A multi scale approach to the study of lime-cement mortars in masonry
Abstract
In masonry constructions, the choice of mortar composition is usually guided by requirements of the final application, which could range from new constructions to conservation projects. Often, lime and cement are combined, to overcome their shortcomings and consequently serve as a suitable binder in masonry mortars. Depending on their proportion in the mixture, it may be possible to obtain the desired range of characteristics in different mechanical properties like strength, stiffness, shrinkage, porosity, and so on. And even though the practice of combining lime and cement in masonry mortars has been around for many years, its benefits have not yet been addressed in a systematic, quantitative manner. Often, seemingly significant differences in the mechanical behavior of mortars do not reflect proportionally in changes in mechanical properties at the masonry level. Thus, the aim of this experimental research is focused on investigating the quantitative benefits of substituting cement with lime in masonry mortars, at the mortar as well as masonry level. Performance indicators have been determined from a structured experimental campaign of mechanical behavior of blended mortars, characterizing several properties for multiple lime-cement mix proportions: workability, compressive and flexural strength, stiffness, drying shrinkage, and open porosity, among others. The factor of aging has also been accounted for, with selected tests being performed up to 365 days, to account for the carbonation of lime in the mortars. Based on the breadth of experimental results that were obtained, patterns were analyzed through regression analyses, to estimate mechanical properties of mix proportions that were not physically tested in the laboratory. From the results obtained, the most suitable proportions were identified and consequently subjected to further experimentation at the masonry level. The response of brick masonry constructed with two different lime-cement mortars was compared with that of cement mortar in masonry, specifically focusing on differences in strength, stiffness, and ductility. In parallel, the influence of lime-cement mortars on the flexural strength of masonry, parallel, and perpendicular to the bed joints was also assessed. The final stage of this research involved quasi-static cyclic loading, to study the in-plane shear response of masonry wall panels, supplemented by experimental information on the shear bond strength of masonry, all focused on assessing the influence of lime-cement mortars compared to a cement mortar.
Full text
Universidade do Minho Escola de Engenharia Meera Ramesh March 2021 A multi scale approach to the study of lime-cement mortars in masonry Meera Ramesh A multi scale approach to the study of lime-cement mortars in masonry UMinho|2021
March 2021 Work conducted under supervision of Professor Doutor Miguel Ângelo Dias Azenha and Paulo José Brandão Barbosa Lourenço Professor Doutor Doctoral Thesis Civil Engineering Universidade do Minho Escola de Engenharia Meera Ramesh A multi scale approach to the study of lime-cement mortars in masonry Universidade do Minho Escola de Engenharia
A multi scale approach to the study of lime-cement mortars in masonry ii DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Licença concedida aos utilizadores deste trabalho Atribuição CC BY https://creativecommons.org/licenses/by/4.0/
A multi scale approach to the study of lime-cement mortars in masonry iii Acknowledgments This journey has been extremely gratifying and deeply joyful, and I am grateful to many wonderful people who have helped me through it. I would like to start by acknowledging Professor Miguel Azenha. In the last four years, he has helped shape not only this work but my approach to research in general. Not once did he shoot down any idea, regardless of how crazy it seemed, and allowed me to discover my dead ends, while gently pushing the research back on track when I got carried away. He has been honest and kind in equal measures and has insisted on paying attention to details. I am equally grateful to Professor Paulo Lourenco, who was one of the first people I sought advice from about pursuing a Ph.D. He has always taken out the time to guide the project and given invaluable technical feedback. By example, he has taught me to network and collaborate and to never lose the bird’s eye view of a project. With both of them, I hope to maintain the mentor-mentee relationship for the rest of my professional career, since it is one of the most valuable things I have obtained from this Ph.D. I would like to acknowledge FCT Portugal for my scholarship and the mortar task force of the European Lime Association, for funding a part of this Ph.D. I would especially like to thank Dr. Peter, Mr. Givens, and Dr. Rompaey, who have guided this research through important technical discussions and crucial feedback. This Ph.D. would have been impossible without the lab technicians in the University of Minho – Marco, Mr. Matos, Mr. Martins, Carlos, Mr. Goncalves, Cesar, and Luciano. I’m ever so grateful to each of them, they’ve all been generous with their time, gracious and kind, and a lot of fun to work with. I would like to thank my friends in Guimaraes, who are now family - Xinyu, Antonio, Telma, Rafael, Pilar, Leslie, Alberto, Elesban, Giorgos, Ioana, and Nicoleta. Things have a way of fading into the fabric of time, and I am certain I will not be able to carry every part of this journey with me. So what I hope remains, are the relationships I have developed here, all the wonderful people I became friends with, and those that I met in passing. They have all made this journey more fun and meaningful. Finally, I would like to thank some of the most important people in my life. My parents, Meenakshi and Ramesh, who have given me wings to fly and roots to come back to, who encourage me to travel and to learn new things, to take risks, fall, and start over. They are my foundation. My sister, Padmini, who is half my heart and encourages me to be more passionate and caring. And to Soham, who has patiently walked this journey with me, adding humor to dull days, and comfort to the difficult ones. I feel extremely privileged and thankful for all the love that these people bring into my life.
A multi scale approach to the study of lime-cement mortars in masonry iv STATEMENT OF INTEGRITY I hereby declare having conducted my thesis with integrity. I confirm that I have not used plagiarism or any form of falsification of results in the process of the thesis elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho.
A multi scale approach to the study of lime-cement mortars in masonry v A multi scale approach to the study of lime-cement mortars in masonry Abstract In masonry constructions, the choice of mortar composition is usually guided by requirements of the final application, which could range from new constructions to conservation projects. Often, lime and cement are combined, to overcome their shortcomings and consequently serve as a suitable binder in masonry mortars. Depending on their proportion in the mixture, it may be possible to obtain the desired range of characteristics in different mechanical properties like strength, stiffness, shrinkage, porosity, and so on. And even though the practice of combining lime and cement in masonry mortars has been around for many years, its benefits have not yet been addressed in a systematic, quantitative manner. Often, seemingly significant differences in the mechanical behavior of mortars do not reflect proportionally in changes in mechanical properties at the masonry level. Thus, the aim of this experimental research is focused on investigating the quantitative benefits of substituting cement with lime in masonry mortars, at the mortar as well as masonry level. Performance indicators have been determined from a structured experimental campaign of mechanical behavior of blended mortars, characterizing several properties for multiple lime-cement mix proportions: workability, compressive and flexural strength, stiffness, drying shrinkage, and open porosity, among others. The factor of aging has also been accounted for, with selected tests being performed up to 365 days, to account for the carbonation of lime in the mortars. Based on the breadth of experimental results that were obtained, patterns were analyzed through regression analyses, to estimate mechanical properties of mix proportions that were not physically tested in the laboratory. From the results obtained, the most suitable proportions were identified and consequently subjected to further experimentation at the masonry level. The response of brick masonry constructed with two different lime-cement mortars was compared with that of cement mortar in masonry, specifically focusing on differences in strength, stiffness, and ductility. In parallel, the influence of lime-cement mortars on the flexural strength of masonry, parallel, and perpendicular to the bed joints was also assessed. The final stage of this research involved quasi-static cyclic loading, to study the in-plane shear response of masonry wall panels, supplemented by experimental information on the shear bond strength of masonry, all focused on assessing the influence of lime-cement mortars compared to a cement mortar. Key words: Cyclic loads; E-modulus; Lime-cement mortars; Mechanical behavior; Unreinforced brick masonry;
A multi scale approach to the study of lime-cement mortars in masonry vi Abordagem multi-escala ao estudo de argamassas mistas de cal e cimento em alvenaria Resumo Nas construções em alvenaria, a escolha da composição da argamassa costuma ser orientada pelos requisitos da aplicação final, seja na construção nova ou reabilitação. É frequente o recurso à combinação de cal e cimento para superar respetivas deficiências individuais obtendo-se um ligante adequado em argamassas de alvenaria. Dependendo das proporções de cal e cimento na mistura, pode ser possível obter uma faixa desejada de características em diferentes propriedades. No entanto, apesar da prática de combinar cal e cimento em argamassas de alvenaria já existir há muitos anos, os seus benefícios ainda não foram analisados de forma sistemática e quantitativa. As diferenças significativas no comportamento mecânico das argamassas não refletem necessariamente alterações proporcionais nas propriedades mecânicas no nível da alvenaria. Assim, o objetivo desta investigação experimental centrase na investigação quantitativa dos benefícios decorrentes da substituição parcial do cimento por cal em argamassas de alvenaria, tanto ao nível da argamassa como ao nível do comportamento da alvenaria. Os indicadores de desempenho foram determinados a partir de uma campanha experimental focada no comportamento mecânico de argamassas mistas de cal e cimento, caracterizando diversas propriedades para múltiplas proporções de mistura cal-cimento: trabalhabilidade, resistência à compressão e flexão, módulo de elasticidade, retração de secagem e porosidade aberta, entre outras. O envelhecimento também foi contabilizado, através de realização de ensaios até aos 365 dias de idade. Com base na amplitude dos resultados experimentais obtidos, foram efetuadas análises de regressão para estimar as propriedades mecânicas das proporções da mistura que não foram testadas fisicamente em laboratório. Foram identificadas as proporções cal-cimento mais adequadas e, consequentemente, escolhidas para o programa experimental à escala da alvenaria. A resposta de provetes de alvenaria construídos com duas argamassas mistas contendo razões cal-cimento diferentes foi comparada com a da alvenaria executada com argamassa de cimento. Paralelamente, foi avaliada a influência das argamassas mistas de cimento-cal na resistência à flexão de provetes de alvenaria. Finalmente, foi estudada a resposta ao corte no plano de painéis de parede de alvenaria com carregamento cíclico quase estático, com enfoque na avaliação da influência das argamassas de cimento-cal comparativamente à argamassa de cimento. Palavras-chave: Alvenaria de tijolo não armada Argamassas; Argamassas mistas de cal e cimento; Carregamento cíclico; Comportamento mecânico; Módulo de Elasticidade;
A multi scale approach to the study of lime-cement mortars in masonry vii Contents Acknowledgments ............................................................................................................................... iii Abstract............................................................................................................................................... v Resumo.............................................................................................................................................. vi List of symbols .................................................................................................................................. xii List of abbreviations ........................................................................................................................... xv List of chemical compounds and phases ............................................................................................ xvi List of figures .................................................................................................................................... xvii List of tables ..................................................................................................................................... xxii 1. Introduction ................................................................................................................................ 1 1.1 Problem statement .............................................................................................................. 1 1.2 Objectives of research ......................................................................................................... 2 1.2.1 Mortar level ................................................................................................................. 2 1.2.2 Masonry level .............................................................................................................. 2 1.3 Scope of research ............................................................................................................... 3 1.4 Methodology of research ..................................................................................................... 4 1.4.1 Research question ....................................................................................................... 4 1.4.2 Variables used ............................................................................................................. 4 1.5 Outline of research .............................................................................................................. 6 2. Lime-cement mortars and their role in masonry ........................................................................... 8 2.1 Introduction......................................................................................................................... 8 2.2 Binders ............................................................................................................................... 8 2.2.1 Cement and hydration ................................................................................................. 8 2.2.2 Air lime and carbonation ............................................................................................ 11 2.3 Lime-cement blended (masonry) mortars ........................................................................... 13 2.3.1 Lime carbonation and cement hydration..................................................................... 14
A multi scale approach to the study of lime-cement mortars in masonry xiv τc Shear strength associated with the cracked section τw Shear strength associated with the whole section fvo Initial shear strength fvko Characteristic initial shear strength fxk1 Characteristic flexural strength, parallel to bed joints fxk2 Characteristic flexural strength, perpendicular to bed joints c Cohesion fvk Shear bond strength γm Partial safety factor ftb Tensile strength of brick Aenv Area under the envelope Hdmax Horizontal force corresponding to maximum displacement dmax Maximum displacement Hcr Lateral force at which the first significant crack occurs Hmax Maximum lateral force dcr Displacement at which the first significant crack occurs dHmax Displacement corresponding to maximum lateral force Hu Ultimate lateral resistance Driftcr Drift at cracking DriftHmax Drift at maximum horizontal capacity Driftmax Drift at maximum displacement EN Normalized cumulative energy Edissipated Energy dissipated in each cycle n Load cycle Hf Lateral flexural capacity r Ratio of height to length of wall Hs Maximum shear capacity ᵧ Safety factor ff,beam Flexural strength corresponding to beam ff,joint Flexural strength corresponding to joint stdevf,joint Standard deviation corresponding to flexural strength of a joint stdevf,beam Standard deviation corresponding to flexural strength of a beam
A multi scale approach to the study of lime-cement mortars in masonry xv j Vertical stress distribution at compressed toe List of abbreviations EN European Standards (Norms) BS British Standards ASTM American Society for Testing and Materials MIP Mercury Intrusion Porosimetry SEM Scanning Electron Microscopy EMM-ARM Elasticity Modulus Measurement through Ambient Response Method IUPAC International Union of Pure and Applied Chemistry TGA Thermogravimetric Analysis UPV Ultrasound Pulse Velocity LOI Loss on ignition PSD Particle size distribution RH Relative humidity IRA Initial rate of absorption y Binder to aggregate, % by volume x Lime content in binder (% by volume) BET Brunauer–Emmett–Teller w/b Water to binder CEM Cement conforming to EN 197-1 IO Immediate Occupancy LS Life Safety CP Collapse Prevention LVDT Linear Variable Differential Transformer UMinho University of Minho MTF Mortar task force EULA European Lime Association CoV Coefficient of variation TSTM Temperature Stress Testing Machine BTJASPE BéTon au Jeune Age, Suivi de la Prise et du module d'Elasticité
A multi scale approach to the study of lime-cement mortars in masonry xvi List of chemical compounds and phases SO3 Sulfur trioxide MgO Magnesium oxide Al2O3 Aluminum oxide Fe2O3 Ferric oxide K2O Potassium oxide SiO2 Silicon dioxide CaO Calcium oxide C2S Dicalcium silicate (Belite) C3S Tricalcium silicate (Alite) TiO2 Titanium dioxide C3A Tricalcium Aluminate C4AF Tetracalcium alumino ferrite AFt Calcium trisulfoaluminate hydrate (Ettringite) AFm Calcium monosulfoaluminate C-S-H Calcium silicate hydrate CH Portlandite Ca(OH)2 Calcium hydroxide Ca Calcium Si Silicon CaCO3 Calcium carbonate CO2 Carbon dioxide C-A-H Calcium Aluminate Hydrate Al Aluminum O Oxygen
A multi scale approach to the study of lime-cement mortars in masonry xvii List of figures Figure 1: Dependent, independent, and moderating variables for mortar level research ........................ 5 Figure 2: Dependent, independent, and moderating variables for masonry level research ..................... 5 Figure 3: Depiction of the process of cement production ..................................................................... 9 Figure 4: Typical composition of Portland cement (CEM I) [33] .......................................................... 10 Figure 5: Illustration of the lime cycle ................................................................................................ 12 Figure 6: Flow value between 155-185 mm for a lime-cement mortar, measured on a flow table according to EN 1015-3 [75] ............................................................................................................................ 16 Figure 7: Cyclic compression, typically used to measure E-modulus................................................... 19 Figure 8: Ottosen model and modification [118] ................................................................................ 22 Figure 9: Vertical and lateral loads on masonry ................................................................................. 28 Figure 10:Depiction of stresses in masonry when unit is stiffer than the mortar [129] ........................ 31 Figure 11: Flexural strength of masonry parallel (𝑓𝑥𝑘1) and perpendicular (𝑓𝑥𝑘2) to bed joints [271] ........................................................................................................................................................ 34 Figure 12: Flexural bond strength of horizontal bed joints in masonry using a bond wrench [195] ...... 35 Figure 13: Triplet masonry specimens used to determine the initial shear strength of masonry .......... 37 Figure 14: In-plane failure mechanisms of masonry subject to combined vertical and horizontal loading ........................................................................................................................................................ 39 Figure 15: Packaging and storage of binders – lime and cement ....................................................... 46 Figure 16: TGA of the binders - lime and cement, at the end of the research period ........................... 47 Figure 17: PSD of aggregates used in experiments for mortar level studies, the results are in Chapter 4 ........................................................................................................................................................ 49 Figure 18: Modified particle size distribution of aggregate (siliceous sand) used in experiments involving masonry specimens, the results of which have been discussed in Chapter 5 ..................................... 50 Figure 19: Clay bricks considered as options for this research campaign, corresponding to Table 7 ... 51 Figure 20: Solid molded clay brick with frogs chosen for the project, supplied by Wienerberger .......... 52 Figure 21: Equipment used in the laboratory for casting mortar mixes, from the brand Matest (a) Mixer E 093 [340]; (b) Flow table E 090 [339]; (c) Mold E 105 [341]; (d) Jolting apparatus E 130 [342]; ...... 53 Figure 22: Overview of casting of lime-cement mortars according to EN 196-1 [344] and curing according to EN 1015-11 [16] .......................................................................................................................... 54 Figure 23: Target consistency aimed for 175±10 mm, for all mortars ................................................ 56
A multi scale approach to the study of lime-cement mortars in masonry xviii Figure 24: Summary of mixes tested for different mechanical properties, for mortar level research pertaining to Chapter 4 ..................................................................................................................... 58 Figure 25: Water expelled (a) during compaction from a trial of the reference cement mix (b) ............ 61 Figure 26: Illustration of process used to construct masonry ............................................................. 65 Figure 27: Time-lapse of construction of masonry specimens ............................................................ 66 Figure 28: Left – Mortar specimens constructed for in situ characterization with masonry | Right – Overview of masonry specimens cast for mechanical characterization ............................................... 67 Figure 29: Illustration summarizing tests performed for masonry level research ................................. 68 Figure 30: Water-binder ratio (by mass) of mixes expressed as a function of lime content in the binder, in the workability range of 175±10 mm, measured using the flow table test according to EN 1015-3 [75]. The values 1:3, 1:4, 1:5 and 1:6 indicate B/Ag ratios. The R-squared values shown are for individual B/Ag ratios (only one mix for 1:6). .................................................................................................... 72 Figure 31: Experimental water binder ratios (Table 13) versus estimated according to (a) Equation 10 (b) Equation 11 ..................................................................................................................................... 74 Figure 32: Images for (a) Flexural test (3-point bending) and (b) Uniaxial (unconfined) compression test for mortar specimens ....................................................................................................................... 75 Figure 33: Evolution of compressive strength with time for different binder-aggregate ratios from 7 to 365 days ................................................................................................................................................. 79 Figure 34: Evolution of flexural strength with time for different binder-aggregate ratios from 7 to 365 days ........................................................................................................................................................ 80 Figure 35: Evolution of normalized compressive strength with time for different binder-aggregate ratios from 7 to 365 days. Strength has been normalized with respect to compressive strength at 90 days .. 81 Figure 36: Evolution of normalized flexural strength with time for different binder-aggregate ratios from 7 to 365 days. Strength has been normalized with respect to flexural strength at 90 days ..................... 82 Figure 37: Evolution of compressive strength with time estimated according to Equation 13 .............. 83 Figure 38: Evolution of flexural strength with time estimated according to Equation 13....................... 84 Figure 39: Evolution of compressive strength with time estimated according to Equation 14 .............. 86 Figure 40: Evolution of flexural strength with time predicted according to Equation 14 ....................... 87 Figure 41: Graphical representation of Equation 14 – Evolution of compressive strength with time, normalized with respect to strength at day 7, for lime v/s cement dominant blended mixes ............... 88 Figure 42: Illustration of a linear relationship between the compressive strength of mortar and lime content in the binder (% by volume) for different B/Ag ratios and at different ages .............................. 89
A multi scale approach to the study of lime-cement mortars in masonry xix Figure 43: Illustration of a linear relationship between the flexural strength of mortar and lime content in the binder (% by volume) for different B/Ag ratios and at different ages .............................................. 89 Figure 44: Illustration of calculation of normalized slope using compressive strength and lime content in the binder as an example (Also applicable to flexural strength because of the linear relationship demonstrated in Figure 43) .............................................................................................................. 90 Figure 45: Illustration of a linear relationship between compressive strength of mortar and B/Ag ratio (% by volume) for lime contents in the binder and at different ages ......................................................... 92 Figure 46: Illustration of a linear relationship between flexural strength of mortar and B/Ag ratio (% by volume) for lime contents in the binder and at different ages ............................................................. 93 Figure 47: Illustration of calculation of normalized slope using flexural strength and B/Ag ratio as an example (Also applicable to compressive strength because of the linear relationship demonstrated in Figure 45) ........................................................................................................................................ 94 Figure 48: Predicted versus experimental values for mechanical strength, corresponding to Equation 15 without the correction factor: (a) 𝑓𝑐 compressive strength (b) 𝑓𝑓 flexural strength............................. 97 Figure 49: Depiction of measurement of ultrasound velocity (UPV) along the length (160 mm) of a mortar specimen ......................................................................................................................................... 97 Figure 50: Evolution of bulk density with time for different lime-cement mixes .................................... 99 Figure 51: Evolution of ultrasound pulse velocity (UPV) with time for different lime-cement mixes ..... 100 Figure 52: Correlation between ultrasound velocity and a function of density and compressive strength for varying lime content in the binder (10, 25, 33.3, 50, 66.7, 75) % by volume, for B/Ag ratio 1:3 101 Figure 53: Unconfined cyclic compression test for measurement of static E-modulus: (a) Load cycle (b) Setup of specimen .......................................................................................................................... 103 Figure 54: Evolution of E-modulus with time, as measured by cyclic compression test (B/Ag 1:3, by volume) .......................................................................................................................................... 105 Figure 55: Evolution of ratio of E-modulus to compressive strength (E/fc) with time ......................... 106 Figure 56: Measurement of Poisson’s ratio - Set up of the specimen with the horizontal layout of LVDT ...................................................................................................................................................... 107 Figure 57: Set up used to measure fracture energy ......................................................................... 108 Figure 58: Illustration of calculation of fracture energy ..................................................................... 109 Figure 59: Evolution of drying shrinkage with time for lime-cement mortars (B/Ag 1:3, by volume) ... 113 Figure 60: Illustration of the preparation for and set up of EMM-ARM to measure E-modulus (continuous monitoring) ..................................................................................................................................... 114
A multi scale approach to the study of lime-cement mortars in masonry xx Figure 61: Evolution of E-modulus of lime-cement mortars obtained from EMM-ARM test (B/Ag 1:3, by volume) .......................................................................................................................................... 115 Figure 62: (a) E-modulus versus lime content in the binder, % by volume (b) Normalized evolution of Emodulus with time for lime-cement mixes (B/Ag 1:3, by volume) ..................................................... 117 Figure 63: Comparison of values of E-modulus (GPa), obtained from EMM-ARM and cyclic compression test at 7 days of age ....................................................................................................................... 118 Figure 64: Mechanical strength of mortars (standard conditions) used for research on masonry ....... 123 Figure 65: Mechanical strength of mortars (in situ conditions) used for research on masonry ........... 124 Figure 66: Comparison of mechanical strength of mortars, used in masonry specimens in standard (Table 30) and in situ (Table 31) conditions at 28 and 90 days of curing age ............................................. 125 Figure 67: E-modulus for mortars (standard and in situ) .................................................................. 126 Figure 68: Different tests performed to characterize the mechanical properties of bricks .................. 128 Figure 69: Schematic representation of set-up used for testing compressive strength and E-modulus of masonry specimens........................................................................................................................ 130 Figure 70: Compressive strength and E-modulus of masonry wallets ............................................... 131 Figure 71: (a) Compressive strength – Masonry v/s mortar; (b) E-modulus – Masonry v/s mortar.... 132 Figure 72: Stress-strain curves for masonry specimens constructed with different mortars ............... 135 Figure 73: (Left) Averaged stress-strain curves; (Right) Normalized stress – averaged strain curves .. 135 Figure 74: Vertical strain at peak stress v/s maximum compressive strength of masonry (Absolute and normalized values).......................................................................................................................... 136 Figure 75: Vertical strain (corresponding to peak and yield stresses) v/s strength of mortars used ... 137 Figure 76: Set up used for the flexural test of masonry (Parallel to the bed joints and perpendicular to the bed joints) ...................................................................................................................................... 139 Figure 77: Force-displacement curves for masonry specimens with different mortars, parallel to bed joints (Note the different vertical scales in the graphs) .............................................................................. 141 Figure 78: Force-displacement curves for masonry specimens with different mortars, perpendicular to bed joints (Note the different vertical scales in the graphs) .............................................................. 141 Figure 79: Setup for testing shear bond strength of masonry ........................................................... 144 Figure 80: Shear stress versus relative slip for different masonry triplets .......................................... 144 Figure 81: Normal stress versus shear stress for masonry triplets with different mortars .................. 145 Figure 82: Maximum shear stress of masonry triplets as a function of compressive strength of mortar ...................................................................................................................................................... 147
A multi scale approach to the study of lime-cement mortars in masonry xxi Figure 83: Normal stress versus shear stress for masonry triplets with different mortars, common linear regression ...................................................................................................................................... 148 Figure 84: Setup used for in-plane cyclic loading of masonry specimens (Image from the laboratory, UMinho) ......................................................................................................................................... 149 Figure 85: Illustration of setup used for in-plane cyclic loading of masonry specimens ...................... 150 Figure 86: Horizontal deformations imposed for in-plane cyclic loading test, labelled with drift (%) .... 151 Figure 87: Lateral force versus lateral displacement in different masonry specimens subjected to in-plane cyclic loads ..................................................................................................................................... 152 Figure 88: Final crack patterns observed in masonry specimens with the same brick and different mortars subjected to in-plane cyclic loads, indicated by principal strains from DIC ........................................ 153 Figure 89: Experimental envelopes of lateral force v/s lateral displacements for the specimens tested ...................................................................................................................................................... 154 Figure 90: Bilinear idealization of the experimental force-displacement curve ................................... 155 Figure 91: (a) Lateral resistance, normalized with respect to maximum capacity v/s lateral drift (b) Stiffness degradation versus lateral drift .......................................................................................... 160 Figure 92: Energy dissipated v/s lateral drift for unreinforced masonry specimens with different mortars ...................................................................................................................................................... 161
A multi scale approach to the study of lime-cement mortars in masonry xxii List of tables Table 1: A summary of values of K, α, and β (Equation 4) as presented by different authors.............. 29 Table 2: Chemical analysis of main components of cement CEM I - 42.5R as provided by the manufacturer Secil ................................................................................................................................................. 45 Table 3: Particle size distribution (PSD) of lime CL 90 - S as provided by the manufacturer Lhoist ...... 46 Table 4: Chemical composition of lime CL 90 - S as provided by the manufacturer Lhoist .................. 46 Table 5: Details of TGA of binders (lime and cement) at the end of the research (Figure 16) .............. 47 Table 6: Information on particle size distribution (PSD) and chemical composition of siliceous filler added to aggregate ..................................................................................................................................... 49 Table 7: Mechanical characterization of clay bricks considered as options for this research campaign 51 Table 8: Composition of blended lime-cement mortars (for 1 m3 of mortar produced) ......................... 55 Table 9: Summary of tests conducted at the mortar level .................................................................. 58 Table 10: Compressive strength of mixes tested for repeatability ....................................................... 59 Table 11: Composition of mortars with modified aggregate (for 1 m3 of mortar produced) .................. 63 Table 12: Summary of tests conducted at the masonry level ............................................................. 67 Table 13: Water-binder ratio (by mass) for different mixes - used, predicted, and difference (%) .......... 73 Table 14: Compressive strength values of lime-cement mortars from 7 days to 365 days .................. 76 Table 15: Flexural strength values of lime-cement mortars from 7 days to 365 days .......................... 77 Table 16: Ratio of compressive strength to flexural strength for mixes of different ages ...................... 77 Table 17: Change (%) in compressive strength of lime-cement mixes for a unit change in lime content in the binder (% by volume) ................................................................................................................... 90 Table 18: Change (%) in flexural strength of lime-cement mixes for a unit change in lime content in the binder (% by volume) ........................................................................................................................ 91 Table 19: Change (%) in compressive strength of lime-cement mixes for every unit change in B/Ag ratio (by volume) ...................................................................................................................................... 94 Table 20: Change (%) in flexural strength of lime-cement mixes for every unit change in B/Ag ratio (% by volume) ............................................................................................................................................ 94 Table 21: Values of coefficients corresponding to equation 15 ........................................................... 96 Table 22: R-square values obtained from linear regression performed on mortars with varying lime contents, at different curing ages from 7-365 days and varying cases of fixed B/ag ratios ................ 102 Table 23: R-square values obtained from linear regression performed on mortars with varying B/Ag ratios, at different curing ages from 7-365 days and varying cases of fixed lime-cement ratios in the binder 102
A multi scale approach to the study of lime-cement mortars in masonry xxiii Table 24: Values of E-modulus as obtained from the cyclic compression test ................................... 104 Table 25: Values of Poisson's ratio and shear modulus for lime-cement mortars (7, 28, and 90 days of age) ............................................................................................................................................... 107 Table 26: Values of fracture energy for lime-cement mortars (7, 28, and 90 days of age) ................. 110 Table 27: Derived values of tensile strength and characteristic length for lime-cement mortars ......... 110 Table 28: Values of open porosity for lime cement mixes (7, 28, and 90 days) ................................ 111 Table 29: Comparison of E-modulus (GPa) at 7 days of age, obtained from EMM-ARM and cyclic compression test ............................................................................................................................ 117 Table 30: Mechanical (compressive and flexural) strength of standard mortars ................................ 123 Table 31: Mechanical (compressive and flexural) strength of in situ mortars .................................... 124 Table 32: E-modulus for mortars (standard and in situ) ................................................................... 125 Table 33: Mechanical characterization of bricks used to construct masonry ..................................... 129 Table 34: Compressive strength and E-modulus of masonry wallets ................................................. 131 Table 35: Values of coefficients for Equation 21 and comparison with Eurocode 6 ........................... 134 Table 36: Peak strain, corresponding secant stiffness, and ductility of different specimens ............... 135 Table 37: Estimated (equation 22) and experimental values of strain at peak stress ......................... 138 Table 38: Values of flexural strength of masonry in the directions parallel and perpendicular to the bed joints .............................................................................................................................................. 140 Table 39: Characteristic values of flexural strength and recommendations of Eurocode 6 ................. 142 Table 40: Values of flexural tensile bond strength of joint estimated from flexural strength measured parallel to bed joints ....................................................................................................................... 143 Table 41: Values of maximum shear stress obtained for masonry specimens for varying levels of vertical pre-compression ............................................................................................................................. 145 Table 42: Joint characteristics for different types of mortars used with brick masonry ...................... 146 Table 43: Values of cohesion obtained for different types of mortars, using the same coefficient of friction for all ............................................................................................................................................. 147 Table 44: Deformations imposed in the horizontal direction on specimens for in-plane cyclic loading 151 Table 45: Data obtained from experimental envelopes corresponding to three characteristic points; cracking (Hcr,dcr), maximum resistance capacity (Hmax,dHmax) and maximum displacement (Hdmax,dmax) ............................................................................................................................ 156 Table 46: Parameters used to describe the bilinear idealized envelopes ........................................... 157 Table 47: Averaged values of different parameters for in-plane cyclic loading test ............................. 157
A multi scale approach to the study of lime-cement mortars in masonry 4 Type of unit used to construct masonry It is acknowledged that the structural performance of masonry depends on the unit used for construction, the type of mortar, and their consequent interaction. However, since the focus of this research is on the mortar, only one type of unit has been selected for studying the mechanical behavior of masonry, i.e., solid frogged clay bricks with a high suction rate. This way, differences in the behavior of masonry could be attributed to the mortar, since all other parameters including type of unit, construction protocols, and the testing methods were the same. Standards used as guidelines for tests It is also recognized, that requirements of the structural performance of masonry vary based on the geographical region in which it is constructed. For this research, the standards used were European, with the EN 1015 series [16] used for the mortar specimens, EN 1052 series [17] adopted for the masonry specimens, and Eurocode 6 [18] for the design recommendations. 1.4 Methodology of research 1.4.1 Research question The global question in this research was to investigate the effect of partial replacement of cement with lime in mortars on the structural performance of unreinforced brick masonry. Improved structural performance of masonry, would depend on the specific property being considered and could mean higher material strength, greater deformation capacity, improved bond strength, and so on. 1.4.2 Variables used The variables used in any research can typically be categorized into three types namely dependent, independent, and moderating. Independent variables are those that are being studied and cause changes in the outcome. Dependent variables are the outcomes that are being studied and are of interest to the research. Moderating variables influence the intensity or strength of the relationship between dependent and independent variables. Since this research is split into different scales of study (mortar and masonry), the variables for each of them have been described below (Figure 1, Figure 2): ➢ Mortar level study 1. Independent variables – The lime-cement ratio in the binder or the amount of lime in the binder (10%, 25%, 33.3%, 50%, 66.7%, 75%, and 90%), and amount of binder in the mix or the B/Ag ratio (1:3, 1:4, 1:5, and 1:6), and time in terms of the number of days at which the mix was tested.
A multi scale approach to the study of lime-cement mortars in masonry 5 2. Dependent variables – All mechanical and physical properties that were studied at the mortar level, such as mechanical strength and stiffness, drying shrinkage, open porosity, and so on. 3. Moderating variable – The amount of water used in each of the mixes or w/b ratio, which was decided based on a target flow table value. Figure 1: Dependent, independent, and moderating variables for mortar level research ➢ Masonry level study 1. Independent variable – The type of mortar used to construct masonry. 2. Dependent variables – All mechanical properties studied such as compressive strength, flexural strength, shear bond strength, and so on. 3. Moderating variable – The type of unit used for masonry construction, in this research, was high suction solid frogged clay brick. Figure 2: Dependent, independent, and moderating variables for masonry level research
A multi scale approach to the study of lime-cement mortars in masonry 6 1.5 Outline of research The outline of the thesis has been highlighted below, with a summary of the contents of each of the chapters: Chapter 2 – Lime-cement mortars and their role in masonry This chapter is a state of the art report, which aims at providing the reader with context for the experiments performed in this research and the discussion on results obtained that ensue in the subsequent chapters. The first section covers a description of the two relevant binders - Portland cement and air lime, a summary of their chemistry and hardening mechanisms. The second section is a summary of limecement mortars. Different mechanical properties such as strength, stiffness, deformability, and porosity have been covered, and wherever available, the influence of the presence of lime on these properties has been discussed. And finally, the third section of this chapter is about the role of mortar in masonry focusing on the mechanical behavior of masonry as a function of lime content in the binder of the mortar. Chapter 3 – Materials and methods The discussion in this chapter covers the choice of materials used as binders and aggregates in the mortars, the unit selected for masonry construction, and curing conditions for mortar and masonry specimens. It also outlines the design of the mortar mixes, based on the workability (target flow table value) selected. The protocols used to cast and cure mortars, and to construct masonry have also been discussed in detail. Finally, this chapter presents a summary of the different experiments performed and the size and number of specimens used for each of the tests. Chapter 4 – Mechanical properties of lime-cement mortars This chapter discusses the results obtained from experiments performed on mortars. The water binder ratio of different mortars as a function of the lime content in the binder and the B/Ag ratio of the mortar has been presented. Subsequently, the results of mechanical strength, ultrasound pulse velocity, and density of 15 lime-cement mortars, have been analyzed. Experimental results have been supplemented with quantification of different factors such as lime-cement ratios, B/Ag ratios, and curing ages, and their impact on the strength of lime-cement mortars. Relations between compressive strength, flexural strength, ultrasound pulse velocity, and density, have also been explored. Thereafter other mechanical properties have been discussed including stiffness at very early ages (0 to 7 days), and up to 90 days, fracture energy, Poisson’s ratio, drying shrinkage, and open porosity. The focus
A multi scale approach to the study of lime-cement mortars in masonry 7 here was to explore the impact of the quantity of lime on those mechanical properties that are not studied frequently in literature, as well as to characterize a few commonly used lime-cement mixes, so that a range of expectable values could be presented, for future references. Chapter 5 – Influence of lime-cement mortars on the mechanical behavior of masonry This chapter begins with a characterization of two lime-cement mixes chosen from the study in the previous chapter, and introduces a reference cement mortar, for the sake of comparison. The mortars are characterized for strength and stiffness, in curing conditions similar to that of the masonry specimens. Furthermore, the mechanical characterization of the chosen clay brick has also been presented. The bulk of the chapter focuses on tests performed on masonry specimens according to European standards. The tests include compression, E-modulus, flexural strength in parallel and perpendicular directions, and finally, the behavior of masonry wall panels when subjected to in-plane cyclic loads. The main goal of this chapter is to understand if lime in the mortar affects the different mechanical properties of masonry. Experimental results have been analyzed and discussed in the said context and have also been compared with the requirements of Eurocode 6. Chapter 6 – Conclusions and recommendations for future work This chapter highlights the main contributions of this research by summarizing the findings at the mortar level as well as masonry level. Based on the conclusions of the present work, recommendations have also been presented for further systematic research in this field of study.
A multi scale approach to the study of lime-cement mortars in masonry 8 2. Lime-cement mortars and their role in masonry 2.1 Introduction The chapter begins with an introduction to the two binders used – lime and cement, along with a summary of their hardening mechanisms, individual and combined, to provide context for the kinetics of the reactions, rate of development of mechanical properties, and optimum curing conditions. Subsequently, masonry mortars have been introduced and accompanied by an overview of the current state of knowledge on lime-cement blended mortars. Following this, a discussion has been provided on important properties at the mortar level; workability, compressive and flexural strength, ultrasound pulse velocity (UPV), E-modulus, Poisson’s ratio, fracture energy, drying shrinkage, and open porosity. The second aspect of this chapter addresses the role of mortars in masonry, with a special emphasis on the behavior of lime-cement mortars. Four important parameters, that are usually used to characterize masonry have been discussed; compressive strength, E-modulus, flexural strength, and shear bond strength. The final section discusses the response of masonry wall panels subject to combined vertical and in-plane cyclic shear loads. 2.2 Binders 2.2.1 Cement and hydration Portland cement is a hydraulic binding material formed by grinding clay and limestone together and calcining the mixture at a temperature of 1450˚C [19]. The mixture obtained after calcination is called clinker which often has a few lumps or nodules and in the final step is ground/crushed into a powder (fineness ≤ 75 µm) along with calcium sulfate to form cement (Figure 3) [5, 20]. The calcium sulfate, in the form of gypsum or otherwise, helps in controlling the hardening of cement after hydration and its rate of setting [2]. The European standard EN 197-1 [21] defines cement as a hydraulic binder that is finely ground, and inorganic, which when mixed with water hardens and gains strength through a series of hydraulic reactions. EN 197-1 [21] recognizes twenty-seven distinct types of cement and groups them into five main types based on their compositions, namely CEM I - Portland, CEM II – Portland composite, CEM III - Blastfurnace, CEM IV - Pozzolanic, and CEM V - Composite. Additionally, different types of cement may also fall into strength classes based on the compressive strength attained by mortars formed by them
A multi scale approach to the study of lime-cement mortars in masonry 9 (cement, water, and standard sand) at 28 days. The classes are 32.5, 42.5, and 52.5 MPa and each of them has a type N (ordinary) and R (high) early strength (at 2 and 7 days). Such categorization also accounts for performance requirements of initial setting time and soundness (expansion). Figure 3: Depiction of the process of cement production In terms of chemical composition, the clinker usually has the elements – silicon, aluminum, calcium, and iron in the form of their oxides (SiO2, Al2O3, CaO, Fe2O3), though the proportions may vary depending on the source of the raw materials employed and based on the properties desired in the end product [20, 22]. In order of importance, the main phases of cement clinker are C3S (Alite), C2S (Belite), C3A (Aluminate), and C4AF (Ferrite) [23]. The quantity of alite present in the clinker varies between 50% to 80% by weight and contributes to early strength gain by resulting in the formation of an almost amorphous phase of calcium silicate hydrate (C-S-H gel), which has a ratio of calcium to silica (Ca/Si) of approximately 1.45 to 1.75 and Portlandite or calcium hydroxide, also represented as CH [24]. Three common crystal structures of alite are known to be possible, depending on the temperature and impurities present; triclinic (with three polymorphs T1, T2, T3), monoclinic (with three polymorphs M1, M2, M3), and rhombohedral (R) (620-980˚C, 980-1070˚C and >1070˚C respectively), which accounts for a total of seven possibilities [25, 26]. As the clinker cools down, two of the monoclinic polymorphs have been observed to prevail, M1 and M3, depending on the presence of sulfate or magnesium impurities respectively [27-29]. C2S or Belite is formed in lower quantities than alite, constituting 15-30% by weight of clinker, and has a relatively more regular crystal structure than alite, with 5 known polymorphs [30]. It is less reactive than alite, and becomes a significant phase only approximately 10 days after hydration [31]. While cement clinkers may have more than one polymorph of belite, the most common type is -belite which is less hydraulic compared to the other belite polymorphs and is known to have monoclinic structure at room
A multi scale approach to the study of lime-cement mortars in masonry 10 temperature [30, 32]. Belite also contributes to the formation of C-S-H gel but at a slower rate and hence contributes to the gain of strength at a later stage [20]. The other two phases of cement clinker are tricalcium aluminate (C3A) and tetracalcium aminopherase (C4AF), present in quantities of 5-10% or 515% of the clinker, respectively (Figure 4). The former tends to react with water rapidly, leading to undesirable early setting of cement (flash set), which is often offset by adding gypsum to the clinker; the latter is not reported to cause such issues [20]. Figure 4: Typical composition of Portland cement (CEM I) [33] The hydration of cement, which results in its hardening, is a complex reaction, the kinetics of which depend on a multitude of factors such as its fineness and specific surface area, and its phase compositions [2]. Broadly speaking, the mechanism of cement hydration occurs through a process of dissolution and precipitation [34]. In the first step of dissolution, calcium sulfate, as well as phases of tricalcium silicates, break down into ions, forming an aqueous solution [2]. For the next step to occur in hydration, it is essential that the resulting products of hydration must be less soluble than the anhydrous phases. Since C-S-H gel is less soluble than alite regardless of the concentration of CH in the solution, alite always hydrates into C-S-H [34]. Typically belite does not hydrate at the same time as when alite is undergoing hydration, because the concentration of the solution tends to be higher than the solubility of belite [34]. In this second phase of precipitation, i.e., recombination of the ions into a solid phase that is energetically more favorable, alite dissolves to form C-S-H around cement particles, C3A dissolves to form an aluminate gel which reacts with sulfates to form ettringite and all the initial reaction collectively cause a large amount of heat release [2]. The first exothermic stage of the hydration reaction usually lasts only for a few minutes and is followed by the dormant period. The most commonly accepted theory for this induction period is that further reaction is inhibited because of the formation of metastable C-S-H on the surface of alite [2]. The next stage is the accelerated hydration of alite leading to the formation of C-S-H C3S 60% C2S 16% C3A 10% C4AF 8% Gypsum 6%
A multi scale approach to the study of lime-cement mortars in masonry 11 gel and Portlandite. This stage is almost exclusively guided by the nucleation and growth of C-S-H [35]. It has been proposed that small nuclei of C-S-H form on alite particle and thereafter, the growth of these nuclei control the hydration kinetics, at a rate proportional to the free area of the nuclei [36]. This period could last between 3 to 24 hours, with 30% of the cement typically expected to have reacted, and is subsequently followed by a stage called deceleration which corresponds to slowing down of relations, and slower gain of strength in cement, with time [37]. In this period, hydration continues but its process becomes diffusion controlled, and the hydration of belite becomes significant but the overall rate of reaction and evolution of heat are reported to reduce [2, 37]. Finally, the main products of Portland cement hydration rare C-S-H, reportedly occupying 50-60% of the volume of solid phase, and CH is reported to make up 20-25% of the volume [2]. Two other phases that occur during the hydration process are AFm and Aft, and are also known as monosulfoaluminate and ettringite [38]. Ettringite is formed from the reaction between gypsum and tricalcium aluminate or calcium ferrite phases and helps in avoiding the formation of hydrogarnet, which leads to flash setting of cement [39]. Ettringite is also attributed to expansion cracks in certain cases [39]. If all the gypsum added, reacts before aluminate, ettringite converts to a more stable phase of monosulfoaluminate with lesser sulfate. This instability occurs due to a sudden drop in the concentration of sulfate ions in the pores. Otherwise, ettringite reacts with aluminate to form monosulfoaluminate through a different chemical reaction [39, 40]. 2.2.2 Air lime and carbonation Calcium carbonate (CaCO3) commonly referred to as limestone, is quarried from the earth in its raw form (marble, chalk, and shell) and treated to produce Calcium Oxide (CaO) quicklime (Figure 5). This treatment, known as calcination takes place in a kiln at around 900˚C and releases carbon dioxide as a result of the endothermic reaction that occurs [41]. Since this form of lime is highly reactive and unstable to handle, typically water is added to it, and the resulting hydrated air lime is then used to prepare mortars. The technology used to produce air lime can influence the properties of the mortars it acts as a binder for, in both the fresh as well as hardened states [2]. If water is added to quicklime to hydrate it in a controlled manner and stoichiometric ratios it gets converted into a powder. This process of hydration (dry process), causing a rapid expansion in volume, results in the formation of calcium hydroxide or Ca(OH)2 called hydrated lime [42]. Depending on the grade of lime, the available lime content could vary from 70% to 90% [43]. Air lime could also be dolomitic, which involves the presence of mainly oxides and
A multi scale approach to the study of lime-cement mortars in masonry 12 hydroxides of calcium-magnesium, ranging from 5% to 30% by volume [43]. Usually, lime that is used in the industry is obtained through the dry process of hydration [10]. Figure 5: Illustration of the lime cycle If excess water is added to lime, the process is referred to as slaking (wet process). It results in the formation of lime putty and has been described as ‘suspension of calcium hydroxide crystals in water’ [2]. Traditionally, it was more common to use lime in the form of putty because it could be stored underwater for long periods (up to 3 years) and this would lead to an improvement of its properties [2]. It has been found that mortars with lime putty result in better workability due to higher water retention and viscosity of the same volume of material used compared to hydrated lime [2, 9]. Rodriguez-Navarro et al. [44], report that this happens because as lime putty ages, pre-existing micrometer-sized prismatic calcium hydroxide crystals get modified to sub micrometer-sized plate-like crystals, due to preferential dissolution of the prismatic faces and secondary crystallization. These new plate-like crystals which are generated on the original prismatic crystals offer a larger specific surface area. Both lime putty and hydrated lime are non-hydraulic binders, that harden by absorbing carbon dioxide (CO2) and result in the formation of calcium carbonate or CaCO3 once again [41]. This process, called carbonation of lime is a diffusion-dissolution mechanism, dependent on carbon dioxide, and to a great extent dependent on the pore structure of the resulting mortar and the amount of water in it [2]. The first step involves the dissolution of calcium hydroxide crystals in the capillary pore water, which also tends to raise the pH to around 12.8 [2]. The rate of this dissolution has been reported to depend on the specific surface area, solid-liquid interfacial area, and concentration of hydroxide ions, and the size of the crystals of calcium hydroxide (smaller crystals have higher solubility than larger crystals) [45, 46]. The process of carbonation though, is really slow because atmospheric carbon dioxide has to diffuse into the pores of
A multi scale approach to the study of lime-cement mortars in masonry 13 the material, after the removal of excess water from the mix [47]. Simultaneously, another process occurs which is the diffusion of carbon dioxide through the pores of the mortar, as pore water dries and evaporates [47]. Diffusion of carbon dioxide gets hindered in saturated conditions because diffusivity of carbon dioxide drops by almost 10,000 times in water compared to that in the air [48, 49]. However, since the controlling factor of carbonation reaction is the dissolution of calcium hydroxide crystals in the aqueous medium of the reaction, the presence of some minimum amount of water is necessary [47]. Based on experimental data, it was reported that the optimum range of relative humidity required for carbonation is between 40% to 80%, in which vapor and liquid form of water co-exist, facilitating the presence of continuous phase in the pores [2, 47]. The diffusion of carbon dioxide in water is reported to be the controlling step in carbonation [50]. Once carbon dioxide has dissolved in water, carbonate ions are formed, which react with calcium ions to form calcium carbonate [2]. The process is essentially that of nucleation and crystal growth, the kinetics of which is reported to be driven by the kinetics of transfer of ion mass [2]. It is widely accepted that carbonation reaction is affected by relative humidity and temperature, and is independent of the concentration of carbon dioxide [2, 47]. Calcium carbonate resulting from the carbonation of calcium hydroxide is known to crystallize in three different polymorphs, namely calcite, aragonite, and vaterite depending on the ambient temperature and humidity conditions, in which the reaction takes place [2]. Calcite has been reported to usually be the predominant polymorph precipitating in lime mortars, with the other two relatively unstable polymorphs eventually transforming into it [51]. Aragonite is found in different depths of mortars and is reported to precipitate at high pressures and temperatures, while vaterite, which is a metastable polymorph, is reported to precipitate in the beginning, along with amorphous calcium carbonate, in the size of nanometer crystals [52]. The formation of different polymorphs is also guided by the domination of kinetic or thermodynamic factors. If the former prevails, the polymorphs likely to be formed are aragonite or vaterite, eventually transforming into the more stable calcite form. However, in the case of thermodynamics being the prevailing factor in the carbonation reaction, it is expected that calcium carbonate will precipitate as calcite [45]. 2.3 Lime-cement blended (masonry) mortars The technical role of bedding mortar in masonry construction is often discussed with regard to properties such as deformability, workability, strength, stiffness, shrinkage, and vapor transmission [13, 53-55]. However, it is difficult for any individual binder to fulfill all requisite criteria of a suitable masonry mortar and so it is common to use more than one type of binder [8, 56-58]. Masonry mortars generally combine
A multi scale approach to the study of lime-cement mortars in masonry 20 appears that there is a general agreement on E-modulus decreasing with an increasing quantity of lime in binder, however similar to the tendencies observed for compressive strength, there is a lack of quantitative correlations in the literature. Furthermore, studies have been almost universally focused on the behavior of mortars, which have gained adequate maturity, generally accepted as 28 days for cement-based materials and at least 90-180 days for lime-based materials [119-121]. Based on the literature review conducted, no research was found to focus on the behavior of lime–cement mortars specifically between 0 and 7 days of curing age. This knowledge is important to bridge the research gap concerning gain of mechanical strength and stiffness in masonry and consequently stresses developed, in early ages. However, E-modulus in early ages is not easy to measure using the cyclic compression test, even for cement-based mortars, mostly for practical reasons such as the material not possessing adequate strength to be tested [122, 123]. For early age testing, other approaches, based on ultrasound wave transmission, bender-extender elements, or resonant frequency identification have been adopted by different researchers, mostly for cement-based materials [124]. Some of the alternatives follow the same principle of cyclic loading but allow tests without demolding the specimens, such as BTJASPE (BéTon au Jeune Age, Suivi de la Prise et du module d'Elasticité) [125] and TSTM (Temperature Stress Testing Machine) [126]. Amongst these, a method that may relevant for monitoring E-modulus in early ages is EMM-ARM (Elasticity Modulus Measurement through Ambient Response Method)) [127]. This technique has been successfully employed in cement and lime-based stabilized soils since it is capable of measuring the changes in stiffness of any material that can be cast into a mold and undergoes significant changes in stiffness with time [128]. The method is based on continuous modal identification of the first flexural resonant frequency of composite beams (specimens of materials to be tested, cast in cylindrical molds), based on vibrations naturally occurring in the environment [124]. The evolution of resonant frequency identified during the experiment can be directly correlated with the E-modulus by using the dynamic equation of motion, and appears to provide values of E-modulus similar to those obtained from the classical cyclic compression tests [115, 129]. The potential of this method has been explored in this research to assess its suitability for lime-cement mortars. 2.3.5 Deformability and Poisson’s ratio Static Poisson’s ratio, defined as the ratio of lateral to longitudinal strain, is a parameter that is widely recognized as needed in assessing the deformability and deflection properties of different materials [130, 131]. While it is not a primary factor that influences design decisions, it is important to investigate
A multi scale approach to the study of lime-cement mortars in masonry 21 Poisson’s ratio of mortars because it is also known to influence the failure mechanisms of masonry [132, 133]. It is common to use values of E-modulus and Poisson’s ratio of both unit and mortar to determine stresses in the different components of masonry [134]. Usually, units are stiffer than the mortar used in masonry, meaning that the unit restrains the mortar when subjected to compressive loads, causing triaxial compression in the mortar and a state of biaxial tension-compression in the brick [135, 136]. Research conducted on the behavior of mortars subjected to triaxial stresses suggests that the failure mechanism of mortars varies based on the type or quantity of binder used, as well as the lateral stresses applied on the mortars [133, 137]. Hayen et al. [138] tested mortars with different compositions (lime putty, hydraulic lime, and lime-cement binders) in triaxial compression and found that the response of mortars is independent of their respective composition, and is influenced primarily by the ratio of lateral to vertical stress (k). Contrastingly, the experimental work of Mohamad et al. [118, 133] indicates that the failure mechanism of mortars under triaxial compression may be dependent on the type and quantity of binder used. From their work on lime-cement mortars, it is possible to observe that as lateral stresses increase, weak mortars (1:1:6, 1:2:9) exhibit an exponential decrease in Poisson’s ratio, whereas strong mortars (4:1:12, 2:1:9), exhibit a linear decrease. And the reason for this difference has been attributed to higher porosity and differing void sizes in the ‘weaker’ i.e., lime based mortars. A linear decrease in values of Poisson’s ratio with increasing lateral stresses, has been confirmed for cement mortars by other authors [139]. In the context of concrete, Ottosen [140] had proposed that Poisson’s ratio remains constant until a stress/strength ratio (β) of 0.8, after which, it increases up to failure. Based on the experimental results obtained by Mohammad et al . [118] a modification to Ottosen’s model was proposed (Figure 8), for two different cases of failure types in mortars. Here, the symbol (β) represents the lateral stress to strength ratio. In both cases, the value of Poisson’s ratio of mortar decreases with increasing stress/strength ratio, up to a threshold value. The behavior thereafter, depends on the strength of the mortar. If the mortar is strong, the Poisson’s ratio increases gradually corresponding to shear failure, marked as (a) in the figure. While in the case of the weak mortar, the Poisson’s ratio increases suddenly, corresponding to pore collapse, marked as (b) in the figure. In general, if the Poisson’s ratio of mortar is less than that of the unit, variations in it are not expected to affect the strength of masonry [141].
A multi scale approach to the study of lime-cement mortars in masonry 22 Figure 8: Ottosen model and modification [118] Considering the contradictory observations in the little data available on this subject, it is important to systematically assess the influence of different factors on Poisson’s ratio of mortars. Experimental values of static Poisson’s ratio of mortars have not been widely researched, authors usually employ average or representative values for analytic expressions [141-143]. It is also possible to find studies focused on measuring dynamic Poisson’s ratio of mortars [144, 145]. However, in the absence of a standard or widely accepted method of correlation between dynamic and static values of Poisson’s ratios, these studies remain at best, only an indication of static values of Poisson’s ratios. The value of Poisson’s ratio usually used for concrete is around 0.2 [146, 147]. And it is possible to observe from literature, that the general range of Poisson’s ratio obtained/used for mortars (lime, cement, or blended) is around 0.2 as well, often ranging between 0.15 to 0.25 [134, 148-150]. 2.3.6 Fracture energy and crack propagation Fracture energy may be defined as the energy required for crack propagation, or to create one unit area of a crack [151, 152]. One of the most famous mathematical descriptions has been provided by Griffith applicable for homogenous, brittle materials under uniaxial tensile stresses, which relies on an energy balance approach and links the extent of plastic deformation associated with the crack extension to a quantitative figure, shown in (Equation 1) [153-155]. Here, ‘a’ indicates the size of the smallest crack that can be detected where ‘𝛾’ is the surface energy, 𝜎 is the associated stress level and E is the Emodulus. σ=√2E a 1
A multi scale approach to the study of lime-cement mortars in masonry 23 While in brittle materials, the energy is released from rupturing of chemical bonds along the plane of cracking, in more ductile materials this energy is associated with plastic flow near the crack tip [152]. Since Griffith’s model was based on balancing surface energy and total strain energy released per unit volume upon the development of the crack; it was found unsuitable in the case of ductile materials [155]. Irwin et al . [156] modified the equation to account for the energy dissipated due to plastic flow in the proximity of the crack tip introducing the parameter critical strain energy release rate Gf, shown in (Equation 2). σ=√2EGf a 2 In this approach, the assumption is that a unit area of crack is formed when the energy released Gf is greater than or equal to the energy absorbed [157]. This critical energy that causes crack propagation can be determined by experimentally obtaining the corresponding critical load that is required to fracture a specimen with a pre-defined crack length (a), typically in a three-point bending test [155, 158]. While several advanced analytical concepts have been proposed over time by different authors for determining fracture energy, the most commonly used method to determine fracture energy in mortar and concrete is the three-point bending tests of samples with a pre-defined notch in them, based on the recommendations of RILEM, which relies on calculating fracture energy as a ratio of work done during the test to the area of the ligament of the specimen being tested as shown in (Equation 3) [158, 159]. W is the work done during the test, b is the width of the specimen, d is the depth and a is the size or length of the notch. In the RILEM recommendation itself, the work done is calculated as a sum of the area under the force-displacement curve and a factor associated with the mass of the setup. Gf=W b(d−a) 3 Elices, Planas, and Guinea [160-162] put forward a series of three articles discussing the influence of different factors such as the experimental procedures being used, the tail of the force-displacement curve, and energy dissipation during the test, and finally put across a more refined expression built on what was proposed by RILEM [158]. The modified expression took into account measured and unmeasured work that was done due to the force-displacement tail, in the numerator (Wm+Wum) and has been used by different authors since to evaluate fracture energy of specimens. Garijo et al. [94] used these modified expressions to evaluate the fracture energy of hydraulic lime mortars with different water-binder ratios and found values to range between 4 to 13 N/m. The fracture energy of air lime mortars was found to
A multi scale approach to the study of lime-cement mortars in masonry 24 range around 5 N/m in the literature [163]. In the case of lime-cement mortars, fracture energy is reported to decrease as the quantity of lime in the binder increase, with values ranging between 7 and 44 N/m for lime-cement mortars with B/Ag ratio 1:6 and 75%, 50% and 25% lime in the binder [164]. Not surprisingly, fracture energy is yet another parameter that has been much more extensively studied in the case of concrete or cement-based mortars, over lime-based materials [165, 166]. From reviewing the tendencies found in the fracture energy of concrete it is known that increasing the size of the aggregates or the specimen, decreasing the water/binder ratio, or decreasing the ratio of the predefined notch to the depth of the beam specimen, lead to an increase in fracture energy [159, 166]. The presence of fiber reinforcements leads to an increase in the fracture energy of the concrete, though this depends on the type of fiber used [166, 167]. In this regard, the presence of fiber reinforcements has also been found to increase the fracture energy of lime-based materials [168]. Broadly though, most studies on fracture energy were found to focus on concrete [159, 169, 170]. Adequate data could not be found to spot contradictions or specific trends on fracture energy as a function of lime content in the binder. 2.3.7 Open porosity Porosity is commonly used to discuss the pore structure of a given material and refers to the volume of void space of a material, as a function of its total volume [171]. It is usually linked with the strength and durability of mortars and the amount of water in the pores not only influences both carbonation and hydration but also affects mechanical properties such as stiffness and shrinkage [172-175]. Furthermore, porosity has been reported to be affected by the type of binder used, the water binder ratio in the mortar, the B/Ag ratio, the mineralogical nature of the aggregate used, its particle size distribution as well as the ambient environment of the mortar [172]. Unless otherwise specified, porosity that is usually referred to, is ‘open porosity’, versus total porosity which is a combination of open and closed porosity. These two types refer to the interconnectivity of pores, with the word ‘open’ implying permeability to gases and liquids, while ‘closed’ refers to isolated pores that do not connect to the main pore structure of the material [172]. The methods that are widely used to measure open porosity and pore size distribution are the water immersion method and mercury intrusion porosimetry (MIP) respectively. The former involves, placing specimens in a vacuum for a predefined period, followed by complete immersion in water, under vacuum [176]. Porosity is then calculated as a function of the saturated weight of the specimen, measured in water and air, and the dried weight of the specimen expressed in percentage. MIP is a technique [176], that involves pressurizing mercury into the pores of the material being tested. Larger pores are expected to fill initially, followed by
A multi scale approach to the study of lime-cement mortars in masonry 25 smaller pores getting filled as the applied pressure increases. This method is capable of providing information on the porosity, pore size distribution, and pore volume of a given material [56, 66]. Porosity in this method is calculated as a function of the volume of mercury that penetrates the specimen, and the weight and bulk density of the specimen. Silva et al. [7], reported that for the same binder-aggregate ratio (1:3), the pore size distribution of lime mortars is bimodal (with peaks at 0.5 and 30 mm) and that of cement mortar is unimodal (with peaks at 0.3 mm). On mixing the two binders, this clear demarcation begins to blur and there is a shift observed in the size of the pores as well as their distribution curves. With increasing hydraulic content in the binder, there is an increase in pores with less than 0.01 mm size and a decrease of pores with 10 mm size. The pore sizes mentioned will vary based on the quantity and type of aggregates used, even if the binder remains unchanged. While increasing the quantity of aggregates leads to an increase in the porosity of cement mortars, in the case of lime mortars, the highest quantity of binder exhibits the highest porosity as well [58, 177]. In lime-cement mortars an increase in open porosity of blended pastes and mortars corresponded with increasing quantities of lime in the binder [2, 45]. However, the literature on the effect of lime or cement in blended mortars is not unanimous. Cizer et al. [48] report that the porosity of mortars increases with increasing quantities of lime in the binder and ranges between 18% and 28%. Contrastingly, Arandigoyen et al. [59] found the open porosity of mortars (around 20%) to be independent of the lime-cement proportion in the binder. This is unexpected since the same research group found the porosity of limecement pastes to increase with increasing quantities of lime [45, 66, 67]. Yet another trend, was reported by Macharia [95], who found the values of porosity of lime-cement mortars to range between 20% and 30%, with an increase in porosity only up to 45% lime in the binder by mass, followed by a subsequent decrease in values of porosity. This suggests that in lime-cement mortars, there might be a desirable or optimum ratio with regard to porosity and pore structure. However, there is a lack of unanimity in trends and quantitative values of porosity of lime-cement mortars. 2.3.8 Drying shrinkage ASTM C596-01 [178] defines drying shrinkage as the change in length of a specimen due to a sum of factors excepting any external applications of force, instated conditions of temperature, relative humidity, and evaporation rates. Drying shrinkage is also associated with volume changes in materials, typically reduction when it is exposed to drying [179]. This phenomenon is of relevance because when it occurs
A multi scale approach to the study of lime-cement mortars in masonry 26 in the presence of external restraints imposed on the material, tensile stresses may develop and lead to cracking [179]. In the case of concrete, where members might be several centimeters thick, moisture gradience between the surface and the bulk of the specimen also plays a role in the development of stresses, however, in the case of mortars in masonry, this factor usually is not of great concern [62, 179]. In cement-based materials, drying shrinkage along with autogenous shrinkage has been extensively researched in the last few years, since it is linked with studies of moisture transfer, porosity, and ultimately durability of the material [180-183]. However, the same attention has not been provided to the drying shrinkage of lime-based mortars [184, 185]. This is surprising since, drying shrinkage can happen in any mortar and the corresponding cracks are bad for the durability of the materials because they allow the ingress of moisture and possibly unwanted harmful chemical salts [186, 187]. For cement mortars, the values of drying shrinkage found in the literature were found to stabilize by 30 days and were in the range of 700-1200 µm/m, depending on their compositions and the incorporations of admixtures [188-190]. One of the articles that were found to discuss lime-cement mortars reported drying shrinkage to reduce in the order of air lime mortar, lime-cement blended mortar, and hydraulic lime mortar, with values in the range of 6000, 2000, and 1500 µm/m respectively [62]. The general range of drying shrinkage values of lime-cement mortars (800-1200 µm/m) was found to be higher than those of cement mortars (600900 µm/m) [185, 191]. Furthermore, within lime mortars, lime putty is reported to exhibit more shrinkage than hydrated lime powder [192]. To the best of the author’s knowledge, no work was found discussing the impact of the quantity of lime in the binder on the drying shrinkage of lime-cement mortars. 2.3.9 Summary The literature on the impact of lime in mortars was found to be primarily focused on strength, stiffness, and porosity. In the case of the former, despite established trends, there is a lack of quantitative assessment [48, 59] and in the case of the latter, there is contradictory information [48, 59, 95]. Based on the review, one may however expect that an increase in lime in a lime-cement mortar would lead to lower strength and stiffness [48, 118][48, 118][48, 118][34, 110] and higher porosity [48, 59]. Regarding stiffness, no work was found discussing the E-modulus of lime-cement mortars in early ages (< 7 days). Regarding properties such as Poisson’s ratio, fracture energy, and drying shrinkage, adequate literature on lime-based mortars was not found to permit discussion on trends or contradictions in them. However, a general range of expectable values was found for each of these properties, either from tests on predominantly cement-based mortars or scattered works on lime-based mortars. For Poisson’s ratio of
A multi scale approach to the study of lime-cement mortars in masonry 27 mortars, values were found to lie between 0.15 and 0.25 [134, 148-150]. In the case of fracture energy, values found varied in the range of 7 – 44 N/m, with values decreasing as the quantity of lime in the mortar increased [164]. And finally drying shrinkage for lime-cement mortars was reported to be around 600 – 900 µm/m [185, 191]. Furthermore, since this collective information comes from multiple sources, with different materials and operators involved in each case, direct comparison becomes difficult. There is a clear lack of a comprehensive study that uses the same materials and protocols to test several mechanical properties focusing on the lime-cement ratio in the binder. Results obtained from such experiments will also serve as useful inputs for numerical modeling. There is also scope for exploring the correlation between different parameters, which could optimize the number of tests required for a given set of materials or serve as a method to cross-check results obtained. 2.4 Influence of lime-cement mortars on the mechanical behavior of unreinforced masonry Masonry elements primarily resist loads in the vertical direction but are also often subject to lateral loads such as wind pressure and earthquakes (Figure 9). The weakness in such cases usually lies at the interface of unit and mortar, and therefore, good bond strength is crucial to ensure adequate resistance of masonry to shear and tension [193]. The influence of mortar on the mechanical behavior of masonry has been addressed in the literature, but most studies discuss the behavior of masonry in compression [13], with a few of them focusing on bond strength [194-196]. However, there is a lack of knowledge on the influence of lime in the mortar on these properties of masonry. Because of this, regarding different lime-cement ratios, there is also a gap in the correlation of different scales of study, from the mortar level to the bond between brick and mortar to wall panels to full scale structures. And this information is crucial to optimize the choice of mortar for a given unit, not just to satisfy design requirements for masonry but also to obtain the best combination possible of strength, porosity, workability and shrinkage characteristics of the mortar [8, 197]. This segment, therefore, discusses the contribution of mortar to the compressive strength, stiffness, flexural resistance, and bond strength of masonry, focusing on the presence of lime in the binder. It also addresses the in-plane shear resistance of masonry wall panels.
A multi scale approach to the study of lime-cement mortars in masonry 28 Figure 9: Vertical and lateral loads on masonry 2.4.1 Masonry in vertical compression 2.4.1.1 Strength Compressive strength of masonry is a parameter that has been investigated widely by researchers and is measured in Europe according to EN 1052-1 [17] which recommends dividing the maximum vertical load applied on the specimen without restraint or eccentricity, by the loaded cross-section of the specimen. Compressive strength is known to be affected by a variety of factors such as the type of unit and mortar used, the relative strength and stiffness of the components, and it naturally also depends on the bond between the unit and mortar [135, 195, 198-207]. The type of masonry bond, i.e. the arrangements of the units and the texture, greatly influences the response of masonry to compressive loads [148, 201, 208]. It has also been reported that the compressive strength of masonry decreases as its height to thickness ratio increases [209-211]. Results on the effect of joint thickness on strength of masonry are conflicting, with some researchers suggesting an optimal thickness of 2 cm [212], while others suggest that strength of masonry consistently decreases as the thickness of joint increases [150, 213, 214]. The use of thinner mortar joints is recommended based on mechanics, since they are expected to reduce lateral tensile stresses in the units and increase the stress confinement in the mortar, consequently increasing its strength [215]. Indeed, it is common to find masonry with joint thickness around 10-12 mm [12, 13, 216], providing this is sufficient to accommodate the geometric tolerance of the masonry units. In addition to these factors, the anisotropic and inhomogeneous nature of masonry provides additional complexity [217]. Different researchers have tried to estimate the compressive strength of masonry based on the properties of its components [136, 193, 218, 219]. One of the earliest attempts at this
A multi scale approach to the study of lime-cement mortars in masonry 29 quantification was made in 1971 by assuming linear elastic behavior of masonry [150]. Since then, considerable development has taken place in this field with the use of non-linear micro-mechanical models, artificial neural networks, and fuzzy logic to predict the behavior of masonry in compression [148, 220-224]. The bottleneck in this development is the need for a wide range of experimental data to calibrate the models [220, 223]. So until these models start producing reliable results for different cases, simple analytic expressions, estimations, and trends are essential to understand the behavior of masonry. From the experiments in the literature, on various units and mortars, it may be also be concluded that the strength of the masonry is dependent on the strength of the unit to a much larger extent than that of the mortar. For joint thicknesses of 10 and 15 mm, an increase in the strength of mortar by 150% leads to an increase in the strength of masonry of only 16% and 36% respectively [225]. In tests on filled and unfilled concrete masonry prisms, an increase in mortar strength of 250% led to an increase in strength of only 35% [11]. It has been shown that in the case of ungrouted prisms, increasing the compressive strength of mortar by almost 72% led to an increase in the strength of masonry by only 20% [226]. One of the most commonly used expressions to estimate the strength of masonry involves the strength of the unit and the strength of mortar and is of the nature as shown in (Equation 4). fk=Kfbαfmβ 4 Here, fb is the compressive strength of bricks and fm that of the mortar, while the values of K, α, and β vary based on the experimental data being used. A summary of these values, from some of the most frequently appearing works in the literature, has been presented in Table 1. From this data, it is possible to observe that almost all the expressions have a higher value of exponent for the strength of unit than the strength of mortar. Only one work uses the same value of exponent 0.5 for unit and mortar [227]. Table 1: A summary of values of K, α, and β (Equation 4) as presented by different authors K α β Source Variable (~0.55) 0.7 0.3 Eurocode 6 [18] 0.83 0.67 0.33 Mann [228] 0.25 1.03 0.28 Sajanthan et al. [229] .275 0.5 0.5 Dayaratnam [227] 0.63 0.49 0.32 Kaushik et al. [13] 0.3 1 - Bennet et al. [230]
A multi scale approach to the study of lime-cement mortars in masonry 36 some researchers [283]. Experimental values of flexural bond strength are reported to be higher than flexural strength (parallel to bed joints) obtained from prisms, which in turn are higher than those obtained from wallets [275]. Furthermore, when tested in bending, the strength of masonry also depends on the number of joints in pure flexure. Based on the weakest link theory, masonry specimens are often treated as beams in bending with joints of different strengths, and since one of them is bound to fail first, the strength of the beam is determined by the strength of its weakest joint [284, 285]. Van der Pluijm [275] showed using order statistics that for a coefficient of variation of 25%, the flexural strength obtained from a masonry specimen with 4 joints in pure flexure, would be 0.7 times the value of flexural bond strength in the joints. 2.4.3 Shear bond strength Shear bond strength is an extremely important parameter of masonry and has been studied by different researchers [195, 286]. It is typically investigated through experiments using couplet or triplet specimens, with the latter being more common [287-289]. In Europe, shear bond strength is tested through triplet specimens and is done according to the recommendations of EN 1052-3 [290]. It involves the application of three distinct levels of perpendicular pre-compression, while a lateral load is applied to shear the unitmortar joint (Figure 13). The maximum shear strength obtained is recorded for each level of perpendicular pre-compression, and experimental data are analyzed to obtain values of initial shear strength or characteristic shear strength of the materials being tested. This method involves the use of Mohr-Coulomb law to describe failure, which establishes a linear relationship between normal stress (σ) and shear stress (τ) (Equation 7) [291]. Here, c is cohesion, also known as initial shear strength (fvo), and tan ɸ is the coefficient of friction. This relationship is valid for only low and moderate normal stresses since at higher values, crushing and cracking of the unit is possible [291]. τ=c+tanɸ.𝜎 7
A multi scale approach to the study of lime-cement mortars in masonry 37 Figure 13: Triplet masonry specimens used to determine the initial shear strength of masonry EN 1052-3 [290] allows the determination of initial shear strength by testing specimens in the absence of normal stresses, this however implies that information about friction between unit and mortar is lost. Another parameter associated which is often studied, in association with shear bond strength is the angle of dilatancy (ψ), the tangent of which (tanψ) is defined as the ratio of normal to shear displacement [286, 292]. It is an indication of the volume change associated with inelastic shearing deformation and is known to linearly decrease with increasing pre-compression [275, 286]. It is possible to find research [196] focusing on the effects of pre-compression on the peak shear stress on (wire-cut, clay) brick masonry and cement mortars of different strengths (10-30 MPa). Shear stress increases with an increase in normal stress, and the relationship tends to become non-linear above 1 MPa of vertical pre-compression. The values of cohesion increase with increasing strength of mortar, and also depended on other factors such as surface roughness and water absorption properties of the unit [196]. On the other hand, a relationship between internal friction and strength of mortar was not established, since the former was found to be independent of the latter to a large extent. The angle of friction was found to be around 45˚ by some researchers [196] but accompanied by significant variation. Initial shear strength has been tested for different lime-cement mortars and concrete blocks [293], with the primary focus being on the relative strength between mortar and units. It was found that both, the strength of mortar and unit contributed to an increase in cohesion, with higher values obtained the quantity of cement in the binder was increased. Another study reached similar conclusions by testing lime-cement mortars (1:2:9, 1:1:6 and 4:1:12) with different types of units, molded clay brick, extruded clay brick and concrete blocks, in triplet specimens. A higher quantity of cement in the binder led to higher shear bond strength for all types of units [282]. However, in both these studies, no vertical precompression was applied and therefore, there is a lack of information on the coefficient of friction. Values
A multi scale approach to the study of lime-cement mortars in masonry 38 of initial shear strength for lime-cement mortars were found to vary widely, ranging between 0.07 to 1 MPa [282, 293]. For lime-based mortars, the values of initial shear strength and coefficient of friction were found to lie between 0.15 to 0.43 MPa and 0.8 to 0.92 respectively [294]. Compared to flexural strength, information on the initial shear strength of masonry is more widely available in the literature, though not many of them focus on the influence of lime-cement ratios in the binder on the bond strength of masonry [282, 293, 294]. Even in the cases where different lime-cement ratios are compared, it is unfortunate that the tests were performed in the absence of vertical precompression and so information on friction, and the effect of lime on it is absent [200, 282]. Research has also highlighted a good correlation between initial shear strength and compressive strength of masonry, emphasizing that good bond strength would lead to an improvement in the compressive strength of masonry [200]. However, there is contradictory information on this, since some researchers found a poor correlation between the two properties through their experimental data [294]. Initial shear strength is also often correlated with flexural tensile bond strength of masonry across different units, with the former being approximately 1.2 times the latter [282]. The Australian standard AS 3700 [295] also recommends a correlation of shear bond strength being equal to 1.25 times the flexural bond strength. Other researchers have also found this to be true [288]. 2.4.4 In-plane shear strength (combined vertical and horizontal loading) Masonry typically fails in shear, when it is subjected to a combination of vertical and horizontal loads, as often happens in earthquakes [296]. And its response may be studied through quasi-static cyclic tests or dynamic shaking table tests [104]. Compared to dynamic tests, quasi-static cyclic tests may be a more suitable option for testing unreinforced masonry, because they facilitate accurate damage measurements, even if dynamic tests are capable of simulating seismic forces more accurately [297]. Quasi-static cyclic tests are also more conservative compared to dynamic tests because they lead to lower lateral capacities and greater damage in the specimens being tested [104]. The lateral strength capacity of unreinforced masonry is an important parameter, however, to evaluate the response of a structure to seismic loads, factors such as stiffness and strength degradation, energy dissipation, and overall ductility are also crucial. A high ductility factor in masonry indicates better non-linear deformation capacity and thereby better performance of masonry under seismic loading [104]. Usually, deformations are compared by accounting for a parameter called lateral drift which is expressed in percentage and is the ratio of lateral displacement and the height of the wall [104]. Energy dissipation is also an important indicator for evaluating the
A multi scale approach to the study of lime-cement mortars in masonry 39 seismic performance of a structure because high energy dissipation implies a potential reduction in demand for ductility [298]. In earthquakes since the direction of lateral loads reverses constantly, typically cyclic tests are designed to simulate the alternating direction of loads [206, 299-301]. The in-plane response of unreinforced masonry walls is known to depend on its geometry, vertical loads, boundary conditions and its mechanical characteristics as well as those of its constituents [296]. Furthermore, it is known that different failure modes are possible, based on the vertical load, quality and bond strength of unit and mortar (Figure 14) [302]. Figure 14: In-plane failure mechanisms of masonry subject to combined vertical and horizontal loading The first type is known as sliding failure, which occurs as a result of low vertical stresses and typically poor mortar quality, causing the wall to shear into two and the sliding occurs between the two parts [303]. The characteristics of this type of failure mode are that it is stable, results in large displacements and energy dissipation, and is usually observed in the upper parts of buildings since they are subjected to lower vertical stresses compared to the bottom parts [296, 303]. The second failure mode is called diagonal shear failure and occurs when the principal tensile strength of masonry is surpassed by the principal stresses (in-plane) generated in it [303]. This failure mode is characterized by the development of diagonal cracks, low displacement capacity and relatively fast dissipation of strength and stiffness, and average energy dissipation, and is typically observed in the bottom part of masonry buildings. The formation of cracks through mortar joints, units, or both depends on the quality of the individual components [296]. The third mode of failure is called rocking-flexural failure and is usually associated with slender walls where the moment to shear ratio is high, low vertical stresses, high displacement capacity, and very little strength degradation and average energy dissipation in hysteresis. The final failure takes place with masonry being crushed at the corners [303].
A multi scale approach to the study of lime-cement mortars in masonry 40 Usually, in cyclic tests performed in the laboratory, flexural mechanisms develop in the beginning due to the low axial tensile strength of masonry and must not be mistaken as the final failure mode [296]. Development of horizontal tensile cracks in the bottom part of the specimens takes place near the supports, potentially accompanied by crushing of corners, but the resistance of the wall usually increases until it fails in shear. It is also important to remark that the first failure mode discussed above, that of sliding failure, is linked with shear bond strength, i.e., a function of cohesion and friction between the unit and mortar joints. The second mode of shear failure mechanism is guided by principal tensile strength [296, 304], and may be expressed analytically (Equation 8) by assuming that masonry behaves in an isotropic, elastic manner until failure [304, 305]. Here, σ is the vertical compressive stress, τmax is the shear stress in the wall corresponding to maximum lateral load capacity and r is a factor that considers the ratio of height to length of the wall. ft=√(σ2)2+(r τmax)2−σ2 8 This failure could occur through the formation of a single diagonal crack, or two of them, and could pass through only the mortar joints, only the units (less common), or involve both mortar joints and units [306308]. Even after the formation of diagonal cracks, the wall is usually expected to have some remaining bearing capacity, especially if it is composed of units and mortar of good quality [309]. Balasubramanian et al. [304] have presented a review of analytic formulae presented by different authors to estimate the shear capacity of masonry subjected to cyclic loads, categorized according to failure modes, which is useful for comparison of experimental results with analytical data. Eurocode 6 [18] recommends the determination of design value of shear resistance based on the Mohr-Coulomb law, and global strength parameters of masonry, with the initial shear strength of masonry depending on the strength class of the mortar and coefficient of friction for the wall equal to 0.4 (Equation 9.1), where lc is the length of the wall under compression. However, it may be argued that this expression (Equation 9.1) is specific to shear failure due to sliding and has no relation with the tensile strength of masonry [206]. Another expression, that is frequently used to express the shear resistance of masonry has been presented by Magenes et al. [310] and is valid for failure through mortar joints (equation 9.2). It employs local (i.e. joint) material properties of cohesion and coefficient of friction, also accounting for the shear ratio (effective height/length of the wall) and influence of head joints (Equation 9.2). To account for the failure which may be initiated by shear-tensile cracking of bricks, equation 9.3 may be considered based on a proposal of Mann et al. [311]. For the rocking-flexural failure mode, one of the most commonly used analytic
A multi scale approach to the study of lime-cement mortars in masonry 41 expressions is based on equilibrium (equation 9.4) and has been proposed by Calvi and Magenes [310]. Here, j accounts for the vertical stress distribution at the compressed toe and usually has a value of 0.85, based on the assumption of an equivalent rectangular stress block. In all the analytic expressions presented, some accuracy has been sacrificed to permit speed and simplicity, however, this simplification allows the evaluation of relative importance of different parameters that contribute to the shear response of unreinforced masonry [297]. Vd=(fvko+ 0.4 σ) t lc γm 9.1 9 Vd=l t τmax where τmax=min (τc, τw) τc=1.5c+μσ 1+3cαv/σ where αv=ho l τw=c+μσ 1+αv 9.2 Vd=lt ftb 2.3(1+αv)√1+σ ftb 9.3 Vd=l t σ 2αv(1−σ jfu) 9.4 Many researchers have investigated the response of masonry to cyclic loads with different types of reinforcements such as glass fiber reinforced polymers (GFRP) and fiber-reinforced cementitious matrices (FRCM) to study how different retrofitting techniques improve the lateral strength and energy dissipation capacity of masonry [306, 308, 312-314]. The response of masonry (stone) to in-plane cyclic loads has also been studied as a function of distinct textural typologies. Ductility was found to decrease as the irregularity of bonds in the masonry walls increased [315]. In lime-based unreinforced masonry, the performance of three lime mortars with B/Ag ratios 1:1, 1:2, and 1:3 was compared in quasi-static cyclic loading and it was found that parameters such as energy dissipation and stiffness increased, and ductility reduced with an increase in the compressive strength of the lime mortar used [307]. The general range of maximum drift capacity (%) for clay brick unreinforced masonry walls in the literature is reported to be between 0.43% and 1.06%, with an average value of 0.6% [316]. Furthermore, maximum drift capacity is known to decrease with increasing values of vertical loads applied [104, 316]. Regarding the maximum lateral capacity of clay brick unreinforced masonry walls, values were found to range between 0.3 and 0.5 MPa [296]. However, to the best of the author’s knowledge research on this topic, focusing specifically on the influence of lime-cement mortars, could not be found.
A multi scale approach to the study of lime-cement mortars in masonry 42 2.4.8 Summary Regarding masonry in vertical compression, it is known so far that including lime in the mortar leads to greater deformability, lower strength, and stiffness of masonry [12, 13]. However, it is also reported that, in general, while mortar does not contribute majorly to altering the strength of masonry, it does significantly influence its stiffness and deformation capacity [11]. In the case of flexural strength, not much research was found to directly focus on the influence of lime on the resistance of masonry parallel and perpendicular to the bed joints, however, there is consensus on flexural strength increasing with an increase in the strength of mortar [275]. Furthermore, it is also widely acknowledged that flexural strength in the perpendicular direction is higher than in the parallel direction to the bed joints [285]. Concerning flexural tensile bond strength, values are reported to increase with increasing strength of the mortar [14, 195]. However, some research has shown that this may not necessarily be true. A lime-cement mortar with lower compressive strength than a cement mortar was found to result in improved flexural bond strength [282]. It is evident, that there is need for more experimental research on this topic. Initial shear strength of unreinforced masonry is a very important property that relies on the bond of materials of the mortar and unit and is often correlated with the flexural bond strength and compressive strength of the material [200, 282, 288]. Values of initial shear strength are reported to increase with the strength of the mortar, and range between 0.07 to 1 MPa in the case of lime-cement mortars [282, 293, 294]. However, the focus of these works is not on the influence of the lime-cement ratios, and even in the cases where different lime-cement ratios are compared the tests were performed in the absence of perpendicular pre-compression and so information on friction, and the effect of lime on it is absent [200, 282]. Finally, the failure of unreinforced masonry walls, subjected to combined in-plane vertical and horizontal loads is important to understand their mechanical behavior when subject to seismic action [104]. While studies specifically focusing on the influence of using different lime-cement ratios in the mortar on this aspect of masonry were not found, the general range of maximum drift capacity (%) for clay brick URM was found to be between 0.43% and 1.06% [316]. This is a wide range, and there is a need for not only deformation capacity, but also for factors, such as lateral strength capacity, stiffness and strength degradation and energy dissipation, to be assessed as a function of different materials used in the mortar. From the discussion above, it is possible to conclude that the influence of lime in mortar on the behavior of masonry in compression has been relatively well explored compared to the behavior of masonry in flexure or shear, and especially under quasi-static cyclic loads [13, 240]. This is not ideal, since masonry
A multi scale approach to the study of lime-cement mortars in masonry 43 is designed to resist loads not just in vertical compression, but also in tension and shear. In the case of seismic activity, it is also important to understand its behavior when lateral loads are reversed. Since lime and cement are commonly used for the construction of masonry around the world, it is crucial to study their impact on the behavior of masonry, to optimize design and facilitate the choice of a compatible mortar for any given unit.
A multi scale approach to the study of lime-cement mortars in masonry 44 3. Materials, mortar compositions & protocols 3.1 Introduction This chapter initially highlights the choices and characteristics of the materials used in this doctoral research, then elaborates on the mortar mixes considered, and subsequently introduces the masonry specimens investigated. The aspects discussed include the mineralogical composition of the binders and aggregates, the protocols adopted to prepare the mortars and, their curing conditions, the repeatability of results, and finally the construction of masonry specimens. To give the reader an overview of the experiments performed and the samples adopted, the chapter also presents a summary of the tests conducted at the mortar and masonry level. Detailed discussions on the methodology used for each test has been discussed in subsequent chapters (Chapter 4 - mortars, Chapter 5 - masonry), within the context of results obtained from each experiment. This chapter is primarily focused on providing the reasoning behind the choices of materials and on the preparation of samples (mortar and masonry). One of the important intents of the present research was to ensure that the mortar mixes selected were representative, in terms of their compositions. A natural consequence of this was the workability of the selected mixes. It was of utmost importance that the mortars investigated in this research program were workable, usable by masons on the construction sites, and so this factor also played a significant role in developing mortar mixes. In particular, representativity contributed to the choice of the reference cement mix. Finally, the choice of materials has been strongly influenced by the intent to ensure repeatability of results in this research campaign, as well as scientific replication and comparison by other researchers. Within the framework of the above-mentioned priorities, the choice of the unit (brick) and possible alternatives for the construction of masonry specimens have also been discussed, together with the reasons that guided the final decision. It is also important to explicitly state the involvement of the Mortar Task Force (MTF) of the European Lime Association (EuLA), in the context of this research program, specifically pertinent to this chapter. Decisions concerning the materials (binders, aggregates, and bricks) were made based on literature review and were complemented by technical discussions between the author, the supervisors of this Ph.D. and the representatives of MTF, EuLA, taking into consideration representativeness and the ability to translate the knowledge obtained from lab work to industrial and construction sites. 3.2 Raw materials
A multi scale approach to the study of lime-cement mortars in masonry 45 3.2.1 Binders The materials used as binders for the mixes were air lime and Portland cement. The type of cement chosen was Portland cement CEM I - 42.5 R. Despite the knowledge that CEM II is more commonly employed in the industry, CEM I was chosen to reduce the number of chemical variables in the mix designs and improve repeatability. The reasoning for this is the composition of different types of cement in EN 197-1 [21]. According to the said standard, CEM II in comparison with CEM I, is allowed to have 30% more variation in its constituents apart from clinker, such as silica fumes, limestone filler, fly ash, calcined Pozzolana, and burnt shale. The composition of these constituents are themselves variable, based on the geographical location they are obtained from and the method of treatment, hence in an attempt to reduce variability, CEM I was chosen. Details of the corresponding batch of cement CEM I – 42.5 R used in this campaign were obtained from the technical data sheet provided by the manufacturer Secil, namely ACM-040/2016. The density and Blaine specific surface of the material specified was 3.12 g/cm3 and 3508 cm2/g respectively. The clinker composition consisted of 12.6% of C2S and 62.2% of C3S. The chemical composition of the main components of cement has been presented in Table 2. The term LOI refers to loss on ignition and was measured based on the recommendations of EN 459-2 [317]. The apparent bulk density measured was equal to 0.93 g/cm3. Table 2: Chemical analysis of main components of cement CEM I - 42.5R as provided by the manufacturer Secil LOI (%) SO3(%) MgO (%) Al2O3 (%) Fe2O3 (%) K2O (%) SiO2(%) CaO(%) 2.05 3.05 1.75 4.27 3.2 0.77 20.55 63.4 The type of air lime chosen for this campaign was CL90-S. Similar to the choice of cement, the selection of the type of lime was based on minimizing the variables influencing the design of mortar mixes. According to EN 459-1 [43], compared to other types of air lime, CL 90-S has the least amount of variation in its chemical composition and the maximum amount of available lime, ≥ 80% by mass, and therefore, it was selected for this research. Lime was provided by Lhoist and details of its composition were obtained from the datasheet provided by the manufacturer for the corresponding batch used; control number 90000998782. The density and BET specific surface area declared were 2.24 g/cm3 and 150000 cm2/g respectively. The mean value of particle size distribution was reported to be in the range of 5.5-6.5 µm. The details of particle size distribution were obtained from laser diffraction (Malvern) (Table 3). The chemical composition of lime as obtained from X-ray fluorescence (Axios Panalytical), expressed as oxide equivalent has been presented in Table 4. LOI referring to loss on ignition was based on the
A multi scale approach to the study of lime-cement mortars in masonry 52 Figure 20: Solid molded clay brick with frogs chosen for the project, supplied by Wienerberger 3.3 Mixes for mortar level study 3.3.1 Preparation and curing protocols The beginning of this research aligned with the timeline of setting up new equipment in the laboratory for casting mortar mixes, especially focusing on the needs of this project (Figure 21). All of the equipment was from the company Matest and the specific models used have been indicated in Figure 21 [339-342]. The equipment included a mixer (E 093) programmed according to protocols of EN 196-1 [340]. According to the protocol, that was used to prepare all the mortar mixes, the binder should be mixed with water at low speed (rotation of 140±5 min-1 and planetary movement of 62±5 min-1) for 60 seconds, followed by high-speed mixing (rotation of 285±10 min-1 and planetary movement of 125±10 min-1) for another 30 seconds. Aggregate is supposed to be added to the mixture, slowly between 30 to 60 seconds. It is then recommended to bring the mixer to rest for 90 seconds. In this interval, the sides of the vessel of the mixer may be scraped to mix the mortar that may be sticking to the sides or bottom of the vessel. Finally, the contents of the mixer are supposed to be mixed at high speed for another 60 seconds, making the total mixing process 4 minutes long. Each batch of mortar was restricted in quantity based on the size of the vessel which had a capacity of 4.7 liters. The rule of thumb used for the quantities of materials, was approximately 3 kgs of aggregates for each casting so that material would not spill out during highspeed mixing. Time ‘zero’ was measured as the moment when water was first brought into contact with the binder. For all batches that were cast in this mixer, from time zero to the moment when the compacted molds were placed in the climatic chamber for curing, it took between 45-60 minutes.
A multi scale approach to the study of lime-cement mortars in masonry 53 Figure 21: Equipment used in the laboratory for casting mortar mixes, from the brand Matest (a) Mixer E 093 [340]; (b) Flow table E 090 [339]; (c) Mold E 105 [341]; (d) Jolting apparatus E 130 [342]; For the studies that took place at the mortar level, raw materials were preconditioned in the same environment to avoid variation in the mortar mixes introduced from changes in temperature or humidity. In this regard, pre-conditioning protocols used by the round-robin testing program of cost action TU 1404 [343] were used as a base, since these guidelines were adopted by them to compare results from the same experiments, across different laboratories in various countries [343]. The binders (cement and lime) were stored in sealed bags at 20±1˚C, for a minimum of 7 days before the actual mixing. Additionally, before each casting, the aggregates were dried completely at 105˚C and cooled down to room temperature (20±1˚C). This was usually done over-night, corresponding to a cooling down period of approximately 15 hours, since the sand would be removed from the oven, an evening before (around 7 PM), and the mixes would be cast the next morning (around 10 AM). The water used for the mixing was also stored in the laboratory at room temperature (20±1˚C) for a minimum of 7 days before casting. With regard to sample preparation at the mortar level, the specimens that were cast in prisms of size 40×40×160 mm3 (molds E 105) were compacted using standard jolting apparatus (E 130 - Figure 21) specified in EN 196-1 [344]. The process consisted of partially filling the molds and jolting them 60 times in exactly 60 seconds. The mold would then be completely filled and jolted another 60 times in 60 seconds. In the case of mortars being cast in a cylindrical shape, each cylindrical specimen had a diameter of 60 mm and a height of 120 mm and was made from a polyvinyl chloride tube with a base of strong adhesive tape. Since this arrangement could not be compacted on the jolting apparatus, a vibrating table was used [345] with specifications 220-240 V, 50 Hz, complying with EN 12390-2 [346]. The
A multi scale approach to the study of lime-cement mortars in masonry 54 cylindrical specimens were subjected to vibration for 10 seconds twice during casting; with the mold first being half-filled and consequently completely filled to assist the removal of air bubbles. During the processes of compaction and vibration, care was taken to observe that no bleeding or segregation of water and mortar mix occurred, because then the mix would be unacceptable and the water-binder ratio would have to be readjusted. An overview of the casting and curing processes used for mortar preparation has been illustrated in Figure 22. Figure 22: Overview of casting of lime-cement mortars according to EN 196-1 [344] and curing according to EN 1015-11 [16] Curing conditions for lime-cement mixes were decided based on recommendations in EN 1015-11 [16]. All lime-cement mortar mixes were cured at 20±1˚C and 95±5% RH for the first 7 days and thereafter at 20±1˚C and 65±5% RH, up to the age of testing. Demolding of lime-cement mortar specimens was also decided based on EN 1015-11 [16], which recommends demolding after 2 days for lime-cement mixes if the amount of lime in the binder is less than 50% by mass, and demolding after 5 days if the amount of lime in the binder is greater than 50% by mass. In the case of the cement-only mix (i.e., no lime present in the binder), specimens were cured according to EN 196-1 [344]. For the first 24 hours, the molds containing cement-only mixes were covered in a plastic bag and placed in a climatic chamber with a temperature of 20±1˚C and RH of 95±5%. Thereafter, the cement-only specimens were submerged in water at 20±1˚C, up to the age of testing.
A multi scale approach to the study of lime-cement mortars in masonry 55 3.3.2 Mortar mix compositions For research at the mortar level, fifteen different mortars were studied to understand the mechanical behavior of lime-cement mixes (Table 8). Binder aggregate (B/Ag) ratios of 1:3, 1:4, 1:5, and 1:6 were tested, expressed in percentage by volume as 33%, 25%, 20%, and 17% respectively, while the quantity of lime in the binder was varied from 10% to 90%, by volume. An effort was made to choose mix compositions that would be representative of what is used on the field, and not just convenient from an academic point of view. For example, the lime content in the binder could have been varied by 10% systematically for regression analyses, i.e., 60%, 70%, and so on. However, out of the lime-cement mortar mixes commonly used on the field 67% lime in the binder is more common (1:2:9), over 60% or 70%. Choosing to directly study a commonly used mix composition, would help translate the research done in the laboratory to practical applications in the field. To guide the process of selecting the most commonly used B/Ag ratios and lime-cement ratios in the binder, the MTF of EuLA was consulted [333], along with identifying the mix compositions, that appeared repeatedly in the literature [2, 9, 59, 95, 347]. Mortar compositions prescribed in the national annexes to Eurocode 6, from different countries were also taken into account [348-355]. The notations adopted, denote the proportion of different constituents of the mix by volume, for example, 1C3L12S represents a 1:3:12 mix ratio of cement C, lime L, and sand S, respectively. Furthermore, to make it convenient for the reader, all graphs and references to different mixes in the text have been provided with the quantity of lime in the binder (by volume) in parenthesis adjacent to the name of the mix, for example, 1C3L12S (75%). All proportions were converted to mass, using apparent bulk densities of cement (0.93 g/cm3), lime (0.36 g/cm3), and sand (1.6 g/cm3), for the sake of consistent measurement of raw materials. The asterisk symbol (*) in Table 8 indicates, that the value lay in the range of 175±10 mm but the exact measurement was not recorded. Table 8: Composition of blended lime-cement mortars (for 1 m3 of mortar produced) Nomenclature (Lime content by volume %) Cement: Lime: Sand (Volume) Cement (kg) Lime (kg) Aggregate (kg) w/b ratio (By weight) w/b ratio (By volume) Flow table value (mm) 9C1L30S (10%) 9:1:30 315.2 13.4 1846.1 0.88 0.77 185 3C1L12S (25%) 3:1:12 262.7 33.4 1846.1 1.00 0.79 165 2C1L9S (33%) 2:1:9 233.5 44.5 1846.1 1.09 0.81 180 1C1L6S (50%) 1:1:6 175.1 66.8 1846.1 1.25 0.81 165 1C2L9S (67%) 1:2:9 116.8 89.0 1846.1 1.58 0.87 165 1C3L12S (75%) 1:3:12 87.6 100.1 1846.1 1.76 0.88 165
A multi scale approach to the study of lime-cement mortars in masonry 56 1C9L30S (90%) 1:9:30 35.0 120.2 1846.1 2.31 0.96 185 3C1L16S (25%) 3:1:16 197.0 25.0 1846.1 1.35 1.07 182.5 2C1L12S (33%) 2:1:12 175.1 33.4 1846.1 1.50 1.11 * 1C1L8S (50%) 1:1:8 131.3 50.1 1846.1 1.72 1.11 180 1C2L12S (67%) 1:2:12 87.6 66.8 1846.1 1.94 1.07 180 2C1L15S (33%) 2:1:15 140.1 26.7 1846.1 1.80 1.33 175 1C1L10S (50%) 1:1:10 105.1 40.1 1846.1 2.21 1.42 180 1C2L15S (67%) 1:2:15 70.1 53.4 1846.1 2.38 1.31 * 1C1L12S (50%) 1:1:12 87.6 33.4 1846.1 2.69 1.73 180 A key aspect of this research program was to ensure that the mixes being studied would be adequate in terms of use in the field. Therefore a target value of 175±10 mm, as per the flow table test [356], was chosen to determine water-binder (w/b) ratios of the mixes (Figure 23). This chosen interval of 175±10 mm also falls in the range of workability assessed as suitable in the work of Hendrickx [9], who presented a doctoral thesis that addresses the workability of mortar mixes, taking into account the technical experience of six international masons, which has been further elaborated on, in Chapter 2, Section 2.3.1. And therefore the water-binder ratio for all mixes was determined, to attain a flow table value of 175±10 mm. In most of the cases, on average, it took 4 trials, (ranging between 2 to 6 attempts) to obtain the desired flow table value of 175±10 mm. For each attempt, the mix was cast, and workability was tested on the flow table according to EN 1015 - 3 [356], using a hand-operated flow table (Figure 21). If the desired flow table value was not attained, the mix was discarded and a new batch of mortar was cast with a modified water-binder ratio. This process was repeated till the flow table value obtained lay in the range of 175±10 mm. The range was also considered wide enough to encompass minor variations that could arise due to changes in room temperature, water temperature or other effects. Figure 23: Target consistency aimed for 175±10 mm, for all mortars
A multi scale approach to the study of lime-cement mortars in masonry 57 In this research, specifically pertaining to mortars, not all tests could be carried out on all 15 mixes. It was decided that the basic properties of mechanical strength (compression and flexure), hardened bulk density, and UPV would be measured for all mixes up to 365 days of age. Thereafter, a few mixes would be selected from the 15 mixes for other tests such as open porosity, E-modulus, and drying shrinkage. Since different tests have been conducted on mortar specimens of various sizes and at multiple ages, an illustration has been provided summarizing the number and composition of mortar mixes that were tested for each of the different mechanical properties (Figure 24). The figure has three columns, with the first two from the left corresponding to discrete measurements, and the third one discussing continuous measurements. Within each column, there are four tiers, the first of which specifies the experiments, and the second tier indicates how many mixes were tested for each of those experiments. The third tier describes the composition of the mixes tested, grouping them by B/Ag ratio, and mentioning the quantity of lime in the binder (% by volume), in parentheses. And the final tier indicates the different ages at which each of those tests were performed, in number of days. A more detailed description of specimen sizes, testing ages, and standards for the experiments performed at the mortar level have been presented in Table 9. All test results were obtained from an average of 3 specimens per mortar type, except for compressive strength in which 6 specimens were used per mortar type and for EMM-ARM (a method used for constant monitoring of E-modulus) in which simultaneous tests were performed on 2 specimens. The details of the methodology, along with a discussion of the results of all mortar level properties will be discussed in Chapter 4. For research at the mortar level alone, it may be noted that results from 409 individual specimens, including prisms and cylinders, were used. Furthermore, mortar was also cast for trial tests, to obtain the right workability and to repeat some of the tests and confirm the results. If those quantities are also accounted for, a conservative estimate would be approximately 650 individual specimens which is a mortar volume of 0.246 m3 implying around 135 batches of mortar cast.
A multi scale approach to the study of lime-cement mortars in masonry 58 Figure 24: Summary of mixes tested for different mechanical properties, for mortar level research pertaining to Chapter 4 Table 9: Summary of tests conducted at the mortar level Name of test Specimens (Units in mm) Mortars tested Recommendations followed Ages of testing (days) Hardened density Prisms 40×40×160 All mortars Based on EN 1015-10 [357] 7, 14, 28, 90, 180, 365 UPV Prisms 40×40×160 All mortars - 7, 14, 28, 90, 180, 365 Flexural strength Prisms 40×40×160 All mortars EN 1015-11[16] 7, 14, 28, 90, 180, 365 Compressive strength Halves of prisms from flexural strength test All mortars EN 1015-11[16] 7, 14, 28, 90, 180, 365 Open porosity Prisms 40×40×160 1C2L9S (67%), 1C1L6S (50%), 3C1L12S (25%) Based on TC 25-PEM [176] 7, 28, 90 Drying shrinkage Prisms 40×40×160 1C3L12S (75%), 1C2L9S (67%), 1C1L6S (50%), 2C1L9S (33%), 3C1L12S (25%) Based on EN 12617-4 [358] 0 to 90 EMM-ARM Cylinder l 550 , ∅ 44 1C3L12S (75%), 1C2L9S (67%), 1C1L6S (50%), 2C1L9S (33%), 3C1L12S (25%) EMM-ARM user manual [359] 0 to 7 E-modulus Cylinder h 120, ∅ 60 1C2L9S (67%), 1C1L6S (50%), 3C1L12S (25%) EN 12390:13 [117] 7, 28, 90 Poisson’s ratio Cylinder h 120, ∅ 60 1C2L9S (67%), 1C1L6S (50%), 3C1L12S (25%) Similar to E-modulus 7, 28, 90 Fracture energy Prisms 40×40×160 1C2L9S (67%), 1C1L6S (50%), 3C1L12S (25%) RILEM 50-FMC [158] 7, 28, 90 3.3.3 Assessment of repeatability Because 15 mortar mixes were being tested for different mechanical characteristics at multiple ages, and since the capacity of the vessel in the mixer was limited to 5 liters, it was inevitable that several batches
A multi scale approach to the study of lime-cement mortars in masonry 59 of the same mortar would have to be cast, on the same day or different days. Therefore, two key factors were paid special attention to, from the point of view of repeatability (consistent repeated measurements under identical conditions including the same operator and machinery): a) Repeatability of casting process of the mix, so that the same mortar was being tested for different mechanical properties. b) Repeatability of experiments. This was also done to acknowledge the learning curve associated with experimentation (casting mixes and testing various mechanical parameters) for the author, who was also the operator. So, experiments from which data was recorded for drawing scientific conclusions, were only accounted for after repeatability was verified. In this regard, before recording any data, there was a special period dedicated to verifying the robustness of the mixing protocol. One of the distinguishing factors involved was the adoption of guidelines from the round-robin testing program of cost action TU 1404 [343] as mentioned in Section 3.3.1, for pre-conditioning of the materials. Additionally, tests were made using two different mortar mixes (cast on different days) for the same property, at the same age, and the results obtained were compared. The parameter chosen for comparison was compressive strength, and the two mixes chosen were 2C1L9S (33%) and 1C2L9S (67%), with less than and more than 50% lime content. The mixes were tested at curing ages of 15 days and 7 days respectively. The results obtained have been shown in Table 10. The difference in the two batches of 1C2L9S (67%) was found to be 1.5% and for the mix 2C1L9S (33%), it was found to be 0.6%. The differences in the results were less than the coefficients of variation in the results. Further, these differences may also very well lie in the range of individual variation of any mortar mix tested for compressive strength even in early ages [59, 360] as will be observed in a detailed discussion on compressive strength in Chapter 4 Section 4.4.2. Another aspect considered was that of quality control, assisted by non-destructive testing, through measurement of ultrasound pulse velocity (UPV). UPV was measured for all mortar specimens, at each age of testing and compared with ‘control specimens’ for their respective mixes, at the same age. It is also possible to observe from values of ultrasound velocity presented in Table 10, that the differences recorded were small. Table 10: Compressive strength of mixes tested for repeatability Mixes 2C1L9S (33%) 1C2L9S (67%) Age of testing (days) 15 7
A multi scale approach to the study of lime-cement mortars in masonry 60 Room temp (˚C) and RH (%) while casting - 1 23.0˚C, 51 % 24.2˚C, 55 % Room temp (˚C) and RH (%) while casting - 2 24.8˚C, 66 % 24.0˚C, 49 % Compressive strength (MPa) – 1 (CoV%) 7.97 (2.4 %) 1.72 (3.7 %) Compressive strength (MPa) – 2 (CoV%) 8.01 (4.3 %) 1.75 (3.3 %) Difference in compressive strength (%) 0.6 % 1.5 % Ultrasound velocity (m/s) – 1 2857 1946 Ultrasound velocity (m/s) – 2 2923 1973 Difference in ultrasound velocity (%) 2.3 % 1.4 % To further utilize the potential of UPV as a non-destructive method to control quality, the same was used for all 15 mixes through 6 different ages, up to 365 days. The specimens that were to be tested for compressive and flexural strength at 365 days of age, were labeled as control specimens for each mix. At each age of testing and for each mix, before testing the specimens for strength, bulk density, and UPV were measured. In parallel, bulk density and UPV were also measured for the control specimens at the same corresponding age. This way, there were two sets of UPV and bulk density measurements available for each mix at each age, except at 365 days because at that age, only the control specimens were left to be tested. The aim of this was to measure the difference in UPV between the control specimens and the actual specimens that were to be tested for strength at each age, serving as a test of coherence in values from different batches of the same mix, aiding the assessment of repeatability and possibly to identify/explain outliers. This process would also serve in monitoring changes in UPV and bulk density over one year in the control specimens of each of the 15 mixes. Amongst the mixes tested, the difference between the UPV values of the control specimens and the actual specimens to be tested at each age was mostly found to be less than 5%, in some cases extending up to 13% (for less than 10 out of 90 cases, Annex-Table 14). The difference in hardened density was also less than 5% mostly, going up to 8% in very few cases Annex-Table 14. These mixes happened to have either large quantities of lime in the binder or high binder-aggregate (B/Ag) ratios. It was thus found reliable to cast several batches of mortar mixes with the mixing protocol defined in Section 3.2.3. Due to the low variations observed, it was also found acceptable to correlate different mechanical properties of UPV, compressive strength, and hardened density of the same mixes cast in different batches. 3.4 Masonry level research
A multi scale approach to the study of lime-cement mortars in masonry 61 3.4.1 Modification in aggregates and reference mix for masonry level research In alignment with the goals of this research campaign (Chapter 1, Section 1.2), one of the objectives was to compare the performance of lime-cement mortared masonry with that of cement mortared masonry. For this purpose, a reference cement mortar with composition 1:5 (Cement: Aggregate) was chosen. The decision to use a cement only mortar as a reference was based on the fact that cement mortars (with B/Ag ratios 1:3, 1:5, and 1:6) are commonly used in the construction industry [13, 361-363]. This reference cement mix 1:5 ended up becoming the reason for modifications in the aggregate. To use this reference mix, it was necessary to find a suitable water-binder ratio that would result in a flow value of 175±10 mm (Section 3.5, Chapter 3). It was found that despite repeated trials, it was not possible to obtain a reference cement mix (1:5) that did not bleed, and yet satisfied the 175±10 mm flow value workability criterion. A mortar that bled would not lead to reliable results since the amount of water in it would always vary based on how much water bled out of it, consequently creating significant variations in the mechanical behavior. Therefore, the next step was to use a plasticizer that would help obtain a mix that would satisfy the workability criterion and prevent bleeding of the mix. For this, various options were considered and tested, the prominent one being Sika Viscocrete (650 duo A). This led to the desired workability in the mix concerning the flow table value of 175±10 mm. However, during the compaction of the mortar into molds, a considerable quantity of water was discarded (Figure 25). This was also not acceptable because there was no way to control or assess the quantity of water being discarded. Next, potential reasons for this problem were analyzed with a special focus on the aggregate. It was possible to pinpoint, that since the aggregate was tailored for this project with a particle size distribution of [0.063,4] mm (Section 3.2.2), it was missing fines less than 63 µm in size. In the case of lime cement blended mixes, this problem of bleeding was not occurring, probably due to the presence of lime and its particle size distribution was in some way serving as fines (Section 3.2.1). Figure 25: Water expelled (a) during compaction from a trial of the reference cement mix (b)
A multi scale approach to the study of lime-cement mortars in masonry 68 Figure 29: Illustration summarizing tests performed for masonry level research
A multi scale approach to the study of lime-cement mortars in masonry 69 4. Mechanical behavior of lime-cement mortars 4.1 Introduction As discussed in Chapter 2, it is possible to find a consensus on the overall trends regarding the impact of the type of binder and quantity of aggregates on the mechanical behavior of mortars. However, there is an evident lack of systematic quantification of such an impact on the basic mechanical properties of mortars, as well as the correlation between different properties. This chapter aims to address that gap, by presenting information on the mechanical behavior of lime-cement mortars and providing a discussion on possible correlations. Following the introduction, there are ten sections (Sections 4.2 to 4.11) in this chapter that cover the methodology and results of different mechanical properties in the order of – workability and corresponding water-binder ratios, mechanical strength, hardened bulk density and UPV, E-modulus, Poisson’s ratio, fracture energy, open porosity, drying shrinkage and EMM-ARM (Elasticity Modulus Measurement through Ambient Response Method). Subsequently, final remarks are presented. The sequence of the sections has been based on the type of measurements (continuous/discrete) and the number of mixes tested in each case (Detailed explanation in Chapter 3). Discrete measurements of workability, mechanical strength (compressive strength and flexural strength), hardened bulk density, and UPV were measured for 15 lime-cement mortars for 6 different ages – 7, 14, 28, 90, 180, and 365 days. The next set of results are for discrete measurements at 7, 28, and 90 days, regarding mechanical properties: E-modulus, Poisson’s ratio, fracture energy, and open porosity have been discussed for three lime-cement mixes. The three mixes (B/Ag ratio 1:3) selected were, one with 50% lime in the binder (1C1L6S (50%)), one with less than 50% lime in the binder (3C1L12S (25%)), and one with more than 50% lime in the binder (1C2L9S (67%)). The choice of these mixes was made based on the selection of other researchers as well as what is often used in the industry, as has been discussed in detail in Chapter 3 [48, 59, 364]. Finally, results from two more mechanical properties, EMM-ARM and drying shrinkage have been presented (continuous measurements), the former from 0 to 7 days, and the latter from 7 to 90 days. Five mixes were studied for these two properties namely 3C1L12S (25%), 2C1L9S (33%) 1C1L6S (50%), 1C2L9S (67%) and 1C3L12S (75%). Details of the specimens and ages of testing have also been presented in Chapter 3. It must be mentioned, that all equations and relationships presented in this chapter are valid only for the materials specified in this thesis. If the nature of any of the materials is changed (lime, cement, or sand) or if other conditions such as ambient temperature or relative humidity are varied, the equations may require recalibration. The analytic expressions presented in this chapter are intended to help better
A multi scale approach to the study of lime-cement mortars in masonry 70 understand the quantitative influence of different factors on mechanical properties. Furthermore, it is also hoped that a consequence of presenting these expressions could be the generalization and development of “rules-of-thumb” that could be used by practitioners in the field or the lab. These correlations could also help cross-check experimentally obtained values or allow estimation of mechanical properties without an actual test, resulting in optimization of time, money, and resources of materials and space. For example, if the strength of a mix is tested at 7 days, the strength of the mix at 365 days could be estimated. 4.2 Methodological aspects considered in statistical correlations a) Regression analyses performed in this thesis used either Origin Pro 9.0 which is a data analysis and graphing software [370], Microsoft Excel, or Python (programming language) on the platform of Google collab [371]. b) Results of different mechanical properties are supplemented with corresponding coefficients of variation (CoV) in parenthesis, wherever applicable. c) Wherever applicable, graphs have been supplemented with error bars that represent the standard error. Standard error was obtained by diving the standard deviation by the square root of the number of specimens used to obtain the average value of that measurement. d) Concerning regression analyses performed in this chapter, the r-squared (R2) value (or the coefficient of determination) was used as an indication of how well, the model or equation fit the data. It is a statistical measure that is used to explain the proportion of variation in the response variable, which can be explained by the regression model being used and varies between 0 and 1. The value 0 indicates that the regression model does not explain any variation, while values approaching 1 indicate that the regression explains most of the variance in the response variable [372]. e) Another characteristic that was taken into account while considering the suitability of the regression analyses was the p-value. A small p-value indicates that the relationship between the dependent and independent variables is not due to chance, thereby allowing the rejection of the null hypothesis. A null hypothesis in turn is the assumption that there is no relation between the dependent and independent variable. Therefore, a p-value smaller than the significance level, which is most commonly set as 0.05, in addition to an r-squared value approaching 1, validates the regression model. In this research, most results presented were found to have a p-value notably less than 0.05, indicating their statistical significance [373]. However, since the p-value is only a probability of the null hypothesis being right, it cannot be treated in isolation and must be interpreted as a part of the
A multi scale approach to the study of lime-cement mortars in masonry 71 entire research accounting for the sample size, sample treatment, and outcome [374]. In a few cases where regression was performed in this research with only three data points available, the pvalue exceeded 0.05, by a small margin, despite reasonable R-squared values (generally > 0.94, in some cases ranging till 0.85). The data was not rejected in these cases because the p-value is dependent on sample size, and having only 3 data points would require exceptionally high R-squared values to result in p-values less than 0.05 Since inadequacy of data points in these cases could lead to unreliable conclusions based on p-values, values > 0.05 were also considered acceptable. f) The F-critical value (also known as F-statistic) was also considered in addition to the p-value, and it is calculated as the ratio of the mean square of the fitted model and the mean square of errors [375]. It indicates if the fitted model performs better than the intercept only model, with zero predictor variables and a constant result. If the F-value obtained from the data is greater than the Fcritical value, this indicates that the coefficients added to the regression model, improve the explanation of the response variable. Most often, F-value being greater than F-critical is indicated by Prob(F-statistic) or p-value, which if less than 0.05, indicates that the F-value obtained from the data is greater than the F-critical value. [372, 373, 375]. 4.3 Workability and water-binder ratios It was of utmost importance that the workability of mixes chosen was representative of what is used on the field by masons and consistent throughout the experimental program, as discussed in detail in Section 3.3.2, Chapter 3. Therefore, workability was measured following EN 1015-3 [75] and the target flow table value chosen was 175±10 mm for all mixes. The water-binder ratio (by mass) for each mix was determined experimentally, by trial and error (Figure 30).
A multi scale approach to the study of lime-cement mortars in masonry 72 Figure 30: Water-binder ratio (by mass) of mixes expressed as a function of lime content in the binder, in the workability range of 175±10 mm, measured using the flow table test according to EN 1015-3 [75]. The values 1:3, 1:4, 1:5 and 1:6 indicate B/Ag ratios. The R-squared values shown are for individual B/Ag ratios (only one mix for 1:6). It was observed that for the same B/Ag ratio, the requisite water for the mix changed each time the quantity of lime in the binder was varied. It was possible to identify a linear pattern that related the amount of lime in the binder (by volume) with the water-binder ratio (Figure 30). It was also discovered that each time the binder-aggregate ratio (by volume) was changed; the linear relation had to be recalibrated. If the quantity of lime in the binder was kept constant and the B/Ag ratio was changed, it was possible to observe the influence of the latter on the w/b ratio as well. Therefore, a multiple linear regression was performed and has been presented to correlate the quantity of lime in the binder (by volume) and the B/Ag ratio (by volume) with the water-binder ratio (by mass) of each mix, which resulted in an adjusted R2 value of 0.93, and F-value of 87.08 (Prob>F-value 1.8e-7). Attempts were made to use the water-binder ratio by volume as the dependent variable for the multiple linear regression. However, a good fit was not obtained and therefore, the water-binder ratio shown is by mass or weight in Equation 11. w b=2.495+0.0168 (Lime in binder %)−0.059 (B Ag%) 11 All 15 mortar mixes (Detailed in Chapter 3) were used as input for the regression performed (Equation 11). Predicted values were compared with actual values (average difference in absolute value was 7.1%),
A multi scale approach to the study of lime-cement mortars in masonry 73 and it was found that the difference was mostly less than 10% (Table 13) except for the mix 1C1L12S (50%) with a difference of around 12.5%, and the mix 9C1L30S (10%) with a difference of around 20%. Since Equation 11 has three terms, a simpler expression has also been provided in equation 12 as an alternative with a common slope for all mixes, where the intercept varies as a function of the B/Ag ratio. w b=0.0158 (Lime in binder %)+f(B Ag) 12 The slope 0.0158 was calculated by averaging the value of slopes of all three linear regressions shown in Figure 30 and the term f(B Ag) is the value of the intercept obtained from the average difference in different intercepts (0.357) from individual regressions, added to the intercept from B/Ag ratio 1:3 (0.567). Therefore, f(B Ag) for 1:3 is 0.567, for 1:4 is 0.924 (which is 0.567+0.357), for 1:5 is 1.282 (which is 0.924+0.357), and for 1:6 it is 1.639 (which is 1.282+0.357). Table 13: Water-binder ratio (by mass) for different mixes - used, predicted, and difference (%) Mortar mixes Lime content in binder (%) B/Ag ratio (%) Actual waterbinder ratio Predicted water-binder ratio Difference (%) Equation 10 Equation 11 Equation 10 Equation 11 9C1L30S 10 33.3 0.88 0.69 0.72 -20.8 -17.4 3C1L12S 25 33.3 1.00 0.95 0.96 -5.2 -3.7 2C1L9S 33.3 33.3 1.09 1.09 1.09 -0.4 0.3 1C1L6S 50 33.3 1.25 1.37 1.36 9.0 8.2 1C2L9S 67.7 33.3 1.58 1.66 1.64 5.3 3.6 1C3L12S 75 33.3 1.76 1.79 1.75 1.2 -0.8 1C9L30S 90 33.3 2.31 2.04 1.99 -11.8 -13.9 3C1L16S 25 25 1.35 1.44 1.32 6.4 -2.4 2C1L12S 33.3 25 1.50 1.58 1.45 5.3 -3.2 1C1L8S 50 25 1.72 1.86 1.71 7.9 -0.5 1C2L12S 67.7 25 1.94 2.15 1.99 10.8 2.5 2C1L15S 33.3 20 1.80 1.87 1.81 4.1 0.5 1C1L10S 50 20 2.21 2.15 2.07 -2.6 -6.3 1C2L15S 67.7 20 2.38 2.45 2.35 3.0 -1.2 1C1L12S 50 16.7 2.69 2.35 2.43 -12.6 -9.7 Average 7.1 4.9 The objective of proposing such equations, (calibrated according to the materials used) is to reduce the number of trials that would otherwise be required to determine the water-binder ratio for a targeted flow
A multi scale approach to the study of lime-cement mortars in masonry 74 value. The values estimated from Equations 10 and 11 and presented in Table 13 have been plotted alongside the actual water-binder ratios used in the mortar mixes, in Figure 31. Figure 31: Experimental water binder ratios (Table 13) versus estimated according to (a) Equation 10 (b) Equation 11 4.4 Mechanical strength 4.4.1 Methodology Compressive strength and flexural strength of the mortar mixes were measured according to EN 101511 [16], on a universal testing machine from the company Lloyd. Three prismatic specimens of size 40 × 40 × 160 mm3 were used in each 3-point bending test to measure flexural strength (Figure 32 (a)). All specimens initially tested for flexural strength and were loaded by controlling the displacement, at a rate of 12 µm/s. Subsequently, the halves obtained after flexural strength testing were used for the test of compressive strength; 6 specimens/measurements contributed to the final value of compressive strength, in each case. For compressive strength, the contact area between the specimen and equipment for load application was 40 mm × 40 mm, and the height of the specimen was 40 mm (Figure 32 (b)). A pre-load of 150 N was applied on all specimens and the rate of loading was 50 N/s.
A multi scale approach to the study of lime-cement mortars in masonry 75 Figure 32: Images for (a) Flexural test (3-point bending) and (b) Uniaxial (unconfined) compression test for mortar specimens 4.4.2 Results of mechanical strength Since the mechanical strength of lime-cement has been expressed as a function of common factors, the factors have been explained below along with their notations. a) Lime content in the binder (10, 25, 33.3, 50, 66.7, 75, 90) % by volume. For example, lime content in the mix 1C2L9S has been expressed as 67% by volume of the binder. b) Binder aggregate ratio or B/Ag ratio (1:3, 1:4, 1:5, 1:6) or (33.3, 25, 20, 16.7) % by volume. For example, the mix 1C2L9S has a binder-aggregate ratio of 1:3 or 1/3 expressed as 33% binder in the mix, by volume. c) Curing age, (7, 14, 28, 90, 180, 365) days. For the 15 mortar mixes specified in Chapter 3, for ages from 7 days to 365 days, values of compressive strength obtained have been presented in Table 14 and range from 0.14 MPa to 15.5 MPa, while values of flexural strength range from 0.04 MPa to 5.22 MPa as presented in Table 15. It is recognized that a direct comparison of the values obtained here with those found in the existing literature is not possible because, among many other factors, researchers use different types of lime or cement. However, a comparison of the general range of values is feasible and must be discussed to validate the strengths obtained in this thesis. Arandigoyen et al. [59] present one of the widest ranges of data of compressive strength available on lime-cement blended mortars (5–13 MPa) and this is similar to the range of values obtained here (2–15 MPa). However, in some cases, the values of strength presented by them are slightly higher than the corresponding values obtained in this campaign. One of
A multi scale approach to the study of lime-cement mortars in masonry 76 the plausible reasons for this could be the use of a lower water-binder ratio by them due to differences in the type of cement (Type II versus Type I in this campaign) or difference in target workability for the mixes. But this is difficult to verify because the water-binder ratio used to prepare their mixes has not been explicitly mentioned. The values of compressive strength presented by Cizer [2] seem to be in the range of 5–35 MPa, considerably higher than those presented here. But the flow value targeted was between 120 and 130 mm which is likely to explain the difference, since the lower the amount of water used in the mix, the greater is the strength of the mix obtained [96]. Data on compressive strength of blended lime-cement mortars presented by Haach et al. was found to be similar to the values obtained in this experimental campaign for approximately the same consistency [11, 96]. Similarly, it was found that if similar consistency (flow value) is targeted, flexural strength attained in blended lime-cement mortars falls in the range of data found in the current experimental campaign, as observed from the work of Macharia [95]. Eurocode 6 [18], together with EN 998-2 [376], characterize mortars into classes based on the minimum compressive strength (in MPa) that the mortar must attain in 28 days. This categorization is based only on strength and does not discuss the composition of the mortar. However, in the different national annexes to the current version of the Eurocode 6 [18], more details are provided on how different compositions of mixes, may lead to the fulfillment of strength requirements for different mortar classes. It was found that in the cases where the same mortar mixes were tested, the strength at 28 days was in the expectable mortar class range. For example, the Polish national annex [377] suggests a mix composition of 1:2:9 in the category of 2.5 MPa. As may be observed from Table 14, the mix 1:2:9 attains a strength of 2.35 MPa in 28 days. The Belgian national annex [350] suggests mix compositions of 1:1:6 and 2:1:9 for the categories of 5 MPa and 8 MPa respectively. In this research, the mixes 1:1:6 and 2:1:9 led to strengths of 4.68 MPa and 8.13 MPa respectively (Table 14). Table 14: Compressive strength values of lime-cement mortars from 7 days to 365 days Mortars fc-7 (MPa) (CoV %) fc-14 (MPa) (CoV %) fc-28 (MPa) (CoV %) fc-90 (MPa) (CoV %) fc-180 (MPa) (CoV %) fc-365 (MPa) (CoV %) 9C1L30S (10%) 8.94 (3.6%) 10.64 (9.6%) 12.11 (4.4%) 12.22 (8.0%) 11.29 (7.4%) 15.46 (5.5%) 3C1L12S (25%) 7.66 (3.5%) 9.91 (5.0%) 9.95 (3.6%) 9.28 (0.6%) 9.92 (6.3%) 12.19 (5.3%) 2C1L9S (33%) 6.09 (2.8%) 7.36 (4.9%) 8.13 (6.6%) 8.57 (6.0%) 8.73 (9.7%) 8.65 (6.7%) 1C1L6S (50%) 4.12 (5.5%) 5.41 (8.8%) 4.68 (6.3%) 6.23 (6.9%) 6.31(3.9%) 7.01 (6.4%) 1C2L9S (67%) 1.48 (6.7%) 1.90 (2.8%) 2.39 (2.6%) 2.45 (6.7%) 2.69 (10.5%) 2.67 (6.9%) 1C3L12S (75%) 0.63 (9.0%) 1.16 (7.2%) 1.37 (3.3%) 1.53 (3.1%) 1.55 (6.7%) 1.61 (3.5%) 1C9L30S (90%) 0.14 (9.8%) 0.22 (11.8%) 0.30 (6.7%) 0.41 (3.5%) 0.45 (5.7%) 0.43 (7.6%)
A multi scale approach to the study of lime-cement mortars in masonry 77 3C1L16S (25%) 4.59 (6.6%) 6.46 (1.4%) 6.01 (2.1%) 7.05 (4.8%) 7.85 (3.5%) 7.77 (5.1%) 2C1L12S (33%) 3.23 (5.7%) 5.03 (3.2%) 5.09 (9.5%) 5.73 (3.5%) 6.34 (2.6%) 5.16 (5.7%) 1C1L8S (50%) 2.31 (1.1%) 3.12 (8.7%) 3.00 (8.2%) 3.53 (5.4%) 3.10 (9.6%) 3.56 (5.3%) 1C2L12S (67%) 0.77 (7.0%) 1.37 (6.6%) 1.41 (9.0%) 1.19 (12.0%) 1.45 (6.1%) 1.62 (5.7%) 2C1L15S (33%) 1.86 (13.0%) 2.83 (9.5%) 3.08 (7.6%) 3.26 (12.4%) 3.69 (10.1%) 3.26 (3.5%) 1C1L10S (50%) 1.31 (3.2%) 1.77 (15.6%) 1.59 (18.9%) 1.74 (17.9%) 1.99 (5.0%) 2.05 (17.2%) 1C2L15S (67%) 0.46 (8.0%) 0.83 (8.9%) 0.85 (7.0%) 0.86 (8.6%) 0.90 (12.8%) 0.96 (6.7%) 1C1L12S (50%) 0.85 (12.2%) 1.34 (22.2%) 1.39 (17.5%) 1.19 (12.0%) 1.34 (14.9%) 1.62 (5.8%) Table 15: Flexural strength values of lime-cement mortars from 7 days to 365 days Mortars ff-7 (MPa) (CoV %) ff-14 (MPa) (CoV %) ff-28 (MPa) (CoV %) ff-90 (MPa) (CoV %) ff-180 (MPa) (CoV %) ff-365 (MPa) (CoV %) 9C1L30S (10%) 2.67 (9.8%) 3.15 (3.6%) 3.89 (12.3%) 3.15 (12.3%) 3.75 (6.7%) 5.22 (2.0%) 3C1L12S (25%) 1.95 (6.6%) 2.58 (2.7%) 2.76 (7.0%) 3.22 (4.5%) 3.64 (4.2%) 4.60 (8.0%) 2C1L9S (33%) 1.52 (5.6%) 2.32 (0.9%) 2.60 (8.6%) 2.52 (2.6%) 3.03 (5.5%) 3.11 (6.8%) 1C1L6S (50%) 1.23 (4.8%) 1.69 (4.0%) 1.96 (6.0%) 2.14 (6.7%) 2.31 (2.0%) 2.54 (1.9%) 1C2L9S (67%) 0.41 (7.7%) 0.70 (11.3%) 0.69 (4.5%) 0.86 (6.3%) 0.99 (2.9%) 0.98 (2.6%) 1C3L12S (75%) 0.28 (4.5%) 0.44 (10.8%) 0.48 (16.0%) 0.58 (7.6%) 0.61 (10.5%) 0.52 (16.9%) 1C9L30S (90%) 0.04 (29.9%) 0.04 (28.3%) 0.19 (18.1%) 0.28 (7.7%) 0.26 (5.2%) 0.21 (4.1%) 3C1L16S (25%) 1.23 (2.3%) 1.72 (11.2%) 2.10 (6.6%) 2.35 (12.6%) 2.95 (3.4%) 2.74 (7.9%) 2C1L12S (33%) 1.12 (9.3%) 1.50 (6.4%) 2.09 (1.7%) 2.09 (2.5%) 2.26 (9.9%) 2.08 (1.4%) 1C1L8S (50%) 0.70 (10.2%) 0.95 (17.3%) 1.08 (11.9%) 1.20 (3.9%) 1.24 (4.7%) 1.17 (5.4%) 1C2L12S (67%) 0.30 (6.1%) 0.53 (8.1%) 0.52 (10.2%) 0.50 (0.5%) 0.51 (13.7%) 0.60 (11.7%) 2C1L15S (33%) 0.66 (5.0%) 0.99 (9.2%) 1.16 0.8%) 1.26 (6.8%) 1.30 (7.2%) 1.32 (2.5%) 1C1L10S (50%) 0.37 (14.7%) 0.58 (11.0%) 0.64 (3.3%) 0.66 (13.8%) 0.74 (9.0%) 0.77 (3.7%) 1C2L15S (67%) 0.21 (2.9%) 0.34 (14.3%) 0.35 (37.9%) 0.36 (10.1%) 0.39 (14.8%) 0.31 (12.3%) 1C1L12S (50%) 0.29 (3.4%) 0.42 (20.9%) 0.50 (17.3%) 0.48 (9.7%) 0.51 (11.7%) 0.60 (11.7%) If all 15 mixes are taken into account at 6 different curing ages, there are 90 data points available for compressive strength and flexural strength each. Therefore, an attempt was made to correlate the two properties. It was found that the ratio of compressive strength to flexural strength was approximately 3 for all mixes, across all ages (Table 16). The general range of the ratios varied between 2.5 to 3.5, and therefore an average ratio was calculated for each mix across different ages and has also been presented in Table 16, along with the coefficient of variation. Table 16: Ratio of compressive strength to flexural strength for mixes of different ages fc ff (MPa/MPa) 7d 14d 28d 90d 180d 365d Average CoV (%) 9C1L30S (10%) 3.3 3.4 3.1 3.9 3.0 3.0 3.3 10.3
A multi scale approach to the study of lime-cement mortars in masonry 84 Figure 38: Evolution of flexural strength with time estimated according to Equation 13 For the proposed model, the two variables that require fitting are a and b. Based on equation 13, in the event, that time (t) tends to infinity (∞), the value of 𝑒1 ∞ would tend to 𝑒0 which is 1, implying that fc or compressive strength would tend to the value of parameter a, which would theoretically be the strength of the mortar once it is mature, between 180 and 365 days of age, based on data presented in Table 14. Furthermore, it was found that the ratio between strength (both 𝑓𝑐 and 𝑓𝑓) at 365 days of age and 7 days of age varied in the range of 1.6 to 2.5 (Annex-Table 5 and Annex-Table 6). Therefore, parameter a could be expressed in the range of 1.6 to 2.5 times, a function of the compressive strength at 7 days of the mixes. This factor ranging between 1.6 and 2.5 could potentially be generalized for different mixes. What makes that interesting is that if then, the strength of a mix is normalized with respect to its strength at 7 days, it is possible to obtain a single curve that could represent the normalized evolution of strength with time for lime-cement mixes. This was achieved through curve fitting and regression and has been shown in equation 14. This data set, therefore, included 14 mixes, at 6 different ages, which means a total of
A multi scale approach to the study of lime-cement mortars in masonry 85 84 data points. For this analysis, all mixes were used except for 1C9L30S (90%) since it was a clear outlier, in terms of the rate of gain of strength as well as the very small absolute values of strength at day 7. The most probable reason for this is the large quantity of lime in its binder, which makes it behave similar to an only lime binder. fc(T) fc(7)=1.7 e−1.2 √T when lime in binder ≤ 50% by volume fc(T) fc(7)=2.4 e−1.7 √T when lime in binder > 50% by volume ff(T) ff(7)=2.2 e−1.8 √T 14 Equation 14, therefore, allows the user to estimate the compressive strength of any of the lime-cement mixes discussed if the strength of that mix is tested at 7 days of age. Based on equation 14, if the estimated values are compared with the experimentally obtained values, it was found that the average difference was around 0.44 MPa or 11.2%, in the case of compressive strength, with the maximum difference going up to 27%, as shown in Annex-Table 7, despite low R-squared values of 0.56 and 0.60 in case of less than and greater than 50% lime in the binder, respectively. In the case of flexural strength, the R-squared value found was 0.72 and the average difference in estimated and experimental values was found to be around 10% as well or 0.15 MPa, as shown in Annex-Table 8. The evolution of compressive and flexural strength with time, based on equation 14, is illustrated in Figure 39 and Figure 40. In the cases of mixes with ≤ 25%, lime in the binder such as 9C1L30S (10%) and 3C1L12S (25%), the functions presented in equation 14 can quite significantly overestimate the strength of the mixes at certain ages, since they behave differently from the other mixes between 90 and 365 days (~up to 30% in compressive strength and even up to 60% in flexural strength Figure 35 and Figure 36). And therefore, at this stage, these equations cannot yet be safely used for extrapolation for all lime-cement mixes. It is recognized that a significantly larger data set would be required, with more mixes tested over time, to obtain better fitting and consequently higher r-squared values, which would ultimately lead to better estimations of strength. The findings reported herein seem to point in the direction of the plausibility of potential generalizations when adequate data is available.
A multi scale approach to the study of lime-cement mortars in masonry 86 Figure 39: Evolution of compressive strength with time estimated according to Equation 14
A multi scale approach to the study of lime-cement mortars in masonry 87 Figure 40: Evolution of flexural strength with time predicted according to Equation 14 Another observation that can be made from equation 14, is that in the case of flexural strength, only one expression was found to be adequate for all the mixes. However, for compressive strength it is possible to distinguish the normalized evolution of strength with time, depending on the mixes having more than or less than 50% lime in the binder, since the values of the constants a and b obtained were different. This has been demonstrated graphically in Figure 41. It may be noted that the two curves shown in Figure 41 do not start from the value 1 because they represent the analytic expression of equation 14, which is obtained from the best fit of the evolution of strength with time for 15 mixes. Based on the values of a and b obtained in equation 14 for the two different curves, as well as Figure 41, it is possible to conclude, that the evolution of strength in lime dominant mixes is slower, and therefore continues for longer. Their normalized growth in strength is also higher than that of cement dominant blended mixes. This may be attributed to the slow kinetics of the carbonation process in lime dominant mortars, which could take
A multi scale approach to the study of lime-cement mortars in masonry 88 several months or even years to progress [47], compared to the relatively faster process of cement hydration, the bulk of which takes place within the first 28 days [2, 37]. The curves in Figure 41, appear to confirm the result of the difference in kinetics of the two phenomena, which have been discussed in detail in Chapter 2. Figure 41: Graphical representation of Equation 14 – Evolution of compressive strength with time, normalized with respect to strength at day 7, for lime v/s cement dominant blended mixes 4.4.2.2 Impact of lime content in the binder on the mechanical strength of mortar This discussion assesses the impact that lime has on the strength of the mortar by varying lime content in the binder of mixes with different B/Ag ratios (1:3, 1:4, and 1:5 by volume). To assess and quantify the influence of this factor, linear regressions were performed on data obtained at different ages. To facilitate comparison and normalization, a benchmark mortar had to be chosen, with the highest feasible strength, which is a mix with the least amount of lime in the binder (10%), for any given B/Ag ratio. Furthermore, 10% lime means 90% cement in the mix, which means it is close to the composition of a cement mortar and yet it retains the label of lime-cement masonry mortar from a practical point of view, because of the presence of 10% lime in the binder. Allowing this benchmark mix to have some amount of lime in it ensures that it is subjected to the same curing conditions, thereby reducing variables in the comparison. Based on the rationale presented, while studying the impact of lime in the binder on the mechanical performance of mortar, the benchmark mix always consists of 10% lime or 90% cement (by volume).
A multi scale approach to the study of lime-cement mortars in masonry 89 It was found that for a given B/Ag ratio at any age, the experimentally obtained values of compressive/flexural strength vary linearly with the quantity of lime present in the binder of the mix. To illustrate this, Figure 42 and Figure 43 show the variation of compressive and flexural strength respectively with lime content in the binder for three different mixes with different B/Ag ratios and at different ages; (a) B/Ag ratio 1:3 at day 7, (b) B/Ag ratio 1:4 at day 90 and (c) B/Ag ratio 1:5 at day 365. Figure 42: Illustration of a linear relationship between the compressive strength of mortar and lime content in the binder (% by volume) for different B/Ag ratios and at different ages Figure 43: Illustration of a linear relationship between the flexural strength of mortar and lime content in the binder (% by volume) for different B/Ag ratios and at different ages In a linear relationship, the slope indicates the change in the dependent variable for a unit change in the independent variable. In this context that would mean the change in compressive/flexural strength of the mix for a unit change in lime content of the binder. If the values of slope obtained are to be normalized across different B/Ag ratios, for comparison, it would be necessary to divide the value of slope by the strength of the benchmark mix and expressed as a percentage. This has been illustrated in Figure 44, using compressive strength as an example but the same is applicable for flexural strength as well. Furthermore, Table 17 and Table 18 summarize the data that was obtained from treating the variation of compressive and flexural strength respectively with lime content in the binder for all ages – days 7, 14, 28, 90, 180, and 365. This has been shown for all three B/Ag ratios 1:3, 1:4, 1:5. Table 17 and Table
A multi scale approach to the study of lime-cement mortars in masonry 90 18 also show the R-squared value for each case, establishing the goodness of fit for the linear relationship in each case. It may be observed that in all cases the R2 value was 0.98 on an average, going to 0.94 and 0.93 in one case each of compressive and flexural strength. It must be noted that in the case of B/Ag ratio 1:3, the mix 1C9L30S (90%) was excluded from the linear regression for both compressive and flexural strength because it reduced the linearity of the rest of the data set, and consequently the accuracy of the estimations. Additionally, the absolute values of strength obtained for this mix were so low, that it was very sensitive to variations in estimations of strength. For example, an average absolute difference of 0.2 MPa, in estimation and actual strength could lead to more than 100% error and so 90% lime was not included in the data set for regression. Figure 44: Illustration of calculation of normalized slope using compressive strength and lime content in the binder as an example (Also applicable to flexural strength because of the linear relationship demonstrated in Figure 43) Table 17: Change (%) in compressive strength of lime-cement mixes for a unit change in lime content in the binder (% by volume) Change (%) in compressive strength for every 1% change in lime content in binder | Benchmark mix has 10% lime and 90% cement in the binder B/Ag ratio 1:3 1:4 1:5 Data points – lime content (% vol) 10, 25, 33.3, 50, 66.7, 75 25, 33.3, 50, 66.7 33.3, 50, 66.7 Age (Curing days) R2 ∆fc(%) R2 ∆fc(%) R2 ∆fc(%) 7 0.99 -1.43 0.97 -1.50 0.99 -1.43 14 0.98 -1.38 0.99 -1.47 1.00 -1.40
A multi scale approach to the study of lime-cement mortars in masonry 91 28 0.99 -1.40 0.99 -1.44 0.96 -1.45 90 0.99 -1.34 1.00 -1.51 0.97 -1.46 180 0.98 -1.30 0.97 -1.55 0.98 -1.49 365 0.98 -1.40 0.94 -1.47 1.00 -1.40 Average of all ages (CoV) 0.99 (0.7%) -1.4 (3.4%) 0.98 (2.2%) -1.5 (2.5%) 0.98 (1.6%) -1.4 (2.5%) Table 18: Change (%) in flexural strength of lime-cement mixes for a unit change in lime content in the binder (% by volume) Change (%) in flexural strength for every 1% change in lime content in binder | Benchmark mix has 10% lime and 90% cement in the binder B/Ag ratio 1:3 1:4 1:5 Data points – lime content (% vol) 10, 25, 33.3, 50, 66.7, 75 25, 33.3, 50, 66.7 33.3, 50, 66.7 Age (Curing days) R2 ∆ff(%) R2 ∆ff(%) R2 ∆ff(%) 7 0.98 -1.41 0.99 -1.40 0.97 -1.40 14 0.99 -1.31 0.99 -1.33 0.97 -1.35 28 0.98 -1.36 0.96 -1.43 0.97 -1.41 90 0.93 -1.23 0.99 -1.46 0.96 -1.44 180 0.96 -1.26 0.98 -1.55 0.98 -1.41 365 0.98 -1.39 0.97 -1.48 0.99 -1.48 Average of all ages (CoV) 0.97 (2.2%) -1.3 (5.5%) 0.98 (1.5%) -1.4 (5.2%) 0.97 (1.2%) -1.4 (3.2%) It was thus found that for every 1% increase in lime content in the binder (by volume), on average there is a 1.4% decrease in compressive and flexural strength of the given mix with respect to the benchmark mix (90% cement in the binder by volume), at any age and for any B/Ag ratio. This analysis makes it possible to theoretically estimate the compressive/flexural strength of any lime-cement mix with a specified quantity of lime in the binder if the strength of the benchmark mix is known. To test if this method leads to a reasonable approximation of the strength of a mix, the compressive strength of the mix 1C2L9S (67%) at 90 days of age, is estimated as an example. The B/Ag ratio would be 1:3 and the benchmark mix would have 90% cement or 10% lime in the binder and therefore it would be the mix 9C1L30S (10%). From experimental testing (Table 14), it is known that the value of the strength of this mix at 180 days of age is 11.29 MPa, and from Table 17 it may be noted that for B/Ag ratio 1:3, at 180 days, the change in strength would be -1.30% for every 1% increase in lime content in the binder (by volume). The increase in lime content is 66.67–10 = 56.67%, therefore the change in strength should be 56.67×(-1.30) = -73.67%. The strength would then be 11.29 – (73.67% of 11.29) = 2.97 MPa. The experimental value recorded in the test was 2.69 MPa (CoV 10.5%), which means a good estimation. Using the same logic, as just presented, values of compressive strength have been estimated for different mixes using Table 17 and compared with the experimental values (Table 14) with the differences (%)
A multi scale approach to the study of lime-cement mortars in masonry 92 presented in Annex-Table 9. The average difference in estimation and the experimental value was found to be 7.5%, with the maximum difference going up to 27%. In the case of flexural strength (Table 15) values were estimated using Table 18 and differences have been presented in Annex-Table 10. The average difference in estimation and the experimental value was found to be 8.8% and the maximum difference was found to be 25%. 4.4.2.3 Impact of binder-aggregate ratio (B/Ag) on the mechanical strength of mortar This section assesses the impact of the quantity of the binder present in a mix, across fixed binder compositions (33.3%, 50%, and 66.7% lime in the binder, by volume). In the case of comparison of different B/Ag ratios, out of the options available (1:3, 1:4, 1:5, 1:6), the binder-aggregate ratio of 1:3, has the most amount of binder in the mix, and for a given binder composition, is likely to result in the highest strength of the mix. And therefore, the benchmark mix would be a mortar with a B/Ag ratio of 1:3, by volume. It is also the strongest composition possible from a practical point of view of what may be used in the field. While studying the impact of binder in the mix on the mechanical performance of mortar, the benchmark mix always consists of 33% binder or binder-aggregate ratio of 1:3 (by volume). The compressive and flexural strengths of different mixes were found to vary linearly with the B/Ag ratio of different mixes (Figure 45 and Figure 46). Figure 45: Illustration of a linear relationship between compressive strength of mortar and B/Ag ratio (% by volume) for lime contents in the binder and at different ages
A multi scale approach to the study of lime-cement mortars in masonry 93 Figure 46: Illustration of a linear relationship between flexural strength of mortar and B/Ag ratio (% by volume) for lime contents in the binder and at different ages The impact of the binder on the strength of mortar mixes was quantified similarly to how the impact of the lime content in the binder was assessed (Section 4.4.2.2). The values of slopes, indicating the change in compressive/flexural strength of the mix for a unit change in B/Ag ratio of the mixes, were normalized, using the benchmark mix and expressed as a percentage (An example has been illustrated in Figure 47 using flexural strength, but the same applies to compressive strength, because of the linear relationship exhibited in Figure 45). The data obtained from treating the variation in mechanical strength with B/Ag ratio has been shown in Table 19 and Table 20 for all ages – days 7, 14, 28, 90, 180, and 365. This has been shown for three lime-cement compositions 33.3%, 50%, and 66.7%, lime in the binder, by volume. Table 19 and Table 20 also show the R2 value for each case, establishing the goodness of fit for the linear relationship in each case. It may be observed that in all cases the R2 value was 0.98 on average, going down to 0.97 in case of compressive and till 0.95 in case of flexural strength.
A multi scale approach to the study of lime-cement mortars in masonry 100 Figure 51: Evolution of ultrasound pulse velocity (UPV) with time for different lime-cement mixes 4.5.2.2 Correlation of UPV, bulk density, and compressive strength It is known from mechanics that the square of UPV is directly proportional to Young’s modulus (E-mod) [105]. Furthermore, from studies of the mechanical behavior of concrete, it is known that a product of density and compressive strength ρafcb is also proportional to E-mod, where the values of a and b may vary based on the experimental data, but are often equal to 0.5 [379, 380]. Based on this, a linear relationship was found between a function of UPV (m/s) and a function involving bulk density (kg/m3) and compressive strength (MPa) (Equation 16). UPV(T)2=k.ρ(T)1.5fc(T)0.5 16
A multi scale approach to the study of lime-cement mortars in masonry 101 The symbol T is time and the symbol k is a constant of proportionality. Since for a given B/Ag ratio, compressive strength decreases linearly with an increase in lime content in the binder (Figure 42), and UPV and density also appeared to decrease with an increase in lime content in the binder, equation 16 was plotted graphically, to check if it had a relation with lime content in the binder for any given B/Ag ratio. An example has been illustrated in (Figure 52) for a B/Ag ratio of 1:3, where the y-axis corresponds to Y(T)=UPV(T)2 and the x-axis corresponds to X(T)=ρ(T)1.5fc(T)0.5 and T is equal to 28 days. The R2 value of 0.97 obtained, verifies the linearity of the relationship. Figure 52: Correlation between ultrasound velocity and a function of density and compressive strength for varying lime content in the binder (10, 25, 33.3, 50, 66.7, 75) % by volume, for B/Ag ratio 1:3 Keeping the B/Ag ratio constant, this linear relationship was thereafter verified for different curing ages. Subsequently, it was tested for different B/Ag ratios as well, and all cases were found to be linear, with high R2 values; the lowest R2 value being 0.96 (Table 22). To assess, whether the linearity would be valid if the binder composition was kept constant and the B/Ag ratio was varied, linear regressions were performed for fixed lime-cement compositions as well (Table 23). In this case, as well, high R2 values were obtained, with the lowest going down to 0.85. It may thus be concluded, that equation 16 has the potential to provide a relatively simple, non-destructive method to estimate the compressive strength of mixes without actually testing them since the measurement of UPV is quite simple and so is the measurement of weight of mortar specimens. For this purpose, however, the value of the constant k would have to be established for a given set of materials, which is possibly not an easy task. Once that is
A multi scale approach to the study of lime-cement mortars in masonry 102 done, equation 16 may even be used for estimating the strength of mixes at later ages, with a reasonable error range. More importantly, this method could serve as an easy technique for quality control in some cases. Table 22: R-square values obtained from linear regression performed on mortars with varying lime contents, at different curing ages from 7-365 days and varying cases of fixed B/ag ratios R2 values obtained from equation 16 applied to different mixes, with varying lime contents Age | B/Ag ratio 1:3 1:4 1:5 Data points (Lime content %) 10, 25, 33.3, 50, 66.7, 75 25, 33.3, 50, 66.7 33.3, 50, 66.7 7 0.99 0.99 0.99 14 0.99 0.99 1.00 28 0.99 0.99 0.96 90 0.99 0.99 0.98 180 1.00 0.98 1.00 365 0.99 0.99 0.99 Table 23: R-square values obtained from linear regression performed on mortars with varying B/Ag ratios, at different curing ages from 7-365 days and varying cases of fixed lime-cement ratios in the binder R2 values obtained from equation 16 applied to different mixes, with varying lime contents Age | Lime content % 33% 50% 67% Data points (Lime content %) 1:3, 1:4, 1:5 1:3, 1:4, 1:5, 1:6 1:3, 1:4, 1:5 7 0.99 0.99 0.99 14 0.98 0.96 0.98 28 0.87 0.95 0.92 90 0.85 0.99 0.95 180 0.95 1.00 0.99 365 0.99 0.99 1.00 4.6 Static E-modulus (Unconfined cyclic compression test) 4.6.1 Methodology Static E-modulus was measured using the recommendations of EN 12390-13 [117] adapted for mortars (Figure 53). For this test, the specimens used were cylinders with 60 mm diameter and 120 mm height. The final value of E-modulus was computed by averaging results from three specimens. To have a smooth
A multi scale approach to the study of lime-cement mortars in masonry 103 surface for even application of load during the test, epoxy resin was used to cap the specimens. A hydraulic actuator with a capacity of 25 kN was used to apply an axial pre-load of 50 N and four continuous loading and unloading cycles. Figure 53: Unconfined cyclic compression test for measurement of static E-modulus: (a) Load cycle (b) Setup of specimen The maximum load applied in the test equaled approximately one-third of the maximum compressive strength of the mortar at that age. This value of maximum uniaxial unconfined compressive strength was attained by testing three additional cylindrical specimens at each age before measuring E-modulus, using displacement control at a rate of 0.012 mm/s. This loading rate/velocity was determined such that the loading cycle would take 60 seconds, followed by a constant ramp of 20 seconds, and then the unloading ramp would take 60 seconds as well (Figure 53 a). The slope of the (stress/average strain) of each of the ascending branches was calculated. Then the average of the second, third, and fourth cycles was used to calculate the E-modulus of the specimen. The setup of the LVDTs adopted was similar to that used by Silva [128] for testing soil specimens stabilized by cement and has been shown in Figure 53 b. Three mixes were tested using this method 3C1L12S (25%), 1C1L6S (50%), 1C2L9S (66.7%) at curing ages of 7 days, 28 days, and 90 days. 4.6.2 Results 4.6.2.1 Evolution of static E-modulus with time, as a function of lime in the binder Evolution of E-modulus, measured by the cyclic compression test, for the mixes 3C1L12S (25%), 1C1L6S (50%), and 1C2L9S (66.7%) has been shown in Figure 54. The global trend observed in mechanical strength concerning the quantity of lime in the binder was found true for E-modulus as well. An increase in the quantity of lime in the binder leads to a decrease in the value of the E-modulus of the mortar, at all
A multi scale approach to the study of lime-cement mortars in masonry 104 ages. This observation was found to be consistent with the literature [271]. Naturally, more products of cement hydration are formed, when the quantity of cement in the binder is greater. One of the most abundant products of this reaction of cement hydration reaction is calcium silicate hydrate (C–S–H) crystals. Networks of C–S–H crystals form strong connections with the solid phase, binding discrete compounds into a cohesive whole and contributing to an increase in the overall strength and stiffness of hydrated cement [24]. While stiffness of all the mixes is observed to evolve with time, the mixes with greater quantities of cement in them seem to gain most of their stiffness in the first 7 days (Figure 54, Table 24). Table 24: Values of E-modulus as obtained from the cyclic compression test Mortars E-mod (GPa) Day 7 CoV (%) E-mod (GPa) Day 28 CoV (%) E-mod (GPa) Day 90 CoV (%) Increase (%) (Days 7 to 90) 1C2L9S (66.7%) 4.8 8.5 5.5 7.7 6.2 4.6 27.1 1C1L6S (50%) 9.8 7.0 10.9 2.6 11.7 10.8 19.0 3C1L12S (25%) 16.3 15.0 16.6 3.5 17.5 1.8 6.8 The increase in stiffness of the mixes between day 7 and day 90 is found to be 7%, 19%, and 27% for the mixes 3C1L12S (25%), 1C1L6S (50%), and 1C2L9S (67%) respectively. One may, therefore, conclude that the greater the amount of lime in the binder of the mix, the larger is the continued increase in stiffness, up to the age of 90 days. The most plausible explanation for this increase in stiffness over time may be attributed to stiffening induced by carbonation of the lime since hydration processes are expected to have been almost completed by 28 days of curing age [2]. While the feasibility of the general range of values obtained could be validated from literature (3 to 24 GPa), existing data classifying E-modulus of the mortar as a function of lime content in the binder could not be identified for a direct comparison [11].
A multi scale approach to the study of lime-cement mortars in masonry 105 Figure 54: Evolution of E-modulus with time, as measured by cyclic compression test (B/Ag 1:3, by volume) 4.6.2.2 Correlation of E-modulus (cyclic compression test) and compressive strength The evolution of the ratio of E-modulus to compressive strength (E/fc) with time, for cylindrical specimens, has been plotted in Figure 55. This ratio was found to lie in the range of 1300-3000, which was similar to the ratios found in the literature [243]. Only one value of the ratio for mix 1C2L9S (66.7%) at day 7 was found to reach 4650 (Figure 55). Mortars with a greater quantity of cement in them, exhibit a lower E/fc ratio. From an engineering perspective, a valid question regarding cracking and capacity to accommodate movements is whether the E/fc or the absolute value of E is more relevant. Given the low values of stresses applied to masonry, the absolute value of E seems more relevant for this purpose. It is also interesting to note that with time, the difference in values of E/fc ratios decreases, and by 90 days, the ratios for all three mortars fall in the range of 1300-2300 which means that the increase of strength with time, is much larger than an increase of stiffness with time.
A multi scale approach to the study of lime-cement mortars in masonry 106 Figure 55: Evolution of ratio of E-modulus to compressive strength (E/fc) with time 4.7 Poisson’s ratio 4.7.1 Methodology The experimental conditions used for Poisson’s ratio were similar to the one used for E-modulus (cyclic compression) (Section 4.6.1, Figure 53). The test of Poisson’s ratio was conducted on the same specimens that were used for E-modulus. However, separate tests were required to accommodate the vertical and horizontal layouts of the LVDTs. Cylindrical specimens with 60 mm diameter and 120 mm height were used for the test and the mixes tested had binder-aggregate ratio 1:3 and lime contents 25%, 50%, and 67%. The loading cycles were also the same as that used for E-mod. Since Poisson’s ratio is defined as lateral strain divided by longitudinal strain, the longitudinal strain was measured by the set-up of E-modulus with 3 LVDTs (Figure 53) and the lateral strain was measured by a metallic ring with a hinge, fastened at the center of the specimen with an LVDT attached in the horizontal direction (Figure 56). The final value of Poisson’s ratio was also computed by averaging results from three specimens for each mix.
A multi scale approach to the study of lime-cement mortars in masonry 107 Figure 56: Measurement of Poisson’s ratio - Set up of the specimen with the horizontal layout of LVDT 4.7.2 Results Poisson’s ratio was recorded for the different mixes and has been displayed along with the individual coefficient of variation of each observation, adjacent to it, in percentages (Table 25). All values were observed to be in the range of 0.13 to 0.23. The global average was found to be 0.18 with a scatter of 18%, regardless of the age or lime content in the binder. Shear modulus (G) for each of the mixes was also calculated using the value of E-modulus (E) and Poisson’s ratio (ʋ) measured (Equation 17). Table 25: Values of Poisson's ratio and shear modulus for lime-cement mortars (7, 28, and 90 days of age) Property Poisson’s ratio (COV %) Shear modulus (GPa) Mix/Age (Days) 1C2L9S (67%) 1C1L6S (50%) 3C1L12S (25%) 1C2L9S (67%) 1C1L6S (50%) 3C1L12S (25%) 7 0.20 (18.2) 0.20 (20.2) 0.18 (12.5) 2.0 4.1 6.9 28 0.19 (10.8) 0.15 (7.4) 0.21 (14.0) 2.3 4.8 6.8 90 0.16 (21.6) 0.13 (3.8) 0.23 (7.9) 2.7 5.2 7.1 It was possible to observe that the greater the quantity of lime in the mix, the lower was the value of shear modulus obtained (Table 25). It may also be noted that the value of shear modulus increased with time for almost all mixes, regardless of binder content. However, the quantity of lime in the binder of the mix was found to influence the extent of the increase in values of shear modulus recorded. For instance, between day 7 and day 90 of curing age, the shear modulus was found to increase by 24%, 21%, and 3% in the mixes 1C2L9S (67%), 1C1L6S (50%), and 3C1L12S (25%) respectively. G = E 2(1+ʋ) 17 Finally, it is worth noting that Poisson’s ratio is a parameter that is difficult to measure very accurately and tends to have high values of dispersion, due to the precision and sensitivity of the setup and
A multi scale approach to the study of lime-cement mortars in masonry 108 instrumentation involved. Coefficients of variation between 8 and 20% have been found, with higher dispersion at 7 days. At 28 and 90 days, it was consistently the mortar with the greatest quantity of cement that exhibited the largest Poisson’s value. 4.8 Fracture energy 4.8.1 Methodology Fracture energy was measured in prismatic specimens of size 40×40×160 mm (Figure 57), with a trapezoid-shaped precast notch; 5 mm in depth, equal length of non-parallel sides inclined at an angle of 15˚and the shorter parallel side being 2.5 mm in length, according to RILEM recommendation 50 FMC [381]. Displacement control was used at 0.006 mm/s with a preload of 50 N. Figure 57: Set up used to measure fracture energy If fracture energy, were to be calculated using the recommendation of RILEM 50 FMC [381], fracture energy would be dependent on the weight of the specimen as well as the portion of the force-displacement curve that is usually not recorded in experiments. The value of fracture energy Gf was therefore obtained using equation(s) 18, a method originally proposed by Elices, Guinea, and Planas [160-162] which eliminates the above-mentioned variations. Gf = Wm+Wum b(d−a) Wum = 2A du F=A(1 d02−1 du2) 18 Here Wm represents the measured energy and is the area under the experimental force-displacement curve, Wum is the unmeasured energy and represents the work that is not recorded because the test is stopped before complete failure of the specimen (Figure 58). The symbol b refers to the width of the
A multi scale approach to the study of lime-cement mortars in masonry 109 prismatic specimen, d refers to the height of the prismatic specimen and a refers to the depth of the notch in the prismatic specimen. The exact procedure for such calculations has been based on the work done by Fallahnejad et al. [382] and Garijo et al. [94]. Wum was calculated based on the ultimate displacement of the specimens in the test and a constant A. The constant A was calculated by fitting the tail end of the experimental curve, at displacement d0 corresponding to a force that is 10% of the maximum load and a displacement du corresponding to load/force of 0 N. Since, it is desirable to have the same value of ultimate displacement for all specimens, the value of ultimate displacement (du) chosen was 0.2 mm. Typically Wum is reported to be < 10% of Wm and was found to be true in this experimental campaign as well [94]. Figure 58: Illustration of calculation of fracture energy 4.8.2 Results The results of fracture energy obtained have been displayed in Table 26, for the mixes 1C2L9S (67%), 1C1L6S (50%), and 3C1L12S (25%) for 7, 28, and 90 days. It may be observed that values of fracture energy tend to increase with time for the mixes 1C2L9S (67%) and 1C1L6S (50%). This was expected since fracture energy is usually found to increase with time for concrete and cement-based mortars [164, 383-385]. However, in the mix 3C1L12S (25%), there appears to be an increase in fracture energy from 7 to 28 days, followed by a drop in 90 days. The reason for this behavior is neither known nor expected. Therefore, a conservative approach has been adopted and no conclusions regarding trends in fracture energy have been drawn.
A multi scale approach to the study of lime-cement mortars in masonry 116 9C1L30S (10% lime in the binder by volume) was used as a benchmark for normalization. It was found that at all curing ages, day 1 to day 7: every 1% increase in the quantity of lime in the binder led to a corresponding 1.3% decrease in stiffness of the mortar. It is also possible to observe that all mortars, regardless of the quantity of lime in the binder, appear to gain approximately 40% of their total stiffness in the first 24 hours, and 80% in the first 72 h hours. After the fourth day, the increase in stiffness of all the mortars was found to be 5% or lesser. Furthermore, stiffness was normalized for the corresponding values attained at day 7 for all mortars and plotted together (Figure 62 b). One may conclude that the dormant period in cement hydration that occurs in the first 3-4 hours, appears to be similar in all the mortars. It may also be observed that curves of almost all the mortars tend to overlap, indicating that the kinetics of hardening between the different mortars is similar. Concerning kinetics of stiffness evolution, it is reported in the literature [2] that cement hydration is the dominant reaction in the first few days since the casting of a lime-cement mix, and therefore, the expectation would be that the greater the quantity of cement in the binder, the faster would be the kinetics of hardening. However, a small difference may be observed in the mix 2C1L9S (33.3%), which exhibits the slowest kinetics, which is contrary to expectation, since it does not have the least amount of cement in the binder. The conclusion may be that the mix 2C1L9S (33.3%) behaves somewhat as an outlier in this test. External factors such as ambient curing temperature or minor differences in the setup may have led to slightly slower kinetics. However, since a significant difference in the environment or procedure of the test was not observed by the operator, there may be other mechanisms at play and this matter would, therefore, merit further investigation. Especially because minor differences in the rate of cement hydration, due to the presence of hydrated lime have been reported in the literature [393, 394]. Fourmentin et al. [393], state that the presence of lime accelerates the process of cement hydration, reducing its dormant period, but to a negligible extent. This phenomenon has been attributed to the high specific surface area of lime, which possibly provides a larger surface area for precipitation of the C–S–H crystals formed during cement hydration. These authors further state that this accelerating effect of lime saturates after a certain quantity. Another explanation is that lime destroys Al–O bonds networks (corresponding to oxides of Aluminium) in tricalcium aluminates, which are formed as a product of cement hydration; resulting in an increase in alkalinity of the mix, consequently accelerating the reaction [2]. It is, therefore, not possible to conclude with certainty whether the mix 2C1L9S (33.3%) is an outlier or it represents a mechanism that is not well understood as of now.
A multi scale approach to the study of lime-cement mortars in masonry 117 Figure 62: (a) E-modulus versus lime content in the binder, % by volume (b) Normalized evolution of E-modulus with time for lime-cement mixes (B/Ag 1:3, by volume) 4.11.2.2 Comparison of E-modulus values from EMM-ARM & cyclic compression test Since values of E-modulus of the mortar after 7 days of curing were obtained from 2 different methods, they were compared. A direct comparison is not possible because of differing moisture curing conditions, since, in the EMM-ARM test, the specimens were kept sealed throughout, while for the cylinders tested using the cyclic compression test, the specimens were demolded and exposed to 90% RH between 2 to 7 days. Regardless, it was expected that the values would be in the same range (Table 29, Figure 63), and was found to be true. All the mixes showed a slightly higher E-modulus value from the test of EMMARM and this may be expected due to two reasons. First, the sealed specimens in EMM-ARM, provide greater RH, promoting conditions for cement hydration [2, 24]. The second reason is that in the cyclic compression (static) test, even though loads applied are in the linear range, deformations do take place, possibly causing micro-changes/damages in the microstructure which does not happen in a dynamic test (EMM-ARM) [395]. Table 29: Comparison of E-modulus (GPa) at 7 days of age, obtained from EMM-ARM and cyclic compression test Mortars Cyclic compression| E-mod (GPa) EMM-ARM| E-mod (GPa) Diff (%) using cyclic compression test as a reference 1C2L9S (66.7%) 4.8 5.8 19.1 1C1L6S (50%) 9.8 12.0 22.0 3C1L12S (25%) 16.3 17.4 6.2
A multi scale approach to the study of lime-cement mortars in masonry 118 Figure 63: Comparison of values of E-modulus (GPa), obtained from EMM-ARM and cyclic compression test at 7 days of age To facilitate a comparison of E-modulus from the classical cyclic compression test and EMM-ARM, with similar moisture curing conditions, one mix was chosen 1C1L6S (50%). Six cylindrical specimens were kept sealed up to the time of testing (6.5 days) by the method of cyclic compression (Section 4.6.2.1) - three for compression and three for E-modulus. E-modulus obtained from EMM-ARM (average value) corresponded to 11.8 GPa and that from cyclic compression test corresponding to 10.9 GPa (obtained from an average of three specimens with a coefficient of variation of 0.2%). The difference of 7.4% in the results was considered acceptable since up to 10% variation was found common in the measurement of static Young’s modulus of mortars, as observed in this research (Table 24) as well as in literature [11]. 4.12 Final remarks This chapter focuses on the characterization and correlation of basic mechanical properties of different lime-cement mortars, including – workability, mechanical strength (compression and flexure), UPV & bulk density, stiffness measured by EMM-ARM as well as the cyclic compression method, fracture energy, Poisson’s ratio, drying shrinkage and open porosity. The following key points summarize the findings of this chapter: 1) For a target workability, the requisite water-binder ratio increases linearly with increasing lime content in the binder, as well as with decreasing B/Ag ratio of the mix. An equation has been presented to
A multi scale approach to the study of lime-cement mortars in masonry 119 estimate the requisite water binder ratio for a given mix, as a function of the composition of the mix (lime-cement ratio, B/Ag ratio), within a 10% error margin (Section 4.3). 2) The evolution of mechanical strength (compression and flexure) with time has been expressed in equations that are a function of the strengths of the corresponding mixes at 7 days (Section 4.4.2.1). 3) It was found that for B/Ag ratios 1:3, 1:4, and 1:5, every 1% increase in the quantity of lime in the binder (by volume), led to a 1.4% reduction in the mechanical strength (compression/flexure) of the mix, with respect to the benchmark mix – with 10% lime or 90% cement in the binder (Section 4.4.2.2). 4) It was found that for binder compositions with 33.3%, 50%, and 66.7% lime in the binder, every 1% decrease in B/Ag ratio (by volume), led to a 5% reduction in the mechanical strength (compression/flexure) of the mix, with respect to the benchmark mix – B/Ag ratio 1:3 (Section 4.4.2.3). 5) Mechanical strength (compression/flexure) could be expressed in an equation as a function of lime content in the binder, B/Ag ratio, and curing age (Section 4.4.2.4) 6) Bulk density was found to decrease till 28 days for almost all the mixes and then stabilize or increase slowly with time (Section 4.5.2.1). UPV values were found to decrease with increasing lime content in the binder and decreasing B/Ag ratios of lime-cement mixes. Furthermore, it was found that the square of UPV (UPV(T)2) and a product of density and compressive strength (ρ(T)1.5fc(T)0.5) varied linearly with lime content in the binder and B/Ag ratios (Section 4.5.2.2). 7) E-modulus measured by the cyclic compression test was found to be in the range of 4-18 GPa for mixes 1C2L9S (67%), 1C1L6S (50%), and 3C1L12S (25%) between 7 to 90 days. The increase in stiffness of the mixes between day 7 and day 90 was found to be 7%, 19%, and 27% for the mixes 3C1L12S (25%), 1C1L6S (50%), and 1C2L9S (67%) respectively (Section 4.6.2.1). Furthermore, for these mortars, the ratio of E-modulus to compressive strength for cylindrical specimens was found to vary from 1300 to 2300 at 90 days of age (Section 4.6.2.2). 8) For the mortars 3C1L12S (25%), 1C1L6S (50%), and 1C2L9S (67%), Poisson’s ratio was found to vary from 0.13 to 0.23 between 7 to 90 days of age, with a global average of 0.18 (18% CoV). (Section 4.7.2). 9) For the mortars 3C1L12S (25%), 1C1L6S (50%), and 1C2L9S (67%), fracture energy was found to range from 5 to 83 N/m depending on the quantity of lime in the binder (Section 4.8.2). 10) Open porosity was measured for the mixes 3C1L12S (25%), 1C1L6S (50%), and 1C2L9S (67%) at 7, 28, and 90 days of age. The value decreased with age for all mixes and increased with the amount
A multi scale approach to the study of lime-cement mortars in masonry 120 of lime content in the binder. The general range of values was found to vary only slightly between 23% and 27% (Section 4.9.2). 11) Drying shrinkage was measured for lime-cement mortars with B/Ag ratio 1:3 and lime content varying from 25% to 75% up to 90 days of curing age. The quantity of lime in the binder did not appear to have an impact on the extent of drying shrinkage. Most values ranged between 550-750 micro strains (Section 4.10.2). 12) E-modulus was measured for lime-cement mortars with B/Ag ratio 1:3 and varying lime content in the binder (25%, 33%, 50%, 67%, 75%) from time 0 to 7 days of age using the method EMM-ARM. Results from the EMM-ARM test indicate that at all curing ages, day 1 to day 7: every 1% increase in the quantity of lime in the binder led to a corresponding 1.3% decrease in stiffness of mortars. It was also observed that all mortars, regardless of the quantity of lime in the binder, appeared to gain approximately 40% of their total stiffness in the first 24 hours, and 80% in the first 72 hours. After the fourth day, the increase in stiffness of all the mortars was found to be 5% or lesser (Section 4.11.2).
A multi scale approach to the study of lime-cement mortars in masonry 121 5. Influence of lime-cement mortars on the mechanical behavior of masonry 5.1 Introduction The goal of this chapter is to describe and discuss the findings of an experimental program, aimed at better understanding how the mechanical behavior of masonry changes when it is constructed with different types of mortars: lime-cement and cement only. All masonry specimens were constructed with solid-frogged clay brick units, chosen according to the reasoning presented in Section 3.2.3, Chapter 3. In total, three mortars were used for this stage of the research; the reference mortar is made of cement with a composition of 1:5 (B/Ag, by volume), while the other two had a B/Ag ratio of 1:3 by volume and 50% and 66.7% lime in the binder, by volume. It may be noted that these mortars are different in composition, from the corresponding mortars in Chapter 4, because of the change in the particle size distribution of the aggregate. This was done to obtain the cement reference mortar (Detailed explanation in Section 3.4.1, Chapter 3). Therefore, to avoid confusion, the mortars used for research on masonry will be referred to as Ref, L50, and L67 corresponding to the compositions 1:5 (Cement: Aggregate), 1:1:6 (Cement: Lime: Aggregate), and 1:2:9 (Cement: Lime: Aggregate), all proportions by volume. Section 5.2 discusses the mechanical characteristics of the components used to construct masonry; the three selected mortars (Ref, L50 and L67) and brick. Section 5.2.1 discusses the mortars used to construct masonry and a study of their mechanical properties including mechanical strength (compression and flexure) and E-modulus, while section 5.2.2 characterizes the mechanical properties of the brick used to construct masonry, such as mechanical strength, E-modulus, IRA and water absorption. Subsequently, sections 5.3 to 5.6 study the contribution of mortar to masonry by addressing different mechanical properties; compressive strength and E-modulus, flexural strength, shear bond strength and in-plane shear wall strength of masonry. Section 5.3 addresses the compressive strength and E-modulus of masonry, as well as compares the strains at peak and ductility under compression. Section 5.4 presents the flexural strength of masonry, in two directions – parallel to the bed joints and perpendicular to the bed joints. Section 5.5 presents the shear bond strength of masonry, and the coefficients of friction. Finally section 5.6 presents results of masonry wall panels subject to combined normal and lateral inplane cyclic loading, and discusses energy dissipation, stiffness degradation and drift capacities. Masonry specimens have been referred to, based on the mortars used to construct them, i.e., Ref, L50 and L67.
A multi scale approach to the study of lime-cement mortars in masonry 122 To measure displacements in the masonry specimens, linear variable differential transformers (LVDTs) from RDP were used [396]. Those with a linear range of ±2.5 mm were of the type D6/02500ARA with a sensitivity of 375 mV/V and linearity of 0.07%. Others with a linear range of ±5 mm were of the type D6/05000A, and had a sensitivity of 700 mV/V and linearity of 0.13%. Further, all surfaces of masonry specimens that came into contact with steel plates were rectified for even distribution of the loads applied, using a bitumen based polyester resin (Sotinco 67-120 [397]). For attaching the LVDTs to the surface of masonry specimens, small metallic plates were used as an interface, along with a thermoplastic adhesive which is commonly referred to as hot glue, known for fast bonding, low costs and ease of availability [398]. 5.2 Components used for masonry construction 5.2.1 Mortars used for masonry construction The rationale behind choosing the three mortar mixes 1:0:5 or (1:5), 1:1:6 and 1:2:9, has been presented in detail in Chapter 3. This decision was based on results obtained from the 15 lime-cement mixes studied in Chapter 4, and a review of compositions that were found to be commonly used in the literature and industry [2, 95, 347, 364]. The reference mortar (1:0:5) has been denoted as ‘Ref’, while the other two mortars (1:1:6 and 1:2:9) have been denoted as ‘L50’ and ‘L67’ corresponding to 50% and 67% lime (by volume) in their respective binders. 5.2.1.1 Characterization of mechanical strength of mortars used in masonry As mentioned in Chapter 3, two different curing conditions were adopted for the mortars, and were assigned the nomenclature of standard and in situ. ‘Standard’ implies that the process respected the recommendations of the European standards, for both mixing protocols as well as curing conditions. ‘In situ’ refers to mortars that were cast in large batches for masonry construction. Samples were taken from the large batches and cured next to the masonry specimens in the same temperature and humidity conditions. (a) Standard mortars Mechanical (compressive and flexural) strength was tested on days 7, 28, and 90 and the results have been presented in Table 30 and Figure 64. By day 90, the reference mix has the highest compressive and flexural strength, closely followed by the mix L50 with a compressive strength that is 17% lower and
A multi scale approach to the study of lime-cement mortars in masonry 123 almost the same flexural strength. The mix L67 on the other hand has a compressive strength 58% lower than the reference mix and a flexural strength 47% lower than the reference. Table 30: Mechanical (compressive and flexural) strength of standard mortars Property Compressive strength (MPa) Flexural strength (MPa) Mortars/Age fc-7 (CoV %) fc-28 (CoV %) fc-90 (CoV %) ff-7 (CoV %) ff-28 (CoV %) ff-90 (CoV %) L67 (1:2:9) 2.28 (7.9) 4.35 (10.8) 4.69 (2.1) 0.85 (6.5) 1.88 (5.0) 1.88 (4.5) L50 (1:1:6) 5.80 (4.9) 9.35 (5.1) 9.28 (5.7) 1.88 (3.9) 3.93 (0.7) 3.42 (3.5) Ref (1:0:5) 7.77 (5.2) 10.27 (7.3) 11.21 (2.7) 2.46 (3.5) 3.14 (2.4) 3.53 (7.1) Figure 64: Mechanical strength of mortars (standard conditions) used for research on masonry Mortars L67 and L50 in standard conditions were tested according to EN 1015-11[16] and have compositions similar to 1C2L9S (67%) and 1C1L6S (50%) respectively, that were studied in Chapter 4. A comparison of their performance in terms of mechanical strength at days 7, 28, and 90 have been presented in the Annexes (Annex-Figure 1), since it would help understand the impact of changing the particle size distribution of the aggregates, on the mechanical strength of mortars. (b) In situ mortars
A multi scale approach to the study of lime-cement mortars in masonry 124 Mechanical (compressive and flexural) strength was tested on days 28 and 90 and the results have been presented in Table 31 and Figure 65. It may be seen that by day 90, the compressive and flexural strengths of all the mortars are slightly higher than in the standard conditions (Figure 66). The global trends, however, remain similar. Once again, on day 90, the reference mix has the highest compressive and flexural strength, closely followed by the mix L50 with a compressive strength that is 17% lower and almost the same flexural strength. And the mix L67 has a compressive strength 56% lower than the reference mix and a flexural strength 49% lower than the reference. Table 31: Mechanical (compressive and flexural) strength of in situ mortars Property Compressive strength (MPa) Flexural strength (MPa) Mortars fc-28 (CoV %) fc-90 (CoV %) ff-28 (CoV %) ff-90 (CoV %) L67 (1:2:9) 4.12 (3.4) 5.30 (5.2) 1.57 (3.5) 1.95 (2.1) L50 (1:1:6) 9.75 (7.6) 10.07 (8.5) 2.99 (9.3) 3.55 (7.8) Ref (1:0:5) 10.88 (8.9) 12.08 (6.0) 3.04 (1.0) 3.78 (11.3) Figure 65: Mechanical strength of mortars (in situ conditions) used for research on masonry
A multi scale approach to the study of lime-cement mortars in masonry 125 Figure 66: Comparison of mechanical strength of mortars, used in masonry specimens in standard (Table 30) and in situ (Table 31) conditions at 28 and 90 days of curing age 5.2.1.2 Characterization of E-modulus of mortars used in masonry E-modulus for the in situ mortars was measured using the cyclic compression test, at ages 28 and 90 days. For the sake of comparison and characterization in standard conditions, E-modulus was also measured for standard mortars at 90 days of age. All results have been shown in Table 32 and Figure 67. If experimental scatter is accounted for, the values of E-modulus between day 28 and day 90 for the in situ conditions may be considered the same. Similarly, considering the coefficient of variation at day 90, if the values of E-modulus are compared for in situ and standard conditions, the values are once again in the same range except for the reference mix. This is expectable since, in the standard conditions, the reference mix is immersed in water until testing which favors cement hydration [24], as opposed to the in situ conditions, in which the reference mix is exposed to the same atmosphere as masonry. Table 32: E-modulus for mortars (standard and in situ) Category In situ Standard Mortars E-mod - 28 (GPa) (CoV %) E-mod - 90 (GPa) (CoV %) E-mod - 90 (GPa) (CoV %) L67 (1:2:9) 6.94 (15.7) 7.90 (5.3) 8.54 (1.2) L50 (1:1:6) 16.47 (16.8) 15.97 (17.5) 14.86 (2.2) Ref (1:0:5) 16.53 (7.6) 15.21 (5.1) 19.47 (10.5)