close

Вход

Забыли?

вход по аккаунту

?

Effects of microwave heating on hardening and properties of cement-based materials

код для вставкиСкачать
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type o f computer printer.
The quality o f this reproduction is dependent upon the quality o f the
copy subm itted.
Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted.
Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back o f the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zed) Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
NORTHWESTERN UNIVERSITY
Effects of Microwave Heating
on Hardening and Properties
o f Cement-Based Materials
A DISSERTATION
SUBMITTED TO THE GRADUATE SCHOOL
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
for the degree
DOCTOR OF PHILOSOPHY
Field of Materials Science and Engineering
By
Donggy Sohn
Evanston, Illinois
June 1998
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 9832693
Copyright 1998 by
Sohn, Donggy
All rights reserved.
UMI Microform 9832693
Copyright 1998, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
© Copyright by Donggy Sohn 1998
All Rights Reserved
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
Effects o f Microwave Heating on Hardening and Properties
o f Cement-Based Materials
Donggy Sohn
The use o f the instrumented penetration test as a means o f monitoring the
hardening process o f cement-based materials was evaluated. The possibility of an effect
o f microwave energy on cement systems with respect to chemical and mechanical
property changes was explored.
The instrumented penetration test makes it possible to investigate materials in-situ
and continuously.
It provides detailed information to establish the scientific and
engineering basis for the hardening process of cement-based materials. It has the ability
to monitor rheological behavior of cement-based materials and the ability to help
determine elevated temperature curing time for cement mortar without strength
degradation.
D-optimal experimental design and multiple correlation analysis were used to
describe complicated effects o f numerous variables, resulting in establishing a cementhardening model. The developed models indicate that the hardening processes of cementbased systems are mainly based on the cement hydration. The most significant factors in
the hardening rate model as well as the initial setting time are temperature, water-to-total
reactive solid ratio and type o f reactive admixtures, while the effect o f sand content is
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
small.
While the degree o f hydration is the principal reason for hardening, the
temperature and water-to-cement ratio also influence the relationships between the degree
o f hydration and the hardness.
The nonlinear dependence o f hardening rate on
temperature found in almost all models implies that the hardening process is very
sensitive and complex with respect to temperature.
Microwave heating alters the chemical and mechanical properties o f cement-based
materials by accelerating cement hydration. Above 60°C, hardened cement paste contains
more porosity, resulting in a weak structure. A properly adjusted curing time, established
through the use of the instrumented penetration test, enables the elimination of
degradation o f 28-day strength at elevated temperatures at or below 60°C.
However,
80°C cannot be accepted as curing temperature. With a proper heating time, the final
setting time at 60°C can be reduced to 35% o f the room temperature value without
degrading strength. Thus, microwave energy will be a potential process in the area of
cement-related manufacturing.
Approved:
Prof. D. Lynn Johnson
Department of Field o f Materials Science and Engineering
The Robert R. McCormick School o f Engineering and Applied Science
Northwestern University
June 1998
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGEMENTS
I would like to express my deepest gratitude to my advisor, Professor D. Lynn
Johnson for unyielding perseverance and guidance throughout my endeavors.
The
countless hours o f discussions have provided the uncommon opportunity to explore and
enjoy the essence o f scientific research. I am greatly indebted to Professor Thomas O.
Mason for the collaborative opportunities he has provided, his helpful observations and
interest in the work. I would like to thank Professor Hamlin M. Jennings and Surendra P.
Shah for making the time and efforts to sit on my examination committee. I would like to
thank Professor Leslie J. Struble and Mr. G. Sun for cooperation in rheology work.
I would like to thank the many members o f Professor Johnson’s, Professor
Mason’s and Professor Jennings’ research groups for their assistance and friendship over
the last few years; Dr. J. Hwang, Dr. S. Song, Dr. S. Ford, Dr. M. Teng, Dr. H. Su, Dr. F.
Kaatz, Dr. R. Olson, John Shane, and Matthew Henrichsen.
I also gratefully thank my father, Seoubjoon Sohn; my mother, Okhwa Park; my
sisters, Seongkyeong, Mi, and Jina. I like to dedicate this thesis to my wife, Nieun and to
my son, Wonil William.
This work was by supported by the National Science Foundation Center for
Science and Technology o f Advanced Cement-Based Materials under grant no. CHE-9120002.
Software provided by Harold S. Haller & Co. (Cleveland, OH) is gratefully
acknowledged.
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
ABSTRACT ...................................................................................................................
iii
ACKNOWLEDGMENTS ...........................................................................................
v
TABLE OF CONTENTS .............................................................................................
vi
LIST OF FIGURES ......................................................................................................
x
LIST OF TABLES ........................................................................................................
xix
CHAPTER 1: Introduction .........................................................................................
1
CHAPTER 2: Background and Literature Review ...................................................
6
2.1. Cement Hydration Mechanism .....................................................................
6
2.1.1. Cement Compositions and Types .................................................
6
2.1.2. Hydration o f Cement .......................................................................
7
2.2. Hardness Development ..................................................................................
11
2.1.1. Setting Phenomena ........................................................................
11
2.2.2. Estimation Methods for Setting Times .........................................
13
2.3. Elevated Temperature Curing .......................................................................
17
2.3.1. Effects o f Elevated Temperature on Cement Hydration .............
18
2.3.2. Applications of Microwave Energy for Curing o f Cement ........
22
2.4. Rheology o f Cement Pastes ...........................................................................
27
2.4.1. Models o f Flow Behavior ..............................................................
28
2.4.2. Viscoelastic Properties ...................................................................
32
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4.3. Rheologicai Behavior o f Cement Pastes .......................................
34
2.5. Pozzolanic Materials ....................................................................................
38
2.5.1. Classification o f Pozzolans ............................................................
38
2.5.2. Effects o f Pozzolans on Hydration Reactions and Rheology .....
41
2.5.3. Strength Changes by Adding Pozzolans .......................................
44
2.5.4. Influence o f Pozzolans on the Concrete Durability ....................
46
CHAPTER 3: Experimental Procedure .....................................................................
52
3.1. Apparatus .........................................................................................................
52
3.1.1. Microwave Oven .............................................................................
52
3.1.2. Instrumented Penetration Test ......................................................
52
3.2. Sample Preparation .........................................................................................
54
3.2.1. Materials ..........................................................................................
54
3.2.2. Experimental Designs .....................................................................
55
3.2.3. Mixing and Curing ..........................................................................
59
3.2.4. Sample Curing for Strength Test ..................................................
60
3.3. Characterization Methods ..............................................................................
62
3.3.1. Vicat Needle Test ............................................................................
62
3.3.2. Rheometer ........................................................................................
63
3.3.3. Loss on Ignition ..............................................................................
63
3.3.4. Impedance Spectroscopy ................................................................
65
3.3.5. Pore Solution Analysis ...................................................................
66
CHAPTER 4: Hardness Development o f Cement-Based Materials ........................
69
4.1. Microwave Heating Effects on Cement Paste ...........................................
69
4.1.1. Temperature Changes ....................................................................
69
4.1.2. Hardness Development of Cement Paste ......................................
75
4.1.3. Setting Times o f Cement Paste ......................................................
78
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.4. Activation Energy .........................................................................
82
4.2. Hardening Rate o f Cement Paste ...................................................................
90
4.2.1. Influence of the Penetration Rate on the Penetration Hardness ..
90
4.2.2. Hardening Rate Equation o f Cement Paste ...................................
92
4.2.3. Relationships between Hardness and Degree o f Hydration .......
119
4.3. Sand Effects on Hardening Behavior ............................................................ 126
4.3.1. Influence of Sand on Setting Time ...............................................
126
4.3.2. Influence of Sand on the Hardening Rate ..................................... 128
4.4. Influence o f Reactive Admixtures on Hardening Behavior ........................ 131
4.4.1. Effects o f Reactive Admixtures on the Initial SettingTime ........ 134
4.4.2. Effects o f Reactive Admixtures on the Hardening Rate .............
137
4.5. Rheologicai Properties o f Cement Paste ....................................................... 153
4.5.1. Rheology Studies by Rheometer and Penetration Test ...............
153
4.5.2. Influence of Mixing on Rheology .................................................. 160
4.5.3. Penetration Rate Effects on Cement Rheology ...........................
164
CHAPTER 5: Influence o f Microwave Curing on Hydration and Properties ........ 171
5.1. Curing Temperature Effects on Hydration o f Cement Paste ......................
171
5.1.1. Degree o f Hydration ........................................................................ 171
5.1.2. Microstructure Changed Monitored by Impedance Spectroscopy
174
5.1.3. Pore Solution Analysis .................................................................... 179
5.2. Temperature Effects on Compressive Strength o f Cement Mortar ...........
193
5.2.1. Curing Conditions ............................................................................ 193
5.2.2. Two-Hour Microwave Curing Effects on Strength ...................... 193
5.2.3. Determination of Curing Time for Enhancing Strength .............
196
5.2.4. Final Setting Time o f Cement Mortar ........................................... 200
5.2.5. Strength Test for Blended Mortars ................................................ 202
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 6: Conclusions and Future Work ............................................................. 207
6.1. Conclusions ...................................................................................................... 207
6.2. Suggestions for Future Work ......................................................................... 209
REFERENCES .............................................................................................................. 211
APPENDIX A: General Theory o f Microwaves .......................................................
222
APPENDIX B: Statistical Methodology ..................................................................... 231
APPENDIX C: Impedance Spectroscopy Study on Cement-Based Materials ......
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
238
LIST OF FIGURES
Figure 1-1 Overview o f the research objectives........................................................
4
Figure 2-1 Typical heat o f hydration curve forcem eat paste...................................
8
Figure 2-2 Vicat Apparatus..........................................................................................
14
Figure 2-3 Typical plot o f penetration resistance.......................................................
16
Figure 2-4 Comparisons in compressive strengths between
microwave and normal curing from several references..................................
25
Figure 2-5 Plausible microwave applications.............................................................
26
Figure 2-6 Flow curves o f several flow models.........................................................
29
Figure 2-7 Hysteresis loop for material under shear stress.......................................
31
Figure 2-8 Type I hysteresis flow curve o f cement paste..........................................
36
Figure 2-9 Phase diagram for ternary systems o f CaO-SiiO-AliOs.........................
39
Figure 2-10 Relative strength developments o f concretes influenced
by replacing pozzolans.......................................................................................
47
Figure 3-1 Schematic diagram o f the instrumented penetration test
unit and microwave oven...................................................................................
53
Figure 3-2 (a) Schematic diagram o f the microwave oven for
curing mortar for strength test (b) Dimensions of the mortar
specimen for strength tests.................................................................................
61
Figure 3-3 Schematic diagram of modifying microwave oven for
impedance spectroscopy measurements............................................................
67
Figure 4-1 Temperature profile of cement pastes cured at room
temperature under adiabatic conditions with various w/c ratios....................
70
x
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4-2 Temperature profile o f cement pastes at various
temperatures by using microwave energy with insulation..............................
72
Figure 4-3 Temperature profile o f cement pastes at various
temperatures by using microwave energy without insulation.........................
74
Figure 4-4 Penetration hardness versus distance for wet silica powder....................
76
Figure 4-5 Penetration hardness development for cement pastes
with w/c = 0.3 cured at various temperatures by microwaves........................
77
Figure 4-6 Setting time change for cement pastes measured by Vicat
and the instrumented penetration test at room temperature............................
79
Figure 4-7 Initial setting time response surface and contour plot for
cement pastes measured by the instrumented penetration test........................
81
Figure 4-8 Initial setting time for cement pastes, representing
the effect o f the interaction between w/c ratio and temperature.....................
83
Figure 4-9 Semi-logarithm plot of penetration hardness...........................................
85
Figure 4-10 Arrhenius plot o f the rate o f hardening at 5 MPa of
the penetration hardness......................................................................................
86
Figure 4-11 Apparent activation energy obtained by the penetration
hardness and degree o f hydration for cement pastes.......................................
89
Figure 4-12 Second order polynomial curve fitting from room
temperature to 60°C at 5 MPa of the penetration hardness.............................
91
Figure 4-13 Effects o f penetration rate on the penetration hardness.........................
93
Figure 4-14 Hardening rate changes caused by the penetration rate.........................
94
Figure 4-15 Penetration hardness developments for cement pastes.........................
95
Figure 4-16 Ln(Hardening rate) at w/c = 0.3 and 6.4 mm/hr o f the
penetration rate for La Farge Type I cement pastes.........................................
98
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4-17 Ln(Hardening rate) at w/c = 0.3 and 6.4 mm/hr o f the
penetration rate for Holnam Type I cement pastes..........................................
99
Figure 4-18 Ln(Hardening rate) response surface and contour plot
with respect to temperature and hardness level for cement pastes,
fixing w/c ratio at 0.3 and the penetration rate at 64 mm/hr..........................
100
Figure 4-19 Ln(Hardening rate) response surface and contour plot
with respect to temperature and hardness level for cement pastes,
fixing w/c ratio at 0.4 and the penetration rate at 6.4 mm/hr.......................... 102
Figure 4-20 Ln(Hardening rate) for cement pastes, representing the
effect o f the interaction between hardness level and w/c ratio....................... 103
Figure 4-21 Ln(Hardening rate) for cement pastes, representing the
effect o f the interaction between hardness level and gear ratio...................... 104
Figure 4-22 Ln(Hardening rate) response surface and contour plot
with respect to w/c ratio and hardness level for cement pastes,
fixing temperature at 30°C and the penetration rate at 6.4 mm/hr................. 106
Figure 4-23 Ln(Hardening rate) response surface and contour plot
with respect to w/c ratio and hardness level for cement pastes,
fixing temperature at 60°C and the penetration rate at 6.4 mm/hr................. 107
Figure 4-24 Ln(Hardening rate) response surface and contour plot
with respect to penetration rate and hardness level for cement
pastes, fixing temperature at 30°C and w/c ratio at 0.4................................... 108
Figure 4-25 Ln(Hardening rate) response surface and contour plot
with respect to penetration rate and hardness level for cement
pastes, fixing temperature at 60°C and w/c ratio at 0.4..................................
109
Figure 4-26 Ln(Hardening rate) for cement pastes, representing the
effect of the interaction between w/c ratio and temperature........................... 110
Figure 4-27 Ln(Hardening rate) response surface and contour plot
with respect to temperature and w/c ratio for cement pastes,
fixing hardness level at 1 MPa and penetration rate at 6.4 mm/hr................
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ill
Figure 4-28 Ln(Hardening rate) response surface and contour plot
with respect to temperature and w/c ratio for cement pastes,
fixing hardness level at 13 MPa and penetration rate at 6.4 mm/hr............... 112
Figure 4-29 Ln(Hardening rate) for cement pastes, representing the
effect o f the interaction between temperature and penetration rate...............
113
Figure 4-30 Ln(Hardening rate) response surface and contour plot
with respect to temperature and penetration rate for cement
pastes, fixing hardness level at 1 MPa and w/c ratio at 0.3............................
114
Figure 4-31 Ln(Hardening rate) response surface and contour plot
with respect to temperature and penetration rate for cement
pastes, fixing hardness level at 13 MPa and w/c ratio at 0.3..........................
115
Figure 4-32 Ln(Hardening rate) response surface and contour plot
with respect to w/c ratio and penetration rate for cement pastes,
fixing hardness level at 13 MPa and temperature at 30°C............................... 117
Figure 4-33 Ln(Hardening rate) response surface and contour plot
with respect to w/c ratio and penetration rate for cement pastes,
fixing hardness level at 13 MPa and temperature at 60°C..............................
118
Figure 4-34 Correlation between degree o f hydration and penetration
hardness for cement pastes cured at 20°C......................................................... 121
Figure 4-35 Correlation between degree o f hydration and penetration
hardness for cement pastes with fixed w/c ratio at 0.3.................................... 122
Figure 4-36 Degree o f hydration response surface and contour plot
with respect to penetration hardness and w/c ratio
for cement pastes cured at 20°C........................................................................
123
Figure 4-37 Degree o f hydration response surface and contour plot
with respect to penetration hardness and temperature for cement
pastes with fixed w/c ratio at 0.3.......................................................................
125
Figure 4-38 Initial setting time response surface and contour plot for
cement mortars measured by the instrumented penetration test..................... 127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4-39 Ln(Hardening rate) response surface and contour plot
for cement mortars with respect to temperature and hardness level.............. 130
Figure 4-40 Ln(Hardening rate) response surface and contour plot
for cement mortars with respect to s/c ratio and hardness level,
fixing penetration rate at 6.4 mm/hr and at 30°C............................................
132
Figure 4-41 Ln(Hardening rate) response surface and contour plot
for cement mortars with respect to s/c ratio and hardness level,
fixing penetration rate at 64 mm/hr and at 30°C.............................................
133
Figure 4-42 Initial setting time for cement pastes containing reactive
admixtures, representing the effects of the interaction between
admixture amount and temperature (a) GGBFS (b) silica fume...................
136
Figure 4-43 Initial setting time response surface and contour plot
for silica fume containing cement pastes (w/r = 0.4)......................................
138
Figure 4-44 Initial setting time response surface and contour plot
for GGBFS containing cement pastes (w/r = 0.4)............................................ 139
Figure 4-45 Initial setting time response surface and contour plot
for Class F fly ash containing cement pastes (w/r = 0.4)................................ 140
Figure 4-46 Ln(Hardening rate) for cement pastes containing silica fume,
representing the effect of the interaction between silica fume
amount and w/r ratio........................................................................................... 143
Figure 4-47 Ln(Hardening rate) response surface and contour plot
with respect to silica fume amount and hardness level for cement
pastes containing silica fume with w/r = 0.3.................................................... 144
Figure 4-48 Ln(Hardening rate) response surface and contour plot
with respect to silica fume amount and hardness level for cement
paste containing silica fume with w/r = 0.4...................................................... 145
Figure 4-49 Ln(Hardening rate) for cement pastes containing GGBFS,
representing the effect of the interaction between GGBFS amount
and w/r ratio......................................................................................................... 146
xiv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4-50 Ln(Hardening rate) response surface and contour plot
with respect to GGBFS amount and temperature for cement
pastes containing GGBFS with w/r = 0.3......................................................... 147
Figure 4-51 Ln(Hardening rate) response surface and contour plot
with respect to GGBFS amount and temperature for cement
pastes containing GGBFS with w/r = 0.4........................................................
148
Figure 4-52 Ln(Hardening rate) for cement pastes containing
Class F fly ash, representing the effect of the interaction
between fly ash amount and temperature.......................................................... 150
Figure 4-53 Ln(Hardening rate) response surface and contour plot
with respect to fly ash amount and temperature for cement
pastes containing Class F fly ash with w/r = 0.3.............................................. 151
Figure 4-54 Ln(Hardening rate) response surface and contour plot
with respect to fly ash amount and temperature for cement
paste containing Class F fly ash with w/r = 0.4...............................................
152
Figure 4-55 Comparisons between elastic storage modulus (G’)
and penetration hardness for cement pastes with w/c = 0.3............................ 154
Figure 4-56 Comparisons between elastic storage modulus (G’)
and penetration hardness for cement pastes with w/c = 0.4............................ 155
Figure 4-57 Comparisons between viscous loss modulus (G”)
and penetration hardness for cement pastes with w/c = 0.3............................ 156
Figure 4-58 Comparisons between viscous loss modulus (G”)
and penetration hardness for cement pastes with w/c = 0.4............................ 157
Figure 4-59 Correlation between the elastic storage modulus (G’)
and the penetration hardness for cement pastes............................................... 159
Figure 4-60 Effects o f mixing method and time on the penetration
hardness developments o f cement pastes.........................................................
161
Figure 4-61 Relationships between the times to reach hardness levels
of 1 or 13 MPa and the degree o f hydration o f cement pastes....................... 162
xv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4-62 Effect o f mixing on the later stage o f cement hydration....................... 163
Figure 4-63 Effects o f changing penetration rate on the penetration
hardness for cement paste o f w/c = 0.3 at 60°C...............................................
165
Figure 4-64 Influences o f changing penetration rates on hardening
curves for cement pastes o f w/c = 0.3................................................................ 167
Figure 4-65 Influences o f delayed onset penetration rates on hardening
Curves for cement pastes o f w/c = 0.3............................................................... 168
Figure 4-66 Force distributions though fluid- or solid-like media............................ 170
Figure 5-1 Degree o f hydration for cement pastes with w/c = 0.4
cured by microwaves..........................................................................................
172
Figure 5-2 Volume fraction o f capillary porosity for cement pastes
with w/c = 0.4 cured by microwaves................................................................. 173
Figure 5-3 Normalized conductivity in early age o f hydration for
cement pastes with w/c = 0.4 cured by microwaves........................................ 175
Figure 5-4 Full scale o f normalized conductivity versus degree of
hydration for cement pastes with w/c = 0.4 cured by microwaves................
176
Figure 5-5 Pore structure factor for cement pastes with w/c = 0.4
cured by microwaves..........................................................................................
178
Figure 5-6 Conductivity o f the pore solution for cement pastes with
w/c = 0.4 cured by microwaves.........................................................................
180
Figure 5-7 pH o f the pore solution for cement pastes with w/c = 0.4
cured by microwaves..........................................................................................
181
Figure 5-8 Alkali ion concentrations o f the pore solution for cement
pastes with w/c = 0.4 cured by microwaves.....................................................
182
Figure 5-9 Sulfate concentrations o f the pore solution for cement pastes
with w/c = 0.4 cured by microwaves................................................................. 183
xvi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5-10 Aluminate concentrations o f the pore solution for cement pastes
with w/c = 0.4 cured by microwaves................................................................
184
Figure 5-11 X-ray diffraction patterns for cement pastes at 7 age days
with w/c = 0.4 cured by microwaves................................................................
186
Figure 5-12 Calcium concentrations o f the pore solution for cement pastes
with w/c = 0.4 cured by microwaves................................................................
187
Figure 5-13 Silicate concentrations o f the pore solution for cement pastes
with w/c = 0.4 cured by microwaves................................................................
188
Figure 5-14 Concentrations o f SiC>2 versus CaO o f the pore solutions
for cement pastes with w/c = 0.4 cured by microwaves.................................
190
Figure 5-15 CaO/Si02 ratio of the pore solution versus degree o f hydration
for cement pastes with w/c = 0.4 cured by microwaves.................................
192
Figure 5-16 Temperature history with respect to thermocouple
positions for cement mortars cured by microwaves........................................ 194
Figure 5-17 28-day compressive strength for room temperature cured
and microwave cured cement and 50% GGBFS mortars...............................
195
Figure 5-18 Temperature history for cement mortars during heating
at various temperature and time by microwaves.............................................. 197
Figure 5-19 Temperature history for cement mortars after heating.......................... 198
Figure 5-20 28-day compressive strength for room temperature cured
and microwave cured cement mortars. Heating times were
determined by the instrumented penetration test.............................................
199
Figure 5-21 Final setting times for room temperature cured and
microwave cured cement mortars...................................................................... 201
Figure 5-22 28-day compressive strength for room temperature cured
and microwave cured mortars containing reactive admixtures.
Heating times were determined by the instrumented penetration te s t
xvii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
203
Figure 5-23 Procedures for microwave curing process on cement
mortar manufacturing.......................................................................................... 206
Figure A -l Electromagnetic spectrum......................................................................... 223
Figure A-2 Interaction o f microwaves with materials................................................ 224
Figure A-3 Absorption o f microwaves to half-power depth, D ................................ 227
Figure A-4 Penetration depth as a function of temperature for pure water.............
228
Figure A-5 Heating Patterns in conventional and microwave furnaces................... 229
Figure C -l Typical Nyquist plot o f cement paste at 28 days..................................... 240
xviii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES
Table 2-1 Cement notation and weight percent in Portland cement.........................
7
Table 2.2 Chemical compositions in pozzolanic materials........................................
42
Table 2.3 Classification o f cement according to ASTM C595.................................
42
Table 2-4 Classification o f cement according to European Standard.....................
42
Table 3-1 Chemical compositions o f cement-based materials..................................
55
Table 3-2 Variables and their levels for studying initial setting time
and hardening rate...............................................................................................
56
Table 3-3 D-optimai design for cement paste.............................................................
57
Table 3-4 D-optimal design for cement mortar...........................................................
57
Table 3-5 D-optimal design for cement paste containing GGBFS...........................
58
Table 3-6 D-optimal design for cement paste containing silica
fume or fly ash....................................................................................................
58
Table 4-1 Apparent activation energy o f the hardening rate for
cement pastes with various w/c ratios.............................................................
87
Table 4-2 Ln(Hardening rate) response surface coefficients and their
t-values for cement pastes.................................................................................
96
Table 4-3 Lists o f coded variables for cement pastes................................................
96
Table 4-4 Effects o f increasing the major variables on
Ln(Hardening rate) o f cement pastes................................................................. 116
Table 4-5 Effects o f the interactions on Ln(Hardening rate) of cement pastes
116
Table 4-6 Experimental designs for investigating the relationships
between degree o f hydration and hardness level.............................................
119
xix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4-7 Ln(Hardening rate) response surface coefficients and their
t-values for cement mortars................................................................................
129
Table 4-8 Lists o f coded variables for cement mortars...............................................
129
Table 4-9 Initial setting time response surface coefficient for
pastes containing reactive admixtures...............................................................
134
Table 4-10 Effects o f adding reactive admixtures on initial setting tim e................
135
Table 4-11 Ln(Hardening rate) response surface coefficients and
their t-values for pastes containing reactive admixtures.................................. 141
Table 4-12 Lists o f coded variables for pastes containing reactive
Admixtures...........................................................................................................
141
Table 4-13 Effects o f adding reactive admixtures on Ln(Hardening rate)
o f pastes containing reactive admixtures........................................................... 149
Table 5-1 Microwave heating time o f cement mortars to reach given
levels o f penetration hardness at several curing temperature........................... 196
Table 5-2 Microwave heating time of mortars containing reactive
admixtures to reach given levels o f penetration hardness
at several curing temperature.............................................................................. 202
Table B -l Advantages and disadvantages for experimental designs........................
xx
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
232
CHAPTER 1
Introduction
Although cement-based materials have become the most widely used construction
material throughout the world, aspects of the hydration process are still unsolved,
compared with other structural materials.
Since the hydration process governs the
physical properties o f hardened cement, increasing our knowledge o f the hydration
reactions will lead to a better understanding o f the important industrial characteristics,
such as strength, creep, and durability, of cement based materials, ranging from cement
paste to mortar to concrete.
Since cement paste becomes solidified during the hydration process causing by
reactions between cement particles and water, this physical change is closely related to
the cement chemistry. However, conventional methods to monitor the physical changes
of cement paste, for instance the Vicat needle test (ASTM C191) and the standard
penetration resistance test (ASTM C403), cannot prove the relationship because they
offer only limited information about the hardening phenomena.
A newly developed instrumented penetration test can describe the hydration
process o f cement paste in detail in terms of the resistance to penetration. This makes it
possible to make continuous, in-situ measurement on one specimen. This is a distinct
advantage over conventional methods because it can reduce labor and cost for testing and
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
minimize experimental error. This test can also be applied to a variable temperature
environment as well as the presence o f mineral admixtures (e.g., pozzolans).
The hydration o f cement can be accelerated by several methods. The simplest of
these methods, steam curing under moisture, involves an external heating source to
increase the hydration rate and impart strength to the cement pastes in less time.
However, it requires a relatively long time for curing since heat must diffuse inward from
the surface and the inherently nonuniform temperature can generate thermal cracking.
Since microwave energy generates heat through interaction with the water, the heat is
deposited in the bulk.
Thus, microwave-enhanced heating is a potentially attractive
method for accelerating cement hydration.
The use o f microwave energy for heating began in the 1950’s as an application of
the magnetron developed during World War H. In spite of many data on the dielectric
properties o f materials and significant development in the design of microwave
generators and applicators, the science o f microwave processing is still in its early stages
of development, especially in the processing o f cement-based system.
In the past,
industry was reluctant to accept the entirely new microwave processing technique. A lack
o f positive research and development made the process even less attractive to industry.
However, microwave processing has obvious advantages over conventional
processing such as steam curing through significant reductions in manufacturing cost and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
processing time, improved microstructures and properties, high quality o f product due to
uniform heating, and synthesis of new materials.
The central goal o f this work was to obtain a fundamental understanding o f the
hardening process and property changes o f cement-based materials under microwave
heating. The objectives were to exploit the instrumented penetration test for monitoring
physical changes during hydration, and to investigate how process variables influence the
hardening process, resulting in establishing quantitative relations between chemical and
physical changes of cement-based materials during early hydration.
Other objectives
were to investigate the effects of microwave heating on chemical and mechanical
properties, and to see if the instrumented penetration test could become an alternative
method for studying the rheology of cement-based materials.
Figure 1-1 shows the
overview o f the research objectives.
There are six chapters and three appendices in this thesis. Chapter 2 presents a
brief description o f the cement hydration mechanism, including hardness development
and temperature effects. Short reviews of rheology and characteristics o f cement-based
materials are also presented. Chapter 3 describes the experimental techniques used in this
study.
Chapter 4 discusses hardening rate and rheological behavior o f cementitious
materials as examined using the instrumented penetration test. Chapter 5 discusses the
effects o f microwave processing on chemical and mechanical properties o f cement-based
materials. Chapter 6 summarizes the major conclusions o f this research and discusses
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
Instrumented
Penetration Test
Hardening
Mechanism
Rheological
Behavior
Microwave
Heating
Mechanical
Property
Hydration
Mechanism
Setting
Tim e
Rheology
Test
A ctivation
Energy
M ixing
Effect
Pore
Structure
H ardening
Rate
Rate
Effect
Ion
Behavior
28-day
Strength
Figure 1-1 Overview of the research objectives.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Degree o f
Hydration
5
future research potential.
Appendix A describes the general theory of
microwaves. Appendix B gives a brief description o f statistical methodology used in this
research. Appendix C shows the recent progress o f applying impedance spectroscopy in
studying cement-based materials.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 2
Background and Literature Review
2.1. Cement Hydration Mechanism
2.1.1. Cement Compositions and Types (Mindess and Young, 1981)
Ordinary Portland cement is composed o f five basic constituents with various
compositions;
tricalcium
silicate
(C3S:
(3 C a 0 )S i0 2),
dicalcium
silicate
(C2S:
(2Ca0)-Si02), tricalcium aluminate (C3A: (3 Ca 0 )-Al20 3 ), tetracalcium aluminoferrite
(C4AF: (4 Ca0 )-Al20 3 'Fe20 3 ), and Gypsum (CSH2: C aS(V 2H2C)). C3S is responsible for
the majority o f physical properties of hardened cement.
C2S resembles C3S in its
development o f long-term strength but reacts more slowly. Although C3A is believed to
develop little strength itself, its extremely fast hydration rate affects significantly the final
cement properties. The hydration o f C4AF is very' similar to that of C3A, but is slower.
Gypsum is added to control violent early hydration o f C3A.
Depending on the percentage of these components, Portland cement can be
divided into five types o f cement, listed in Table 2.1. Among these five types, Type I
cement is the most commonly used. Type III contains a high concentration o f C3S which
accelerates the hydration and has the greatest fineness among the five types, causing early
strength. This characteristic would be beneficial when pouring cement in cold weather or
when rapid removal from molds is desired. Type IV cement has the lowest C3S content
and low fineness and is used when heat dissipation is a concern. Type V has a low C3A
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
content, but provides better resistance to sulfate attack. Type II cement is also sulfate
resistant, but gains strength at a higher rate than Type V.
Table 2-1 Cement notation and weight percent in Portland cement (ASTM C 595).
(3Ca0)*Si0 2
(2Ca0)-Si0 2
(3Ca0)-AJ20 3
(4Ca0)-Al20 3Fe 20 3
C aS0 4-2H20
Blaine Fineness (nr/kg)
Symbol
C3S
C2S
C3A
C4AF
csh2
I
50
25
12
8
5
350
n
in
45
30
7
60
15
12
10
8
5
350
5
450
IV
25
50
5
V
40
40
4
12
10
4
300
4
350
2.1.2. Hydration o f Cement (Mindess and Young, 1981; Taylor, 1990)
The setting and hardening of cement paste result from chemical and physical
processes that take place between cement and water. An understanding of the chemistry
o f hydration is necessary for a full appreciation o f the properties o f cement pastes. Each
phase has unique hydration kinetics, the order of reactivity being C3A>C3S>C4AF>C2S.
Since all o f these reactions occur simultaneously, one way to view cement hydration is to
treat each phase separately.
Although this treatment is not completely valid, it is
reasonable in most cases.
The actual chemical processes of cement hydration are usually categorized into
five distinct stages. Figure 2-1 shows these stages, superimposed on the heat of hydration
curve. As cement comes into contact with water, ions are immediately dissolved from
cement particles, generating heat. This is called stage I. These reactions are indicated as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
IV
m
C
O
+-*
3
O
>
W
<u
e
«4-H
o
<D
15 min.
2-4 hrs.
8-10 hrs.
30 hrs.
Time after M ixing
Figure 2-1 Typical heat o f hydration curve for cement paste divided into the five stages
o f hydration.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
follows:
C3S => C2S + C a 2+ + OH*
(Eq. 2-1)
C3A + 6H + 3CSH2 => 6Ca2+ +2A1(0H)4' + 3S042+ + 40H '
(Eq. 2-2)
The dissolved ions form ettringite according to the following reaction during stage I.
(Eq. 2-3)
6Ca2+ + 6H +2A1(0H)4*+ 3 S 042+ + 40H* => C6AS3H 32
Under normal conditions the dissolving reactions proceed very slowly during the
following 2-4 hours in stage II, known as the induction period. The inactivity of the
induction period is believed to be the result o f a nucleation and growth phenomena or the
formation o f a protective layer around the cement grains. At the end o f the induction
period, most o f the network structure is believed to be formed and cement paste begins to
stiffen; this stage is called initial set. An acceleration period (stage III) follows, during
which cement hydrates rapidly and reaches its maximum rate o f heat evolution. In this
period, the main hydration reactions begin to occur. The final set occurs in 4 to 8 hours.
Cement paste then begins to harden and hydration slows to a steady state (stage IV and
V) after 12 to 24 hours.
The hydration products o f cement are calcium silicate hydrate (C-S-H), calcium
hydroxide (CH), and several sulfates like ettringite or monosulfate.
Since calcium
silicates comprise about 75% by weight o f ordinary Portland cement, the principal
hydration products are C-S-H and CH. The C-S-H has a variable stoichiometry and is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
highly amorphous, thus the C-S-H designation. On the other hand CH is crystalline and
has a fixed composition. The C-S-H and CH are produced by thehydration o f tri or
dicalcium silicate, shown in Eq. 2-4 and 2-5.
Also, CH isreacted
to C-S-H by a
pozzolanic reaction when pozzolan is added; this will be discussed in section 2-5.
Tricalcium Silicate:
2 C3S + 6 H =>
C-S-H + 3CH
(Eq. 2-4)
Dicalcium Silicate:
2 C 2S + 4H =>
C-S-H + CH
(Eq. 2-5)
Tricalcium aluminate (C 3A) is the most reactive o f all the phases and gypsum
must be added to reduce flash set formation. Ettringite can be formed immediately as a
result o f C 3A and gypsum hydration, shown in Eq. 2-6.
2 C 3A + 26H + 3 CSH2 =>
C6AS 3H32 (Ettringite)
(Eq. 2-6)
However, this product is only stable with an adequate supply of sulfate. If gypsum is
consumed before all C3A has hydrated, the ettringite acts as a sulfate source to
decompose to monosulfate, shown in Eq. 2-7.
2 C3A + 4H + C6AS 3H32
=> 3 C4ASH 12 (Monosulfate)
(Eq. 2-7)
The other minor products o f C3A are calcium aluminate hydrates (C-A-H) and
hydrogamet. C4AF is hydrated similarly to that of C3A but is slower and produces less
heat. Since this phase is generally the minor phase, it will not be discussed. The reaction
o f C4AF is shown in Eq. 2-8.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
Tetracalcium Aluminoferrite: C4AF + 3 CSH 2 => C6(A,F)S3H32 + (A,F)H3
(Eq. 2-8)
2.2. Hardness Development
The solidification o f cement paste is physically manifested during three
consecutive process stages after mixing with water: induction, setting and hardening.
Although some hydration products like ettringite and CH begin to be formed immediately
after mixing, the paste still remains plastic and workable until reaching the initial setting
time, which represents the end o f the dormant period and the beginning o f the main
reactions o f the cement paste. Because the paste becomes unworkable after the dormant
period, it should be placed prior to the initial setting time.
After setting, the paste
continues to harden and gains strength with time, defined as the hardening process
(Soroka, 1979).
2.2.1. Setting Phenomena
The term setting has been used to describe the onset of rigidity in fresh cement
paste and concrete (Neville, 1996; Mindess and Young, 1981; Mehta and Monteiro,
1993). From the morphological point o f view, setting is controlled by the hydration of
cement grains, especially C 3S. The paste remains workable within the induction period
because ettringite and CH coatings, which form around the cement grains after the initial
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
contact with water, retard further hydration. These coatings are broken up at the end of
the dormant period by an osmotic pressure due to differing ion concentrations inside and
outside the coating (Powers, 1961). Since the rupture o f the gel coating exposes the
cement grains, hydration is resumed, and the setting process takes place (Soroka, 1979).
Thermodynamically, the initial set is marked by a rapid temperature rise of
cement pastes, which corresponds roughly to the beginning o f the acceleration stage.
This temperature rise will reach a maximum rate around the final set. However, the
precise relationships between the strength gain and the temperature rise by chemical
reactions are not proven yet. Although the paste acquires some strength during setting, it
is important to distinguish setting from hardening, which refers to the development of
useful and measurable strength. Therefore, setting is thought to be a transitional period
between states o f true fluidity and true rigidity (Mehta and Monteiro, 1993). Setting is
also related to a decrease in electrical conductivity and an increase in the velocity of
sound waves propagating through paste (Mindess and Young, 1981).
The phenomena o f the setting can be altered by some factors, such as water-tocement ratio (Bilodeau and Malhotra, 1992), curing temperature (Eren, et al., 1995)
pozzolan amounts (Gebler and Klieger, 1986) or admixtures.
Concretes, mortars or
cement pastes which have higher water contents tend to set more slowly because of
diluted ion concentrations by additional water. Higher curing temperature reduces the
setting time by accelerating cement hydration. Setting times are increased when cement
is replaced by slag or fly ash, but the effects o f these pozzolans varies significantly due to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
wide variations in chemical compositions. The chemical nature o f the admixture and the
amount used also change the setting nature. Accelerators, retarders, and superplasticizers
are examples o f commonly used admixtures. Accelerators such as calcium chloride have
the ability to reduce the initial and final setting times and accelerate the hardening of
concrete; retarders like sugar extend these setting times. Generally, for normal dosages of
superplasticizers some retardation occurs, but according to the type and the amount of the
superplasticizers, the setting times are reduced, extended or unaffected (Ramachandran,
etal., 1981).
2.2.2. Estimation Methods for Setting Times
The setting times of cement pastes are usually determined by the Vicat needle
test. This test measures the resistance to a penetration of a needle (<j) = 1 mm) under a
300 g load o f plunger alone without any external force (ASTM C187, ASTM C191). The
Vicat needle apparatus is shown in Figure 2-2. The time at 25 mm penetration indicates
the initial set and no visible penetration indicates the final set. The Gillmore test is also
able to obtain the setting times of cement pastes as detailed elsewhere (ASTM C266).
The Vicat test also estimates the setting times o f mortars by using a 2 mm in diameter
penetration needle instead. The time at 10 mm penetration indicates the initial set for
mortar test and no visible penetration indicates the final set (ASTM C807).
For concrete, the setting times are approximately defined by the penetration
resistance method (ASTM C403) instead o f the Vicat test. This test requires mortar from
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Movable Rod
Adjustable Indicator
Set Screw
ZZZ
Removable Needle
<j>= 1 mm
Conical Ring
1
Glass Plate -------
Figure 2-2 Vicat Apparatus.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
sieving the test concrete through a No.4 screen. Any one o f several standard sizes of pins
•y
(1, 1/2, 1/4, l / l 0, 1/20, 1/40 in") may be used. Penetration resistance is measured at
various times using several identical mixtures. The time at 500 and 4000 psi (3.5 and
27.6 MPa) penetration resistance are defined as the initial and final setting times,
respectively. Figure 2-3 shows a typical plot of penetration resistance versus elapsed
time and hand fit curve used to determine setting times. Polivka and Klein found that
penetration resistance P could be expressed as an exponential function o f time t, shown in
the following equation (Polivka and Klein, 1960):
P = aeb' : a and b are coefficients
(Eq. 2-9)
Eren suggested that a power function would be appropriate (Eren, et al., 1995), that is:
P = At8 : A and B are coefficients
(Eq. 2-10)
However, the setting times merely define two arbitrary points in the general
relationship between the time o f water addition and penetration resistance gain; thus they
have no fundamental significance and do not indicate the real strength o f samples. In
fact, at the initial setting time the compressive strength is zero; at final setting time the
compressive strength is about 100 psi. This is in great contrast to penetration resistance
of 500 and 4000 psi respectively (Mehta and Monteiro, 1993).
Recently, the instrumented penetration test was developed, which is similar to the
penetration resistance test for concrete, as mentioned above (Croft, 1996). By using this
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
Penetration Resistance (psi)
5000
-3 0
Final Setting
4000
-2 5
*0
3
CD
cd
pa
o ’
3
3000“
-2 0
-1 5
73
CD
03
55’
pa
3
O
CD
-10
1000
“
Initial Setting
180
210
240
*0
Outlier
270
300
330
-5
360
pa
420
Time after Mixing (min)
Figure 2-3 Typical plot o f penetration resistance values versus elapsed time and hand fit
curve used to determine time o f setting (Note: Not drawn to actual scale)
(ASTM C403).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
test, the setting times of cement paste were successfully measured. This will be discussed
later. This penetration test has several advantages over the conventional Vicat needle test
or the penetration resistance test, as follows:
(1 )/« situ and continuous measurement o f the hardness development.
(2) Ability to get more information regarding cement hydration process.
(3) Application under controlled laboratory conditions as well as under field conditions.
(4) Only a single specimen is tested (This saves the labor and cost o f producing and
testing several specimens).
(5) Reduction of the experimental error because o f no need for handling during test.
2.3. Elevated Temperature Curing
Some important objectives in precast concrete processing are reducing the curing
time, cost and workshop area. Since the curing o f concrete can be accelerated by raising
its temperature, the use o f external heating sources has been investigated. This can be
done in practice by passing an electrical current though concrete or by placing concrete in
hot environments, such as steam, hot water or ovens (D.F. Orchard, 1973).
Among them, the steam curing process has widely been investigated due to its
simplicity.
It operates at atmospheric pressure and less than 100°C.
This process
accelerates cement hydration to make the curing time shorter (Mindess and Young,
1981). Thus, steam curing can offer the precast industry an attractive substitute for
normal curing. Unfortunately, steam curing has some drawbacks; although steam cured
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
specimens show a great increase in early strength, long-term strength is lower than
specimens cured under normal conditions (Ravina and Shalon, 1971; Alexanderson,
1973).
2.3.1. Effects o f Elevated Temperature on Cement Hydration
Since concrete properties are dominated by its cement hydration, it is important to
investigate the effects o f temperature on the cement reaction mechanism so that changes
of concrete properties can be determined. Verbeck and Helmuth (Verbeck and Helmuth,
1968) first tried to explain the temperature effects in view o f the relationships between
the hydration mechanism and the microstructural development of cement paste. At low
temperature, there is ample time for the hydration products to diffuse and precipitate
uniformly throughout the interstitial space between the cement grains.
However,
elevated temperature does not allow time for such diffusion, resulting in high
concentration of hydration products surrounding the hydrating cement grain and large
pores between the grains.
The coarsening of pores at elevated temperature was proved by BET (BrunauerEmmett-Teller) results (Skalny and Odler, 1972).
The surface area o f the room
temperature cured specimen continuously increases with hydration, while that of
specimens treated at higher temperature decreases after reaching a maximum to become
smaller than for room temperature cured materials.
The reduction o f surface area
indicates growing pore size. Large populations of coarsened pores caused by the elevated
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
temperature was also confirmed by MIP (Mercury Intrusion Porosimetry) (Sellevold,
1974). Another effect o f the elevated temperature is the reduction o f the overall degree
o f hydration for later age, which was investigated by x-ray diffraction analysis (AlunnoRossetti, 1974) and loss on ignition experiments (Parcevaux, 1984).
Backscattered
electron microscopic work revealed that four different phases were found in hydrated
cement pastes: unhydrated cement, CH, other hydration products, and pores (Scrivener,
1992).
Using microscopy, other investigators found that the phase distribution became
less homogeneous and the volume fraction o f large pores was increased as temperature
increased (Kjellsen, et al., 1990; Kjellsen, et al., 1991). The C-S-H hydration products
around unhydrated cement grains became denser with increased curing temperature. The
annular ring-shape structures around unhydrated cement grains were produced by
changing temperature. However, the compositions o f the dark and the bright zone were
not significantly different (Kjellsen, 1996). The dense “shell” structure (or darker zone)
was also detected by x-ray study (Alunno-Rossetti, 1974) and is believed to serve as
diffusion barrier. This barrier reduces the ultimate degree of hydration and coarsens the
pore structure in the interstitial space, resulting in a detrimental effect on the strength
even for the same degree of hydration.
The CH crystal morphology was changed from lamellar to more coarse crystalline
as temperature increased (Patel, et al., 1996). The coarsening of pores due to elevated
temperature detrimentally influences the strength (Bentur, et al. 1979) and durability
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
(Detwiler, et al., 1991). Additionally, microcracks became more prevalent (Patel, et al.,
1996) and elastic modulus is decreased (Nassar, 1973).
The curing temperature also altered the nature o f silicate formation.
TMS
(Trimethylsilyl) method (or gel permeation chromatography) and 29Si NMR with MASS
(magic angle sample spinning) were used to study the development of the silicate units in
C-S-H (Young, 1988). Silicate dimers predominate during room temperature curing, but
elevated temperature tends to favor formation of oligomer silicate units. Monosilicate
formation occurs only at very low temperature (= 2°C). Therefore, the formation o f
hydration products is affected by curing temperature.
There are reports that the temperature range from 70°C to 90°C may be an
important transition range for cement hydration. This was proven by increase in the
threshold diameter (MIP), change in the pore size distribution (Parcevaux, 1984), and an
irreversible modulus transition (Radjy and Richards, 1973).
Therefore, it became a
critical issue for steam curing not to overheat into this range. Considering this point, a
plausible design o f the curing system was suggested (Wainwright and Tolloczko, 1983).
This result was to heat and control the environmental temperature with a feedback signal
from the real specimen temperature, making sure the temperature matched throughout the
curing process. However, the heating rate o f 6 °C/hour obtained by this process was too
slow to be practical.
Generally, as mentioned earlier, the long-term strength decreased as curing
temperature increased. However, it is unclear to state the temperature effects on the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
cement hydration without the knowledge of temperature history because the cement
hydration mechanism, in itself, is characterized by its heat evolution and time-dependent
reactions. Typically, in the steam curing process the specimen is heated after a 2-4 hour
presteaming period and held for about 16 hours at a set temperature and then cooled
down to room temperature for storage before testing. The presteaming increases the
heating rate to 30°C/hour (Mindess and Young, 1981). Without this presteaming period,
the maximum heating rate is restricted to 10°C/hour (Taylor, 1990). This restriction may
be related to overheating due to the exothermic hydration reactions.
Interestingly, strength values reported when specimens were cured differently
were inconsistent. Concrete heated by steam immediately after mixing and held for four
hours at a temperature less than 60°C had strengths close to that of concrete cured at
room temperature (Lapinas, 1973). Kjellsen also found that the strengths o f mortars were
not significantly degraded when the specimens were cooled after holding for 4-hours at
50°C (Kjellsen, et al., 1991). Gjorv also reported no strength degradation with using dry
aggregate (Gjorv, et al.,1994)
The common factors in these studies which report no
degradation in strength are as follows:
(1) Heating rates were much higher than a typical steam curing.
(2) Maximum temperatures were less than 60°C.
(3) No overheating was found because of the careful heating control.
(4) Holding times at the elevated temperature were less than 4 hours.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
Consequently, a high temperature curing apparatus which provides scientific
insight into the true effects o f the elevated temperature curing on cement hydration while
precisely controlling specimen temperature is desired. Other reasons for pursuing the
technique include cost-efifectiveness, simplicity of operation, and industrial application.
2.3.2. Applications o f Microwave Energy for Curing of Cement
The steam curing process relies on the thermal conductivity o f the specimen, with
heat flowing from the exterior to the interior. The larger the material, the slower the heat
penetration to the interior and the less uniform the heating throughout the material.
However, high frequency electromagnetic heating, such as microwave enhanced heating,
is able to reduce such nonuniformity due to its superior penetration depth of microwaves
(Sutton, 1989).
Also microwave energy enables evaporation o f residual free water,
which is added for better workability and easier molding, but which is believed to make
the physical properties worse, prior to setting. The evaporation of water by microwave
heating is uniform and rapid throughout the whole body of the specimen, developing
irreversible and consistent plastic shrinkage (Wu, et al., 1987).
Microwave processing has several advantages over steam curing in the precast
concrete industry: first, microwave energy can heat a specimen uniformly and
volumetrically because it is less dependent of the thermal conductivity o f the specimen.
Second, it can more easily enhance the evaporation rate and better control energy
absorption, and thereby optimize the overall heating process before demolding. Lastly,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
the final performance o f the cement-based materials can potentially be improved
(Christo, 1989).
For the microwave curing process, the curing period was much shorter than that
o f normal curing, but the improvement in the quality, represented as 28-day compressive
strength, is still inconsistent.
This may be related to uncertainty in temperature
measurements. One shortcoming of previous studies is uncertainty about the thermal
history o f the specimen during curing; typically the power, rather than the temperature,
was fixed. Isothermal heating, achieved best by feedback temperature control, is required
to properly investigate the temperature effects on the properties. The pioneering work of
Watson shows that 28-day compressive strength o f microwave-cured concretes displayed
only half the strength o f the normally cured concretes (Watson, 1968). However, his
results were uncertain because the temperature of specimens might have fluctuated due to
the pulsed microwave energy which he used. Also, an internal temperature o f 90°C was
reached at which cracks could be generated, resulting from the escape o f steam from the
interior (Christo, 1989). However, Wu and coworkers reported that microwave curing
improved the 28-day compressive strength o f mortar as much as 3-7% as well as
enhancing short-term strength (Wu, et al., 1987). They emphasized optimization o f the
internal temperature and the final water-to-cement ratio o f the specimen, controlling
processing time and microwave power.
Too much microwave energy could cause a
decrease o f strength due to overevaporation and overheating.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
Other independent results confirmed the possibility o f improving the 28-day
compressive strength by microwave processing (Christo, 1989; Hutchison, et al., 1991;
Chang, 1994; Li, 1994). Recently, Leung and Pheeraphan reported that the microwave
curing technique can potentially produce concrete with very high early strength and little
deterioration in its long-term performance by controlling material compositions and
microwave power (Leung and Pheeraphan, 1995). In a subsequent study, they found the
most important parameter to control was specimen temperature during the curing process.
They accomplished temperature control by either a feedback temperature control or
discrete power application based on the data from feedback temperature control. Both
ways can reduce overheating and thermal shock during curing (Leung and Pheeraphan,
1997). The compressive strength results of microwave cured specimens, comparing with
specimens cured at room temperature, are summarized in Figure 2-4. The effect of the
microwave enhanced heating process on cement-based materials needs more careful and
precise work that includes the monitoring o f process variables that will affect the
properties o f the final products.
Leung and Pheeraphan suggested plausible designs for microwave applicators in
industry, shown in Figure 2-5 (Leung and Pheeraphan, 1995). Figure 2-5 (a) shows the
application to the precast masonry block industry. Being loaded on the production line,
the normally cast blocks are heated appropriately by the microwave generators until
reaching the gate to storage. When a cured block is moved out, a new uncured block
enters the oven simultaneously. Depending on the optimum condition for each process,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
T3
<
u,D
ZS
U
<u
>
<33
25
70
as
a,
/
60■
s
50-
A
£
O
•/ "
40O
00
c
a>
s -l
S
w '
/
.
■
s
s
□
a
•
A
s
A
o
O
A
/
o
o
'
30-
/
/
/
s
/
4-
/
V
v
T
□
•*->
C/3
#1
#2
#3
#4
#5
#6
#7
#8
#9
#10
#11
#12
-F
i
20
20
|
i
30
|
i
40
|
50
i
|
i
60
70
Strength of Normal Cured (MPa)
#
Material
w/c
Time Tmax Days
Reference
(Min) (°C)
1 Concrete 0.6
1:2:4
Constant I kW
90
60
28
Watson, 1968
*2
Mortar 0.44 1:2.5:0 Constant 150 W 30
58
28
Wu, 1987
Mortar 0.44 1:2.5 :0 Constant 150 W 90 N/A 28
Wu, 1987
4
Mortar 0.44 1:2.5:0 Constant 300 W 30 N/A 28
Wu, 1987
5
Mortar 0.44 1:2.5 :0 Constant 300 W 60 N/A 28
Wu, 1987
*6
Mortar 0.44
1 :2:0
Constant 50 W
40
65
28 Hutchison, 1991
*1 Concrete 0.45 1:1.5:3 Constantl50 W 45
60
28
Li, 1994
8 Concrete 0.45 1:1.5:3 Constant 150 W
90 N/A 28
Li, 1994
9
Mortar 0.5
1 :2:0
Feedback control 45
7
80
Leung, 1997
10
Mortar 0.5
1 :2:0
Feedback control 90
7
60
Leung, 1997
11 Concrete 0.4
1 : 1 : 1 .5 Feedback control 45
80
7
Leung, 1997
*12 Concrete 0.4
1 : 1 : 1 .5 Discrete power
45
60
7
Leung, 1997
*: Strengths o f microwave cured specimens (solid symbols) were equal or better.
Figure 2-4
C:S:A
Conditions
Comparisons in compressive strengths between microwave and normal
curing from several references (See above table).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
Microwave Generator
Chamber
Door
From
Molding
Machine
To Storage
Specimens
»
Roller
Chamber
Door
o~o (To a
:
£ED (2 j D
(a) Precast concrete blocks
A Section
of the
Microwave
Applicator
Microwave
Generator
Metallic
Enclosure
A Section o f
the Precast Bed
for Concrete Slab
Concrete, Reinforced
or Prestressed
(b) Precast concrete slab
Microwave Generator
Shield
Diaphragm
with Opening
Surrounding
Metal Skirt
Old Concrete
in Pavement
New Concrete Patch
(c) Pavement repair
Figure 2-5 Plausible microwave applications (Leung and Pheeraphan, 1995).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
the length o f the oven, the speed o f belt or the number o f microwave applicators can be
determined. A conceptual design for the casting o f a long concrete slab is shown in
Figure 2-5 (b). The casting operation can determine the applicator length and the number
o f microwave generators that are placed on top o f the metallic bed. The simultaneous
casting procedure o f the whole slab needs an applicator as long as the bed, while the
casting procedure from one side requires a shorter applicator with an appropriate amount
o f microwave energy. The conceptual applicator for pavement repair should consist o f a
source, an external shield, an internal diaphragm with openings and a surrounding skirt,
shown in Figure 2-5 (c), because it is impossible to completely enclose the repair patch in
an oven. Since the impedance o f concrete changes as curing proceeds, the diaphragm
position and opening size need to be mechanically adjusted to keep the efficiency high
during the whole curing process. The surrounding skirt reduces microwave radiation
from the ground to meet safety requirements.
2.4. Rheology o f Cement Pastes
Rheology is defined as “the science o f the deformation and flow o f matter.”
(Tattersall and Banfill, 1983) In practice, rheology is concerned with materials whose
flow properties are more complicated than those o f a simple fluid or ideal elastic solid.
Rheology is interested in establishing relationships between stress, deformation and rate
o f deformation o f fluids (Stein, 1986). The primary objective of studying rheology in the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
cement area is to provide fundamental information for explanation and prediction of
concrete properties, which relate to rheological properties.
2.4.1. Models o f Flow Behavior
‘Newtonian’ fluids like water or honey exhibit the simplest relationships between
shear stress, r, and shear rate, y :
Newton’s Law: r = rjy
(Eq. 2-11)
where the ratio o f shear stress to shear rate, q, is defined as viscosity. The SI unit of
viscosity is Pa-s (1 Pa-s = 1000 centipoise (cp)).
The viscosity o f water at room
temperature is about 1 cp and that o f honey is about 10,000 cp. Generally, the viscosity
is very sensitive to temperature.
If stress and strain rate are not linearly related, viscosity is a more complex
function o f shear rate.
Non-Newtonian fluid viscosity is dependent on y and shear
history. The simplest class o f Non-Newtonian materials is ‘Bingham’, whose flow curve
is a straight line but does not pass through the origin (see Figure 2-6), shown in Eq. 2-12.
Bingham: r = r 0 + Tjy
(Eq. 2-12)
The more complicated case is that the flow curve may not be linear at all, like power-Iaw.
Power-Iaw: z = k y n
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Eq. 2-13)
29
Shear-Thickening
Shear Stress (t)
Shear-Thinning
Bingham
Newtonian
o
Shear Rate (dy/dt)
Figure 2-6 Flow curves of several flow models.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
where k is the plastic viscosity.
A material whose flow curve is concave towards the stress axis is said to be 'shear
thickening’ (or ‘dilatant’) because it becomes more difficult to flow at higher shear rate.
The opposite case is said to be 'shear thinning’ (or ‘pseudoplastic’). In the power-Iaw
model, when the power n is greater than 1, the material is shear thickening; when it is less
than 1, it is shear thinning. Figure 2-6 shows these flow curves. The more general
expression o f the Non-Newtonian flow is ‘Hershel-Buckley’ (Tattersall and Banfill,
1983).
Hershel-Buckley: r = T0 + k y n
(Eq. 2-14)
Shear thinning is o f considerable interest in the concrete related area. The reason
for thinning (the slope decreases as the shear rate increases) is that the shearing forces are
breaking some structure in the material, the progress of the destruction being greater the
higher the shear rate, the ‘up-curve’ on Figure 2-7. The ‘down-curve’ on Figure 2-7
indicates a trace o f the shear rate as this shearing force is reduced.
If the structural
breakdown is immediately and instantaneously reversible, a decrease in shear rate (or
down-curve) will result in a progressive build-up o f structure to the same state as it
increases (or up-curve). Thus, the down-curve and up-curve will be identical.
However, for fluids for which time is required for the structure to rebuild, or
which does not rebuild at all, the curves do not coincide, resulting in a hysteresis loop, as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Shear Stress (x)
31
up-curve
down-curve
o
Shear Rate (dy/dt)
Figure 2-7 Hysteresis loop for material suffering structure breakdown under shear stress.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
shown in Figure 2-7. This hysteresis loop is typical of what is known as a ‘thixotropic’
material, that is, a material that becomes thinner when it is disturbed and thickens up
again when it is left alone. The reverse case is said to be ‘antithixotropic’.
Flocculated cement pastes have shear thinning or shear thickening behavior, but
pastes act more nearly Newtonian when well dispersed by water reducers or
superplasticizers (Struble, 1991). The time-dependent behaviors o f cement pastes are
both thixotropic and antithixotropic, depending on material conditions (Struble, 1991).
2.4.2. Viscoelastic Properties
One o f the experimental procedures to characterize viscoelastic materials like
cement pastes is oscillatory shear.
When a sinusoidal shear strain is applied to a
viscoelastic material, the shear stress is obtained as a sinusoidal function with a phase
angle (Tadmor and Gogos, 1979). The ratio o f shear stress to strain can be calculated
from equations, called a complex modulus, G*.
Shear Strain: y (t ) =y 0 e ,a>'
(Eq. 2-15)
Shear Stress: r ( / ) = r0 e '(a"+<n
(Eq. 2-16)
Complex Modulus: G* = — = — (cos5 + / sin<5) = G’ + iG"
r
To
(Eq. 2-17)
The real part o f the modulus, G \ is called the elastic storage modulus because it
defines the energy stored in the specimen due to the applied strain. The imaginary part,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
G”, is called viscous loss modulus which indicates the dissipation o f the energy. For a
Hookean solid, the phase angle is 0° and the loss modulus is zero; whereas for a fluid, the
phase angle is 90°, then the storage modulus is zero. Viscoelastic materials have phase
angles between 0° -90° and both moduli are non-trivial.
A special viscoelastic material like cement paste shows a solid-like behavior
below a critical strain which strongly depends on the structure of the material. Below the
critical strain, particles in a flocculated suspension can easily recover elastically to their
equilibrium positions, thus maintaining the structural integrity o f the flocculated network.
The storage modulus, G \ is then independent o f the applied strain amplitude. However,
when particles are displaced far enough from their equilibrium position due to applied
strain above the critical value, the attractive interaction is no longer effective, resulting in
losing the structural integrity of the flocculated network. Then, the material behaves
fluid-like (Schultz and Struble, 1993).
Yielding phenomena can occur in viscoelastic materials and there are two types of
yield stress, namely dynamic and static. Dynamic yield stress refers to a stress which
must be exceeded in order to initiate flow, for instance, the Bingham yield stress is a
dynamic yield stress.
The static yield stress is the critical stress beyond which the
material becomes nonlinear, for instance, a true power-Iaw fluid.
Since the dynamic
yield stress o f a viscoelastic material appears when the structure of the suspension has
broken down and flows, while the static yield stress is at the critical stress, dynamic
yielding occurs before static yielding (Yang, 1994).
Both yield stresses provide
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
information concerning the microstructure o f flocculated suspensions.
They are not
interpreted by the behavior o f viscosity because the latter is measured after breaking the
network structure.
2.4.3. Rheological Behavior of Cement Pastes
Tattersall and Banfill explained the origin o f cement rheology during hydration.
Dry cement particles are in close contact and held together by weak surface forces.
Within a few seconds o f the initial contact between cement and water, a membrane of
gelatinous hydration product like calcium silicate-sulfoaluminate hydrate is formed
around these cement particle surfaces.
The membrane holds particles much more
strongly than the weak surface force in the dry powder.
When enough o f a shear force is applied, the bridging membrane will be ruptured
to separate the particles. Once this happens, the hydration process resumes instantly on
the newly exposed surface to precipitate new membranes. Thus, the bridge is unable to
reform after breaking.
This is an irreversible process o f structural breakdown. The
irreversible breakdown caused by rupturing membranes due to shear force must be
distinguished from the reversible Bingham behavior where the structure, due to its
attraction force between particles is able to hold particles together when shear stops
(Tattersall and Banfill, 1983).
Since the rheological behavior o f cement paste includes both fluid-like and solid­
like properties, flow curves o f cement pastes are complicated. Three types o f hysteresis
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
loops are observed in cement paste flow curves (Tattersall and Banfill, 1983). Since
variations in flow curves are closely related to the time intervals o f cycles, these
differences might be explained by a qualitative model based on two simultaneous
irreversible process - shear dependent structural breakdown and time dependent build-up
of structure due to hydration. Figure 2-8 shows a typical flow curve of cement paste,
appearing especially with short cycle times. The mismatch between the up-curve and the
down-curve indicates decreasing resistance to the applied shear stress, which indicates
irreversible structural breakdown o f the cement paste. In spite o f variations in the shape
o f the up-curve in the three flow curves, their down curves are very similar, and are
typically Bingham flow. For this reason, the Bingham yield stress and plastic viscosity
values are mostly determined from the down-curve.
The flow behavior o f cement paste is influenced by several factors, such as shear
history during mixing and rheological testing, water-to-cement ratio, cement fineness,
chemical composition, and age o f the paste. Prolonged and intensive mixing may cause
significant structural breakdown, resulting in reduction o f Bingham yield stress, plastic
viscosity and the amount o f hysteresis (Roy and Asaga, 1979; Banfill, 1981). Longer
mixing time or vibration produced a reversible flow curve, reducing both Bingham yield
stress and plastic viscosity to equilibrium values (Banfill, 1981; Johns and Taylor, 1977).
Summarizing the results o f researchers, Bingham yield stress and plastic viscosity
exponentially decrease as water-to-cement ratio increases or fineness decreases
(Tattersall and Banfill, 1983).
It was found that Bingham yield stress and plastic
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
Bingham Flow
Shear Rate (dy/dt)
Figure 2-8 Type I hysteresis flow curve o f cement paste (Tattersall and Banfill, 1983).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
viscosity were power law functions of fineness and the stress showed a higher power
index than the viscosity (Vom Berg, 1979), which indicates the stress is more strongly
dependent on attractive forces between surfaces than is viscosity.
The variations in chemical compositions o f normal Portland cement paste
insignificantly affect the rheological behavior o f cement paste compared to the effects of
water-to-cement ratio and fineness (Odler, et al., 1978).
However, when the rate of
hydration is much accelerated or retarded by adding admixtures, Bingham yield stress
and plastic viscosity will be influenced gready. These phenomena are closely related to
changing C3A reactivity, because the hydration of C 3A occurs prior to other reactions
when the paste is still viscoelastic. It is proved that the addition of alkalis, which raise
the reactivity o f C3A, increases both Bingham yield stress and plastic viscosity of pastes
(Greszczyk, 1991).
Due to the existence of the induction and acceleration periods during cement
hydration, the effect o f age on rheological behavior is to be expected. In the induction
period, the slow processes taking place in the paste cause the paste to stiffen, and both
Bingham yield stress and plastic viscosity slightly increase with time. However, after the
induction period, the stiffening accelerates as the paste sets, and both parameters increase
much more rapidly (Tattersall and Banfill, 1983).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
2.5. Pozzolanic Materials
A ‘pozzolan’ is a finely divided siliceous or siliceous and aluminous material that
reacts chemically with slaked lime at ordinary temperature and in presence o f moisture to
form a strong slow-hardening cement (Merriam Webster’s Collegiate Dictionary, 10th
ed., 1993). The reasons for using these pozzolans are usually economic: they are cheaper
than Portland cement because they exist as natural deposits requiring little processing or
because they are a byproduct or waste from industry. The other advantages o f utilizing
pozzolans are ( 1) decreasing overall energy consumption, (2 ) improving workability,
impermeability, durability, and resistance to chemical attack, and (3) reducing alkaliaggregate reaction. Although manufacturers in the USA have produced little blended
cement due to marketing concerns and corporate strategy, they should be o f interest soon
because there are sufficient good quality pozzolans available in the USA. (Malhortra and
Hemmings, 1995).
2.5.1. Classification of Pozzolans
Since summarizing all pozzolans is too extensive, three artificial pozzolans,
namely blast furnace slag, silica fume, and fly ash, will be reviewed in this study. These
are widely used cement-based materials, composed of the same oxides as cement clinker,
but in different proportions and mineralogical compositions. The phase diagram for
ternary systems o f CaO-SiaO-AUCb with compositions o f Portland cement and pozzolans
is shown in Figure 2-9.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
SiO
Silica Fume
Natural Pozzolan
Fly Ash
Class F
Class C
Blastfurnace
Slag
C2S
Portland
Cement
CS
Alumina Cement
Bauxite , j
Limestone
CaO
CA
Figure 2-9 Phase diagram for ternary systems o f CaO-SiiO-AhCh.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ALO
40
Blast furnace slag is a waste product in the manufacture o f pig iron, about 300 kg
o f slag being produced for each ton o f pig iron. Blast furnace slag varied in physical
structure depending on the processes used and on the method o f cooling of the slag. For
use in the manufacture o f slag cement, the slag has to be quenched so that it solidifies as
glass (crystallization being largely prevented) then ground.
Thus it is called ground
granulated blast furnace slag (GGBFS). The specific gravity o f GGBFS is about 2.9 (cf.
Portland cement is 3.15) and surface area is usually greater than 350 m 2/kg.
Silica fume is a by-product o f the manufacture o f silicon and ferrosilicon alloys
from high purity quartz and coal in a submerged-arc electric furnace.
The escaping
gaseous SiO oxidizes and condenses in the form o f extremely fine spherical particles of
amorphous silica (Si0 2 ). Glass formed silica is highly reactive, and the fineness o f the
particles speeds up the reaction with CH produced by the hydration of Portland cement.
The extremely small particles of silica fume can enter space between cement particles,
and thus improve packing. The specific gravity o f silica fume is 2.20 and the specific
surface area determined by nitrogen absorption is about 20000 m 2/kg. Silica fume is
available in the form o f micropellets and slurry.
Fly ash, (or pulverized-fuel ash) is the ash precipitated elecrostatically or
mechanically from the exhaust gases of coal-fired power stations. The fly ashes are
spherical particles and have a very high fineness. The specific surface is between 250
and 600 m 2/kg.
The typical specific gravity is 2.35.
Class F fly ash, derived from
bituminous coal, is mainly siliceous. Sub-bituminous coal and lignite result in high-lime
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
ash, known as Class C fly ash. Fly ash has lots o f variations in glass content, carbon
content, particle shape and size distribution due to the conditions o f power plants.
Chemical compositions o f slag (GGBFS), silica fume, fly ash (Class F and C) are
compared to the compositions of Portland cement in in Table 2.2.
When other pozzolanic materials are included, they are called blended hydraulic
cements, defined as “A hydraulic cement consisting o f two or more inorganic constituents
that contribute to the strength-gaining properties of the cement, with or without other
constituents, processing additions and functional additions” (ASTM C l 157). Table 2.3
and 2.4 show standards o f compound composition for blended cements in the USA and
Europe.
2.5.2. Effects o f Pozzolans on Hydration Reactions and Rheology
The final products o f pozzolan-lime reactions are the same as those of Portland
cement hydration, such as C-S-H (but C/S « 1), C4AH 13, C 8AFH 26 and C4ASH 12 (Sersale,
1983). Also, a pozzolan reacts with CH produced during cement hydration to form
additional C-S-H due to the additional silica source. This is called pozzolanic reaction or
pozzolanic activity (Mindess and Young, 1981).
CH+S+H
=>
C-S-H (Calcium Silicate Hydrates)
(Eq. 2-18)
The effect o f the pozzolanic reaction is then to increase the proportion of C-S-H
in the hydrated paste at the expense o f CH. When sufficient reactive
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
Table 2-2 Chemical compositions in pozzolanic materials.
Component
CaO
S i0 2
A120 3
FeO (Fe20 3 )
MgO
Na20
K20
S03
L.O.I.
OPC
Type I
65.0
20.3
4.6
Blast furnace
Slag
24.2
33
13
2.8
1.2
6.0
1.9
0.2
0.5
2.5
1.5
Silica Fume
Fly Ash
(F)
2.4
49
28
9.5
0.08-0.3
94-98
0.1-0.4
0.02-0.15
0.3-0.9
0.1-0.4
0.2-0.7
0.4
0.7
0.04
0.42
Fly Ash
(C)
13-25
28-48
11-22
1.5
4.2
5-10
2.9-1.9
0.2-13
0.5-1.5
—
1.2
1-12
0.8-1.5
0.3
0 . 1- 1.8
1.6
Table 2-3 Classification o f cement according to ASTM C595.
Traditional Description
Portland-Blast Furnace Slag Cement
Portland-Pozzolan Cement
Slag Cement
ASTM
Type IS or Type I (SM)
Type IP or Type I (PM)
Type S
Table 2-4 Classification o f cement according to European Standard.
Type
Designation
I
II/A
II/B
n/A
II/B
II/A
Portland
Portland
Slag
Portland
Fly ash
Portland
Silica Fume
Portland
Composite
Blast furnace
Slag
II/A
II/B
IE/A
III/B
III/C
TV/A
rv/B
Pozzolanic
V 'ass percenltage o f materia s
Clinker
Silica Fume GGBFS
Fly ash
95-100
80-94
6-20
65-79
21-35
_
80-94
6-20
65-79
21-35
90-94
6-10
—
—
—
—
—
—
—
___
—
—
___
80-94
65-79
35-64
20-34
5-19
65-89
45-64
___
<---------<----------
<
<
_
_
___
___
—
—
11-35
36-55
---------- >
6-20
21-35
----------------------------
>
36-65
66-80
81-95
>
>
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
_
—
43
alumina is also present in a pozzolan like fly ash, it can react analogously to silica:
CH+A+H
=>
C-A-H (Calcium Aluminate Hydrates)
(Eq. 2-19)
Since calcium aluminate hydrates can react expansively with sulfate to form ettringite,
pozzolans low in alumina should be used to improve sulfate resistance.
The various pozzolanic materials affect the progress o f hydration and properties
in consequence o f their chemical composition, reactivity, particle size distribution, and
particle shape. Pozzolans except silica fume will lower the total heat of hydration and
also the rate o f liberation o f heat, because the reactivity o f pozzolans is generally quite
low.
These can be used to advantage in warm weather to minimize cold joint,
particularly in large pours, though stiffening due to water evaporation will still occur
(Mindess and Young, 1981). Therefore, one of the major advantages of using blended
cements with high replacement levels is to modify temperature profiles in mass concrete
sections (Miller, 1993).
Concrete containing pozzolans has different rheological properties, being
generally more fluid and having extended stiffening times. The inclusion o f fly ash, and
to a lesser extent GGBFS, improves cohesion and workability o f the mix because o f a
ball bearing effect based on the spherical nature o f particles. Different characteristics are
observed with concrete containing silica fume, where the dispersion of the materials
gives it a thixotropic nature: It appears to be stiff but when vibrated is very workable and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
readily placed by all conventional methods. It is necessary to use superplasticizers with
silica fume when designing higher strength concretes (Miller, 1993).
One obvious difference between Portland cement and blended cement is the color
of the resultant concrete, depending on individual materials (Miller, 1993). Fly ash and
silica fiime impart a darker color. With the addition o f GGBFS, the concrete initially has
a bluish hue when first demolded due to the presence o f sulfide, but the color soon reverts
to a more standard tone.
2.5.3. Strength Changes by Adding Pozzolans
The most acknowledged property of blended cements is the difference in rates of
gain in both short and long term strengths, these being due to the differences in hydration
chemistry o f pozzolans. The higher the replacing amounts, the lower the early strength,
except silica fume with which strength is similar to 100% Portland cement concrete
(Neville, 1996). However, the latent strengths beyond 28 days are increased depending
on the type and quantity of pozzolans present (Ramezanianpour and Malhotra, 1995).
The initial hydration o f GGBFS is very slow because it depends on the
breakdown o f the glass by the hydroxyl ions released during the hydration o f Portland
cement. However, because a mixture o f Portland cement and GGBFS contains more
silica and less lime than Portland cement alone, hydration o f the blended cement
produces more C-S-H, resulting in denser microstructure. The progressive release of
alkalis by the GGBFS with the formation o f CH by Portland cement results in a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
continuous reaction o f GGBFS over a long period. Thus, long-term strength will be
greater than Portland cement alone, in spite o f lower early strength gain (Hogan, 1981).
The study o f the strength development o f mortar containing varying proportions of
GGBFS suggests an optimum GGBFS content o f about 50% by weight from a strength
standpoint (Roy and Idom, 1982).
The hydration o f fly ash is chemically sensitive to the alkalinity o f the pore water
(Fraay, et al., 1989), which depends on the properties o f the mixed Portland cement.
Additionally, fly ash has a physical packing effect at the interface of coarse aggregate
particles to improve microstructure (Neville, 1996), which is absent in mortars. For both
reasons, quantified predictions o f the influence o f fly ash on strength of mortar or
concrete may be impossible. Typical compressive strength results of fly ash-blended
concrete shows very slow early strength gain, similar to GGBFS concretes, but the
strength at 28 days is still lower that that o f normal concretes, unlike GGBFS concretes
(Ramezanianpour and Malhotra, 1995). Class C fly ash concretes are a little bit stronger
than Class F concrete (Neville, 1996). However, the fly ash concretes become stronger
beyond I year when replacing less than 30% (Odler, 1991). The maximum dosages of
fly ash for obtaining the appropriate long-term strength are 20% by weight for Class C
and 30% for Class F fly ash.
Silica fume concrete has some different hydration characteristics due to the very
finely divided nature o f the reactive silica fume. The extremely fineness of silica fume,
which provides more nucleation sites for CH, contributes to early strength development.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
Since silica fume dissolves in a saturated solution of CH within a few minutes, C-S-H can
be formed on the surface o f silica fume particles, as soon as the pore solution is saturated
with CH during Portland cement hydration (Roy, 1987). This initial reaction proceeds at
a high rate, but subsequent reaction is very slow (Roy, 1992). A consequence o f the
rapidity o f the early reactions in concretes containing silica fume is that the development
of heat o f hydration may be as high as when rapid-hardening Portland cement (Type III)
is used alone (Roy, 1987).
Another contribution o f silica fume to the early strength development is through
improvement o f the interfacial zone with the aggregate (Benz, et al., 1992).
The
strengths o f silica fume concretes are also higher than normal concretes at 28 days.
However, the strength development o f silica fume containing concrete ceases much
earlier than normal concrete resulting from self-desiccation phenomena due to being
difficult for water supply through its dense microstructure.
Therefore, the strength
improvement by adding silica fume could disappear after 1 year, compared with normal
concretes (Hooton, 1993). The optimum amount o f silica fume replacement is 10% by
weight (Neville, 1996). Figure 2-10 shows the relative strength developments o f concrete
influenced by appropriate replacements o f GGBFS, fly ash, or silica fume.
2.5.4. Influence o f Pozzolans on the Concrete Durability
Durability in the concrete related area can be broadly defined as the ability of
performing a planned function for a designated period with minimal need for repair or
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
Relative Compressive Strength (%)
175
150
125
100
75
50
— &— 50 w t%
10 w t%
25 w t%
— 25 w t%
25
GGBFS
Silica Fume
Class F Fly Ash
Class C Fly Ash
0
1
10
100
1000
Age (days)
Figure 2-10 Relative strength development o f concretes influenced by replacing
50 wt% GGBFS, 10 wt% silica fume, or 25 wt% fly ash
(Reference for Data: Neville, 1996).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
maintenance (Miller, 1993).
The permeability o f concrete is fundamental to most
durability factors, that is the ability for foreign agents to enter the concrete and cause
damage to the structure.
Although a general correlation might be expected between
permeability and porosity, pozzolanic cement pastes show higher porosity and lower
permeability (Hooton, 1986). This apparent contradiction can be explained through the
following model. After the hydration products o f cement are formed and develop on site,
pozzolan particles are deposited on the remaining part because pozzolans hydrates behind
cement hydration. When the pozzolans start reacting, the particles are surrounded by
porous, but already stiff, structures. There being no evidence of a growing pressure, their
hydration products may precipitate in capillary pores. This precipitation is too small to
fill the larger pores, but is enough to obstruct the thin connections existing between the
large pores. As a consequence, the porosity o f pozzolanic cement is still higher whereas
permeability is reduced (Massazza, 1993).
The durability o f concrete depends on a number o f chemical and physical causes,
the most outstanding of which are (1) leaching, (2) chlorides, (3) sulfates, (4) reactive
aggregate, (5) carbonation, and (6 ) freeze/thaw. As far as items (1 )-(4) are concerned,
pozzolanic cements behave better than Portland cement, whereas with respect to items
(5)-(6) Portland and pozzolanic cements behave substantially the same (Massazza, 1993).
Leaching indicates the decomposition of the hydrated compounds by water. If the
concrete is porous and water is abundant, all the lime including that present in silicate and
aluminate will be leached. As leaching proceeds, concrete becomes more porous and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
permeable and then weaker. As compared to Portland cements, hardened pozzolanic
cement pastes are more leach-resistance due to low contents o f CH, more gel-like C-S-H
or C-A-H, and lower permeability.
Chloride ions can harmfully affect concrete and, more dangerously, steel
reinforcement.
When the Cl'/OH* ratio near the reinforcing steel increases over a
threshold limit o f 0.3, the passivation layer o f the steel is destroyed and then corrosion
becomes inevitable with the presence of water and oxygen (Diamond, 1986). Therefore,
less permeable concretes and adequate cover thickness are required to prevent this
corrosion. The positive effect of the pozzolans is suggested to be based mainly on their
capacity to bind chloride to the hydration product (or lower diffusion rate) and their
influence on the concrete permeability (Kouloumbi, 1994).
Sulfates are harmful to concrete because they can lead to concrete swelling and,
consequently, cracking. This sulfate attack causes formation o f expansive ettringite from
the reactions between sulfate and aluminate sources, and is initially by reaction between
sulfate ions and CH to form gypsum (Mindess and Young, 1981).
Initial Reaction:
CH + SO42' (aq)
=>
CSH2 (Gypsum) + 2 0 H ' (aq)
Secondary Reaction: C4ASH 12 + CSH 2 + 16 H
=>
C 6AS 3H 32 (Ettringite)
(Eq. 2-20)
(Eq. 2 -2 1 )
The sulfate attack depends on the permeability o f concrete and the diffusivity of
sulfate ions. Since the amounts o f CH and permeability o f pozzolanic concrete are lower
than normal concrete, the resistance to sulfate attack is believed to be improved by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
replacements o f pozzolans. Also, pozzolans (except fly ash) contain low C3A, which
enables minimization o f the risk o f sulfate expansion.
Some types o f silica (opal, chalcedony and tridymite) react with cement alkali to
form alkaline and alkaline-earth silicates which tend to absorb water and therefore to
swell (Massaza, 1993). The reductions in permeability and CH levels are closely related
to the decrease in the tendency of alkali-aggregate reaction in pozzolanig concretes
(Chatteiji, 1979).
Air contains CO 2 which reacts with hydrated cement in the presence o f water to
form carbonation. W ith respect to durability, the importance o f carbonation lies in the
fact that it reduces the pH o f the pore solution in hardened Portland cement paste. When
the low pH front reaches the vicinity o f the surface o f the reinforced steel, the passive
layer is removed and corrosion can take place (Massaza, 1993). The effects of pozzolans
on the carbonation are two-fold (Neville, 1996). Due to the small amount o f CH present
in the hydrated cement paste, CO 2 is not fixed near the surface o f the concrete so that
there is no pore-blocking formation o f calcium carbonate. This would tend to increase
the depth o f carbonation, but it is counteracted by the low permeability o f pozzolanic
concrete.
Due to the expansion o f water when it freezes, repeated freezing/thawing cycles
accelerate concrete decay. A suitable protection against frost can easily be obtained by
introducing microscopic air bubbles to allow water movement by the ice formation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
Having denser microstructures, pozzolans may reduce the resistance to freezing/thawing.
However, the results are inconsistent (Neville, 1996).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 3
Experimental Procedure
3.1. Apparatus
3.1.1. Microwave Oven
A highly overmoded cylindrical microwave oven (MMT 101, Oak Ridge, TN: 750
mm in diameter and 1200 mm long) with a power range of 0 to 6 kW was used for the
curing process o f cement pastes and mortars. Temperature o f specimens was measured
by a shielded type-K thermocouple. The thermocouple output was fed to a controller
which regulated the power input to the oven.
3.1.2. Instrumented Penetration Test
The in situ hardness development of cement paste or mortar was measured at
room temperature and with microwave heating, using an alumina flat-tip penetrator o f a
diameter o f 2.8 mm and a length o f about 360 mm.
A schematic diagram o f the
instrumented penetration test is shown in Figure 3-1. A polyethylene specimen container
was inserted into styrofoam at each end and supported by an alumina tube. The load cell
(Omega LCC A-25, Stamford, CT: 25 lbf max.) on the end of the penetrator measured the
applied load and the load data was recorded by a personal computer via an analog-digital
converter (Omega D 1521, Stamford, CT). The drive mechanism consisted of an electric
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
-Supporter
Thermocouple
(K-Type)
Microwave Chamber
(MMT101)
PC
Load cell
(Omega LCCA-25)
A/D Converter
(Omega D 1521)
Samole
Motor & Gear
reduction system
Styrofoam
Microwave
Generator
(Cober S6F)
Drive
Mechanism
Controller
(Micristar 828E)
Alumina Penetrator
Figure 3-1 Schematic diagram o f the instrumented penetration test unit and microwave
oven.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
motor controlled by a rheostat. A variable gear reduction system stepped down the motor
rotation. A constant motor output o f 561 rpm (±5) was reduced to 0.34 rpm at the
threaded shaft with the gear reduction system. A gear reduction setting o f 100:1, 50:1, or
10:1 resulted in a penetrator speed o f 6.4, 12.8, or 64 mm/hr, respectively. The testing
was performed 30 minutes after the initial contact o f cement and water to the time at
which the penetration resistance reached about 20 Ibf (89.0 N) for all penetration tests.
In order to investigate temperature rise and its effects on the hydration of cement
pastes with or without insulation, the apparatus was modified to satisfy the individual
experiment purpose.
When the insulation was needed, the sample placed in the
microwave oven in Figure 3-1 was covered by styrofoam insulation.
3.2. Sample Preparation
3.2.1. Materials
Two commercially available Type I Portland cements produced by Holnam
Cement Inc. (Clarksville, MO) and La Farge Co. (Milwaukee, WI) and ASTM C778
graded silica sand (AGSCO, EL) were used for preparing cement paste and mortar
samples. La Farge cement was used only for studying the hardening process of cement
paste and mortar, while Holnam cement was used for all research.
Exploring the
possibility o f hardening process alternation and property improvements, appropriate
amounts o f pozzolanic materials such as ground granulated blast furnace slag, silica fume
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
and Class F fly ash were substituted for a portion o f the Portland cement. The chemical
compositions o f these materials except silica fume are shown in Table 3-1.
The
compositions o f silica fume slurry were 50 wt% water and about 50 wt% SiC>2 with some
carbon. The Blaine fineness o f silica fume was about 200,000 cm 2/g.
Table 3-1 Chemical compositions and Blaine fineness of cement-based materials.
Component
CaO
Si0 2
A120 3
FeO (Fe20 3)
MgO
Alkalies (Na20 or K20 )
S03
L.O.I.
Blaine fineness (cm2/g)
Type I
(La Farge)
64
Type I
(Holnam)
65
21
20
39
38
5.4
2.3
3.7
0.5
4.6
8
2.8
0.5
1.9
0.5
2.5
1.5
3800
11
20
1.2
0.7
N/A
0.4
5600
3.7
2.9
2.9
2890
2.8
1.1
3680
GGBFS
Fly Ash
(class F)
1.2
47
23
3.2.2. Experimental Designs
In exploring the initial setting and the hardening rate, the experimental set was
determined by D-optimal experimental design software (EDO, Harold S. Haller &
Company, 1992). It enables a reduction in the total experimental sets with a reasonable
statistical basis (See Appendix C .l). The experimental results were analyzed by Multiple
Correlation (MC, Harold S. Haller & Company, 1992) to establish relationships between
variables. Statistical terminology for MC is shown in Appendix C.2.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
The list o f Variables and their levels are shown in Table 3-2. Note that some
variables o f this list were chosen according to the purpose of each experiment.
Table 3-2 Variables and their levels for studying initial setting time and hardening rate.
Variable Names
1/Temperature (1/K)
w/c ratio
Gear ratio
s/c ratio
Silica fume amount (wt%)
GGBFS amount (wt%)
Class F Fly ash (wt%)
Levels
1/303, 1/318, 1/333
0.3, 0.35, 0.4
10, 50, 100
0,
1,
2
0, 5, 10
0, 25, 50
0 , 10 , 20
The w/c ratio indicates water-to-cement ratio.
When reactive admixtures
(pozzolans) were substituted for cement, water-to-total reactive solid (w/r) ratio was used
instead o f w/c ratio. The s/c ratio indicates sand-to-cement ratio. Note that the gear
ratios are inversely proportional to penetration rate: gear ratios o f 10, 50 and 100 were
equivalent to penetration rates o f 64, 12.8 and 6.4 mm/hr, respectively.
Table 3-3 shows the D-optimal design of cement paste, which consists of three
variables, w/c ratio, temperature and gear ratio. Table 3-4 shows the design of cement
mortar whose variables are s/c ratio, temperature and gear ratio. The w/c ratio for the
mortar study was fixed at 0.4. Tables 3-5 and 3-6 show the designs of cement paste
containing reactive admixtures, which have an additional variable, namely admixture
content. Note that one experiment, 30°C curing temperature and 64 mm/hr, was restricted
from the design because it would require a longer specimen than our setup permitted.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
Table 3-3 D-optimal design for cement paste.
# o f Experiment
1
2
3
4
5
6
7
8
9
10
11
12
w/c ratio
0.4
0.4
0.3
0.3
0.35
0.4
0.35
0.4
0.4
0.3
0.3
0.3
T emperature(K)
333 (60°C)
303 (30°C)
3 0 3 (30°C)
333 (60°C)
303 (30°C)
3 3 3 (60°C)
3 1 8 (45°C)
303 (30°C)
318 (45°C)
318 (45°C)
3 3 3 (60°C)
3 1 8 (45°C)
Gear Ratio (Penetration Rate: mm/hr)
100 (6.4)
100 (6.4)
100 (6.4)
100 (6.4)
50(12.8)
10 (64)
100 (6.4)
50 (12.8)
10 (64)
10(64)
10 (64)
50 (12.8)
Table 3-4 D-optimal design for cement mortar.
# o f Experiment
s/c ratio
1
2
2
2
2
0
2
1
1
0
0
0
1
0
3
4
5
6
7
8
9
10
11
12
Temperature (K)
3 3 3 (60°C)
3 3 3 (60°C)
303 (30°C)
3 3 3 (60°C)
3 1 8 (45°C)
3 3 3 (60°C)
318 (45°C)
318 (45°C)
303 (30°C)
333 (60°C)
318 (45°C)
3 0 3 (30°C)
Gear Ratio (Penetration Rate: mm/hr)
100 (6.4)
10 (64)
100 (6.4)
100 (6.4)
50 (12.8)
50 (12.8)
100 (6.4)
10(64)
50 (12.8)
10(64)
10 (64)
100 (6.4)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
Table 3-5 D-optimal design for cement paste containing GGBFS.
#
w/r ratio
1
2
0.35
0.3
0.35
0.4
0.4
0.3
0.4
0.35
0.3
0.35
0.3
3
4
5
6
7
8
9
10
11
GGBFS
(wt%)
25
0
50
50
25
25
0
50
50
0
0
t emperature
(K)
303 (30°C)
303 (30°C)
3 3 3 (60°C)
303 (30°C)
318 (45°C)
333 (60°C)
333 (60°C)
303 (30°C)
318 (45°C)
318 (45°C)
318 (45°C)
Gear Ratio
(Penetration Rate: mm/hr)
50 (12.8)
100 (6.4)
50 (12.8)
50 (12.8)
100 (6.4)
10 (64)
10 (64)
100 (6.4)
10 (64)
10 (64)
50 (12.8)
Table 3-6 D-optimal design for cement paste containing silica fume or fly ash.
#
1
2
3
4
5
6
7
8
9
10
11
w/r ratio
0.35
0.4
0.35
0.4
0.3
0.35
0.35
0.3
0.4
0.3
0.3
Silica Fume; Fly Ash Temperature (K)
(wt%)
5: 10
318 (45°C)
333 (60°C)
10 ; 20
10 ; 20
3 3 3 (60°C)
3 1 8 (45°C)
10 ; 20
3 3 3 (60°C)
0; 0
333 (60°C)
5; 10
0
303 (30°C)
0;
3 0 3 (30°C)
5; 10
303 (30°C)
0; 0
318 (45°C)
0; 0
10 ; 20
333 (30°C)
Gear Ratio
(Penetration Rate: mm/hr)
10 (64)
50 (12.8)
10 (64)
10 (64)
10 (64)
100 (6.4)
50 (12.8)
50 (12.8)
100 (6.4)
100 (6.4)
100 (6.4)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
3.2.3. Mixing and Curing
Materials were mixed with deionized water using a paddle mixer (kitchenaid,
model K5ss) for 10 minutes at speed level 1. Pastes were cast into 5 cm in diameter and
10 cm in length polyethylene cylinders, vibrated to eliminate air bubbles, and then capped
with polyethylene covers.
The application o f microwave energy began 30 minutes after initial water contact
and then was varied in accordance with the setpoint temperature, which was raised at a
constant heating rate o f 2°C/min to each preset curing temperature and maintained at that
temperature for the preset time.
The setpoint temperatures were varied from room
temperature (or 20°C) to 90°C. Room temperature means no microwave application on
the specimen. The length o f microwave heating was varied according to the individual
experimental purposes.
Especially for chemical analysis, specimens were naturally cooled at 6 hours from
the initial mixing, and then were sealed with duct tape and stored in an 100 % humidity
chamber at room temperature to reduce water evaporation. They were removed later to
obtain aqueous fluid for the pore solution analysis and crushed powder for loss-onignition and x-ray experiments.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
3.2.4. Sample Curing for Strength Test
28-day compressive strength tests were performed with blended cement mortar as
well as cement mortar.
For producing blended mortar, 50% in weight GGBFS, 10%
silica fume, or 20% class F fly ash was substituted. The total reactive solid/sand ratio and
water-to-total reactive solid ratio for all mortars were fixed at 1:2 and 0.4, respectively.
Figure 3-2 (a) shows a schematic diagram o f the microwave oven for curing o f 28day compressive strength specimens. Temperature was measured by a thermocouple
inserted directly into the specimen. The control thermocouple was placed 22 mm from
the bottom o f the specimen, out of the portion to be used in strength testing. Additional
thermocouples were placed in some specimens at 37 and 67 mm from the specimen
bottom to determine temperature uniformity during heating.
The application of
microwave energy began 30 minutes after the water was added and then temperature was
raised at a constant heating rate of 7°C/min to each preset curing temperature and
maintained at that temperature.
After curing, the specimen was stored in 100% RH
containers for 1 day, then removed from the mold and immersed in lime saturated water
at 20°C for 27 more days until testing at 28 days. Both room temperature-cured and
microwave-cured specimens were cut at 25mm from both ends and ground before testing.
The final dimension of the strength test specimen was 50 mm x 50 mm, as shown in
Figure 3-2 (b) (gray region).
The compressive strength test of 28-day cured mortars was performed on an MTS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
Microwave Chamber
(MMT 101)
Styrofoam
Thermocouple
Thermocouple
Controller
Microwave
Generator
(Micristar 828 E)
(a)
50 mm
(b)
Figure 3-2 (a) Schematic diagram o f the microwave oven for curing mortar for strength
test (not to scale).
(b) Dimensions o f the mortar specimen. The gray region was used for
strength tests.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
million pound machine (Material Testing System, Minneapolis, MN). Load was applied
with a constant stroke rate of 1.02 x 10'3 mm/sec until reaching the maximum strength.
3.3. Characterization Methods
3.3.1. Vicat Needle Test
To estimate setting times o f cement pastes, the Vicat needle test was employed.
The schematic diagram of the Vicat test was already shown in Figure 2-2. According to
ASTM C191, a cement paste specimen was placed in a plastic mold with a moist
environment. The test began at 1 hour after mixing by using a 1-mm needle and every 15
minutes thereafter until a penetration o f 25mm or less obtained. For the test, the needle
was lowered until it rested on the surface o f the cement paste. After setting the indicator,
the rod was released quickly by releasing the set screw, and the needle was allowed to
settle for 30 seconds.
No penetration test was made closer than 1/4 inch from any
previous penetration and no penetration test was made closer than 3/8 inch from the
inside o f the mold. The time when a penetration o f 25mm was obtained was taken as the
initial setting time.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
3.3.2. Rheometer
The rheological measurements utilized a computer-operated constant strain rate
rheometer (Bohlin variable oscillatory rheometer: VOR, Bohlin Reologi AB, Sjobo,
Sweden). This work has done in collaboration with Professor L.J. Struble and Mr. G.K.
Sun at the University o f Illinois at Urbana-Champaign. All measurements were made
using Couette (cup and bob) geometry with narrow gap and smooth surface. The outer
cylinder (or cup) o f the rheometer is made to rotate, and the torque is measured on the
inner cylinder (or bob). The radius of the inner cylinder was 7.0 mm and the gap was 0.7
mm. The measurements were carried out in the oscillatory mode with 1 Hz frequency.
Data were collected at 3-minute intervals.
Throughout Theological testing, the
temperature was maintained at 20°C by means of a circulating water bath. Cement paste
was mixed with deionized water by a paddle mixer for 10 minutes and then only a small
portion o f it was transferred into the rheometer. All tests were begun 15 minutes after
cement and water were first mixed.
3.3.3. Loss on Ignition
The degree o f hydration of cement pastes can be calculated by weight changes
under ignition, or loss-on-ignition. The experimental procedure o f the loss-on-ignition
test is as follows: ground pastes are weighed after 1 day in an oven at 105°C, and then
heated to 1000°C for 3 hours. A final weight is obtained when the pastes return to 105°C
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
for 1 day. D-drying (or oven drying at 105°C) removes evaporable water in both capillary
and gel pores.
Nonevaporable water is subsequently lost when a paste is ignited to
1000°C. Nonevaporable water measures the amount o f water combined structurally in the
hydration products.
The nonevaporable water (w„) is proportional to the amount of
hydration (Mindess and Young, 1981);
w
Wt
n
a ‘=^
=
_ -W t
105°C
024 x Wt
1000°C
1000"C
„
I000°C
.
..
p”
(Eq' J - ‘)
where a p is uncorrected degree of hydration o f paste and etc is degree of hydration o f dry
cement. When the cement is fully hydrated (a = 1), 0.24 g o f nonevaporable water are
combined with each gram o f cement.
Using the assumption of the Powers-Brownyard model (Taylor, 1990), the
microstructure of cement pastes consists o f unhydrated cement, hydration products and
capillary porosity.
The ratio of hydration product to unhydrated cement is given by
(Christensen, 1993):
g = volume o f hydration product = 2 J ± Q 1
volume of unhydrated cement
and the initial fraction of solid is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
/ = T - ------- ~
l r -p r
^ P cement
(Hq. 3-3)
^ *0
Therefore, the volume fractions o f product and unhydrated cement are
^cement —(1 ~&)fj
^product = S C tf
(Eq. J-4)
and the volume fractions of capillary porosity and gel porosity are given by.
^capillary
1 “^product' ^cement—1—”
_ _, ' T
I + 32 (w / c)
d> = ---- ----------gel0313 + w / c
(Eq. 3-5)
(Eq. 3-6)
4
where S = 2 3 and pcement= 3.2 g/cm 3 as general values of cement.
3.3.4. Impedance Spectroscopy
Impedance response was obtained using an HP 4192A low frequency impedance
analyzer (Hewlett-Packard, Santa Clara, CA) interfaced to a PC via an IEEE-488
interface.
Frequencies from 11 MHz to 5 Hz were swept in a logarithmic manner,
collecting 20 data points per decade o f frequency. After being corrected to m inim ize
inductance effects at high frequencies, impedance spectra were analyzed using the
EQUIVCRT.PAS program (Boukamp, 1988).
elsewhere (Christensen, et al., 1994).
The correcting procedure is described
When the corrected data were loaded into the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
program, the low frequency arc, representing electrode response, was removed and the
partial non-linear least squares (NLLS) routine was used over the frequency range o f 10
kHz to 3.55 MHz to obtain the resistance o f the specimen for all cases. The resistance
was converted to conductivity using appropriate geometrical factors.
Details o f
application o f impedance spectra to cement-based materials are given in Appendix C.
Molds for impedance spectroscopy measurement had stainless steel foil at both
ends as electrodes, shown in Figure 3-3. One specimen was used to control microwave
power according to its temperature and the other was used to obtain impedance spectra.
The temperature difference between the two samples was less than 4°C. To minimize the
temperature difference caused by microwave power disturbance, two electrodes are
inserted in both samples.
3.3.5. Pore Solution Analysis
Since the aqueous phase inside o f cement pastes plays a significant role in
hydration, pore solution analysis is helpful to investigate the mechanism of cement
hydration under elevated temperature curing conditions.
Generally, the pore solution
from cement pastes has soluble ions such as Na+, K.+, Ca++, Si, A1 and SO 4 - as well as
high OH' ion concentration (pH =1 3 ) with various concentrations, depending upon the
degree of hydration. To investigate o f the effects o f microwave-enhanced curing on the
aqueous phase o f pastes, pore fluid was extracted from cured specimens by means o f a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
Microwave Chamber (MMT 101)
______________
Electrode
Styrofoam
Impedance
Analyzer
M icrow ave
G enerator
(HP 4192 A)
(Cober S6F)
Controller
(Micristar 828E)
Figure 3-3 Schematic diagram o f modifying microwave oven for impedance
spectroscopy measurements.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
high pressure steel die.
Details o f the construction o f the die are given elsewhere
(Bameyback, and Diamond, 1981). Samples were demolded and crushed in the die. Pore
fluid was captured in 10 ml syringes and then sealed until testing. The conductivity o f the
pore fluid was obtained using the same impedance analyzer. The pore fluid was placed in
a 5 mm X 50 mm plastic tube with an electrode at each end and connected to the
analyzer. The intercept o f the impedance response with the real axis was termed the
resistance o f the pore fluid.
The ionic concentration o f aqueous solutions was determined via ICP
(inductively-coupled plasma)
method
in
a
Perkin-Elmer Plasma 40
Emission
Spectrometer (Perkin-Elmer, Eden Prairie, MN). A sample solution is introduced into a
pump and injected as a fine aerosol into an argon-supported, inductively-coupled plasma,
which ionized the sample. The ionic concentration can then be determined by comparing
the emission wavelengths and intensities with those of standards. A filtered (0.1 pm)
aqueous solution was used for ICP analysis. The pH value of each pore solution was also
measured by a pH-meter.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 4
Hardness Development of Cement-Based Materials
4.1. Microwave Heating Effects on Cement Paste
4.1.1. Temperature Changes
Figure 4-1 shows temperature rise o f cement pastes cured at room temperature
under adiabatic conditions (or covered by styrofoam insulation) with various w/c ratios.
The size o f the specimen was 5 cm in diameter and 10 cm in length, and the
thermocouple was placed at the center of the specimen. Although there is no external
heat source, due to the exothermic reactions o f cement paste, the specimen temperature
increased to about 90°C, nearly the boiling point of water in the case o f a w/c ratio o f 0.3.
The cement paste shows faster hydration reactions and greater temperature rise as the w/c
ratio is reduced. Since the main exothermic reactions begin when dissolved ions reach
threshold concentration, cement pastes with lower w/c ratios heat up faster because the
threshold concentrations are achieved sooner. Also, more water is believed to absorb
hydration heat, thereby reducing the temperature rise. Therefore, reactions begin faster
with a rapid rise in temperature in pastes with lower w/c ratios.
When specimens were heated by an external heating source, the temperature
profiles were changed. Specimens were covered by styrofoam insulation and heated by
microwave energy until reaching various set point temperatures.
The temperature
profiles o f cement pastes o f various w/c ratios and setpoint temperatures are shown in
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
100
w/c=0.3
w /c=0.4
80-
w/c=0.5
Temperature (°C)
w /c=0.7
60-
20
-
0
5
10
15
20
Time after Mixing (Hours)
Figure 4-1 Temperature profile o f cement pastes cured at room temperature under
adiabatic conditions with various w/c ratios.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
Figure 4-2. Because of heating, all cement pastes reacted faster, indicated by a decrease
o f the plateau interval around each setpoint temperature. When paste was initially heated
to 40°C by the microwave, it was heated later by its reactions to about 90°C for both 0.3
and 0.4 w/c ratios and to 60°C for 0.5 after the plateau intervals, which are shorter than
the intervals at room temperature. At 60°C, the rate o f temperature rise became greater
and pastes were heated beyond 100°C, while at 80°C, no further enhancement was found.
Therefore, above 60°C, the chemical reaction mechanism could be different or not
affected by temperature anymore.
This will be discussed with activation energy in
section 4.1.4.
To study the effects o f curing temperature on cement-related work, specimen
temperature should be controlled accurately at the setpoint temperature. Without careful
consideration o f real specimen temperature, it is hard to estimate temperature effects on
cement properties, because the real specimen temperature could be much higher than the
chamber temperature, which could be regarded as the temperature variable.
When an uninsulated specimen was heated by microwave energy with feedback
control, the real specimen temperature was nicely maintained at the set point temperature,
as shown in Figure 4-3. Additionally, temperature was uniform throughout the specimen.
The 2.45 GHz microwave used can penetrate through the specimen which had a 2 cm
radius.
Since the specimen is composed of water and an insulator (cement), the
penetration depth o f 2.45 GHz microwave would be greater than the 1.3 cm depth value
for pure water. Therefore, all specimens afterward were prepared without insulation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
120
Temperature (°C)
RT
100
40°C
-
60°C
80“
80°C
60-
20
“
(a) w /c = 0.3
0
5
10
15
Time after Mixing (Hours)
120
Temperature (°C)
RT
40°C
60°C
80°C
40-
20(b) w /c = 0.4
0
5
10
15
Time after Mixing (Hours)
Figure 4-2 Temperature profile of cement pastes at RT, 40°C, 60°C, and 80°C by using
microwave energy with styrofoam insulation at (a) w/c = 0.3 (b) w/c = 0.4 (c)
w/c = 0.5. Microwaves turned off as soon as temperature reached target
value.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
120
RT
100
_
40°C
-
u
V
3
60°C
80-
(-1
60<L>
O,
B
a>
H
40-
20
-
(c) w /c = 0.5
0
5
10
Time after Mixing (Hours)
Figure 4-2 Continued.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
74
100
80°C
80 -
U
60°C
o
r
t-i
6° -
3
j—
<u
Q,
£
<L>
H
40°C
40 -
20
-
0
1
2
3
4
6
Time after Mixing (Hours)
Figure 4-3 Temperature profile o f cement pastes o f w/c = 0.4 at 40°C, 60°C, and 80°C
by using microwave energy without insulation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
4.1.2. Hardness Development of Cement Paste
As shown in Figure 3-1, the penetrator was driven into each specimen as paste
hydration proceeds. The reaction force against the penetrator was detected by the load
cell at the end o f the penetrator. For this study, the penetration hardness is defined as
follows:
Force
Penetration Hardness (H) = —------ —------- — ----------Frontal Area o f Penetrator
(Eq. 4-1)
This definition is similar to the penetration resistance o f ASTM C 403.
Before performing the penetration test, it was necessary to measure any frictional
effects between the sample and the penetrator. As shown in Figure 4-4, the penetration
o f wet silica (silica particle size = 14 micron) showed no frictional effects (Croft, 1996).
If any significant friction existed, the response should increase as the penetrator proceeds
because the friction would be linearly dependent on contact area. Also, the rate of 2200
mm/hr is much greater than that chosen for the main study, thus the frictional force can
be ignored. Note also, the penetration hardness increased with the penetration rate and
solid amount.
Since fine silica powder does not react with water, it is seen that the
penetration rate as well as the water-to-solid ratio is able to influence the penetration
hardness.
Figure 4-5 shows the penetration hardness development for cement pastes of w/c
ratio o f 0.3 cured at various temperatures. The hardness level does not indicate the real
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
0.1
0.09
Penetration Hardness (MPa)
w/s = 0.27, Rate = 2200 mm/hr
0.08
0.07
0.06
0.05
0.04
w/s = 0.3, Rate = 2200 mm/hr
0.03
0.02
w/s = 0.3, Rate = 150 mm/hr
0.01
0
t
0
20
1-- - - - - - - - 1- - - - - - - - - 1-- - - - - - - - 1- - - - - - - - - 1-- - - - - - - - - r
40
60
80
100
Penetrated Distance (mm)
Figure 4-4 Penetration hardness versus penetrated distance for wet silica powder with
various water-to-solid ratio and the penetration rate (Croft, 1996).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
RT
35°C
Penetration Hardness (MPa)
40°C
50°C
15 -
60°C
10
"
70°C
x
80°C
xl-
-
0
60
120
180
240
300
Time after M ixing (M inutes)
Figure 4-5 Penetration hardness development for cement pastes with w/c = 0.3 cured at
various temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
strength o f the specimen, but is generally much higher than its real strength (Croft, 1996).
The penetration hardness begins to develop after some time, which may closely
correspond to the beginning of the acceleration period in the cement-water reaction.
Higher curing temperatures are associated with shorter starting times and faster hardness
development.
Above 60°C curing temperature, no further significant change in the
response was found, which were also found at 0.25 and 0.4 w/c ratio. This trend is very
similar to the observations o f the internal temperature rises in section 4.1.1. It is thought
that the hydration mechanism above 60°C could be different from the lower temperature
ranges, independent o f the w/c ratio, thus it may be less dependent on the curing
temperature.
4.1.3. Setting Times o f Cement Paste
Initial and final setting time were defined at the penetration hardness o f 3.5 MPa
and 27.6 MPa, respectively, according to the definition o f the setting times on the ASTM
C403. To verify the setting times, the conventional Vicat needle test was also performed.
Especially, the final setting time was calculated from the initial setting time obtained by
the Vicat test by the following equation (Neville, 1996):
[Final setting time (in minute)] = 90 + 1.2 x [Initial setting time (in minute)]
(Eq. 4-2)
The setting times determined by both methods are in good agreement, as shown in
Figure 4-6. Therefore, the instrumented penetration test enables one to measure the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
500
RT Curing
t/3
400 -
<D
3
C
00
300 -
c
'*
i-l
<U
<ts
cd
<u
B
H
200
-
100
-
0
Initial Setting time by V icat Test
-■ - Initial Setting tim e by penetration (at 3.5MPa)
"0— Final setting tim e by F (min)=90-H .2 x I (min)
- s - Final setting time by penetration (at 27.6M Pa)
T
0.25
0.3
T
T
0.35
0.4
w /c ratio
Figure 4-6 Setting time change for cement pastes with w/c ratio in comparison of the
initial and final setting times measured both by the conventional Vicat and
the instrumented penetration test at room temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
setting times for cement pastes. Since the setting time determination by the instrumented
penetration test is similar to the standard test for the setting time o f concrete, it may also
be an alternative method for estimating the setting times for mortar and concrete.
The initial setting time was statistically analyzed by Multiple Correlation analysis
(MC, Harold S. Haller & Company, 1992). MC takes the independent variables that most
affect the dependent variable and computes t-values according to their relative
importance. The greater the effect that an independent variable has the higher its t-value.
However, since MC analysis considers variables only in view of statistics, the variables
in the developed equation should be confirmed by physical sense. The equation of the
initial setting time of cement paste is as follows (see Appendix B for explanation of
terms):
Time (min) = 105 + (21 ± 2)(w/c) + (38 ± 2)(1000AT)
+ (16 ± 4)(1000/T)2 + (11 ± 2)(w/c)(1000/T)
(Eq. 4-3)
(R2 = 0.9939; Modified R2 = 0.9964; D.O.F = 3; Sy.x = 3.9628; STest = 2.5538)
All independent variables were coded as +1 for w/c = 0.4, 1/T = 1/303 and -1 for w/c =
0.3, 1/T = 1/333. The 95% confidence interval o f each coefficient is also given. D.O.F.
means the degree of freedom o f this model. Since it was found that gear ratio does not
affect the initial setting time, some experiments in Table 3-3 become redundant. The
estimate of Stest given above was computed from these pseudo-replicates.
Figure 4-7 shows the predicted response surface o f the initial setting time with
various curing temperatures and w/c ratios. It was found that the initial setting time
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
Temperature
□ 180-210
Bl 150-180
120-150
El 90-120
60-90
0.30 031 032 033 034 035 036 037 038 039 0.40
w/c Ratio
Figure 4-7 Initial setting time response surface and contour plot for cement pastes (Holnam
Type I) with various w/c ratios and curing temperatures measured by the
instrumented penetration test.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
becomes longer at high water content and low temperature. The effects of water content
and temperature on the initial setting time correspond to those effects on the heat
evolution (Figure 4-2). Perhaps, the ions are concentrated by the temperature rise and the
lower amount o f water, to reach the threshold values to end the induction period and
begin the setting process sooner.
Eq. 4-3 shows an interaction between w/c ratio and 1/T. An interaction exists
between two variables if the response o f one o f them depends upon the value of the other.
To understand the effects of the interaction, the initial setting time equation was
differentiated with respect to each variable involved. Note all variables are coded.
w/c ratio:
d (Time)
= 2 6 + 14 (1000 / T)
d (Time)
Temperature: ^ q q q / ^ = 39 +16 (1 0 0 0 /T) + 14
(Eq. 4-4)
(
w
/
c
)
(Eq. 4-5)
Figure 4-8 shows the effect of the interaction. It can be seen from this figure and
the equations that the w/c ratio effect is less at higher temperature, while the temperature
effect is less at lower w/c ratio.
4.1.4. Activation Energy
The activation energy of hydration for cement paste was calculated according to
the following steps. First, the hardness development plot like Figure 4-5 was converted
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Initial Setting Time (Minutes)
240
30°C
180 -
120
-
60°C
60 ■
0.3
0.35
0.4
Initial Setting Time (Minutes)
w/c ratio
240
w/c = 0.4
180
120
-
w /c = 0.3
60 -
30
40
50
60
Temperature (°C)
Figure 4-8 Initial setting time for cement pastes, representing the effect of the
interaction between w/c ratio and temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
to a semi logarithm plot, for instance Figure 4-9. Hardness values below 0.4 MPa were
neglected. Next, the hardness was assumed to be related only to the degree o f hydration
and independent o f temperature at a given amount o f hydration. Then, the penetration
hardness development can be represented as a function of time as follows:
In H = A + Bt + Ct2
(Eq. 4-6)
where H is the penetration hardness, t is the time after mixing and A, B, and C are
constants. Because o f the curvature o f each hardness line in Figure 4-9, it is reasonable
to consider the 2nd order time term.
The hardening rate can be represented:
(Eq* 4-7)
where (d In HIdt) is the slope of the semi logarithm plots o f hardness versus time.
If the microstructure (or degree o f hydration) at a given degree of hardness is
independent o f temperature, the apparent activation energy can be estimated. Arbitrarily
choosing 5 MPa for H, the hardening rates at various w/c ratios and curing temperatures
are shown on an Arrhenius plot in Figure 4-10. Below 60°C, indicated as region I, the
hardening rates o f all w/c cases are approximately linearly dependent on 1000/T with
negative coefficients. These coefficients are directly related to the activation energy of
the hardening. However, above 60°C, region II, the hardening rates o f all w/c data
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
100 i
RT
35°C
Penetration Hardness (MPa)
40°C
50°C
60°C
0
+
70°C
*
80°C
60
120
180
240
300
Time after Mixing (Minutes)
Figure 4-9 Semi-logarithm plot o f penetration hardness for cement pastes with w/c = 0.3
cured at various temperatures by microwaves, which corresponds to Figure
4-5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
1
dH/dt
At 5 MPa
o.i
2.7
•
w/c=0.25
□
w/c=0.3
±
w/c=0.4
2.8
2.9
3
3.1
3.2
3.3
3.4
1000/T
Figure 4-10 Arrhenius plot o f the rate o f hardening at 5 MPa o f the penetration hardness
measured by the instrumented penetration test for cement pastes with
various w/c ratios cured by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
deviated from the line made in region I to converge to a single point at 80°C. Ishida and
coworkers found very similar results from fine P-C2S work, that all data points in the
range from 20°C to 60°C are perfectly matched with lines for estimating activation
energy of each degree o f hydration on the Arrhenius plot, but at 80°C, the points deviated
(Ishida, et al., 1992).
The difference in behavior at 80°C was also found in the
temperature rise in section 4.1.1. It is thought to be caused by a change in reaction
chemistry at the increased temperature. Specifically, ettringite cannot be formed at 80°C
(Orchard and Barnett, 1971).
The apparent activation energy of the hardening rate o f cement paste was
calculated from this Arrhenius plot within the temperature range from room temperature
to 60°C and is shown in Table 4-1.
Table 4-1 Apparent activation energy of the hardening rate for cement pastes
with various w/c ratios.
w/c ratio
0.25
0.3
0.4
Activation energy (KJ/mol)
26.8
31.1
35.2
The estimated activation energy of the hardening rate varied with water content.
However, it may be unreasonable that the activation energy o f cement hydration is
changed by the water content, even though the hardening rate is associated with the rate
of hydration.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
The activation energy o f hydration was obtained by the loss-on-ignition
technique, and is shown in Figure 4-11, superimposed on the activation energy results
calculated from the penetration hardness. Details on the degree o f hydration will follow
in Figure 5-1. The degree o f hydration o f cement paste with a w/c ratio o f 0.4 is roughly
0.07 when the penetration hardness reaches 5 MPa. These activation energies are in the
similar range o f about 33-34 kJ/mol.
According to Fujii and Kondo’s assumption,
cement hydration is divided into two stages with differing activation energies, nucleationgrowth and diffusion (Fujii and Kondo, 1974).
Considering the degree o f hydration
(Figure 5-1) and temperature rise (Figure 4-2), the boundary o f these two stages could be
roughly at a degree of hydration o f 0.4. So, up to about a 0.4 degree o f hydration, the
activation energy should be relatively constant. Thus, if the activation energy at 0.1
degree o f hydration were precisely calculated by degree of hydration data, it should be
close to that o f penetration hardness. Additionally, beyond 0.5 o f degree o f hydration,
the calculated activation energy may represent the diffusion mechanism o f cement
hydration. From Figure 4-11, the activation energy o f diffusion is about 21.7 kJ/mol.
Interestingly, Fujii and Kondo reported similar values for the activation energy o f
C 3S hydration, 31.4 kJ/mol for nucleation-growth stage and 20.9 for diffusion stage (Fujii
and Kondo, 1974). Also, Ishida calculated the activation energy o f P-C2S hydration as
about 57 kJ/mol when at 0.5 water-to-solid ratio. Since the hydration o f P-C 2S is slower
than C 3S, it is reasonable to need more energy for hydration of P-C2S than for cement.
He could not find the change o f hydration mechanism corresponding to the progress in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
Degree o f Hydration (LOI)
Penetration Hardness
40 -
30 -
20
~
10
-
0
0.1
0.2
0.3
0.4
0.5
0.6
a
Figure 4-11 Apparent activation energy obtained by penetration hardness and degree of
hydration for cement paste with w/c = 0.4.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
hydration (Ishida, 1992).
However, estimating the activation energy by the assumption that the rate of
hardening is linearly dependent on inverse temperature is somewhat risky, because the
hardening rate may not be a simple function of temperature. Figure 4-12 shows that a
second order polynomial fits the Arrhenius plot better than a straight line.
Possible
reasons for the complicated temperature dependence o f the hardening rate may be the
multiplicity o f cement hydration reaction and possible temperature effects on rheology.
Actually, the response o f the penetration hardness may be influenced by the
temperature effect on rheological behavior as well as cement hydration. This temperature
effect, discussed further in section 4.2.3, would give an apparent activation energy of
hardening that is less than that for hydration.
4.2. Hardening Rate of Cement Paste
4.2.1. Influence o f the Penetration Rate on Penetration Hardness
As mentioned earlier, temperature and water amount influence the hydration
mechanism o f cement paste, resulting in alteration of the hardening rate. However, the
penetration rate (or speed o f penetration) is expected to have an affect on the apparent
hardening rate. From Figure 4-4, the penetration hardness of wet silica powder increased
with only changing the penetration rate. Thus, the investigation o f the penetration rate
effect is required to establish an overall scientific basis for utilizing the instrumented
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
At 5 MPa
-0.5 -
"O
52
*o
'c5
-2
-
w/c=0.25
Q— w/c=0.3
-*— w /c=0.4
-2.5
2.9
3
3.1
3.2
3.3
3.4
1000/T
Figure 4-12 Second order polynomial curve fitting from room temperature to 60°C at 5
MPa o f the penetration hardness.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
penetration apparatus.
Figure 4-13 shows a typical example o f the influence o f the penetration rate on
the measured penetration hardness.
The faster the penetration rate, the greater the
hardening rate. This effect becomes greater as cement hydration proceeds, represented as
a higher penetration hardness level, shown in Figure 4-14. This behavior was found at all
curing temperatures. Therefore, the penetration rate is one o f the variables for studying
hardening rate. The penetration rate effects on cement rheology will be discussed in
section 4.5.3.
4.2.2. Hardening Rate Equation of Cement Paste
To establish a hardening rate equation, two Type I Portland cements
(manufactured by La Farge and Holnam) were investigated.
Figure 4-15 shows that
HoInam-manufactured cement pastes hydrated faster than La Farge-manufactured cement
pastes at both 30 and 60°C because o f differences in their starting materials.
The dependent variable for MC analysis was the logarithm o f the hardening rate
(Ln(Hardening rate)).
The additional independent variable was penetration hardness,
whose levels are 1, 5, 9, and 13 MPa.
The response surface coefficients and the
corresponding t-values of Ln(Hardening rate) for the two cement pastes are listed in
Table 4-2. The coefficients were based on coded values o f each variable, listed in Table
4-3.
The equations o f both cements greatly resembled each other with respect to
significant variables and their signs and values, although the development o f hardness is
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
6.4 m m /hr
-
64 m m /hr
128 m m /hr
Penetration Hardness (MPa)
10
w/c = 0.3, at 50 C
90
100
110
120
130
140
Time after Mixing (Minutes)
Figure 4-13 Effects o f penetration rate on the penetration hardness for cement pastes
with w/c = 0.3 at 50°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
1
w/c = 0.3, at 50°C
13 M Pa
0.5
Ln (dH/dt)
0
M Pa
-0.5
MPa
1
-1.5
2
1 M Pa
-2.5
0
30
60
90
120
150
mm/hr
Figure 4-14 Hardening rate changes caused by the penetration rate at several levels of
penetration hardness for cement pastes with w/c = 0.3.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
Penetration Hardness (MPa)
20
30°C (La Farge)
60°C (La Farge)
30°C (H olnam )
60°C (Holnam )
15
10
0
0
60
120
180
240
Time after Mixing (Minutes)
Figure 4-15 Penetration hardness development for cement pastes manufactured by La
Farge Type I and Holnam Type I.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
Table 4-2 Ln(Hardening rate) response surface coefficients and their t-values
for cement pastes.
Variable
Intercept
Hardness (H)
w/c Ratio (W)
lOOO/TCT)
Gear Ratio (R)
Hz
fP
rHR
TR
HW
WT
Holnam Type I
R2 = 0.9971
D.O.F. = 36
Coefficient
t-value
-6.41E-01
6.78E-01
15.96
-1.05E-01
-10.20
-5.81E-01
-46.25
-8.01E-02
-6.40
-6.34E-01
-31.68
4.16E-01
9.29
-1.45E-01
-6.77
-8.78E-02
-6.37
3.52E-02
2.28
5.28E-02
4.04
-6.18E-02
-5.04
—
La Farge Type I
R2 = 0.9943
D.O.F. = 38
Coefficient
t-value
-6.66E-01
7.31E-01
14.71
-1.82E-01
-14.51
-4.85E-01
-33.76
-1.53E-01
-12.32
-6.72E-01
-27.83
4.36E-01
8.28
-6.79E-02
-2.65
-6.92E-02
-4.10
-3.84E-02
-2.44
—
—
—
—
-
Table 4-3 Lists o f coded variables for cement pastes.
Hardness Level
w/c Ratio
1000/T
Gear Ratio
Coded
Uncoded
Coded
Uncoded
Coded
Uncoded
Coded
Uncoded
-0.333
-1
0.333
1
5
9
1
13
0
1
-1
0.35
0.3
0.4
-0.047
1
-1
1/333 (60°C) 1/318 (45°C)
1/303 (30°C)
-0.111
-1
1
50
100
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
different for the two cement batches, as discussed previously. The equations indicate that
the hardening rate increases with hardness level and temperature but is inversely
proportional to water content and gear ratio. Although interaction terms are different due
to the accuracy o f data in analysis, these interactions have significant physical meanings.
Figures 4-16 and 4-17 show the comparison of actual data and the predicted lines.
The solid lines in both figures represent the predicted values o f Ln(Hardening rate) with
the 95% confidence bands superimposed in the case o f a w/c ratio of 0.3 and the
penetration rate o f 6.4 mm/hr. Since it is found that a very good fit exists between the
actual data and the predicted values, the predicted equations appears to be valid.
As expected, high temperature enhances the rate o f hardening.
Another
observation is the confirmation of non-linear temperature effects on the hydration
process, which was discussed in the activation energy (section 4 .1.4.), because the 2nd
order terms o f temperature are included in the hardening rate equations. High water
amounts make the material softer and also reduce the hardening rate. The penetration
rate (inversely proportional to gear ratio) is not expected to change the hardening process
but it has significance in the equation due to the nature o f the penetration test. The higher
the penetration rate, the greater the hardening rate.
Because the correlation involves four variables, two o f them must be fixed to
generate a 3-dimensional surface. Typical plots o f the Ln(Hardening rate) as a function
o f temperature and hardness level are shown in Figure 4-18 in case o f low water amount
(w/c = 0.3) and high penetration rate (64 mm/hr) for cement paste, which is expected to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
1
La Farge Type I
13 M Pa
0
9 M Pa
Ln (dH/dt)
5 M Pa
1
1 M Pa
•2
3
■4
2.9
3
3.1
3.2
3.3
3.4
1000/T
Figure 4-16 Ln(Hardening rate) at w/c = 0.3 and 6.4 mm/hr o f the penetration rate for
La Farge Type I cement pastes. Solid lines indicate the predicted values.
Symbols are actual data points. Dashed lines are 95% confidence intervals.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
1
Holnam Type I
13 M Pa
9 M Pa
0
Ln (dH/dt)
5 M Pa
1
1 M Pa
2
•3
•4
2.9
3
3.1
3.2
3.3
3.4
1000/T
Figure 4-17 Ln(Hardening rate) at w/c = 0.3 and 6.4 mm/hr of the penetration rate for
Holnam Type I cement pastes. Solid lines indicate the predicted values.
Symbols are actual data points. Dashed lines are 95% confidence intervals.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
0 .5 "
0-0.5p
-l-
2,
JS
-1.5-2-
-2.5-3-
3 4 Penetration
(M Pa)
1000/T
9.4 »
■
S
□ 0-0.5
■ -0.5-0
■ -1-0.5
0 -1.5-1
■ -2-1.5
□ -2.5—2
■ -3-2.5
3.00 3.03 3.06 3.09 3.12 3.15 3.18 321 324 327 3.30
1000/T
Figure 4-18 Ln(Hardening rate) response surface and contour plot with respect to
temperature and hardness level for Holnam Type I cement pastes, fixing
w/c ratio at 0.3 and the penetration rate at 64 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
show the highest surface. Figure 4-19 shows the case of high water (w/c=0.4) and low
penetration rate (6.4 mm/hr), the lowest surface. La Farge paste results look similar.
Since the equation o f Ln(Hardening rate) for Holnam paste contains more
interactions, the equation o f Holnam paste was differentiated with respect to each
variable to explore the interaction effects. Note all variables are coded, as shown in
Table 4-3.
Hardness:
d ( ln(dH/dt))
,
;
■ = 0.68 - 126(H) + 125(H2) + 0.053(w / c) - 0.088(R)
5(H )
(Eq. 4-8)
w/c ratio:
d (ln(dH / dt))
g~^ / c j
= - 0.11 + 0.053(H) - 0.062(1000 / T)
(Eq. 4-9)
1000/T:
d (ln(dH / dt))
g (10Q01 T)
Gearratio:
°'58 ~ 0 2 9 (10001T ) + 0-035(R) - 0.062(w / c)
d (ln(dH / dt))
^
= -0 .0 8 -0.088(H ) + 0.035(1000 /T )
(Eq.4-10)
(Eq.4-11)
Figure 4-20 shows the hardness-w/c ratio interaction effect on the Ln(Hardening
rate) equation. At high hardness level, the curves o f two w/c ratios become slightly
closer to each other.
For the hardness-gear ratio interaction, the curve o f higher
penetration rate (lower gear ratio) is slightly below the lower penetration rate, but the
higher crosses over the lower to become greater as hardness levels goes up, as shown in
Figure 4-21. Since there is no interaction between hardness and temperature, these trends
are the same at the entire temperature range. The Ln(Hardening rate) response surface
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
Penetration
Hardness
1000/T
■
- -0 .5
-1.5-1
2-1.5
3.00 3.03 3.06 3.09 3.12 3.15 3.18 3.21 3.24 3.27 330
1000/T
Figure 4-19 Ln(Hardening rate) response surface and contour plot with respect to
temperature and hardness level for Holnam Type I cement pastes, fixing
w/c ratio at 0.4 and the penetration rate at 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
1
Penetration Rate = 6.4 mm/hr
In (dH/dt)
o
w/c = 0.3
1
■2
w/c = 0.4
3
-4
0
2
4
6
8
10
12
14
12
14
Penetration Hardness (MPa)
Penetration Rate = 64 mm/hr
In (dH/dt)
0
w /c = 0.3
1
•2
w/c = 0.4
•3
4
0
2
4
6
8
10
Penetration Hardness (MPa)
Figure 4-20 Ln(Hardening rate) for Holnam Type I cement pastes, representing the
effect of the interaction between hardness level and w/c ratio at fixed
temperature o f 30°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
1
w/c = 0.3
64 mm/hr
0
1
2:
"O
6.4 mm/hr
•2
•3
4
0
2
4
6
8
10
12
14
Penetration Hardness (MPa)
l
w/c = 0.4
o
T3
64 mm/hr
1
iT3
•2
6.4 mm/hr
3
4
0
2
4
6
8
10
12
14
Penetration Hardness (MPa)
Figure 4-21 Ln(Hardening rate) for Holnam Type I cement pastes, representing the
effect o f the interaction between hardness level and gear ratio (or
penetration rate) at fixed temperature o f 30°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
with respect to hardness versus w/c ratio are shown in Figures 4-22 (at 30°C) and 4-23 (at
60°C). Figures 4-24 and 4-25 show the Ln(Hardening rate) response surface with respect
to hardness and penetration rate at 30°C and 60°C, respectively.
Figure 4-26 shows the effects o f w/c ratio-temperature and w/c ratio-hardness
level interactions on the Ln(Hardening rate) equation. The tendency of decreasing the
Ln(Hardening rate) with w/c ratio is more noticeable at low temperature for all hardness
levels. Comparing the upper and lower figures reveals that the w/c ratio has less effect
on hardening rate at higher hardness.
Interestingly, w/c ratio does not affect the
Ln(Hardening rate) at high temperature (60°C) and high hardness level (13 MPa).
Figures 4-27 and 4-28 show Ln(Hardening rate) response surface with respect to w/c
ratio and temperature at the hardness levels o f 1 and 13 MPa, respectively. However, the
w/c ratio-hardness and w/c ratio-temperature interactions were not found in La Farge
paste (see Table 4-2).
The last significant interaction on the Ln(Hardening rate) is temperature-gear ratio
(or penetration rate).
Figure 4-29 shows the effects of this interaction.
Due to the
positive sign in the TR term (Table 4-2), the distance between the two curves is slightly
reduced as temperature decreases regardless of the hardness level and w/c ratio.
However, the sign o f this term is negative in the La Farge paste case (see Table 4-2).
Figures 4-30 and 4-31 show the Ln(Hardening rate) response surface with respect to
temperature and penetration rate at hardness level o f 1 MPa and 13 MPa, respectively.
The w/c ratio was fixed at 0.3 for both response surfaces.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
Penetration
Hardness
(MPa)
w/c Ratio
•13
-10.6
•9.4
■8.2
Penetration
Hardness
(MPa)
-11.8
7
5.8
4.6
□ -1-0.5
B -1.5-1
B -2-1.5
3.4
0 -2.5-2
22
B -3-2.5
1
□ -3.5-3
0.30 03 1 032 033 034 035 036 037 0.38 039 0.40
w/c Ratio
Figure 4-22 Ln(Hardnening rate) response surface and contour plot with respect to w/c
ratio and hardness level for Holnam Type I cement pastes, fixing
temperature at 30°C and the penetration rate at 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
Penetration
Hardness
w/c Ratio
□ 0-0.5
B -0.5-0
- -0 .5
□ -1.5-1
2-1.5
□ -2.5-2
030 031 032 033 034 035 036 037 038 039 0.40
w/c Ratio
Figure 4-23 Ln(Hardnening rate) response surface and contour plot with respect to w/c
ratio and hardness level for Holnam Type I cement pastes, fixing
temperature at 60°C and the penetration rate at 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
6 12 u 24 29
35 41
Penetration
Hardness
(MPa)
47 52 58 64
Penetration Rate (nun/hr)
-13
-11.8
9.4
82
Penetration
Hardness
(MPa)
-10.6
7
5.8
4.6
12
18
24
29
35
41
47
52
58
B -1.5-1
B -2-1.5
3.4
□ -2.5-2
22
B -3-2.5
1
6
□ -1-0.5
□ -3.5-3
64
Penetration Rate (mm/hr)
Figure 4-24 Ln(Hardnening rate) response surface and contour plot with respect to
penetration rate and hardness level for Holnam Type I cement pastes,
fixing temperature at 30°C and w/c = 0.4.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
0.50-0.5p
-1-
§
=
-1.5-2-2.56 12 18 24 29 35 41 a i ^
^
35 41 47 52 58 ^
Penetration
Hardness
(MPa)
1
Penetration Rate (mm/hr)
EL s -
*9
*» « 3a
CO o
3
_
CO
□ 0-0.5
B -0.5-0
o
1
'■
B -1.5-1
B -2-1.5
□ -2.5—2
6
12
18
24
29
35
41
47
52
58
64
Penetration R ate (mm/hr)
Figure 4-25 Ln(Hardnening rate) response surface and contour plot with respect to
penetration rate and hardness level for Holnam Type I cement pastes,
fixing temperature at 60°C and w/c = 0.4.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
Penetration Hardness = 1 MPa
In (dH/dt)
-1.5
-2.5 -
30°C
-3.5
0.28
0.3
0.32
0.34
0.36
0.38
0.4
0.42
0.4
0.42
w/c Ratio
Penetration Hardness = 13 MPa
In (dH/dt)
0.5
60°C
-0.5 -
30°C
0.28
0.3
0.32
0.34
0.36
0.38
w/c Ratio
Figure 4-26 Ln(Hardening rate) for Holnam Type I cement pastes, representing the
effect o f the interaction between w/c ratio and temperature at the
penetration rate o f 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ill
-2.75-
w/c Ratio
1000/T
-0.4
-029
■028
■027
■026
1
n
73
S3
o*
0.35
024
□ -2-1.75
H -2.25-2
023
B -2 5 -2 2 5
022
0 -2.75-2.5
021
B -3-2.75
02
3.00 3.03 3.06 3.09 3.12 3.15 3.18 3.21 324 327 320
□ -3 2 5 -3
1000/r
Figure 4-27 Ln(Hardening rate) response surface and contour plot with respect to
temperature and w/c ratio for Holnam Type I cement pastes, fixing
hardness level at 1 MPa and the penetration rate at 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
-025-
w/c Ratio
1000/T
■0.4
-039
■
■038
n
■037
S3
■036
50
a.
c
035
034
033
□ 0-025
B -025-0
■ -0.5-025
0.32 0 -0.75-0.5
031 ■-1-0.75
□-125-1
03
3.00 3.03 3.06 3.09 3.12 3.15 3.18 321 324 327 3.30
1000/T
Figure 4-28 Ln(Hardening rate) response surface and contour plot with respect to
temperature and w/c ratio for Holnam Type I cement pastes, fixing
hardness level at 13 MPa and the penetration rate at 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-1
Penetration Hardness = 1 MPa
In (dH/dt)
-1.5 -
-2
-
-2.5 6.4 mm/hr
-3 -
64 mm/hr
-3.5 -
1 ---------------1-------------- 1-------------- 1-------------- 1---------------1---------------f
3
3.1
3.2
3. 3
1000/T
1
Penetration Hardness = 13 MPa
In (dH/dt)
0.5
0
-0.5
6.4 mm/hr
-1
64 mm/hr
-1.5
T
3
1-----------1
-----------1------------1
------------'------------r
3.1
3.2
3.3
1000/T
Figure 4-29 Ln(Hardening rate) for Holnam Type I cement pastes, representing the
effect o f the interaction between temperature and penetration rate at w/c
ratio = 0.3.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
114
Penetration
Rate
(m m /hr)
1000/T
64
58
47
41
Penetration
Rate
(nun/hr)
52
35
29
24
Si
□ -2-1.7 5
B -2 2 5 -2
18
B -2.5-225
12
□ -2.75-2.5
6
3.00 3.03 3.06 3.09 3.12 3.15 3.18 321 324 327 330
B -3-2.75
1000/T
Figure 4-30 Ln(Hardening rate) response surface and contour plot with respect to
temperature and penetration rate for Holnam Type I cement pastes, fixing
hardness level at 1 MPa and w/c ratio at 0.3.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
115
Penetration
Rate
(mm/hr)
1000/T
~
3*
w
o
33 JO 3
»
3 ti ?
- S’3
s:.
j-i
3
□ 0.25-0.5
S 0-0.25
■ -0.25-0
0 -0.5-0.25
■ -0.75-0.5
□ -1-0.75
3.00 3.03 3.06 3.09 3.12 3.15 3.18 3.21 3.24 3.27 3 30
1000/T
Figure 4-31 Ln(Hardening rate) response surface and contour plot with respect to
temperature and penetration rate for Holnam Type I cement pastes, fixing
hardness level at 13 MPa and w/c ratio at 0.3.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
116
Other Ln(Hardening rate) response surfaces with respect to w/c ratio and
penetration rate are shown in Figures 4-32 (at 30°C) and 4-33 (at 60°C). The hardness
level was fixed at 13 MPa for both surfaces.
Table 4-4 summarizes the effects o f major variables on the Ln(Hardening rate)
equation for Holnam paste. The summary o f interaction effect is shown in Table 4-5.
Table 4-4 Effects o f increasing the major variables on Ln(Hardening rate) of Holnam
cement pastes.
Variables
Hardness Level
w/c ratio
Temperature
Penetration Rate
Ln(Hardening rate)
Increase
Decrease
Increase
Increase
Tendency
Cubic
Linear
Squared
Linear
Table 4-5 Effects o f the interaction on Ln(Hardening rate) o f Holnam cement pastes.
Partial Derivative
5(ln(dH / dt))
3(H)
5(ln(dH / dt))
5(w / c)
5(ln(dH / dt))
5(1000/T )
Conditions
High w/c ratio
Low w/c ratio
High penetration rate
Low penetration rate
High temperature
Low temperature
Effect on 5(ln(dH /dt))
High penetration rate
Low penetration rate
Negative
Positive
Positive
Negative
Positive
Negative
Positive
Negative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
117
Penetration
R ate
(nun/hr)
w/c Ratio
□ -0.7—0.6
H -0.8—0.7
■ -0.9-0.8
□ -1-0.9
■ - 1. 1 - 1
□ - 1 2 - 1.1
030 031 032 033 034 035 036 037 038 039 0.40
w/c Ratio
Figure 4-32 Ln(Hardening rate) response surface and contour plot with respect to w/c
ratio and penetration rate for Holnam Type I cement pastes, fixing hardness
level at 13 MPa and temperature at 30°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Rate
(mm/hr)
w/c Ratio
a?
JO A
» S'
S' 3
s-.
o
s
□ 0.4-0.5
■ 0.3-0.4
■ 0.2-0.3
0
0.1-03
■ 0-0.1
030 031 032 033 034 035 036 037 038 039 0.40
w/c Ratio
Figure 4-33 Ln(Hardening rate) response surface and contour plot with respect to w/c
ratio and penetration rate for Holnam Type I cement pastes, fixing hardness
level at 13 MPa and temperature at 60°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
119
4.2.3. Relationships between Hardness and Degree o f Hydration
Previously, it was assumed that hardness is only related to degree o f hydration (a)
o f cement paste.
Since a viscoelastic material usually becomes softer as temperature
rises, temperature should be considered in establishing the relationships between
hardness and degree o f hydration.
relationship, too.
Additionally, water content may change the
For this test, cement pastes with various w/c ratios were cured at
various temperatures until the penetration hardness reach preset points, then the loss-onignition procedures were immediately performed. Exploring the effects o f w/c ratio (I)
and temperature (II) was done separately, as shown in Table 4-6.
T able 4-6 Experimental design for investigating the relationships between
degree o f hydration and hardness level.
#
1
2
3
4
5
6
7
8
9
I: Fixed Temperature at 20°C
W/c ratio
Hardness (MPa)
0.3
1
0.3
5
0.3
13
0.4
1
0.4
5
0.4
13
0.5
1
0.5
5
0.5
13
H: Fixed w/c ratio at 0.3
Temperature (°C) Hardness (MPa)
20
1
20
5
20
13
40
1
40
5
40
13
60
1
60
5
60
13
Equations were established by MC analysis, shown in Eq. 4-12 and 13.
I: a = 0.063 +0.016 (Ln(H)) + 0.016 (w/c) + 0.008 (Ln(H))(w/c)
: R2 = 0.9998, D.O.F. = 4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Eq. 4-12)
120
II: a = 0.067 + 0.016 (Ln(H)) - 0.020 (1000/T) - 0.008 (Ln(H))( 1000/T)
: R2 = 0.9999, D.O.F. = 3
(Eq. 4-13)
All variables are coded as +1 for Ln(H) = Ln(13 MPa), w/c = 0.5, 1/T = 1/293 and -1 for
Ln(H) = Ln(l MPa), w/c = 0.3, 1/T = 1/333.
From the equations, hardness is logarithmically related to the degree o f hydration.
The w/c ratio and temperature are significant factors. Figures 4-34 and 4-35 show the
correlation between the degree o f hydration and the hardness for I and II experiments,
respectively. Actual data points are superimposed upon the predicted lines based on the
equations.
Since the predicted lines are in good agreement with the real values, the
prediction is valid.
Since the free water (or capillary pores) amount is not considered for calculating
the degree of hydration, the amount o f this free water increases with w/c ratio regardless
o f the level of the degree o f hydration. If a paste has more free water, the penetration
hardness would be lower. Thus, although two pastes have the same degree o f hydration,
the penetration hardness level is also dependent on the free water amount. Although free
water is used as a water source for cement hydration, the reduction in the free water
amount is independent the initial w/c ratios. This means that a higher degree o f hydration
is required to obtain a given penetration hardness at higher w/c ratios. This is seen in
Figure 4-36, which shows the degree o f hydration response surface and its contour plot
with respect to w/c ratio. When the surface is projected onto the degree of hydration-w/c
ratio plane o f this figure, the slopes o f the projected lines at constant hardness become
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
121
0.12
Degree of Hydration (a)
at 20°C
0 .0 8 -
0 .0 4 -
0 .0 2 -
0
5
•
w/c = 0.3
■
w/c = 0.4
A
w/c = 0.5
10
15
Penetration Hardness (MPa)
Figure 4-34 Correlation between degree o f hydration and penetration hardness
at various w/c ratios for cement pastes cured at 20°C. Actual data points
are superimposed on the predicted line. The error bars indicate 95%
confidence bands based on three specimens.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
122
0.12
Degree of Hydration (a)
w/c=0.3
0.08 -
0.04 -
20°C
0.02
-
40°C
60°C
0
5
10
15
Penetration Hardness (MPa)
Figure 4-35 Correlation between degree o f hydration and penetration hardness
at various temperatures for cement pastes with fixed w/c ratio at 0.3.
Actual data points are superimposed on the predicted line.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
123
•9
0.0813.0
Penetration
Hardness
(MPa,
log scale)
1.7
w/c Ratio
ST 13-0
•
10.1
h7.8
-
6.0
-4.7
?
o
ora
»
?
s.
Eg
S
|
a2
S’-
3 s §
-3.6
2.8
□ 0.10-0.12
-23.
n 0.08-0.10
h 1-7
■ 0.06-0.08
-
•13
El 0.04-0.06
■ 0.02-0.04
•
1.0
0.30 032 034 036 0.38 0.40 0.42 0.44 0.46 0.48 0.50
w/c Ratio
Figure 4-36 Degree o f hydration response surface and contour plot with respect to
penetration hardness and w/c ratio for cement pastes cured at 20°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
124
greater with hardness level. This indicates the hardness-w/c ratio interaction.
Figure 4-37 shows the degree o f hydration response surface and contour plot with
respect to temperature. Because o f the same w/c ratio (0.3), the free water content should
be unchanged with temperature at a given degree o f hydration. However, the degree of
hydration is influenced by temperature. The response surface shows that cement paste
needs to be more hydrated to reach a certain level o f hardness at higher temperature. In
other words, the matrix becomes softer due to high temperature.
This indicates
temperature effects on the Theological behavior (viscosity) o f cement paste.
The
influence o f temperature becomes more dominant at high hardness level because of the
hardness-temperature interaction.
This interaction effect proves that the influence of
temperature on the viscosity o f cement paste becomes greater when cement paste is more
solid-like due to hydration. This is reasonable because the viscosity of most viscoelastic
materials, for instance silicate glasses, will be rapidly dropped by raising temperature,
when they are in the solid state, and thereafter smoothly reduced after passing the
softening point, which is associated with more fluid-like behavior (Doremus, 1994).
Thus, the relationships among hardness, degree o f hydration, and temperature for
cement paste are successfully established. Temperature and water content are important
factors in these relationships, due to additional Theological effects.
The temperature
effects on rheology were also found in the activation energy analysis in section 4.1.4.
Therefore, careful consideration o f rheological behavior is required for a better
understanding o f hardening.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
125
0.12
s
.o
*«
•a
>»
S
e
o
e
OX)
13.0
Penetration
Hardness
(MPa,
log scale)
Q
10
1000/T
1.0
1.1
s
3
7
^
-S3 S S 3 w
S’ 2. 2.
"09 1os 1371
n
S3
09
O
S
5
i
□ 0. 10-0.12
I
B 0.08-0.10
7
5
B 0.06-0.08
El 0.04-0.06
B 0.02-0.04
)
3.00 3.04 3.08 3.13 3.17 3.21 325 329 323 3.37 3.41
1000/T
Figure 4-37 Degree o f hydration response surface and contour plot with respect to
penetration hardness and temperature for cement pastes with fixed w/c ratio
at 0.3.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
126
4.3. Sand Effects on H ardening Behavior
The instrumented penetration test can be applied to study mortar as well as
cement paste. In this section, sand effects on the hardening behavior will be discussed.
4.3.1. Influence o f Sand on Setting Time
The initial setting time was also assumed as the time when the penetration
hardness reaches 3.5 MPa. Using MC analysis, the equation o f initial setting time was
developed:
Time (min) = 120 - (8 ± 3)(s/c) + (44 ± 3)(1000/T) + (16 ± 6)(1000/T)2
(Eq. 4-14)
(R2 = 0.9923; Modified R2 = 0.9940; D.O.F = 3; SyJt = 4.1807; STest = 1.9760)
All variables were coded as +1 for s/c = 2 , 1/T = 1/303 and -1 for s/c =0, 1/T = 1/333.
The 95% confidence interval o f each coefficient is given.
The response surface and its contour plot o f the initial setting time for cement
mortar are shown in Figure 4-38. The initial setting time was reduced with increased
temperature and sand amount. Similar to paste, the hydration of mortar is accelerated by
raising the temperature, resulting in a shorter initial setting time. Due to the rigidity o f
sand particles, mortar can resist the penetrator more easily than paste when cement
hydration products tightly link sand particles. However, neat cement paste is expected
not to show such resistance at the same time because the rigidity o f hydrated cement
paste is less than that of linked sand particles at around the initial setting time. Therefore,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
127
Temperature
s/c Ratio
□ 180-210
■
B 150-180
120-150
B 90-120
60-90
0.0 0.2 0.4 0.6 0.8 1.0 12 1.4 1.6 1.8 10
s/c Ratio
Figure 4-38 Initial setting time response surface and contour plot for cement mortars
(Holnam Type I) with various s/c ratios and curing temperatures measured
by the instrumented penetration test.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
128
increased sand content decreases the initial setting time, but not greatly.
4.3.2. Influence o f Sand on the Hardening Rate
Table 4-7 shows coefficients and their t-values o f Ln(Hardening rate) equations
for Holnam and La Farge cement mortar by MC analysis. Ail variables are coded, as
shown in Table 4-8. The equations indicate that the hardening process of cement mortar
is very similar to cement paste because the signs o f the major factors, such as hardness,
temperature and penetration rate, are exactly the same as those shown in the paste
equation.
Also, the equations are not significantly dependent o f manufacturer.
The
hardness-penetration rate interaction also exists with the same sign in both mortar
equations, as with the paste equations. Sand affects the equations as the 1st order or
interactions, but its level o f importance is small.
As expected in the equations, the Ln(Hardening rate) response surface for cement
mortar results (Figure 4-39) looks similar to that of cement paste (Figure 4-19; w/c = 0.4
and penetration rate = 6.4mm/hr), with respect to the levels o f Ln(Hardening rate) as well
as the shape. This implies that sand contributes little to the hardening process of cement
hydration.
Since sand reduces the initial setting time but does not influence the
hardening, it only acts as a filler. However, if a reactive sand is used, sand can influence
the hardening process.
The effect o f the s/c ratio-gear ratio interaction can be estimated by partial
differentiation, as follows (Holnam mortar case):
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
129
Table 4-7 Ln(Hardening rate) response surface coefficients and their t-values
for cement mortars.
Variable
Intercept
Hardness (H)
s/c Ratio (S)
1000/T (T)
Gear Ratio (R)
H*
H3
T
HR
HS
SR
ST
Holnam Type I
R2 = 0.9932
D.O.F. = 38
Coefficient
T value
-7.14E-01
6.75E-01
10.86
-5.15E-02
-3.27
-5.91E-01
-33.27
-1.01E-01
-5.99
-6.34E-01
-21.65
4.08E-01
6.23
-1.18E-01
-3.97
-7.09E-02
-3.52
—
—
—
5.14E-02
2.93
—
La Farge Type I
R2 = 0.9954
D.O.F. = 35
Coefficient
T value
-6.88E-01
7.13E-01
13.20
—
—
—
—
-5.52E-01
-9.94E-02
-6.48E-01
4.19E-01
-2.42E-01
-6.37E-02
-5.28E-02
6.74E-02
-6.56E-02
-34.60
-6.67
-24.87
7.32
-9.11
-3.50
-2.83
4.21
-3.97
Table 4-8 Lists o f coded variables for cement mortars.
Hardness Level
s/c Ratio
1000/T
Gear Ratio
Coded
Uncoded
Coded
Uncoded
Coded
Uncoded
Coded
Uncoded
-1
1
-1
0
-1
1/333 (60°C)
-1
10
-0.333
5
0.333
9
0
1
-0.047
1/318 (45°C)
-0.111
50
1
13
1
2
1
1/303 (30°C)
1
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
130
Penetration
Hardness
(M Pa)
1000/T
rl 3
-10.6
-9.4
-8.2
-7
-5.8
-4.6
-3.4
-22
-1
3.00 3.03 3.06 3.09 3.12 3.15 3.18 3.21 32 4 32 7 330
Penetration
Hardness
(MPa)
-11.8
□ 0-0.5
■ -0.5-0
■ -1 -0 .5
0 -1.5-1
■ -2 -1 .5
□ -2 5 -2
■ -3-2.5
0 -3.5-3
1000/T
Figure 4-39 Ln(Hardening rate) response surface and contour plot with respect to
temperature and hardness level for Holman Type I cement mortars, fixing
s/c ratio at 2 and the penetration rate at 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
131
s/c ratio:
d (ln(dH/dt))
' = - 0.052 + 0.051(R)
d (s / c)
(Eq. 4-15)
From Eq. 4-15, increasing the s/c ratio reduces the Ln(Hardening rate) at low gear
ratio (or high penetration rate), while it has negligible effect at the highest gear ratio.
This is seen in the Ln(Hardening rate) response surfaces for Holnam cement mortar
shown in Figures 4-40 (low penetration rate) and 4-41 (high penetration rate).
The
presence o f the s/c ratio-gear ratio interaction implies that the sand content may influence
penetration hardness Theologically rather than chemically. La Farge mortar shows more
complicated interactions between s/c ratio and other variables. However, the influence of
the s/c ratio-gear ratio interaction on the Ln(Hardening rate) is similar because o f the
same positive sign and similar coefficient of this interaction.
4.4. Influence of Reactive Admixture on Hardening Behavior
Reactive admixtures (pozzolanic materials) are expected to influence the
hardening process o f cement paste due to the differences in chemical composition and
particle sizes. The maximum amounts o f substituted admixtures were chosen according
to the industrial requirements: 50 wt% for GGBFS, 10 wt% for silica fume and 20 wt%
for fly ash.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
132
3-4 Penetration
Hardness
s/c Ratio
-13
■10.6
9.4
Penetration
Hardness
(MPa)
-11.8
82
7
5.8
□ -1.5-1
4.6
3.4
■ -2.5—2
22
0 -3 -2 .5
1
0.0 0 2
0.4 0.6
0.8
1.0 12
■ -2 -1 .5
■ -3.5-3
1.4 1.6 1.8 2.0
s/c Ratio
Figure 4-40 Ln(Hardening rate) response surface and contour plot with respect to s/c
ratio and hardness level for Holman Type I cement mortars, fixing
penetration rate at 6.4 mm/hr and temperature at 30°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
133
34 Penetration
Hardness
(M Pa)
“
s/c Ratio
□ - -0 .5
2-1.5
El -1 5 -2
-3-2.5
□ -3.5-3
0.0 02. 0.4 0.6 0.8
1.0
12
1.4 1.6
1.8 10
s/c Ratio
Figure 4-41 Ln(Hardening rate) response surface and contour plot with respect to s/c
ratio and hardness level for Holman Type I cement mortars, fixing
penetration rate at 64 mm/hr and temperature at 30°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
134
4.4.1. Effects o f Reactive Admixtures on the Initial Setting Time
The coefficients o f the equations o f initial setting time for pastes containing
different reactive admixtures are shown in Table 4-9.
Table 4-9 Initial setting time response surface coefficients for pastes
containing reactive admixtures.
Variable
Intercept
w/r Ratio (W)
Admixture amount (A)
1000/T (7)
T
WT
AT
Silica Fume
R2 = 0.9990
D.O.F. = 4
99
19 ± 1
-6 ± 1
34 ±1
18 ± 2
14 ± 2
-
GGBFS
R2 = 0.9972
D.O.F. = 4
127
24 ± 5
20 ± 4
52 + 5
14 ± 8
14 ± 6
15 ± 4
Class F Fly Ash
R2 = 0.9991
D.O.F. = 5
112
21 ± 2
4±2
38 + 2
13 ± 4
10 ± 3
—
Note all variables are coded as +1 for w/r = 0.4, l/T = 1/303, SF % = 10%, GGBFS % =
50%, FA % = 20% and -1 for w/r = 0.3, l/T = 1/333, SF % = 0%, GGBFS % = 0%, FA
% = 0%. The 95% confidence interval on each coefficient is also given.
Compared with cement paste (Eq. 4-3), the tendency o f w/c ratio (or w/r ratio),
temperature and their interaction (WT) effects on the initial setting time is unchanged by
substituting reactive admixtures.
The initial setting time increases as the w/r ratio
increases and temperature decreases. The effect of the w/r ratio-temperature interaction
behaves similarly to cement paste, as shown in Figure 4-8.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
135
When reactive admixtures are substituted for cement in paste, they alter the initial
setting time, depending on their type. The particle sizes o f GGBFS and Class F fly ash
used are not markedly different from that o f cement as shown in Table 3-1, thus the
effects o f the surface area o f these materials can be neglected.
The extremely small size o f silica fume particles causes reduction o f the initial
setting time because pores between cement particles filled with silica fume contribute
nucleation sites or make the paste denser. Since the reaction of GGBFS is generally
delayed compared with that o f cement because of its inertness, paste with GGBFS
substitutions reaches the initial set later. Also, a GGBFS amount-temperature interaction
with a positive sign was found. This indicates the retardation o f the initial setting time
due to adding GGBFS becomes greater at low temperature, as shown in Figure 4-42 (a),
compared with no such interaction for silica fume (Figure 4-42 (b)). Class F fly ash
retards the initial setting time, but not as much as GGBFS. Table 4-10 summarizes the
effects o f adding reactive admixtures on the initial setting time, including interactions.
Table 4-10 Effects o f adding reactive admixtures on initial setting time.
Variables
w/r ratio
Admixture
Amount
Temperature
Condition
HighT
LowT
HighT
LowT
High w/r
Low w/r
High Amount
Low Amount
Silica fume
Less Increase
More Increase
Decrease
(No interaction)
More Decrease
Less Decrease
Decrease
(No interaction)
GGBFS
Less Increase
More Increase
Less Increase
More Increase
More Decrease
Less Decrease
More Decrease
Less Decrease
Fly ash
Less Increase
More Increase
Increase
(No interaction)
More Decrease
Less Decrease
Decrease
(No interaction)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Initial Setting Time (Minutes)
136
300
30°C
240
180 -
120
60°C
-
60 ■
(a) GGBFS
0
10
20
30
40
50
8
10
Initial Setting Time (Minutes)
GGBFS %
240
30°C
180
60°C
60 •
(b) Silica Fume
0
2
4
6
Silica Fume %
Figure 4-42 Initial setting time for cement pastes containing reactive admixtures,
representing the effect of the interaction between admixture amount and
temperature (a) GGBFS (b) silica fume.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
137
The initial setting time response surfaces are shown in Figures 4-43 (silica fume),
4-44 (GGBFS) and 4-45 (fly ash).
4.4.2. Effects o f Reactive Admixtures on the Hardening Rate
The results of MC analysis o f Ln(Hardening rate) for pastes containing pozzlans
were shown in Table 4-11. All variables are coded according to Table 4-12.
All previously found major coefficients (such as H, W, T, R, H2, H3 and HR) from
cement paste (Table 4-2) also exist with the same sign when cement was substituted by
reactive admixtures. The squared T term was found in silica fume and fly ash, while
being insignificant in GGBFS. The hardening rate was influenced differently by the type
of reactive admixtures. GGBFS reduces the hardening rate, while silica fume increases
it. The amount of fly ash affected the hardening rate less significantly in itself, but it
interacts with other variables to change the hardening rate.
To investigate the effects of admixtures on Ln(Hardening rate), the equations are
partially differentiated, as follows (all variables are coded; A = amount o f admixtures):
Silica fume paste:
GGBFS paste:
Fly ash paste:
5 ( ln(dH/dt))
5(A ) =
5(ln(dH/dt))
+ ^*054 (H) + 0.065 (w / c)
= -0 .1 7 6 -0 .0 8 1 ( w / c ) -0.043 (1000/T )
5(ln(dH /dt))
5 (A)
= 0,036 ( w / c ) ~°-Q78 (1Q0Q 1 ^
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Eq. 4-16)
(Eq.4-17)
(Eq.4-18)
9 10
Silica Fume wt%
■30
•33
36
38
41
£
3
Tt
3
e
3
44
4/
□ 180-210
50
B 150-180
54
B 120-150
57
0 90-120
B 60-90
60
2
3
4
5
6
7
Silica Fume wt%
Figure 4-43 Initial setting time response surface and contour plot for cement pastes
containing silica fume (w/r = 0.4) with various silica fume wt% and
temperature, measured by the instrumented penetration test.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
139
'S'
■§ 240E
H
BX)
210-
e
s
180-
<Z5
150-
is
■a
120-
Temperature
90-
GGBS wt%
£
3
B
IS
3
c
3
□ 240-270
m 210-240
■ 180-210
0 150-180
■ 120-150
□ 90-120
20
30
GGBS wt%
Figure 4-44 Initial setting time response surface and contour plot for cement pastes
containing GGBFS (w/r = 0.4) with various GGBFS wt% and temperature,
measured by the instrumented penetration test.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
140
210-'
JB*
180Si
B
H
150-
M
e
’£
120-
'■ S
90-
o
Cft
'B
60-
Temperature
10 12 14 16 lg 2Q
Fly Ash wt%
3
*8
3
s
3
□ 180-210
B 150-180
1
B 120-150
0 90-120
2
4
6
8
10
12
14
16
18 20
Fly Ash wt%
Figure 4-45 Initial setting time response surface and contour plot for cement pastes
containing class F fly ash (w/r = 0.4) with various fly ash wt% and
temperature, measured by the instrumented penetration test.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
141
Table 4-11 Ln(Hardening rate) response surface coefficients and their t-values
for pastes containing reactive admixtures.
Variable
Intercept
Hardness (H)
w/r ratio (fV)
Amount (A)
1000/T (T)
Gear Ratio (R)
Hl
H3
rHR
HA
WA
AT
TR
Silica Fume
(R2 = 0.9983)
D.O.F. = 31
Coefficient T value
-6.56E-01
6.64E-01
20.57
-1.68E-01
-18.67
3.50E-02
3.90
-4.78E-01
-41.67
-9.92E-02
-10.61
-6.31E-01
-40.68
4.09E-01
12.00
-1.23E-01
-6.80
-6.93E-02
-6.29
5.39E-02
4.71
6.48E-02
4.36
—
—
—
-
“
GGBFS
(R2 = 0.9976)
D.O.F. = 33
Coefficient T value
-8.66E-01
6.39E-01
15.579
-1.57E-01
-11.27
-1.76E-01
-16.52
-35.64
-5.59E-01
-8.93E-02
-5.65
-6.35E-01
-32.91
4.07E-01
9.44
—
Class F Flv Ash
R2 = 0.9988
D.O.F. = 29
Coefficient T value
-7.94E-01
6.05E-01
20.61
-2.32E-01
-17.07
—
—
—
-5.58E-01
-6.20E-02
-6.16E-01
4.06E-01
-6.17E-02
-4.38E-02
-54.80
-7.09
-44.91
13.17
-3.70
-4.54
—
—
-1.22E-01
-8.35
—
—
—
—
-8.13E-02
-4.28E-02
-4.84
-2.45
3.60E-02
-7.78E-02
5.32E-02
2.66
-5.68
4.85
-
—
Table 4-12 Lists o f coded variables for pastes containing reactive admixtures.
Hardness Level
w/r ratio
Amount o f Reactive
Admixtures
1000/T
Gear Ratio
Coded
Uncoded
Coded
Uncoded
Coded
for Silica Fume
for GGBFS
for Fly Ash
Coded
Uncoded
Coded
Uncoded
-1
1
-0.333
5
-1
0.3
-1
0%
0%
0%
-1
1/333 (30°C)
-1
10
0.333
9
0
0.35
0
5%
25%
10%
-0.047
3 0 3 (30°C)
-0.111
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
13
1
0.4
1
10%
50%
20%
1
303 (30°C)
1
100
142
Figure 4-46 shows the effects of silica fume content on the Ln(Hardening rate),
interacting with other variables. The enhancement o f the hardening rate by adding silica
fume becomes greater at high w/r ratio as hydration proceeds.
This trend is not
influenced by temperature. When silica fume is added into high water content paste, the
silica fume particles will exist in pores, resulting in increasing viscosity and acting as
nucleation sites for hydration. Also, it acts as an additional silica source for hydration.
Therefore, the hydration process can be enhanced by adding silica fume. However, the
silica fume effect on hydration becomes less as the water content decreases.
When
cement is mixed with less water, the pore solution in the paste is saturated more easily.
Thus, it may not be necessary to have additional nucleation sites, so the influence of
adding silica fume can be insignificant. The corresponding Ln(Hardening rate) response
surfaces are shown in Figures 4-47 and 4-48.
The contribution o f GGBFS content is different from that o f silica fume; it makes
the hardening process slower as well as the initial setting time longer. These could be
associated with the characteristics o f GGBFS such as slow ion dissolution and chemical
reactions and low heat evolution.
Figure 4-49 shows the interaction effects on the
hardening rate. The more water content, the greater the reduction o f Ln(Hardening rate)
due to GGBFS content. Since the admixture amount-temperature interaction exists, this
reduction tendency decrease as temperature increase. The corresponding Ln(Hardening
rate) response surface are shown in Figures 4-50 and 4-51.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
143
2
In (dH/dt)
Penetration Hardness = 1 MPa
w /r = 0.3
3
w /r = 0.4
•4
0
2
4
6
8
10
Silica Fume (wt%)
0
Penetration Hardness =13 MPa
In (dH/dt)
w /r = 0.3
1
w /r = 0.4
2
0
2
4
6
8
10
Silica Fume (wt%)
Figure 4-46 Ln (Hardening rate) for cement pastes containing silica fume, representing
the effect o f the interaction between silica fume amount and w/r ratio at
two levels o f penetration hardness at 30°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
144
-0.5-1____^
|3
-1.5-
1
-2-
-3-
Penetration
Hardness
(MPa)
Silica Fume (wt%)
-1-0.5
2.5-2
-3-2.5
3
4
5
6
7
8
Silica Fume (wt%)
Figure 4-47 Ln(Hardening rate) response surface and contour plot with respect to silica
fume amount and hardness level for cement pastes containing silica fume
with w/r = 0.3 at 30°C and 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
145
Penetration
Hardness
(MPa)
Silica Fume (wt%)
r 13
ki.8
-9.4
-8 2
Penetration
Hardness
(MPa)
-10.6
-7
-5.8
-4.6
-3.4
-2.2
□ -1 -0 .5
■ -1.5-1
■ -2 -1 .5
0 -2 .5 -2
■ -3-2.5
□ -3.5-3
-1
Silica Fume (wt%)
Figure 4-48 Ln(Hardening rate) response surface and contour plot with respect to silica
fume amount and hardness level for cement pastes containing silica fume
with w/r = 0.4 at 30°C and 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
146
In (dH/dt)
0
w/r = 0.3
1
w/r = 0.4
•2
0
10
20
30
40
50
40
50
GGBFS (wt%)
l
In (dH/dt)
w/r = 0.3
0
w/r = 0.4
1
0
10
20
30
GGBFS (wt%)
Figure 4-49 Ln (Hardening rate) for cement pastes containing GGBFS, representing the
effect o f interactions between GGBFS amount and w/r ratio, and between
amount and temperature at two levels o f temperatures at 13 MPa.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
GGBS (wt%)
-3.00
-3.03
-3.06
•3.oy
o
o
o
=
9
■3.12
3.15
3.18
3.21
□ 0-0.25
H -025-0
■ -0.5-025
324
10
15
20
25
30
35
40
45
0 -0.75-0.5
327
■ -1-0.75
3,30
□ -1.25-1
50
GGBS (wt%)
Figure 4-50 Ln(Hardening rate) response surface and contour plot with respect to
GGBFS amount and temperature for cement pastes containing GGBFS with
w/r = 0.3 at 13 MPa and 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
GGBS (wt%)
-3.00
-3.03
3.09
1000/T
•3.06
3.12
3.15
□ -025-0
3.18
■ -0.5-025
321
■ -0.75-0.5
0 -1-0.75
324
■ -1 2 5 -1
327
□ -1.5-1.25
3 30
15
20
25
30
35
40
45
■ -1.75-1.5
50
GGBS (wt%)
Figure 4-51 Ln(Hardening rate) response surface and contour plot with respect to
GGBFS amount and temperature for cement pastes containing GGBFS
with w/r = 0.4 at 13 MPa and 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
149
The Ln(Hardening rate) o f fly ash-containing paste shows a two-fold temperaturedependent behavior, unlike other admixtures. At low temperature, the hardening rate
decreases with fly ash content, while it increases at high temperature for all hardness
levels, as shown in Figure 4-52.
The contribution of w/r ratio-admixture amount
interaction is opposite to that o f GGBFS but less than the temperature-admixture
interaction effects. The corresponding response surfaces are shown in Figures 4-53 and
4-54.
Consequently, the overall hydration process of reactive admixture-containing
pastes is very similar to that of cement paste but will be varied a little with the chemical
and physical nature o f admixtures.
Table 4-13 summarizes the reactive admixture
effects.
Table 4-13 Effects o f adding reactive admixtures on Ln(Hardening rate) o f pastes
containing reactive admixtures.
Variable
A*
Conditions
High hardness level
Low hardness level
5(ln(dH / dt))
High w/r ratio
5(A)
Low w/r ratio
High temperature
Low temperauture
* A = Reactive admixture amount
Silica Fume
Increase
+
—
+
—
No
interaction
GGBFS
Decrease
No
interaction
Class F Fly Ash
T, w/r dependent
No
interaction
—
+
+
+
+
-
—
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
—
150
1
w/r = 0.3
In (dH/dt)
60°C
0
30°C
1
0
5
10
15
20
Class F Fly Ash (wt%)
In (dH/dt)
0
60°C
-1
30°C
w/r = 0.4
.?
0
5
10
15
20
Class F Fly Ash (wt%)
Figure 4-52 Ln (Hardening rate) for cement pastes containing class F fly ash,
representing the effect of the interaction between fly ash amount and
temperature at two levels of w/r ratios at 13 MPa of penetration hardness.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Class F Fly Ash (wt%)
□ 0-0.25
n - 025-0
■ - 0. 5-025
□ -0.75-0.5
■ -1-0.75
4
6
8
10
12
14
16
18
20
Class F Fly Ash (wt%)
Figure 4-53 Ln(Hardening rate) response surface and contour plot with respect to class F
fly ash amount and temperature for cement pastes containing class F fly ash
with w/r = 0.3 at 13 MPa and 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
152
-0.25-0.55
3
.
-0.75.
s
-l-1.25-1.5-
1000/T
10 12 14 I6
18 20
Class F Fly Ash (wt%)
□ -0.25-0
-0.5-0.25
0.75-0.5
0 -1 -0 .7 5
-1.25-1
□ -1.5—1.25
4
6
8
10
12
14
16
18
20
Class F Fly Ash (wt%)
Figure 4-54 Ln(Hardening rate) response surface and contour plot with respect to Class
F fly ash amount and temperature for cement pastes containing class F fly
ash with w/r = 0.4 at 13 MPa and 6.4 mm/hr.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
153
4.5. Rheological Properties o f Cement Paste
Since penetration hardness development is the transition o f cement paste from
liquid to solid due to its hydration, the instrumented penetration test can offer information
about the rheology o f cement paste, which can be compared with results obtained with a
variable oscillatory rheometer (VOR).
4.5.1. Rheology Studies by Rheometer and Penetration Test
The variable oscillatory rheometer is able to monitor rheological properties of
viscoelastic materials, like cement pastes, in terms of elastic storage and viscous loss
modulus (Struble, 1991). Because both the moduli and the penetration resistance have
the same dimensions, it is possible to compare their responses in the same figure, in spite
o f no theoretical relationship between them. From Figures 4-55 and 4-56, the increment
o f both the elastic storage modulus and the penetration rate became greater at lower w/c
ratio because lower water content in cement paste causes more solid-like behavior. The
elastic storage modulus was measurable prior to the penetration hardness, with some
overlap. This can be explained by the comparisons between the viscous loss modulus
and the penetration hardness, shown in Figures 4-57 and 4-58. The viscous loss modulus
increased initially and then decreased post-peak to zero. The chemistry o f the cement
hydration process can explain the changes o f the viscous loss modulus.
Since the
particles begin to be dissolved in the pore solution immediately after mixing with water,
cement paste behaves more fluid-like resulting in an increase in the viscous loss modulus.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
154
500
G1(Rheom eter)
H (Penetration)
400
£
300
X
u,
o
O
200
100
0
0
30
60
90
Time after Mixing (Minutes)
Figure 4-55 Comparisons between elastic storage modulus (G’) obtained by the
rheometer and penetration hardness measured by the instrumented
penetration for cement pastes with w/c =0.3.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
120
155
500
■s— G' (Rheometer)
* — H (Penetration)
400 -
300 -
X
i-4
o
O
200 -
100
-
0
30
60
90
120
150
180
Time after Mixing (Minutes)
Figure 4-56 Comparisons between elastic storage modulus (G’) obtained by the
rheometer and penetration hardness measured by the instrumented
penetration for cement pastes with w/c = 0.4.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
210
156
■©— G" (Rheom eter)
G" or H (kPa)
■ — H (Penetration)
50 -
25 -
0
30
60
90
120
Time after Mixing (Minutes)
Figure 4-57 Comparisons between viscous loss modulus (G”) obtained by the rheometer
and penetration hardness measured by the instrumented penetration for
cement pastes with w/c =0.3. Vertical separation o f H points is the
resolving power o f the apparatus.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
157
G" (Rheom eter)
H (Penetration)
G" or H (kPa)
40 -
30 -
0
30
60
90
120
150
180
Time after Mixing (Minutes)
Figure 4-58 Comparisons between viscous loss modulus (G”) obtained by the rheometer
and penetration hardness measured by the instrumented penetration for
cement pastes with w/c = 0.4. Vertical separation o f H points is the
resolving power o f the apparatus.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
158
However when the pore solution is saturated, the dissolved ions react with water to
generate solid hydration products. The more solid products the cement paste has, the
more solid-like the behavior, resulting in a decrease in the viscous loss modulus. The
decrease o f the viscous loss modulus is well matched with the beginning o f the
penetration hardness development, regardless o f the w/c ratio.
A linear relationship can be established between the elastic storage modulus and
the penetration hardness in cement paste, as shown in Figure 4-59. To establish this
relationship, the penetration response was chosen from the time when the viscous loss
modulus decreases because the penetration hardness during the particle dissolution (or
more fluid-like) stage could be inaccurate because o f the finite resolving power of the
apparatus. The resultant correlation equations are:
w/c = 0.3 => G’ (kPa) = -30.9+ 0.174 H (k P a ):R 2 = 0.9650
(Eq. 4-19)
w/c = 0.4 => G’ (kPa) = -15.6 + 0 .181 H (kP a): R2 = 0.9167
(Eq. 4-20)
The correlation coefficients o f in the equations of two levels o f w/c ratio are
similar, about 0.18.
This indicates that development of the storage modulus and the
penetration hardness can be successfully correlated regardless of w/c ratio. However, the
intercept values are different. One o f possible reasons for this difference is the resolution
o f the instrumented penetration test. If a better resolution load cell were used, a more
precise correlation would result. As a result, the instrumented penetration test is able to
extend the rheology study to greater degrees of hydration.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
159
100
— w/c=0.3: y = -30.9 + 0.174x RT = 0.9650
Penetration Hardness (kPa)
80 -
" w/c=0.4: y = -15.6 + 0.181x 1^= 0.9167
40 -
20
-
0
100
200
300
400
500
G' (kPa)
Figure 4-59 Correlation between the elastic storage modulus (G’) obtained by the
rheometer and the penetration hardness measured by the instrumented
penetration test for cement pastes with w/c = 0.3 and 0.4.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
160
4.5.2. Influence o f Mixing on Rheology
Since the beginning o f cement hydration is dominated by the concentration o f
ions dissolved from cement particles, reducing the particle size (or increasing contacting
surface area) may accelerate the reactions. Thus, severe mixing (i.e. blender mixing)
breaks particles finer and shows accelerates hydration. For this purpose, two types o f
mixers, a paddle mixer and a laboratory blender (model 31BL91-7010, Waring, New
Hartford, CT), with varied mixing time, were employed. The paddle mixing time was 10
minutes, whereas for the blender, cement paste was mixed for 0.5, 2, or 4 minutes after 3
minutes o f paddle mixing.
To prevent any temperature rise due to blender mixing,
cooling water was passed beneath the blender canister.
The temperatures of cement
pastes after mixing were about 23°C for the paddle mixing and about 21°C for blender
mixing.
Blender mixing accelerates hydration o f cement paste, resulting in faster
development o f penetration hardness, as shown in Figure 4-60.
As mixing time
increased, the hardness curve became steeper. It is thought that links initially formed
between particles were broken by the more severe mixing, generating more sites for
hydration.
The acceleration of cement reactions caused by mixing was confirmed.
Pastes mixed differently have similar degree of hydration at the same hardness levels in
spite o f the variation in time to reach the levels, as shown in Figure 4-61.
However, the various mixing methods, as shown in Figure 4-62, did not
significantly influence the degree of hydration at later stages o f hydration.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
161
cd
a.
C/3
C/3
a>
a
"O
cd
X
15 -
10
Paddle 10 minutes
B lender 0.5 m inutes
B lender 2 minutes
B lender 4 minutes
-
c
o
«s-
-*—>
4>
(a) w /c = 0.4
C
4)
Q*
60
0
120
180
240
300
360
Time after Mixing (Minutes)
cd
cu
C/3
C/3
15-
4>
c
Paddle 10 minutes
B lender 0.5 m inutes
B lender 2 minutes
B lender 4 minutes
T3
cd
K
c
10-
o
'SUi
4)
(b) w /c = 0.5
C
4)
CU
0
60
120
180
240
300
360
420
480
Time after Mixing (Minutes)
Figure 4-60 Effects o f mixing method and time on the penetration hardness
developments o f cement pastes with (a) w/c = 0.4 (b) w/c = 0.5. The tests
were performed at room temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
162
0.16
w/c=0.5 at 13 M Pa
Degree of Hydration (a)
0.14
0.12
w /c=0.4 at 13 M Pa
0.1
0.08
0.06
w/c=0.5 at 1 M Pa
0.04
w /c=0.4 at 1 M Pa
0.02
a
Blender 4 minutes
c
B lender 0.5 minutes
o
Paddle 10 minutes
0
100
200
300
400
500
Time after Mixing (minutes)
Figure 4-61 Relationships between the times to reach hardness levels of 1 or 13 MPa
and the degree o f hydration o f cement pastes with w/c = 0.4 and 0.5. The
tests were performed at room temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
163
0.8
Degree of Hydration (a)
0.7
0.6
0.5
0.4
0.3
- w /c=0.4 Paddle 10 minutes
- w /c=0.4 B lender 0.5 minutes
0.2
- w /c=0.4 B lender 4 minutes
■♦■■■ w /c=0.5 Paddle 10 minutes
w /c=0.5 B lender 0.5 minutes
—-ir— w /c=0.5 B lender 4 minutes
0.1
I I I I I
0
10
100
1000
Time after Mixing (Hours)
Figure 4-62 Effect o f mixing on the later stage o f hydration for cement pastes with w/c
= 0.4 and 0.5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
164
4.5.3. Penetration Rate Effects on Cement Rheology
As mentioned earlier, the hardening rate o f cement pastes and mortars is a
function of the penetration rate. However, it is thought that the external force induced by
the penetrometer could significantly influence rheology, not cement chemistry. Thus, the
investigation o f the further penetration rate effects could be useful for studying rheology.
A continuous variable penetration rate experiment, using one specimen, was
conducted to measure the penetration rate effect. After penetration hardness began to
develop, the penetration rate was increased by a factor o f 10 (6.4 mm/hr => 64 mm/hr).
After the hardness response had stabilized, the penetration rate was reduced to its initial
value (6.4 mm/hr). This was repeated several times so the effect at various hardness
levels could be observed.
The results o f this experiment are shown in Figure 4-63. Clearly, a change in
penetration rate has a significant effect on the measured hardness value. Increasing or
decreasing the rate by a factor o f 10 causes the measured hardness to increase or decrease
about 10%.
As the penetration rate decreased, the hardness also decreased, but
concavely, for a few seconds. This phenomena is related to the material relaxing (Brett,
1996). Increasing the penetration rate causes an instant linear rise followed by a convex
rise in the measured hardness, which indicates a sort o f yielding process occurring. The
presence of the rate effect indicates that this test is capable o f measuring the nonNewtonian behavior o f cement paste.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Penetration Hardness (MPa)
165
10
-
76
78
80
82
84
86
88
90
Tim e after M ixing (M inutes)
Rate (mm/hr)
6.4 => 64
64 => 6.4
6.4 => 64
64 => 6.4
6.4 => 64
64 => 6.4
Before (MPa)
3.53
5.02
5.02
12\
10.55
14.32
After (MPa)
4.19
4.41
5.42
6.36
11.96
12.91
%
+ 19
-12
+8
-12
+13
-10
Figure 4-63 Effects o f changing penetration rate on the penetration hardness for cement
paste o f w/c = 0.3 cured at 60°C by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
166
Figure 4-64 shows the effects o f changing penetration rate on hardening curves
with different paste specimens cured at room temperature. The test began after mixing
with the rate being 6.4 mm/hr, and the rate was raised to 64 or 640 mm/hr at 220 minutes
after mixing. The variations in the hardness before changing the rate could be caused by
the variable nature o f cement hydration, but are not significant for this purpose. The
hardness curves increased quickly, responding to the rate change. The faster the rate, the
greater the increment. Relaxation occurs post-increase, being different from previous
results at elevated temperatures. It is thought that there is enough time for redistributing
solid phase material because of the slower reaction rate at room temperature than at
elevated temperatures. The following bumping on the faster rate indicates that cement
pastes hydrating in the acceleration period are still heterogeneous mixtures o f soft and
hard phases, which are unable to smoothly move at this rate. As expected, the slope of
the penetration hardness was shifted upward when the rate increased because of the
positive coefficient o f the penetration rate in the hardening rate equation of cement paste,
mentioned in section 4.2.2.
It was reported that delayed onset tests until the continuous test reached 1 MPa,
does not affect test results (Croft, 1996). However, the response o f the longer delayed
onset test (at 5 MPa o f the continuous test) was in disagreement, shown in Figure 4-65.
The hardness at the lowest rate of 6.4 mm/hr o f the delayed onset test developed slowly
but did not reach the hardness for the continuous run, while at the highest rate of 640
mm/hr hardness developed very rapidly to reach immediately the continuous hardness.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
167
9
— Continuous: 6.4 m m /hr
—- 6.4 m m /h r => 640 mm/hr
— 6.4 m m /hr => 64 m m /hr
Penetration Hardness (MPa)
8
7
6
.+
■XT
R ate Changed
4
3
216
218
220
222
224
226
Time after Mixing (Minutes)
Figure 4-64 Influences o f variable penetration rates on the hardening curves for cement
pastes o f w/c = 0.3 cured at room temperature. The hardness was measured
continuously but with rate change.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
168
12
Continuous: 6.4 m m /hr
-*— 640 mm/hr
* — 64 m m /hr
^ — 6.4 m m /hr
Penetration Hardness (MPa)
10
8
6
4
2
0
215
220
225
230
235
240
245
Time after Mixing (Minutes)
Figure 4-65 Influences of variable penetration rates on the hardening curves for cement
pastes o f w/c = 0.3 cured at room temperature. The hardness was measured
with delayed onset.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
169
The hardness o f 64 mm/hr developed between those rates, but did not reach the
continuous hardness.
Another interesting aspect is a distinct change in slope o f the
hardness curve at the 64 mm/hr rate, which does not exist on the hardness curve o f 640
mm/hr. This feature is also thought to exist, but is hard to distinguish, in the case o f the
6.4 mm/hr rate.
Figure 4-66 illustrates the compaction mechanism by the penetrator.
When a
material is fluid-like, the force driven by the penetrator is distributed through the media
by water movements indicated by white arrows (see Figure 4-66 (a)). Cement paste
behaves fluid-like from immediately after mixing, and it is still fluid enough to deform
readily before the initial setting time. Thus, the delayed onset test does not show any
difference from the continuous run.
However, after the initial set time, the paste will lose this fluidity continually as
hydration proceeds, to be gradually compressed by the external force o f the continuous
running penetrator. Figure 4-66 (b) roughly illustrates this compressed area in shading.
The spreading angle and distance may be dependent of penetration rate and the degree of
solidity.
With the delayed onset test, the hardness development is expected to be
dependent on the penetrator rate because o f differing compressed area. The fast rate can
compress media sooner to obtain fast hardness development (see Figure 4-66 (c)). On the
other hand, the slow rate requires more time to compact the media (see Figure 4-66 (d)).
Therefore, the hardness develops slowly.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
170
Fluid-like media
Continuous
(a)
Penetrator
Solid-like media
(b)
Continuous
Solid-like media
Delayed onset
(C)
Fast R ate
Solid-like media
Delayed onset
(d)
S low Rate
Figure 4-66 Force distributions through fluid- or solid-like media. White arrows
indicate fluid movements against driving penetrator. Shaded area indicates
compressed region by penetrator.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTERS
Influence of Microwave Curing on Hydration and Properties
5.1. Curing Temperature Effects on Hydration o f Cement Paste
5.1.1. Degree o f Hydration
The results confirmed that the rate o f hydration is accelerated at elevated
temperature, particularly in the first hours after mixing. Figure 5-1 shows the degree of
hydration of each elevated temperature-cured sample at 0.4 w/c ratio obtained by loss-onignition. The differences in the degree o f hydration among RT, 40°C and 60°C-cured
specimens becomes smaller and smaller during hydration and is not significant after 1
day. The degree o f hydration o f the 90°C-cured specimen is consistently higher than the
others up to 28 days.
The volume fractions of the capillary porosity, calculated from the degree of
hydration by using Eq. 3-5, are shown in Figure 5-2. The accelerated hydration due to the
elevated temperature causes a reduction in the amount o f the capillary porosity. The
capillary porosity o f the 90°C-cured specimen is also lower than the others. Considering
the evaporation o f water due to high temperature, the real w/c ratio could be lower than
the initial ratio. The amount of capillary water then becomes lower than the plotted
values, especially for the higher temperature such as 90°C.
171
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
172
0.8
0.7 -
0.6
-
0.3 -
0.2
cn
-
1
10
100
1000
Time after Mixing (Hours)
Figure 5-1 Degree o f hydration calculated by loss-on-ignition for cement pastes with w/c
= 0.4 cured at various temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
173
0.6
X
40°C
0.5 -
60°C
•
*90°C
0 .4 -
cap 0.3 -
3.
0 .2 -
1
10
100
1000
Time after M ixing (Hours)
Figure 5-2 Volume fraction of capillary porosity (calculated by Eq. 3-5) versus time for
cement pastes with w/c = 0.4 cured at various temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
174
5.1.2. Microstructure Changes Monitored by Impedance Spectroscopy
When cement paste undergoes temperature changes during monitoring impedance
spectra, the normalization should be performed with careful consideration o f temperature.
McCarter found that the bulk conductivity o f fully hydrated cement mortar increased with
temperature, which indicates that temperature could affect ion conductivity in the pore
solution, without making any change in the structure o f the pore network (McCarter,
1995).
Therefore, the bulk conductivity should be normalized by the pore solution
conductivity at exactly the same temperature as the bulk. If the temperature effects for
the normalization are ignored, an unexpected drop on the conductivity line can be seen at
the point o f temperature decrease.
Figure 5-3 shows the change o f the normalized conductivity of each specimen at
early times. The normalized conductivity decreased with increasing hydration time and
became lower more quickly as the processing temperature increased.
This can be
explained by accelerating hydration due to heating. However, for 90°C, the values were
almost the same as those o f 60°C, which indicates no further increment on the normalized
conductivity in spite of more hydration. This is the first evidence of different behavior of
the 90°C-cured specimen. Figure 5-4 is a full scale plot of the normalized conductivity
versus degree o f hydration.
The normalized conductivity decreases as the degree of
hydration increases because the amount of free water is reduced to make more hydration
products and the electric pathway becomes more tortuous. The normalized
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
175
10
Cooling
a (T) / a (T)
Heating
'
“® ^3""®" ®*
^ 4L w
•
0
2
4
6
8
10
12
Time after Mixing (Hours)
Figure 5-3 Normalized conductivity in early age o f hydration for cement pastes with w/c
= 0.4 cured for 6 hours at various temperature by microwaves. Pore
conductivity was modified by actual specimen temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
176
10°
cj
/ a
lO*1
n
10
o
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Figure 5-4 Full scale of normalized conductivity versus degree o f hydration for cement
pastes with w/c = 0.4 cured for 6 hours at various temperature by
microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
177
conductivity is lower as the curing temperature increases up to 60°C over the range o f a =
0.2 to 0.5. This implies a difference in microstructure. Beyond a = 0.5, these differences
disappeared. This indicates that nearly identical hydration structures were formed at each
temperature. However, the normalized conductivity o f the 90°C-cured specimen does not
agree with the others. It is consistently higher than that of RT and, moreover, remains
higher at the later ages.
For better understanding of the temperature effects on the microstructure, the
relationship
(Eq.
C l)
was
employed to
calculate
the
pore
structure
factor
(P; P = CT/(a0O cap)) which indicates the degree of tortuousity in the pore network. Figure
5-5 exhibits the pore network formation in terms o f the electrical response and similar
trends can be found. Up to 60°C, P decreases sooner as the curing temperature increases,
ultimately converging on the RT line at later ages. The pore network can be formed faster
due to heating at early age, but may not be influenced changed by the curing temperature
at later age. The 90°C specimen also shows rapid formation o f pore structure but has a
higher P (lower tortuosity) at the later ages. It can be predicted that the 90°C specimen
has more open and coarser pore structure that is detrimental to the strength and durability
of the cement pastes. The different behavior implies different hydration products or at
least a different structure o f the pore network at this temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
178
P
- » •>
' ~ ~E -------
(3 = cj / a O
10
"
l
10
100
1000
Time after Mixing (Hours)
Figure 5-5 Pore structure factor (p) for cement pastes with w/c = 0.4 cured for 6 hours at
various temperature by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
179
5.1.3. Pore Solution Analysis
The conductivity o f the pore solutions represents the amount o f mobile ions,
shown in Figure 5-6. The solid lines represent the predicted values o f the conductivity of
the pore solution with the dashed 95% confidence bands superimposed. The actual data
points are well matched with the predicted lines. The regression line o f the conductivity
o f 90°C was lower than other temperatures. If the pore solution conductivity was only a
function o f the degree o f hydration, specimens of the same degree o f hydration, in spite of
different curing temperature, should have the same pore solution conductivity, resulting
in the same formation and amount of hydration products. Therefore, The different pore
solution conductivity behavior at 90°C is an evidence of a different hydration mechanism
at this temperature. However, the distinction of the 90°C regression line could not be
seen on the pH plot and the sum o f the alkaline ion concentrations (Na+ and K+) plot,
shown in Figures 5-7 and 5-8, respectively. As expected, Na+, K+ and OFT concentrations
increase with the degree o f hydration for all curing temperatures.
The concentrations of sulfate and aluminate also show the dissimilarity o f 90°C
curing, shown in Figures 5-9 and 5-10. Cement pastes cured at various temperatures
except at 90°C exhibit similar behaviors.
The sulfate concentrations continuously
decrease and go to a low value after a degree of hydration of 0.45, while the aluminate
concentrations decrease initially and then increase again at later age. The sharp reduction
o f the sulfate ions is related to the formation o f ettringite with consumption o f mobile
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
180
10
(S/m)
8
6
4
2
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Figure 5-6 Conductivity o f the pore solution for cement pastes with w/c = 0.4 cured at
various temperature by microwaves with regression analysis. Dashed lines
show the 95% confidence intervals.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
181
RT
40°C
13.8
Best fit for the others
60°C
90°C
13.6
13.2
-A,
A'
Best fit for 90°C
12.8
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
a
Figure 5-7 pH o f pore solution for cement pastes with w/c = 0.4 cured at various
temperature by microwaves with regression analysis. Solid lines indicate the
individual best fits for 90°C and the others. Dashed lines show the 95%
confidence intervals (the longer dash is for only 90°C, while the shorter dash
is for the others).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
182
600
RT
40°C
500
60°C
Best fit for 90°C
Na+K (mMol/L)
90°C
400
<n
300
200
Best fit for the others
100
0
0.1
0.2
0.4
0.3
0.5
0.6
0.7
a
Figure 5-8 Sum o f alkali ion concentration in pore solution for cement pastes with w/c =
0.4 cured at various temperature by microwaves with regression analysis.
Solid lines indicate the individual best fits for 90°C and the others. Dashed
lines show the 95% confidence intervals (the longer dash is for only 90°C,
while the shorter dash is for the others).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
183
Sulfate Concentration (mMol/L)
100
80 -
60 -
40 -
20
-
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
a
Figure 5-9 Sulfate concentrations o f the pore solution versus degree o f hydration for
cement pastes with w/c = 0.4 cured at various temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
184
0.15
o
B
c
o
•4—
>
e
<u
o
r"
O
U
« 0.05 c
S .
S
3
<
0
0.1
0.2
0.4
0.3
0.5
0.6
0.7
a
Figure 5-10 Aluminate concentrations o f the pore solution versus degree of hydration for
cement pastes with w/c = 0.4 cured at various temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
185
aluminate ions, which occurs before reaching 0.4 of degree of hydration or less than 10
hours after mixing.
The later increase of aluminate concentration can be explained.
Since aluminate ions are still dissolving from C3A, but the sulfate source in the solution is
used up, the ettringite formation will be stopped and then the dissolved aluminate ions
will exist in the solution. However, the cement pastes cured at 90°C show that sulfate ion
concentration was not much changed from the initial value during the whole of hydration,
and a lot o f sulfate ion still exists in the pore solution. The aluminate ion concentration
was lower at the early stage and never rises up at later degree of hydration. Over 70°C,
ettringite formation is prohibited and sulfate and aluminate ions are mainly incorporated
into the C-S-H, though some o f the sulfate enters the pore solution (Taylor, 1990). Since
some gypsum still existed and no ettringite peak was found in the 90°C-cured sample by
x-ray diffraction at 7 days, shown in Figure 5-11, sulfate concentration in the pore
solution may still be high enough to prevent dissolution o f sulfate from gypsum. This
may be explained by evaporation of water by too high curing temperature, which leads to
sample self-desiccation.
Any differences in other hydration products among cement
pastes cured at various temperatures were not detected by the x-ray diffraction analysis.
Figures 5-12 and 5-13 show CaO and SiOi concentrations of the pore solution for
cement pastes with w/c ratio of 0.4 at various curing temperatures.
As hydration
proceeds, CaO concentrations were reduced and Si0 2 concentrations were increased,
regardless o f curing temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
186
ICH+C S
0 Gypsum
C-S-H+C S4C S I
CH !
r ■i i
c si
1
C-S-H
;
CS
90 C
CH
CH,
iCH
60°C
40 C
RT
10
15
20
25
30
35
40
45
50
55
60
26
Figure 5-11 X-ray diffraction patterns for cement pastes with w/c = 0.4 cured at various
temperature at 7 age days.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
187
CaO Concentration (mMol/L)
25 -
20
-
15 -
10
-
0
0.1
0.2
0.4
0.3
0.5
0.6
0.7
a
Figure 5-12 Calcium concentrations o f the pore solution versus degree o f hydration for
cement pastes with w/c = 0.4 cured at various temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
188
0.5
■<— *
c
<D
O
oC
0.2
-
U
At
O
0
0.1
0.2
0.4
0.3
0.5
0.6
0.7
a
Figure 5-13 Silicate concentrations o f the pore solution versus degree o f hydration for
cement pastes with w/c = 0.4 cured at various temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
189
F igure 5-14 is a m etastable phase diag ram fo r th e C a 0 - S i 0 2 -H 20 system b a se d on
the h y d ratio n o f C 3 S (Jennings, 1986; G artn er a n d Je n n in g s, 1987). T his diag ram w as
m ad e by c o lle ctin g m an y published d a ta fo r th e c o n c en tra tio n o f lim e an d s ilic a in
so lu tio n s in c o n ta c t w ith C -S-H , prepared in se v era l lab o rato ries o v e r a period o f 50
years.
If m eta sta b le eq u ilibrium w ere established, th e c o n c en tra tio n d a ta w ou ld fall on
c u rv e A a n d so m e tim e s o n curve B, w ith relativ ely fe w in b etw een.
W hen u nhydrated
C 3 S rem ains, d a ta fall o n curve B; w h ile w h en C 3 S a re c o m p le te ly consum ed, d a ta fall on
cu rv e A.
C u rv e A c a n be associated w ith C -S -H (I), w h ic h is structurally related to
to b erm o rite (C 5 S 6 H 5 ) a n d has a C a:Si ratio in the a p p ro x im a te range 0.8 to 1.3. H ow ever,
cu rv e B is m o st u n lik e ly to represent th e m etastable so lu b ility o f C 3 S. It m eans th at C-S-
H form ed o n c u rv e B m ig h t be a m ixture o f to b o m o lite -lik e (C -S -H (I)) and je n n ite -lik e
(C -S -H (II); C 9 S 6 H 1 1 ) form s o f C -S -H (Jennings, 1986). A n o th e r explanation o f c u rv e B
is th at it rep re sen ts a quasisteady state in w hich th e rate o f d isso lu tio n o f OH* ions from
C 3 S and th e rate o f precip itatio n o f C -S -H are equal (B a rre t a n d B ertrandie, 1988).
The equations for these curves are as follows:
1-2.03697
Curve A: [S i02] = 952.503 * [Ca]
(Eq. 5-1)
Curve B: [Si02] = 16825.3 * [Ca]1-1.98724
(Eq. 5-2)
CH Curve: [S i02] = 8.02075e*25 * [Ca]*17917
(Eq. 5-3)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
190
10'
SiC>2 Concentration (jiMol/L)
A queous Phase
+ SiO ( H O )
RT
40°C
C-S-H
+ Aqueous Phase
60°C
90°C
Curve B
f CH
+ C-S-H
-+- Aqueous
Phase
101
Curve A
10°
Ca(OH)
>i
A queous Phase
0
5
10
15
A queous Phase
+ CH
20
25
30
CaO Concentration (mMol/L)
Figure 5-14 Concentrations o f S 1O2 versus CaO of the pore solutions for cement pastes
with w/c = 0.4 cured at various temperatures by microwaves superimposed
on the metastable CaO-Si0 2 - H2O phase diagram.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
191
T h e c o n c en tra tio n d a ta in p o re so lu tio n s o b tain e d fro m ro o m tem perature an d
m ic ro w a v e c u re d c e m e n t p a ste s a re su p e rim p o se d o n th e m eta sta b le phase diagram . A t
ea rly h y d ratio n , d a ta fall o n curve B , th en m o v e to cu rv e A as h y dration proceeds. T he
m o v e m e n t tra c es o f all tem p eratu res are c lo se e n o u g h to b e ro u g h ly identical, in spite o f
so m e sc atter.
H ow ever, th is p lo t ca n n o t d isc lo se th e effects o f tem perature o n the
in d iv id u al C aO /SiC >2 ratio s in th e C -S -H b e c au se th ere is no in fo rm a tio n about the degree
o f h y d ratio n .
Figure 5-15 shows the changes o f the Ca 0 /Si 0 2 ratio in the pore solution as a
function o f degree o f hydration. Those data points which fall near curve A and B are
identified on the plot. For less than 0.2 of degree o f hydration, all data fall near curve B,
regardless o f the temperature. This corresponds to the high reactivity of unhydrated C3S
at early hydration stage. The range o f 0.2-0.4 corresponds to the area between the two
curves. Above 0.4, the data up to 60°C fall near curve A, which may indicate that all C3S
have been hydrated or C3S particles are covered by dense hydration products, being
retarded from further hydration. However, 90°C shows some exceptions. The data still
exist in the middle at a degree o f hydration of about 0.6. It is thought that 90°C might
alter or influence the Ca 0 /Si 0 2 ratios in C-S-H, and this is another abnormal behavior of
the 90°C paste.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
192
1000
RT
Near Curve B
40°C
60°C
90 C
100
o
O
«s
u
Near Curve A
0
o.i
0.2
0.3
0.4
0.5
0.6
0.7
a
Figure 5-15 Ca0/Si02 ratio o f the pore solution versus degree o f hydration for cement
pastes with w/c = 0.4 cured at various temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
193
5.2. Temperature Effects on Compressive Strength of Cement Mortar
5.2.1. Curing Conditions
The
three
thermocouples
in preliminary
specimens
demonstrated
good
temperature uniformity. Figure 5-16 shows the temperature profiles measured by these
thermocouples at the set temperatures of 40, 60 and 80°C.
At the beginning, the
temperature differences were about 1-4°C among the three thermocouples, being larger at
the higher temperature. However, these differences were considered to be too small to be
significant.
Overheating was not found.
Thus, all o f the specimens could be cured
isothermally at the proposed curing temperature used in this study.
5.2.2. Two-Hour Microwave Curing Effects on Strength
Two hours was arbitrarily chosen for the microwave heating time. Figure 5-17
shows the 28-day compressive strength of Type I and 50% GGBFS mortars under two
hours microwave curing at various temperatures. 95% confidence intervals based on
three identical specimens are also superimposed on the average value. At 40°C, the
compressive strengths o f both cement mortar and 50% GGBFS mortar are similar to
those cured at room temperature, considering the confidence intervals. On the contrary,
specimens cured at 80°C have only half o f the room temperature strength for both. In the
case o f 60°C, specimens show 80% strength of the room temperature cured specimens.
However, it was thought that two hours may be unnecessarily long for heating, and that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
194
100
90
Temperature (°C)
80‘
70
6
0
=
50
40
30
20
0
30
60
90
120
Time during microwave heating (minutes)
Figure 5-16 Temperature history for cement mortars measured by thermocouples placed
(1) 22 mm (2) 37 mm (3) 67 mm from the bottom o f the specimen at curing
temperatures o f 40,60, and 80°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28-day compressive strength (MPa)
195
Type I Mortar
50% GGBFS Mortar
102%
RT
40°C 60°C 80°C
RT
40°C 60°C 80°C
Figure 5-17 28-day compressive strength for room temperature cured and microwave
cured cement (Portland Type I) and 50% GGBFS mortars at 40, 60, and
80°C with sandrcement (of total reactive solid) = 2:1, w/c (or w/r) = 0.4 and
heating time = 2 hours. The error bars indicate 95% confidence intervals
based on three specimens.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
196
less strength degradation might be realized for shorter heating times.
5.2.3. Determination o f Curing Time for Enhancing Strength
For the variable heating time experiments, three levels o f the penetration
hardness, namely 1, 5, and 13 MPa were arbitrarily chosen to establish the heating time
for all curing temperatures. Table 5-1 indicates the time needed to reach the given levels
o f penetration hardness for cement mortar at the various curing temperatures. Specimens
for strength tests were cured according to Table 5-1.
Table 5-1 Microwave heating time (minute) o f cement mortars to reach given levels of
penetration hardness at various curing temperatures.
Penetration
Hardness (MPa)
1
5
13
40°C
76
100
120
Curing Temperature
60°C
48
62
73
80°C
39
50
58
In other words, for example, at 1 MPa and 40°C, the specimen was heated 76 minutes at
40°C, and then allowed to be cooled naturally to room temperature. The temperature
histories were shown in Figures 5-18 and 5-19.
Figure 5-20 shows the 28-day strength o f cement mortars heated at 40, 60, and
80°C until the penetration hardness reached the stated values (see Table 5-1). The results
o f curing at 40°C or 80°C are very close to the 2-hour curing results on Figure 5-17. The
strength values are unchanged under 40°C curing temperature until the hardness reached
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
197
100
40°C (1 M Pa)
40°C (13 MPa)
60°C (1 M Pa)
80-
60°C (13 MPa)
80°C (1 M Pa)
oU
80°C (13 MPa)
<
D
S s—
D
O,
60-
s
<D
H
40"
0
60
120
180
Time after Mixing (minutes)
Figure 5-18 Temperature history for cement mortars during heating at various
temperatures and time by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
240
198
100
40°C (1 M Pa: at 113 min)
40°C (13 M Pa: a tl5 0 min)
60°C (1 M Pa: at 78 min)
60°C (13 M Pa: at 103 min)
80°C (1 M Pa: at 69 min)
O
80°C (13 M Pa: at 88 min)
<D
c3
S -t
(U
O.
£U
H
i
o
60
i
|
120
i1 - i
i
i
i
|
i " i ■
"r ■
i
180
Tim e after M icrowave Curing (m inutes)
Figure 5-19 Temperature history for cement mortars after heating at various
temperatures by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
\
240
199
80
cd
Oh
I
1 M Pa
I
5 M Pa
j
13 M Pa
60 -
RT Curing
00
a
<
L>
J
—
■4—
»
art
<U
>
40 -
'a i
Oi
<U
D,
£
o0
&
20
-
*o1
OO
<N
0
40°C
60°C
80°C
Figure 5-20 28-day compressive strength for room temperature cured and microwave
cured cement mortars at 40, 60, and 80°C with sandrcement = 2:1 and w/c =
0.4. Specimens were heated until the penetration hardness reached 1, 5, or
13MPa.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
200
13 MPa, but at 80°C they are drastically decreased. However, the behavior at 60°C curing
is different.
When specimens were cured at 60°C until the penetration hardness reach 1 MPa,
or 48 minutes curing time for cement mortar, their 28-day strength was unchanged
compared with that o f room temperature-cured specimens, which are indicated by the
dotted line on the same figure. As the curing time was longer than 48minutes (or I MPa),
the strength decreased. Although the strength decreased to 88% o f room temperature at
13 MPa (73 minutes), it still is higher than that o f the 2-hour cured specimens (79%;
Figure 5-17).
From this study, the best curing times for cement mortar without
degradation o f 28-day compressive strength are 120 minutes (or 13 MPa) at 40°C and 48
minutes (or 1 MPa) at 60°C.
5.2.4. Final Setting Time of Cement Mortar
Since the penetrometer load cell was incapable o f loads equivalent to the ASTM
C403 final setting criterion (27.6 MPa), the times to reach this hardness were estimated
by extrapolation. The penetration test was started at the beginning o f the heating time to
continue until 13 MPa o f the penetration hardness, regardless o f whether the microwaves
were on or off and the specimen was cooling down. The penetration hardness curves
were extrapolated from 13 MPa to 27.6 MPa. The extrapolated times are shown in Figure
5-21. The casting time (30 minutes for all) indicates the time need for mixing, casting,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
201
§
I
Cooling Tim e
Heating Tim e
Casting Tim e
60°C
1 MPa
40°C
13 MPa
40°C
1 MPa
RT
Curing
0
1
2
3
4
5
6
7
Time after Mixing (Hours)
Figure 5-21 Final setting times for room temperature cured and microwave cured cement
mortars with sand:cement = 2:1 and w/c = 0.4.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
202
and placing inside the microwave chamber.
To reach the final setting time, room
temperature cured cement mortars required at least 6 hours, but the time was reduced to
half o f this at 40°C and, to just over one-third of this at 60°C. Therefore, the appropriate
microwave curing technique enables a significant reduction in the time to demolding.
5.2.5. Strength Tests for Blended Mortar
The curing times were varied due to the nature o f the pozzolans, as shown in
Table 5-2.
Table 5-2 Microwave heating time (minute) o f mortars containing reactive admixtures
to reach given levels of penetration hardness at various curing temperatures.
50% GGBFS
10% Silica
Fume
20% Class F
Fly Ash
Penetration
Hardness (MPa)
1
5
13
1
5
13
1
5
13
40°C
105
141
172
60
90
109
92
122
147
Curing Temperature
60°C
53
72
88
39
56
66
61
78
89
80°C
39
52
61
46
54
43
54
62
GGBFS and fly ash blended mortars need longer microwave heating time for the
target hardness levels than cement mortar for all temperatures, but silica fume-blended
mortars needed shorter times.
Figure 5-22 shows the 28-day compressive strength results o f blended mortars.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
203
1 MPa
^
5 M Pa
C
13 MPa
- RT Curing
20
< 50% GGBFS > < 10% Silica Fume x 20% Fly ash (F) >
Figure 5-22 28-day compressive strength for room temperature cured and microwave
cured 50 wt% GGBFS, 10 wt% silica fiime, and 20 wt% class F fly-ash
cement mortars with sandrtotal reactive solid = 2:1 and w/r = 0.4 cured at
40, 60, and 80°C by microwave. Mortar specimens were heated until the
penetration hardness reached to 1, 5, or 13MPa.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
204
GGBFS or silica fume raised the strength when specimens were cured at room
temperature, but fly ash decreased it.
All blended mortars show no degradation of
strength when cured at 40°C until the penetration hardness reaches 13 MPa. Additionally,
adding silica fume raised the strength more than for room temperature curing. At 80°C
curing, strength was greatly decreased, regardless of the type of admixtures. At 60°C
curing, the strength o f the specimen heated until 1 MPa is similar to the room
temperature-cured one, but deceased as heating time increased for the GGBFS and silica
fume cases. This trend is less when fly ash is added. Therefore, heating times associated
with penetration hardness values o f 13 MPa at 40°C or 1 MPa at 60°C are the maximum
times for which the 28-day compressive strength is not degraded with or without
admixtures. Since times to reach a certain penetration hardness are different for each
admixture, because o f its nature, the penetration hardness test made it possible to
establish the optimum curing time regardless of the amount or type o f reactive
admixtures.
The potential usefulness o f microwave heating o f concrete has been proven, but
the energy efficiency is still low because microwave energy is consumed continuously or
intermittently during curing.
process.
Figure 4-2 suggest a possible highly energy efficient
The temperature profiles show that insulation can hold the temperature
considerably without further microwave energy. This holding time is longer than the
optimized curing times in Table 5-1. The total time of microwave energy application was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
205
only a few minutes. Thus, utilizing proper insulation can reduce the consumption of
microwave energy. Consequently, Figure 5-23 suggests a plausible microwave process
for cement mortar manufacturing.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
206
STEP I:
Mixing & Casting Stage
O
STEP II:
Microwave Heating Stage
STEP III:
Insulation Stage
O
STEP IV:
Cooling Stage
Figure 5-23 Procedures for microwave curing process on cement mortar manufacturing.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 6
Conclusions
6.1. Conclusions
This project has shown that an instrumented penetration test is a viable tool to
study hydration mechanisms o f cement-based materials and that utilizing microwave
energy has a potential application in cement-related areas. The instrumented penetration
test provides useful information about the early hardening process for cement-based
materials, including setting time.
The in-situ and continuous measurements enable
reduction o f experimental errors as well as labor and other costs. Additionally, it can be
used to determine the optimum curing time of microwave treatment on cement products
without degradation in strength.
Microwave heating enhances curing of cement paste and mortar significantly.
The heating rate is accelerated thermally, not by athermal microwave effects.
The
apparent activation energy o f cement hardening is affected by the w/c ratio, but the
hardening rate is not a simple function of temperature.
D-optimal experimental design and multiple correlation analysis revealed that the
most significant factors in the hardening rate models are hardness level, temperature, and
water-to-cement (or total reactive solid) ratio. The hardness level is closely associated
with the degree o f hydration, and temperature and w/c ratios also influence this
207
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
208
relationship.
Generally, raising temperature or reducing water content make the
hardening rate greater and the initial setting time shorter. Adding sand influences the
hardening rate rheologically rather than chemically, which implies cement hydration is
the main reason for hardening. Also, the hardening rate is varied with adding pozzolanic
materials, depending on their type and amounts. Silica fume tends to increase the rate
due to its extreme fineness, while GGBS decreases it because of its late hydration. Class
F fly ash shows a two-fold temperature dependence.
Compared with the conventional rheometer tests, the instrumented penetration test
extends the investigation o f the rheological behavior o f cement paste to higher levels of
hydration.
Also, this test makes it possible to explore the rheology o f cement-based
materials in various ways from other type o f materials, for instance mortar, to changing
force conditions like delayed onset test. It was found that the more severe mixing makes
early hydration faster, but has little influence on the ultimate degree o f hydration.
Higher temperature enhanced the hydration of cement, as mentioned earlier, and
altered several properties o f the pore solution. Considering the level o f each degree of
hydration, any additional effects on the pore solution properties could not be expected.
However, when the temperature was 80°C or 90°C, abnormal behavior was found in
normalized bulk and pore solution conductivity, pore structure factor (P) and sulfate and
aluminate concentrations.
This abnormal behavior was also detected in hardness
development and 28-day strength. The possible reasons for these differences may be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
209
associated with generating more porous structure or the restriction of ettringite formation
due to too high temperature.
Curing temperature and time are important parameters for improvement o f 28-day
strength o f microwave cured pure cement or blended mortars.
The instrumented
penetration test can be used to determine the optimum microwave heating time of
mortars. Additionally, microwave curing with feedback temperature control significantly
reduced the final setting times of mortars.
The optimum curing conditions are at
penetration hardness levels o f 13MPa at 40°C or IM Pa at 60°C regardless of type o f
reactive admixtures. However, 80°C was unsuitable for microwave curing. Therefore,
microwave energy is potentially useful for processing cement-based materials.
6.2. Suggestions for future work
This project has provided a solid background for investigating hardening process
and microwave heating effects on cement-based materials. It has also raised a number of
questions that need to be answered.
First, it is necessary to figure out experimental conditions elaborately to utilize the
instrumented penetration test in the practical area. A 500 psi of penetration hardness was
chosen as a initial setting time for this study, matching the ASTM standard test.
However, reconsidering this definition is essential for more precise results because the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
210
hardness responses o f the instrumented penetration test may tend to develop faster than
the standard test, applying continuous compressive force on the material. Also, matching
the standard, the penetrator should move and test vertically to reduce settling effects.
Thus, a vertical instrumented penetration test should be designed.
Second, some aspects o f physical changes during cement hydration are not
completely understand. Although many interaction effects on the hardening process o f
cement-based materials were found in this study, the significance o f these interactions in
terms o f microstructure is still unknown. Also, a more comprehensive study o f hardness
response in terms o f rheology would be in order. This would help to better establish the
correlation between the instrumented penetration test and conventional rheology tests,
resulting in finding a novel technology. Thus, it is necessary to extend understanding of
the hardening process with respect to knowledge o f microstructural or rheological
aspects. A further investigation on the hardness development to find the relationships
between chemical and physical changes during cement hydration is required, resulting in
establishment of a scientific basis for the hydration mechanisms o f cement-based systems.
Lastly, a comprehensive study for exploring the cause o f strength enhancements
with respect to microstructure changes is necessary.
This study would realize the
utilization of microwave energy on cement-based systems.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES
Alexanderson, J., “Strength Losses in Heat Curing - Causes and Countermeasures,”
Behavior o f Concrete under Temperature Extremes. ACI SP 39, Detroit, MI, 91-107
(1973).
Alunno-Rossetti, V., Chiocchio, G., and Collepardi, M., “Low Pressure Steam Hydration
o f Tricalcium Silicate,” Cement and Concrete Research, Vol. 6,279-298 (1974).
ASTM C l87, “Standard Test Method for Normal Consistency o f Hydraulic Cement,”
American Society for Testing and Materials. Philadelphia, PA.
ASTM C191, “Standard Test Method for Time of Setting o f Hydraulic Cement by Vicat
Needle,” American Society for Testing and Materials. Philadelphia, PA.
ASTM C266, “Standard Test Method for Time o f Setting o f Hydraulic Cement by
Gillmore Needle,” American Society for Testing and Materials. Philadelphia, PA.
ASTM C403, “Standard Test Method for Time o f Setting o f Concrete Mixture by
Penetration Resistance,” American Society for Testing and Materials. Philadelphia,
PA.
ASTM C595, “Standard Specification for Blended Hydraulic Cements,” American
Society for Testing and Materials. Philadelphia, PA.
ASTM C778 “Specification for Standard Sand,” American Society for Testing and
Materials. Philadelphia, PA.
ASTM C807, “Standard Test Method for Time o f Setting o f Hydraulic Cement Mortar by
Modified Vicat Needle,” American Society for Testing and Materials. Philadelphia,
PA.
ASTM C l 157 “Performance Specification for Blended Hydraulic Cement,” American
Society for Testing and Materials. Philadelphia, PA.
Banfill, P.F.G., “Viscometric Study o f Cement Pastes Containing Superplasticizers with a
Note on Experimental Techniques,” Magazine o f Concrete Research, Vol. 33, No.
114,37-47(1981).
211
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
212
Bameyback,Jr., R.S., and Diamond, S., “Expression and Analysis o f Pore Fluid from
Hardened Cement Pastes,” Cement and Concrete Research, Vol. 11,279-285 (1981).
Barret, P. and Bertrandie, D., “Comment on “Aqueous Solubility Relationships for Two
Types o f Calcium Silicate Hydrates”,” Journal o f the American Ceramic Society, Vol.
71, No. 2, C113-C115 (1988).
Bentur, A., Berger, R.L., Kung, J.H., Milestone, N.B. and Young, J.F., “Structural
Properties o f Calcium Silicate Pastes: II, Effect o f Curing Temperature,” Journal o f
the American Ceramic Society, Vol. 62, No. 7, 362-366 (1979).
Benz, D.P., Stutzman, P.E., and Garboczi, E.J., “Experimental and Simulation Studies o f
the Interfacial Zone in Concrete,” Cement and Concrete Research, Vol. 22, 891-902
(1992).
Bilodeau, A. and Malhotra, V.M., “Concrete Incorporating High Volumes of ASTM
Class F Ashes: Mechanical Properties and Resistance to Deicing Salt Scaling and to
Chloride Penetration,” Proceedings, 4th International Conference on Fly Ash, Silica
Fume, Slag and Natural Pozzolans in Concrete, ACI SP 132, Detroit, MI, Vol. 1,
319-350(1992).
Boukamp, B.A., “Equivalent Circuit (EQUTVCRT.PAS),” University o f Twente, Dept, o f
Chemical Technology, Enschede, the Netherlands (1988).
Chang, J.T.Y., “Microwave Characterization and Thermal Processing of Cementitious
Materials,” Ph.D. Thesis, Northwestern University, Evanston, EL (1994).
Chatteiji, S., “The Role of Ca(OH)2 in the Breakdown o f Portland Cement Concrete due
to Alkali-Aggregate Reaction,” Cement and Concrete Research, Vol. 9, 185-188
(1979).
Christensen, B.J., “Microstructure Studies o f Hydrating Portland Cement-Based Materials
Using Impedance Spectroscopy,” Ph.D. Thesis, Northwestern University, Evanston,
IL (1993).
Christensen, B.J., Coverdale, R.T., Olson, R.A., Ford, S.J., Garboczi, E.J., Jennings,
H.M., and Mason, T.O., “Impedance Spectroscopy o f Hydrating Cement-Based
Materials: Measurement, Interpretation, and Application,” Journal o f the American
Ceramic Society, Vol. 77, No. 11,2789-2802 (1994).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
213
Christensen, B.J., Mason, T.O., and Jennings, H.M., “Influence of Silica Fume on the
Early Hydration o f Portland Cement Using Impedance Spectroscopy,” Journal o f the
American Ceramic Society, Vol. 75, No. 4, 939-945 (1992).
Christo, S.M., “Electromagnetic Interaction with Calcium Silicate Compounds
Microwave Characterization and Processing,” M.S. Thesis, Northwestern University,
Evanston, IL (1989).
Croft, B.A., “/« Situ Strength Development o f Microwave Cured Ordinary Portland
Cement Pastes,” M.S. Thesis, Northwestern University, Evanston, IL (1996).
Detwiler, R.J., Kjellsen, K.O., and Gjerv, O.E., “Resistance to Chloride Intrusion of
Concrete Cured at Different Temperatures,” AC I Materials Journal, Vol. 88, No. 1,
Jan.-Feb., 19-24(1991).
Diamond, S., “Chloride Concentrations in Concrete Pore Solutions Resulting from
Calcium and Sodium Chloride Admixture,” Cement and Concrete Aggregates, Vol. 8,
97-102 (1986).
Doremus, R.H., Glass Science. 2nd ed., John Wiley & Sons Inc., New York, NY (1994).
Eren, O., Brooks, J.J., and Celik, T., “Setting Times of Fly Ash and Slag-Cement
Concretes as Affected by Curing Temperatures,” Cement, Concrete, and Aggregate,
Vol. 17, No. 1, 11-17(1995).
Fraay, A.L.A., Bijen, J.M., and De Hann, Y.M., “The Reaction of Fly Ash in Concrete: A
Critical Examination,” Cement and Concrete Research, Vol. 19, No. 2, 235-246
(1989).
Fujii, K., and Kondo, W., “Kinetics of the Hydration o f Tricalcium Silicate,” Journal o f
the American Ceramic Society, Vol. 57, No.l 1,492-497 (1974).
Garboczi, E.J., “Permeability, Diffusivity, and Microstructural Parameters: A Review,”
Cement and Concrete Research, Vol. 19, 327-332 (1989).
Gartner, E.M., and Jennings, H.M., “Thermodynamics o f Calcium Silicate Hydrates and
Their Solutions,” Journal o f the American Ceramic Society, Vol. 70, No. 10, 743-749
(1987).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
214
Gebler, S.H., and Klieger, P., “Effects of Fly Ash on Physical Properties of Concrete,”
Proceedings, 2nd International Conference on Fly Ash, Silica Fume, Slag and
Natural Pozzolans in Concrete, ACISP 91, Detroit, MI, Vol. 1, 1-50 (1986).
Gjorv, O.E., Tan, K., Monteiro, P.J.M., “Effect o f Elevated Curing Temperature on the
Chloride Permeability o f High-Strength Lightweight Concrete,” Cement, Concrete,
and Aggregate, Vol. 16, No. 1, 57-62 (1994).
Greszczyk, S., and Kucharska, L. “The Influence o f Chemical Composition of Cement on
the Rheological Properties,” Rheology of Fresh Cement and Concrete, Banfill, P.F.G.,
ed., University Press, Cambridge, MA, 27-36 (1991)
Gu, P., Xie, P., Beaudoin, J.J., and Brousseau, R., “A.C. Impedance Spectroscopy (I): A
New Equivalent Circuit Model for Hydrated Portland Cement Paste,” Cement and
Concrete Research, Vol. 22, 833-840 (1992).
Gu, P., Xie, P., Beaudoin, J.J., and Brousseau, R., “A.C. Impedance Spectroscopy (II):
Microstructural Characterization of Hydrating Cement-Silica Fume System,” Cement
and Concrete Research, Vol. 23, 157-168 (1993).
Gu, P., Xu, Z., Xie, P., and Beaudoin, J.J., “Application o f A.C. Impedance Techniques
in Studies o f Porous Cementitious Materials (I): Influence o f Solid Phase and Pore
Solution on High Frequency Resistance,” Cement and Concrete Research, Vol. 23,
531-540(1993).
Harold S. Haller & Company, “Experimental Design Optimizer,” Harold S. Haller &
Company, Cleveland, OH (1992)
Harold S. Haller & Company, “Multiple Correlation,” Harold S. Haller & Company,
Cleveland, OH (1992)
Hogan, F.J., and Meusel, J.W., “Evaluation for Durability and Strength Development of a
Ground Granulated Blast Furnace Slag,” Cement, Concrete and Aggregate, Vol. 3,
No. 1,40-52(1981).
Hooton, R.D., “Permeability and Pore Structure of Cement Pastes Containing Fly Ash,
Slag, and Silica Fume,” Blended Cement, ASTM STP 897, American Society for
Testing and Materials. Philadelphia, PA, 128-143 (1986).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
215
Hooton, R.D., “Influences o f Silica Fume Replacement of Cement on Physical Properties
and Resistance to Sulfate Attack, Freezing and Thawing, and Alkali-Silica
Reactivity,” ACIMaterials Journal, Vol. 90, No. 2, 143-141 (1993).
Hutchison, R.G., Chang, J.T., Jennings, H.M., and Brodwin, M.E., “Thermal
Acceleration o f Portland Cement Mortars with Microwave Energy,” Cement and
Concrete Research, Vol. 21, 795-799 (1991).
Ishida, H., Sasaki, K., Okada, Y., and Mitsuda, T., “Highly Reactive (3-Dicalcium
Silicate: IE, Hydration Behavior at 40-80°C,” Journal o f the American Ceramic
Society, Vol. 75, No. 9, 2541-2546 (1992).
Jennings, H.M., “Aqueous Solubility Relationships for Two Types of Calcium Silicate
Hydrates,” Journal o f the American Ceramic Society, Vol. 69, No. 8, 614-618 (1986).
Jennings, H.M., “Reply to Comment on “Aqueous Solubility Relationships for Two
Types o f Calcium Silicate Hydrates”,” Journal o f the American Ceramic Society, Vol.
71, No. 2, C115-C116 (1988).
Jennings H.M., and Lange, D.A., “Toward High Performance Cement Based Materials”
Proceedings o f the American Society o f Civil Engineers National Convention, St.
Louis, MO, October 1988, American Society of Civil Engineers, New York, 18-21
(1989).
Johns, T.E.R., and Taylor, S., “A Mathematical Model Relating the Flow Curve of a
Cement Paste to its Water/Cement Ratio,” Magazine o f Concrete Research, Vol. 29,
No. 101,207-212(1977).
Katz, A.J., and Thompson, A.H., “Quantitative Prediction o f Permeability in Porous
Rock,” Physics Review B, Vol. 30, No. 11, 8179-8181 (1986).
Kjellsen, K.O., “Heat Curing and Post-Heat Curing Regimes of High-Performance
Concrete: Influence on Microstructure and C-S-H Composition,” Cement and
Concrete Research, Vol. 26,295-308 (1996).
Kjellsen, K.O., Detwiler, R.J., and Gjorv, O.E., “Backscattered electron Imaging of
Cement Pastes Hydrated at Different temperatures,” Cement and Concrete Research,
Vol. 20,308-311 (1990).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
216
Kjellsen, K.O., Detwiler, R.J., and Gjorv, O.E., “Development o f Microstructures in Plain
Cement Pastes Hydrated at Different Temperatures,” Cement and Concrete Research,
Vol. 21, 179-189 (1991).
Kouloumbi, N., Batis, G., and Malami, C., “The Anticorrosive Effect o f Fly Ash, Slag,
and a Greek Pozzolan in Reinforced Concrete,” Cement and Concrete Composites,
Vol. 16,253-260 (1994).
Lapinas, R.A., “Hot Concrete for Achievement o f High Early Strength,” Behavior of
Concrete under Temperature Extremes. SP 39, Detroit, MI, 109-138 (1973).
Leung, C.K.Y., and Pheeraphan, T., “Determination o f Optimal Process for Microwave
Curing o f Concrete,” Cement and Concrete Research, Vol. 27,463-472 (1997).
Leung, C.K.Y., and Pheeraphan, T., “Microwave Curing of Portland Cement Concrete:
Experimental Results and Feasibility for Practical Application,” Construction and
Building Materials, Vol. 9, No. 2,67-73 (1995).
Leung, C.K.Y., and Pheeraphan, T., “Very High Early Strength of Microwave Concrete,”
Cement and Concrete Research, Vol. 25, 136-146 (1995).
Li, D., and Wu, X., “A Study on the Application o f Vacuum Microwave Composite
Dewatering Technique in Concrete Engineering,” Cement and Concrete Research,
Vol. 24, 159-164(1994).
Malhotra, V.M. and Hemmings, R.T., “Blended Cements in North America - A Review,”
Cement and Concrete Composites, Vol. 17,23-35 (1995).
Massazza, F., “Pozzolanic Cements,” Cement and Concrete Composites, Vol. 15, 185214(1993).
McCarter W.J., and Garvin, S., “Dependence o f Electrical Impedance o f Cement-Based
Materials on Their Moisture Condition,” Journal o f Physics D: Applied Physics, Vol.
22, 1773-1776(1989).
McCarter, W.J., “Effects of Temperature on Conductivity and Polarization in Portland
Cement Mortar,” Journal o f the American Ceramic Society, Vol. 78, No. 2, 411-415
(1995).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
217
McCarter, W.J., “The A.C. Response of Hardened Cement Paste,” Cement and Concrete
Research, Vol. 20, 891-900 (1990).
McCarter, W.J., and Afshar, A.B., “Further Studies on the Early Hydration o f Portland
Cement Paste,” Journal o f Material Science Letter, Vol. 4,405-408 (1985).
McCarter, W.J., and Afshar, A.B., “Some Aspects o f the Electrical Properties o f Cement
Paste,” Journal o f Material Science Letter, Vol. 3, 1083-1086 (1984).
McCarter, W.J., Garvin, S., and Bouzid, N., “Impedance Measurements on Cement
Paste,” Journal o f Material Science Letter, Vol. 7, No. 10, 1056-1057 (1988).
Mehta, P.K., and Monteiro, P.J.M., Concrete: Structure. Properties, and Materials. 2nd
ed., Prentice-Hall Inc., Englewood Cliffs, NJ (1993).
Metaxas, A.C., and Meredith, R.J., Industrial Microwave Heating. Peter Peregrinus Ltd.,
London, United Kingdom (1983).
Miller, E.W., “Blended Cements - Applications and Implications,” Cement and Concrete
Composites, Vol. 15, 237-245 (1993).
Mindess, S. and Young, J.F., Concrete. Prentice-Hall Inc., Englewood Cliffs, NJ (1981).
Mehta, P. K. and Monteiro, P. J. M. Concrete: Structure. Properties, and Materials.
Prentice-Hall Inc., Englewood Cliffs, NJ (1993).
Moukwa, M., Brodwin, M.E., and Shah, S.P., “Microwave Heating Applied to the Curing
o f Mortars,” International Conference on Microwave and High Frequencies, Nice,
France, 209-212(1991).
Nasser, K.W., “Elevated Temperature Effect on the Structural Properties of Air-Entrained
Concrete,” Behavior o f Concrete under Temperature Extremes. ACI SP 39, Detroit,
MI, 139-148 (1973).
Neville, A. M., Properties o f Concrete. John Wiley & Sons, New York, NY (1996).
Odler, I., Brecker, T.L., and Weiss, B., “Rheological Properties of Cement Pastes,” R
Cemento, Vol. 75, 303-310 (1978).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
218
Odler, I., “ Final Report o f Task Group I, 68-MMH Technical Committee on Strength of
Cement,” Materials and Structures, Vol. 24, No. 140, 143-157 (1991).
Orchard, D. F., Concrete Technology, vol.2, 3rd ed., John Wiley & Sons, New York, NY
(1993).
Orchard, D. F. and Barnett, A. M., “The Electrocuring o f Concrete Test Specimens,”
Journal o f Materials, JMLSA, Vol. 6, 617-642 (1971).
Oriol, M., and Pera, J., “Pozzolanic Activity of Metakaolin under Microwave Treatment,”
Cement and Concrete Research, Vol. 25, 265-270 (1995).
Parcevaux, P., “Pore Size Distribution o f Portland Cement Slurries at Very Early Stages
o f Hydration,” Cement and Concrete Research, Vol. 14, 419-430 (1984).
Patel, H.H., Bland, C.H., and Poole, A.B., “The Microstructure o f Steam-Cured Precast
Concrete,” Advanced in Cement Research, Vol. 8, No. 29, 11-19 (1996).
Polivka, M., and Klein, A., “Symposium on Effect o f Water-Reducing Admixtures on
Properties o f Concrete,” ASTM STP 266, American Society for Testing and Materials
Philadelphia, PA, 124-149 (1960).
Powers T. C., J. Res. Dev. Labs. Portland Cement Association, Vol. 3, No. I, 47-56
(1961).
Powers, T.C., The Properties o f Fresh Concrete. John Wiley and Sons, New York, NY
(1968).
Radjy, F., and Richards, C.W., “Effect of Curing and Heat Treatment History on the
Dynamic Mechanical Response and the Pore Structure o f Hardened Cement Paste,”
Cement and Concrete Research, Vol. 3, 7-21 (1973).
Ramachandran, V. S., Feldman, R. F. and Beaudoin, J. J., Concrete Science. Heyden &
Son, Philadelphia, PA (1981).
Ramezanianpour, A.A., and Malhotra, V.M., “Effects o f Curing on the Compressive
Strength, Resistance to Chloride-Ion Penetration and Porosity of Concretes
Incorporating Slag, Fly Ash or Silica Fume,” Cement and Concrete Composites, Vol.
17, 125-133 (1995).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
219
Ravina, D., and Shalon, R., “The Effect of Elevated Temperature on Strength o f Portland
Cements,” Temperature and Concrete. ACISP 25, Detroit, MI, 275-289 (1971).
Ross, J. P., Taguchi Techniques for Quality Engineering. McGraw-Hill, New York, NY
(1988)
Roy, D.M., “Hydration o f Blended Cements Containing Slag, Fly Ash, or Silica Fume,”
Proceedings o f Meeting Institute o f Concrete Technology, Coventry, United
Kingdom, 29-40 (1987).
Roy, D.M., “The Effect o f Blast Furnace Slag and Related Materials on the Hydration and
Durability o f Concrete,” Durability of Concrete. International Symposium, ACI SP
131, Detroit, MI, 195-208 (1992).
Roy, D.M., and Asaga, K., “Rheological Properties o f Cement Mixes: III Effects of
Mixing Procedures on Viscometric Properties o f Mixes Containing Superplasticizer,”
Cement and Concrete Research, Vol. 9, 731-739 (1979).
Roy, D.M., and Idom, G.M., “Hydration, Structure, and Properties o f Blast Furnace Slag
Cements, Mortars, and Concrete,” ACI Journal, Vol. 79, No. 6, Nov.-Dee., 444-457
(1982).
Scrivener, K.L., “The Effect o f Heat Treatment on Inner Product C-S-H,” Cement and
Concrete Research, Vol. 22, 1224-1226 (1992).
Scuderi, C.A., Mason, T.O., and Jennings, H.M., “Impedance Spectra o f Hydrating
Cement Pastes,” Journal o f Material Science, Vol. 26, 349-353 (1991).
Sellevold, E.J., “Mercury Porosimetry of Hardened Cement Paste Cured or Stored at
97°C,” Cement and Concrete Research, Vol. 4, 399-404 (1974).
Sersale, R., “Aspects o f the Chemistry of Additions,” Advances in Cement Technology.
S.N. Ghosh, ed., Pergamon Press, New York, NY (1983)
Shane, J.D., Hwang, J.H., Sohn, D., Mason, T.O., Jennings, H.M., and Garboczi, E.J.,
“Recent Developments in the Measurement o f Transport Properties in Cement-Based
Materials: Emerging Transport Property Measurement Techniques,” Mechanisms o f
Chemical Degradation in Cement-Based Materials, in press
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
220
Schultz, M.A., and Struble L.J., “Using Creep and Recovery to Study Flow Behavior of
Fresh Cement Paste,” Cement and Concrete, Vol. 23,273-282 (1993).
Skalny, J., and Odler, L., “Pore Structure o f Calcium Silicate Hydrates,” Cement and
Concrete Research, Vol. 2, 387-400 (1972).
Soroka, I., Portland Cement Paste and Concrete. Macmillan Press, London, United
Kingdom (1979).
Stein, H.N., “Rheological Behavior o f Suspensions,” Encyclopedia o f Fluid Mechanics
V, Cheremisinoff, N.P. ed., Gulf Publishing Company, Houston, TX, 4-48 (1986)
Struble L.J., “The Rheology of Fresh Cement Paste,” Advances in Cementitious
Materials. Mindess, S. ed., The American Ceramic Society, Westerville, OH, 1-26
(1991).
Struble L.J., and Lei, W.-G. “Rheological Changes Associated with Setting o f Cement
Paste,” Advanced in Cement Based Materials, Vol. 2, N o.6,224-230 (1995).
Sutton, W.H., “Microwave Processing o f Ceramic Materials,” The American Ceramic
Society Bulletin, Vol. 68, No. 2, 376-386 (1989).
Tadmor, Z., and Gogos, C.G., Principles o f Polymer Processing. John Wiley & Sons,
New York, NY (1979)
Tashiro, C., Ishida, H., and Shimamura, S., “Dependence o f the Electrical Resistivity on
Evaporable Water Content in Hardened Cement Pastes,” Journal o f Material Science
Letter, Vol. 6, 1379-1381 (1987).
Tattersall, G.H., and Banfill, P.F.G., The Rheology o f Fresh Concrete. Pitman Advanced
Publishing Program, Academic Press, London, United Kingdom (1983).
Taylor, H.F.W., Cement Chemistry. Academic Press, London, United Kingdom (1990).
Thomas, J.J., “Microwave Processing o f Reaction-Bonded Silicon Nitride,” Ph.D. Thesis,
Northwestern University, Evanston, IL (1994).
Verbeck, G.J., and Helmuth, R.H., “Structures and Physical Properties o f Cement
Pastes”, Proceedings o f the 5th International Symposium o f the Chemistry o f Cement,
Tokyo, Japan, 1-32 (1968).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
221
Vom Berg, W, “Influences of Specific Surface and Concentration o f Solids upon the
Flow Behavior o f Cement Pastes” Magazine o f Concrete Research, Vol. 31, No. 109,
211-216(1979).
Wainwright, P.J., and Tolloczko, J.J.A., “A Temperature-Matched Curing System
Controlled by Microcomputer,” Magazine o f Concrete Research, Vol. 35, No. 124,
164-169 (1983).
Watson, A., “Curing o f Concrete,” Microwave Power Engineering. Vol. 2, E.C. Okress
ed. Academic Press, New York, NY, 108-110, (1968).
Whittington, H.W., McCater, W.J., and Forde, M.C., “The Conduction of Electricity
through Concrete,” Magazine o f Concrete Research, Vol. 33, No. 114, 48-60 (1981).
Wu, X., Dong, J., and Tang, M., “Microwave Curing Technique in Concrete
Manufacture,” Cement and Concrete Research, Vol. 17, 205-210 (1987).
Xu, Z., Gu, P., Xie, P., and Beaudoin, J.J., “Application o f A.C. Impedance Techniques
in Studies o f Porous Cementitious Materials (II): Relationship between ACIS
Behavior and the Porous Microstructure,” Cement and Concrete Research, Vol. 23,
853-862 (1993).
Yang, M., “A Study o f the Rheological Behavior and Microstructure of Cement Paste,”
Ph.D. Thesis, Northwestern University, Evanston, IL (1994).
Young, J.F., “ Investigations o f Calcium Silicate Hydrate Structure Using Silicon-29
Nuclear Magnetic Resonance Spectroscopy,” Journal o f the American Ceramic
Society, Vol. 71, No. 3, Cl 18-C120 (1988).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX A
General Theory of Microwaves
Microwaves are coherent and polarized electromagnetic waves with a frequency
range o f 0.3 to 300 GHz and corresponding wavelengths ranging from 1 m to 1 mm
(Metaxas and Meredith, 1983), shown in Figure A l. They can be transmitted, absorbed ,
or reflected, depending on the material type (Sutton, 1989), shown in Figure A2. The
material’s complex permittivity, s* (F/m), which is related to the degree o f microwave
absorption, can be divided into real and imaginary components:
e* = s ’ - j e” =
So
( sr’ - j
S e ff”
)
(Eq. A l)
where s0 is the free space permittivity (= 8.86X1 O'12 (F/m)), sr’ is the relative dielectric
constant, seff” is the effective relative dielectric loss factor, and j = (-1)1/2.
The penetration and propagation o f microwaves through a dielectric material
induces internal electric fields within the affected volume, resulting in translational
motions o f free or bound charge and rotation o f dipoles.
At lower frequencies, the
dipoles have ample time to follow the alternating field, hence a high dielectric constant.
As the frequency increases, the dipoles are unable to keep up, resulting in a sharp
decrease o f the dielectric constant and a corresponding rise in loss factor. This frequencydependent resistance to the induced motions due to inertial, elastic and frictional forces
222
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
223
---------
1
11
1----1
1
1
Ix-rays! u.v.
1 SHF VHF MF VLF
1
!EHF, r UHF, r H F , r L F , r
i.r.
I l l
100 A
/
1 (i
100 ly 1 cm
1m
1... V . ..J..
100
10 km
Wavelength
3 x l0 16 3 x l0 14 3xL012 3 x l0 10 3xl08 3xl06\3 x l0 4 Frequency (Hz)
/
\
/
\
/
\
/
\
/
\
/
Dielectric heating
\
/
frequencies
/
\
Millimeter
waves
Radio frequencies
Microwaves H"
I
Radar bands —
•
K
|x|
S
1L
1 cm
10 cm
1m
10 m
100 m
Wavelength
3 x l0 10
3x109
3x108
3 x l0 7
3xl06
Frequency (Hz)
tt t
2.45 GHz
433.9 MHz
13.56 MHz
27.12 MHz
900 MHz
Principal
frequencies
allocated
for industrial
use
F igure A -l Electromagnetic spectrum (Metaxas and Meredith, 1983).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
224
v
/
V
W
/ \ / \ / \ / ^
O || <=> *
O || o
Material type
Penetration
Transparent
(Low loss
insulator)
Total
Opaque
(Conductor)
None
(Reflected)
Absorber
(Lossy
insulator)
Partial
to Total
Absorber
(Mixed)
Partial
to Total
(a) Matrix : low loss insulator
(b) Particles : absorbing materials
Figure A-2 Interaction o f microwaves with materials (Sutton, 1989).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
225
causes loss and attenuation o f the electric field, generating heat rapidly throughout the
material. The loss tangent is used to describe these losses as
tan 5 = eeff” / £r’ = cr / (2n f eQsr’)
(Eq. A2)
where ct is the total effective conductivity a n d /is the frequency (GHz). When tan 8 »
I,
materials are considered lossy, whereas when tan 8 « 1 , they show low-Ioss nature.
The relative dielectric constant and the loss tangent are the two most widely used
and measured parameters to describe the behavior o f a dielectric under the influence of a
microwave field. The relative dielectric constant represents the amount o f stored energy
in the material caused by polarization under an applied field, whereas the loss tangent
indicates the amount o f energy dissipation within the material by all o f the loss
mechanisms relevant to high frequency heating (Thomas, 1994).
The power absorbed per unit volume P (W/m3), provides the following basis for
heating:
P = <r E2 = 2 n f e 0 er’ tan 8 E2
(Eq. A3)
where E is the magnitude o f the internal field. Theparameters off, er’, tan 8 and E are all
independent, and E is dependent on the size, geometry, and location o f the material within
a microwave cavity and on the design and volume o f the cavity. However, the equation
provides a useful approximation o f P and describes the basic relationships between these
four variables (Sutton, 1989).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
226
Attenuation o f the electric fields as the microwaves penetrate and propagate
through an absorbing material are described by the penetration depth D, which indicates
the length where the incident power is reduced by one half, shown in Figure A3 (Sutton,
1989). There is a simplified equation o f D for low loss materials such as pure water (tan
5«1)
rpr
D = -------------3—0
8.686 71 tan5 (e r / s 0) 1/2
(Eq. A4)
where X0 is the incident or ffee-space wavelength.
The penetration depth is dependent on the microwave frequency and the type of
materials.
In the low frequency range, the penetration depth is too large to generate
enough heat because o f low loss.
As the frequency increases, the penetration depth
become smaller and the heating necessarily increases. For pure water, the penetration
depth increases with increasing temperature. For the commercial microwave frequency
o f 2.45 GHz, the penetration depth is 13 mm at 20°C and increases to 30mm at 60°C,
shown in Figure A4 (Metaxas and Meredith, 1983).
Microwave heating is fundamentally different from conventional heating.
In
microwave processes, heat is generated internally within the material instead of
originating from external heating sources, resulting in volumetric heating. Due to the
internal and volumetric heating, heat flow and thermal gradients in microwave-processed
materials are the reverse o f those in materials processed by conventional heating (Figure
A5) (Sutton, 1989). Consequently, microwave processing makes it possible to heat both
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
227
D = ----------------- ^
------=
( 8.686 n ) ( tan 0 )
_
s'r / s 0
D : Penetration depth at 1/2 power
A.0 : Free space wavelength
tan d : Dielectric loss tangent
s ’r/s0 : Dielectrric constant
t
frequency (GHz)
0.915
2.45
o
Dwater ( C H l )
7.6
1.3
32.8
12.3
Figure A-3 Absorption o f microwaves to half-power depth, D (Sutton, 1989).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
228
Water
1000
500
200
£
100
B,
CL
Q
- -a
434 MHz
- 915 MHz
- - a - - 2450 MHz
20
40
60
T(°C)
Figure A-4 Penetration depth as a function of the temperature for pure water
(Metaxas and Meredith, 1983).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
229
Conventional
Heating Element
Sample
-
-►
-
-►
-
-►
Furnace
Insulation
Microwave
Microwave port
I
Sample
Insulation
Metal shell
Cavity
Figure A-S Heating Patterns in conventional and microwave furnaces (Sutton, 1989).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
230
small and large shapes very rapidly and uniformly, to efficiently remove volatile
constituent, i.e. binders, moistures, etc., from thick sections, and to reduce thermal
stresses that cause cracking during processing.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX B
Statistical Methodology
B.l. Experimental Design (Harold S. Haller & Co, 1992)
Experimental design is one o f the most widely discussed topics in the area of
quality and process improvement. Many types o f experimental design exist.
In this
section, a brief introduction o f three more popular designs, such as full factorial, Taguchi,
and D-optimal, with their strengths and weaknesses, is given.
First, ‘Full Factorial’ designs experiments to run all possible combinations of the
experimental conditions. As an example, when 3 variables have 3 levels, the full factorial
sets are 27. The designs have very low Average Errors o f Predictions which indicate that
designs are too inefficient to be practical for most applications.
Taguchi (or three level Fractional Factorial) is the most well-known method of
experimental design.
It is a fractional of a Full Factorial design with each variable
examined at three levels. This design can quantify non-linear effects to make it possible
for the analysis to detect optimums.
The D-optimal design is more flexible because it accommodates factors that have
a different numbers o f levels and allows the operator to place constraints on the design
space. Table B.l summarizes advantages and disadvantages of these designs and details
are given elsewhere (Ross, 1988; Harold S. Haller & Co, 1992).
231
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
232
Table B -l Advantages and disadvantages for experimental designs
(Harold S. Haller & Co, 1992).
Advantage
Full Factorial
Taguchi
Disadvantage
Uncover complicated effects
Inefficient
Least distortion
Expensive
Easy to analyze data
Time-consuming
Three levels allow
Difficult to look at interaction
Averages out error quite well
Often inefficient
Easy to analyze data
Cannot build on existing data
Cannot exclude undesirable set
D-optimal
Less effort to look as variables
Sacrifice o f some information
Averages out variation
Need multiple correlation
Can build on existing data
Need more computing time
Easy to study interactions
Can exclude undesirable set
Allows ordering o f experiment
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
B.2. Multiple Correlation Analysis
Multiple Correlation (MC) determines the relation between a dependent variable
(Y) and independent variables (X’s) to predict Y from the X’s in an accurate manner.
MC analysis can distinguish the independent variables that significantly affect the
dependent variable to establish a functional relationship between them. Also, it allows to
quantify the effects o f main variables and their interactions, resulting in accurate
predictions from the relationships obtained.
Since MC analysis uses a Taylor expansion to approximate the response, it is
necessary to choose significant Taylor expansion terms in a statistically reasonable way.
The MC software (MC, Harold S. Haller & Co, 1992) provides this method.
Prior to generating the MC model, it is highly recommended to code the
independent variables, with -1 and +1 representing the minimum and maximum,
respectively. The encoding and decoding functions for variable X are as follows:
„
,.
Encoding
„
2 ( X - X min) ,
X c = —----- — ----- -- 1
^m ax
n
in
Decoding
m in
y = -------------+ ^ (Xmax ” Xmin^+. vX min
X
The useful statistical terminology are listed below in alphabetical order.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
234
Correlation Matrix: The correlation coefficient (denoted by the letter r) for two
variables is a measurement o f the linear relationship between them in a range between -1
and +1.
If r is close to the absolute value o f 1, two variables are much correlated,
whereas if r is close to zero, the variables are independent each other. In general, a high
correlation between independent variables is not desirable. The way to avoid this is to
collect data using a appropriate experimental design.
Degree o f Freedom (D.O.F): The degree o f freedom indicates overfitting. The degree o f
freedom is calculated by:
Degree o f Freedom = n - p -1 ,
where n = number o f rows in the date set, p = number o f terms in the model. For instance,
when the degree o f freedom goes to zero, the model is highly likely overfitted. More than
3 o f the degree o f freedom is highly recommended to avoid forcing a fit.
E.S. Residual (Externally Studentized Residual): This has a true t-distribution. The
standard deviation term used in calculating the E.S. residual is the standard deviation of
the regression without the i* data point. If a row has an E.S. Residual greater than 3, that
raw is considered to be an outliner.
Residual:
E.S. Residual: = ----- .===-=
S ;V ^H 7
|(n - p) S j , - (R esidual,)2/(I - H J
‘“ V
(n-p-1)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
235
where a = number o f rows in the date set, p = number o f terms in the model.
H value: The H-value is a measurement o f how good the experimental design is for ail
variables in the model. A high H value indicates that the data point is much scattered
from others. However, it is not recommended that a data row be deleted solely on the
basis o f its H-value.
Interaction: An interaction exists between two independent variables if the response of
one o f them depends on the value o f the other. Interactions between variables are often
crucial in doing multiple correlation.
Modified R2: This is a measure of how much of the variance in the dependent variable,
exclusive of the test variance, is explained by the model. This is calculated by:
M
]p 2
. ( n - p - 1 ) (S2x - S2Mt)
Modified R = 1 —------------- ;---------^-----(n -l)(S ^ -S ^ )
where n = number o f rows in the date set, p = number of terms in the model.
Outlier: When a row o f data is somehow different from the rest of the data, it is
considered to be an outiiner.
Finding out what caused this row to be different may
suggest other variables to be included in the model. It should be considered to delete any
outliners because they distort the correlation and lead to erroneous conclusions. Here are
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
236
some criteria for determining outliners (but these are not absolute). If a row violates
more than one o f these criteria, it should be deleted.
1 .1Residual | > 3 Syjc
2. | Residual | > | 1.5 * (next largest residual) |
3. | Standardized residual | > 3
4. | E.S. Residual | > 3
R2: The residual square (R2) calculation is for the model, known as the coefficient of
multiple determination. This is a measure o f how much o f the variation in the dependent
variable is explained by the current model. A value of 1 would means all the variation
was explained, while 0 indicates none. It is calculated by:
(n -p -l)s ;,
( n - l ) S | 0Ol
where n = number of rows in the date set, p = number of terms in the model. Note that a
large R2 may sometimes indicate overfitting.
Residual: It can determine an outliner. The residual for a given row is calculated by:
Residualj = (Observed Value)* - (Fitted Value)*
S tandard E rror: This is the standard deviation o f a predicted value. It can be used to
establish confidence intervals for the estimate o f the average effect at that point. For a
sample containing n specimens, for which the standard deviation is S, the standard error
is S.E. = S/Vn.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
237
68% Confidence: Predicted Value ± Standard Error
95% Confidence: Predicted Value ± 2 * Standard Error
99% Confidence: Predicted Value ± 3 * Standard Error
Standardized Residuals: These are calculated by:
S tan dardized Re sidual i =
Residual,y.x
If a raw has the value greater than 3, it is an outlier.
Sxest-' This is a measure o f the variability in the test procedure for the dependent variable.
Knowing Sxest is necessary in order to use Modified R2 statistics. The calculation is
S Test = ^(weighted average var iance for duplicated tests)
Sy.x: This measures the amount o f variability in dependent variables not explained by the
given model. If no variables were found in the model, Sy.x is equal to standard deviation
of the dependent variables. Sy x in the model is expected to be close to, but not below,
Sxest- It is highly recommended that Sy.x <1.7 Sxest-
t-value: The t-value indicates the significance o f an independent variable. Generally, a tvalue > |2| is considered to be significant. The t-value for the i1*1variable is given by:
t = bj/s(bj)
where bi = coefficient for the i1*1variable, s(bj) = standard deviation of the i* coefficient.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX C
Impedance Spectroscopy Study on Cement-Based Materials
As the microstructure and the amount o f porosity changes with time, the
expectation is that the electrical properties would also change. These changes in the
electrical properties o f cement-based materials can be related to known changes in the
microstructure, thereby providing a useful probe for studying the changing microstructure
(Whittington, et al., 1981; McCarter, et al.; 1984, McCarter, et al., 1985; McCarter, et al.,
1988). Electrical measurement offers a non-destructive method, involving continuous
and in situ measurement on the same sample without altering the microstructure. This is
a distinct advantage over such conventional methods o f microstructure investigation as
SEM, MEP, BET, because the microstructure of the cement is so sensitive to changes in
the humidity o f its surroundings. Testing methods that require drying before analysis
have left many questions o f the real microstructure being tested in these methods and the
validity o f the results obtained (Christensen, et al., 1994).
Cement paste begins as a plastic mixture of solids and water. Depending on the
hydration time, a number o f conduction paths are conceivable. The main conduction
paths are through capillary porosity and gel porosity because the solid hydration products
and unhydrated cement have much higher resistivities (Tashiro, et al., 1987; McCarter
and Garvin, 1989).
Cement pastes containing silica fume exhibit a higher electrical
238
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
239
resistance than normal cement paste at the same water-to-cement ratio and time of
hydration (Christensen, et al., 1992).
This is most likely due to the decreasing
conductivity o f the pore fluid (Christensen, et al., 1992), in conjunction with the reported
formation o f discontinuous capillary pathways in these materials (Jennings and Lange,
1989).
Gain and phase angle measurements are made over a broad range o f frequencies
in Nyquist plots (real impedance versus imaginary impedance) as shown in Figure C-l
(Christensen, et al., 1994). The high frequency arc is associated with bulk phenomena,
while the partial low frequency arc is an electrode interface phenomenon.
Recently,
considerable attention has been given to the bulk arc which occurs in the kHz to MHz
range (Scuderi, et al., 1991; McCarter, 1990; Gu, et al., 1992; Gu, et al., 1993; Gu, et al.,
1993; Xu, et al., 1993). Attempts have been made to correlate particular aspects of the
bulk arc with capillary porosity, composition o f the pore fluid, threshold pore diameter
and a measure of the connectivity of the pore structure. These various microstructural
factors can be incorporated into the relationship (Garboczi, 1989)
ct- cto thcap (3
(Eq. Cl)
where a is the overall conductivity of paste, ct0 is the conductivity o f the aqueous phase in
the porosity, Ocap is the volume fraction of capillary porosity, and P is connectivity or
inverse tortuosity factor. This allows one to characterize the microstructure, independent
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
240
1200
-Imaginary Impedance
1000
Bulk
' Electrode
Response Response
-
800-
600“
Bulk
MHz
400-
200
5 Hz
-
0
200
400
600
800
1000
Real Impedance
Figure C -l Typical Nyquist plot o f Portland cement paste at 28 days.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1200
241
o f alkali content, admixture addition, water content, etc. Since none o f these terms is a
constant, it is necessary to measure each component as a function o f time to accurately
understand the changing microstructure. Since both Ocap and (3 decrease with hydration
time, the normalized conductivity (<j /<j 0) o f paste also decrease (Christensen, 1993).
The normalized conductivity is used to predict the microstructure-related
properties such as permeability (Katz and Thompson, 1986) and diffusivity (Shane, 1995)
using the following equations:
Katz-Thompson (Permeability)
Nemst-Einstein (Diffusivity):
k=
f
1
CT
\
226.
D_
vc r 0 J
where dc is the threshold pore size from MIP,
ct/ ct0
lD n
(Eq. C2)
(Eq. C3)
is normalized conductivity, D is the
diffusivity o f the species of interest in the paste and D0 is the intrinsic diffusivity in pure
water. Permeability is closely related to durability. The diffusivity is a measure of how
quickly an ion moves through a material under a concentration gradient. The diffusion of
ions through cement paste is a critical design issue in steel-reinforced concrete and in the
containment o f landfill or nuclear wastes. These relationships must be tested in a more
rigorous manner on a common set o f cement samples. However, it seems to indicate that
electrical and fluid transports are related and a reasonable estimate can be made by these
methods.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IMAGE EVALUATION
TEST TARGET ( Q A - 3 )
1.0
£ 1 2 12
12
|££
if |g£
l.l
S la
1 22
\m
1.8
1.25
1.4
1.6
150mm
IIW lG E .Inc
1653 East Main Street
Rochester, NY 14609 USA
Phone: 716/482-0300
Fax: 716/288-5989
0 1993. Applied Image. Inc.. AD Rights Reserved
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
7 992 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа