 865-871.

Collier, N. C., Sharp, J. H., Milestone, N. B., Hill, J., and Godfrey, I. H. (2008). "The influence of water removal techniques on the composition and microstructure of hardened cement pastes." Cement and Concrete Research 38: 737-744.

– 193 –

CSA (2008). Cementitious materials compendium, CSA. CSA-A3000.

Li, D., Shi, H., Deng, J., and Xu, Y. (2003). "Study on the thermal decomposition kinetics of nano-sized calcium carbonates." Journal of Zhejiang University 4(3): 363-368.

Parrott, L. J. (1996). "Some effects of cement and curing upon carbonation and reinforcement corrosion in concrete." Materials and Structures 29: 164-173.

Pera, J., Husson, S., Guilhot, B. (1999). "Influence of finely ground limestone on cement hydration." Cement and Concrete Composites 21(2): 99-105.

Ramachandran, V. S., and Beaudoin, J. J. (2001). Handbook of Analytical Techniques in Concrete Science and Technology, William Andre Publishing/Noyes

Rostami, V., Shao, Y., and Boyd, A. J. (2011). "Durability of concrete pipes subjected to combined steam and carbonation curing." Construction and Building Materials 25: 3345-3355.

Sawada, K. (1997). "The mechanism of crystallization and transformation of calcium carbonates." Pure and Appl. Chem. 69(5): 921-928.

Shideler, J. J. (1955). "Investigation of the moisture-volume stability of concrete masonry units." Portland Cement Association Research and Development Laboratories Bulletin D3.

Toennies, H. T. (1960). "Artificial carbonation of concrete masonry untis." American Concrete Institute Journal(56-42): 737-755.

Toennies, H. T., and Shideler, J. J. (1963). "Plant drying and carbonation of concrete block - NCMA-PCA cooperative program." American Concrete Institute Journal 60(33): 617-632.

– 194 – TABLES Table 8.1: Chemical composition of OPC and PLC concretes

Constituent (%) Blaine Fineness 2 Cement CaO SiO2 Al2O3 Fe2O3 MgO Na2O CO2 SO3 (m /kg) OPC 63.10 19.80 4.90 2.00 2.00 0.85 1.72 3.80 390 PLC 59.81 21.64 4.50 2.67 1.70 0.78 6.97 3.20 500

Table 8.2: CO2 uptake and content (%) of OPC and PLC concretes

CO2 Uptake CO2 Content CO2 Mass Mass Furnace Uptake Cement Sample # Gain Curve Analysis TG/DTG HCl Average 0a (Ref) - - 1.9 4.3 4.9 0 OPC 18a+4c 24.2 24.4 27.4 26.8 27.3 23.80 18a+4c+sp 25.7 24.2 27.7 28.9 27.9 24.66 0a (Ref) - - 6.52 6.62 7.01 0 PLC 18a+4c 18.34 18.23 25.31 26.06 25.98 18.75 18a+4c+sp 18.29 18.09 25.85 26.54 26.55 19.03

Table 8.3: Percentage of produced poorly and well crystalline carbonated in OPC and PLC concretes Poorly crystalline Well Crystalline Cement Sample 1-day 28-day 1-day 28-day 18a+4c 72.6 51.9 27.4 48.1 OPC 18a+4c+sp 74.3 61.0 25.7 39.0 18a+4c 50.5 47.4 49.5 52.6 PLC 18a+4c+sp 56.1 52.7 43.9 47.3

– 195 – FIGURES

Fig. 8.1: Schematic of freeze-thaw test

30 OPC 1-day PLC 1-day OPC 28-day PLC 28-day 25

20 18.2 17.9 15.3 14.1 14.3 15 12.8 9.9 10.510.2 8.8 10 7.9 7.0

5 Compressive Strength (MPa) CompressiveStrength

0 0a (Ref) 18a+4c 18a+4c+sp Sample #

Fig. 8.2: Compressive strength of OPC and PLC concretes

– 196 – 20 OPC PLC

15 13.8 13.0 12.2 12.8 11.6 11.3 10

5 Water Absorption (%) WaterAbsorption

0 0a (Ref) 18a+4c 18a+4c+sp Sample #

Fig. 8.3: Water absorption of 28-day OPC and PLC concretes

Fig. 8.4: Freeze-thaw resistance of 28-day OPC and PLC concretes

– 197 –

Fig. 8.5: Hydrated OPC concrete after 6 freeze-thaw cycles (0a-Ref)

Fig. 8.6: Hydrated PLC concrete after 6 freeze-thaw cycles (0a-Ref)

– 198 –

Fig. 8.7: Commercial concrete block after 6 freeze-thaw cycles

Fig. 8.8: Steamed OPC concrete after 6 freeze-thaw cycles (2a+4s)

– 199 –

Fig. 8.9: Carbonated OPC concrete after 10 freeze-thaw cycles (18a+4c)

Fig. 8.10: Carbonated OPC concrete with water compensation after 10 freeze-thaw cycles (18a+4c+sp)

– 200 –

Fig. 8.11: Carbonated PLC concrete after 10 freeze-thaw cycles (18a+4c)

Fig. 8.12: Carbonated PLC concrete with water compensation after 10 freeze-thaw cycles (18a+4c+sp)

– 201 –

Fig. 8.13: XRD patterns of 1-day OPC and PLC concretes

Fig. 8.14: XRD patterns of 28-day OPC and PLC concretes

– 202 –

Fig. 8.15: DTG curves of 1-day OPC and PLC concretes

Fig. 8.16: DTG curves of 28-day OPC and PLC concretes

– 203 –

Fig. 8.17: SEM micrograph of OPC and PLC hydrated reference (0a) after 1 day

Fig. 8.18: SEM micrograph of OPC and PLC hydrated reference (0a) after 28 days

Fig. 8.19: SEM micrograph of carbonated OPC and PLC concrete (18a+4c) after 1 day

– 204 –

Fig. 8.20: SEM micrograph of carbonated OPC and PLC concrete (18a+4c+sp) after 28 days

Fig. 8.21: Carbonation curing model: (a) After casting; (b) After initial air curing; (c) After carbonation; (d) After water compensation and subsequent hydration

– 205 – Chapter 9

–

CONCLUSIONS

Static and dynamic carbonation curing at early age were proposed as alternative methods to steam curing for lightweight concrete masonry units. The performance of concrete slab samples as well as 20-cm concrete masonry units exposed to carbonation was investigated. OPC and PLC concrete were compared in terms of CO2 uptake, compressive strength, and microstructural changes. To facilitate carbonation at early age, pre-conditioning with air curing of concrete was vital to remove excess free water and produce capillaries. Two types of initial hydration regimes were employed. Chapter 3-6 employed an environmental chamber (25ºC and 50% RH) for durations ranging from 4 to 18 hours, while Chapter 7 used an open-air curing concept, in which samples were air cured in open air (24±1ºC and 50±10% RH) for a duration of only 2 hours. The proposed carbonation curing system of Chapters 3-7 was addressed as either static or dynamic. During static carbonation (Chapters 3-6), the relative humidity of the chamber gradually increased to a point of water condensation on the chamber walls. At this point, carbonation reactivity would hinder due to excessive amount of water surrounding the concrete surface. To remove the released water, the dynamic system (Chapter 7) was employed using a circulation pump and a desiccant material. Further inspection of the carbonated samples showed improved performance, durability, and permanent sequestration of the CO2 gas. The current research proved that carbonation would be greatly beneficial in concrete masonry unit production. Based on the study, the following conclusions were drawn: 1. An attempt at the carbonation of fresh OPC concrete in Chapter 3 showed that excess surface free water limited the carbonation reactivity. Thus, initial air curing

was introduced from 4 to 18 hours to facilitate the penetration of CO2 and promote carbonation. Since concrete internal relative humidity could not be reduced to an ideal 50-60%, water loss was chosen as a more reliable parameter to

– 206 – justify the condition for carbonation. The respective water loss during 4-8 hours and 18 hours of initial air curing for OPC concrete was 32-33% and 51%. Further 5-10% water was lost during 4-hour carbonation of OPC concrete. In addition, Chapter 7 performed carbonation of OPC concrete in a dynamic system that combined initial air curing with carbonation. The total mass loss during 2 hours of open-air curing followed by 4-hour dynamic carbonation reached 39%. 2. To produce truly green concrete blocks, PLC was used as a binder in the carbonated concrete. Chapter 5 discussed the possibility of carbonation of PLC concrete. Similar to OPC concrete, PLC underwent an initial air curing process of 4-8 hours and 18 hours, and resulted in a water loss of 37-39% and 51% respectively. Upon subsequent carbonation, 4-6% of initial water was lost due to the exothermic reaction. Lower water loss during carbonation of PLC concrete in comparison to OPC was associated with a lower reactivity.

3. The CO2 uptake was greatly influenced by initial air curing. OPC concrete was carbonated for 4 hours after 4-18 hours of initial air curing, and the resulting uptake ranged between 21-25%. Longer exposure to carbonation to up to 4 days increased the uptake to 35% with a carbonation efficiency of 70%. It was conclusive that longer air curing permitted the evaporation of free water and the

production of capillary pores. This mechanism allowed the penetration of CO2 gas and the precipitation of carbonation products. In addition, the dynamic carbonation of OPC concrete was performed for 4 and 18 hours, resulting in a respective uptake of 13% and 22%. It was conclusive that the dynamic system could be as efficient as the static system with a combination of air curing and carbonation under controlled relative humidity. In comparison to freshly

carbonated OPC concrete (0a+4c), the CO2 uptake was increased through dynamic carbonation by 50% owing to the constant RH, which ensured an ideal condition for carbonation. 4. A similar static carbonation mechanism was employed in the carbonation of PLC concrete, but the uptake ranged from 16% to 18% for similar curing conditions. Substitution of cement clinker, which represents the reactants in the carbonation reaction, produced concrete with less degree of carbonation.

– 207 – 5. The effect of initial air curing and carbonation on compressive strength of OPC and PLC was investigated. Hydrated reference without air curing had the highest strength among all other hydrated concrete. With 18-hour air curing followed by 4-hour carbonation (18a+4c), the carbonated concrete could not maintain a comparable late strength due to the formation of capillary pores, which were partially filled with carbonation products. However, with water compensation, other parts of this pore space were occupied by hydration products and the strength was comparable to the hydrated. In comparison to hydrated OPC concrete, PLC strength was lower by 20-30% at 1 and 28 days. Limestone replaced 15% of the cement clinker, which represents the hydration reactants, thus producing less hydration products. It was also possible that the crushed limestone absorbed part of the mixing water required for subsequent hydration. The addition of a ternary material such as silica fume could improve the strength. 6. Unlike cement paste and mortar, concrete mixes contain more than 80% aggregates, which can lead to distorted results in microstructure analysis. A technique was developed to determine the cement content (CC) in the concrete powder. This method would allow the quantification of hydration and carbonation products using TG. Findings of the TG analysis showed that carbonates produced in OPC concrete were more in the poorly crystalline form than those in PLC. The size of carbonates differed in the two concretes ranging down to nanometers for the former and up to 5 microns for the latter. More poorly crystalline and smaller well crystalline carbonates were associated with the higher strength of OPC concrete. The 28-day subsequent hydration of carbonated OPC concrete resulted in a transformation of poorly crystalline aragonite and vaterite into well crystalline calcite. PLC concrete, on the other hand, did not show such a transformation, but the produced carbonates remained in the same degree of crystallinity. However, the final carbonates produced in the carbonation of OPC decomposed at higher temperatures than those of PLC, indicating a higher degree of crystallinity. In the dynamic carbonation system, the carbonates produced in 4- hour dynamic carbonation were more stable and had less tendency to transform into well crystalline than those produced after 18-hour.

– 208 – 7. While consuming calcium hydroxide, calcium silicates, and ettringite, carbonation had produced poorly crystalline aragonite and vaterite and well crystalline calcite. However, in case of carbonated OPC concrete (static and dynamic), the final carbonation products were found to be intermingled with hydration products. A discussion on the constancy of the ratio of poorly bound water in C-S-H (105- 200ºC) to that of poorly crystalline carbonates (470-720ºC) showed that these two components are unified as calcium silicate hydrocarbonate. PLC concrete had not shown this constancy, and it was conclusive that in PLC concrete the reaction products were intermingled but were not formed into one well-defined phase. 8. The freeze-thaw durability of concrete masonry units (CMU) subjected to hydration, carbonation, and steam curing was investigated. Although there was no significant difference in water absorption, the freeze-thaw resistance was quite different due to the curing scheme employed. Carbonated CMU had much higher freeze-thaw resistance than hydrated and steam counterparts, with the steam curing being the worst. It was found that the freeze-thaw resistance was not dominated by the voids in porous masonry units, but by the microstructure of the binding matrix. The complete curing process of initial air curing, carbonation, and subsequent hydration could provide the concrete with a carbonate-reinforced cement matrix. The created free space due to air curing, which was not filled by reaction products, may have been beneficial to freeze-thaw resistance as entrained air if it were coated with a carbonation product layer. It showed no negative effect on strength. However, in steam-cured concrete, these uncoated voids have shown to reduce both strength and freeze-thaw resistance. Carbonates produced in carbonation played a critical role in improving freeze-thaw resistance. 9. Chapter 3, 5, and 7 discussed the production of 20-cm CMU in the laboratory.

CO2 uptake and compressive strength were measured in order to realize whether the proposed static and dynamic carbonation system could be applied in industrial scale. The results showed similar results between concrete slab samples and masonry units, and therefore, it was conclusive that laboratory work on carbonation could be scaled up to commercial block production. Based on 24%

CO2 uptake, the CMU production in US and Canada is able to sequester 2 million

– 209 – tons of CO2 per year. In these two countries, the annual cement production is about 100 million tons per year with CO2 emission of about 80 million tons. The utilization and sequestration of 2 million tons of CO2 per year by CMU industry will make contribution to CO2 emission reduction by 2.5%.

– 210 – SUGGESTIONS FOR FUTURE WORK Based on the presented work, some suggested subjects require further study: 1. Static and dynamic carbonation were performed in a pressurized chamber at 1 bar above atmospheric. An attempt at dynamic carbonation at a pressure of 0.5 bar proved to have no significant effect on the reaction and results. Carbonation curing at atmospheric pressure would be more feasible in industrial practice. Further investigation is suggested to study the durability improvement performing innovative early-age carbonation at atmospheric pressure. Early carbonation at atmospheric pressure could offer a more attractive curing practice to the CMU production with minimum required changes in existing equipment. 2. Based on CMU literature review, the ideal relative humidity for carbonation ranged between 50-60%. The dynamic system used a manual monitoring system to control the relative humidity at 53%. It is proposed that other RH be considered within the previously mentioned range to optimize carbonation efficiency. 3. The dynamic system is a newly devised carbonation system. As a preliminary study of the effect of such a system on the reaction, only 1 block was carbonated per run. A detailed investigation should be performed in order to industrialize this method, including the development of an approach that would carbonate more blocks using the dynamic system and the possibility of recycling the remaining gas in the chamber for further carbonation.

4. The use of 99.5% pure CO2 may be energy intensive due to the extraction process. Flue gas utilization is simpler in that it can involve only gas compression and avoid the effort of recovering and transporting carbon dioxide. It is, therefore,

suggested that the use of flue gas be investigated further. Improving the CO2 uptake achievable with the use flue gas would greatly benefit the concept of carbonation curing. 5. The energy calculated in this study is preliminary and only to show a comparison among static carbonation and steam curing. It is proposed that if more blocks are cured in a larger system, the energy used for the environmental chamber or circulation pump would be much less. Therefore, it is suggested that low energy

– 211 – and low cost initial curing methods should be developed to make carbonation curing economically feasible, sustainable, and competitive to steam process. 6. PLC concrete samples showed less carbonation degree and compressive strength where it was expected to perform otherwise. Literature review showed that the size and fineness of the added limestone is of great importance. For this reason, it is suggested that future work would attend to the effect of carbonation of OPC concrete with added limestone of different sizes. Another approach would be to use more or less limestone (other than 14% in PLC) to study the effect of the quantity of limestone on early-age carbonation. The results also showed a lag in compressive strength of PLC concrete, which can be improved by using a ternary material system including silica fume. 7. While other research addressed cement, the present investigation studied lightweight concrete with expanded slag aggregates. Experiments performed on expanded shale or granite as coarse aggregates with river sand as fine aggregates showed lower uptakes and different compressive strengths. It is important to study the effect of carbonation on different materials system including the normal weight concrete mixes.

– 212 – STATEMENT OF ORIGINALITY This work presents a feasibility study on replacing conventional steam curing by an alternative carbonation curing in lightweight concrete masonry units production. Incorporating a carbonation treatment within a normalized 24-hour curing period, the study makes contributions to the technology advances relating to the following important findings: 1. A controlled carbonation process was developed for curing of fresh lightweight concrete masonry units (CMU) in a manner that promotes strength development, durability improvement, and permanent carbon dioxide storage and sequestration. The tested approach is cost-effective and contributes to ongoing global efforts aimed at alleviating anthropogenic emissions. Although the early and late-age strength of carbonated CMU is comparable to steam-cured concrete, the freeze- thaw resistance of carbonated CMU is much higher. It is possibly associated with the formation of hydro-carbonates in CSH and the capillary pore space created in initial air curing.

2. Capillary water on the surface of freshly prepared concrete hindered CO2 gas penetration and, hence, overall carbonation. Initial air curing was found to be

necessary to remove free water and better promote CO2 diffusion. 3. Both static and dynamic systems were established to further optimize the proposed carbonation curing method, and demonstrate commercial feasibility. Implementing electronic control systems in a dynamic system that regulate

relative humidity and the recycling of CO2 gas give way to better and more efficient sequestration especially for fresh concrete curing. 4. To promote the production of greener concrete products, Portland Limestone Cement was investigated as a substitute to conventional ordinary Portland cement (OPC) binder in CMU production. PLC contains less embodied energy as it comprises 15% limestone in its composition. 5. A unique method was developed to quantify the cement content in concrete powder for thermogravimetric analysis. Hydration and carbonation products in carbonated concrete can be quantified by the use of the method.

6. Carbonation efficiency was maximized and a CO2 uptake of 24% was achieved

– 213 – based on dry cement mass. If cement content in a single CMU is assumed to be

10% by mass, the total CO2 utilization and storage will reach 2 million tons per year by US and Canada CMU production. Based on an annual CO2 emission of 80 million tons from the production of 100 million tons of cement in these two countries, the CO2 emission can be reduced by 2.5% through carbonation process alone. The proposed carbonation process for CMU production has shown technical, economical, and environmental benefits.

– 214 –