ULTRA-LIGHTWEIGHT CEMENT COMPOSITES

AND PHOTOCATALYTIC COATING

FOR ENERGY EFFICIENT BUILDINGS

WU YUNPENG

SUPERVISOR: PROFESSOR ZHANG MIN-HONG

EXAMINERS:

ASSOCIATE PROFESSOR TAM CHAT TIM

DR. ONG GHIM PING, RAYMOND

PROFESSOR LI ZONGJIN, UNIVERSITY OF MACAU

NATIONAL UNIVERSITY OF SINGAPORE

2018
 ULTRA-LIGHTWEIGHT CEMENT COMPOSITES

AND PHOTOCATALYTIC COATING

FOR ENERGY EFFICIENT BUILDINGS

WU YUNPENG

(B.ENG., HARBIN INSTITUTE OF TECHNOLOGY, CHINA)

A THESIS SUBMITTED

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF CIVIL AND ENVIRONMENTAL

ENGINEERING

NATIONAL UNIVERSITY OF SINGAPORE

2018
 Declaration

I hereby declare that this thesis is my original work and it has been written by me in its

entirety. I have duly acknowledged all the sources of information which have been used in the

thesis.

This thesis has also not been submitted for any degree in any university previously.

_______________

WU YUNPENG

23 January 2018
 Acknowledgement

First and foremost I want to express my sincere appreciation and gratitude to

my supervisor, Professor Zhang Min-Hong for her supervision, invaluable

guidance, constructive advices, and ultimate patience throughout the entire course

of this research. It is my greatest fortune to have her as my supervisor.

I would also like to express my special gratitude to Associate Professor Yu

Liya E. for her valuable advices and suggestions on my research.

Associate Professor Tam Chat Tim and Dr. Ong Ghim Ping, Raymond,

National University of Singapore, and Professor Li Zongjin, University of Macau,

deserve special thanks as my thesis committee members for their constructive

suggestions, helpful comments, and valuable discussions.

I want to express my thanks to laboratory staffs and technicians at Structural

and Concrete Laboratory for their assistance and facilitation to experiments. My

special thanks to Mr. Ang Beng Oon, who assisted me in my experimental work.

My thanks also go out to Mr. Wong Kai Wai, Mr. Kamsan Bin Rasman, and Ms.

Li Wei from Structural and Concrete Laboratory, Department of Civil and

Environmental Engineering, Mr. Raghavan Sacadevan from Air Conditioning

Laboratory, Department of Mechanical Engineering, Dr. Chen Fangzhi, Dr.

Shabunko Veronika, and Mr. Chen Qiqiang from Solar Energy Research Institute

of Singapore for their assistance for some of the experiments.

Dr. Krishnan Padmaja and Dr. Wang Junyan have offered me a lot of

guidance and valuable suggestions who deserve my great gratitude. I also want to

express my appreciation to my group members and labmates, Dr. Le Hoang

i
Thanh Nam, Dr. Feng Huajun, and Dr. Liu Minghui, Dr. Du Hongjian, Dr. Huang

Zhenyu, and Mr. Zhang Fengling for their valuable advice and help. My thanks

also go to undergraduates Ms. Ang Xin Yi, Ms. Wang Juan, Ms. Luo Siwei, Mr.

Yi Qiang, and Ms. Wan Cheng for their assistance for some experiments.

This study would not be possible without the unconditioned love, support,

and encouragement from my family for these years.

The research scholarship provided by National University of Singapore is

greatly acknowledged. The financial support from the Republic of Singapore’s

National Research Foundation through a grant to the Berkeley Education Alliance

for Research in Singapore (BEARS) for the Singapore-Berkeley Building

Efficiency and Sustainability in the Tropics (SinBerBEST) Program is also

acknowledged.

ii
 Table of Contents

List of Tables.................................................................................................... viii

List of Figures .................................................................................................... xi

List of Abbreviations ......................................................................................... xv

List of Notations .............................................................................................. xvii

Summary .......................................................................................................... xx

Chapter 1. Introduction ..................................................................................... 1

1.1 Backgrounds ............................................................................................. 1

1.2 Objectives, scope, and organization of the thesis ....................................... 7

1.2.1 Objectives .......................................................................................... 7

1.2.2 Scope of research ............................................................................... 8

1.2.3 Thesis organization .......................................................................... 10

Chapter 2. Development of ultra-lightweight cement composites with low

thermal conductivity and high specific strength ............................................ 11

2.1 Introduction............................................................................................. 11

2.2 Experimental details ................................................................................ 15

2.2.1 Materials .......................................................................................... 15

2.2.2 Mix proportions and specimen preparation ....................................... 17

2.2.3 Experimental methods ..................................................................... 18

2.3 Results and discussion ............................................................................. 21

2.3.1 Compressive strength, flexural tensile strength, and elastic modulus 21

2.3.2 Thermal conductivity ....................................................................... 23

2.3.3 Estimation of thermal conductivity of cenospheres and ULCCs ....... 25

iii
 2.3.4 Effect of some ingredients in the ULCCs on mechanical properties and

thermal conductivity ................................................................................. 29

2.3.5 Balance of thermal conductivity and mechanical properties ............. 32

2.4 Summary and conclusions ....................................................................... 34

2.5 Tables ..................................................................................................... 37

2.6 Figures .................................................................................................... 47

Chapter 3. Effects of silica fume and fly ash on properties of plain and fiber

reinforced ultra-lightweight cement composites ............................................ 55

3.1 Introduction............................................................................................. 55

3.2 Experimental details ................................................................................ 59

3.2.1 Materials used .................................................................................. 60

3.2.2 Stability of cenospheres and steel fibers in ultra-lightweight cement

composites................................................................................................ 60

3.2.3 Effect of silica fume and fly ash on the Ca(OH)2 content and

properties of hardened ultra-lightweight cement composites ..................... 64

3.3 Results and discussion ............................................................................. 66

3.3.1 Effect of silica fume and fly ash on the rheological parameters,

flowability, and stability of fresh ultra-lightweight cement composites ..... 66

3.3.2 Effect of silica fume and fly ash on Ca(OH) 2 content of cement pastes

................................................................................................................. 70

3.3.3 Effect of fly ash on properties of hardened plain and fiber-reinforced

ultra-lightweight cement composites ......................................................... 73

3.4 Summary and conclusions ....................................................................... 76

iv
 3.5 Tables ..................................................................................................... 79

3.6 Figures .................................................................................................... 84

Chapter 4. Using photocatalytic coating to maintain solar reflectance and

lower cooling energy consumption of buildings.............................................. 90

4.1 Introduction............................................................................................. 90

4.2 Experimental details ................................................................................ 97

4.2.1 Materials, specimens, and coatings .................................................. 97

4.2.2 Deposition of black carbon on coated specimens .............................. 99

4.2.3 Exposure of the specimens with black carbon to simulated solar

irradiation ................................................................................................. 99

4.2.4 Determination of color and solar reflectance .................................. 100

4.2.5 Determination of heat transfer through the specimens .................... 101

4.3 Results and discussion ........................................................................... 104

4.3.1 Solar reflectance and gray scale L* ................................................ 104

4.3.2 Heat transfer through the specimens and surface temperatures at

different time points ............................................................................... 108

4.3.3 Correlation between solar reflectance and heat transfer through the

specimens ............................................................................................... 110

4.3.4 Practical significance ..................................................................... 111

4.4 Summary and conclusions ..................................................................... 113

4.5 Tables ................................................................................................... 116

4.6 Figures .................................................................................................. 122

v
Chapter 5. Effect of ultra-lightweight cement composite and photocatalytic

coating on heat transfer through opaque building envelope ........................ 128

5.1 Experimental studies ............................................................................. 128

5.1.1 Solar reflectance of the concrete and lightweight cement composite

with and without coating......................................................................... 129

5.1.2 Surface temperatures and heat gain through the specimen exposed to

the simulated sunlight ............................................................................. 130

5.2 Empirical equations development .......................................................... 137

5.2.1 Methodology ................................................................................. 137

5.2.2 Results and discussions .................................................................. 139

5.3 Summary and conclusions ..................................................................... 142

5.4 Tables ................................................................................................... 144

5.5 Figures .................................................................................................. 147

Chapter 6. Conclusions and recommendations ............................................ 151

6.1 Summary and conclusions ..................................................................... 151

6.1.1 Development of ultra-lightweight cement composites with low

thermal conductivity and high specific strength for energy efficient

buildings................................................................................................. 151

6.1.2 Effects of silica fume and fly ash on properties of plain and fiber

reinforced ultra-lightweight cement composites ...................................... 153

6.1.3 Using photocatalytic coating to maintain solar reflectance and lower

cooling energy consumption of buildings ................................................ 155

vi
 6.1.4 Effect of ultra-lightweight cement composite and photocatalytic

coating on heat transfer through opaque building envelope ..................... 157

6.2 Significance of this research .................................................................. 158

6.3 Recommendations for future research .................................................... 159

References ...................................................................................................... 161

Appendix ......................................................................................................... 184

Publications from the research ......................................................................... 188

Curriculum vitae .............................................................................................. 190

vii
 List of Tables

Table 2.1 Thermal conductivities of concrete and its constituent ....................... 37

Table 2.2 Thermal conductivity of concrete under different moisture conditions 38

Table 2.3 Summary of lightweight aggregate concrete in literature.................... 39

Table 2.4 Chemical and mineral compositions of cement and silica fume (% by

mass) from the manufacturers ............................................................................ 40

Table 2.5 Properties of polyethylene (PE) fibers ............................................... 40

Table 2.6 Mix Proportions and workability of ultra-lightweight cement

composites (ULCCs), cement pastes, and concrete............................................. 41

Table 2.7 Test methods and specimen information ............................................ 42

Table 2.8 Density, mechanical properties, and thermal conductivity .................. 43

Table 2.9 Comparison of experimental and calculated elastic modulus of the

ULCCs .............................................................................................................. 44

Table 2.10 Estimation of thermal conductivity of the cenospheres by Series,

Parallel, Cubic, and Hashin-Shtrikman (H-S) models ......................................... 45

Table 2.11 Comparison of experiment and estimated thermal conductivity of the

ULCCs .............................................................................................................. 45

Table 2.12 Comparison of compressive strength, density, and thermal

conductivity of ULCCs and lightweight concrete reported in literature .............. 46

Table 3.1 Test methods and specimen information ............................................ 79

Table 3.2 Chemical and mineral compositions of cement, fly ash, and silica fume

(% by mass) from the manufacturers .................................................................. 80

Table 3.3 Properties of steel fibers provided by manufacturer ........................... 80

viii
Table 3.4 Effect of silica fume and fly ash content on rheological properties of

ULCCs (Bingham model) * ................................................................................. 81

Table 3.5 Ca(OH)2 content of cement paste with different content of fly ash and

silica fume by thermogravimetric analysis ......................................................... 82

Table 3.6 Density, mechanical properties, and thermal conductivity of various

ULCCs .............................................................................................................. 82

Table 3.7 Mix Proportions and flowability of ULCCs * ...................................... 83

Table 4.1 Solar reflectance of common roof surfaces ...................................... 116

Table 4.2 Effect of soiling on solar reflectance and thermal performance of

various coatings ............................................................................................... 117

Table 4.3 Information of mortar specimens ..................................................... 118

Table 4.4 Characteristics of photocatalytic TiO2 .............................................. 118

Table 4.5 Test methods and specimen information .......................................... 118

Table 4.6 Experimental conditions for determining the heat transfer through

specimens exposed to a sun simulator .............................................................. 119

Table 4.7 Solar reflectance and color of specimens with various coatings at

various time points .......................................................................................... 120

Table 4.8 Heat gain in the chamber and surface temperatures of the specimens

with various coatings exposed to the sun simulator for 9 hrs ............................ 121

Table 5.1 Density, thermal conductivity, and solar reflectance of the concrete and

ULCC with and without photocatalytic coating ................................................ 144

Table 5.2 Heat gain and surface temperatures of the concrete and ULCC with and

without coating exposed to simulated sunlight ................................................. 145

ix
Table 5.3 Experimental results on material properties and heat gain ................ 146

Table 5.4 The subset regression analysis of the heat gain (Qt) on oven dry density

( ), thermal conductivity (λ), and solar reflectance (β) ............................... 146

Table 5.5 Coefficients of various models ........................................................ 146

x
 List of Figures

Figure 2.1 Particle size distributions of the two types of cenospheres used ........ 47

Figure 2.2 Image of cenospheres (Exlite) .......................................................... 47

Figure 2.3 Photograph of the ASTM C1609-07 (2007) test setup ...................... 48

Figure 2.4 Average curve of the load-deflection curves by using OriginPro ...... 48

Figure 2.5 Schematic experimental setup for determining thermal conductivities

of by heat flow meter (a), guarded hot plate (b), and hot box (c) ........................ 49

Figure 2.6 28-day compressive strength vs wet density after demoulding .......... 50

Figure 2.7 (a) Flexural tensile strength vs compressive strength; (b) Flexural

tensile strength vs wet density after demoulding ................................................ 50

Figure 2.8 (a) Elastic modulus vs compressive strength, (b) Elastic modulus vs

wet density after demoulding ............................................................................. 51

Figure 2.9 Difference between experimental and calculated elastic modulus

versus (a) wet density at 1d and (b) 28d compressive strength ........................... 51

Figure 2.10 Thermal conductivity vs oven-dry density ...................................... 52

Figure 2.11 Loose packed particulate sample for thermal conductivity test (not to

scale) ................................................................................................................. 52

Figure 2.12 Scanning electron microscope image of a ULCC with cenospheres 53

Figure 2.13 Flexural stress-deflection curves of fiber-reinforced ULCCs (ULCC

7 and 8) in comparison to that without fibers (ULCC 2) ..................................... 53

Figure 2.14 (a) Density, compressive strength, and thermal conductivity of the

ULCCs in comparison to concrete with various lightweight aggregates reported in

literature, (b) Merit number of various lightweight concretes mentioned in (a) ... 54
xi
Figure 3.1 Particle size distributions of the cement, fly ash, and cenospheres used

.......................................................................................................................... 84

Figure 3.2 Sketch of samples used to investigate the stability of (a) cenospheres

in plain ULCCs and (b) steel fibers in fiber reinforced ULCCs (not to scale) ..... 85

Figure 3.3 Shear stress versus shear rate of ULCC with different silica fume (a)

and fly ash (b) content ....................................................................................... 86

Figure 3.4 Effect of silica fume on yield stress and plastic viscosity of ULCCs . 86

Figure 3.5 Comparison of the densities of the samples from the top and bottom of

(a) plain and (b) fiber reinforced ULCC specimens with the density of the

specimen before cutting. .................................................................................... 87

Figure 3.6 Binary images of cross sections of fiber reinforced prism (100 × 100-

mm) without incorporation of silica fume (black dot-steel fibers; blue-ULCC

matrices) ............................................................................................................ 87

Figure 3.7 Ca(OH)2 (CH) content of cement pastes with different contents of

silica fume and fly ash ....................................................................................... 88

Figure 3.8 Compressive strength of ULCCs with fly ash and steel fibers at

different ages ..................................................................................................... 88

Figure 3.9 Flexural performance of fiber reinforced ULCCs with different fly ash

content............................................................................................................... 89

Figure 4.1 XRD pattern of the photocatalytic TiO2 used in this study .............. 122

Figure 4.2 (a) Locations of heat flux sensors, thermocouples, and testing points

for solar reflectance and surface temperature; (b) Schematic experimental setup

for determining the heat transfer and surface temperature of the specimen exposed

xii
to a sun simulator (not to scale); (c) Image of the testing chamber and the

locations of some sensors................................................................................. 123

Figure 4.3 Changes of (a) solar reflectance and (b) gray scale L* of the

specimens with various coatings and black carbon deposition exposed to

simulated solar irradiation in accelerated weathering chamber ......................... 124

Figure 4.4 Images of specimens at different time points .................................. 124

Figure 4.5 Correlation between solar reflectance and gray scale L*................. 125

Figure 4.6 Heat flux (a) and surface temperatures (b) of the specimens with

various coatings exposed to a sun simulator for 9 hrs ....................................... 125

Figure 4.7 Effect of black carbon deposition on the heat flux (a) and surface

temperatures of the specimens with TSC (b), WSC (c), and PWSC (d) exposed to

the sun simulator for 9 hrs ............................................................................... 126

Figure 4.8 Effect of black carbon removal on the heat flux (a) and surface

temperature (b) of the PWSC specimen before and after 300 hrs exposure to

simulated solar irradiation................................................................................ 126

Figure 4.9 Comparison of the heat flux (a) and surface temperature (b) of the

PWSC specimen before black carbon deposition and after black carbon removal

........................................................................................................................ 127

Figure 4.10 Relationship between solar reflectance and the heat gain (a) and

surface temperatures of the specimens (b) after 9 hrs of exposure to the sun

simulator ......................................................................................................... 127

Figure 5.1 Image of cenospheres, cement, silica fume, and sand and their solar

reflectance (SR) values .................................................................................... 147

xiii
Figure 5.2 Sketch of heat transfer through opaque building envelope exposed to

simulated sunlight (not to scale) ....................................................................... 147

Figure 5.3 Heat flux (a) and surface temperatures (b) of the concrete and ULCC

specimens ........................................................................................................ 148

Figure 5.4 Heat flux (a) and surface temperatures (b) of the ULCC specimen with

and without coating, under the exposure to the simulated sunlight ................... 148

Figure 5.5 Heat flux (a) and surface temperatures (b) of the ULCC specimen with

photocatalytic coating and concrete, under exposure to the simulated sunlight . 148

Figure 5.6 Normal plot of residuals for equations contain solar reflectance and

thermal conductivity (a) or oven dry density (b)............................................... 149

Figure 5.7 Prediction error of heat gain with equations contain solar reflectance

(SR) and oven dry density or thermal conductivity (TC) .................................. 149

Figure 5.8 Predicted heat gain with equations contain solar reflectance and oven

dry density or thermal conductivity .................................................................. 150

xiv
 List of Abbreviations

BC Black carbon

BET Brunauer-Emmett-Teller

CH, Ca(OH)2 Calcium hydroxide

cm Cementitious materials

CP Cement paste

EPR Electron paramagnetic resonance

FA Fly ash

FTIR Fourier transform infrared spectroscopy

H-S Hashin-Shtrikman

PE Polyethylene

PVA Polyvinyl alcohol

PWSC White silicate coating incorporating photocatalyst

RH Relative humidity

SF Silica fume

SP Superplasticizer

SR Solar reflectance

SRA Shrinkage reducing admixture

ST Steel fibers

TC Thermal conductivity

TSC Transparent silicate coating

ULCC Ultra-lightweight cement composite

xv
VMA Viscosity modifying agent

WSC White silicate coating

xvi
 List of Notations

A Area, m2

Adj-R2 Adjusted coefficient of determination

Cp Specific heat capacity, J/kg·oC

Cs Stokes drag coefficient

E Elastic modulus, GPa

fc Compressive strength, MPa

g Acceleration due to gravity, m/s2

I Solar irradiance at the exposure surface of the specimens, W/m2

k Total variables considered plus the constant

L* Gray scale

M Merit number, equals to compressive strength / (thermal

conductivity × density), 104·K·m3/W·s2

Mean square error of model with all variables

Mean square error of model with p variables

n Number of the experimental sets

p Number of variables and constant in the model

qt Heat flux through the specimens, W/m2

Q Heat reached the exposure surface (or energy input), J/m2

Ql Heat loss from exposure surface due to convection and

radiation, J/m2

Qr Heat reflected by exposure surface, J/m2

xvii
Qs Heat stored in the specimens, J/m2

Qt Heat gain through the specimens, J/m2

R Radius of the spherical particle, m

R2 Coefficient of determination

t Exposure time, s

Tt Average temperature of the specimen after exposed to the

simulated sunlight, °C

T0 Initial temperature of the specimen, °C

U Velocity of settlement in the fluid, m/s

V Volume of specimen, m3

Volume ratio of cenospheres in ULCC

Volume ratio of cement paste in ULCC

w/b Water to binder ratio

w/c Water to cement ratio

Heat gain value of individual sample determined by experiment,

kJ/m2

Predicted heat gain value of individual sample, kJ/m2

Mean of the heat gain of experimental data, kJ/m2

Yg Dimensionless parameter

Resistance to soiling

β Solar reflectance of exposure surface

Aged solar reflectance

Solar reflectance value before soiling

xviii
 Solar reflectance of an thick opaque soiled surface

Shear rate, s-1

Temperature gradient, K/m

η Plastic viscosity, Pa·s

λ Thermal conductivity, W/m·K

Thermal conductivity of ULCCs, W/m·K

Thermal conductivity of cement pastes, W/m·K

ceno Thermal conductivity of cenospheres, W/m·K

ρ Density, kg/m3

ρdry Oven dry density, kg/m3

Density of the fluid, kg/m3

Density of the particles, kg/m3

τ Shear stress, Pa

τ0 Yield stress, Pa

xix
 Summary

Cooling energy consumption of buildings correlates to heat transfer through

opaque building envelope which can be reduced by decreased thermal

conductivity and increased solar reflectance of building envelope. This thesis

focuses on the development and evaluation of high performance materials with

low thermal conductivity and high solar reflectance for energy efficient buildings.

Ultra-lightweight cement composites (ULCCs) with low thermal conductivity

but high specific strength for structural applications were developed by

incorporation of hollow cenospheres. The ULCCs had 1-day densities ranged

from 1155 to 1470 kg/m3 and 28-day compressive strengths ranged from 33.0 to

69.4 MPa. With similar 28-day compressive strength, the thermal conductivity of

the ULCC was 80% lower than that of the concrete.

The yield stress and plastic viscosity of fresh ULCCs generally increased

with the increase of silica fume content from 0 to 12% by the mass of

cementitious materials, whereas the incorporation of fly ash by 26% and 40% did

not affect these two rheological parameters of the fresh ULCCs. The segregation

of cenospheres was not observed in the ULCCs with and without silica fume,

whereas at least 4% silica fume is essential for the stability of steel fibers in the

ULCCs.

Mineral admixtures such as fly ash can be used to reduce Portland cement

content in the ULCCs as the strength of ULCCs is mainly governed by

cenospheres. With a reasonable cement content of about 460 kg/m3, the ULCC

with 26% fly ash and 8% silica fume achieved 28-day compressive strength of 40
xx
MPa, density of < 1250 kg/m3, and post cracking performance comparable to that

of the control mixture without fly ash. Experimental results from

thermogravimetric analysis also revealed reduced pozzolanic reaction of fly ash

due to the presence of silica fume, and indicated that fly ash content beyond

certain extent may not lead to significant increase in the pozzolanic reaction and

improvement in mechanical properties.

The ability of a photocatalytic coating to maintain the solar reflectance of

opaque building envelope by removing deposited pollutants from opaque surface

was experimentally demonstrated. The experimental results indicated that

photocatalytic TiO2 was able to remove black carbon and restore the solar

reflectance of the mortar specimen after 300 hrs of exposure to simulated solar

irradiation. The inner surface temperature and heat gain of the specimen with

photocatalytic coating after the exposure were 7.2 °C and 53% lower than those

contaminated specimen with black carbon deposition.

Compared with concrete without coating, the ULCC with lower thermal

conductivity and photocatalytic coating with higher solar reflectance reduced the

heat gain by 33% and 54%, respectively, while their combination reduced the heat

gain by 69% when exposed to simulated sunlight for 9 hours. With reduced heat

gain through opaque building envelope, energy consumed by cooling system to

keep constant indoor temperature will decrease. Therefore, the cooling energy

consumption of buildings can be reduced significantly by the use of the ULCC,

photocatalytic coating, or a combination of them. Linear empirical equations to

xxi
predict the heat gain through homogeneous opaque building envelope with

thermal conductivity, solar reflectance, and oven dry density were developed.

In summary, this study has developed ultra-lightweight cement composites

with low thermal conductivity and high specific strength and investigated

photocatalytic coating with high solar reflectance and self-cleaning capacity.

Their individual and combined effects on heat transfer through opaque building

envelope are quantitatively demonstrated under controlled experimental condition.

xxii
 Chapter 1. Introduction

1.1 Backgrounds

Energy consumed by building sectors mainly occurs in five stages: (1)

manufacture of building materials, (2) transportation of building materials to

construction sites, (3) construction process, (4) during operation, and (5)

demolition process (Jones, 1998). Among these five stages, energy consumed

during the operation period by lighting, cooling, and heating etc is more than 70%

(United Nations Environment Programme, 2007). On the other hand, existing

buildings consume more than 40% of total primary energy (from natural sources,

such as energy from coal and natural gas) in the world and are responsible for 30%

of greenhouse gases (such as CO2 and NOx) emission (International Energy

Agency, 2008). Therefore, reducing energy consumption during operation period

is one of the most desirable approaches to save energy.

The concept of “Energy Efficient Building” has become more attractive in

recent years due to the urgent desire to reduce the greenhouse gases emission and

worldwide reliance on fossil fuel. An energy efficient building is defined by

achieving satisfactory internal environment with minimum energy consumption

(Chartered institution of building services engineers, 2004). One well-known

concept of energy efficient building is the “Zero Energy Building” which is

characterized by generation of sufficient renewable energy to fulfill its own

consumption requirements (Department of Energy, 2017a). The zero energy

consumption is mainly achieved by generation of renewable energy through, e.g.

1
photovoltaic panels and solar water heating systems, and by reduction of building

energy consumption through various technologies such as installing sun shading

devices, natural ventilation, and lighting systems (Behind the success of zero

energy building, 2017). Recently, Building Construction Authority of Singapore

has retrofitted an existing three-story building to a zero energy building. It is

revealed that this building generates more renewable energy than it needs (Behind

the success of zero energy building, 2017) which is equivalent to annual energy

cost saving of $60,000 compared to an ordinary office building in Singapore

(What is zero energy building?, 2017).

In tropical countries with high temperatures and relative humidity and in the

summer season of temperate countries, a large amount of building energy

consumption is due to cooling the building interior. In Singapore (a tropical

country), for example, approximately 40-50% of the building electricity

consumption is due to air-conditioning (Kua and Wong, 2012). A significant part

of energy consumed by air conditioning is caused by heat gain through the

building envelope. A study conducted on residential buildings in Singapore by

Chua and Chou (2010) reveals the energy consumed by air conditioning is

approximately linear to the heat gain through building envelope. Although

substantial amount of the heat gain is through windows, the heat gain through

opaque building envelope, e.g. walls and roofs, is also responsible for a

significant amount of cooling energy consumption of buildings in tropical regions

(Chua and Chou, 2010) especially for buildings of masonry and reinforced

concrete structures with low window-to-wall ratios.

2
 Various methods have been proposed to reduce the heat gain through opaque

building envelope such as reducing thermal conductivity, increasing solar

reflectance, and increasing thermal mass of opaque building envelope. Among

them, increasing thermal mass is only effective to reduce heat gain through

opaque building envelope in areas with high diurnal temperature fluctuation (Al-

Homoud, 2005; Balaras, 1996) whereas reducing thermal conductivity and

increasing solar reflectance are more effective and economical for tropical regions

with low diurnal and seasonally temperature fluctuations (Al-Homoud, 2005;

Simpson and Mcpherson, 1997). Therefore, this study focuses reducing the heat

gain through opaque building envelopes by reducing thermal conductivity and

increasing solar reflectance.

As one of the most widely used opaque building envelope materials, concrete

has a thermal conductivity of 1.5 - 3.0 W/m·K (Mindess et al., 2003). Lightweight

aggregates with low thermal conductivities such as perlite and expanded

polystyrene beads can be incorporated in concretes to reduce densities and

corresponding thermal conductivities (Schackow et al., 2014; Sengul et al., 2011;

Yu et al., 2015). Although incorporating lightweight aggregates in concrete

reduces its thermal conductivity and increases its insulation capacity, the

mechanical properties, such as compressive strength and elastic modulus, are

generally compromised. Many lightweight concretes reported in literature are not

able to meet the modern-day requirement for structural applications due to low

compressive strengths (Wu et al., 2015), and are purely used for partitioning and

insulation purposes. Therefore, one objective of this study is to develop ultra-

3
lightweight cement composites (ULCCs) with low thermal conductivities and

high specific strengths (compressive strength / density). To achieve this,

cenospheres were selected as micro lightweight aggregates in the ULCCs.

Cenospheres are byproduct from coal combustion at thermal power plants

(Montgomery and Diamond, 1984; Wandell, 1996). The process of burning coal

in thermal power plants produces fly ash which includes mainly solid particles

and a small amount of hollow particles (cenospheres). Cenospheres are formed

during the coal burning process when the gas evolved is entrapped in a viscous

molten glass matrix (Clayton and Back, 1989). They generally have hollow

interior covered by a thin shell with thicknesses in the order of 5-10% of its

diameter (Cardoso et al., 2002).

Typical ULCCs include binders (e.g. cement and mineral admixtures),

cenospheres, fibers, water, and chemical admixtures (Liu et al., 2016; Wang et al.,

2013; Wu et al., 2015). The performance of fiber reinforced ULCCs is dependent

on the dispersion and stability of the cenospheres and fibers in the composites,

and would be reduced significantly with the segregation of the cenospheres and

fibers from the composites. The incorporation of silica fume may improve the

stability of cenospheres and steel fibers by increased yield stress and plastic

viscosity but may reduce the flowability of cement composites (Nehdi et al., 1998;

Park et al., 2005). Another concern is that Portland cement content in the ULCCs

is much higher (600 - 850 kg/m3) than that in conventional concrete of

comparable compressive strength due to the elimination of coarse aggregate and

smaller maximum aggregate size with cenospheres acting as micro-aggregate. As

4
the compressive strength of ULCCs is governed primarily by the strength of the

hollow cenospheres, fly ash or slag may be used to partially replace Portland

cement in the ULCCs with economical, environmental, and technical benefits.

However, the silica fume incorporated for stability purpose may consume

significant amount of calcium hydroxide (Ca(OH)2) by pozzolanic reaction and

limit the amount of fly ash which can participate the pozzolanic reaction.

Therefore, the contents of silica fume and fly ash need to be optimized by

considering not only stability and flowability of fresh ULCCs but also pozzolanic

reactions and strength development of hardened ULCCs.

Another approach to reduce the heat gain through opaque building envelope

is to increase the solar reflectance of building surfaces. Conventional concrete has

a solar reflectance of about 0.4 - 0.5 (Levinson and Akbari, 2002; Marceau and

Vangeem, 2008), which can be increased to 0.8 with various coatings (Levinson

et al., 2010; Song et al., 2014; Synnefa and Santamouris, 2013). However, the

solar reflectance of building envelope with or without coating is significantly

reduced due to soiling by pollutants (Cheng et al., 2012; Levinson et al., 2005;

Synnefa and Santamouris, 2013; Takebayashi et al., 2016) such as black carbon

(or soot). Such reduction in the solar reflectance due to soiling increases the heat

gain through opaque building envelope and increases the cooling energy

consumption of buildings (Paolini et al., 2014). To mitigate such negative impacts

of soiling, coatings with high resistance to soiling have been developed based on

various principles such as increasing hydrophobicity and incorporating

photocatalysts (Krishnan, 2015; Krishnan et al., 2013b; Sleiman et al., 2014b).

5
Although high hydrophobicity may increase the resistance to the deposition of

pollutants on building surfaces, substantial reduction of solar reflectance is still

observed with time (Sleiman et al., 2014b). In contrast, some experimental results

(Krishnan, 2015; Krishnan et al., 2013b) indicate that coatings with photocatalyst

can maintain the solar reflectance of building surfaces by removing deposited

pollutants such as black carbon and organic matters.

Furthermore, effects of the ULCC with low thermal conductivity and

photocatalytic coating with high solar reflectance on the heat transfer through

building envelope need to be studied systematically to determine their individual

and combined contributions to thermal performance of buildings.

Based on the introduction above, this study focuses on reducing heat transfer

through opaque building envelope by reducing thermal conductivity and

increasing solar reflectance. Considering the compromise between thermal

conductivity and compressive strength, ULCCs with high specific strength and

low thermal conductivity will be developed in this study. However, the

cenospheres and steel fibers tends to segregate in fresh ULCCs due to the

significant differences in their densities versus those of cement paste matrices.

Silica fume may be used to increase the stability by increasing yield stress and

plastic viscosity of cement pastes. In addition, the cement content of ULCCs is

usually high, leading to high risk of thermal cracking, which may be solved by

partial replacement of cement with fly ash. The solar reflectance of building

surfaces decreases with time due to soiling and photocatalytic coating may be

used to remove black carbon and maintain the solar reflectance. Therefore, it is

6
expected that ULCCs with low thermal conductivity and photocatalytic coating

with high solar reflectance and self-cleaning ability can effectively reduce the

heat transfer through opaque building envelope and thus cooling energy

consumption of buildings in tropical countries.

1.2 Objectives, scope, and organization of the thesis

1.2.1 Objectives

(a) Develop ultra-lightweight cement composites and investigate their

thermal conductivities in comparison to that of cement pastes with

comparable water/binder ratios (w/b) and a concrete with comparable 28-

day compressive strength.

(b) Investigate the effect of silica fume on the stability of cenospheres and

steel fibers and flowability of fresh ULCCs.

(c) Investigate the amount of fly ash which can be used to replace Portland

cement in the ULCCs with silica fume from the perspectives of

pozzolanic reactions and strength development.

(d) Evaluate the ability of a photocatalytic coating with titanium dioxide

(TiO2) to remove black carbon from opaque mortar surface, maintain

solar reflectance, lower heat gain through opaque building envelope, and

reduce cooling energy consumption.

(e) Investigate the individual and combined effects of the ULCC with low

thermal conductivity and photocatalytic coating with high solar

reflectance on the heat transfer through opaque building envelope and

7
 develop empirical equations to predict the heat transfer through

homogeneous opaque building envelope with different thermal

conductivities and solar reflectance.

1.2.2 Scope of research

This thesis focuses on the development and evaluation of ULCCs with low

thermal conductivity and photocatalytic coating with high solar reflectance and

self-cleaning ability for opaque building envelope. Density, thermal conductivity,

and mechanical properties were used to evaluate the performance of plain and

fiber reinforced ULCCs, while solar reflectance and self-cleaning efficiency were

used to evaluate the performance of photocatalytic coatings. The individual and

combined effects of ULCC and photocatalytic coating on heat transfer through

opaque building envelopes were also studied under controlled measurement

conditions. The specific scopes for each objective are as follows:

For objective (a), ULCCs with densities of 1150 - 1470 kg/m3 after

demoulding and 28-day cube compressive strengths of 33.0 - 69.4 MPa were

developed. Experiments on the thermal conductivity and properties such as

density, compressive strength, flexural tensile strength, and elastic modulus of the

materials were conducted. Effects of the cenospheres, chemical admixtures, and

fibers on the performance of the ULCCs are discussed. The experimental results

of the thermal conductivities of the ULCCs are compared with those estimated

from composite models and an empirical model.

For objective (b), effect of silica fume content from 0% to 12% on the

rheological parameters and flowability of the ULCC matrices and their influence
8
on the stability of the cenospheres and steel fibers were studied. The stability of

the cenospheres and steel fibers was evaluated by image analysis and densities of

specimens from the top and bottom of the hardened ULCCs.

For objective (c), Ca(OH)2 content of cement pastes with silica fume content

of 0% and 8% and fly ash content of 0%, 26%, and 40% and mechanical

properties of the ULCCs with 8% silica fume content but various fly ash contents

(0%, 26%, and 40%) were determined at various ages and discussed.

For objective (d), Portland cement mortar specimens which represent typical

concrete wall and roof surfaces were coated with transparent silicate coating

(TSC), white silicate coating (WSC), and white silicate coating incorporating

photocatalyst (PWSC). The solar reflectance, color, heat gain, and surface

temperatures of the specimens exposed to a sun simulator for 9 hrs were

determined at three different time points: (1) before black carbon deposition; (2)

after black carbon deposition (8 μg/cm2); and (3) after 300 hrs exposure to

simulated solar irradiation in an accelerated weathering chamber. The correlation

between the solar reflectance and heat transfer through the specimens in a

controlled experimental environment was investigated and discussed. Effect of

black carbon on the heat transfer through opaque building envelope is also

discussed.

For objective (e), a normal weight concrete and a ULCC with comparable

compressive strength and solar reflectance but significantly different thermal

conductivities were selected to determine the effect of thermal conductivity on the

heat transfer through panel specimens. The photocatalytic coating with solar

9
reflectance of approximately 0.8 was applied on the ULCC specimen to increase

its solar reflectance. The heat gain and surface temperatures of the specimens of

the concrete and ULCC with and without coating exposed to simulated sunlight

for 9 hrs were compared and analyzed. Empirical relationships between the heat

gain through the specimens and materials properties of oven dry density, thermal

conductivity, and solar reflectance were established based on statistical methods.

1.2.3 Thesis organization

Chapter 1 introduces the backgrounds, objectives, scope, and organization of

this thesis. Relevant literature review and motivations for each objective are

presented at the beginning of each chapter.

Chapter 2 presents the experimental study on the development of ULCCs

with low thermal conductivity and high specific strength for structural

applications.

Chapter 3 investigates the effect of silica fume and fly ash on properties of

plain and fiber reinforced ULCCs.

Chapter 4 experimentally demonstrates the ability of photocatalytic coating

with TiO2 to maintain solar reflectance of mortar specimens and lower cooling

energy consumption of buildings.

Chapter 5 investigates the individual and combined effects of a ULCC with

low thermal conductivity and a photocatalytic coating with high solar reflectance

on the heat transfer through panel specimens.

Chapter 6 summarizes the conclusions of this study and identifies the gaps

for future research.

10
 Chapter 2. Development of ultra-lightweight cement composites

with low thermal conductivity and high specific strength

2.1 Introduction

Thermal conductivity is a measure of the rate at which heat is transferred

through a material, and it is defined as the ratio of heat flux (rate of heat transfer

through a specimen per unit time per unit area) to temperature gradient (Equation

2.1).

λ = qt (2.1)

Where, λ - thermal conductivity of the specimens, W/m·K;

qt - heat flux through the specimens, W/m2 ;

- outer and inner surface temperature gradient, K/m.

Reducing thermal conductivity of building envelope can be achieved by

incorporating low thermal conductive insulation materials such as mineral wool

and polyurethane foam in walls and roofs of buildings (Alvarado et al., 2009;

Cheung et al., 2005; Han et al., 2009; Huang et al., 2013). The most widely used

method to incorporate insulation materials is by adding insulation layer into

building envelope. For example, Huang et al. (2013) investigated cooling energy

consumption of buildings with and without insulation layers and found that the

installation of a 50-mm extruded polystyrene board (0.035 W/m·K) in walls

reduced annual cooling energy consumption by more than 80% in both Hong

Kong and Singapore. In another study based on the climate condition of Hong

Kong (Cheung et al., 2005), by installing 100-mm extruded polystyrene plate onto
11
the inside of the a concrete wall, an annual saving of cooling energy by about 20%

could be achieved.

However, the application of insulation materials layer is limited mainly due

to low fire resistance of organic insulation materials such as extruded polystyrene

and polyurethane foam and concerns about human health (e.g. carcinogenicity) of

inorganic insulation materials such as fiber glass and rock wool (Papadopoulos,

2005; Schiavoni et al., 2016). In addition, low resistance to water absorption of

some insulation materials such as perlite and mineral wool boards leads to

significant increase in thermal conductivity during operation (Whetsell et al.,

2015). To ensure satisfactory insulation, a building envelope often includes

multiple layers, e.g. a structural layer, an insulation layer, a fire protection layer,

and a waterproofing layer. This complicates the construction process and

increases the thickness of building envelope and duration and cost for the

installation (Jelle, 2011).

To simplify the construction process and reduce the duration and cost of

installation, it is desirable to incorporate insulation materials into the structural

layer and achieve insulation performance with a single layer building envelope.

Lightweight concrete may be used to achieve thermal insulation, sufficient

structural capability, and durability with a single layer.

As one of the most widely used opaque building envelope materials, concrete

has a thermal conductivity of 1.5 - 3.0 W/m·K (Mindess et al., 2003). Thermal

conductivity of concrete is affected primarily by the thermal conductivity of the

raw materials used, mixture proportion, void content, and moisture condition of

12
the concrete (ACI-122R, 2002). Decreasing thermal conductivity of mixing

materials in concrete and increasing the proportion of materials with low thermal

conductivity will reduce the thermal conductivity of concrete. The thermal

conductivity of most natural rock aggregates is much higher than that of the

cement pastes, and it varies depending on the rock type (Table 2.1) (Mehta and

Monteiro, 2005; Mindess et al., 2003). Studies demonstrate that the thermal

conductivity of concrete increased with the increase of aggregate volume ratio

(Kim, 2003). In addition, rocks with crystalline structure show higher thermal

conductivity than amorphous and vitreous rocks of the same composition

(Harmathy, 1970; Zoldners., 1971). The thermal conductivity of cement pastes

increases with the reduction of water to cement ratio (w/c) and capillary porosity

(Mindess et al., 2003). Moisture content is another major factor that influences the

thermal conductivity of concrete (Table 2.2 (Scanlon and McDonald, 1994))

because the thermal conductivity of water (0.5 W/m·K) is higher than that of air

(0.03 W/m·K), as shown in Table 2.1. Valore (1980) suggested a relation of 6%

increase of in thermal conductivity with 1% increase of in free or evaporable

moisture by weight in relation to oven dry density. The thermal conductivity of

concrete is less affected by the measurement temperature and curing (Kim, 2003).

Due to the lower thermal conductivity of air, the low thermal conductivity of

concrete can be achieved by incorporation of air voids in concrete. There are

several ways to introduce air voids into concrete: (1) voids in aggregates of

various sizes, (2) voids in cement paste, (3) elimination of sand in the concrete

mixture leading to voids in between the coarse aggregate, and (4) combinations of

13
the above. Among them, introducing air voids by porous lightweight aggregate is

typically used to achieve higher specific strength and low permeability. In recent

years, lightweight concrete mixtures with low thermal conductivities have been

developed using lightweight aggregates. Examples of the lightweight aggregates

include expanded clay (FIP, 1983), expanded shale (FIP, 1983), foamed slag (FIP,

1983), pumice ( opçu and Uygunoğlu, 2007; Uysal et al., 2004), perlite (Gül et

al., 2007; Tandiroglu, 2010), cenospheres (Blanco et al., 2000; Huang et al., 2013),

polyurethane foam (Mounanga et al., 2008), diatomite ( opçu and Uygunoğlu,

2007), expanded glass (Yu et al., 2013), aerogel (Gao et al., 2014), and high-

impact polystyrene (Wang and Meyer, 2012). Some lightweight concrete in

literature with various lightweight aggregates are summarized in Table 2.3.

Although introducing voids in concrete reduces its thermal conductivity and

increases its insulation capacity, the mechanical properties, such as compressive

strength and elastic modulus, are generally compromised (ACI-122R, 2002), as

shown in Table 2.3. The lightweight concrete with traditional lightweight

aggregates such as expanded clay and shale are not included because information

on their corresponding strengths is not available. It is observed that many of these

lightweight concretes in Table 2.3 are not able to meet the modern day

requirement for structural applications due to low compressive strengths. These

materials can only be used in non-structural applications. For example, Sengul et

al. (2011) developed lightweight concrete with a thermal conductivity lower than

0.6 W/m·K by using expanded perlite. Yu et al. (2015) developed ultra-

lightweight concrete with a thermal conductivity approximately 0.1 - 0.2 W/m·K

14
by incorporating synthesized porous aggregate from recycled glass. However, the

28 days compressive strength of these lightweight concrete were below 20 MPa.

Construction materials with better performance in terms of both strength and

thermal conductivity is desired in practice.

Therefore, ultra-lightweight cement composites (ULCCs) with low thermal

conductivity and high specific strength have been developed in this study. The

thermal and mechanical properties of various ULCCs in comparison to those of

cement pastes with comparable w/b and a concrete with comparable 28-day

compressive strength are discussed in this chapter.

2.2 Experimental details

2.2.1 Materials

Ordinary Portland cement and undensified silica fume 1 (SF) were used for

most mixtures including ULCCs and cement pastes except for the concrete in

which only the Portland cement was used. The chemical compositions of the

cement and silica fume are summarized in Table 2.4. According to the report

provided by manufacturer, the cement used in this study meets the specification of

EN CEM I 52.5 N in EN 197-1:2014 and ASTM type I in ASTM C150: 2017.

Two types of cenospheres 2 (QK300 and Exlite) with average particle

densities of approximately 908 and 615 kg/m3, respectively, were used in ULCCs

as micro-aggregates. Particle size distributions of the cenospheres determined by

1
Elkem Microsilica Grade 920E, Elkem Materials, Norway
2
Sun Microspheres, China
15
a particle size analyzer 3 are given in Figure 2.1, and the results show that most of

the particles had sizes from 10 to 300 μm. The image of cenospheres, shown in

Figure 2.2, revealed the cenospheres have hollow interior covered by a thin shell.

According to the manufacturer, the cenospheres are extracted from fly ash by

floatation method during which particles with densities lower than water were

collected. A previous study (Wang et al., 2012b) showed that the cenospheres had

low content of CaO but high combined content (approximately 90%) of SiO2 and

Al2O3. X-ray diffraction analysis of the cenospheres indicated that they contained

a large amount of amorphous material and small amounts of quartz and mullite

crystals (Wang et al., 2012b). Barbare et al. (2003) determined the uptake and loss

of water of cementitious composites with cenospheres and observed smaller water

penetration than normal concrete, indicating the cenospheres are not open except

those broken during the mixing process. The experimental results according to

ASTM C227-10 (2010) and ASTM C1260-14 (2014) demonstrated the stability of

cenospheres in terms of alkali silica reaction (Wang et al., 2012b).

Crushed granite with a nominal maximum size of 10 mm and natural sand

with a fineness modulus of 2.66 were used in the concrete mixture. Apparent

densities of the coarse and fine aggregates were 2650 and 2630 kg/m3,

respectively.
4
A polycarboxylate based superplasticizer (SP) was used for most of the

mixtures except for the concrete in which a naphthalene based SP 5 was used. The

reason for the use of polycarboxylate based superplasticizer with higher water

3
Laser Scattering Particle size Distribution Analyzer LA-950V2, Horiba Instruments, Japan
4
ADVA 181N, W.R. Grace Asia Pte Ltd, Singapore
5
Daracem® 100, W.R. Grace Asia Pte Ltd, Singapore
16
reducing capacity in the ULCCs and cement paste was because of their large

amount of fine materials (cenospheres and cementitious materials). However,

such superplasticizer is dosage sensitive to the stability of the mixtures. Thus,

naphthalene based superplasticizer was used for the concrete mixture.

6
Shrinkage reducing admixture (SRA) was used for some mixtures. A

previous study (Wang et al., 2012a) indicates that the SRA reduces the surface

tension of deionized water and synthetic pore solution. As a result, it can reduce

entrapped air content in the mixtures and improve the performance of the ULCCs

(Wang et al., 2013).

7
In addition, a viscosity modifying agent (VMA) which is often used to

increases plastic viscosity and produces more cohesive concrete was used in some

mixtures to adjust workability. Silane oligomers (3-glycidyloxypropyl

trimethoxysilane (C9H20O5Si)) 8 and a de-foaming admixture (Tributyl phosphate)

were used in a ULCC mixture reinforced with polyethylene (PE) fibers

(properties given in Table 2.5) to evaluate their effects. The silane oligomers were

prepared by heating the silane at 120 oC for 4 hrs (Ishida, 1993; Monticelli et al.,

2006).

2.2.2 Mix proportions and specimen preparation

Nine ULCCs, two cement pastes, and one concrete samples were included in

this study, and mix proportions are given in Table 2.6. The w/b of the ULCCs

ranged from 0.35 to 0.56. For comparison, two cement pastes with w/b of 0.35

6
Eclipse® Floor, W.R. Grace Asia Pte Ltd, Singapore
7
V-MAR@ 10P, W.R. Grace Asia Pte Ltd, Singapore
8
Silane (oligomers), Evonik Degussa (SEA) Pte Ltd, Singapore
17
and 0.45 were also prepared. However, the cement paste sample with w/b of 0.56

was not successful due to excessive bleeding. The ULCCs and cement pastes

contained 8% silica fume by mass of cemenentitious materials. Mix ULCC-7 and

Mix ULCC-8 contained 0.5% of polyethylene fibers by volume of the ULCCs to

reduce their brittleness. A concrete with a target 28-day compressive strength of

65 MPa was prepared for comparison.

The ULCCs and cement paste mixtures had good flowability determined

according to BS EN 1015-3 (1999) test, and no segregation was observed. For

most of the mixtures, the flow value was within a range of 190 to 215 mm with

only a few exceptions: due to larger amounts of the cenospheres used in Mixes

ULCC-4 and ULCC-6 in order to achieve lower densities, higher flow values of

240 to 250 mm were used to properly cast the specimens.

The specimens shown in Table 2.7 were made for each mixture. They were

covered with plastic sheets, and demoulded within 48 hrs after casting. The

specimens were then cured in a fog room at temperatures of 28 - 30 °C

(simulating tropical environment) until testing at 7 and 28 days, or other testing

ages with drying after 28-day moist curing.

2.2.3 Experimental methods

The test methods, relevant standards, testing ages, specimen sizes, and

specimen numbers are summarized in Table 2.7.

18
2.2.3.1 Density and mechanical properties

The density of all the cube specimens was determined, after demoulding,

using the water displacement method. The air content in the cement pastes or

mortar matrices (excluding the voids in the cenospheres) was estimated according

to ASTM C138 -13 (2013). Compressive strength was determined using 100-mm

cubes according to BS EN 12390-3 (2009). The compressive strength in this

thesis refers to the compressive strength of 100 mm cubes. Flexural performance

was determined using 75 × 100 × 400 -mm prisms with third-point loading (4-

point bending, span length 300 mm) as shown in Figure 2.3 (ASTM C1609-07,

2007). An Instron closed-loop, servo-controlled testing system was used for the

flexural test with a loading rate of 0.05 mm/min. Three specimens were tested for

each mixture and the load-deflection curves were averaged with software

OriginPro 2016 9. An example of the load-deflection curves of three specimens

and their average are presented in Figure 2.4 and the variability of the test results

appears low.

For measuring the elastic modulus, two strain gauges were pasted at mid

height of each specimen on two opposite sides (ASTM C469-14, 2014). A

Denison 3000 kN compression testing equipment was used with a loading rate of

0.01 mm/min. For Mixes ULCC-1 - ULCC-4 and CP 0.35, the specimens were

tested at around 70 days after casting (due to a problem of the equipment). During

the period from 28 to 70 days, these specimens were covered with plastic sheets

to prevent moisture loss.

9
OriginPro 2016, OriginLab Corporation, USA
19
2.2.3.2 Thermal conductivity

Thermal conductivity of the ULCCs specimens was determined according to

ASTM C518-10 (2010) by a heat flow meter 10 (Figure 2.5(a)). The applicable

range of the equipment was 0.002 - 0.500 W/m·K and the results reported were

the average from two specimens.

The thermal conductivities of the cement pastes and concrete were

determined following ASTM C177-13 (2013) using a guarded hot plate

equipment 11 (Figure 2.5(b)). The applicable testing range of the equipment was

0.02 - 2.00 W/m·K and two 300 × 300 × 50 -mm specimens were required for

each test.

The thermal conductivity of the concrete was also determined by a hot box

apparatus (Figure 2.5(c)) according to ASTM C1363-11 (2011) because the result

obtained using the guarded hot plate equipment was beyond the testing range of

the equipment.

The 300 × 300 × 50 -mm specimens used were first moist cured for 28 days,

and then dried in laboratory air for two weeks followed by drying in an oven at

105 °C until constant weight was achieved. The specimens were first put in an

oven with a temperature 60 °C for one day before moving to another oven with a

temperature of 105 °C to avoid thermal cracking. In general, about two months

were needed to achieve a constant weight.

10
Heat Flow Meter HFM 436/3 Lambda, NETZSCH, Germany
11
Guarded Hot Plate GHP 300, Holometrix Micromet Company, USA
20
2.3 Results and discussion

The results for density, mechanical properties, and thermal conductivity of

the ULCCs, cement pastes, and concrete are summarized in Table 2.8. It is

observed that the mixes with QK300 generally has higher estimated air content

than mixes with Exlite. This may be explained by higher viscosity of the mixtures

with QK300 (qualitatively observed during the experiment) due to larger particle

size of QK300. According to Banfill (2006), plastic viscosity of mortar increases

with the increase of aggregate size when all others factors (e.g. volume of

aggregate and w/c) remain constant due to increased inter-particle contact, surface

inter-locking, and inability of aggregate to be sheared.

2.3.1 Compressive strength, flexural tensile strength, and elastic modulus

Figure 2.6 clearly shows that the density of the ULCCs and concrete is

affected by the density of the aggregates embedded in the cement paste matrices.

The ULCCs had significantly lower density (1154 to 1471 kg/m3 after

demoulding) due to the incorporation of hollow cenospheres (particle densities =

615 and 908 kg/m3) as micro aggregates compared with the concrete (2341 kg/m3

after demoulding) which included granite coarse aggregate and natural sand with

densities of 2650 and 2630 kg/m3, respectively.

The compressive strength of the ULCCs, cement pastes, and concrete

samples was affected by their density. The compressive strength of the ULCCs

and cement pastes was reduced with the decrease in density (Figure 2.6).

Nevertheless, 28-day compressive strength of 69.4 MPa was achieved for the

21
mixture ULCC-1, similar to that of the cement paste with w/b of 0.35 and the

concrete. The specific strength (strength/density) of the ULCC-1 is 0.047

MPa/kg/m3, which is equivalent to a normal weight concrete with compressive

strength of about 110 MPa.

Flexural tensile strength of the ULCCs, cement pastes, and concrete increased

with the increase in the compressive strength (Figure 2.7(a)). The flexural tensile

strength of the ULCCs was about 8.4% to 10.9% of their 28-day compressive

strength, which is comparable to that of the cement pastes and concrete. The

flexural tensile strength of the ULCCs and cement pastes also decreased as the

density decreased (Figure 2.7(b)).

The elastic modulus of cement composites and concrete was generally

affected by their compressive strength and density. As shown in Figure 2.8(a), the

elastic modulus of the ULCCs increased with an increase in the compressive

strength. With the 28-day compressive strength within a range from 33.0 to 69.4

MPa, the elastic modulus of the ULCCs, cement pastes, and concrete decreased

with the density (Figure 2.8(b)). The elastic modulus of the ULCCs and concrete

were also affected by the density of cenospheres and aggregates incorporated in

the cement paste matrices.

Equation 2.2, given by ACI 318-11 “Building code requirements for

structural concrete and commentary” (2011), estimates the elastic modulus E of

concrete based on the compressive strength fc and density ρ of concrete. The

equation is applicable to concrete with compressive strengths < 41 MPa and

densities from 1440 to 2480 kg/m3. It is noted that the compressive strength fc in

22
the code refers to the compressive strength of cylinder specimens. From the

results of preliminary study, ULCCs cylinders and cubes have comparable

compressive strength. Therefore, the cube strength was used in this equation to

estimate the elastic modulus of ULCCs.

E = 0.043ρ1.5 (2.2)

The experimental values and the calculated values of the elastic modulus of

the ULCCs, together with their difference, are shown in Table 2.9. The difference

were plotted against 1 day density and 28 days compressive strength and shown in

Figure 2.9(a) and (b), respectively.

From Figure 2.9, It is clear that the equation overestimates the elastic

modulus of the ULCCs when the density > ~1230 kg/m3 and compressive

strength > ~43 MPa but underestimates the elastic modulus when the density <

~1230 kg/m3 and compressive strength < ~43 MPa. The difference between the

calculated and experimental elastic modulus of the ULCCs seems to increase

when the density and compressive strength are increased beyond ~1230 kg/m3

and ~43 MPa, respectively. It should be noted that although the equation from

ACI 318-11 (2011) is applicable to concrete, the equation may be used to estimate

the elastic modulus of the ULCCs with reasonable accuracy based on the results.

2.3.2 Thermal conductivity

Thermal conductivity of the ULCCs, cement pastes, and concrete reduced

with the decrease in oven dry density (Figure 2.10). The ULCCs had significantly

lower thermal conductivity than the cements pastes and concrete due to the

incorporation of the hollow cenospheres. Comparing the mixture ULCC-2,

23
cement paste (CP 0.35), and concrete of similar 28-day compressive strength, the

thermal conductivity of the ULCC-2 (0.39 W/m·K) was 54% lower than that of

the cement paste (0.84 W/m·K) and 80% lower than that of the concrete (1.98

W/m·K) (Table 2.8). With the decrease of oven dry density to 966 kg/m3, the

thermal conductivity of the ULCC was further reduced to 0.28 W/m·K (Table 2.8).

It should be mentioned that the thermal conductivity of 2.11 W/m·K was

obtained for concrete from the guard hot plate test. Although the result was

beyond the testing range of the equipment (0.02 - 2.00 W/m·K), this method

generally gives more accurate result than the hot-box method. The thermal

conductivity of the concrete was also affected by the source of the aggregate.

Thermal conductivities of 2.6 - 2.7 W/m·K and 3.1 W/m·K for concretes with

granite aggregates are mentioned in textbooks by Mehta and Monteiro (2005) and

Mindess et al. (2003), respectively.

The thermal conductivities of the cement pastes with w/b of 0.35 and 0.45

with oven dry densities of 1715 and 1479 kg/m3 were 0.84 and 0.80, respectively,

which are higher than those reported by Hajmohammadian Baghban et al (2013).

They reported a good correlation between the thermal conductivity and oven dry

density ranged from about 1050 to 1730 kg/m3 using experimental results of

cement pastes with w/c from 0.84 to 0.36. According to their relationship, the

thermal conductivities of the cement pastes with densities comparable to those in

our study should be about 0.73 and 0.60. The higher thermal conductivity values

of the cement pastes obtained in this study may be due to the incorporation of 8%

silica fume in the cement pastes which produces denser microstructure.

24
 The relationship between thermal conductivity and dry density of concrete

has been explored and reported in literature (Demirboğa and Gül, 2003; Gül et al.,

2007; opçu and Uygunoğlu, 2007; Valore, 1980). The thermal conductivity may

be predicted by oven dry density of concrete, however, the relationship is affected

by the type of aggregate used in concrete (ACI-122R, 2002).

Based on the results of the ULCCs from this study, a regression equation

(Equation 2.3) was obtained with R2 of 0.92 (plotted in Figure 2.10). Equation 2.3

is in similar form to an equation proposed by Valore et al. (1980) (Equation 2.4)

for predicting the thermal conductivity of concrete. The thermal conductivity of

the ULCCs can also be estimated by using Valore’s equation with a multiplication

factor of about 1.2 (Equation 2.5).

λ =0.11e( /1000)
(2.3)

λ =0.072e(1.25× /1000)
(2.4)

λ =0.086e(1.25× /1000)
(2.5)

where, λ - the thermal conductivity, W/m·K;

- the oven dry density, kg/m3.

2.3.3 Estimation of thermal conductivity of cenospheres and ULCCs

As described in Section 2.2.3, approximately 2 months are needed to dry a

ULCC specimens (300×300×50 -mm) to constant weight for determining their

thermal conductivity. Therefore, the estimation of thermal conductivity of ULCCs

using composite models is meaningful. Considering ULCCs as a discrete two-

phase composite material with the cenosphere particles as micro aggregate

25
embedded in cement paste matrix, its thermal conductivity may be estimated if

the thermal conductivity of each component (cement paste and cenospheres) and

their relative proportions are known. Some commonly used models for estimating

the thermal conductivity of ultra-lightweight cement composites include Series

and Parallel models, Hashin-Shtrikman (H-S) models (Hashin and Shtrikman,

1962) and cubic model (ACI-122R, 2002).

For fine particulate materials such as cenospheres, loosely packed samples

are often used to determine their thermal conductivities. The thermal conductivity

values of the samples obtained in such tests depend on the thermal conductivity of

the particulate materials and void space among the particles as shown in Figure

2.11. As the test sample includes particles and air voids among various particles,

the values may underestimate the actual thermal conductivities of the particles

because of the low thermal conductivity of air.

The thermal conductivity of the cenospheres QK300 was, therefore,

estimated using parallel model (Equation 2.6), series model (Equation 2.7),

Hashin-Shtrikman models (Equation 2.8 and 2.9) and cubic model (Equation

2.10). The calculation and analyses are based on the assumptions that (1) ULCC

is two phase composites of cement paste and cenospheres, and (2) volume sum of

the cement paste and cenospheres equal to 1, and (3) volume of the SRA, SP,

VMA and air voids is considered as a part of the cement paste. These two phase

models were selected considering that the volume of ITZ is relatively smaller than

those of cenospheres and cement pastes. Even though the study on determining

the volume of ITZ in ULCCs is not available, the secondary SEM images in this

26
study (Figure 2.12) and back-scattering SEM images in Reference (Rheinheimer

et al. 2017) indicate negligible thickness of ITZ layer between cenospheres and

cement pastes.

Parallel model:

(2.6)

Series model:

1 V Vceno
= λ cp (2.7)
λU cp λceno

H-S model (upper bound):

Vceno
λU =λcp 1 Vcp (2.8)
λceno -λcp 3λcp

H-S model (lower bound):

Vcp
λU =λceno 1 Vceno (2.9)
λcp -λceno 3λceno

Cubic model:

V2ceno
3
λU =λcp Vceno (2.10)
V2ceno
3
-Vceno ( )
λceno V2 3
ceno 1-V2 3
λcp ceno

Where, λU - thermal conductivity of ULCCs, W/m·K;

λcp - thermal conductivity of cement pastes, W/m·K;

ceno - thermal conductivity of cenospheres, W/m·K;

Vcp - volume ratio of cement paste in ULCCs.

Vceno - volume ratio of cenospheres in ULCCs.

The ULCC mixtures 1, 2, and 4 and corresponding cement pastes (CP 0.35

and CP 0.45) were used for estimating the thermal conductivity of the

27
cenospheres QK300, and results are shown in Table 2.10. The ULCC Mixtures

4(VMA), 7, and 8 were not used in the estimation due to possible influence of

VMA, fibers, and silane on the thermal conductivity. The ULCC-3 was also not

included because of different w/b. However, these four mixtures were used to

verify if the estimated thermal conductivity of the cenospheres QK300 can be

used to estimate the thermal conductivity of the ULCCs.

From Table 2.10, the thermal conductivity of the cenospheres QK300

estimated from the Series model and H-S lower bound give reasonably consistent

values from 0.201 to 0.216 W/m·K and 0.127 to 0.137 W/m·K, respectively,

whereas those estimated from Parallel, H-S upper bound, and cubic models give

negative values. The negative values indicate that the Parallel, H-S upper bound,

and cubic models may not be suitable in this case. The estimated average thermal

conductivity of the cenospheres is 0.209 W/m·K from Series model and 0.133

W/m·K from H-S lower bound. These estimated thermal conductivity values of

the cenospheres QK300 are significantly higher than that provided by

manufacturer (0.08 W/m·K) which is likely due to the effect of void space among

the cenosphere particles mentioned earlier (Figure 2.11).

Based on the estimated thermal conductivity of the cenospheres above, the

thermal conductivities of the ULCCs are calculated using the Series model and H-

S lower bound, which are further compared with corresponding experimental

values. Since the thermal conductivity of the cement paste is not affected

significantly by the w/b between 0.35 and 0.45, the thermal conductivity of the

cement paste for ULCC-3 (w/b = 0.37) was estimated to be 0.83 W/m·K using

28
linear interpolation. Comparison of the calculated and experimental values of the

thermal conductivities of ULCCs is given in Table 2.11. The error of estimation

using the H-S lower bound and Series model are both less than 10%. Overall, the

H-S lower bound gives lower error of estimation than the Series model in

estimating the thermal conductivity of the ULCCs.

2.3.4 Effect of some ingredients in the ULCCs on mechanical properties and

thermal conductivity

2.3.4.1 Effect of cenospheres density and amount

A scanning electron microscope image of a typical ULCC is shown in Figure

2.12. The incorporation of hollow cenospheres in the ULCCs reduced the density

and thermal conductivity substantially but maintained the compressive strength

compared to the cement paste of similar w/b although the elastic modulus was

reduced. For example, with w/b of 0.35 (compare ULCC-2 and CP 0.35), the

incorporation of the cenospheres reduced the density and thermal conductivity by

27% and 54%, respectively; with w/b equal to 0.45 (compare ULCC-4 and CP

0.45), the incorporation of the cenospheres reduced the density and thermal

conductivity by 33% and 59%, respectively. For both w/b, the elastic modulus

was reduced by about 20%.

The good strength properties of the ULCCs in comparison to that of the

cement pastes may be attributed to (1) hard and rigid cenospheres, (2) “controlled”

void sizes in the cenospheres, and (3) dense cement paste matrices provide “three

dimensional” confinement to the cenospheres.

29
 For a given type of cenospheres used (Mixes ULCC-1, -3, and -4 or Mixes

ULCC-5 and -6) (Table 2.8), the increase in the amount of the cenospheres and

w/b leads to a reduction in the thermal conductivity and mechanical properties.

For a given w/b, the ULCC-5 had lower porosity in the cement paste matrix but

higher porosity in the cenospheres (Exlite) compared with ULCC-3 (Table 2.8).

As the two ULCCs had similar density, their mechanical properties and thermal

conductivity were similar.

2.3.4.2 Effect of shrinkage reducing admixture and viscosity modifying admixture

Although the shrinkage reducing admixture (SRA) reduces the surface

tension of deionized water and synthetic pore solution of cement paste (Wang et

al., 2012a) and, consequently, reduce the air content of the ULCCs (Wang et al.,

2013), the results shown in Table 2.8 indicate that the SRA did not have

significant influence on the density, mechanical properties, and thermal

conductivity of the ULCC-1 and ULCC-2. This may be due to the different

superplasticizer used in this study.

Mix ULCC-4 had a higher flow value of 240 mm in comparison to other

ULCCs mixtures, therefore the VMA was used in Mix ULCC-4VMA (flow value

= 195 mm) to determine if the flow value and VMA can affect the density,

thermal conductivity, and mechanical properties of the ULCC. The Results shown

in Table 2.8 indicate that the density and mechanical properties were somewhat

reduced in the mixture with VMA; however, the thermal conductivity was not

affected.

30
2.3.4.3 Effect of fibers and silane

Fibers were incorporated to reduce the brittleness of the ULCCs. Figure 2.13

shows flexural stress versus deflection curves of the fiber-reinforced ULCC-7 and

ULCC-8 in comparison to that of the plain ULCC-2. The absence of post-peak

curve of the ULCC-2 under bending is due to limitations in the experimental

setup. When the flexural strength was reached, each specimen of the plain ULCC

was broken into two pieces without further load carrying capability, whereas the

specimens of the fiber-reinforced ULCCs were still able to carry bending load as

shown in the Figure 2.13. This indicates reduced brittleness of the fiber-reinforced

ULCCs.

However, the incorporation of the polyethylene fibers (Mix ULCC-8)

reduced the density, strengths, and elastic modulus of the ULCC, even though it

did not affect the thermal conductivity significantly. To determine the effect of

the fibers at given density of the ULCC, the silane and defoaming agent described

earlier were used in Mix ULCC-7. The fiber-reinforced ULCC-7 had density,

strengths, and elastic modulus similar to those of the plain ULCC-2 without fibers.

However, the thermal conductivity of the former was higher than the latter. The

higher thermal conductivity of the fiber-reinforced ULCC-7 in comparison to the

plain ULCC-2 was explained by the higher thermal conductivity of the

polyethylene fibers (0.46-0.52 W/m·K (Mark, 2007)).

It should be noted that the silane may also affect the thermal conductivity.

Hajmohammadian Baghban et al. (2013) reported decreased thermal conductivity

for oven dried cement paste with alkyl alkoxysilane, whereas Xu and Chung

31
(2000) reported increased thermal conductivity of cement paste with amino vinyl

silane in comparison to the control paste with similar density.

2.3.5 Balance of thermal conductivity and mechanical properties

A decrease in the density of concrete generally results in reduced thermal

conductivity but also decreases the mechanical properties, such as strengths and

elastic modulus. The former is beneficial to reduce heat gain through opaque

building envelope but, on the other hand, sufficient mechanical properties are

essential for structural applications.

Since the thermal conductivity of concrete is significantly affected by

moisture content, the thermal conductivity is generally determined based on oven

dry specimens. In the following discussion to compare results from this study to

those reported in literature, if the oven dry density is not reported, it is then

estimated based on mix proportions provided according to ASTM C567-00 (2000).

In this study, high compressive and flexural tensile strengths of 69.4 MPa and

7.3 MPa, respectively, were achieved for the ULCC-1 sample, comparable to

those of the concrete sample. The ULCC-1 also had a thermal conductivity of 0.4

W/m·K, substantially lower than that of the concrete. The low thermal

conductivity of the ULCCs is due to hollow cenospheres which effectively

introduce voids and decrease density of the ULCCs, whereas the high strengths

obtained may be attributed to the hard and rigid cenospheres under confinement

of dense cement paste matrices and controlled void sizes in the cenospheres as

mentioned earlier. The combination of low thermal conductivity, lightweight, and

32
high specific strength of the ULCCs is significant improvement compared with

those reported in literature (Figure 2.14(a) and Table 2.12).

As shown in Figure 2.14(a) and Table 2.12, most low thermal conductive

concrete (thermal conductivity < 0.4 W/m·K) have compressive strength less than

30 MPa. Although the composites reported by Huang et al (2013) have low

thermal conductivity of < 0.4 W/m·K, the composites have lower compressive

strength (< 50 MPa) and higher estimated oven dry density (≥1480 kg/m3)

compared with the ULCCs obtained in this study. The lightweight concretes

reported by Tandiroglu (2010) have high compressive strengths of 60 - 80 MPa

but also higher thermal conductivity (>1.4 W/m·K) and oven dry density (>1798

kg/m3).

Blanco et al. (2000) also used cenospheres to make lightweight concretes and

obtained low densities comparable to those of the ULCCs in this study and low

thermal conductivities of 0.36 - 0.46 W/m·K. However, the compressive strength

of most of their lightweight concretes were only 5 to 30 MPa (Table 2.3), much

lower than that of the ULCCs obtained in this study. Another study incorporated

cenospheres is Huang et al. (2013) which is discussed above.

For comparison between various mix proportions, it is desirable to define a

merit number to take compressive strength, density, and thermal conductivity into

consideration. Since high specific strength and low thermal conductivity are

desired, a merit number “M” is defined as specific strength divided by thermal

conductivity. Even though this number doesn’t have physical meaning, it provides

an approach to directly evaluate the performance of concrete in terms of strength,

33
density, and thermal conductivity. Higher values of the merit number indicate

better optimization between density, mechanical and thermal properties. The

histogram shown in Figure 2.14(b) indicates that the merit number of the ULCC is

basically one order of magnitude higher than most of the results from Sengul et al.

(2011), Mounanga et al. (2008), Gül et al. (2007), and Yun et al. (2013).

2.4 Summary and conclusions

Mechanical properties and thermal conductivity of ultra-lightweight cement

composites (ULCCs) with 1-day density from 1154 to 1471 kg/m3 and 28-day

compressive strength from 33.0 to 69.4 MPa were studied. The low densities of

the ULCCs were achieved by incorporating hollow cenospheres from fly ash

generated in thermal power plants. The properties of the ULCCs were compared

with those of cement pastes with w/b of 0.35 and 0.45 and those of a concrete

with 28-day compressive strength of 67.6 MPa. Based on the experimental results,

numerical calculation, and discussion, following conclusions can be drawn:

1. Compressive strength, flexural tensile strength, and elastic modulus of the

ULCCs were reduced with decrease in density. However, the compressive and

flexural tensile strength of 69.4 and 7.3 MPa were achieved for the ULCC-1,

respectively, similar to those of the cement paste with w/b of 0.35 and the

concrete. The specific strength of the ULCC-1 is 0.047 MPa/kg/m3, equivalent to

that of a normal weight concrete with compressive strength of about 110 MPa.

The density and elastic modulus of the ULCCs and concrete were affected by the

density of aggregates embedded in the cement paste matrices. The ratio of the

34
flexural tensile strength/compressive strength of the ULCCs is comparable to that

of the cement pastes and concrete.

2. Thermal conductivity of the ULCCs was reduced with decrease in the

oven-dry density and it was significantly lower than that of the cement pastes and

concrete due to the incorporation of hollow cenospheres. With similar 28-day

compressive strength, the thermal conductivity of the ULCC was 54% and 80%

lower than that of the cement paste and concrete, respectively.

3. Compared to properties from various lightweight concretes reported in

literature, the ULCCs have desirable combination of low thermal conductivity,

lightweight, and high specific strength. The low thermal conductivity of the

ULCCs is due to the incorporation of hollow cenospheres as micro-aggregates,

which effectively introduce voids and decrease density of the ULCCs. The high

specific strength of the ULCCs may be attributed to (1) the presence of hard and

stiff shell present in the cenospheres, (2) the development of “controlled” void

sizes in the cenospheres, and (3) the creation of strong cement paste matrices that

provide “three dimensional” confinement to the cenospheres.

4. The incorporation of polyethylene fibers reduced brittleness observed in

the ULCC samples, but did not have significant effect on the strength and elastic

modulus if the density and w/b were kept constant. However, the fiber reinforced

ULCC had higher thermal conductivity than that without fibers, which may be

due to the higher thermal conductivity of the polyethylene fibers. The shrinkage

reducing admixture and viscosity modifying agent did not have significant effect

on the mechanical properties and thermal conductivity of the ULCCs.

35
 5. he thermal conductivity of the U can be estimated by using Valore’s

equation: λ =0.072e(1.25× /1000)

with a multiplication factor of about 1.2.

6. The estimated average thermal conductivity values of the cenospheres

based on H-S lower bound and Series model are significantly higher than that

provided by manufacturer due to the effect of void space among the cenosphere

particles.

7. Based on the estimated thermal conductivity of cenospheres, the thermal

conductivity of the ULCCs can be estimated using H-S lower bound and Series

model with error of estimation less than 10%. Overall, the H-S lower bound gives

lower error of estimation than the Series model.

36
 2.5 Tables

Table 2.1 Thermal conductivities of concrete and its constituent

Thermal Conductivity,
Reference
W/m·K

Granite 3.1
Basalt 1.4

Limestone 3.1
Dolomite 3.6 Mindess et al., 2003
Aggregate
Sandstone 3.9

Quartzite 4.3

Marble 2.7

Quartz 5.8 ACI 122R, 2002

w/c = 0.4 1.3

Cement paste
(saturated w/c = 0.5 1.2
condition)
w/c = 0.6 1.0
Mindess et al., 2003
Water 0.5

Air 0.03

Concrete 1.5 - 3.5

Granite aggregate 2.6 - 2.7

Basalt aggregate 1.9 - 2.2

Limestone aggregate 2.6 - 3.3 Mehta and

Concrete
Dolomite aggregate 3.2 Monteiro, 2005

Quartzite aggregate 3.5

Rhyolite aggregate 2.2

37
 Table 2.2 Thermal conductivity of concrete under different moisture conditions
(Scanlon and McDonald, 1994)

Moisture condition Thermal conductivity (W/m·K)

Moisture 2.2

Limestone concrete RH: 50% 1.7

Dry 1.4

Moisture 2.9

Sandstone concrete RH: 50% 2.2

Dry 1.4

Moisture 3.3

Quartz gravel concrete RH: 50% 2.7

Dry 2.3

Moisture 0.85
Expanded shale concrete RH: 50% 0.79
Dry 0.62
RH: Relative humidity

38
 Table 2.3 Summary of lightweight aggregate concrete in literature

Oven dry Compressive Moisture condition Thermal

Aggregate information Wet density,
References w/b density, strength, at which TC was Conductivity,
Type / Maximum size kg/m3
kg/m3 MPa determined W/m·K

Moist cured 0.46-0.60

Blanco et al. 2000 Cenosphere / 4mm 0.3 1090-1415 1050-1350* 5.0-30.1
Dried at 100 oC 0.36-0.46
0.38-
Gao et al. 2014 Aerogel /4mm 1000-2050 900-1950* 8.3-60* NA 0.26-1.9**
0.39
Gül et al. 2007 Perlite /16mm 0.7 1773-1984 1590-1800* 11.3-25.1 NA 0.82-1.23
Cenosphere /600um & 1483-
Huang et al. 2013 0.26 1649-2001 44.3-48.1 NA 0.29-0.37
Iron ore tailings / 300um 1890***
Fine bottom ash /0.6mm
Kim et al. 2012 & expanded shale / 0.47 1800-1553 1500-1210** 22-8 0% 0.54-0.36
19mm
Mounanga et al. Polyurethane foam waste NA 1.44-0.31
0.6-0.7 1261-2165 900-1980* 23.0-3.2
2008 / 10mm Saturated 2.34-0.68
Tandiroglu 2010 Perlite / 15mm 0.4-0.65 NA 1798-1883 60-80 0% 1.47-1.76**
Diatomite /4mm 900 6 0.13
Topçu and
0.2 NA 0%
Uygunoğlu 2007
Pumice / 4mm 1500 9 0.44

Uysal et al. 2004 Pumice /16mm NA 2270-1329 NA NA 0% 1.46-0.78

Wang and Meyer High impact polystyrene
0.55 NA 1980-1560 37-19 NA 0.61-0.27
2012 /2.36mm
0.38-
Yu.et al. 2013 Expanded glass / 4.0mm 1280-1460 1100-1380* 23-30 0% 0.49-0.85
0.59
Glass Bubble / (Median
Yun et al. 2013 0.55 2011-2370 1900-2260* 43.9-24.6 NA 2.25-1.41
diameter: 65um)
NA - not available
*
estimated based on ASTM C567-00 (2000) where mix proportions are available
**
obtained from Figures in literature based on aggregate in saturated surface dry condition.
***
air dry density was used here because it’s lower than estimated oven dry density using AS M 567-00 (2000).

39
 Table 2.4 Chemical and mineral compositions of cement and silica fume (% by mass) from the manufacturers

CaO SiO2 Al2O3 Fe2O3 MgO Na2O K2O SO3 LOI C3S C2S C3A C4AF

Portland
63.8 20.4 4.9 3.6 4.3 0.3 0.4 2.3 1.5 60.1 13.2 6.9 11.0
cement
Silica fume NA 87.8 NA NA NA 0.3 0.9 0.5 3.0 NA NA NA NA

Table 2.5 Properties of polyethylene (PE) fibers

Tensile strength, Elastic modulus,

Length, mm Diameter, µm Aspect ratio Density, kg/m3
MPa GPa
Polyethylene fibers 2610 79 12 39 308 970

40
 Table 2.6 Mix Proportions and workability of ultra-lightweight cement composites (ULCCs), cement pastes, and concrete

Coarse Binder PE Flow/

Water Sand * Cenosphere SP SRA VMA Silane**
aggr Fibers Slump
Cenosphere
Mix ID Variables w/b Vol,%
type kg/m3
kg /m3† † in L/ m3† kg/m3† mm
ULCCs
Max
Concrete 0.42 --- 172 946 810 410 --- --- 5.4 0 0 0 0 95 ††
aggr=10mm
ULCC-1 0.35 QK300 302 --- --- 909 348 38.3 4.9 10.5 0 0 0 200

ULCC-3 w/c, 0.37 QK300 282 --- --- 796 402 44.3 5.2 9.8 0 0 0 210
cenosphere
ULCC-5 0.37 Exlite 287 --- --- 795 268 43.6 5.9 8.9 0.202 0 0 215
type and
ULCC-4 density 0.45 QK300 282 --- --- 660 442 48.7 5.6 9.8 0 0 0 240

ULCC-6 0.56 Exlite 290 --- --- 542 317 51.6 6.6 9.1 0.202 0 0 250
ULCC-2 0.35 QK300 305 --- --- 920 352 38.7 5.2 0 0 0 0 195
SRA
ULCC-1 0.35 QK300 302 --- --- 909 348 38.3 4.9 10.5 0 0 0 200
ULCC-4 VMA, 0.45 QK300 282 --- --- 660 442 48.7 5.6 9.8 0 0 0 240
ULCC-4 control
workability 0.45 QK300 282 --- --- 659 442 48.6 6.7 9.7 0.176 0 0 195
(VMA)
ULCC-2 0.35 QK300 305 --- --- 920 352 38.7 5.2 0 0 0 0 195
Fiber and
ULCC-7 silane + 0.35 QK300 304 --- --- 842 350 38.6 3.6 0 0 4.2 5.3 200
defoamer
ULCC-8 0.35 QK300 301 --- --- 904 346 38.1 4.3 0 0 0 5.7 190

CP 0.35 0.35 --- 499 --- --- 1473 --- --- 1.3 14.9 0 0.0 0.0 200
Cement
pastes, w/c
CP 0.45 0.45 --- 561 --- --- 1282 --- --- 0.0 16.9 0.912 0.0 0.0 160
*
binder contained 8% silica fume (SF) by mass of cementitious materials except for concrete which contains no SF.
**
5% percent (by weight of silane) of tributyl phosphate was added together with silane to avoid excessive foaming.
†
for 1m3 of all materials added.
††
slump
41
 Table 2.7 Test methods and specimen information

Specimen
Property Test standard Testing age
Type and size Number
Workability (flow table)
BS EN 1015-3 (1999) Right after mixing --- ---
(ULCCs and cement pastes)

Workability (slump test)

ASTM C143-03 (2003) Right after mixing --- ---
(concrete)
Density of hardened specimens Cube, 100×100×100
BS EN 12390-7 (2009) 1-2 day 6
after demoulding mm

Estimation of porosity ASTM C138-13 (2013) --- --- ---

Cube, 100×100×100
Compressive strength BS EN 12390-3 (2009) 7 & 28 days 3
mm
Flexural performance ASTM C1609-07 (2007) 28 days Prism,75×100×400 mm 3
Cylinder,
Elastic modulus ASTM C469 (2014) 28 days 3
Ø100×200 mm
Thermal conductivity of ULCCs ASTM C518-10 ( 2010)

Thermal conductivity of cement Slab,

ASTM C177-13 (2013) About 100 days* 2
pastes 300×300×50 mm
Thermal conductivity of ASTM C177-13 (2013) and
concrete ASTM C1363-11 (2011)
*
28d moist curing followed by drying in lab air for 14d and then drying in oven at 105 °C to constant weight

42
 Table 2.8 Density, mechanical properties, and thermal conductivity

Estimated air 28 day

Oven 7 day
1day content in Thermal
Aggr Other dry comp Comp Flexural
Mix ID Variable w/b density, paste matrix, conductivity,
type information density, strength,
kg/m3 expressed strength, strength, E, GPa W/m·K
kg/m3 MPa
as % in 1m3 MPa MPa
Max
Concrete aggr=10m 0.42 Granite --- 2341 (8) 2251 2.9 59.2 (1.4) 67.6 (0.9) 6.5 (0.3) 35.5 (1.3) 1.98
m
---
ULCC-1 0.35 QK300 1471 (2) 1303 6.6 57.8 (3.4) 69.4 (7.0) 7.3 (0.3) 17.0 (0.4) 0.40±0.00

ULCC-3 0.37 QK300 --- 1357 (2) 1198 9.2 40.4 (5.2) 56.9 (3.5) 5.4 (0.1) 14.8 (0.2) 0.36±0.00
w/c,
ULCC-5 cenosphere 0.37 Exlite --- 1355 (3) 1178 1.2 42.8 (0.8) 55.9 (2.1) 4.7 (0.3) 13.9 (0.1) 0.35±0.00
type and
density ---
ULCC-4 0.45 QK300 1240 (4) 1087 11.3 38.7 (2.8) 49.8 (1.7) 4.4 (0.2) 13.0 (0.1) 0.33±0.01
---
ULCC-6 0.56 Exlite 1154 (4) 966 1.0 24.5 (1.3) 33.0 (0.8) 3.6 (0.1) 10.4 (0.3) 0.28±0.00

ULCC-2 0.35 QK300 No SRA 1457 (3) 1298 7.8 57.7 (5.1) 66.1 (2.1) 7.0 (0.3) 17.0 (0.1) 0.39±0.00
SRA
ULCC-1 0.35 QK300 With SRA 1471 (2) 1303 6.6 57.8 (3.4) 69.4 (7.0) 7.3 (0.3) 17.0 (0.4) 0.40±0.00

ULCC-4 VMA, 0.45 QK300 No VMA 1240 (4) 1087 11.3 38.7 (2.0) 49.8 (1.7) 4.4 (0.2) 13.0 (0.1) 0.33±0.01
ULCC-4 control
workability 0.45 QK300 With VMA 1196 (4) 1042 14.4 32.0 (2.6) 40.9 (0.9) 4.1 (0.2) 12.0 (0.6) 0.31±0.00
(VMA)
ULCC-2 0.35 QK300 No fiber 1457 (3) 1298 7.8 57.7 (5.1) 66.1 (2.1) 7.0 (0.3) 17.0 (0.1) 0.39±0.00
Fiber and With fiber
ULCC-7 silane+ 0.35 QK300 1455 (2) 1300 7.8 57.2 (3.5) 66.5 (3.3) 6.3 (0.6) 16.0 (0.1) 0.43±0.01
& silane
defoamer
With fiber
ULCC-8 0.35 QK300 1361 (2) 1219 13.4 43.6 (2.5) 54.4 (0.5) 5.2 (0.3) 14.3 (0.4) 0.39±0.00
no silane
Cement
CP 0.35 0.35 --- 2000 (3) 1715 0.0 52.5 (5.1) 63.1 (8.5) 6.7 (1.1) 21.6 (0.5) 0.84
Cement paste
pastes, w/c CP, with
CP 0.45 0.45 --- 1854 (12) 1479 0.4 37.8 (0.3) 51.3 (1.8) 3.5 (1.2) 16.5 (1.2) 0.80
VMA
* the values included in the bracket refer to standard deviation of the testing results, while the values behind “±” refer to testing range. The number of testing results
used for variation evaluation is detailed in Table 2.7. The testing results ranges of CP0.35 and CP0.45 are not presented because only one value was obtained by
using averaged heat flow and temperature gradient of two specimens (Section 2.2.3.2 and Figure 2.5 (b)). Only one sample of concrete was used to determine the
thermal conductivity (Section 2.2.3.2 and Figure 2.5 (c)).
43
 Table 2.9 Comparison of experimental and calculated elastic modulus of the ULCCs

1-day 28-day Experimental Difference between

Aggregate Calculated elastic
Mix ID w/b density, compressive Elastic modulus, experimental and
type modulus, GPa
kg/m3 strength, MPa GPa calculated E, %

ULCC-1 0.35 QK300 1471 69.4 17.0 20.2 18.9

ULCC-2 0.35 QK300 1457 66.1 17.0 19.4 14.4

ULCC-3 0.37 QK300 1357 56.9 14.8 16.2 9.6

ULCC-4 0.45 QK300 1240 49.8 13.0 13.2 1.9

ULCC-4
0.45 QK300 1196 40.9 12.0 11.4 -5.2
(VMA)
ULCC-5 0.37 Exlite 1355 55.9 13.9 16.0 15.4

ULCC-6 0.56 Exlite 1154 33.0 10.4 9.7 -6.9

ULCC-7 0.35 QK300 1455 66.5 16.0 19.5 21.6

ULCC-8 0.35 QK300 1361 54.4 14.3 15.9 11.4

44
Table 2.10 Estimation of thermal conductivity of the cenospheres by Series, Parallel, Cubic, and Hashin-Shtrikman (H-
S) models

Thermal conductivity
ceno, W/m·K, estimated by using
Mix ID w/b λULCC, W/m·K cp, W/m·K Vceno
Parallel H-S upper H-S lower
Series model Cubic Model
model bound bound
ULCC-1 0.35 0.398 0.835 0.383 -0.306 0.216 -0.118 -0.056 0.137
ULCC-2 0.35 0.388 0.835 0.387 -0.320 0.210 -0.128 -0.066 0.127
ULCC-4 0.45 0.327 0.804 0.487 -0.175 0.201 -0.040 -0.006 0.136

Table 2.11 Comparison of experiment and estimated thermal conductivity of the ULCCs

Thermal Estimated thermal conductivity of ULCCs

cp, conductivity
Mix ID w/b W/m· Vceno of ULCCs From H-S lower bound From Series model
K (Experiment), Error of Error of
W/m·K ceno = 0.133W/m·K estimation ceno = 0.209 W/m·K estimation
ULCC-
0.45 0.804 0.486 0.308 0.323 4.8% 0.337 9.5%
4(VMA)
ULCC-3 0.37 0.829* 0.443 0.355 0.352 0.9% 0.358 0.9%
ULCC-7 0.35 0.835 0.386 0.425 0.390 8.3% 0.387 8.9%
ULCC-8 0.35 0.835 0.381 0.390 0.393 0.8% 0.390 0.0%
*
calculated by linear interpolation

45
 Table 2.12 Comparison of compressive strength, density, and thermal conductivity of ULCCs and lightweight
concrete reported in literature

Oven dry density, Compressive strength, Thermal Conductivity,

References
kg/m3 MPa W/m·K

ULCCs in this section 966-1306 33.0-69.4 0.28-0.43

Blanco et al. 2000 1050-1350* 5.0-30.1 0.36-0.46

Gao et al. 2014 900-1950* 8.3-60* 0.26-1.9**

Gül et al. 2007 1590-1800* 11.3-25.1 0.82-1.23
***
Huang et al. 2013 1483-1890 44.3-48.1 0.29-0.37
Kim et al. 2012 1500-1210** 22-8 0.54-0.36

Mounanga et al. 2008 900-1980* 23.0-3.2 1.44-0.31

Tandiroglu 2010 1798-1883 60-80 1.47-1.76**

900 6 0.13
opçu and Uygunoğlu
2007 1500 9 0.44
Wang and Meyer 2012 1980-1560 37-19 0.61-0.27
Yu.et al. 2013 1100-1380* 23-30 0.49-0.85
Yun et al. 2013 1900-2260* 43.9-24.6 2.25-1.41
* estimated based on ASTM C567-00 (2000) where mix proportions are available
** obtained from Figures in literature based on aggregate in saturated surface dry condition.
*** air dry density was used here because it’s lower than estimated oven dry density using ASTM C567-00 (2000).

46
 2.6 Figures

Figure 2.1 Particle size distributions of the two types of cenospheres used

Figure 2.2 Image of cenospheres (Exlite)

47
 Figure 2.3 Photograph of the ASTM C1609-07 (2007) test setup

Figure 2.4 Average curve of the load-deflection curves by using OriginPro

48
Figure 2.5 Schematic experimental setup for determining thermal conductivities of by heat
flow meter (a), guarded hot plate (b), and hot box (c)

49
 Figure 2.6 28-day compressive strength vs wet density after demoulding

Figure 2.7 (a) Flexural tensile strength vs compressive strength; (b) Flexural
tensile strength vs wet density after demoulding

50
Figure 2.8 (a) Elastic modulus vs compressive strength, (b) Elastic modulus vs wet density
after demoulding

Figure 2.9 Difference between experimental and calculated elastic modulus

versus (a) wet density at 1d and (b) 28d compressive strength

51
 Figure 2.10 Thermal conductivity vs oven-dry density

Figure 2.11 Loose packed particulate sample for thermal conductivity test (not to scale)

52
Figure 2.12 Scanning electron microscope image of a ULCC with cenospheres

Figure 2.13 Flexural stress-deflection curves of fiber-reinforced ULCCs (ULCC

7 and 8) in comparison to that without fibers (ULCC 2)

53
 Figure 2.14 (a) Density, compressive strength, and thermal conductivity of the
ULCCs in comparison to concrete with various lightweight aggregates reported in
literature, (b) Merit number defined as M = compressive strength /(thermal
conductivity × density) = MPa/(W/m·K) × (kg/m3) = 104·K·m3/W·s2 of various
lightweight concretes mentioned in (a)

54
 Chapter 3. Effects of silica fume and fly ash on properties of

plain and fiber reinforced ultra-lightweight cement composites

3.1 Introduction

Various fibers such as steel and polymer fibers, e.g. polyvinyl alcohol (PVA)

and polyethylene (PE) fibers, have been incorporated in the ULCCs to reduce

their brittleness (Liu et al., 2016; Wang et al., 2013; Wu et al., 2015). Compared

with polymer fibers, steel fibers may be more effective in improving the

mechanical properties of cement composites such as flexural strength and residual

strength after cracking due to their higher elastic modulus (Bentur and Mindess,

2006). For example, Wang et al. (2013) reported that the ULCCs with steel fibers

had better post cracking performance (toughness and residual strength) than that

with comparable amount of polyethylene fibers. The performance of fiber

reinforced ULCCs is dependent on the dispersion and stability of the cenospheres

and fibers in the composites, and would be reduced significantly with the

segregation of the cenospheres and fibers from the composites.

Beris et al. (1985) investigated the settlement of a spherical particle in

Bingham fluid and found out that a spherical particle would settle in a Bingham

fluid when the dimensionless parameter Yg defined in Equation 3.1 is less than

0.143 (Beris et al., 1985), and the velocity of settlement may be predicted by

Stokes drag (Equation 3.2) (Petrou et al., 2000).

3τ0
g = 2R(ρ -ρ )g (3.1)
s f

-
(3.2)
55
 Where, Yg - dimensionless parameter;

- yield stress, Pa;

R - radius of the spherical particle, m;

- density of the particles, kg/m3 ;

- density of the fluid, kg/m3 ;

g - the acceleration due to gravity, m/s2 ;

U - velocity of settlement in the fluid, m/s;

η - plastic viscosity of the fluid, Pa·s;

Cs- stokes drag coefficient.

From these equations, it is clear that reduction in the maximum aggregate

size and density difference between aggregates and paste matrices may reduce the

potential segregation. For given aggregate size and density, potential segregation

of the cement composite may be reduced with the increase in the yield stress and

plastic viscosity of the matrices.

A number of studies on lightweight aggregate concrete demonstrate that the

segregation of aggregates from mortar matrices can be described by the Equations

3.1 and 3.2 (Chia et al., 2005; Navarrete and Lopez, 2016) although the

lightweight aggregate tends to float to top in concretes.

For ULCCs, the density of the cenospheres is significantly lower than that of

the paste matrices, whereas the steel fibers have substantially higher density than

that of the matrices. Thus, the stability of the cenospheres and steel fibers in the

ULCCs needs to be taken into consideration even though the particle sizes of the

cenospheres are relatively small. While recognizing that these equations are

56
intended to describe the stability or segregation of the composites with density of

particulate inclusions higher than that of the matrices, the principles can be

applied to the stability of cenospheres and fibers in the ULCCs.

The fresh ULCCs can be considered as composites with cenospheres

embedded in cementitious matrices. Various studies demonstrate that the water to

cement ratio (w/c), particle size distributions of aggregates and cementitious

materials, hydration time, mineral and chemical admixtures, temperature, and

mixing procedure affect the rheological behavior of fresh cement composites

(Banfill, 2006; Bentz et al., 2012; El-Chabib and Nehdi, 2006; Nehdi et al., 1998;

Shaughnessy and Clark, 1988; Struble et al., 1998; Taylor, 1990). For a cement

composite with a given w/c, its yield stress and plastic viscosity may be increased

by incorporating materials with higher specific surface area than that of cement

(Khayat, 1998; Nehdi et al., 1998). Therefore, silica fume has been used to

increase the yield stress and plastic viscosity of cement composites (Park et al.,

2005) in order to reduce potential segregation. ACI committee 234R-06: “Guide

for the use of silica fume in concrete” (2006) states that the incorporation of silica

fume increased the cohesiveness of concrete and segregation resistance of

aggregate. On the other hand, the yield stress and plastic viscosity are correlated

to the flowability of fresh cement composites. Experimental studies indicate that

the slump value of fresh concrete decreases with the increase in the yield stress

(Chia and Zhang, 2004; Wallevik, 2006). Therefore, the content of silica fume in

cement composites needs to be optimized by considering both flowability of fresh

composites and stability of ingredients such as cenospheres and fibers.

57
 Another concern is that Portland cement content in the ULCCs (600-850

kg/m3) is much higher than that in conventional concrete of comparable strength

due to the elimination of coarse aggregate. Since the strength of ULCCs is mainly

governed by the cenospheres, the strength decrease of cement paste matrix may

not lead to significant decrease in the strength of the U s. According to “FIP

manual of lightweight aggregate concrete” (FIP, 1983), there is a “limit strength

value” for lightweight concrete determined by the strength and elastic modulus of

aggregates. Once the “limit strength” of lightweight concrete is achieved, further

increase of matrix strength does not increase the strength of lightweight concrete

significantly (FIP, 1983). Therefore, fly ash or slag can be used to partially

replace the Portland cement in the ULCCs, which have economical,

environmental, and technical benefits.

In the ULCCs with different pozzolanic admixtures, example, both silica

fume and fly ash, these mineral admixtures may compete for Ca(OH)2 for their

pozzolanic reactions. Since silica fume has much finer particle sizes and higher

reactivity than fly ash, the Ca(OH)2 from cement hydration may be consumed by

the pozzolanic reaction of the silica fume at early age, which reduces the

alkalinity of the pore solution (Shehata and Thomas, 2002) and may affect the

pozzolanic reaction of the fly ash. Weng et al. (1997) reported that 10% of silica

fume in cement paste with a water to binder ratio (w/b) of 0.4 reduced the

Ca(OH)2 content by approximately 50% at 7 days. This indicates that the

incorporation of silica fume may reduce the pozzolanic reaction of fly ash, and

58
further increase in the fly ash replacement may not contribute to improved

mechanical properties of ULCCs.

Based on the background information above, the incorporation of silica fume

may improve the stability of cenospheres and steel fibers but may reduce the

flowability of fresh ULCCs. In addition, the incorporation of silica fume may

affect the pozzolanic reaction of fly ash. The objectives of the research described

in this chapter are (1) to study the effect of silica fume and fly ash on the

rheological parameters and flowability of the ULCCs and their influence on the

stability of cenospheres and steel fibers, and (2) to investigate the amount of fly

ash which can be incorporated in the ULCCs from the perspectives of pozzolanic

reaction and mechanical properties. The mix ULCC-4 in Chapter 2 was selected

as control mix considering the viscosity of this mix was observed as lowest in all

ULCC mixes with cenosphere QK300. Minor adjustment was made on w/c to

achieve comparable compressive strength, according to preliminary experiment.

3.2 Experimental details

In this section the materials used for ULCCs and cement pastes as well as

experimental methods are described. The test methods, relevant standards, testing

ages, specimen sizes, and specimen numbers are summarized in Table 3.1. The

mix proportions of the mixtures are presented together with the results and

discussion in Section 3.3.

59
3.2.1 Materials used

ASTM Type I Portland cement, Class F fly ash, and undensified silica fume

were used in this study. The chemical compositions of the cement, fly ash, and

silica fume are summarized in Table 3.2. The particle size distributions of the

cement and fly ash determined by a particle size analyzer are shown in Figure 3.1

and the median particle size of the cement and fly ash are 12.6 μm and 11.2 μm,

respectively. Silica fume has a specific surface area of 15 - 35 m2/g, according to

the manufacturer.

Cenospheres (QK300) with an average particle density of approximately 970

kg/m3 was used in ULCCs as micro-aggregates. Particle size distributions of the

cenospheres in comparison to that of the cement and fly ash is shown in Figure

3.1. Most of the cenosphere particles have sizes from 10 to 300 μm with a median

particle size of 160 μm.

Steel fibers 12 with the properties summarized in Table 3.3 were used in some

ULCC mixtures. The polycarboxylate based superplasticizer (SP) and a shrinkage

reducing admixture (SRA) detailed in Section 2.2.1 were also used in the ULCC

mixtures of this chapter.

3.2.2 Stability of cenospheres and steel fibers in ultra-lightweight cement

composites

Effects of silica fume and fly ash on the rheological parameters are

determined in order to evaluate their influence on the stability of cenospheres and

steel fibers in the ULCCs. In addition, the flowability of ULCCs with different
12
Dramix@ OL13/.16, Bekaert, Belgium
60
silica fume and fly ash contents is determined according to BS EN 1015-3 (1999)

test to evaluate their effect.

The dry ingredients i.e., cement, silica fume or fly ash, and cenospheres were

mixed for 5 min (at a speed of 140±5 min-1) in a Hobart mixer, followed by the

addition of water and SP and mixed for another 5 min (at a speed of 280±10 min-
1
). The rheological parameters of the ULCCs were determined according to

ASTM C1749-12 (2012) at 10 min and 30 min after the water had been added.

Before each test, the mixture was remixed in the Hobart mixer for 1 min.

A rheometer 13 with coaxial cylinder configuration was used to determine the

rheological parameters of the ULCCs. The inner cylinder has a length of 55 mm

and a diameter of 38 mm and the gap between the inner and the outer cylinders is

2.7 mm. ASTM C1749-12 (2012) suggests the gap be more than 10 times of

maximum particle size of the cement composites to avoid unexpected high shear

stress and locking. Even though the gap in this study is about 7 times of the

maximum cenospheres diameter, the unexpected high shear stress and locking

were not observed during the experiment. The surfaces of the inner and outer

cylinders are serrated so that the slippage between the ULCC samples and the

surfaces of the cylinders is minimized. During the test, the shear rate first

increased from 0.5 to 50 s-1 in 50 s and then decreased to 0.5 s-1 in another 50 s

and the corresponding shear stress at various shear rates was determined. The

descending curves (from 50 to 0.5 s-1) with 100 data points (pairs of shear stress

and shear rate) were used for analyses. The yield stress and plastic viscosity of the

13
Haake RotoVisco, Thermo Electron GmbH, Germany
61
ULCCs were determined by regression fitting the descending curve with Bingham

model (Equation 3.3).

τ = τ0 η (3.3)

Where, τ - shear stress, Pa;

τ0 - yield stress; Pa,

- shear rate, s-1;

η - plastic viscosity, Pa·s.

To study the stability of cenospheres and steel fibers in the fresh ULCCs,

plain and steel fiber reinforced ULCCs with different silica fume contents were

prepared, and they were filled into three cylinder moulds of Ø100 × 200 -mm and

two prism moulds of 100 × 100 × 400 -mm, respectively, and vibrated to remove

entrapped air. Even though the sizes of the specimens are limited, the significant

differences between the particle densities of cenospheres (~970 kg/m3) and steel

fibers (7850 kg/m3), versus cement pastes (~2000 kg/m3) makes it reasonable to

determine the segregation of cenospheres and fibers in cement composites. In

addition, the dimension of the specimens in this study is sufficient to represent the

case of slabs and horizontally precast wall panels. The power and duration of the

vibration were kept constant for various mixtures. About 24 hrs after casting,

these specimens were taken out from the moulds. The top and bottom of the three

cylinder specimens were grinded for image analysis and three pairs of digital

images from the top and bottom of the three cylinders of each mix were taken

(Figure 3.2(a)). A cross section of 100 × 100 -mm was cut from the middle of the

fiber reinforced ULCC prisms (Figure 3.2(b)) and two images of the cross

62
sections were taken from the two prisms of each mix. These digital images were

converted to binary images according to color differences of the cenospheres and

cement pastes for the plain samples or steel fibers and ULCC matrices for the

fiber reinforced samples to study the possible segregation of the cenospheres and

steel fibers from the ULCCs, respectively. No significant difference was observed

between the samples from the same mixture and the same location.

To quantitatively analyze the stability of cenospheres in the ULCCs, samples

with a thickness of approximately 15 mm were cut from the top and bottom of the

hardened cylinder specimens with a high speed saw (Figure 3.2(a)). The densities

of the cylinder specimen before cutting and the samples from the top and bottom

of the cylinder were determined with water displacement method. According to

preliminary experiment, the water absorption of the ULCCs was low. Thus, the

determination of the density would not be affected significantly.

As shown in Figure 3.2(b), two fiber reinforced ULCC prisms of 100 × 100 ×

400 -mm were cut into eight cubes of approximately 100 × 100 × 100 -mm to

enable further cutting with the high speed saw. Samples with a thickness of

approximately 15 mm were cut from the top and bottom of three cube specimens

from the middle of the prisms for density measurement. The reason for the use of

only three cube specimens was due to the low variability of the density measured.

The densities of the cube specimens and the samples from the top and bottom of

the cubes were determined.

63
 For both plain and fiber reinforced ULCCs, the average densities of the

samples from the top and bottom of specimens were reported and compared with

the average density of the cylinder and cube specimens.

3.2.3 Effect of silica fume and fly ash on the Ca(OH)2 content and properties of

hardened ultra-lightweight cement composites

Thermogravimetric analysis method was used to determine the Ca(OH)2

content in the cement pastes at various ages, which is an indicator of cement

hydration and pozzolanic reactions of silica fume and fly ash. Wang et al. (2012b)

reported limited pozzolanic reactivity of cenospheres at ambient temperature,

whereas Hanif et al. (2017) reported noticeable pozzolanic reaction which may be

due to the different sources of the cenospheres used and approximately 7% of fine

particles less than 10 µm in the study reported in (Hanif et al., 2017). The

pozzolanic reactivity of cenospheres may be affected by the composition of coals

and burning process in power plants. Considering the low amount of fine particles

in the cenospheres of this study (Figure 3.1), the pozzolanic reaction is expected

to be limited, and the analysis of the cement pastes in this section may also be

applied to the ULCC mixtures.

The fresh cement pastes were filled in 50 × 50 × 50 -mm moulds, compacted,

and placed in a fog room at temperatures of 28 - 30 oC for curing. Samples in the

interior of the cube specimens were obtained with a slow speed saw at the ages of

7, 28, 56, and 120 days. The sample was grinded to powder manually and placed

in ethanol for 24 hrs, followed by drying in an oven at 105 oC for additional 24

hrs to stop cement hydration and pozzolanic reactions. The sample was sealed in
64
plastic bag and placed in a vacuum dessicator and used for the experiment within

two days to minimize possible carbonation. A simultaneous thermal analyzer 14

was used to determine the Ca(OH)2 content and a powder sample of about 35 mg

was heated in a nitrogen environment from 30 to 950 oC at a rate of 10 oC/min.

The mass loss due to the decomposition of Ca(OH)2 was determined from the

thermogravimetric curve and used to calculate the Ca(OH)2 content, which is

expressed as mass percentages of anhydrous sample and Portland cement for

discussion. The mass of anhydrous cement paste sample was taken as that heated

to 600 oC (Scrivener et al., 2015). The Ca(OH)2 content per gram of Portland

cement was calculated by dividing Ca(OH)2 content per gram of anhydrous

sample by mass ratio of cement in total amount of cementitious materials. Two

tests were conducted for some cement paste mixtures and the results showed good

consistency. Therefore, only one test was conducted for the rest of the mixtures.

The ULCC specimens with or without steel fibers were prepared to evaluate

the effects of fly ash on their compressive strength development up to 270 days,

and elastic modulus flexural behavior, and thermal conductivity at 28 days.

Cement, silica fume, fly ash, cenospheres, water, SRA, steel fibers, and SP were

mixed in a Hobart Mixer and filled into various moulds. The specimens were

covered with plastic sheets to prevent moisture loss, and demoulded at 24 hrs

followed by curing in a fog room to 180 days and by exposure to laboratory air

after that. The density, compressive strength, elastic modulus, flexural

14
Linseis L81-II, Linseis Incorporation, USA
65
performance, and thermal conductivity of the ULCCs were determined according

to relevant standards, as summarized in Table 3.1.

3.3 Results and discussion

3.3.1 Effect of silica fume and fly ash on the rheological parameters, flowability,

and stability of fresh ultra-lightweight cement composites

To determine the effect of silica fume and fly ash on the rheological

parameters, flowability, and stability of cenospheres and steel fibers, ULCC

mixtures with various contents of silica fume (0%, 4%, 8%, and 12%) and fly ash

(0%, 26%, and 40% by mass of total cementitious materials) as cement

replacement were prepared and evaluated. These ULCCs had a w/b of 0.41, an SP

dosage of 0.3% by mass of total cementitious materials, and a cenosphere content

of 30% by volume selected based on preliminary experiments to achieve

sufficient flowability for experiment using the rheometer.

The shear stress versus shear rate curves of the fresh ULCCs with various

silica fume and fly ash contents are presented in Figure 3.3 with their yield stress,

plastic viscosity, and flowability summarized in Table 3.4. It is observed that

Bingham model can describe the relations between the shear stress and shear rate

of the ULCCs very well with R2 over 0.99. Negative yield stress was observed for

the control cement paste and cement pastes with fly ash due to their slight shear

thickening behavior. Negative yield stresses have been widely observed when

fitting shear stress versus shear rate curves with Bingham model (Cyr et al., 2000;

Feys et al., 2008; Larrard et al., 1998). It is believed that the shear thickening is

66
caused by the formation of reversible hydro-clusters due to great hydrodynamic

forces among particles of cementitious materials (Feys et al., 2009; Maranzano

and Wagner, 2001).

At 10 min after the addition of water, the yield stress of the ULCCs was

increased with the increase in silica fume content from 0% to 12% (Figure 3.4 and

Table 3.4). However, the plastic viscosity was only increased with the silica fume

content from 0 to 8%, but no further increase from 8% to 12%. The results at 30

min show similar trends except that no result could be obtained from the

rheometer test for the ULCC with 12% silica fume (U-SF12) due to loss of

workability with shear stress exceeded the testing range of the rheometer. The

increased yield stress and plastic viscosity of the ULCCs with silica fume might

be attributed to small particle sizes and large surface area of the silica fume which

might have increased adsorption of SP which subsequently increased “gelation”

(formation of a continuous three-dimensional network) and resistance to flow

(Nehdi et al., 1998; Struble et al., 1998). Nehdi et al. (1998) investigated the

effect of silica fume on the rheology of concrete and reported that silica fume can

adsorb multi layers of SP molecules on its surface and reduce the dispersing

efficiency of SP. The above may also explain the more significant increase in the

yield stress of the mixtures with silica fume from 10 min to 30 min than that of

the control mixture without silica fume. Compared with the yield stress, the

increase in the plastic viscosity from 10 min to 30 min seems to be less significant

because the plastic viscosity is less affected by the “gelation” which can be easily

broken by shear (Struble et al., 1998).

67
 The incorporation of fly ash in the ULCCs did not affect the yield stress and

plastic viscosity significantly (Figure 3.3(b) and Table 3.4) due to its similar

particle size distributions to that of the cement (Figure 3.1), and their yield stress

increased only slightly from 10 min to 30 min after the addition of water.

The effect of silica fume and fly ash on rheological parameters of fresh

ULCCs seems to be consistent with that for cement paste (Park et al., 2005) but

different from that of concrete (Banfill, 2006) reported in literature, which may be

explained by the small sizes of the cenospheres in the ULCCs.

Since silica fume had significant effect on the rheological parameters of the

ULCCs, its effect on the stability of cenospheres and steel fibers in the ULCCs

was studied. The density of the samples from the top and bottom of the specimens

in comparison to the density of the corresponding cylinder and cube specimens is

presented in Figure 3.5(a) and (b).

The differences between these densities of the plain ULCC specimens with

and without silica fume were within 5% (Figure 3.5(a)). By examining and

comparing the images of the cross sections of the ULCCs from the top and bottom

of the cylinders, no apparent segregation of cenospheres was observed with naked

eyes. These indicate that the cenospheres are reasonably stable in the ULCCs with

and without silica fume which may be explained by the small sizes of the

cenosphere particles even though their densities are substantially lower than that

of the paste matrices.

To investigate the stability of steel fibers in the ULCCs, mixtures were

designed by incorporating 0.5% steel fibers by volume of the ULCCs based on the

68
mixtures shown in Table 3.4. According to the results in Figure 3.5(b), the density

of the sample from the bottom of the ULCC without silica fume was

approximately 12% higher than the density of the cube specimen, whereas the

density of the sample from the top was approximately 10% lower than the density

of the cube specimen. With the incorporation of 4% silica fume, the density

differences between the samples from the top and bottom of the ULCC and the

cube specimen were decreased to approximately 2% indicating improved stability

of the steel fibers. Increase in the silica fume content from 4% to 12% did not

affect the density differences significantly. Based on binary images of the cross

sections of the specimens, severe segregation of the steel fibers was observed in

the specimen without silica fume in which the steel fibers were sank to the bottom

as shown in Figure 3.6. However, no apparent segregation was observed in the

specimens with silica fume regardless of the silica fume content from 4% to 12%

which may be explained by increased yield stress and plastic viscosity of the

matrices discussed earlier. The results indicate that silica fume content of at least

4% is essential in these steel fiber-reinforced ULCCs.

The flowability of the ULCCs appears reduced with the increase in the silica

fume content from 0 to 12% due to the increased yield stress and plastic viscosity.

However, the ULCCs with 8% and 12% silica fume had flowability of 200 mm

and 175 mm, respectively, which were sufficient to cast specimens without

difficulty. In addition, it was observed in preliminary experimentation that the

ULCC without silica fume tends to be harsh and the cohesiveness was improved

with the increase of silica fume content. The flowability of the mixtures with fly

69
ash was beyond the testing range of the flow table. However, the insignificant

effect on the yield stress and plastic viscosity indicates that fly ash content may

not have significant effect on the flowability of the ULCCs.

Since the incorporation of silica fume reduces the segregation and improved

the stability of steel fibers, it is an essential ingredient in the steel fiber-reinforced

ULCCs. However, the amount of silica fume needs to be optimized by

considering both the stability of steel fibers and flowability of fresh ULCCs.

Since the rheological behavior of ULCCs are also affected by other factors

such as w/b and cenosphere content, silica fume content required for stability of

steel fibers in ULCCs may vary with different mix proportions.

3.3.2 Effect of silica fume and fly ash on Ca(OH)2 content of cement pastes

Cement pastes with a w/b of 0.41 and various contents of silica fume (0% and

8% by weight of total cementitious materials) and fly ash (0%, 26%, and 40%) as

cement replacements were prepared to investigate effects of silica fume and fly

ash on Ca(OH)2 (CH) content of cement pastes. Although the incorporation of 4%

silica fume can eliminate the segregation of steel fibers in the fiber-reinforced

ULCC as discussed in Section 3.3.1, 8% of silica fume was selected in the

mixtures from the conservative consideration of the cohesiveness and stability.

The CH content was determined at 7, 28, 56, and 120 days, and the results are

presented in Table 3.5. The CH in the cement pastes with silica fume, fly ash, or a

combination of both reflects two processes in the pastes with cement hydration

generating CH and pozzolanic reaction consuming CH.

70
 The control cement paste (CP) contained 15.6% CH at 7 days, and the CH

content increased to 20.6% at 56 days with no further increase at 120 days. The

incorporation of silica fume, fly ash, or a combination of both reduced the CH

content compared with that of the control cement paste at various ages as

expected due to pozzolanic reactions of silica fume and fly ash as well as dilution

effect by reduced cement content.

To exclude the dilution effect, the CH content was normalized to cement

basis (Table 3.5), and the effects of silica fume and fly ash on the CH content are

presented in Figure 3.7(a) and (b), respectively. The cement paste with silica fume

(CP-SF8) showed lower CH content than that without silica fume (CP) at various

ages from 28 to 120 days (Figure 3.7(a)). Even though silica fume has finer

particle sizes and higher pozzolanic reactivity, leading to substantial pozzolanic

reaction before 7 days (Lothenbach et al., 2011), the CH content of the cement

paste with silica fume had comparable or slightly lower CH content than that in

the control mix (Table 3.5). This may be explained by the possible acceleration of

cement hydration at early age by silica fume due to its fine particle size which

produces more CH than the control without silica fume (Lothenbach et al., 2011).

The cement pastes with fly ash (CP-FA26 and CP-FA40) have comparable

CH content at 7 days, but lower CH contents at 28, 56, and 120 days, in

comparison to those of the control cement paste (CP) (Figure 3.7(b)), indicating

that the pozzolanic reaction of fly ash mainly occurred after 7 days. As expected,

the higher the fly ash content, the lower the CH content at 28, 56, and 120 days

due to more CH consumed by pozzolanic reaction.

71
 The CH content, on cement basis, in cement paste with both silica fume and

fly ash was higher than combined CH content in the paste with silica fume or fly

ash alone at the age of 56 and 120 days. For example, the incorporation of 8% of

silica fume reduced the CH content by 9.0% (20.3 - 11.3%), and the incorporation

of 26% fly ash reduced the CH by 9.6% (20.3 - 10.7%) at 120 days (Table 3.5).

Assuming that the pozzolanic reactions of silica fume and fly ash be independent,

the reduction of CH by incorporating 8% silica fume and 26% fly ash is expected

to be 18.6% (9.0 + 9.6%) with residual CH of 1.7% (20.3 - 18.6%) in the paste.

This estimated value of 1.7% is lower than 5.5% CH experimentally determined

in the paste mixture CP-SF-FA26. The higher CH in the CP-SF-FA26 might be

explained by reduced pozzolanic reaction of fly ash in the paste with silica fume.

Since silica fume has substantially smaller particle sizes and higher pozzolanic

reactivity than fly ash, the pozzolanic reaction of fly ash occurs after that of silica

fume as mentioned earlier. Thus, there is less CH available in the system for the

pozzolanic reaction of fly ash, and the lower alkalinity of pore solution also

reduces the pozzolanic reactivity of fly ash (Lothenbach et al., 2011). This

indicates that the pozzolanic reaction of fly ash was reduced by the incorporation

of silica fume. The same applies to the paste with 8% silica fume and 40% fly ash

at 56 and 120 days.

After 120 days of moisture curing, there was only 3.6% and 2.2% CH

remained in the paste samples CP-SF-FA26 and CP-SF-FA40, respectively, on

anhydrous sample basis. Since silica fume is necessary for the stability of steel

fibers as discussed in Section 3.3.1, the above results indicate that further increase

72
in fly ash content beyond certain extent may not lead to significant increase in the

pozzolanic reaction and calcium silicate hydrates in the paste and thus

performance of the corresponding ULCCs.

It is noted that the CH content in the cement pastes and the effect of mineral

admixtures are affected by w/b and characteristics of the mineral admixtures (e.g.

particle size distribution and amorphous silica content). Further research is needed

when considering of ULCCs with different w/b.

3.3.3 Effect of fly ash on properties of hardened plain and fiber-reinforced ultra-

lightweight cement composites

The incorporation of 26 and 40% fly ash has insignificant effect on the

thermal conductivity of ULCCs, whereas the steel fibers increase the thermal

conductivity of the ULCCs (Table 3.6).

Effect of fly ash content up to 40% by mass of cementitious materials on the

mechanical properties of plain and fiber-reinforced ULCCs with 8% of silica

fume is evaluated, and the results are summarized in Table 3.6. The mix

proportions and workability of these ULCCs are given in Table 3.7 with w/b and

cenosphere volume controlled at 0.41 and 47%, respectively, to achieve flowable

ULCCs with density of about 1250 kg/m3. The SRA dosages were 2.5% and 1.5%

(by mass of cementitious materials) for the ULCCs with and without steel fibers,

respectively. The higher dosage of SRA was used to reduce the entrapped air

content in the fiber-reinforced ULCCs. The workability in terms of flow value

determined according to BS EN 1015-3 (1999) is from 180 to 200 mm for the

plain ULCCs and 145 to 150 mm for the fiber-reinforced ULCCs. Although the
73
fiber-reinforced ULCCs had lower flow values than those of the corresponding

plain ones, they could be compacted without difficulty.

The plain ULCCs had densities from about 1205 to 1250 kg/m3,

approximately 50% lower than that of normal weight concrete. The fiber-

reinforced ULCCs had higher densities than the plain ones due to the

incorporation of steel fibers. In general, the incorporation of fly ash reduced the

density of the ULCCs due to the lower density of the fly ash than that of the

Portland cement.

Figure 3.8 shows the compressive strength development of the ULCCs cured

in moist condition up to 180 days followed by exposure to laboratory air from 180

to 270 days. The incorporation of steel fibers did not have significant effect on the

compressive strength of the ULCCs. The compressive strength of the ULCCs

decreased with the increase in the fly ash content, especially at 3 and 7 days, as

expected. With 26% and 40% fly ash, the 28-day compressive strength of the

plain ULCCs was 14% and 27% lower than that of the control one, respectively,

and the 28-day compressive strength of the corresponding steel fiber reinforced

ULCCs was 13% and 18% lower than that of the corresponding control without

fly ash, respectively. Although the strength decreased with the increase in the fly

ash content, the ULCCs with 26% fly ash achieved compressive strength of about

40 MPa at 28 days with a Portland cement content of about 460 kg/m3, reasonable

in comparison to that in conventional concrete. At 180 days the ULCCs with 26%

fly ash (ULCC-FA26 and ULCC-FA26-ST) achieved compressive strengths

comparable to those of the control ones without fly ash (ULCC-control and

74
ULCC-control-ST), whereas those with 40% fly ash (ULCC-FA40 and ULCC-

FA40-ST) still had 10-20% lower strengths due to their lower cement contents. It

is also observed that the compressive strength of the ULCCs with 40% fly ash

increased by approximately 12-16% from 28 to 180 days, even though the

Ca(OH)2 content in the corresponding cement pastes (Section 4.2.2) was reduced

by only about 1% on anhydrous sample basis from 28 to 120 days with no

significant change from 56 to 120 days. This may be explained by the fact that the

strength is affected not only by the Ca(OH)2 content due to pozzolanic reaction of

the mineral admixtures, but also by the consequent formation of additional

calcium silicate hydrates, refinement of pore structure, and improvement of the

interface and bonding between cenospheres and matrices. The increase in the

compressive strength of the ULCCs from 180 to at 270 days might be attributed

partly to the drying of the specimens.

At 28 days, the incorporation of 26% fly ash did not have significant effect

on the elastic modulus of the plain ULCC, whereas the increase in the fly ash

content to 40% decreased the elastic modulus of ULCC by 13%. For the ULCCs

with steel fibers, the incorporation of 26% and 40% fly ash reduced the elastic

modulus by about 15%. According to ACI 318M-11 (2011) and FIP manual of

lightweight aggregate concrete (1983), elastic modulus of lightweight concrete

correlates to its density and compressive strength. The incorporation of fly ash in

ULCCs may lead to decrease in elastic modulus. ACI 232.2R-02 “Use of fly ash

in concrete” (2002), reported that fly ash played a minor role in the elastic

modulus of concrete compared with characteristics of cement and aggregate.

75
Therefore, more results may be needed for statistical analysis to further elaborate

the effect of fly ash on the elastic modulus of ULCCs.

The flexural strength of the ULCC with or without steel fibers was decreased

with the increase in the fly ash content. The steel fibers increased the flexural

tensile strength and offset the strength reduction due to fly ash to certain extent

depending on the fly ash content. The fiber-reinforced ULCC with 26% fly ash

had post-cracking behavior comparable to that of the control ULCC, even though

its flexural tensile strength was lower (Figure 3.9). The post-cracking behavior of

fiber reinforced cement composites is mainly determined by the bonding between

fibers and matrices (Bentur and Mindess, 2006). The ULCC with 26% fly ash

may have more homogeneous distribution of hydration products and less Ca(OH)2

at the interface between the steel fibers and the matrix and better bonding than

that of the control mixture without fly ash due to the lower rate of cement

hydration and pozzolanic reaction by the incorporation of fly ash. This may offset

the weakening of the matrix in the ULCC with 26% fly ash. Further increase in

the fly ash content to 40% reduced both flexural strength and post cracking

behavior.

3.4 Summary and conclusions

This chapter investigates the effect of silica fume and fly ash on the stability

of cenospheres and steel fibers and flowability of fresh ULCCs and the amount of

fly ash which can be used to replace Portland cement in ULCCs with silica fume

from the perspectives of pozzolanic reactions and strength development. Based on

the discussion above, the following conclusions can be drawn:

76
 1. The yield stress and plastic viscosity of fresh ULCCs generally increased

with the increase of silica fume content from 0% to 12% by the mass of

cementitious materials, whereas the incorporation of fly ash by 26% and 40% did

not affect these two rheological parameters of the fresh ULCCs.

2. No segregation of cenospheres was observed in the ULCCs with and

without silica fume, whereas severe segregation of steel fibers was observed in

the ULCC without silica fume. Silica fume content of at least 4% is essential in

reducing the segregation and improving the stability of steel fibers in the ULCCs.

Even though the flowability is reduced with the increase in silica fume content,

the ULCCs with up to 12% silica fume had sufficient flowability to cast

specimens without difficulty.

3. The Ca(OH)2 content in cement paste with both silica fume and fly ash, on

cement basis, was higher than combined Ca(OH)2 content in the paste with silica

fume or fly ash alone at the age of 56 and 120 days due to reduced Ca(OH)2

available and possibly lower alkalinity of pore solution in the former. After 120

days of moisture curing, there were 3.6% and 2.2% Ca(OH)2 remained in the

paste mixtures with both 8% silica fume and 26% or 40% fly ash, respectively, on

anhydrous sample basis. This indicates that further increase in fly ash content

beyond certain extent may not lead to significant increase in the pozzolanic

reaction of the fly ash.

4. Incorporation of fly ash reduced the density, compressive and flexural

strengths, and elastic modulus of the ULCCs with given silica fume content at 28

days in general. However, the ULCCs with 26% fly ash and 8% silica fume

77
achieved 28-day compressive strength of 40 MPa and density of < 1250 kg/m3

with a reasonable cement content of about 460 kg/m3. In addition, the fiber-

reinforced ULCC with 26% fly ash had post cracking performance under bending

comparable to that of the control without fly ash.

Based on the above, optimization of the silica fume and fly ash contents is

needed by considering the flowability, stability of steel fibers, pozzolanic

reactions, and strength development of the ULCCs.

78
 3.5 Tables

Table 3.1 Test methods and specimen information

Specimen
Property Test standard Testing age
Type and size Number
Workability (flow table) (ULCCs Right after
BS EN 1015-3 (1999) --- ---
and cement pastes) mixing

10 & 30 mins
Rheological parameters ASTM C1749-12 --- 1*
after mixing

Image of cross section -- 1 day --- 3

Density of hardened specimens

BS EN 12390-7 (2009) 1 day Cube, 100×100×100 mm 6
after demoulding
7, 28, 56, & 120
Ca(OH)2 content --- --- 1*
days
3, 7, 28, 90, 180
Compressive strength BS EN 12390-3 (2009) Cube, 100×100×100 mm 3
& 270 days
Flexural performance ASTM C1609-07 (2007) 28 days Prism,100×100×400 mm 3

Elastic modulus ASTM C469 (2014) 28 days Cylinder, Ø100×200 mm 3

Slab,
Thermal conductivity ASTM C518-10 ( 2010) About 100 days* 2
300×300×50 mm
* Two samples were selected in preliminary study and the difference is within 5%, therefore, one sample was selected
for this study.

79
Table 3.2 Chemical and mineral compositions of cement, fly ash, and silica fume (% by mass) from the manufacturers

CaO SiO2 Al2O3 Fe2O3 MgO Na2O K2O SO3 LOI C3S C2S C3A C4AF

Portland
63.8 20.4 4.9 3.6 4.3 0.3 0.4 2.3 1.5 60.1 13.2 6.9 11.0
cement
Fly ash 3.9 46.3 28.5 18.5 1.8 0.2 0.6 0.2 2.3 NA NA NA NA
Silica fume NA 87.8 NA NA NA 0.3 0.9 0.5 3.0 NA NA NA NA
NA: not available

Table 3.3 Properties of steel fibers provided by manufacturer

Tensile strength, Elastic modulus, Diameter, Aspect Elongation, % Density,

Length, mm
MPa GPa µm ratio kg/m3
Steel fibers 2500 200 13 150 87 < 3.5 7850

80
 Table 3.4 Effect of silica fume and fly ash content on rheological properties of ULCCs (Bingham model) *

10 min after adding water 30 min after adding water

Silica fume content by Fly ash content by
Mix ID Flowability, mm
cm mass, %† cm mass, % Yield stress, Plastic viscosity, Yield stress, Plastic viscosity,
Pa Pa·s Pa Pa·s
U-control 0 0 -0.2 1.06 3.8 1.11 >300

U-SF4 4 0 20.5 2.21 57.4 2.50 >300

U-SF8 8 0 105.0 3.06 138.0 3.13 200
U-SF12 12 0 158.0 3.03 NA NA 175
U-control 0 0 -0.2 1.06 3.8 1.11 >300
U-FA26 0 26 -1.3 0.98 4.8 1.30 >300
U-FA40 0 40 -2.0 0.89 1.5 1.14 >300
*
w/b=0.41, Volume of cenospheres=30%, SP/cm=0.3%
†
cm: cementitious materials

81
 Table 3.5 Ca(OH)2 content of cement paste with different content of fly ash and silica fume by thermogravimetric analysis

Ca(OH)2 content per gram of anhydrous Normalized cementitious Ca(OH)2 content per gram of
Cementitious materials mass ratio
cementitious materials, % materials mass ratio† cement, %*
Variable Cement Fly ash Silica fume 7d 28d 56d 120d Cement Fly ash Silica fume 7d 28d 56d 120d
CP 100 0 0 15.6 19.0 20.6 20.3 100 0 0 15.6 19.0 20.6 20.3
Silica fume
CP-SF8 92 0 8 13.4 13.4 11.8 10.4 100 0 9 14.6 14.6 12.8 11.3
CP 100 0 0 15.6 19.0 20.6 20.3 100 0 0 15.6 19.0 20.6 20.3
CP-FA26 Fly ash 74 26 0 12.2 11.9 10.1 8.0 100 35 0 16.5 16.1 13.7 10.7
CP-FA40 60 40 0 9.2 5.7 2.7 2.4 100 67 0 15.4 9.6 4.6 4.0
CP-SF8 92 0 8 13.4 13.4 11.8 10.4 100 0 9 14.6 14.6 12.8 11.3
Fly ash +
CP-SF8-FA26 66 26 8 7.5 7.4 5.9 3.6 100 39 12 11.4 11.2 8.9 5.5
silica fume
CP-SF8-FA40 52 40 8 4.4 3.3 2.0 2.2 100 77 15 8.4 6.4 3.8 4.2
†
equals to columns “cementitious materials mass ratio” divided by cement percentage in total cementitious materials
*
equals to columns “Ca(OH)2 content per gram of anhydrous cementitious materials” divided by cement percentage in total cementitious materials

Table 3.6 Density, mechanical properties, and thermal conductivity of various ULCCs

Fly ash Steel fibers Compressive strength, MPa Thermal

Flexural
content content by 1d density, conductivity,
Mix ID E, GPa strength,
by cm total kg/m3 3d 7d 28d 90d 180d 270d* W/m·K
MPa
mass, % vol, %
ULCC-control 0 0 1251 (5) 35.9 (2.1) 42.0 (2.2) 46.9 (2.3) 48.2 (1.7) 48.6 (1.9) 50.9 (2.0) 12.1 (1.5) 4.1 (0.1) 0.34

ULCC-FA26 26 0 1223 (5) 24.2 (0.6) 31.2 (3.1) 40.5 (3.7) 44.5 (0.5) 47.1 (2.7) 48.8 (1.1) 12.0 (0.4) 3.7 (0.6) 0.33
ULCC-FA40 40 0 1205 (6) 18.0 (0.7) 26.0 (0.4) 34.4 (2.3) 39.3 (2.6) 39.8 (1.3) 41.0 (3.6) 10.5 (1.0) 2.4 (0.2) 0.33
ULCC-control-
0 0.5 1277 (11) 35.6 (0.2) 43.2 (1.4) 46.2 (1.9) 49.0 (1.5) 47.4 (1.6) 50.7 (3.0) 14.1 (0.6) 5.3 (0.5) 0.37
ST
ULCC-FA26-ST 26 0.5 1242 (7) 26.2 (0.7) 34.0 (2.3) 40.2 (0.3) 42.8 (2.7) 46.4 (1.3) 47.6 (2.7) 12.3 (0.2) 4.3 (0.1) 0.36

ULCC-FA40-ST 40 0.5 1218 (4) 19.5 (0.4) 27.8 (0.7) 37.7 (0.5) 39.5 (0.5) 42.4 (1.9) 44.5 (3.1) 11.9 (0.5) 3.0 (0.3) 0.36
*
cured in moist fog room from demoulding to 180 days and cured in laboratory air from 180 to 270 days; the values included in the bracket refer to
standard deviation of the testing results, while the values behind “±” refer to testing range.
82
 Table 3.7 Mix Proportions and flowability of ULCCs*

Water Cement Fly ash Silica fume ST fiber Cenosphere SRA SP Flowability
Fly ash content by Steel fibers content
Mix ID
cm mass, % by total vol, % kg /m3† L/ m3† mm

ULCC-control 0 0 283 660 0 57 0.0 455 10.8 6.2 200

ULCC-FA26 26 0 274 461 178 56 0.0 455 10.4 6.2 180

ULCC-FA40 40 0 269 355 272 55 0.0 455 10.2 5.9 200

ULCC-control-ST 0 0.5 281 654 0 57 39.3 455 17.8 5.1 145

ULCC-FA26-ST 26 0.5 265 456 176 55 39.3 455 17.2 4.6 150

ULCC-FA40-ST 40 0.5 260 352 270 54 39.3 455 16.9 5.1 150
*
w/b=0.41, silica fume content is 8% by cementitious materials (cm) mass, Volume of cenospheres=47%
†
for 1m3 of all materials added.

83
 3.6 Figures

Figure 3.1 Particle size distributions of the cement, fly ash, and cenospheres used

84
Figure 3.2 Sketch of samples used to investigate the stability of (a) cenospheres
in plain ULCCs and (b) steel fibers in fiber reinforced ULCCs (not to scale)

85
 Figure 3.3 Shear stress versus shear rate of ULCC with different silica fume (a) and fly ash
(b) content
- the results of cement paste with 12% silica fume after adding water of 30 min were not
obtained because it exceeded the detectable range of rheometer.

Figure 3.4 Effect of silica fume on yield stress and plastic viscosity of ULCCs

86
 Figure 3.5 Comparison of the densities of the samples from the top and bottom of (a) plain
and (b) fiber reinforced ULCC specimens with the density of the specimen before cutting.
(The density of the specimen before cutting is considered as 1)

Figure 3.6 Binary images of cross sections of fiber reinforced prism (100 × 100-mm) without
incorporation of silica fume (black dot-steel fibers; blue-ULCC matrices)

87
Figure 3.7 Ca(OH)2 (CH) content of cement pastes with different contents of silica fume and
fly ash

Figure 3.8 Compressive strength of ULCCs with fly ash and steel fibers at different ages

88
Figure 3.9 Flexural performance of fiber reinforced ULCCs with different fly ash content

89
 Chapter 4. Using photocatalytic coating to maintain solar

reflectance and lower cooling energy consumption of buildings

4.1 Introduction

Solar reflectance is the ability of a material to reflect incoming solar radiation

by its surface under exposure to sunlight (Synnefa and Santamouris, 2013). It is a

surface property on a scale of 0 to 1, with “1” indicating that all the solar energy

striking a surface is reflected back to the atmosphere and “0” indicating that none

is reflected.

Building surfaces with high solar reflectance reduce the heat transfer through

building envelope and cooling energy consumption, increasing the indoor thermal

comfort of building without air-conditioning, and extending the service life of

building surfaces by reducing surface temperature (Department of Energy, 2007b).

Studies demonstrate that increasing the solar reflectance of the rooftops from 0.1 -

0.2 to about 0.6 leads to more than 20% reduction of the cooling cost for

buildings in California and Florida, which translates to savings of more than one

billion USD per year in the United States (Kolokotsa et al., 2013). Further, high

solar reflectance of building surfaces benefits the environment by reducing local

air temperature, lowering peak electricity demand which helps prevent power

outages, and reducing emission of CO2 and SO2 from power plants. Increasing the

solar reflectance of roofs and pavements world-wide by 0.1 is equivalent to

reduce 44 Gt of CO2 emission (Akbari et al., 2009).

90
 A summary of commonly used roofing materials and their solar reflectance

are presented in Table 4.1 (Parker et al., 1993). As one of the most widely used

materials for opaque building envelope, conventional concrete has a solar

reflectance of 0.4 - 0.5, depending on the solar reflectance of ingredient materials

used, their relative proportions, and exposure condition (Levinson and Akbari,

2002; Marceau and Vangeem, 2008). Solar reflectance of the concrete surface can

be increased by increasing the solar reflectance of the mix components except

coarse aggregate, such as replacing ordinary Portland cement with white cement.

It is noted that the solar reflectance of coarse aggregate has insignificant effect on

the solar reflectance of concrete because coarse aggregate is not exposed to

surface in most cases (Levinson and Akbari, 2002). Studies demonstrated that

solar reflectance of concrete increased with the curing ages from one week to six

weeks and remained consistent afterwards (Levinson and Akbari, 2002; Marceau

and Vangeem, 2008), which may be explained by the hydration reaction of

cement (Marceau and Vangeem, 2008). Wet surface has lower solar reflectance

than dry surface because changing the medium surrounding the particles from air

to water decreases their relative refractive index and increases the absorption of

incident photons (Twomey et al., 1986). The solar reflectance of building surfaces

can also be increased by various coatings (Levinson et al., 2010; Qin et al., 2014;

Synnefa and Santamouris, 2013; Synnefa et al., 2007a). In 2010, a “White Roof

Project” was initiated in New ork where white coatings were applied on the

rooftop of buildings to increase the solar reflectance and reduce cooling energy

consumption of buildings (White roof project, 2016).

91
 The solar reflectance of building surfaces may decrease with time due to

weathering and soiling when exposed to ambient environment (Cheng et al., 2012;

Levinson et al., 2005; Synnefa and Santamouris, 2013). Weathering refers to the

exposure to moisture, ultraviolet (UV) radiation and diurnal temperature cycles,

while soiling is caused by the deposition of particulate matters and

microbiological growth (Sleiman et al., 2014b). Compared to weathering, soiling

plays more important role in decreasing solar reflectance of building surfaces

(Favez et al., 2006; Sleiman et al., 2014b). Among various soiling pollutants,

black carbon plays a major role in the darkening and decreasing in the solar

reflectance of building surfaces because it absorbs solar radiation strongly in the

wavelengths from 280 nm to 2500 nm (Berdahl et al., 2002; Favez et al., 2006;

Sleiman et al., 2014b). Black carbon, refers to light-absorbing refractory

carbonaceous matter, is released during the combustion of fossil fuels and

biomass from various industrial and residential sources (Buseck et al., 2012).

Favez et al. (2006) studied the composition of soiling pollutants on glass surfaces

exposed to climate conditions of six cities in Europe and found that the deposited

pollutants was comprised of dust mineral (28-66%), soluble salts (18-27%),

organic matter (8-36%), and black carbon (4-12%). Laboratory experimental

results by Sleiman et al. (2014b) demonstrated that solar reflectance of a roofing

membrane linearly decreased with the increase of black carbon loading on the

surface and 18% reduction in solar reflectance was observed at the black carbon

loading of 0.8 μg cm2. Experimental results reported in literature (Levinson et al.

2015) revealed that rinsing was not able to completely remove the black carbon

92
deposited on roofing surfaces. Further, several buildings bear testimony that the

BC deposited on building surfaces cannot be completely removed by rain

considering that darkening and reduction of solar reflectance are widely observed

on building surfaces in various cities around the world (Paolini et al. 2014;

Levinson et al. 2015; Kelen et al. 2015).

Relevant laboratory and field studies summarized in Table 4.2 (Kelen et al.,

2015; Levinson et al., 2005; Paolini et al., 2014; Shi et al., 2013; Takebayashi et

al., 2016; Xue et al., 2015) demonstrate the reduced solar reflectance when

exposed to ambient environment and simulated soiling processes. For example,

Levinson et al. (2005) studied the effect of soiling on the solar reflectance of 15

roof membranes from 10 cities in the United States and found 11-59% reduction

of solar reflectance after five to eight years of outdoor exposure. Takebayashi et al.

(2016) studied the effect of soiling on the solar reflectance of rooftops with

coatings from different cities in Japan, and found up to 30% reduction of solar

reflectance after four years of exposure.

To better evaluate the thermal performance of building envelope, “aged solar

reflectance”, which is defined as the solar reflectance after e posure to ambient

environment, has been widely used. For e ample, the “2016 Building Energy

Efficiency Standards for Residential and Nonresidential Buildings” from

California Energy Commission (2015) requires that the 3-year “aged solar

reflectance” be higher than 0.63 for roofing products (e.g. coatings and

membranes) on low sloped roofs of non-residential buildings. A linear equation

(Equation 4.1) has been proposed to estimate the “aged solar reflectance” of

93
building surfaces exposed to soiling (Sleiman et al., 2011). It is observed from the

equation that the “aged solar reflectance” of a building surface is affected by the

nature of soiling agents, solar reflectance before soiling, and the resistance to

soiling of building surfaces.

- (4.1)

Where, - aged solar reflectance;

soil - solar reflectance of thick opaque soiled surface, e.g. 0.20

adopted by “2016 Building energy efficiency standards for residential and

nonresidential buildings” by California Energy Commission (2015);

- resistance to soiling, 0 < <1;

new - solar reflectance value before soiling.

However, many coatings reported in literature and available in the market

may not be able to meet this requirement without regular and frequent

maintenance, e.g. washing or bleaching (Table 4.2). Experimental results

indicated the reduction of solar reflectance mainly occurs within the first year or

even first two months after installation (Bretz and Akbari, 1997). This means

annual or even seasonal cleaning the building surfaces is necessary to maintain

the satisfactory solar reflectance, which may not be cost effective (Bretz and

Akbari, 1997).

It is observed from the Equation 4.1 that the “aged solar reflectance” can be

increased by the increase in the resistance to soiling. Therefore, research has been

conducted to develop coatings with high resistance to soiling for energy efficient

buildings (Aoyama et al., 2016; Levinson and Hunter, 2014; Shi et al., 2013;

94
Sleiman et al., 2014a). For example, “Stay Clean” roof coatings with high “dirt

pickup resistance” have been developed by incorporating solid or hollow beads

and chemical additives (Sleiman et al., 2014a). However, approximately 20%

reduction of the solar reflectance was observed on these coatings after exposure to

a simulated soiling process (Sleiman et al., 2014a).

Photocatalytic coatings with photocatalysts (e.g. TiO2, ZnO, and ZrO2) have

been developed in recent decades which can remove particulate pollutants

deposited on the building surfaces (Krishnan et al., 2013b; Smits et al., 2013).

Photocatalytic coatings generally consist of organic or inorganic binders,

photocatalysts, and additives. Considering the organic binders may be

decomposed by photocatalytic reaction (Geiss et al., 2012), inorganic binders are

preferred in photocatalytic coatings (Krishnan et al., 2013a; Krishnan et al.,

2013b). Among various photocatalysts, TiO2 is the most widely used due to its

high photocatalytic reactivity, stability, and low cost (Chen and Poon, 2009). TiO2

mainly comes in two crystalline forms: anatase and rutile. Experimental studies

have shown that TiO2 in anatase phase has higher photoactivity than rutile but a

combination of both presents higher photoactivity than that of pure anatase (Sun

and Xu, 2010).

The reaction efficiency of photocatalytic TiO2 is affected by various factors,

such as particle size, light irradiation, temperature, and humidity (Krishnan, 2015).

Jiang et al. (2008) studied the effect of particle size on the reactivity of TiO 2 and

reported the highest activity for 30 nm particles and the reactivity remained

constant with a particle size higher than 30 nm. The photocatalytic activity is

95
decreased with the decrease of particle size from 30 to 10 nm and less affected

with further decrease of particle size (Jiang et al., 2008). Studies reported that the

photocatalytic degradation rate of TiO2 increased with the increase in light

intensity (Puddu et al., 2010; Yamazaki et al., 1999). Temperature and humidity

may have dual influence on the photocatalytic activity of TiO 2 depending on the

reaction mechanism (Krishnan, 2015).

The degradation of organic matters such as toluene and rhodamine B and

inorganic gases such as SO2 and NOx by photocatalytic TiO2 has been widely

studied (Krishnan et al., 2013a; Krishnan et al., 2013b; Ramirez et al., 2010). It

has also been observed and reported that photocatalytic coatings with TiO2 are

able to remove black carbon deposited on the surfaces of mortar and glass

(Krishnan, 2015; Lee and Choi, 2002; Mills et al., 2006; Smits et al., 2013; M.

Smits et al., 2012). For example, Smits et al. (2013) investigated the degradation

of black carbon (25 μg cm2) deposited on mortar surface with a photocatalytic

coating containing appro imate 250 μg cm2 of TiO2, and experimental results of

image analyses and infrared absorption spectra indicated that the black carbon

was completely degraded after 300 hrs of exposure to simulated solar irradiation.

Based on the discussion above, it is reasonable to hypothesize that the

photocatalytic coating with TiO2 be able to remove black carbon, self-clean

building surfaces, maintain solar reflectance, and reduce cooling energy

consumption. Therefore, the ability of a photocatalytic coating with TiO2 to

maintain the solar reflectance of mortar surfaces and lower cooling energy

consumption by removing black carbon is experimentally studied in this chapter.

96
4.2 Experimental details

Experiments were designed and conducted with three types of coatings

applied on mortar specimens which represent typical wall and roof surfaces: (1)

transparent silicate coating, (2) white silicate coating, and (3) white silicate

coating incorporating photocatalyst. The specimens with these coatings have

comparable surface texture and porosity. The black carbon was used as surrogate

of atmospheric black carbon in particulate pollutants deposited on building

surfaces and the photocatalytic removal of the black carbon was achieved through

simulated solar irradiation.

4.2.1 Materials, specimens, and coatings

Mortar specimens were made by normal Portland cement and natural sand

detailed in Chapter 2. The materials were mixed in a Hobart mixer and filled into

plastic moulds of 300 × 300 × 50 -mm vertically to obtain specimens with

uniform thickness. After about 24 hrs curing in the moulds, the specimens were

demoulded and cured in a fog room at temperatures of 28 - 30 °C to 28 days,

followed by 2-week drying in laboratory air and subsequent drying in an oven at

105 °C until constant weight was achieved. Mix proportion, density, and 28-day

compressive strength of the mortar specimens are given in Table 4.3.

A commercially available liquid transparent silicate coating 15 was applied on

a mortar specimen as a control. According to manufacturer, the silicate coating

contains potassium silicate, silica sol, and organic additives with a solid content of

13.5% and a pH of approximately 11. One layer of the transparent silicate coating
15
Keim Concretal Fixative (Concretal Dilution – Fixativ Concretal), KEIM Mineral Paints, UK
97
(denoted as “ S ”) was applied on one side of the specimen (300 × 300 -mm)

with a brush. The TSC is used because it has negligible effect on the solar

reflectance of mortar specimens and thus, the specimen with TSC can represent

conventional reinforced concrete and masonry building surfaces in terms of solar

reflectance.
16
A commercially available white silicate coating diluted with the

transparent silicate mentioned above (1:0.5 by mass) was applied on one side of a

2nd specimen also with a brush for one layer. he coating is denoted as “WS ”,

and the purpose of the dilution was for better workability.

The white silicate coating with photocatalyst (PWSC) included three layers: a

basecoat of WSC (one coat), a layer of TSC (one coat), and a photocatalytic top

layer (five coats). The photocatalytic top layer was prepared by mixing the

transparent silicate coating with 1.46% photocatalytic TiO2 17 (by volume) using a

probe ultra-sonicator at a power of 30W for 10 min. Characteristics of the

photocatalytic TiO2 provided by the manufacturer are summarized in Table 4.4.

According to X-ray diffraction (XRD) pattern (Figure 4.1) determined by a X-Ray

18
Diffractometer , the photocatalytic TiO2 used in this study includes two

crystalline phases: anatase and rutile, which is consistent with the information

from the manufacturer and reported in literature (Hernández-Ramírez and

Medina-Ramírez 2015). Before a new coat was applied, the specimen was

allowed to dry. The purpose of the basecoat is to increase the solar reflectance

similar to that of the specimen with WSC. The transparent silicate coating layer

16
Beecko-Sol Fine, Beeck'sche Farbwerke GmbH, Germany
17
Degussa P-25, Evonik Industries AG, Germany
18
X-Ray Diffractometer LabX XRD-6000, Shimadzu Corporation, Japan
98
between the basecoat and top layer was added to reduce interaction between

organic additives in the basecoat and the top layer. Experimental results in

literature show that photocatalytic TiO2 may degrade the organic components in

the coatings (Auvinen and Wirtanen, 2008; Geiss et al., 2012). Photocatalytic

performance of the top layer in terms of self-cleaning and degrading

environmental pollutants was reported in references (Krishnan, 2015; Krishnan et

al., 2013a; Krishnan et al., 2013b).

4.2.2 Deposition of black carbon on coated specimens

19
A commercially available black carbon generated from incomplete

combustion of heavy aromatic feedstock in a furnace was used as surrogate of

atmospheric black carbon. Black carbon suspension was prepared by dispersing

50 mg of black carbon powder in 100 g of deionized water by using the probe

ultra-sonicator at a power of 30 W for 10 min. The black carbon suspension with

a loading of 8 μg/cm2 was then applied on the surface of the coated specimens.

Based on the concentration of elemental carbon in airborne fine particulates

measured in Singapore, the loading of black carbon of 8 µg/cm2 is equivalent to

60 months of deposition of fine particulates without loss, for example, through

rainfall washing (Krishnan, 2015).

4.2.3 Exposure of the specimens with black carbon to simulated solar irradiation

In this study, the specimens with the black carbon deposition were exposed to

simulated solar irradiation with wavelength ranging from 295 to 3000 nm in an

19
Carbon Black Monarch 120, Cabot Corporation, USA
99
accelerated weathering chamber 20 for 300 hrs, which is equivalent to about 80

days of outdoor exposure in Singapore (Krishnan, 2015). In the accelerated

weathering chamber, the black panel temperature (the maximum temperature that

a metal plate covered with a black coating will experience), air temperature, and

relative humidity were controlled at 65 °C, 50 °C, and 80%, respectively. The

intensity of the simulated solar irradiation was 0.55 W/m2/nm at a wavelength of

340 nm. The air temperature, relative humidity, and irradiation intensity were

determined according to a standard accelerated exposure test for coatings in

ASTM D7356M (2013). It is noted that the “simulated solar irradiation” in this

thesis refers to the exposure detailed here.

4.2.4 Determination of color and solar reflectance

Color of the coated surfaces was determined by a spectrophotometer 21 and is

expressed in Commission International d’Eclairage AB system. Reading *

defines gray scale, readings a* and b* denote chromaticities (a* for red/green and

b* for yellow/blue). In this study only gray scale (L*) value is used to evaluate the

color of the coated surfaces because a* and b* are not significantly affected by the

coatings and black carbon. Black has L* reading of 0, whereas diffuse white has

L* reading of 100. Readings were taken at 12 locations (Table 4.5), each with

approximately 10 mm in diameter, on the surface of the specimen as shown in

Figure 4.2(a) and average result and standard deviation are reported. A plastic

20
Xenon Test Chamber Q-SUN XE-3, Q-Lab Corporation, USA
21
Spectrophotometer CM 600d. Konica Minolta, Japan
100
sheet template was used to ensure the measurements were taken at the same

locations on the specimen at the different time points.

Solar reflectance of the coated surfaces was determined according to ASTM

C1549-09 (2009) using a solar spectrum reflectometer 22. Readings were taken at

the same 12 locations on the specimens (Table 4.5) as the color except that the

measurement area was approximately 25 mm in diameter.

The solar reflectance and L* of the mortar specimens before coating were

determined and given in Table 4.3. The solar reflectance and L* of the coated

specimens were determined at different time points: (1) before black carbon

deposition; (2) after the black carbon deposition; and (3) after every 50 hrs

exposure to simulated solar irradiation in accelerated weathering chamber.

4.2.5 Determination of heat transfer through the specimens

Equipment setup shown in Figure 4.2(b) consists of a sun simulator, a test

panel (2 × 2 m2), and a sealed chamber (2 × 2 × 1 m3). The 300 × 300 × 50 -mm

specimen was mounted in the centre of the test panel and cotton filling the gap

between the specimen and test panel to ensure good insulation and air tightness.
23
An Arrimax lamp , located 10 m from the test panel, is able to provide

irradiation (wavelength from 300 to 1700 nm) for an uniform area of 1 × 1 m2 at

the centre of the test panel (Chen et al., 2012). The “sun simulator” in this thesis

denotes the lamp detailed here. The air temperature outside the chamber is

comparable to noon time outdoor air temperature in Singapore (28.0 - 31.6 °C)

22
Solar Spectrum Reflectometer Devices & Services Company, USA
23
Arrimax 18/12 lamp system with 18 kW metal halide lamp HMI 18000W/SE/GX51, ARRI, Germany
101
(Roth, 2015), while the air temperature inside the chamber is used to simulate

temperature of an air conditioned room. The incident angle of solar irradiation is

0o, which represents a scenario for the maximum heat gain through the specimen.

The surface of the specimen exposed to the sun simulator is referred to as

“exposure surface” while the other surface inside the chamber is referred to as

“inner surface”. The solar irradiance during the test was monitored by a

pyranometer installed on the exposure surface of the test panel near the specimen

(Figure 4.2(c)). The solar irradiance at the centre of the specimens is from 373 to

392 W/m2, which were similar to the average monthly intensity between 9 am to 5

pm in Singapore (384 W/m2) (Roth, 2015).

Two heat flux sensors were pasted at the centre of the inner surface of the

specimen Figure 4.2(a) to measure the heat flux (W/m2) through the specimen.

The heat flux is defined as the amount of heat gain per unit area per unit time

through the specimen. The instantaneous heat flux (W/m2) is then integrated vs.

time to obtain the total heat gain in the chamber through the specimens (J/m2) in 9

hrs. The duration of the experiment was selected so that the total energy reached

the surface of the specimens (12,085 - 12,690 kJ/m2) is comparable to the average

daily solar energy exposure in Singapore (12,436 kJ/m2) (Roth, 2015). The test

duration is more than sufficient to achieve steady state and further increase in the

test duration would not change heat flux and surface temperatures of the

specimens significantly.

Nine thermocouples were pasted on each side of the specimen to measure the

temperatures of its inner and exposure surfaces as shown in Figure 4.2(a), and the

102
average of the nine data points is used for analysis. Temperature and speed of air

flow in front of the test panel were monitored by a thermocouple and an

anemometer, respectively, placed about 0.5 m in front of the panel (Figure 4.2(c)),

while those in the chamber were monitored by 16 thermocouples and an

anemometer. The temperature in the chamber was controlled at 24 °C by chilled

water circulation and an electrical heater, whereas the air flow was regulated by

fans. About 15 hrs prior to the experiment, the electrical heater, chiller water

circulation, and monitoring system were switched on in order to achieve steady-

state air temperatures in and outside the chamber and surface temperatures of the

specimen. More details of the experimental setup can be found in references

(Chen and Wittkopf, 2012; Chen et al., 2012).

Information of the experiment i.e. intensity of solar irradiance, temperatures,

heat flux, and airflow speeds were collected every 10 seconds from the sensors

during the experiment, and the measurement conditions including solar irradiance,

air temperatures, and airflow speeds are summarized in Table 4.6. Although the

air flow speed outside the chamber varied slightly due to the central air-

conditioning system in the laboratory, the effect is negligible considering the

relatively low speed.

The heat gain through the specimens in the chamber is equivalent to the heat

removed by a cooling system to achieve a constant temperature of 24 °C in the

chamber. The cooling energy consumption is linearly dependent on the heat gain

in the chamber (Brodowicz and Dyakowski, 1993; Chua and Chou, 2010).

103
 Effects of solar reflectance and color on the heat gain and surface

temperatures of these specimens were determined and compared at three different

time points: (1) before black carbon deposition; (2) after the black carbon

deposition; and (3) after 300 hrs exposure to simulated solar irradiation in the

accelerated weathering chamber.

4.3 Results and discussion

In this section, paired sample t-test has been used to compare the difference

of solar reflectance, L*, and surface temperature of mortar specimens with

different TSC, WSC, and PWSC at various time points. The data used in paired t-

test is summarized in Appendix Table A1 and A2, and the statistical significance

used in the hypothesis test is 0.05.

4.3.1 Solar reflectance and gray scale L*

4.3.1.1 Effect of coatings

The solar reflectance and L* of the mortar specimens with TSC, WSC, and

PWSC measured at different times are summarized in Table 4.7 and Figure 4.3.

Transparent silicate coating has negligible effect on the solar reflectance and L*

while WSC and PWSC significantly increases the solar reflectance and L* of the

mortar specimens. Compared with the specimen with TSC, the specimens with

WSC and PWSC appear much whiter with approximately 60% higher solar

reflectance and 27% higher L* value. Results in Table 4.7 also indicate that the

photocatalytic top layer did not affect the solar reflectance and L* of the specimen

significantly.
104
4.3.1.2 Effect of black carbon deposition

The specimens with black carbon deposition appear dark with significant

reduction in the solar reflectance and L* of the specimens (Figure 4.4 and Table

4.7). For example, the solar reflectance and L* of the specimen with PWSC are

reduced by 52% and 28% (from 0.79 and 95.4 to 0.38 and 69.1), respectively. The

specimen with WSC has reductions of solar reflectance and L* similar to those

with PWSC with the deposition of black carbon. It is observed that with the given

black carbon loading, the absolute reduction in the solar reflectance of the

specimens with higher solar reflectance (i.e. WSC and PWSC) is more significant

than the specimen with lower solar reflectance. The observations are consistent

with results reported by other researchers (Cheng et al., 2012; Kelen et al., 2015;

Sleiman et al., 2011).

4.3.1.3 Effect of simulated solar irradiation up to 300 hrs on the specimens with

black carbon deposition

Paired t-test, with the data in Appendix Table A1, reveals that the solar

reflectance and L* of the specimens with TSC and WSC do not change

significantly after 300 hrs exposure as expected. However, the solar reflectance

and L* of the specimen with the PWSC increase with exposure time (Figure 4.3)

and the specimen becomes whiter, indicating removal of black carbon due to

photocatalytic reactions. After the exposure to simulated solar irradiation for 300

hrs, both the solar reflectance and the L* of the specimen with the PWSC (0.78

and 94.0) were recovered to 99% of their original values (0.79 and 95.4),

105
significantly higher than those of the specimens with TSC (0.24 and 54.8) and

WSC (0.38 and 70.8). The results of paired t-test indicates comparable solar

reflectance of the specimens with PWSC before BC deposition and after exposure

to simulated solar irradiation. Images of the specimens at different time points

also indicate black carbon removal and color recovery of the PWSC specimen

exposed to simulated solar irradiation for 300 hrs (Figure 4.4).

The removal of black carbon by photocatalytic TiO2 has been explored and

reported in literature (Lee and Choi, 2002; Mills et al., 2006; Smits et al., 2012;

Smits et al., 2013). For example, Mills et al. (2006) reported complete

degradation of 0.5 μm thick soot by 0.9 μm thick iO2 film on glass slides after

exposure to UV irradiation for 45 days based on image analyses and infrared

absorption spectra.

The mechanism of photocatalytic degradation of black carbon has been

investigated to a certain extent. When light photons of suitable wavelength (i.e.

with energy greater than or equal to the electron band gap of TiO 2) strikes the

surface of TiO2, electrons will be promoted from the filled valence band to the

empty conduction bands (Equation 4.2) (Chen and Poon, 2009). This generates

electron-hole pairs, which act as initiator for reduction-oxidation reactions to

occur at/near the surface of TiO2. The electrons act as acceptors, commonly

reducing oxygen to produce superoxides (Equation 4.3) while the holes oxidize

donors such as water molecules to form hydroxyl radicals (Equation 4.4) (Chen

and Poon, 2009). Carbonaceous compounds such as black carbon can be oxidized

by the holes or hydroxyl radicals to form CO2. Two pathways are proposed by

106
Chin et al. (2007) for the oxidation of soot/black carbon. The first pathway

assumes single step oxidation of the black carbon to CO2, whereas the second

pathway assumes the formation of intermediates before final oxidation to CO 2.

While the steps involved in the degradation are still under investigation, fourier

transform infrared (FTIR) spectroscopy and gas chromatography with photo-

ionization detection (GC-PID) has revealed that the products from the soot

oxidation is CO2 (Equation 4.5) (Pozo-Antonio and Dionisio, 2017; Smits et al.

2012).

h
iO2 h e- (4.2)

-
e- O2 O2 (4.3)

-
h O O (4.4)

O2 O2 (4.5)

While the above photocatalytic principles are generally applicable, the

degradation process is dependent on a number of factors such as type of TiO 2,

amount of exposed TiO2 exposed to solar irradiation, type and concentration of

the pollutant, intensity and duration of light irradiation, temperature, relative

humidity etc. Therefore, the literature results and our experimental results may

not be comparable due to the lack of standard test method and differences in TiO2

type and content, black carbon loading, and intensity of simulated solar irradiation.

The solar reflectance values of the coated specimens, specimens with black

carbon deposition, and those after exposure to simulated solar irradiation are

plotted against L* (Figure 4.5), and a linear relation between these two parameters

is established with a correlation coefficient of 0.96. Therefore, the discussion in

107
the following section will be focused on the effect of solar reflectance on heat

transfer through specimens.

4.3.2 Heat transfer through the specimens and surface temperatures at different

time points

4.3.2.1 Effect of coatings

Temporal changes of the heat flux and surface temperatures of the specimens

with TSC, WSC, and PWSC are shown in Figure 4.6(a) and Figure 4.6(b),

respectively. The heat flux, surface temperatures, and temperature rise rate of the

WSC and PWSC specimens are comparable, but significantly lower compared

with those of the TSC specimen due to the higher solar reflectance of the formers.

For example, after 9 hrs exposure to the sun simulator, the inner and exposure

surface temperatures of the WSC and PWSC specimens are 5.0 °C and 5.9 °C

lower than those of the specimen with TSC. Furthermore, the heat gain in the

chamber through the WSC and PWSC specimens in 9 hrs is almost 50% lower

than that through the TSC specimen, indicating that substantial reduction of

cooling energy consumption can be achieved by simply increasing the solar

reflectance of building surfaces.

4.3.2.2 Effect of black carbon deposition

Temporal changes of the heat flux and surface temperatures of the specimens

before and after black carbon deposition are shown in Figure 4.7(a) and Figure

4.7(b-d), respectively. Because of the reduction in the solar reflectance by black

carbon deposition, the heat flux and surface temperatures of the specimens are
108
increased mainly due to less heat reflected by the exposure surface and more heat

absorption by the black carbon. After 9 hrs exposure to the sun simulator, the heat

gains through the TSC, WSC, and PWSC specimens with black carbon deposition

are increased by 27% (1000 kJ/m2), 109% (2111 kJ/m2), and 123% (2336 kJ/m2),

and the inner surface temperatures of the corresponding specimens are increased

by 2.6 °C, 5.5 °C, and 6.0 °C, respectively (Table 4.8). This indicates substantial

increase in the cooling energy consumption to keep constant chamber temperature

due to the soiling. The results also indicate more significant increases in the heat

gain and surface temperatures for the specimens with higher solar reflectance, i.e.

WSC and PWSC, corresponding to more significant reduction of the solar

reflectance due to black carbon deposition mentioned in Section 4.3.1.2.

The solar reflectance and L* of the PWSC specimen with black carbon

deposition were determined before and after the heat transfer test. The difference

is negligible because of the short duration (9 hrs) of the exposure to the sun

simulator which is insufficient to degrade black carbon and change solar

reflectance and color significantly.

4.3.2.3 Effect of photocatalytic removal of black carbon

Temporal changes of the heat flux and surface temperatures of the PWSC

specimen with black carbon deposition before and after 300 hrs of exposure to

simulated solar irradiation are shown in Figure 4.8. With the removal of black

carbon and increase in the solar reflectance, the heat flux and surface

temperatures of the PWSC specimen are significantly lower compared with those

with the black carbon. After 9 hrs of exposure to the sun simulator, the exposure
109
and inner surface temperatures of the PWSC specimen are reduced by 5.9 °C and

7.2 °C, respectively, and the heat gain in the chamber is reduced by 53 % (Table

4.8). In contrast, the surface temperatures and heat gain through the TSC and

WSC specimens do not change significantly before and after the exposure to

simulated solar irradiation in the accelerated weather chamber.

With photocatalytic removal of the black carbon and recovery of the solar

reflectance of the PWSC specimen, its heat gain and surface temperatures are

comparable to those of the specimen before the black carbon deposition (Figure

4.9 and Table 4.8), as supported by paired t-test. This indicates that the PWSC can

be used to remove the black carbon and self-clean building surfaces, thus

reducing cooling energy consumption.

4.3.3 Correlation between solar reflectance and heat transfer through the

specimens

The effect of solar reflectance on the heat gain in the test chamber and

surface temperatures on the specimens after 9 hrs of exposure to the sun simulator

is shown in Figure 4.10(a) and (b), respectively. Regression analyses indicate

linear dependence of the heat gain and surface temperatures on the solar

reflectance. This seems to be consistent with the finding for opaque roof systems

by Suehrcke et al. (2008). Thus, the solar reflectance shown in this study may be

used as one of the parameters to predict the heat transfer through opaque building

envelope with given materials, design, and exposure condition.

Based on the information in Figure 4.10, an increase in the solar reflectance

by 0.1 reduces the heat gain through the specimen by 490 kJ/m2 and reduces
110
internal surface temperature by 1.3 °C. Comparing with the TSC which represents

opaque building surfaces with exposed concrete in reinforced concrete structures,

the PWSC increases the solar reflectance of the specimen by approximately 65%

(from 0.48 to 0.79), which leads to almost 50% lower heat gain in the chamber

and 5 °C lower temperature on the inner surface. Since the solar irradiance and

total energy reached the exposure surface of the specimens in this study are

comparable to those of day light exposure in Singapore, such a coating may

reduce cooling energy consumption significantly.

4.3.4 Practical significance

The black carbon loading of 8 μg/cm2 is estimated as the amount of elemental

carbon continually depositing on the rooftop of a building in Singapore for 60

months without loss. On the other hand, the complete removal of the black carbon

can be achieved after 300 hrs of solar irradiation in the accelerated weathering

chamber, equivalent to about 80 days of outdoor exposure in Singapore. Since the

rate of the black carbon removal is much faster than the rate of deposition, it is

likely that the specimen coated with PWSC under outdoor exposure may be able

to maintain the high solar reflectance when exposed to continuous soiling by

emissions of automobile traffic, industrial and domestic combustion, etc,

considering the abundant solar irradiation, high temperature and humidity in

tropical environments for the photocatalysis. This is also supported by a field

study in Singapore. A reinforced concrete panel (2.8 m × 2.8 m), used for the field

study, was divided vertically into two halves. One-half of the panel was coated

with photocatalytic coating and the other side was bare. It was observed that the
111
panel without coating, i.e. bare panels showed a decrease in the solar reflectance

from 0.45 to 0.39 during 2.5 years outdoor exposure. However, the photocatalytic

coating could effectively maintain the solar reflectance (without decrease) during

the outdoor exposure period (Krishnan et al., 2017). Singapore has a relatively

“clean” urban environment, as demonstrated by the air quality inde , i.e. the the

Pollutant Standard Inde (PSI) which recorded to be in the “Good” and

“Moderate” range for 99% of days in year 2016. herefore, it is expected that the

decrease in solar reflectance would not be substantial. Several studies show

substantial decreases in the solar reflectance under outdoor exposures (Kelen et al.,

2015; Levinson et al., 2005; Paolini et al., 2014; Takebayashi et al., 2016), as

discussed in Section 4.1. For example, acrylic coatings on ceramic material

showed solar reflectance decrease from 0.54 - 0.90 to 0.46 - 0.63 within 18

months of exposure in Brazil (Kelen et al., 2015). Although the decrease in the

solar reflectance is highly dependent on pollutant concentrations and ambient

weather conditions in a particular location, the use of PWSC can be beneficial

because they can prevent soiling on the surface and maintain the solar reflectance.

Overall, building surfaces with the PWSC can reasonably maintain its high

solar reflectance and contribute to aesthetic appearance and energy efficiency of

buildings. Compared to building surfaces with ordinary coatings, the application

of the photocatalytic coatings with comparable solar reflectance may reduce cost

from both maintenance such as washing and bleaching and cooling energy

consumption of buildings in tropical countries and in summer months in

temperate countries. Cooling energy demand is not elaborated in the manuscript

112
because it is affected by various parameters such as building design and

environment conditions. Nevertheless, this study demonstrates the capacity of the

photocatalytic coating to reduce the heat gain through building envelope,

indicating the reduction of cooling energy demand. Besides the reduced heat gain

and inner surface temperature, the temperature rise on the inner surfaces of the

PWSC specimen is likely slower than that of the specimens with TSC and WSC

under outdoor exposure, suggesting smaller indoor temperature fluctuation of

buildings without air conditioning. This may postpone the arrival of peak

temperature, increase indoor thermal comfort, and shorten the cooling period.

Notwithstanding the above advantages, additional work is required for

formulating photocatalytic coatings which can fulfil the desired service life in

terms of degradation ability and durability before large-scale application.

4.4 Summary and conclusions

The ability of photocatalytic coatings to maintain solar reflectance of opaque

building envelope and lower the cooling load was demonstrated experimentally

by comparing the heat transfer and surface temperature of Portland cement mortar

specimens with TSC, WSC and PWSC under controlled experimental conditions.

The solar reflectance, gray scale L*, heat gain, and surface temperatures of the

specimens exposed to a sun simulator were determined at three different time

points: (1) before black carbon deposition; (2) after black carbon deposition; and

(3) after 300 hrs exposure to simulated solar irradiation in an accelerated

weathering chamber. Based on the results and discussion, following conclusions

are drawn:
113
 1. Specimens with the WSC and PWSC have higher solar reflectance and L*

and appeared whiter than the specimen with TSC. After 9 hrs of exposure to the

sun simulator, the heat gain through the WSC and PWSC specimens is almost 50%

lower than that through the TSC specimen, and the temperatures of the inner and

exposure surfaces of the formers are about 5 °C and 6 °C lower than those of the

latter. The lower heat gain and surface temperatures of the WSC and PWSC

specimens can be attributed to their higher solar reflectance.

2. Black carbon deposition reduces the solar reflectance and L* of the coated

specimens. After 9 hrs of exposure to the sun simulator, the heat gain in the

chamber through the specimens with black carbon deposition is increased by 27%,

109%, and 123% for the specimens with TSC, WSC, and PWSC, respectively,

and the temperatures of the inner and exposure surfaces of the specimens are also

increased significantly, indicating increased cooling energy consumption with

soiling of building surfaces. The greater increase in the heat gains through the

WSC and PWSC specimens may be attributed to more significant reduction of

their solar reflectance by black carbon deposition.

3. After 300 hrs of exposure to simulated solar irradiation, the solar

reflectance and L* of the PWSC specimen with black carbon deposition were

almost completely recovered to the initial values. The surface temperatures and

heat gain through the PWSC specimens are comparable to those before the black

carbon deposition. The inner surface temperature and heat gain of the specimen

with PWSC after 300 hrs exposure to simulated solar irradiation are reduced by

7.2 °C and 2242 kJ/m2 (53%), respectively, compared to the specimen after black

114
carbon deposition. The solar reflectance and L* of the TSC and WSC specimens,

however, do not change significantly after the exposure to the simulated solar

irradiation, and their heat gain and surface temperatures are comparable to those

after the black carbon deposition.

The above suggests that photocatalytic coating can remove black carbon and

maintain solar reflectance of opaque building surfaces. As a result, it can reduce

cooling energy consumption of buildings in tropical countries and in summer

months in temperate countries.

115
 4.5 Tables

Table 4.1 Solar reflectance of common roof surfaces (Marceau and Vangeem,
2008; Parker et al., 1993)

Roof surfaces Solar reflectance

Asphalt shingles 0.03-0.20

Asphalt shingles-white 0.25-0.31

Cement tile 0.25

Cement shingles-white 0.77

Concrete-new 0.41-0.52

Concrete tile-white 0.73

Concrete tile-colored 0.21-0.67

Metal roofing-white 0.56-0.67

Roof membrane-white 0.69-0.81

Roofing membrane-dark color 0.06-0.23

Smooth bitumen 0.06

Unpainted aluminum 0.71

116
 Table 4.2 Effect of soiling on solar reflectance and thermal performance of various coatings

Exposure Solar Solar

Initial solar Exposure Maintenance
Reference Specimens Color duration, reflectance reflectance after
reflectance environment method
years after exposure maintenance

Dornelles et Ceramic plates with acrylic Light Outdoor exposure in

0.54-0.90 1.5 0.46-0.63 NA NA
al. 2015 coating color São Paulo, Brazil

Wiping,
Levinson et Light Outdoor exposure in rinsing,
PVC membranes 0.63-0.80 5-8 0.32-0.70 0.63-0.82
al. 2005 color 10 US cities washing,
bleaching
Roofing membranes of
Paolini et al. Various Outdoor exposure in
modified bitumen, PVC or 0.23-0.85 2 0.23-0.75 NA NA
2014 color 2 Italian cities
polyolefin
Cement based board with Exposed to
Shi et al. coating of acrylic emulsion simulated soiling Washing
White 0.87-0.88 NA 0.80 0.86-0.87
2013 incorporating TiO2 rutile with graphite ash with soap
and glass microspheres slurry
Roofs with high reflectance Washing
Takebayashi Outdoor exposure in
paint, composition White 0.73-0.87 0.2-7.0 0.60-0.79 with wet 0.67-0.88
et al. 2016 5 Japanese cities
unknown cloth

Films with Styrene acrylate Exposed to

Xue et al.
copolymer, white cement, White 0.82 simulated soiling of NA 0.62-0.63 NA NA
2015
and TiO2 rutile. etc graphite ash slurry

“PVC” denotes polyvinyl chloride

117
 Table 4.3 Information of mortar specimens

Size 300 × 300 × 50 -mm

Mix proportion (by mass) Water: Cement: Sand= 0.5:1:2.5
3
Demoulded density, kg/m 2160 (10)
Oven dry density, kg/m3 1975 (10)
28-day compressive strength, MPa 39.0 (3.2)
Solar reflectance 0.52 (0.03)
Gray scale L* 75.2 (1.6)

Table 4.4 Characteristics of photocatalytic TiO2

Specific surface area (BET), m2/g 50±15

Average primary particle size, nm 21
Density, g/L Approximately 130
Ignition loss (2 h at 1000 oC), wt.% ≤2.0
pH-value 3.5 - 4.5
TiO2-content (based on ignited material), wt.% ≥ 99.5

Table 4.5 Test methods and specimen information

Specimen
Property Test standard
Type and size Number
ASTM C1549-09
Solar reflectance 12
(2009)

Color --- 12
Slab,
300×300×50 mm
Heat flux --- 2

Surface temperatures --- 9

Airflow speed --- --- 1

Solar irradiance --- --- 1

118
 Table 4.6 Experimental conditions for determining the heat transfer through
specimens exposed to a sun simulator

Outside the Inside the

chamber chamber
Airflow speed, m/s 0 - 0.15 0.25 - 0.27

Initial air temperature, °C 22.5 - 23.4 24.0

Initial surface temperatures of specimen, °C 23.6 - 24.1

Air temperature during the experiment, °C 29.3 - 30.7 24.0

Incident angle of solar radiation, ° 0 ---

Solar irradiance, W/m2 373 - 392 ---

Total energy reached the exposure surface in 9 hrs,
12,085 - 12,690 ---
kJ/m2

119
 Table 4.7 Solar reflectance and color of specimens with various coatings at various time points

Exposure duration to the simulated solar irradiation, hrs

Solar reflectance
Before BC After BC
50 100 150 200 250 300
deposition deposition
TSC 0.48 (0.03) 0.21 (0.03) 0.22 (0.02) 0.23 (0.02) 0.23 (0.03) 0.24 (0.03) 0.24 (0.03) 0.24 (0.03)
WSC 0.76 (0.02) 0.37 (0.05) 0.37 (0.05) 0.38 (0.05) 0.38 (0.05) 0.38 (0.05) 0.38 (0.05) 0.38 (0.05)
PWSC 0.79 (0.01) 0.38 (0.03) 0.52 (0.05) 0.60 (0.05) 0.68 (0.06) 0.72 (0.04) 0.76 (0.03) 0.78 (0.02)
Color/ gray scale (L*)
Before BC After BC
50 100 150 200 250 300
deposition deposition
TSC 75.1 (2.1) 50.9 (4.2) 52.3 (3.7) 52.7 (3.8) 53.0 (3.8) 53.9 (4.0) 54.1 (3.8) 54.8 (3.9)
WSC 95.4 (1.0) 70.0 (3.7) 70.4 (3.6) 70.3 (3.8) 70.4 (3.8) 70.7 (3.8) 70.8 (3.7) 70.8 (3.6)
PWSC 95.4 (0.7) 69.1 (2.7) 78.7 (3.4) 84.2 (3.2) 88.4 (3.0) 91.0 (2.3) 92.8 (1.9) 94.0 (1.2)
“TSC”, “WSC”, and “PWSC” denote specimen with transparent silicate coating, white silicate coating, and white silicate coating
incorporating photocatalyst, respectively. “BC” denotes black carbon
-data presented are average (standard deviation).

120
 Table 4.8 Heat gain in the chamber and surface temperatures of the specimens
with various coatings exposed to the sun simulator for 9 hrs

Surface temperature after 9 hrs, °C

Solar Heat gain,
reflectance kJ/m2
Inside chamber Outside chamber
Before black carbon deposition
TSC 0.48 (0.03) 36.8 40.1 3746
*
WSC 0.76 (0.02) 31.8 34.2 1928
PWSC 0.79 (0.01) 31.8 34.2 1900
After black carbon deposition
TSC 0.21 (0.03) 39.4 43.5 4746
WSC 0.37 (0.05) 37.3 40.9 4039
PWSC 0.38 (0.03) 37.8 41.4 4236
Exposure to simulated solar irradiation for 300 hrs in accelerated weathering
chamber
TSC 0.24 (0.03) 38.8 42.4 4473
WSC 0.38 (0.05) 37.0 40.3 3909
PWSC 0.78 (0.02) 31.9 34.2 1994
*
the heat gain through the specimen with WSC before black carbon deposition
was measured for 8.75 hrs, and the data presented in this table were obtained by
assuming no change on the heat flux and surface temperatures from 8.75 to 9 hrs.

121
 4.6 Figures

Figure 4.1 XRD pattern of the photocatalytic TiO 2 used in this study

122
Figure 4.2 (a) Locations of heat flux sensors, thermocouples, and testing points
for solar reflectance and surface temperature; (b) Schematic experimental setup
for determining the heat transfer and surface temperature of the specimen
exposed to a sun simulator (not to scale); (c) Image of the testing chamber and
the locations of some sensors

123
 Figure 4.3 Changes of (a) solar reflectance and (b) gray scale L* of the specimens with
various coatings and black carbon deposition exposed to simulated solar irradiation in
accelerated weathering chamber
-“BC” denotes black carbon

Figure 4.4 Images of specimens at different time points

-“ TSC”, “WSC”, and “PWSC ” denote specimen with transparent silicate coating, white
silicate coating, and white silicate coating incorporating photocatalyst, respectively.

124
 Figure 4.5 Correlation between solar reflectance and gray scale L*

Figure 4.6 Heat flux (a) and surface temperatures (b) of the specimens with various coatings
exposed to a sun simulator for 9 hrs

125
Figure 4.7 Effect of black carbon deposition on the heat flux (a) and surface temperatures of
the specimens with TSC (b), WSC (c), and PWSC (d) exposed to the sun simulator for 9 hrs

Figure 4.8 Effect of black carbon removal on the heat flux (a) and surface
temperature (b) of the PWSC specimen before and after 300 hrs exposure to
simulated solar irradiation

126
 Figure 4.9 Comparison of the heat flux (a) and surface temperature (b) of the
PWSC specimen before black carbon deposition and after black carbon removal

Figure 4.10 Relationship between solar reflectance and the heat gain (a) and
surface temperatures of the specimens (b) after 9 hrs of exposure to the sun
simulator

127
 Chapter 5. Effect of ultra-lightweight cement composite and

photocatalytic coating on heat transfer through opaque building

envelope

The low thermal conductivity of ULCCs and high solar reflectance and self-

cleaning ability of photocatalytic coating have been demonstrated in previous

chapters. It is necessary to further explore the effects of the ULCC with low

thermal conductivity and photocatalytic coating with high solar reflectance on the

heat transfer through opaque building envelope. Considering the duration and cost

of the experiments, it is also attractive to develop empirical equations to predict

and evaluate the heat transfer through construction materials with different

thermal conductivities and solar reflectances.

5.1 Experimental studies

A normal weight concrete and a ULCC (ULCC-2) detailed in Chapter 2 were

selected for this part of the study because they have comparable compressive

strength and solar reflectance but significantly different thermal conductivities.

The photocatalytic coating detailed in Chapter 4 was applied on the ULCC

specimen to increase its solar reflectance. The heat gain and surface temperatures

of the specimens of the concrete and ULCC with and without coating exposed to

simulated sunlight for 9 hrs were compared and analyzed in this section. The

experimental methods used were the same as those described in Section 4.2.

Experimental results of the thermal conductivity and solar reflectance of the

concrete and ULCC with and without photocatalytic coating are summarized in
128
Table 5.1, and those of the heat gain and surface temperatures of these specimens

exposed to simulated sunlight are summarized in Table 5.2.

The thermal conductivity of the ULCC (0.39 W/m·K) is approximately 80%

lower than that of the concrete (1.98 W/m·K). The thermal conductivity of the

ULCC specimen with and without the photocatalytic coating does not differ,

indicating that a thin coating on the ULCC specimen has negligible effect on the

thermal conductivity.

5.1.1 Solar reflectance of the concrete and lightweight cement composite with and

without coating

As shown in Table 5.1, the solar reflectance of the concrete is 0.40,

comparable to that of the ULCC (0.41). Paired t-test failed to reject the null

hypothesis that the mean solar reflectance of concrete and ULCC is equal,

indicating the solar reflectance of concrete and ULCC was not statistically

different. The solar reflectance of the concrete and ULCC specimens is affected

by the raw materials used (exposed on surface), mix proportions of the mixtures,

and moisture condition (Levinson and Akbari, 2002), as mentioned above. The

solar reflectance of the materials used in this study in an ascending order is:

cenospheres (0.27) < cement (0.35) < sand (0.43) < silica fume (0.56) (Figure 5.1).

It is noted that the silica fume used in this study was whitish with lower carbon

content compared to typical silica fume. Although the concrete does not contain

whitish silica fume, its solar reflectance is comparable to that of the ULCC due to

the higher solar reflectance of the sand than that of the cenospheres and higher

water to binder (w/b) ratio. Published research (Marceau and Vangeem, 2008)
129
indicates that increasing w/b ratio leads to higher degree of cement hydration and

higher solar reflectance. The solar reflectance of the concrete (0.40) is consistent

with that reported by Marceau and Vangeem (2008).

The photocatalytic coating almost doubled the solar reflectance value of the

ULCC from 0.41 to 0.78 mainly due to the white base layer (Section 4.3.1).

Paired t-test revealed that the solar reflectance of the ULCC with coating was

significantly higher than that without coating. The solar reflectance of the

photocatalytic coating is comparable to those of “cool coatings” for concrete

surfaces reported in literature (Levinson et al., 2010; Song et al., 2014; Synnefa

and Santamouris, 2013). However, the photocatalytic coating can remove

deposited black carbon and maintain high solar reflectance of building surfaces

with time, as discussed in Chapter 4.

5.1.2 Surface temperatures and heat gain through the specimen exposed to the

simulated sunlight

One dimensional heat transfer through a wall of an opaque chamber is shown

in Figure 5.2. The energy input (Q) can be calculated according to Equation 5.1,

which equals to the sum of the heat reflected by the surface (Q r), heat loss on the

exposure surface due to convection and radiation (Ql), heat stored in the specimen

(Qs), and heat gain through the specimen (Qt) (Equation 5.2). The heat reflected

by the exposure surface and the heat stored in the specimen can be calculated by

Equations 5.3 and 5.4, respectively, whereas the heat gain through the specimen

can be obtained from the integration of the heat flux versus time (Equation 5.5)

(Callister and Rethwisch, 2008; Young et al., 2008). From the information, heat
130
loss due to convection and radiation (Ql) can be estimated as the difference

between the input energy (Q) and the sum of heat terms Qr, Qs, and Qt.

Q = It (5.1)

Q = Qr + Ql + Qs + Qt (5.2)

Qr = β Q (5.3)

Qs =Cp ρdry V(Tt-T0)/A (5.4)

Qt = ∫ qt dt (5.5)

Where, Q - heat reached the exposure surface (or energy input), J/m2;

I - solar irradiance at the exposure surface of the specimens, W/m2 ;

t - exposure time, s;

Qr - heat reflected by exposure surface, J/m2;

Ql - heat loss from exposure surface due to convection and radiation,

J/m2 ;

Qs - heat stored in the specimens, J/m2 ;

Qt - heat gain through the specimens, J/m2 ;

β - solar reflectance of exposure surface;

Cp - specific heat capacity of concrete or ULCC, J/kg·°C;

ρdry - oven dry density, kg/m3 ;

V - volume of specimen, m3 ;

A - area of the specimen, m2;

Tt - average temperature of the specimen exposed to the simulated

sunlight, °C;

T0 - initial temperature of the specimen, °C;

131
 qt - heat flux through the specimens, W/m2.

According to these equations, the heat transfer through the panel specimen

exposed to the sunlight is affected by the solar reflectance, thermal conductivity,

density, and specific heat capacity of the specimen. Although the effect of the

specific heat capacity is not negligible for non-steady state heat conduction, the

specific heat capacity of oven-dried lightweight concrete does not differ

significantly from that of normal weight concrete (ACI-122R, 2002; FIP, 1983).

The oven-dry density is correlated to the thermal conductivity of the specimen.

Therefore, the solar reflectance and thermal conductivity can be considered as the

main parameters that influence the heat transfer through the specimens exposed to

the sunlight.

Studies demonstrate that cooling energy consumption of a building is linearly

dependent on the heat gain through building envelope (Chua and Chou, 2010). To

reduce the energy consumption of cooling systems, a lower heat gain through the

building envelope is desired. For a given energy input (Q), the heat gain through

the specimen (Qt) can be reduced by increasing the heat reflected by the exposure

surface (Qr), heat stored in the specimens (Qs) and heat loss due to convection and

radiation (Ql) on the exposure surface.

Based on the equations above, values of Qr, Ql, Qs, and Qt were calculated

and summarized in Table 5.2. Since the solar irradiance at the exposure surface of

the specimen was approximately 373 - 392 W/m2 in this study (Table 4.6), the

average value of 383 W/m2 from the three tests was used for the calculations.

According to “A I-122R: guide to thermal troperties of concrete and masonry

132
systems” (2002), specific heat capacity of concretes with densities ranging from

1280 to 1920 kg/m3 and from 2080 to 2240 kg/m3 is 0.88 and 0.92 kJ/kg·°C,

respectively. In this study the values of 0.88 kJ/kg·°C and 0.92 kJ/kg·°C,

therefore, were used for the calculation of Q s for the ULCC and concrete,

respectively, based on their densities (Table 5.1). The volume and surface area of

the specimen are 4.5 × 10-3 m3 and 0.09 m2, respectively. The temperature of the

specimen in Equation 5.4 is assumed to be the average temperature of the

exposure and inner surfaces. Heat flux in Equation 5.5 was obtained from heat

flux sensors on the inner surface of the specimens (Section 4.2.5).

For example, for the concrete exposed to the simulated sunlight for 9 hrs,

Q = It = 383 × 9 × 3600 = 12,409 kJ/m2;

Qr = βQ = 0.40 × 12,409 = 4964 kJ/m2 ;

Qs =Cp ρdryV(Tt-T0)/A = 0.92 × 2251 × 4.5 × 10-3 × ((40.5+37.4)/2 - 23.6) /

0.09 = 1589 kJ/m2;

Qt = ∫ qt dt= 3689 kJ/m2 (from the integration of heat flux);

Ql = Q - Qr - Qs - Qt = 2167 kJ/m2.

5.1.2.1 Effect of thermal conductivity

For given solar reflectance of the specimen surface, the surface temperatures

and heat gain through the concrete and ULCC specimens are mainly affected by

their thermal conductivity. Lower thermal conductivity of the ULCC leads to 33%

less heat gain through the specimen, 3.7 °C lower inner surface temperature

(Table 5.2), and lower rate of temperature rise on the inner surface after 9 hrs

exposure to simulated sunlight than those of the corresponding concrete specimen

133
(Figure 5.3). Paried t-test has been conducted and the results demonstrated the

reduction of inner surface temperature due to low thermal conductivity. However,

the temperature and rate of temperature increase on the exposure surface of the

ULCC are higher than those of the concrete specimen (Figure 5.3). Because of

less heat gain through the ULCC specimen, more heat from the simulated sunlight

is accumulated on its exposure surface which resulted in higher temperature and

greater heat loss from the exposure surface due to convection and radiation. As a

result, after 9 hrs of exposure to the simulated sunlight, the exposure surface

temperature of the ULCC is 2.7 °C higher than that of concrete specimen.

5.1.2.2 Effect of solar reflectance

The temporal changes in the heat flux and surface temperatures of the ULCC

specimens with and without photocatalytic coating are presented in Figure 5.4.

With the increase in the solar reflectance from 0.41 (ULCC without coating) to

0.78 (ULCC with coating), the heat flux, surface temperatures and rate of

temperature increase are reduced significantly. For example, after 9 hrs exposure,

the exposure and inner surface temperatures of the ULCC with coating are 7.8 °C

and 4.3 oC lower than those without coating, respectively (Table 5.2). The coating

reduced heat gain through the specimen by 54% due to the significant increase in

the solar reflectance and the heat reflected. The heat loss on the exposure surface

the ULCC with coating is lower than those of without coating due to the lower

exposure surface temperature. ULCC with coating has lower heat stored than

ULCC because of lower inner and exposure surface temperatures. These results

are consistent with those reported in literature (Revel et al., 2014; Synnefa et al.
134
2007b; Uemoto et al., 2010; Zingre et al., 2015). For example, Uemoto et al.

(2010) conducted an experimental study on the heat transfer through a cement

based roofing specimen and found 37% reduction of heat flux through the

specimen by increasing the solar reflectance from 0.47 to 0.78.

5.1.2.3 Combined effect of thermal conductivity and solar reflectance

The combined effect of reduced thermal conductivity and increased solar

reflectance was investigated by comparing the heat transfer and surface

temperatures of the concrete without coating and ULCC with coating shown in

Figure 5.5. The ULCC specimen with coating had 80% lower thermal

conductivity and 90% higher solar reflectance compared with the concrete

specimen, leading to significant reduction of heat flux, surface temperatures and

rate of inner temperature rise. For example, after 9 hrs exposure, the inner surface

temperature of the ULCC with coating was 8.0 °C lower than that of concrete.

The heat gain was reduced by 69% mainly due to the increase in the heat reflected.

5.1.2.4 Practical implication

Since the solar irradiance and energy reaching the exposure surface in the

experiment are comparable to those in natural environment in Singapore, the

application of the ULCC, photocatalytic coating, and their combination may

reduce the heat gain and cooling energy consumption of buildings significantly

assuming other factors remain similar. In the natural environment, the temperature

of the exposure surfaces may be higher during the peak period (e.g. 9:00-10:00

for building surfaces facing east and 11:00-14:00 for building surfaces facing

135
south/north in Singapore (Lim, 1988)) due to higher solar irradiance. Therefore, it

is expected that the reduction in the heat gain and inner surface temperature due to

the use of the ULCC, photocatalytic coating, and their combination may be more

significant during the peak period.

In addition, reduced thermal conductivity, increased solar reflectance, and

their combination lead to lower inner surface temperature and slower rate of

temperature rise. The lower inner surface temperature indicates reduced

temperature of the building interior, while the slower rate of temperature rise

leads to smaller indoor temperature fluctuation of buildings without air

conditioning. These may postpone the arrival of peak temperature, increase indoor

thermal comfort, and shorten the cooling period.

Both photocatalytic coating and ULCC can be used for new buildings, but the

photocatalytic coating can also be used for existing buildings as well. Compared

with ULCC, the application of photocatalytic coating is more feasible because it

can achieve significant reduction of heat gain and cooling energy consumption

without affecting structural design of buildings. On the other hand, the increased

solar reflectance on the building envelope in densely populated city may reduce

the outdoor thermal comfort. Thus, both advantages and disadvantages of ULCC

and photocatalytic coating need to be considered comprehensively in the practical

application.

136
5.2 Empirical equations development

5.2.1 Methodology

Heat gain through building envelope was selected as a parameter to be

predicted because it has direct correlation with cooling energy consumption

(Brodowicz and Dyakowski, 1993; Chua and Chou, 2010). Based on the idealized

one-dimensional heat transfer discussed in Section 5.1.2, the heat gain through

opaque building envelope is affected by solar reflectance, density, and thermal

conductivity of the materials used for the building envelope. Since oven dry

density and thermal conductivity are correlated, the empirical equations are

developed based on solar reflectance and thermal conductivity or oven dry density.

As discussed in Section 4.3.3, the heat gain through mortar specimens appears to

be linearly dependent on their solar reflectance. Therefore, the discussion in this

section starts from assuming linear relationship between heat gain and solar

reflectance, density, and thermal conductivity.

The best subset regression method (Montgomery, 2011) was used to identify

parameters and models that produced the highest adjusted coefficient of

determination (adj-R2) from the full set of predictor variables. The adj-R2 is

adjusted for the number of predictors in the model (Equations 5.6 and 5.7) and

increases only if the new term improves the model more than that would be

expected by chance. In cases where several models have comparable adj-R2,

Mallows’ p value (Equations 5.8 - 5.10) and standard error of the regression “S”

value (Equation 5.11) can be used as additional criteria to determine which model

fits the experimental results better. Mallows’ p is used to compare the precision
137
and bias of the model with all predictors to models with the subsets of predictors.

A Mallows’ p value closer to the number of variables plus the constant indicates

that the model is relatively precise and unbiased in estimating the true regression

coefficients and in predicting future responses. In this study, Mallows’ p value

closer to 4 is preferred (three variables and one constant). The “S” value evaluates

difference between the predicted value and experimental results and equals to the

square root of the mean square error. The smaller the “S” values, the better the

model fits the experimental data.

-
- (5.6)
-

2 n-1
adj-R = 1- n-p 1- (5.7)

- -
Mallows’ (5.8)

n 2
1 (yi -yi )
(5.9)
n-p

n 2
1 (yi -yi )
(5.10)
n-k

S= (5.11)

Where, n - number of the experimental sets, equals to 12 in this study;

yi - heat gain value of ith sample determined by experiment, kJ/m2;

yi - predicted heat gain value of ith sample, kJ/m2 ;

y - mean of the heat gain of experimental data, kJ/m2;

- mean square error of model with p variables;

- mean square error of model with all variables;

- number of variables in the model plus the constant;

138
 k - total variables considered plus the constant, equals to 4 in this

study;

Regression analysis is used to investigate and model the relationship between

a response variable (heat gain in this study) and predictors (subset of thermal

conductivity, oven dry density, and solar reflectance). In this study, linear

equations were used to fit the experimental results with Minitab 16 software 24

and the coefficients are determined based on the least squares method.

The experimental results of oven dry density, thermal conductivity, solar

reflectance, and heat gain from Chapter 4 and current chapter are summarized in

Table 5.3. Variable selections according to the best subset method and multiple

regression analysis are conducted with Minitab 16 software based on 12 sets of

experimental data.

5.2.2 Results and discussions

The fittings of the subset regression results with linear equations are shown in

Table 5.4. The individual parameter of the oven dry density, thermal conductivity,

and solar reflectance are insufficient to predict the heat gain based on the low adj-

R2 value. The maximum adj-R2 and smallest “S” values are observed in the subset

with the thermal conductivity and solar reflectance, and this subset also has the

Mallows’ p closest to 4 (number of variables plus a constant). It is observed that

the subset with oven dry density and solar reflectance may also provide

reasonable prediction with an adj-R2 of 0.95. Considering that the oven dry

density can be determined more conveniently than the thermal conductivity of

24
Minitab 16, Minitab Incorporation, USA
139
material, the prediction model with oven dry density and solar reflectance is also

discussed here.

Two regression equations below (Equation 5.12 and 5.13) are assumed and

the relevant coefficients were determined by Minitab software with least squares

method, as summarized in Table 5.5.

Qt = a + bλ + cβ (5.12)

Qt = a + b + cβ (5.13)

Where, Qt - heat gain, kJ/m2 ;

λ - thermal conductivity, W/m·K;

- oven dry density, kg/m3;

β - solar reflectance.

The normal plot of residuals in Figure 5.6 shows that the difference between

the predicted values from both equations and experimental results appears to be

random and unbiased. The prediction errors are plotted in Figure 5.7 and it is

observed that most of predictions had errors within 10% from the equation with

the thermal conductivity and solar reflectance (Equation 5.12) with one exception

which had an error of approximately 17%. Nine predictions from the equation

with the oven dry density and solar reflectance (Equation 5.13) had errors less

than 10% and three had errors between 10-15%. All these analyses revealed that

the heat gain can be predicted with the solar reflectance and thermal conductivity

or oven dry density with errors less than 20%. Even though the equation with the

oven dry density and solar reflectance has lower adj-R2, it can be used to predict

the heat gain when the thermal conductivity data are not available.

140
 The specimens used in the experiments have a thermal conductivity range of

0.39 - 1.98 W/m·K and a solar reflectance range of 0.21 - 0.79. The above thermal

conductivity covers a wide range for most of structural concrete and the solar

reflectance covers a wide range for most of ordinary opaque building surfaces.

This indicates that the prediction models derived above are applicable to most

concrete building envelope.

The heat gain of the specimens with different thermal conductivities (from

0.39 to 1.98 W/m·K) and solar reflectance (from 0.21 to 0.79) predicted by

Equation 5.12 and 5.13 are plotted in Figure 5.8. It is observed that the heat gain

is decreased with the decrease in thermal conductivity but with the increase in

solar reflectance, and the heat gain is more sensitive to change of solar reflectance

due to the higher coefficient values. With the increase of solar reflectance by 0.1,

the heat gain can be reduced by 481 kJ/m2, which is equivalent to reduction in the

thermal conductivity by approximately 0.6 W/m·K (Equation 5.12) or reduction

in the oven dry density by 350 kg/m3 (Equation 5.13). Considering that coating

does not affect the structural design of buildings, it may be more feasible to

reduce the heat gain and cooling energy consumption by increasing solar

reflectance.

Even though the prediction models may be limited to the measurement

conditions (e.g. solar irradiance and air temperature) in this study, they can be

used to evaluate the effect of material properties on the heat transfer through

concrete building envelope.

141
5.3 Summary and conclusions

The effects of a structural ULCC with low thermal conductivity and a

photocatalytic coating with high solar reflectance on the heat transfer through

panel specimens were investigated experimentally in a controlled environment.

Empirical equations to predict the heat gain through construction materials with

different thermal conductivities and solar reflectance were developed. Based on

the results and discussion, following conclusions can be drawn:

1. Compared with a conventional concrete of similar 28-day compressive

strength, the ULCC has 80% lower thermal conductivity (0.39 W/m·K) which

reduces the heat gain through the ULCC by 33% and decreases the inner surface

temperature of the specimen by 3.7 °C after exposure to simulated sunlight for 9

hours.

2. The photocatalytic coating increases the solar reflectance of the ULCC

specimen from 0.41 to 0.78 which reduces the heat gain by 54% and decreases the

inner surface temperature of the specimen by 4.3 °C after exposure to simulated

sunlight for 9 hrs.

3. The combination of the ULCC with low thermal conductivity and

photocatalytic coating with high solar reflectance reduces the heat gain and inner

surface temperature by 69% and 8.0 °C, respectively, after 9 hrs exposure to

simulated sunlight.

4. The linear empirical model with the thermal conductivity and solar

reflectance provides good prediction of the heat gain with adj-R2 of 96.6%. In

addition, the model with the oven dry density and solar reflectance can also

142
provide reasonable prediction for the heat gain with adj-R2 of 95.3% when

thermal conductivity data are not available.

5. Based on the coefficients of the regression equations, the heat gain is more

sensitive to the change of the solar reflectance than that of the thermal

conductivity and oven dry density. Increase in the solar reflectance by 0.1 is

equivalent to reduction in the thermal conductivity by approximately 0.6 W/m·K

or reduction in the oven dry density by approximately 350 kg/m3.

143
 5.4 Tables

Table 5.1 Density, thermal conductivity, and solar reflectance of the concrete and
ULCC with and without photocatalytic coating

Oven dry Thermal conductivity, Solar

Coating
density, kg/m3 W/m·K reflectance*

Concrete No 2251 1.98 0.40 (0.01)

ULCC No 1303 0.39 0.41 (0.02)
ULCC+C Yes 1305 0.39 0.78 (0.02)
*
the paired t-test with data from the 9 test locations (Appendix A3) indicates that
the solar reflectance of concrete and ULCC are not significantly different, while
that of ULCC+C is statistically higher.

144
 Table 5.2 Heat gain and surface temperatures of the concrete and ULCC with and without coating exposed to simulated sunlight

Surface temperature after 9 hrs

Thermal * Heat reflected Qr Heat loss Ql Heat stored Qs Heat gain Qt
Oven dry density, exposure, °C
conductivity, Solar reflectance
kg/m3
W/m·K Outside
Inside chamber kJ/m2
chamber

Concrete 2251 1.98 0.40 (0.01) 40.5 (0.3) 37.4 (0.5) 4964 2167 1589 3689
ULCC 1303 0.39 0.41 (0.02) 43.2 (0.9) 33.7 (0.7) 5088 3985 857 2479
ULCC+C 1305 0.39 0.78 (0.02) 35.4 (0.7) 29.4 (0.8) 9679 1100 494 1136
*
paired t-test with data from 9 test locations revealed that the surface temperatures of the concrete, ULCC, and ULCC+C are statistically
different.

145
 Table 5.3 Experimental results on material properties and heat gain

Oven dry Thermal Heat gain in

Solar
density , conductivity λ, 9 hrs Qt,
reflectance β
kg/m 3
W/m·K kJ/m2†
TSC-1 0.48 3746
TSC-2 0.21 4746
TSC-3 0.24 4473
WSC-1 0.76 1928
*
WSC-2 1975 1.73 0.37 4039
WSC-3 0.38 3909
PWSC-1 0.79 1900
PWSC-2 0.38 4236
PWSC-3 0.76 1994
Concrete 2251 1.98 0.40 3689
ULCC 1303 0.39 0.41 2479
ULCC+C 1305 0.39 0.78 1136
*
data from ref (The Engineering Toolbox, 2017).
†
the total energy reaches the exposure surface is approximately 12,409 kJ/m2.

Table 5.4 The subset regression analysis of the heat gain (Qt) on oven dry density
( ), thermal conductivity (λ), and solar reflectance (β)

n* Adj-R2, % Mallows’ Cp S λ β
†
1 20.2 269.8 1089.6 x
1 21.2 266.5 1083.1 x
1 84.9 44.7 474.5 x
Qt
2 12.4 268.4 1141.5 x x
2 95.3 8.8 265.2 x x
2 96.6 4.7 225.4 x x
*
“n”, number of considered predictors;
†
“x” denotes considered predictors in the subset regression analysis.

Table 5.5 Coefficients of various models

Obtained coefficients for the parameters from the

Model regression analysis Adj-R2, %
a b c
Equation 5.12:
4397 773 -4810 96.6
Qt=a+bλ+cβ
Equation 5.13:
2951 1.39 -4814 95.3
Qt=a+b +cβ

146
 5.5 Figures

Figure 5.1 Image of cenospheres, cement, silica fume, and sand and their solar reflectance
(SR) values

Figure 5.2 Sketch of heat transfer through opaque building envelope exposed to
simulated sunlight (not to scale)

147
Figure 5.3 Heat flux (a) and surface temperatures (b) of the concrete and ULCC
specimens

Figure 5.4 Heat flux (a) and surface temperatures (b) of the ULCC specimen with and
without coating, under the exposure to the simulated sunlight

Figure 5.5 Heat flux (a) and surface temperatures (b) of the ULCC specimen with
photocatalytic coating and concrete, under exposure to the simulated sunlight

148
Figure 5.6 Normal plot of residuals for equations contain solar reflectance and
thermal conductivity (a) or oven dry density (b)

Figure 5.7 Prediction error of heat gain with equations contain solar reflectance (SR) and
oven dry density or thermal conductivity (TC)

149
Figure 5.8 Predicted heat gain with equations contain solar reflectance and
oven dry density or thermal conductivity

150
 Chapter 6. Conclusions and recommendations

6.1 Summary and conclusions

6.1.1 Development of ultra-lightweight cement composites with low thermal

conductivity and high specific strength for energy efficient buildings

Mechanical properties and thermal conductivity of ultra-lightweight cement

composites (ULCCs) with 1-day density from 1154 to 1471 kg/m3 and 28-day

compressive strength from 33.0 to 69.4 MPa were studied. The low densities of

the ULCCs were achieved by incorporating hollow cenospheres from fly ash

generated in thermal power plants. The properties of the ULCCs were compared

with those of cement pastes with w/b of 0.35 and 0.45 and those of a concrete

with 28-day compressive strength of 67.6 MPa. Based on the experimental results,

numerical calculation, and discussion in Chapter 2, following conclusions can be

drawn:

1. Compressive strength, flexural tensile strength, and elastic modulus of the

ULCCs were reduced with decrease in density. However, the compressive and

flexural tensile strength of 69.4 and 7.3 MPa were achieved for the ULCC,

respectively, similar to those of the cement paste with w/b of 0.35 and the

concrete. The specific strength of the ULCC is 0.047 MPa/kg/m3, equivalent to

that of a normal weight concrete with compressive strength of about 110 MPa.

The density and elastic modulus of the ULCCs and concrete were affected by the

density of aggregates embedded in the cement paste matrices. The ratio of the

151
flexural tensile strength/compressive strength of the ULCCs is comparable to that

of the cement pastes and concrete.

2. Thermal conductivity of the ULCCs was reduced with decrease in the

oven-dry density and it was significantly lower than that of the cement pastes and

concrete due to the incorporation of hollow cenospheres. With similar 28-day

compressive strength, the thermal conductivity of the ULCC was 54% and 80%

lower than that of the cement paste and concrete, respectively.

3. Compared to properties from various lightweight concretes reported in

literature, the ULCCs have a desirable combination of low thermal conductivity,

lightweight, and high specific strength. The low thermal conductivity of the

ULCCs is due to the incorporation of hollow cenospheres as micro-aggregate

which effectively introduce voids and decrease density of the ULCCs. The high

specific strength of the ULCCs may be attributed to (1) the presence of hard and

stiff shell present in the cenospheres, (2) the development of “controlled” void

sizes in the cenospheres, and (3) the creation of strong cement paste matrices that

provide “three dimensional” confinement to the cenospheres.

4. The incorporation of polyethylene fibers reduced brittleness observed in

the ULCC samples, but did not have significant effect on the strength and elastic

modulus if the density and w/b were kept constant. However, the fiber reinforced

ULCCs had higher thermal conductivity than that without fibers, which may be

due to the higher thermal conductivity of the polyethylene fibers. The shrinkage

reducing admixture and viscosity modifying agent did not have significant effect

on the mechanical properties and thermal conductivity of the ULCCs.

152
 5. The thermal conductivity of the ULCCs can be estimated by using

Valore’s equation: λ =0.072e(1.25×ρdry /1000) with a multiplication factor of about 1.2.

6. The estimated average thermal conductivity values of the cenospheres

based on Hashin-Shtrikman lower bound and Series model are significantly

higher than that provided by manufacturer due to the effect of void space among

the cenosphere particles.

7. Based on the estimated thermal conductivity of cenospheres, the thermal

conductivity of the ULCCs can be estimated using Hashin-Shtrikman lower

bound and Series model with error of estimation less than 10%. Overall, the

Hashin-Shtrikman lower bound gives lower error of estimation than the Series

model.

6.1.2 Effects of silica fume and fly ash on properties of plain and fiber reinforced

ultra-lightweight cement composites

Effect of silica fume and fly ash on the stability of cenospheres and steel

fibers and flowability of fresh ULCCs and the amount of fly ash which can be

used to replace Portland cement in ULCCs with silica fume from the perspectives

of pozzolanic reactions and strength development were explored in Chapter 3.

Based on the discussion, the following conclusions can be drawn:

1. The yield stress and plastic viscosity of fresh ULCCs generally increased

with the increase of silica fume content from 0% to 12% by the mass of

cementitious materials, whereas the incorporation of fly ash by 26% and 40% did

not affect these two rheological parameters of the fresh ULCCs.

153
 2. No segregation of cenospheres was observed in the ULCCs with and

without silica fume, whereas severe segregation of steel fibers was observed in

the ULCC without silica fume. Silica fume content of at least 4% is essential in

reducing the segregation and improving the stability of steel fibers in the ULCCs.

Even though the flowability is reduced with the increase in silica fume content,

the ULCCs with up to 12% silica fume had sufficient flowability to cast

specimens without difficulty.

3. The Ca(OH)2 content in cement paste with both silica fume and fly ash, on

cement basis, was higher than combined Ca(OH)2 content in the paste with silica

fume or fly ash alone at the age of 56 and 120 days due to reduced Ca(OH)2

available and possibly lower alkalinity of pore solution in the former. After 120

days of moisture curing, there were 3.6% and 2.2% Ca(OH)2 remained in the

paste mixtures with both 8% silica fume and 26% or 40% fly ash, respectively, on

anhydrous sample basis. This indicates that further increase in fly ash content

beyond certain extent may not lead to significant increase in the pozzolanic

reaction of the fly ash.

4. Incorporation of fly ash reduced the density, compressive and flexural

strengths, and elastic modulus of the ULCCs with given silica fume content at 28

days in general. However, the ULCCs with 26% fly ash and 8% silica fume

achieved 28-day compressive strength of 40 MPa and density of < 1250 kg/m3

with a reasonable cement content of about 460 kg/m3. In addition, the fiber-

reinforced ULCC with 26% fly ash had post cracking performance under bending

comparable to that of the control without fly ash.

154
 Based on the above, optimization of the silica fume and fly ash contents is

needed by considering the flowability, stability of steel fibers, pozzolanic

reactions, and strength development of the ULCCs.

6.1.3 Using photocatalytic coating to maintain solar reflectance and lower

cooling energy consumption of buildings

The ability of photocatalytic coatings to maintain solar reflectance of opaque

building envelope and lower the cooling energy consumption was demonstrated

experimentally in Chapter 4 by comparing the heat transfer and surface

temperature of Portland cement mortar specimens with transparent silicate coating

(TSC), white silicate coating (WSC), and white silicate coating with photocatalyst

(PWSC) under controlled experimental conditions. The solar reflectance, gray

scale L*, heat gain, and surface temperatures of the specimens exposed to a sun

simulator were determined at three different time points: (1) before black carbon

deposition; (2) after black carbon deposition; and (3) after 300 hrs exposure to

simulated solar irradiation in an accelerated weathering chamber. Based on the

results and discussion, following conclusions are drawn:

1. Specimens with the WSC and PWSC have higher solar reflectance and L*

and appeared whiter than the specimen with TSC. After 9 hrs of exposure to the

sun simulator, the heat gain through the WSC and PWSC specimens is almost 50%

lower than that through the TSC specimen, and the temperatures of the inner and

exposure surfaces of the formers are about 5 °C and 6 °C lower than those of the

latter. The lower heat gain and surface temperatures of the WSC and PWSC

specimens can be attributed to their higher solar reflectance.

155
 2. Black carbon deposition reduces the solar reflectance and L* of the coated

specimens. After 9 hrs of exposure to the sun simulator, the heat gain in the

chamber through the specimens with black carbon deposition is increased by 27%,

109%, and 123% for the specimens with TSC, WSC, and PWSC, respectively,

and the temperatures of the inner and exposure surfaces of the specimens are also

increased significantly, indicating increased cooling energy consumption with

soiling of building surfaces. The greater increase in the heat gains through the

WSC and PWSC specimens may be attributed to more significant reduction of

their solar reflectance by black carbon deposition.

3. After 300 hrs of exposure to simulated solar irradiation, the solar

reflectance and L* of the PWSC specimen with black carbon deposition were

almost completely recovered to the initial values. The surface temperatures and

heat gain through the PWSC specimens are comparable to those before the black

carbon deposition. The inner surface temperature and heat gain of the specimen

with PWSC after 300 hrs exposure to simulated solar irradiation are reduced by

7.2 °C and 2242 kJ/m2 (53%), respectively, compared to the specimen after black

carbon deposition. The solar reflectance and L* of the TSC and WSC specimens,

however, do not change significantly after the exposure to the simulated solar

irradiation, and their heat gain and surface temperatures are comparable to those

after the black carbon deposition.

The above suggests that photocatalytic coating can remove black carbon and

maintain solar reflectance of opaque building surfaces. As a result, it can reduce

156
cooling energy consumption of buildings in tropical countries and in summer

months in temperate countries.

6.1.4 Effect of ultra-lightweight cement composite and photocatalytic coating on

heat transfer through opaque building envelope

In Chapter 5, effects of a structural ULCC with low thermal conductivity and

a photocatalytic coating with high solar reflectance on the heat transfer through

panel specimens were investigated experimentally in a controlled environment.

Empirical equations to predict the heat gain through construction materials with

different thermal conductivities and solar reflectance were developed. Based on

the results and discussion, following conclusions can be drawn:

1. Compared with a conventional concrete of similar 28-day compressive

strength, the ULCC has 80% lower thermal conductivity (0.39 W/m·K) which

reduces the heat gain through the ULCC by 33% and decreases the inner surface

temperature of the specimen by 3.7 °C after exposure to simulated sunlight for 9

hours.

2. The photocatalytic coating increases the solar reflectance of the ULCC

specimen from 0.41 to 0.78 which reduces the heat gain by 54% and decreases the

inner surface temperature of the specimen by 4.3 °C after exposure to simulated

sunlight for 9 hrs.

3. The combination of the ULCC with low thermal conductivity and

photocatalytic coating with high solar reflectance reduces the heat gain and inner

surface temperature by 69% and 8.0 °C, respectively, after 9 hrs exposure to

simulated sunlight.
157
 4. The empirical model with the thermal conductivity and solar reflectance

provides good prediction of the heat gain with adj-R2 of 96.6%. In addition, the

model with the oven dry density and solar reflectance can also provide reasonable

prediction for the heat gain with adj-R2 of 95.3% when thermal conductivity data

are not available.

5. Based on the coefficients of the regression equations, the heat gain is more

sensitive to the change of the solar reflectance than that of the thermal

conductivity and oven dry density. Increase in the solar reflectance by 0.1 is

equivalent to reduction in the thermal conductivity by approximately 0.6 W/m·K

or reduction in the oven dry density by approximately 350 kg/m3.

6.2 Significance of this research

This study has developed and evaluated the ultra-lightweight cement

composites and photocatalytic coating for energy efficient buildings.

Compared with properties of various lightweight concretes reported in

literature, the ULCCs developed in this study have desirable combinations of low

thermal conductivity, lightweight, and high specific strength. The thermal

conductivity of the ULCC was 80% lower than that of concrete of similar 28-day

compressive strength.

Effects of silica fume and fly ash on properties of fresh and hardened ULCCs

was explored, which provides guideline to optimized mix design of ULCCs with

mineral admixtures. The incorporation of mineral admixtures has economical,

technical, and environmental benefits.

158
 The photocatalytic coating with TiO2 has been used on mortar surfaces to

maintain solar reflectance and lower cooling energy consumption by removal of

deposited black carbon. This part experimentally demonstrated that the

photocatalytic coating can be used on opaque building surfaces to maintain solar

reflectance and reduce cooling energy consumption. Compared with building

surfaces with ordinary coatings, the application of the photocatalytic coatings of

comparable solar reflectance may reduce cost from both maintenance such as

washing and bleaching and cooling energy consumption of buildings in tropical

countries and in summer months in temperate countries.

Individual and combined effects of the ULCC with reduced thermal

conductivity and photocatalytic coating with increased solar reflectance on the

heat transfer through building envelope were studied and empirical equations

were established. The effectiveness of reducing heat gain through building

envelope by increasing solar reflectance and reducing thermal conductivity was

compared, which provides guideline for future development of materials for

energy efficient buildings. In addition, with the empirical equations developed,

the heat gain through opaque building envelope under controlled measurement

condition can be easily estimated based on its thermal conductivity, solar

reflectance, and oven dry density.

6.3 Recommendations for future research

Ultra-lightweight cement composites with significant lower density and

thermal conductivity but comparable compressive strength to conventional

concrete were developed. However, the elastic modulus of ULCCs was

159
significantly lower due to the incorporation of hollow cenospheres which needs to

be taken into consideration in design and the application of the ULCCs for

structures.

Since the rheological behavior of the ULCCs and Ca(OH)2 content in cement

pastes are affected by w/b and the former is also affected by cenosphere content,

silica fume content required for the stability of steel fibers in the ULCCs and its

effect on the pozzolanic reaction of fly ash may vary with different mix

proportions. More systemic research is needed to obtain information on the

necessary silica fume content for the stability of steel fibers and its effect on the

pozzolanic reaction of fly ash in the ULCCs with various mix proportions.

The removal of black carbon by photocatalytic coating and heat transfer

through panel specimens were investigated under controlled condition in

laboratory to simulate natural environment. However, ambient meteorological

conditions and air quality may affect the degradation ability of the photocatalytic

coatings. While this work provides an insight into the prospect of the

photocatalytic coatings, additional efforts are needed to formulate economical and

optimal photocatalytic coatings for field applications. Further, large scale testing

is needed to develop coating method and quality control approach for practical

applications. In addition, the long-term performance of the ULCC and

photocatalytic coating exposed to the ambient environment needs to be evaluated.

160
 References

ACI-122R (2002). Guide to thermal properties of concrete and masonry systems.

1-21. American Concrete Institute.

ACI-232.2R (2002). Use of fly ash in concrete. 1-34. American Concrete Institute.

ACI-234R (2006). Guide for the use of silica fume in concrete. 1-63. American

Concrete Institute.

ACI 318-11. (2010). Building code requirements for structural concrete.

American Concrete Institute.

Akbari, H., Menon, S., & Rosenfeld, A. (2009). Global cooling: increasing world-

wide urban albedos to offset CO2. Climatic Change, 94(3), 275-286.

Al-Homoud, M. S. (2005). Performance characteristics and practical applications

of common building thermal insulation materials. Building and

Environment, 40(3), 353-366.

Alvarado, J. L., Terrell, W., & Johnson, M. D. (2009). Passive cooling systems

for cement-based roofs. Building and Environment, 44(9), 1869-1875.

Aoyama, T., Sonoda, T., Nakanishi, Y., Tanabe, J., & Takebayashi, H. (2016).

Study on aging of solar reflectance of the self-cleaning high reflectance

coating. the 4th International Conference on Countermeasures to Urban

Heat Island, Singapore.

ASTM C138-13. (2013). Standard test method for density (unit weight), yield,

and air content (gravimetric) of concrete. 1-4. ASTM International.

ASTM C150-17. (2017). Standard Specification for Portland cement. 1-9. ASTM

International.
161
ASTM C177-13. (2013). Standard test method for steady-state heat flux

measurements and thermal transmission properties by means of the

guarded-hot-plate apparatus. 1-23. ASTM International.

ASTM C227-10. (2010). Standard test method for potential alkali reactivity of

cement-aggregate combinations (mortar-bar method). 1-6. ASTM

International.

ASTM C518-10. (2010). Standard test method for steady-state thermal

transmission properties by means of the heat flow meter apparatus. 1-16.

ASTM International.

ASTM C567-00. (2000). Standard test method for determining density of

structural lightweight concrete. 1-3. ASTM International.

ASTM C1363-11. (2011). Standard test method for thermal performance of

building materials and envelope assemblies by means of a hot box

apparatus. 1-44. ASTM International.

ASTM C1549-09. (2009). Standard test method for determination of solar

reflectance near ambient temperature using a portable solar reflectometer.

1-6. ASTM International.

ASTM C143-03. (2003). Standard test method for slump of hydraulic-cement

concrete. 1-4. ASTM International.

ASTM C469-14. (2014). Standard test method for static modulus of elasticity and

poisson’s ratio of concrete in compression. 1-5. ASTM International.

ASTM C1260-14. (2014). Standard test method for potential alkali reactivity of

aggregates (mortar-bar method). 1-5. ASTM International.

162
ASTM C1609-07. (2007). Standard test method for flexural performance of fiber-

reinforced concrete (using beam with third-point loading). 1-9. ASTM

International.

ASTM C1749. (2012). Standard guide for measurement of the rheological

properties of hydraulic cementious paste using a rotational rheometer.

ASTM International.

ASTM D7356. (2013). Standard test method for accelerated acid etch weathering

of automotive clearcoats using a Xenon-Arc exposure device. 1-10.

ASTM International.

Auvinen, J., & Wirtanen, L. (2008). The influence of photocatalytic interior paints

on indoor air quality. Atmospheric Environment, 42(18), 4101-4112.

Balaras, C. A. (1996). The role of thermal mass on the cooling load of buildings.

An overview of computational methods. Energy and Buildings, 24(1), 1-

10.

Banfill, P. F. G. (2006). Rheology of fresh cement and concrete. Rheology

Reviews, 61-130.

Barbare, N., Shukla, A., & Bose, A. (2003). Uptake and loss of water in a

cenosphere-concrete composite material. Cement and Concrete Research,

33(10), 1681-1686.

Behind the success of zero energy building. (2017).

https://www.nccs.gov.sg/climatechallenge/issue05/green-tech.html.

Bentur, A., & Mindess, S. (2006). Fibre reinforced cementitious composites. CRC

Press.

163
Bentz, D. P., Ferraris, C. F., Galler, M. A., Hansen, A. S., & Guynn, J. M. (2012).

Influence of particle size distributions on yield stress and viscosity of

cement-fly ash pastes. Cement and Concrete Research, 42(2), 404-409.

Berdahl, P., Akbari, H., & Rose, L. S. (2002). Aging of reflective roofs soot

deposition. Applied Optics, 41(12), 2355-2360.

Beris, A. N., Tsamopoulos, J. A., Armstrong, R. C., & Brown, R. A. (1985).

Creeping motion of a sphere through a Bingham plastic. Journal of Fluid

Mechanics, 158(1), 219-244.

Blanco, F., GarcõÂa, P., Mateos, P., & Ayala, J. (2000). Characteristics and

properties of lightweight concrete manufactured with cenospheres. Cement

and Concrete Research, 30, 1715-1722.

Bretz, S. E., & Akbari, H. (1997). Long-term performance of high-albedo roof

coatings. Energy and Buildings, 25(2), 159-167.

Brodowicz, K., & Dyakowski, T. (1993). Heat pumps. Boston;Oxford [England];.

Butterworth-Heinemann.

BS EN 197-1. (2014). Cement. Composition, specifications and conformity

criteria for common cements. 1-50. British Standards Institute.

BS EN 1015-3. (1999). Methods of test for mortar for masonry - Part 3:

determination of consistence of fresh mortar (by flow table). 1-10. British

Standards Institute.

BS EN 12390-3. (2009). Testing hardened concrete Part 3: compressive strength

of test specimens. 1-22. British Standards Institute.

164
BS EN 12390-7. (2009). Testing hardened concrete Part 7: density of hardened

concrete. 1-14. British Standards Institute.

Buseck, P., Adachi, K., Gelencsér, A., Tompa, É., & Pósfai, M. (2012). Are black

carbon and soot the same? Atmospheric Chemistry and Physics

Discussions, 12(9), 24821-24846.

California Energy Commission. (2015). 2016 Building energy efficiency

standards for residential and nonresidential buildings.

Callister, W. D., & Rethwisch, D. G. (2008). Fundamentals of materials science

and engineering: an integrated approach (3rd edition). Hoboken, N.J. John

Wiley & Sons.

Cardoso, R. J., Shukla, A., & Bose, A. (2002). Effect of particle size and surface

treatment on constitutive properties of polyester-cenosphere composites.

Journal of Materials Science, 37, 603-613.

Chartered institution of building services engineers (CIBSE). Energy efficiency in

buildings: CIBSE guide F. (2004). London.

Chen, F., & Wittkopf, S. K. (2012). Summer condition thermal transmittance

measurement of fenestration systems using calorimetric hot box. Energy

and Buildings, 53, 47-56.

Chen, F., Wittkopf, S. K., Ng, P. K., & Du, H. (2012). Solar heat gain coefficient

measurement of semi-transparent photovoltaic modules with indoor

calorimetric hot box and solar simulator. Energy and Buildings, 53, 74-84.

165
Chen, J., & Poon, C.-S. (2009). Photocatalytic construction and building materials:

from fundamentals to applications. Building and Environment, 44(9),

1899-1906.

Cheng, M.-D., Miller, W., New, J., & Berdahl, P. (2012). Understanding the long-

term effects of environmental exposure on roof reflectance in California.

Construction and Building Materials, 26(1), 516-526.

Cheung, C. K., Fuller, R. J., & Luther, M. B. (2005). Energy-efficient envelope

design for high-rise apartments. Energy and Buildings, 37(1), 37-48.

Chia, K. S., Kho, C. C., & Zhang, M. H. (2005). Stability of fresh lightweight

aggregate concrete under vibration. ACI Materials Journal, 102(5), 347-

354.

Chia, K. S., & Zhang, M. H. (2004). Effect of chemical admixtures on rheological

parameters and stability of fresh lightweight aggregate concrete. Magazine

of Concrete Research, 56(8), 465-473.

Chin, P., Roberts, G. W., & Ollis, D. F. (2007). Kinetic modeling of

mhotocatalyzed soot oxidation on titanium dioxide thin films. Industrial &

Engineering Chemistry Research, 46(23), 7598-7604.

Chua, K. J., & Chou, S. K. (2010). Energy performance of residential buildings in

Singapore. Energy, 35(2), 667-678.

Clayton, R. M., & Back, L. H. (1989). Physical and chemical characteristics of

cenospheres from the combustion of heavy fuel oil. Journal of

Engineering for Gas Turbines and Power, 111(4), 679.

166
Cyr, M., Legrand, C., & Mouret, M. (2000). Study of the shear thickening effect

of superplasticizers on the rheological behaviour of cement pastes

containing or not mineral additives. Cement and Concrete Research, 30(9),

1477-1483.

Demirboğa, R., & Gül, R. (2003). he effects of expanded perlite aggregate, silica

fume and fly ash on the thermal conductivity of lightweight concrete.

Cement and Concrete Research, 33(5), 723-727.

Department of Energy. (2017a). A common definition for zero energy buildings.

https://energy.gov/eere/buildings/downloads/common-definition-zero-

energy-buildings.

Department of Energy. (2017b). Cool roofs. https://energy.gov/energysaver/cool-

roofs.

El-Chabib, H., & Nehdi, M. (2006). Effect of mixture design parameters on

segregation of self-consolidating concrete. ACI Materials Journal, 103(5),

374-383.

Favez, O., Cachier, H., Chabas, A., Ausset, P., & Lefevre, R. (2006). Crossed

optical and chemical evaluations of modern glass soiling in various

European urban environments. Atmospheric Environment, 40(37), 7192-

7204.

Federation internationale de la, precontrainte (FIP). (1983) FIP manual of

lightweight aggregate concrete (2nd edition). New York; Glasgow. Surrey

University Press.

167
Feys, D., Verhoeven, R., & De Schutter, G. (2008). Fresh self compacting

concrete, a shear thickening material. Cement and Concrete Research,

38(7), 920-929.

Feys, D., Verhoeven, R., & De Schutter, G. (2009). Why is fresh self-compacting

concrete shear thickening? Cement and Concrete Research, 39(6), 510-

523.

Gao, T., Jelle, B. P., Gustavsen, A., & Jacobsen, S. (2014). Aerogel-incorporated

concrete: an experimental study. Construction and Building Materials, 52,

130-136.

Geiss, O., Cacho, C., Barrero-Moreno, J., & Kotzias, D. (2012). Photocatalytic

degradation of organic paint constituents-formation of carbonyls. Building

and Environment, 48, 107-112.

Gül, R., Okuyucu, E., ürkmen, İ., & Aydin, A. . (2007). hermo-mechanical

properties of fiber reinforced raw perlite concrete. Materials Letters,

61(29), 5145-5149.

Hajmohammadian Baghban, M., Hovde, P. J., & Jacobsen, S. (2013). Analytical

and experimental study on thermal conductivity of hardened cement pastes.

Materials and Structures, 46(9), 1537-1546.

Han, J., Lu, L., & Yang, H. (2009). Investigation on the thermal performance of

different lightweight roofing structures and its effect on space cooling load.

Applied Thermal Engineering, 29, 2491-2499.

168
Hanif, A., Parthasarathy, P., Ma, H., Fan, T., & Li, Z. (2017). Properties

improvement of fly ash cenosphere modified cement pastes using nano

silica. Cement and Concrete Composites, 81, 35-48.

Harmathy, T. (1970). Thermal properties of concrete at elevated temperatures.

Journal of Materials, 47-74.

Hashin, Z., & Shtrikman, S. (1962). A variational approach to the theory of the

effective magnetic permeability of multiphase materials. Journal of

Applied Physics, 33(10), 3125-3131.

Hernández-Ramírez, A. & Medina-Ramírez, I. (2015). Photocatalytic

Semiconductors: Synthesis, Characterization, and Environmental

Applications, (2015th edition), Springer International Publishing, Cham.

Huang, X., Ranade, R., Zhang, Q., Ni, W., & Li, V. C. (2013). Mechanical and

thermal properties of green lightweight engineered cementitious

composites. Construction and Building Materials, 48, 954-960.

Huang, Y., Niu, J.-L., & Chung, T.-M. (2013). Study on performance of energy-

efficient retrofitting measures on commercial building external walls in

cooling-dominant cities. Applied Energy, 103, 97-108.

International Energy Agency. (2008). Promoting energy efficiency investments:

case studies in the residential sector. Paris. OECD Publishing.

Ishida, F. (1993). The interfacial interactions in polymeric composites. Kluwer

Academic Publisher.

169
Jelle, B. P. (2011). Traditional, state-of-the-art and future thermal building

insulation materials and solutions - Properties, requirements and

possibilities. Energy and Buildings, 43(10), 2549-2563.

Jiang, J., Oberdorster, G., Elder, A., Gelein, R., Mercer, P., & Biswas, P. (2008).

Does nanoparticle activity depend upon size and crystal phase?

Nanotoxicology, 2(1), 33-42.

Jones, D. L. (1998). Architecture and the environment: bioclimatic building

design. London. Laurence King.

Kelen, D., Rosana, C., & Eduvaldo, S. (2015). Natural weathering of cool

coatings and its effect on solar reflectance of roof surfaces. Energy

Procedia, 78, 1587-1592.

Khayat, K. H. (1998). Viscosity-enhancing admixtures for cement-based materials:

an overview. Cement and Concrete Composites, 20(2-3), 171-188.

Kim, H. K., Jeon, J. H., & Lee, H. K. (2012). Workability, and mechanical,

acoustic and thermal properties of lightweight aggregate concrete with a

high volume of entrained air. Construction and Building Materials, 29,

193-200.

Kim, K.-H. (2003). An experimental study on thermal conductivity of concrete.

Cement and Concrete Research, 33, 363-371.

Kolokotsa, D.-D., Santamouris, M., & Akbari, H. (2013). Advances in the

development of cool materials for the built environment. Bentham Science

Publishers

170
Krishnan, P. (2015). Photocatalytic constrction materials for degradation of air

pollutants. (Ph.D.), National University of Singapore, Singapore.

Krishnan, P., Zhang, M.-H., Cheng, Y., Riang, D. T., & Yu, L. E. (2013a).

Photocatalytic degradation of SO2 using TiO2-containing silicate as a

building coating material. Construction and Building Materials, 43, 197-

202.

Krishnan, P., Zhang, M.-H., Yu, L., & Feng, H. (2013b). Photocatalytic

degradation of particulate pollutants and self-cleaning performance of

TiO2-containing silicate coating and mortar. Construction and Building

Materials, 44, 309-316.

Krishnan, P., Zhang, M.-H., & Yu, L. E. (2017). Removal of black carbon using

photocatalytic silicates: laboratory and field studies. Journal of Cleaner

Production (Submitted).

Kua, H. W., & Wong, C. L. (2012). Analysing the life cycle greenhouse gas

emission and energy consumption of a multi-storied commercial building

in Singapore from an extended system boundary perspective. Energy and

Buildings, 51, 6-14.

Larrard, F. d., Ferraris, C. F., & Sedran, T. (1998). Fresh concrete a Herschel

Bulkley material. Materials and Structures, 31, 494-498.

Lee, M. C., & Choi, W. (2002). Solid phase photocatalytic reaction on the soot

TiO2 interface: the role of migrating O radicals. Journal of Physical

Chemistry Part B, 106(45), 11818-11822.

171
Levinson, R., & Akbari, H. (2002). Effects of composition and exposure on the

solar reflectance of portland cement concrete. Cement and Concrete

Research, 32, 1679-1698.

Levinson, R., Akbari, H., Berdahl, P., Wood, K., Skilton, W., & Petersheim, J.

(2010). A novel technique for the production of cool colored concrete tile

and asphalt shingle roofing products. Solar Energy Materials and Solar

Cells, 94(6), 946-954.

Levinson, R., Berdahl, P., Asefawberhe, A., & Akbari, H. (2005). Effects of

soiling and cleaning on the reflectance and solar heat gain of a light-

colored roofing membrane. Atmospheric Environment, 39(40), 7807-7824.

Levinson, R., & Hunter, S. (2014). CERC-BEE cool roofs and urban heat islands:

infrastructure and anti-soiling coatings, 2014 building technologies office

peer review. Retrieved November, 2016.

http://www.energy.gov/eere/buildings/downloads/urban-heat-islands-anti-

soiling-cool-roof-coatings.

Lim, B. B. (1988). Control of the external environment of buildings: selected

papers on the protection of the external surfaces of buildings in warm

humid climate. National University of Singapore Press.

Liu, X., Zhang, M.-H., Chia, K. S., Yan, J., & Liew, J. Y. R. (2016). Mechanical

properties of ultra-lightweight cement composite at low temperatures of 0

to -60 °C. Cement and Concrete Composites, 73, 289-298.

172
Lothenbach, B., Scrivener, K., & Hooton, R. D. (2011). Supplementary

cementitious materials. Cement and Concrete Research, 41(12), 1244-

1256.

Maranzano, B. J., & Wagner, N. J. (2001). The effects of particle size on

reversible shear thickening of concentrated colloidal dispersions. The

Journal of Chemical Physics, 114(23), 10514-10527.

Marceau, M. L., & Vangeem, M. G. (2008). Solar reflectance value for concrete.

Concrete international, 52-58.

Mark, J. E. (2007). Physical properties of polymers handbook (2 nd edition).

Springer Science.

Mehta, P. K., & Monteiro, P. J. M. (2005). Concrete: microstructure, properties,

and materials. New York. McGraw-Hill.

Mills, A., Wang, J., & Crow, M. (2006). Photocatalytic oxidation of soot by P25

TiO2 films. Chemosphere, 64(6), 1032-1035.

Mindess, S., Young, J. F., & Darwin, D. (2003). Concrete. Upper Saddle River,

NJ. Prentice Hall.

Montgomery, D., & Diamond, S. (1984). The influence of fly ash cenospheres on

the details of cracking in fly ash-bearing cement pastes. Cement and

Concrete Research, 14, 767-775.

Montgomery, D. C. (2011). Applied statistics and probability for engineers (5th

edition, International). Chichester;Hoboken, N.J;. Wiley.

Monticelli, F., Toledano, M., Osorio, R., & Ferrari, M. (2006). Effect of

temperature on the silane coupling agents when bonding core resin to

173
 quartz fiber posts. Mechanics of Time Dependent Materials, 22(11), 1024-

1028.

Mounanga, P., Gbongbon, W., Poullain, P., & Turcry, P. (2008). Proportioning

and characterization of lightweight concrete mixtures made with rigid

polyurethane foam wastes. Cement and Concrete Composites, 30(9), 806-

814.

Navarrete, I., & Lopez, M. (2016). Estimating the segregation of concrete based

on mixture design and vibratory energy. Construction and Building

Materials, 122, 384-390.

Nehdi, M., Mindess, S., & A tcin, P. C. (1998). Rheology of high-performance

concrete: effect of ultrafine particles. Cement and Concrete Research,

28(5), 687-697.

Paolini, R., Zinzi, M., Poli, T., Carnielo, E., & Mainini, A. G. (2014). Effect of

ageing on solar spectral reflectance of roofing membranes: natural

exposure in Roma and Milano and the impact on the energy needs of

commercial buildings. Energy and Buildings, 84, 333-343.

Papadopoulos, A. M. (2005). State of the art in thermal insulation materials and

aims for future developments. Energy and Buildings, 37(1), 77-86.

Park, C. K., Noh, M. H., & Park, T. H. (2005). Rheological properties of

cementitious materials containing mineral admixtures. Cement and

Concrete Research, 35(5), 842-849.

174
Parker, D. S., Parker, D. S., McIlvaine, J., Barkaszi, S. F., & Beal, D. (1993).

Laboratory testing of the reflectance properties of roofing materials.

Building Design Assistance Center, Florida Solar Energy Center.

Petrou, M. F., Wan, B., Gadala-Maria, F., Kolli, V. G., & Harries, K. A. (2000).

Influence of mortar rheology on aggregate settlement. ACI Materials

Journal.

Pozo-Antonio J. S., Dionisio A., (2017). Self-cleaning property of mortars with

TiO2 addition using real diesel exhaust soot. Journal of Cleaner

Production, 160, 850-859.

Puddu, V., Choi, H., Dionysiou, D. D., & Puma, G. L. (2010). TiO 2 photocatalyst

for indoor air remediation: influence of crystallinity, crystal phase, and

UV radiation intensity on trichloroethylene degradation. Applied Catalysis

B: Environmental, 94(3), 211-218.

Qin, J., Song, J., Qu, J., Xue, X., Zhang, W., Song, Z., Shi, Y., Jiang, L., Li, J., &

Zhang, T. (2014). The methods for creating building energy efficient cool

black coatings-part I: preparation from white and black pigments. Energy

and Buildings, 84, 308-315.

Ramirez, A. M., Demeestere, K., De Belie, N., Mäntylä, T., & Levänen, E. (2010).

Titanium dioxide coated cementitious materials for air purifying purposes:

Preparation, characterization and toluene removal potential. Building and

Environment, 45(4), 832-838.

Revel, G. M., Martarelli, M., Emiliani, M., Celotti, L., Nadalini, R., Ferrari, A. D.,

Hermanns, S., & Beckers, E. (2014). Cool products for building envelope

175
 - Part II: experimental and numerical evaluation of thermal performances.

Solar Energy, 105, 780-791.

Roth, M. (2015). Geography weather station.

https://inetapps.nus.edu.sg/fas/geog/ajxdirList.aspx.

Scanlon, J. M., & McDonald, J. E. (1994). Thermal properties. Significance of

Tests and Properties of Concrete and Concrete-Making Materials. ASTM

International.

Schackow, A., Effting, C., Folgueras, M. V., Güths, S., & Mendes, G. A. (2014).

Mechanical and thermal properties of lightweight concretes with

vermiculite and EPS using air-entraining agent. Construction and Building

Materials, 57, 190-197.

Schiavoni, S., D‫׳‬Alessandro, F., Bianchi, F., & Asdrubali, F. (2016). Insulation

materials for the building sector: a review and comparative analysis.

Renewable and Sustainable Energy Reviews, 62, 988-1011.

Scrivener, K., Snellings, R., & Lothenbach, B. (2015). A practical guide to

microstructural analysis of cementitious materials. Crc Press.

Sengul, O., Azizi, S., Karaosmanoglu, F., & Tasdemir, M. A. (2011). Effect of

expanded perlite on the mechanical properties and thermal conductivity of

lightweight concrete. Energy and Buildings, 43, 671-676.

Shaughnessy, R., & Clark, P. E. (1988). The rheological behavior of fresh cement

pastes. Cement and Concrete Research, 18, 327-341.

176
Shehata, M. H., & Thomas, M. D. A. (2002). Use of ternary blends containing

silica fume and fly ash to suppress expansion due to alkali-silica reaction

in concrete. Cement and Concrete Research, 32(3), 341-349.

Shi, Y., Song, Z., Zhang, W., Song, J., Qu, J., Wang, Z., Li, Y., Xu, L., & Lin, J.

(2013). Physicochemical properties of dirt-resistant cool white coatings

for building energy efficiency. Solar Energy Materials and Solar Cells,

110, 133-139.

Simpson, J. R., & Mcpherson, E. G. (1997). The effects of roof albedo

modification on cooling loads of scale model residences in Tucson,

Arizona. Energy and Buildings, 25, 127-137.

Sleiman, M., Ban-Weiss, G., Gilbert, H. E., François, D., Berdahl, P., Kirchstetter,

T. W., Destaillats, H., & Levinson, R. (2011). Soiling of building envelope

surfaces and its effect on solar reflectance-Part I: analysis of roofing

product databases. Solar Energy Materials and Solar Cells, 95(12), 3385-

3399.

Sleiman, M., Destaillats, H., & Levinson, R. (2014a). Stay-clean and durable

white elastomeric roof coatings, 2014 building technologies office peer

review. Retrieved November, 2016.

http://www.energy.gov/eere/buildings/downloads/stay-clean-and-durable-

white-elastomeric-roof-coatings.

Sleiman, M., Kirchstetter, T. W., Berdahl, P., Gilbert, H. E., Quelen, S., Marlot,

L., Preble, C. V., Chen, S., Montalbano, A., Rosseler, O., Akbari, H.,

Levinson, R., & Destaillats, H. (2014b). Soiling of building envelope

177
 surfaces and its effect on solar reflectance-Part II: development of an

accelerated aging method for roofing materials. Solar Energy Materials

and Solar Cells, 122, 271-281.

Smits, M., Chan, C. k., Tytgat, T., Craeye, B., Costarramone, N., Lacombe, S., &

Lenaerts, S. (2013). Photocatalytic degradation of soot deposition: self-

cleaning effect on titanium dioxide coated cementitious materials.

Chemical Engineering Journal, 222, 411-418.

Smits, M., Ling, Y., Lenaerts, S., & Van Doorslaer, S. (2012). Photocatalytic

removal of soot: unravelling of the reaction mechanism by EPR and in situ

FTIR spectroscopy. Chemphyschem, 13(18), 4251-4257.

Song, Z., Qin, J., Qu, J., Song, J., Zhang, W., Shi, Y., Zhang, T., Xue, X., Zhang,

R., Zhang, H., Zhang, Z., & Wu, X. (2014). A systematic investigation of

the factors affecting the optical properties of near infrared transmitting

cool non-white coatings. Solar Energy Materials and Solar Cells, 125,

206-214.

Struble, L., Szecsy, R., Lei, W.-G., & Sun, G.-K. (1998). Rheology of cement

paste and concrete. Cement, concrete and aggregates, 20(2), 269-277.

Suehrcke, H., Peterson, E. L., & Selby, N. (2008). Effect of roof solar reflectance

on the building heat gain in a hot climate. Energy and Buildings, 40(12),

2224-2235.

Sun, Q., & Xu, Y. (2010). Evaluating intrinsic photocatalytic activities of anatase

and rutile TiO2 for organic degradation in water. The Journal of Physical

Chemistry C, 114(44), 18911-18918.

178
Synnefa, A., & Santamouris, M. (2013). White or light colored cool roofing

materials. Advances in the Development of Cool Materials for the Built

Environment (pp. 33-71).

Synnefa, A., Santamouris, M., & Akbari, H. (2007a). Estimating the effect of

using cool coatings on energy loads and thermal comfort in residential

buildings in various climatic conditions. Energy and Buildings, 39(11),

1167-1174.

Synnefa, A., Santamouris, M., & Apostolakis, K. (2007b). On the development,

optical properties and thermal performance of cool colored coatings for

the urban environment. Solar Energy, 81(4), 488-497.

Takebayashi, H., Miki, K., Sakai, K., Murata, Y., Matsumoto, T., Wada, S., &

Aoyama, T. (2016). Experimental examination of solar reflectance of

high-reflectance paint in Japan with natural and accelerated aging. Energy

and Buildings, 114, 173-179.

Tandiroglu, A. (2010). Temperature-dependent thermal conductivity of high

strength lightweight raw perlite aggregate concrete. International Journal

of Thermophysics, 31(6), 1195-1211.

Taylor, H. F. W. (1990). Cement chemistry. Academic Press.

The Engineering Toolbox. (2017). Thermal conductivity of some common

materials and gases. http://www.engineeringtoolbox.com/thermal-

conductivity-d_429.html

opçu, İ. B., & Uygunoğlu, . (2007). Properties of autoclaved lightweight

aggregate concrete. Building and Environment, 42(12), 4108-4116.

179
Twomey, S. A., Bohren, C. F., & Mergenthaler, J. L. (1986). Reflectance and

albedo differences between wet and dry surfaces. Applied Optics, 25(3),

431-437.

Uemoto, K. L., Sato, N. M. N., & John, V. M. (2010). Estimating thermal

performance of cool colored paints. Energy and Buildings, 42(1), 17-22.

United Nations Environment Programme. (2007). Buildings and climate change

status: challenges and opportunities.

Uysal, ., Demirboğa, R., Şahin, R., & Gül, R. (2004). he effects of different

cement dosages, slumps, and pumice aggregate ratios on the thermal

conductivity and density of concrete. Cement and Concrete Research,

34(5), 845-848.

Valore, R. C. (1980). Calculation of U-values of hollow concrete mosnory.

Concrete International, 2(2), 40-63.

V. Rheinheimer, Y. Wu, T. Wu, K. Celik, J.-Y. Wang, L.D. Lorenzis, P. Wriggers,

M.-H. Zhang, P. J.M. Monteiro. (2017). Multi-scale study of high-strength

low-Thermal conductivity cement composites containing cenospheres.

Cement and Concrete Composites, 80, 91-103.

Wallevik, J. E. (2006). Relationship between the Bingham parameters and slump.

Cement and Concrete Research, 36(7), 1214-1221.

Wandell, T. (1996). Cenospheres: from waste to profits. The American Ceramic

Bulletin, 75(6), 79-81.

180
Wang, J.-Y., Banthia, N., & Zhang, M.-H. (2012a). Effect of shrinkage reducing

admixture on flexural behaviors of fiber reinforced cementitious

composites. Cement and Concrete Composites, 34(4), 443-450.

Wang, J.-Y., Chia, K.-S., Liew, J. Y. R., & Zhang, M.-H. (2013). Flexural

performance of fiber-reinforced ultra lightweight cement composites with

low fiber content. Cement and Concrete Composites, 43, 39-47.

Wang, J.-Y., Zhang, M.-H., Li, W., Chia, K.-S., & Liew, J. Y. R. (2012b).

Stability of cenospheres in lightweight cement composites in terms of

alkali-silica reaction. Cement and Concrete Research, 42(5), 721-727.

Wang, R., & Meyer, C. (2012). Performance of cement mortar made with

recycled high impact polystyrene. Cement and Concrete Composites,

34(9), 975-981.

Weng, J. K., Langan, B. W., & Ward, M. A. (1997). Pozzolanic reaction in

portland cement, silica fume, and fly ash mixtures. Canadian Journal of

Civil Engineering, 24(5), 754-760.

What is zero energy building? (2017). https://www.bca.gov.sg/zeb/whatiszeb.html.

Whetsell, J., Saha, M. C., Altan, M. C., Liang, J., & Pan, C. (2015). Investigation

of hygrothermal effects on the thermal conductivity characteristics of

insulation materials. ASHRAE Transactions, 121.

White roof project. (2016). http://www.whiteroofproject.org/.

Wu, Y., Wang, J.-Y., Monteiro, P. J. M., & Zhang, M.-H. (2015). Development of

ultra-lightweight cement composites with low thermal conductivity and

181
 high specific strength for energy efficient buildings. Construction and

Building Materials, 87, 100-112.

Xu, Y., & Chung, D. D. L. (2000). Cement of high specific heat and high thermal

conductivity, obtained by using silane and silica fume as admixtures.

Cement and Concrete Research, 30, 1175-1178.

Xue, X., Yang, J., Zhang, W., Jiang, L., Qu, J., Xu, L., Zhang, H., Song, J., Zhang,

R., Li, Y., Qin, J., & Zhang, Z. (2015). The study of an energy efficient

cool white roof coating based on styrene acrylate copolymer and cement

for waterproofing purpose-Part I: optical properties, estimated cooling

effect and relevant properties after dirt and accelerated exposures.

Construction and Building Materials, 98, 176-184.

Yamazaki, S., Tanaka, S., & Tsukamoto, H. (1999). Kinetic studies of oxidation

of ethylene over a TiO2 photocatalyst. Journal of Photochemistry and

Photobiology A: Chemistry, 121(1), 55-61.

Young, H. D., Freedman, R. A., & Sears, F. W. (2008). Sears and Zemansky's

university physics: with modern physics (12th edition). San Francisco.

Pearson Addison-Wesley.

Yu, Q. L., Spiesz, P., & Brouwers, H. J. H. (2013). Development of cement-based

lightweight composites - Part 1: Mix design methodology and hardened

properties. Cement and Concrete Composites, 44, 17-29.

Yu, Q. L., Spiesz, P., & Brouwers, H. J. H. (2015). Ultra-lightweight concrete:

conceptual design and performance evaluation. Cement and Concrete

Composites, 61, 18-28.

182
Yun, T. S., Jeong, Y. J., Han, T.-S., & Youm, K.-S. (2013). Evaluation of thermal

conductivity for thermally insulated concretes. Energy and Buildings, 61,

125-132.

Zingre, K. T., Wan, M. P., Tong, S., Li, H., Chang, V. W. C., Wong, S. K., Thian

Toh, W. B., & Leng Lee, I. Y. (2015). Modeling of cool roof heat transfer

in tropical climate. Renewable Energy, 75, 210-223.

Zoldners., N. (1971). Thermal properties of concrete under sustained elevated

temperatures. ACI Publication, SP-25, 1-31.

183
 Appendix

Table A1 Solar reflectance and color of mortar specimens with various coatings at different
time points

After 300 hrs exposed to

Before BC deposition After BC deposition
simulated solar irradiation
Testing
TSC WSC PWSC TSC WSC PWSC TSC WSC PWSC
points
1 0.47 0.76 0.80 0.18 0.40 0.42 0.26 0.40 0.81
2 0.45 0.79 0.81 0.24 0.35 0.35 0.26 0.36 0.76
3 0.41 0.71 0.79 0.17 0.32 0.42 0.20 0.32 0.81
4 0.47 0.76 0.78 0.20 0.40 0.34 0.21 0.40 0.77
5 0.48 0.79 0.80 0.19 0.30 0.41 0.20 0.31 0.76
Solar 6 0.45 0.77 0.78 0.23 0.37 0.39 0.25 0.38 0.77
reflectance 7 0.50 0.75 0.79 0.26 0.41 0.38 0.29 0.41 0.80
8 0.49 0.79 0.80 0.25 0.30 0.39 0.27 0.31 0.79
9 0.49 0.78 0.77 0.22 0.43 0.39 0.26 0.44 0.75
10 0.50 0.78 0.78 0.23 0.45 0.40 0.26 0.45 0.80
11 0.51 0.79 0.80 0.22 0.38 0.35 0.25 0.39 0.76
12 0.49 0.77 0.79 0.19 0.40 0.36 0.22 0.40 0.77
73.8 95.9 95.7 44.8 70.5 74.1 53.0 71.9 95.7
1
73.8 95.9 95.7 44.8 70.5 74.1 53.0 71.9 95.7
73.6 96.2 96.0 54.3 66.7 64.7 56.5 66.8 93.2
2
73.5 96.2 96.0 54.3 66.7 64.7 56.5 66.8 93.2
69.7 92.9 95.4 44.2 68.0 69.0 48.7 70.4 95.4
3
69.7 92.9 95.4 44.2 68.0 69.0 48.7 70.3 95.4
75.0 95.9 95.3 52.0 73.2 67.0 54.6 72.8 94.3
4
75.0 95.9 95.3 52.0 73.1 67.0 54.6 72.8 94.3
75.8 96.3 96.1 47.7 65.0 70.9 50.4 66.2 93.1
5
Color L* 75.8 96.3 96.0 47.7 65.0 70.9 50.4 66.2 93.1
73.7 94.7 94.7 51.0 69.8 71.5 51.2 70.8 93.1
6
73.7 94.7 94.7 51.0 69.8 71.5 51.2 70.8 93.1
77.3 94.1 95.1 55.8 72.5 67.8 58.4 72.9 94.9
7
77.3 94.1 95.1 55.8 72.5 67.8 58.4 72.9 95.0
75.7 96.3 96.0 55.9 62.8 68.1 59.2 63.3 95.0
8
75.7 96.3 96.0 55.9 62.8 68.1 59.2 63.3 95.0
76.4 94.9 95.4 56.7 72.9 69.5 60.2 73.6 93.2
9
76.4 94.9 95.4 56.7 72.8 69.5 60.2 73.6 93.1
10 76.6 96.1 93.6 48.2 76.8 72.9 53.8 77.2 94.8

184
 76.6 96.1 93.6 48.2 76.8 72.9 53.8 77.2 94.8
76.9 96.0 95.9 53.0 71.0 65.5 60.1 71.7 91.5
11
76.9 96.0 95.8 53.0 71.0 65.5 60.1 71.7 91.5
76.7 95.3 95.8 47.6 71.3 67.7 51.5 72.5 93.7
12
76.7 95.3 95.8 47.6 71.3 67.7 51.5 72.5 93.7

185
Table A2 Surface temperatures of mortar specimens with various coatings at different time
points

After 300 hrs exposed to

Before BC deposition After BC deposition
simulated solar irradiation
Testing
TSC WSC PWSC TSC WSC PWSC TSC WSC PWSC
points
1 37.2 31.6 31.8 39.8 37.6 37.8 39.2 37.1 31.9
2 35.7 31.4 31.3 38.9 36.6 37.4 37.9 36.2 31.6
Inner 3 37.0 32.3 32.6 39.4 37.5 38.2 38.9 37.2 32.6
surface
temperature 4 37.0 31.3 31.5 39.6 37.0 37.7 38.7 37.0 31.5
after 9 hrs 5 36.0 31.3 31.2 38.9 36.9 37.1 38.2 36.4 31.3
exposed to 6 36.9 32.2 32.3 39.8 37.8 38.4 39.2 37.2 32.2
sun
simulator 7 37.1 31.5 31.6 39.1 37.1 37.7 39.0 37.3 31.8
8 36.4 32.5 31.5 38.8 37.2 37.3 38.7 36.6 31.6
9 37.5 31.5 32.7 39.8 38.0 38.5 39.3 37.5 32.5
1 41.0 34.5 34.3 44.0 41.4 42.0 43.0 40.5 34.4
2 39.5 34.2 33.9 43.3 40.4 41.4 41.4 39.5 34.0
Outer 3 40.0 34.6 35.0 43.3 40.5 41.4 42.2 40.3 34.8
surface
temperature 4 40.6 33.7 34.0 43.8 41.3 41.7 42.9 40.6 33.9
after 9 hrs 5 39.4 33.9 33.6 43.1 40.8 41.0 42.0 40.0 33.7
exposed to 6 40.0 34.3 34.4 43.6 40.9 41.7 42.3 40.0 34.4
sun
simulator 7 40.7 33.9 34.2 43.7 41.4 41.1 42.8 40.8 34.3
8 39.7 33.6 34.0 43.2 40.7 40.8 42.2 40.2 33.8
9 40.3 34.7 34.7 43.4 40.7 41.3 42.4 39.8 34.9

186
 Table A3 Solar reflectance and surface temperatures of concrete, ULCC, and
ULCC with photocatalytic coating

Before BC deposition
Testing
Concrete ULCC ULCC+C
points
1 0.43 0.38 0.81
2 0.40 0.43 0.80
3 0.40 0.43 0.78
4 0.41 0.41 0.78
5 0.38 0.43 0.75
6 0.39 0.43 0.76
Solar reflectance
7 0.41 0.41 0.78
8 0.40 0.40 0.79
9 0.41 0.42 0.77
10 0.40 0.41 0.81
11 0.40 0.41 0.79
12 0.43 0.42 0.80
1 37.4 33.9 29.5
2 36.8 33.2 29.0
3 37.7 34.6 30.5
Inner surface 4 37.3 33.9 28.9
temperature after 9
5 36.9 33.0 28.4
hrs exposed to sun
simulator 6 38.1 34.2 30.2
7 37.8 33.7 29.0
8 37.1 32.5 29.0
9 38.0 34.5 30.6
1 40.4 44.0 35.9
2 40.1 42.1 34.9
3 40.6 43.5 36.4
Outer surface 4 40.6 44.3 35.0
temperature after 9
5 40.1 41.8 34.1
hrs exposed to sun
simulator 6 41.0 43.3 35.7
7 40.7 43.9 35.2
8 40.4 42.7 35.2
9 40.9 42.9 36.2

187
 Publications from the research

Journals:

[1] Y. Wu, J.-Y. Wang, P.J.M. Monteiro, M.- . Zhang, “Development of ultra-

lightweight cement composites with low thermal conductivity and high specific

strength for energy efficient buildings”, Construction and Building Materials, 87

(2015) 100-112.

[2] Y. Wu, P. Krishnan, Liya E Yu, M.-H. Zhang, “Using lightweight cement

composite and photocatalytic coating to reduce cooling energy consumption of

buildings”, Construction and Building Materials, 145 (2017) 555-564.

[3] V. Rheinheimer, Y. Wu, T. Wu, K. Celik, J.-Y. Wang, L.D. Lorenzis, P.

Wriggers, M.-H. Zhang, P. J.M. Monteiro, “Multi-scale study of high-strength

low-Thermal conductivity cement composites containing cenospheres”, Cement

and Concrete Composites, 80 (2017) 91-103.

[4] Y. Wu, P. Krishnan, M.-H. Zhang, Liya E Yu, “Using photocatalytic coating

to maintain solar reflectance and lower cooling energy consumption of buildings”,

Energy and Buildings, (Accepted in Jan 2018).

Conferences:

[1] Y. Wu, J.-Y. Wang, P. J. M. Monteiro, M.-H. Zhang, “Ultra-lightweight

cement composites with low thermal conductivity for energy efficient

buildings”, International Conference on Civil and Environmental Engineering

- Cappadocia, 20-23, May, 2015, Nevsehir, Turkey.

188
[2] Y. Wu, “E perimental study on thermal conductivity of ultra-lightweight

cement composites”, 40th Conference on Our World in Concrete & Structures,

26-28, Aug. 2015, Singapore.

[3] Y. Wu, “Increasing building energy efficiency with lightweight cement

composites (poster presentation)”, Gordon Research Conference:Advanced

Materials for Sustainable Infrastructure Development, 01-05, Aug. 2016, Hong

Kong, China.

[4] Y. Wu, M.-H. Zhang, “Effect of low thermal conductivity of ultra-lightweight

cement composite on heat transfer through concrete panels”, International High-

Performance Built Environments Conference (iHBE), 17-18, Nov. 2016, Sydney,

Australia.

189
 Curriculum vitae

Wu Yunpeng

wuyunpeng@u.nus.edu

wuyunpeng0908@gmail.com

Ph.D. Aug. 2013- Present

Specialization: Civil Engineering (Construction Materials)

Department of Civil & Environmental Engineering

National University of Singapore, Singapore

Bachelor of Engineering Sep. 2009- Jul. 2013

Specialization: Road & Bridge Engineering (Construction Materials)

Department of Transportation Science & Engineering

Harbin Institute of Technology, China

190