 

 

 

 

 

 

 

     

.. . . u .. .. n . , v- :1. ..,.‘. A. . v . . . » v... . .4v:~y..~‘..0. .Q‘v.p-.$-.ilvly“‘-.|9-

UNERSI TY LIBRARIE

Illllllll\IlHlllHlLolll llllll [ll

3 129300

llll

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the
dissertation entitled

Reinforcement of Thin Cement Products
with Recycled Wastepaper Fibers

presented by

Zahir Shah

has been accepted towards fulﬁllment
of the requirements for

Ph.D. Civil Engineering

degree in

 

 

, 1 Q _
313 Major professor

Date 2.. /7/ l993

MS U i: an Affirmative Action/Eq ual Opportunity Institution 0.12771

_ .__._-—_. .___..__. __ .. _

 

r LIBRARY
Michigan State
University

X .J

—

 

 

PLACE IN RETURN BOX to remove this checkout from your record.

TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

____i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
czmma-pA

 

REINFORCEMENT OF THIN CEMENT PRODUCTS
WITH RECYCLED WASTEPAPER FIBERS

By

Zalu'r Shah

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Civil & Environmental Engineering

1993

 

ABSTRACT

REINFORCEMENT OF THIN CEMENT PRODUCTS WITH RECYCLED
WASTEPAPER FIBERS

By

Zalu’r Shall

The main intent of this research was to determine the technical feasibility of utilizing mag-
azine wastepaper ﬁbers, obtained through dry processing of paper, as reinforcement in thin
cement products. Dry-processed magazine papers have high levels of non-cellulosic impu-

rities, and the recycling process also breaks and damages the ﬁbers.

In order to produce wastepaper ﬁber-cement composites, ﬁrst the inﬂuential variables in
slurry-dewatering method of processing the composites were identiﬁed in an experimental
study based on fractional factorial design. Among the proportioning and processing vari-
ables investigated, ﬁber mass fraction, level of substitution of virgin ﬁbers with recycled
ones, and ﬁber reﬁnement conditions were found to have statistically signiﬁcant effects on
the ﬂexural performance of composites. Subsequently, response surface analysis tech-
niques were used to devise an experimental program which helped determine the optimum
combinations of the selected inﬂuential variables based on ﬁexural performance and cost.
The optimized recycled composites were then technically evaluated versus virgin compos-
ites. They were shown to possess acceptable ﬂexural strength, dimensional stability, den-
sity, water absorption and moisture content Speciﬁc size distribution of recycled ﬁbers
(with higher ﬁne contents) were used to justify their differences with virgin ﬁbers in

cement composites.

The effects of moisture and weathering on the performance of recycled wastepaper ﬁber-

cement composites were investigated through accelerated laboratory tests simulating the

effects of wet-dry and freeze-thaw cycles as well as carbonation and chemical interaction,
in natural weathering. Microstructural studies were conducted in order to establish the
mechanisms of ageing in the composite material. These mechanisms provided the basis
for selection of certain reﬁnements in the matrix composition, which were successfully
evaluated for the control of weathering eﬁ‘ects on the composite material structure and
properties. The eﬁ‘ects of ageing and moisture on composites were best controlled by mea-
sures which reduced the calcium hydroxide content of hydration products and improved
the water-tightness and the structure of interface zones; these reﬁnements were made
using relatively high levels of replacing cement with silica fume or through full substitu-

tion of Portland cement with a special cement.

A cost analysis was performed on the optimized wastepaper ﬁber-cement composites ver-
sus alternative siding building materials. The optimized recycled composites were found
to present the best initial and life-cycle cost positions among commercially available sid-
ing materials.

 

Dedicated to my mother “Hussan Bano”

iv

ACKNOWLEDGEMENTS

First of all I wish to thank God Almighty who enabled me to undertake and complete this
research studies. I wish to express my sincere appreciation to Dr. Parviz Soroushian for
the advice, guidance and assistance given at every stage of this research. I wish to extend
my gratitude to other members of the committee; Dr. Ronald S. Harichandran, Dr. Nicho-
las J. Alterio, and Dr. Roy V. Erickson for their interest and comments during this

research.

Financial support for the performance of this research was provided by Michigan Depart-
ment of Natural Resources and Research Excellence Fund of the State of Michigan. The
ﬁbers used in this research were provided by The American Fillers and Abrasives and
Interﬁbe Corporation. These contributions are greatly acknowledged. The technical sup-
port provided by the Composite Materials and Structure Center of Michigan State Univer-
sity is also gratefully acknowledged.

I wish to extend my special acknowledgment to Engineers in Chier Branch, Pakistan for
providing me with an opportunity to undertake the doctoral studies.

I also wish to acknowledge the efforts of my father Ambar, my mother Bano, and my elder

brother Musammar for their love and encouragement at all stages.

I wish to thank my wife Ismat who shared all the moments with me during this research
program for her constant love, care, patience and support, and also my three loving sons
(Sarosh, Uzair and Zuhair) for their love and my other members of family for their kind-

BESS.

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... Vii
LIST OF FIGURES ....................................................................................................... .xi
CHAPTER 1 INTRODUCTION .................................................................................... l
1.1 Environmental Aspects of Wastepaper ................................................. 1
1.2 Wastepaper Applications in Cement Products ....................................... 6
1.3 Aim, Scope and Signiﬁcance of Research Study ................................. 16
cm 2 LITERATURE REVIEW .................................................................... 17
2.1 Wastepaper Fibers in Cement. ............................................................. 17
2.2 Potential Problems and Research Needs .......................................... 23
2.3 Theoretical Considerations ................................................................. 37

CHAPTER 3 DETERMINATION OF INFLUENTIAL VARIABLES IN THE
PROCESSING OF RECYCLED CELLULOSE FIBER-CEMENT..55

 

 

 

 

 

 

3.1 Introduction ....................................................... -- 55
3.2 Variables and Experimental Design ............................ - 56
3.3 Recycled Wastepaper Frbers . ...... 65
3.4 Experimental Set Up ................................................. .. 72
3.5 Test Results and Statistical Analysis ................................................... 73
3.6 Discussion of Results .......................................................................... 92
3.7 Summary and Conclusions .................................................................. 92
CHAPTER 4 OPTIMIZATION OF INFLUENTIAL VARIABLES ...................... 94
4.1 Introduction .......................................................................................... 94
4.2 Optimization Experimental Program ................................................... 95
4.3 Test Results and Analysis .................................................................... 96
4.4 Evaluation of the Optimized Composite ........................................... 106
4.5 Technical Evaluation of Recycled Composites .................................. 113
4.6 Summary and Conclusions .............................................................. 114
CHAPTER 5 DURABILITY & MOISTURE SENSITIVITY. .............................. 115
5. 1.1ntroduction....... ................................... 115.
5. 2 Experimental Methods ............................... 116
5.3 Moisture and Ageing Effects on Engineering Properties .................. 123

5.4 Microstructural and Compositional Changes Under Moisture

and Ageing Effect - _ - ............................................... 151

 

5.5 Ageing Mechanisms .......................................................................... 165
5.6 Improvement of Durability ................................................................ 167
5.7 Summary and Conclusions ................................................................ 187
CHAPTER 6 COST ANALYSIS ................................................................................. 194
.6.1 Cost of Recycled Wastepaper Frber-Cement Composites ................ 194
6.2 Comparative Cost Analysis .............................................................. 195
6.3 Life Cycle Costs ................................................................................ 195
6.4 Summary and Conclusions ................................................................ 198
CHAPTER 7 SUMMARY AND CONCLUSIONS ................................................... 199
APPENDIX I THEORETICAL CONSIDERATIONS . ........................................ 209
APPENDIX 11 NOTATIONS AND SYMBOLS ..................................................... 219
APPEHIDIH III STANDARD SPECIFICATIONS ............................................. 220
BIBLIOGRAPH! ............................................................................................................ 221

vii

Table 1.1

Table 2.1

Table 2.2

Table 2.3

Table 3.1

Table 3.2

Table 3.3

Table 3.4

Table 3.5

Table 3.6

Table 4.1

Table 4.2

Table 4.3

Table 4.4

Table 5.1

Table 5.2

LIST OF TABLES

Comparison of Cost and Strength of Virgin Cellulose

and Other Fibers [l9] ................................................................................ 13
Fiber Length Data [26] .............................................................................. 18
Wood Fiber-Cement Composites: Hot Water Soak Test [32] .................... 29

Properties of Wood Fiber Cement Products Exposed to Different Ageing

Conditions [34] ......................................................................................... 36
Fractional Factorial Design of Experiments .............................................. 62
Sand Gradation .......................................................................................... 64
Properties of Binders ................................................................................. 64
Recycled Fiber Length Distribution (weight%) ........................................ 66

Flexural Performance of Recycled Wastepaper
Fiber—Cement Composites ......................................................................... 73

Results of the Analysis of Variance (Flexural Strength, Toughness and

Initial Stiffness) ........................................................................................ 90
Optimization Experimental Program ......................................................... 95
Flexural Performance ................................................................................ 97

Results of the Analysis of Variance (Flexural Strength, Toughness and

Initial Stiffness) ...................................................................................... 103
Technical Evaluation of Optimized Composite“: ................................... 113

Flexural Performance of Recycled Wastepaper Fiber-Cement Composite
at Diﬂ'erent Moisture Conditions ............................................................. 124

Results of the Analysis of Variance

viii

Table 5.3

Table 5.4

Table 5.5

Table 5.6

Table 5.7

Table 5.8

Table 5.9

Table 5.10

Table 5.11

Table 5.12

Table 5.13

Table 5.14

Table 5.15

Table 5.16

Table 5.17

(Flexural strength, toughness and stiffness) ............................................ 126

Effects of Repeated Wetting-Drying on Flexural Performance of Recycled
and Virgin Fiber-Cement Composites .............................. - 129

 

Results of the Analysis of Variance
(Flexural strength, toughness and stiffness) ............................................ 131

Effects of Repeated Freeze-Thaw Cycles on Flexural Performance of
Recycled and Virgin Fiber-Cement Composites ..................................... 133

Results of the Analysis of Variance
(Flexural strength, toughness and stiffness) ............................................ 135

Effects of Repeated Wetting-Drying and Carbonation on Flexural
Performance of Recycled and Virgin Fiber-Cement Composites ......... 137

Results of the Analysis of Variance
(Flexural strength, toughness and stiffness) ............................................ 139

Effects of Hot Water Bath Immersion on Flexural Performance of Recycled
and Virgin Fiber-Cement Composites....... - - 141

 

Results of the Analysis of Variance
(Flexural strength, toughness and stiffness) ............................................ 143

Results of the Analysis of Variance of Accelerated Ageing Tests

(Flexural strength, toughness and stiffness) ............................................ 146
Permeability Coefﬁcient of Virgin and Recycled Composites ................ 149
Results of Analysis of Variance of Permeability coefﬁcient ................... 150
Thermogravimetric Compositional Analysis .......................................... 151

Flexural Performance of Recycled Wastepaper Fiber-Cement

Composite After Higher Silica Fume Substitution ................................. 169
Results of the Analysis of Variance
(Flexural strength, toughness and stiffness) ............................................ 172

Flexural Performance of Virgin and Recycled Wastepaper Fiber-Cement

 

Table 5.18

Table 5.19

Table 5.20

Table 5.21

Table 6.1

Table 6.2

Table 6.3

Table 6.4

 

Composite Using Regular and Special Cement. ...................................... 178

Results of the Analysis of Variance

(Flexural strength, toughness and stiffness) ............................................ 181
Thermogravimetric Compositional Analysis .......................................... 185
Results of Analysis of Variance (T GA) ................................................... 186
Results of Analysis of Variance .............................................................. 187
Cost Comparison with Alternate Siding Materials in Market ................ 195
Life Cycle Cost Analysis ......................................................................... 196
Loss in Product Value In Design Life ...................................................... 196

Net Product Value After 50 years ............................................................ 197

Figure 1.1

Figure 1.2

Figure 1.3.

Figure 1.4

Figure 1.5

Figure 1.6

Figure 1.7

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 2.8

Figure 2.9

LIST OF FIGURES

State Landﬁll Capacity Map [3] ................................................................. 3
Landﬁll Capacity in the Year 2000 [3] ........................................................ 4
Application Areas of Cement Products Reinforced with Cellulose

Fibers [21] ................................................................................................. 10
Reinforcement Action of Fibers in Cement-Based Materials [23] ........... 11

Flexural Performance of Cellulose Fiber—Cement Composites Compared

with Glass Fiber Reinforced Cement and Asbestos Cement [18] ......... 12
Structure of Wood and Wood Fiber [16] .................................................. l4

Geometry and Appearance of Major Types of Fibers in Softwood and

Hardwood [24] .......................................................................................... 15
Flexural Strength Versus Frber Content [26] ............................................ 20
Fracture Toughness Versus Fiber Content [26] ......................................... 21
Water Absorption and Density[26] ........................................................... 22

Schematic Sketch of the Decomposition of Natural Fibers
in the Alkaline Pore Water of concrete [31] ............................................. 25 .

CBI Climate Box [31] ............................................................................... 26

Flexural Strength of the Composite Reinforced with Sisal Fibers
After Wetting-Drying Cycles [31] ............................................................. 28

Flexural Strength and Young’s Modulus of Air Cured and
Autoclaved Products Exposed to Natural Weathering [34] ...................... 31

Scanning Electron Micrographs of Fractured Surfaces
After Accelerated Weathering in Ambient Environment [35] .................. 33

Effect of Carbonation on Wood Fiber Reinforced Cement Sheets [32]...34

Figure 2.10

Figure 2.11

Figure 2.12

Figure 2.13

Figure 2.14

Figure 2.15

Figure 2.16

Figure 2.17

Figure 2.18

Figure 2.19

Figure 2.20
Figure 3.1
Figure 3.2
Figure 3.3

Figure 3.4

Brittle Fracture in a Composite After Accelerated Ageing in
a C02 Rich Environment [34] -- ......... - ........... 37

 

The ‘Pinching Eﬂect’ and Interfacial Shear Stress Distribution Predicted
by In the Arrest of Crack Propagation in a Matrix Between

 

the Fibers [57] .............................. .............................. 41
Effect of Fiber Spacing on First Crack Stress Ratio [61] .......................... 43
Complex Crack Patterns at the Interaction of an Advancing
Crack and a Fiber Lying in its Path [63] ................................................... 44

Idealized Representation of an Advancing Crack and the
Stress Field Around it, in a Fiber Reinforced Cement [64] ....................... 44

Schematic Description of a Traction Free Crack with a Closing Pressure to
Model the Fracture Behavior of Fiber Reinforced Cement [60] .............. 45

Schematic Description of the Model Used to Consider the Pull Out
Problem In Terms of Fracture Mechanics Concepts, With a Propagating
Bonding Crack of Length b [65] ............................................................. 47

Experimental Vs. Theoretical Predictions of Different Properties of
Cellulose Fiber Cement Composites [71] ................................................. 49

Relation Between Experimental and Calculated Tensile Strength in
Coir Fiber Reinforced Cement[72]. ........................... 50

 

Typical Load (P) Against Displacement (8) Records For Crack Propagation
in Cellulose Fiber Cements [73] ............................................................... 52

Crack Growth Resistance (K) Plotted Against Crack Extension (Aa) [73154
Components of the Laboratory Scale Manufacturing Process ............... 57
Manufacturing Process - Slurry Dewatering ............................................ 65
Fiber Length Distribution: Average Values and 95% Conﬁdence

Intervals ................................. ~ ................................... 67
Scanning Electron Micrographs of Virgin and Recycled Cellulose Fibers68

 

xii

Figure 3.5
Figure 3.6

ﬁgure 3.7
Figure 3.8
Figure 3.9
Figure 3.10
Figure 3.11
Figure 3.12
Figure 3.13
Figure 3.14
Figure 3.15
Figure 3.16
Figure 3.17
Figure 3.18
Figure 3.19
Figure 3.20
Figure 3.21
Figure 4.1
Figure 4.2

Figure 4.3

Fiber Compositional Analysis by Thermogravirnetry ............................... 70

 

 

 

 

Flexural Test Setup - .......... - -72
Flexural Strength Test Results ................................................................... 76
Flexural Toughness Test Results ............................................................... 77
Initial Stiffness Test Results ...................................................................... 77
Typical Flexural Load-Deﬂection Curves ................ - - - 78
Trends in Fiber Source Eﬂ'ects .................................................................. 79
Trends in Fiber Mass Fraction Eﬁ‘ects ........ ...................... 80
Trends in Fiber Reﬁnement Eﬁ'ects .......................................................... 81
Trends in Fiber Substitution Level Effects ............................................... 82.
Trends in Sand Maximum Size Effects ..................................................... 83
Trends in Sand/Binder Ratio Effects ......................................................... 84
Trends in Silica Fume/Binder Ratio Effects .............................................. 85
Trends in Flocculating Agent/Binder Ratio Effects .................................. 86
Trends in Vacuum Level Effects ............................................................... 87
Trends in Compaction Pressure Effects ..................................................... 88
Trends in Curing Condition Effects .......................................................... 89
Typical Load-Deﬂection Curves ............................................................... 96
Optimization: Response Surface Analysis ................................................ 98
Flexural Performance of the Optimized Recycled Composite

versus Virgin Composite .......................................................................... 104

 

Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8
Figure 4.9
Figure 4.10

Figure 4.11

Figure 4.12
Figure 4.13
Figure 5.1
Figure 5.2
Figure 5.3
Figure 5.4
Figure 5.5

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

Environmental Chamber For Conditioning of Test Specimens ............... 105

Flexural Load-Deﬂection Curves ........................................................... 106
Flexural Strength Vs. Fiber Mass Fraction .............................................. 107
Flexural Toughness Vs. Fiber Mass Fraction .......................................... 108
Initial Stiﬂ'ness Vs. Fiber Mass Fraction ................................................. 108
Density Vs. Fiber Mass Fracion .............................................................. 109
Water absorption Vs. Fiber Content. ....................................................... 110

Correlation between Density and Water Absorption in

 

Recycled Composite _- ....................................... 110
Dimensional Stability Test Results .......................................................... 111
Moisture Content .................................................................................... 112
Wetting/Drying Experimental Set Up ..................................................... 118
Freeze/Thaw Test Apparatus ................................................................... 119

Carbonation Chamber Producing Rich Carbon Dioxide Environment...120
Hot Water Bath ........................................................................................ 121
Direct Water Permeability Test Set Up .................................................. 122

Moisttue Effects on Flexural Behavior of Recycled and
Virgin Wood Fiber-Cement Composites - - - ................ 127

 

Eﬂ‘ect of Repeated Wetting and Drying Cycles on Flexural
Behavior ................................................................................................... 132

Eﬁ‘ects of Repeated Freeze-Thaw Cycles on Flexural
Behavior ................................................................................................... 136
Effects of Repeated Wetting - Drying and Carbonation on Flexural

 

Figure 5.10
Figure 5.11
Figure 5.12

Figure 5.13

Figure 5.14
Figure 5.15
Figure 5.16
Figure 5.17

Figure 5.18

Figure 5.19

Figure 5.20

Figure 5.21

Figure 1.1

Figure 1.2

Figure 1.3

Figure 1.4

Behavior .................................................................................................. 140

Effects of Hot Water Bath Immersion on Flexural Behavior .................. 144
Comparison of Different Ageing Methods ............................................. 147
Water Permeability Coefﬁcient Test Results ........................................... 151

Scanning Electron Micrographs of Fracture Surface Under Various

 

Accelerated ageing Conditions .......................................... 152
Thermogravimetric Analysis: Typical Weight Loss Curves .................... 156
Correlation of Composition with Engineering Properties ....................... 158
Correlation Between Various Engineering Properties ............................ 162

X-Ray Diﬁ'raction Pattern ........................................................................ 165

Eﬁect of High Silica Fume Content on Flexural Performance,

 

 

Durability and Moisture Sensitivity - ..... 174.
TGA Silica Fume Modiﬁed .................................................... 176
Eﬁ‘ect of Using Special Cement on Flexural Performance,

Durability and Moisture Sensitivity ....................................................... 183
TGA Special Cement Composite ............................................................ 185

Spacing of Fibers .................................................................................... 211

Empirical Relationship Between the Tensile Strength of Steel Fiber
Reinforced Concrete and the Spacing of the Fibers[89] ......................... 213
Virgin Versus Recycled Cellulose Fibers Arresting Cracks .................... 214

Assumed Critical Crack Model Controlling Fracture of Fiber-Reinforced

Concrete Composite ................................................................................. 218

XV

CHAPTER 1

INTRODUCTION

1.1 ENVIRONMENTAL ASPECTS OF WASTEPAPER

1.1.1 Magnitude of Solid Waste Problem

Approximately 250 million tons of residential, commercial and industrial wastes are gen-
erated in the US. each year [1]. The corresponding number for Michigan is 11.8 million
tons per year [2]. leaving industrial discard aside, EPA estimates that residential and com-
mercial wastes account for around 160 million tons. This ﬁgure is projected to reach about
193 million tons by the year 2000 [3].

In 1986 only 10 percent of all municipal solid waste was recycled and 10 to 15 percent
was incinerated (mostly with energy recovery), while almost 80 percent - about 130 mil-
lion tons — was disposed of in landﬁlls. The Michigan recycling (about 12 percent) and in-

cineration (about 19 percent) levels roughly correspond to the national values [4].

1.1.2 Wastepaper
1.1.2.3 Contribution of the Paper and Paper Products to the Solid Waste Stream: The

available data on different materials excavated from landﬁlls indicate that paper and paper
products have increased steadily and now comprise approximately 55 percent by volume
(and almost one-half by weight) of the materials excavated [3]. Information on speciﬁc
types of paper products excavated from landﬁlls suggest that glossy magazine paper has

increased steadily to comprise 2.5 percent of landﬁlled municipal solid waste.

U.S. paper industry represents 30% of world capacity [5]. U.S. primary paper and paper
board products in 1989 were 82.44 million tons, and are expected to be 91.07 million tons
in 1992 [5]. The increase in waste paper use is expected to rise twice as fast as other ﬁbers
[5]. Within two years Michigan is expected to enjoy a rise in total paper and paper board
capacity from 17.9 to 22.3%. Annual waste paper use in Michigan will increase from
20.87 million tons to 24.98 million tons in 1992 [5]. In Michigan there are total of 35 pa-
per mills producing 9621 tons of paper per day of which 27 mills use recycled paper utiliz-
ing 4986 tons per day [6].

On the world-wide basis, recycled waste paper is already the largest ﬁber stream used for
paper making. Fiber consumption by grade in 1987 was: recycled ﬁber 31%; bleached
kraft 26%; un-bleached kraft 15%; mechanical pulp 14%; semi-chemical pulp 4%; sulﬁte
pulp 5%; and non-wood 5%. Recycling rates are low -- around 10% - in ﬁber rich coun-
tries such as Finland and Canada, and range upward to 50 percent in ﬁber poor countries,

such as Europe and Japan. The US. falls in between these extremes, at about 27%.

Although there are over 50 different grades of wastepaper, these grades are generally di-
vided into ﬁve categories [7]: (1) News-predominantly old newspaper (ONP), (2) Corru-
gated- includes old corrugated containers (OCC), (3) Mixed- color papers, magazine, and
envelopes, (4) pulp substitutes, and (5) High grade deinking. Paper’s inherent qualities
(strength, durability, printability etc.), and ultimately its utility, are determined by its ﬁber
composition e. g., softwood or hardwood, bleached or unbleached, virgin or recycled. Soft-
wood ﬁbers, for example, impart strength and tear resistance and printability. Newsprint
generally includes 80-85% ground pulp (mechanical wood pulp) [8]. Fiber length is short-
ened in the grinding process and 1i gnin, which remain in the unprocessed pulp, hastens its
deterioration. Papers made from the ground wood pulp are weak to begin with and deteri-
orate easily upon ageing. Magazine paper ﬁbers are mostly from chemical pulp. Generally,

a mix of softwood and hardwood ﬁbers are used to get the desired performance.

Recycling is the ﬁrst step to reduce the solid waste dilemma, because both the landﬁlls and
incinerators are becoming increasingly costly because of new requirements and siting dif-
ﬁculties. The important point to recognize is that recycling is not the only solution for the
solid waste problem. Recycling is only part of what must be a comprehensive plan to bat-

tle solid waste.

Magazine papers are often coated for aesthetic reasons. Coated paper is smoother, has a
ﬁner pore structure and, therefore, the quality of print is much enhanced. Conventional
coatings consist of clay and pigment particles bound with a latex or a soluble binder such
as starch [8,9]. The use of coated waste papers in the paper mills results in the production
of large amounts of sludge, which can amount to as much as 30 percent of the input by

weight [1]. This sludge then becomes industrial waste.

1.1.2.2 Landﬁll Crisis: Landﬁlling has been the most available disposal method, but
many areas of the country are experiencing shortfalls of permitted landﬁll capacity and
rising landﬁll costs. The “capacity crises” has become a signiﬁcant concern around the
country, particularly in the Northeast (Figure 1.1). According to EPA, for example, more
than half of our existing landﬁlls will reach their capacity within eight years.

',.

1..
‘1 j
I,‘ . : ~
.. ‘nn
1» ', J19.“ vr.‘
. r,
.1411], . . :
. 'l ‘- bh
v . u ‘l
.v. ' '
I‘ll-
rl ‘

a}:
331;?

  

I Less Than 5 Years
5-10 Years
[3 Greater Than 10 Years

 

Figure 1.1 State Landﬁll Capacity Map [3].

Assuming that the existing landﬁlls close at their current rate (11.2 million tons per year of
lost capacity) and the new facilities are built at the same pace as in the recent years (4 mil-

lion tons annually), at today’s recycling and recovery levels, our disposal requirements

will exceed existing capacity by around 1998 [1,3]. For every ton of recycled paper pro-
duced, landﬁll space is reduced by three cubic yards [7]. To bring our disposal needs
roughly into balance with the supply of landﬁll space, other steps must be taken, such as
meeting EPA’s national recycling goal (25 percent, which is also the median statewide
goal in Michigan) within the next two or three years, and tripling the number of waste-to

energy plants (Figure 1.2).

1.1.2.3 Recycling of Paper: Paper and paperboard account for a larger fraction of munici-
pal solid waste than any other single category of material. In 1987, total wastepaper recov-
ery (including pre-consumer waste) in the United States reached 24 million tons, a
recovery rate of 28.5 percent [1]. Recovered wastepaper, or secondary ﬁber, is used to pro-
duce new paper products, construction materials, animal bedding, insulation, etc. Current-
ly in Michigan old corrugated cardboard is recovered at the highest rate, 59 percent of the
generation [10]. The recovery rates are much less for glossy magazine paper, coated old
corrugated cardboard, and mixed paper [10]. Certain trends in recovery and utilization of
wastepaper together with some technical and economic factors, as described below, are

making it necessary to look beyond paper mills for the utilization of various grades of

wastepaper.
2N1
- 1! Cam“
3;...
1 r incl-almanac:
g: I tad-.mmommm
2
5 NIH z 25%“
_._ i /’ 2591mm
. J zssm
504 355mm

 

Figure 1.2 Landﬁll Capacity in the Year 2000 [3].

In North America, as more local recycling programs come on line, the old newsprint mar-

ket is being saturated. Within the next ﬁve years, the supply of old newsprint will reach 20

million tons annually. About one-fourth of this total will be recycled by domestic mills.
Combined with export of old newsprint, this means recycling old news into new paper

only accounts for one-third of the total output [11].

As higher amounts of old newsprint are being recovered, this relatively cleaner wastepa-
per is substituting other grades (e. g., mixed paper) in mills producing boxboard and roof-
ing products, and also in the export markets [11]. True progress in solid waste
management can not be made by substituting 1 ton of mixed paper for 1 ton of old news-

print in our landﬁlls.

Paper is not a closed-loop recyclable like glass. It can not be used indeﬁnitely for its origi-
nal use, because the cellulose ﬁbers break down into smaller pieces each time the paper is
recycled [11, 12].

The use of glossy magazines and coated old corrugated cardboard for the reproduction of
paper products in paper mills results in the production of large amounts of sludge, which
can amount to as much as 30 percent of the input by weight [1]. The wet sludge, produced
while separating the coating material currently has no commercial value and has been re-
moved for disposal in landﬁlls [13]. This generally makes the use of glossy magazines and

coated old corrugated cardboard in paper mills uneconomical.

Markets for paper are historically volatile, causing major ﬂuctuations in the prices of vir-
gin wood pulp; prices for wastepaper are even more volatile than those for wood pulp (this

kind of relationship is typical of secondary materials) [1,14].

Other barriers to increased recycling of various grades of paper for reproducing paper
products include capital investment in equipment, competitiveness of market, higher lev-
els of contaminants found in news supplies of wastepaper, and restrictive speciﬁcations
and rules [1,10]

Given the above constraints against recycling wastepaper into paper products, and the in-
creased collection of wastepaper, we have to look beyond the paper industry and broaden
the base of paper recycling. Advantages can be taken of the insulating, absorptive, and re-
inforcing qualities of recycled wastepaper to produce a whole range of products from in-
sulation to cement boards [11, 12, 13]. These non-paper applications can grow in

importance as increasing supplies of recovered wastepaper becomes available.

 

1.2 WASTEPAPER APPLICATIONS IN CEMENT PRODCUTS

1.2.1 Why Look into Cement Products Than Other Forms of Recycling

Recycling of wastepaper into non-paper products generally starts with grinding up the pa-
per with augers or hammer mills and screening them to various sizes for different prod-
ucts. In some applications chemicals are added to give cellulose ﬁbers different properties.
Non-paper products that can be produced using wastepaper include cellulose insulation,
animal bedding, mulch, cat litter, ﬁre logs, worm bedding, rooﬁng materials, absorbent for
hazardous wastes and sludges, and thin sheet cement products. These non-paper markets

for secondary ﬁbers are brieﬂy analyzed below.

1.2.1.1 Cellulose Insulation is made by grinding newsprint and then adding some addi-
tives as ﬂame retardants. Old newsprint seems to be more suitable than other grades of
wastepaper (e. g., magazine) because of its insulation properties. The market for cellulose
insulation has declined in the recent years due to the popularity of ﬁberglass insulation,
and also because of the loss of a healthy retroﬁt market after the federal government elim-
inated the energy tax credits in December, 1985 [11]. Also cellulose insulation is inferior

to glass ﬁber, rock wool and polyurethane insulations [15].

1.2.1.2 Mulch made from wastepaper is generally made by shredding and ﬂufﬁng the pa-
per, mixing dye and water with the ﬂuff, and then adding the seed prior to spraying the
mulch. All grades of wastepaper are reported to be used. Currently the majority of seeding
uses straw mulch and, partly due to economic factors, it is highly unlikely that State’s use
of recycled mulch will increase much beyond the one to ﬁve percent (less than 1000 tons

per year) currently being used [10].

1.2.1.3 Cat Litter is another non-paper product that can take advantage of the high absor-
bency and biodegradability of wastepaper; particularly old newsprint. Sand and wastepa-
per are the primary ingredients of the litter. The materials are combined, then formed into
granules. The impact that this product could have on the Michigan waste stream appears

insigniﬁcant, mainly because its potential market is very small [10].

 

1.2.1.4 Other Products: Fire Logs are also being produced from old newsprint to replace
raw ﬁrewood for residential uses (Goldgerg, 1989). A new product made with wastepaper
is warm bedding, which is used to ﬁll boxes in which the “herds” are shipped. Some re-
search has also been done on using old newsprint to clean up hazardous liquids and slud-
ges. Rooﬁng products such as shingles, ply felts, base ply and cap sheets are also made
from wastepaper. This market, however, has declined sharply in the recent years mainly

due to the popularity of ﬁberglass rooﬁng products [13].

In summary, non-paper products, other than thin-sheet cement products, appear to have lit-
tle impact on the solid waste stream [10, 11]. These markets are either too small or they
are declining, and can only be effective in some small or concentrated areas. Many of
these markets also use old newsprint as the dominant wastepaper grade, and will not be
suitable for other grades such as glossy magazine papers. Increased recovered volumes of
wastepaper and the limited capacity that can be offered by paper mills for wastepaper in
general and coated grades of wastepaper (glossy magazine and coated corrugated card-
board) in particular, however, provide strong incentives for searching alternative markets
for wastepaper [1,11]. The use of recycled paper as reinforcing ﬁbers in thin—sheet cement
products presents an attractive market for wastepaper. Relatively large volumes of differ-
ent wastepaper grades can potentially replace some costly and energy-intensive virgin ﬁ-
bers currently used by the thin-sheet cement products industry, which is still suffering
from the elimination of asbestos ﬁbers from the US market. Potentials for the utilization

of secondary cellulose ﬁbers in thin-sheet cement products are discussed below.

1.2.1.5 Thin-Sheet Cement Products: World use of hydraulic cements is close to one bil-
lion tons per year [16] and, along with steel and wood, they are the most important con-
struction materials used today. It has been proposed that the use of cement could be
doubled by the year 2000. The low cost and ready availability of raw materials (limestone,
clay, etc.), the fact that the energy consumed for the manufacture of cement is consider-
ably less than for metals and plastics, and that hardening takes place with water at ordinary
temperatures, provide the incentives for widespread use of cement products. Cement-
based materials, however, suffer from one common shortcoming, they fail in a brittle man-
ner under tensile stresses or impact loads. An effective approach to resolve this problem
involves the use of short, randomly distributed reinforcing ﬁbers in cement-based materi-
als.

 

The reason why brittle materials like cement are made stronger by very small addition of
ﬁbers is that cracks are stopped or deﬂected by the presence of ﬁbers and, as a result, the
toughness and tensile strength are dramatically increased. Fibrous cement materials, with
their desirable toughness characteristics and cracking resistance, have found broad appli-
cations for the construction of various thin-sheet products including cladding panels, parti-
tion components, ceilings and walls, garden fencing, silo lining, green house panels,
ducting, drainage and irrigation channels, tiles, pipes, water troughs and ﬁttings, and labo-
ratory surfaces. The global use of thin-sheet ﬁbrous cement products is estimated at 2.5

million tons each year, consuming about 0.5 million tons of reinforcing ﬁbers annually.

1.2.1.6 Replacing Asbestos: World—wide, the asbestos cement sheet industry has been
searching for an alternative reinforcing ﬁber owing to the health risk associated with the
use of asbestos [1?]. Consumption of asbestos for cement reinforcement reached 1.5 mil-
lion metric tons of ﬁber in 1970’s [18]. Usage was principally for factory-made cement
cladding panels and pipes produced in some 800-900 manufacturing units operating virtu-
ally in every country of the world. The asbestos replacement activity in recent years has
resulted in vast world-wide research into alternative cement reinforcement ﬁbers. Virgin
cellulose, glass and polyethylene are among the ﬁbers considered to substitute asbestos in
cement products. Cellulose ﬁbers derived from softwood and hardwood, being fairly
strong and stiff as well as cheap and relatively energy-efﬁcient, have emerged as the dom-

inant ﬁber types currently used in non-asbestos thin-sheet cement products [19].

Preliminary research studies conducted in Japan and Australia have demonstrated the po-
tentials of recycled ﬁbers obtained from various grades of wastepaper to substitute virgin
cellulose ﬁbers in thin sheet cement products [20, 21]. While Japanese workers uses a
combination of recycled cellulose and asbestos ﬁbers[l9, 20, 21], the research in Australia
concentrated on full substitution of all virgin ﬁber types with recycled wastepaper (includ-
ing coated magazine paper) in thin-sheet cement products [19]. Wastepaper ﬁber rein-
forced cement has shown to provide mechanical and physical performance characteristics
close to those obtained with virgin cellulose ﬁbers. The relatively high ﬁne contents of re-
cycled ﬁbers act more like ﬁllers than reinforcing ﬁbers, and thus higher ﬁber contents
were found to be needed in the case of recycled ﬁbers to reach optimum levels of mechan-
ical performance comparable to those obtained with virgin ﬁbers. The promising results of

the preliminary studies conducted so far on the use of recycled wastepaper as reinforcing

ﬁbers in cement products encourage more thorough investigation of on such critical as-
pects of the material behavior as long-term durability, moisture-sensitivity and dimension-
al stability. The resulting technical information should be accompanied with economical
feasibility studies in order to facilitate commercial applications of wastepaper ﬁber rein-
forced cement products in applications where recycled wastepaper presents a viable alter-

native to the virgin ﬁber types.

The uses of wood ﬁber reinforced cement sheets are diverse. They range from major

components in industrial manufacturing to uses in commercial, residential and agricultural
construction (Figure 1.3). The desirable ﬂexural strength and toughness characteristics, di-
mensional stability, ﬁre resistance and impact strength of wood ﬁber-cement composites
suggest that they could be valuable in areas of application with demanding requirements on
materials. Heat shields and spray booths, sound barriers and modular ﬂooring, duct lining
and air shafts, gaskets and seals, laboratory tops and splashbacks, and ﬁre walls in dry kilns
are some of the typical industrial components made of wood ﬁber reinforced cement com-
posites. Commercial and residential uses of wood ﬁber reinforced cement is mainly for the
production of ﬂat and corrugated sheets rooﬁng elements, exterior and interior wall panel-
ling, equipment screens, fascias, facades and sofﬁts. substrate for tiles, window sills and
stools, stair treads and risers, substrate for applied coatings, and utility building cladding
panels. Agricultural uses of wood ﬁber reinforced cement composites are mainly for farm
buildings sidings, stalls and walls, poultry houses and incubators, green house panels and
work surfaces, and fencing and sunscreens.
The use of recycled paper as reinforcing ﬁbers in thin-sheet cement products, where cur-
rently about 0.5 million tons of virgin ﬁbers are used annually, presents an attractive mar-
ket for wastepaper, with potentials to divert about 10 percent of the waste magazine paper
from US. landﬁlls.

 

 

 

 

 

 

 

 

 

i

 

: - "--~ ' , ¥ I
SUBSTRATE FOR APPUED COATINGS EQUIPMENT SCREENS, FBCIAS s FACADES

Figure 1.3 Application Areas of Cement Products Reinforced with Cellulose Fi-
bers[21].

 

1.2.2 Concepts of Fiber-Reinforced Cement

Reinforcement with short, randomly distributed ﬁbers presents an effective approach to
the solution of problems with the brittle nature of failure in cement products [21, 22, 23].
These ﬁbers are effective in stopping and deﬂecting the cracks propagating inside cemen-
titious matrices, there-by substantially enhancing the toughness characteristics and crack—

ing resistance of the material (Figure 1.4).

Fibrous cement products using virgin cellulose, polyethylene and glass ﬁbers have found
broad applications for the construction of various thin-sheet products such as cladding
panels, sidings and sofﬁts, rooﬁng tiles, ducting, drainage and irrigation channels, tiles-
substrate, pipes, water troughs and laboratory surfaces. The reinforcement action of the

relatively low—cost cellulose ﬁbers in cementitious matrices is quite good relative to other

ﬁbers such as glass (Figure 1.5).

 

 

m
Will!

1010

WITH!

 

 

 

 

 

 

 

(a) Crack Arrest (b) Tensile Behavior

Figure 1.4 Reinforcement Action of Fibers in Cement-Based Materials [23].

 

12

Stress (MP1)

l

r’
I I

//3—\

° °'5 ‘ t5 2 2.5 a a5
Deflection (cm)

 

Glass Rem forced Cement

 

 

   

 

 

 

Cellulose-Cement

 

Figure 1.5 Flexural Performance of Cellulose Fiber-Cement Composites Compared
with Glass Fiber Reinforced Cement and Asbestos Cement [18].

1.2.2.1 Fiber Types

Cementitious materials are relatively brittle with relatively low tensile strength. Diﬂ'erent
types of ﬁber have been used to overcome this shortcoming. Asbestos, steel, glass, carbon
and synthetic and natural ﬁbers are among those ﬁber types successfully used for the rein-
forcement of cement-based matrices. Properties of virgin cellulose ﬁbers are compared in
Table 1.1 with those of other ﬁber types used in cement products. It is evident from the ra-
tio of cost to load carried by ﬁbers that virgin cellulose ﬁbers are highly cost effective.

This illustrates why they are now dominating the thin cement-products market [19].

1.2.2.2 Cellulose Fibers.

Figure 1.6 brieﬂy illustrates the structure of wood and wood ﬁber. If a piece of lumber is
considered, it may have defects (knots, cracks etc.); by selection, a piece of clear wood
(near macro defect-free) could be obtained with a tensile strength of say 70 Mpa (9.31
Ksi). However, single ﬁber which constitutes the reinforcing unit of bulk wood has been
tested and found to have tensile strengths greater than 700 Mpa (93.1 Ksi [16]. If one con-

siders cellulose as the basic molecule which makes up the ﬁber, and if one could express

the strength of the chemical bonds which make up the structure of cellulose in terms of

tensile strength, an even greater value of around 7000 Mpa (931 Ksi) would be recorded.

Table 1.1 Comparisons of Cost and Strength of Wood Fibers with Other Fibers [19]

 

 

 

 

 

 

 

 

Rel. Cost Speciﬁc Tensile Rel. Cost per
Fiber per Unit Gravity Strength“ ft/ 56 unit Weight
Weight (SG) (f‘, MPa) (ft/ 86)
= =

Wood 1 1.5 500 333 1
(Kraft Pulp)
Glass Rav- 4 2.5 1,400 560 2.2
ings
Steel 1.4 7.9 2,100 267 1.6
Kevlar Pulp 20 1.5 2.800 1.867 3.3
Asbestos 1.2 2.6 700 269 1.3
(M 5R)

 

 

 

 

 

 

 

 

" Realistic tensile strength values for commercial ﬁbers

Trees serve as the major raw material for cellulose ﬁbers. The trees harvested for the pro-
duction of cellulose ﬁbers are known commercially as “softwoods” and “hardwoods.”
Among commercial trees, softwoods are the source of so-called ‘long ﬁbers.’ The unbro-
ken cellulose ﬁbers in important softwoods range in length from about 2.5 mm (0.098 in.)
up to 7 mm (0.28 in.), but the vast majority of these ﬁbers average in length between 3 and
5 mm (0.12 and 0.20 in.). Even within the same tree species, ﬁber lengths can vary consid-
erably. Softwood cellulose ﬁbers have widths, or diameters, that range form about 15 to 80
microns (30 to 45 microns for most softwoods) [24]. Hardwoods yield cellulose ﬁbers
that, on an average are about 1/3 to 1/2 the length and about 1/2 the width of softwood ﬁ-
bers. Cellulose ﬁbers produced from hardwood also have higher ﬁnes content when com-

pared with those obtained from softwood.

14

 

 

 

Figure 1.6 Structure of Wood and Wood-Fiber [16].

Figure 1.7 provides information on the geometry and appearance of the major ﬁber types
in softwoods and hardwoods. All diagrams are at the same magniﬁcation to show the rela-

tive sizes of these elements.

The major chemical components of wood are cellulose, hemicellulose, lignin and a very
small fraction of extractives. The cells in their natural arrangement in solid softwoods and
hardwoods are bonded together by a layer of amorphous cementing material. It is this
bonding that must be broken in the cellulose ﬁber production (pulping) process, by either

chemical or mechanical means.

Pulping processes are classiﬁed as either chemical, semi-chemical, or mechanical. This
classiﬁcation refers to the nature of the deﬁberization process. In mechanical pulping, the
reduction of logs or chips to ﬁbers occurs by mechanical action which is usually aided by
thermal softening of the lignin between wood cells. No chemicals are added in mechanical
pulping to dissolve the lignin or any other wood component.Semi-chemical processes use
a combination of both chemical reactions and mechanical power [24]. Chemical Pulps
also called kraft pulps are commonly used in the production of book paper and writing,

while mechanical pulps are regularly used for the manufacture of newsprint.

15

1.2.2.3 Recycled Coated Magazine Paper Fibers

Recycled coated magazine paper ﬁbers are different from virgin cellulose ﬁbers in some
aspects. These have broader length distribution (mostly from 0.1 mm to 5 mm) and have
ﬁbers that are damaged and shortened as a result of recycling. In composition these “6-
bers” actually a combination of ﬁbers and impurities (e.g. clay) [25]. Cellulose ﬁber is the
major constituent comprising about 80% of mass. These ﬁbers are formed by mechanical
processing and are grey in color because of the coating pulverization. Recycled cellulose
ﬁber have surface area of 6-7 sq.m lgm and the pH value in 5% slurry has been found to be
about 7.2.

MALYPYUS m

“AWD
men

    

OM
TIAC’ED

 

 

Figure 1.7 Geometry and Appearance of Major Fiber Types of Fibers in Softwood
and Hardwood [24].

16

1.3 OBJECTIVE, SCOPE AND SIGNIFICANCE OF THIS RESEARCH STUDY

1.3.1 Objective: The main thrust of this research is to assess the technical feasibility of
using recycled wastepaper as reinforcing ﬁbers in thin-sheet cement products (to substi-
tute more expensive virgin ﬁbers such as glass, polyethylene and virgin cellulose). Rec-
ommendations are also made for the production and use of cement products reinforced

with recycled paper ﬁbers.

1.3.2 Scope: This research dealt with magazine wastepaper ﬁbers, which cause problems
in recycling back into papers due to their high level of impurities. Dry-processed maga-
zine paper ﬁbers were evaluated as reinforcement in thin-cement products. Short-term and
long-term performance characteristics of the composites were evaluated, and microstruc-
tural changes associated with the ageing processes were investigated. Reﬁned composites
were produced with recycled magazine ﬁbers with a desirable balance of short-term engi-

neering properties and long-term weathering resistance.

1.3.3 Signiﬁcance: Recycling in construction provides opportunities for long-term di-
rection of major volumes of market-limited (impure) waste from landﬁlls, and also for the
development of lower-cost and energy-efﬁcient construction products. A large-volume
component of the municipal solid waste stream (magazine paper) is targeted in this re-
search, which concerns high-value utilization of this waste product in thin cement prod-

ucts for residential and commercial construction.

CHAPTER 2

LITERATURE REVIEW

Every composite of materials represents an individual chemical system with its own set of
problems. In discussions on cement composites reinforced with recycled paper ﬁbers we
are focussing on cementitious environments which are alkaline in nature and a paper pulp
(wood ﬁber) which may be acidic and thus unstable in alkaline environments (if excess of
alum was used in its manufacture [9]) or basic. Whether paper lasts indeﬁnitely or brieﬂy
depends on the materials and methods used in its manufacture as well as the environments
in which it is stored. Since the early observations of Murray and the practical solutions
suggested by Sutenneister and Barrow it has been repeatedly demonstrated that additives
which create acidity within paper hasten its deterioration [9]. Acidic environments cata-
lyze hydrolytic degradation of the polymeric cellulose molecules, reducing their chain
length; even a few chain scissions per molecule cause a substantial loss of physical prop-
erties. Mildly basic environments such as calcium or magnesium carbonate minimize the
acid concentration and therefore the rate of the acid hydrolysis reaction. The cellulose
molecule can also suffer hydrolytic cleavage in an alkaline environment. Residual lignin

in wood pulp can also accelerate the degradation of paper [9].

2.1 WASTEPAPER FIBERS IN CEMENT

Wastepaper is shredded mechanically by a dry process to get the wastepaper ﬁbers. These
ﬁbers as compared to virgin cellulose ﬁbers are shorter, splintered and ﬂattened or col-
lapsed. The use of recycled magazines and newsprint as source of cellulose ﬁbers for the

reinforcement of cement products has been reported by Coutts [26] and Hirajirna et al

17

18

[22]. The main ﬁber constituents of the wastepaper ﬁber used by Coutts [26] were kraft
pulped P. radiata and mixed eucalypts, therrnomechanical pulped P. radiata, neutral sul-
phite semichemical pulped mixed eucalypts, and other ﬁber types present in small
amounts in waste products. The distribution of ﬁber lengths in the wastepaper sample is
shown in Table 1 along with the data for kraft P. radiata and kraft E. regnans virgin ﬁbers.
At least 10000 ﬁbers were analyzed for the wastepaper and E. regnans pulps, and in excess
of 5000 ﬁbers for the P. radiata pulp. The wastepaper ﬁbers are observed to have smaller
average lengths, and to contain a relatively a relatively large fraction of ﬁnes. For the pro-
duction of wastepaper ﬁber reinforced cement, the wastepaper was treated in the laborato-

ry in a Valley beater; the matrix was prepared from ordinary portland cement.

Table 2.1 Fiber Length Data [26]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Weighted
Distribution
(%)
Length (mm) P. radiata E. regnan Wastepaper
= ; L =
<0.2 2.6 2.0 6.6
0.2-0.6 4.0 12.8 21.2
0.6-1.2 9.1 74.2 40.0
1.2-2.0 18.0 9.0 15.3
2.0-3.0 28.0 1.5 11.0
3.0-4.0 25.0 - 5.2
4.0-5.0 11.0 - 1.0
>5.0 3.0 - -
Weighted Average 3.2 1.0 1.9
(mm)

 

 

 

Composites were prepared from the wastepaper ﬁbers using different ﬁber fractions (2-

16% ﬁber by mass). In earlier studies of cellulose ﬁber reinforced cement matrices, ﬂexur-

19

al strength was at a maximum value at about 8-10% ﬁber mass fraction [27, 28, 29]. The
wastepaper ﬁber reinforced cement (WPFRC) developed maximum ﬂexural strength of 18
MPa (2,600 psi) at about 12% ﬁber content by mass, when tested in a controlled atmo-
sphere (Figure 1). The effect of moisture on samples is to reduce ﬂexural strength to ap-
proximately 47-75% of the dry strength (Figure 1). This is in general agreement with the
reduction in strength observed with other air-cured wood ﬁber reinforced cement compos-

ites when tested both wet and dry [28, 29].

The wastepaper considered by Coutts [24] contained softwoods and hardwoods, but these
ﬁbers had experienced considerable damage due to processing and recycling as shown by
high ﬁnes content (Table 2.1). Much of the damaged material acts more as a ﬁller-diluent
than as a reinforcing ﬁber. The ﬂexural strength increased up to higher than usual ﬁber
loadings before the maximum value was reached at 12% ﬁber content by mass. This could
be due to the fact that much of the short material (say <0.6 mm) offers little reinforcement
to the composite and so a greater mass of ﬁber must be added to achieve sufﬁcient num-
bers of the longer reinforcing ﬁbers. The ﬂexural strength values were lower than those of
cement composites reinforced with the short E. regnans ﬁbers (Figure 2.1). This might be
attributed to the fact that although the numbers of long ﬁbers are sufﬁcient to provide high
strength, inefﬁcient ﬁber packing had already taken place due to the total volume of 6-

brous material.

A similar shift of the maximum ﬂexural strength value, with respect to ﬁber content, was
noted for P. radiata therrnomechanical pulp reinforced air-cured cement [30], the maxi-
mum being around 10% by mass. In that instance, the higher ﬁber content required and the
lower ﬂexural strength achieved were related to the lower number of ﬁbers for a given

mass (due to the extra mass of extractives not removed during the pulping process).

In addition to reduced ﬁber length, the altered surface properties of physically and chemi-
cally recycled ﬁbers my also inﬂuence both ﬂexural strength and fracture toughness prop-

erties of ﬁber-cement composites.

20

 

Fl EXURAL smeuam (uPal

 

I l
o l
l . o
20 l '
l
i
1. l ’3\ i
C 3 Q l
10 I
3 WAQTI 'I'II W'HC IN TI'TED ‘ ‘
l i
i / 0 WA." 'APII W'NC WET TEST!!! ‘
14 l . O i
. / . E. REGNANS W'NC IN TESTED (FIFTY, l
/ .
. / A i
12 i // i
2/
to I "’
a ' -/ \
6 I
I
‘ 1
O 2 ‘ . . IO 12 l 4 I.

FIBRE CONTENT ('6 BY MASS)

Figure 2.1 Flexural Strength Versus Fiber Content [26].

The mechanisms that take place during the fracture of a composite include ﬁber breakage
and ﬁber pullout. The latter can have a considerable inﬂuence on the value of fracture
toughness. If the ﬁber is short, then the energy used up in pulling the ﬁber through the ma-
trix after the ﬁber to matrix bond is broken can contribute little to the dissipation of energy
contained in the advancing crack. Therefore, the crack continues to propagate through the
sample and the material appears brittle. In Figure 2.2 a comparison is made between the
fracture toughness versus ﬁber content of wastepaper and E. regnans ﬁber reinforced ce-
ments The E. regnans ﬁber provides fracture toughness values almost twice those of
wastepaper ﬁber at comparable ﬁber mass fractions. This would conﬁrm the presence of
shortened ﬁbers derived from wastepaper. Of course, other factors such as reﬁning, which
takes place during paper manufacture, would have an effect on ﬁber performance; this has
been reported elsewhere [25]. Kraft E. regnans ﬁber reinforced cement composites [29]
produce fracture toughness values which are almost half those of the P. radiata ﬁber rein-

forced cement [30]. This would support the hypothesis that ﬁber length is highly important

21

in increasing of fracture toughness, noting that the average ﬁber lengths of E. regnans and
P. radiata being approximately 1.0 to 3.2 mm (0.04 to 0.1 in), respectively (Table 2.1). Al-
though the average ﬁber length in the wastepaper pulp is 1.9 mm (0.07 in), the amount of
ﬁne material (<0.6mm) constitutes almost 28% of the mass of the ﬁber whereas P. radiata

and E. regnans have values of only 6.6% and 14.8%, respectively (Table 2.1).

When the wastepaper ﬁber reinforced samples are tested wet, there is the expected in-
crease in fracture toughness value [28, 29]. The increase is, however, very small, never ex-
ceeding 40% of the dry value. With P. radiata reinforced samples [29], increases of up to
150% of dry values have been recorded at certain ﬁber loadings (6-8% by mass).

 

1 ‘3
LI!
9 l
1.61 . f

/ a 3 i

 

 

E

-

ﬁ

1. .

U 3.: t /

"’ /

m l

I“ ' Q a o

E 1.0 g _. O

-

d l

3 ' /

g 0.. l I 3 3

a l /

= I A

q

:- 0.‘ I _

2 i I 0 war: nun write an Turn

3 | /;///c a man emu mac we? rum ,

g 0.6 l ‘
i o I. neon“: we: rm rune (nan
: ./ I
i "‘ .

3 z 4 I O to 11 to 1.

F18!!! courarr 5 BY MAS!)

Figure 2.2 Fracture Toughness Versus Fiber Content [26].

The physical pr0perties of air-cured cellulose-ﬁber reinforced cement containing E. regn-

ans ﬁbers are compared with wastepaper reinforced cement composites in Figure 2.3.

Density, water absorption and porosity are all inter-related in so far as their magnitude de-

22

pends upon the free space or void volume present in the material. The higher density of
WPFRC could be attributed to the ability of the ﬁne ﬁber fragments to pack more efﬁ-
ciently with the matrix and so to produce fewer voids. This in turn would account for the
lower water absorption of WPFRC compared to the E. regnans reinforced material at any
given ﬁber content.

Hirajima et al [20] have also used reﬁned pulps recovered from municipal solid waste for
the production of thin-sheet cement boards. The results showed that there was no need to
modify or change the production process adopted for virgin cellulose ﬁber in the case of
newsprint waste utilization as reinforcing ﬁbers. Thin-sheet cement boards produced with
wastepaper ﬁbers fulﬁlled Japanese speciﬁcations. Reﬁned wastepaper pulp proved to be a

suitable reinforcing ﬁber for thin-sheet cement boards.

 

NATII AIIOI'TION
/ .

3

‘1‘

DENSHV to um

 

(‘5) NOUdUOSUV UiIVM

o at“?! up" wnc , .
l

O I. IIGMNS MIC (Ill?)

 

 

1 .
. ‘ s ’ to :2 n‘ ..

WERE CCNTENT B BY MASS)

Figure 2.3 Water Absorption and Density [26].

 

23

2.2 POTENTIAL PROBLEMS - WEATHERING EFFECTS

It is important to ensure that the improvements achieved in the properties of cement based
materials though recycled cellulose ﬁber reinforcement would be retained over a long pe-
riod of time in actual exposure conditions. In particular, one should be careful about the
afﬁnity of cellulose ﬁbers for moisture, their durability in the alkaline environment of ce-
ment, and the possibility of biological attacks on wood ﬁbers. As far as the biological at-
tacks are concerned, it is worth mentioning that no evidence is available to indicate that
natural ﬁbers can be decomposed biologically when used in cement materials [31]. There
are no test data available on durability characteristics of wastepaper ﬁbers reinforced ce-
ment composites. However, studies carried out on natural and virgin cellulose ﬁbers in ce-
mentitious matrix provide insight into general durability characteristics of lignocellulosic
ﬁber-cement composites. A brief review of the mechanisms of deterioration of natural
and virgin cellulose ﬁbers in cement-based matrices along with long-term performance of

composites is presented below.

2.2.1 Mechanisms of Deterioration

Virgin kraft cellulose ﬁbers are the key ﬁber type used in cement composites; they have
minimum lignin content and, noting the susceptibility of lignin to alkaline attack, have
been used in applications involving outside exposure. Different weathering conditions ac-
tually increase the ﬁexural strength and modulus of elasticity of the composite. However,
weathered kraft cellulose ﬁber reinforced cement composites are more brittle than the

original composites [32, 33, 34].

Much of the available test data on long-term durability of wood ﬁber-cement composites
deals with the use of natural ﬁbers (e. g., sisal) in cement-based materials. In this case, un-
less measures ate taken to reduce the alkalinity of cement matrix or to protect the natural
ﬁbers, the repeated action of wetting and drying results in the transport of alkaline pore
water to ﬁbers and the removal of neutralized pore water (which would be produced in the
vicinity of ﬁbers) as well as the decomposed products from these' ﬁbers, causing decompo-
sition of some natural ﬁbers like sisal [31]. This can lead to the embrittlement and loss of
ﬂexural strength in natural ﬁber reinforced cement composites. Repeated wetting and dry-

ing is a key factor accelerating this deterioration process of natural ﬁbers, and it depends

24

on temperature and humidity history in the vicinity of the cement product. Interior expo—
sure conditions or continuous immersion in water at ambient temperatures can not pro-
duce cycles of wetting and drying which are particularly harmful to the composite. Two
years of exposure in such conditions did not lead to any embrittlement of sisal reinforced

cement composites in test results reported by Gram (1983) [31].

A description of the mechanisms leading to the embrittlement of natural ﬁber reinforced
concrete under the action of repeated wetting and drying (rain-heat) is presented in Gram
(1983) [31]. According to this description, the alkaline pore water in concrete reacts with
the lignin and hemicellulose existing in the middle of larnellae causing decomposition of
these constituents of ﬁbers. This leads to the weakening of the link between individual 6-
ber cells in natural ﬁber (Figure 2.4). External changes in moisture and temperature which
can provide a supply of water to and removal of water from concrete pores generate the
moisture movement needed for alkaline pore water to reach and progressively decompose

the natural ﬁbers, leading to the embrittlement of the composite material.

Wood kraft ﬁbers, as mentioned earlier, contain negligible amounts of lignin and thus they
can withstand the alkaline pore water attack better than natural ﬁbers. The reduced tough-
ness accompanied with the strength gain in these composites with ageing may be associ-
ated with the petriﬁcation process (ﬁlling of the core of the ﬁber with hydration products)
and the consequent changes in ﬁber failure mode [35, 36, 37, 38, 39]. The ﬁlling of ﬁber
cores and possibly cell wall pores with hydration products is expected to result in an
increase in strength and stiffness. In addition to that, it seems that the petriﬁed ﬁbers are
more stable dimensionally and no separation and debonding between the ﬁbers and the
matrix could be observed. In short, this petriﬁcation process increases the stiffness,
strength and bond strengths of ﬁbers, but reduces their ductility; these conditions lead to

improved stiffness and strength and increased brittleness of the composite.

25

 

Figure 2.4 Schematic Sketch of the Decomposition of Natural (sisal) Fibers in the
Alkaline Pore Water of Concrete [31].

2.2.2 Durability Test Results

2.2.2.1 Natural Fibers

Gram (1983) [31] has reported test results on the durability characteristics of sisal and coir
ﬁber reinforced mortars, in which a typical matrix had l:2:0.5 binder: aggregate: water
proportions. ﬁber content ranged from 0.5-4% by volume. Tests were performed under
natural ageing and also accelerated ageing conditions. In order to reproduce weathering
effects simulating the alkaline pore water attack on natural ﬁbers (e.g. conditions involv-
ing repeated exposure to rain and sun shine), an accelerated wetting-drying test equipment
was developed (Figure 2.5). The panel specimens used in this “climate box” measure 10
mm (0.4 in) in thickness and are subjected to half-cycles of moistening and cooling by
spraying them with water, followed by half cycles of heating with the temperature reach-
ing 105 deg. C (221 deg. F) and maintained at this level for a sufﬁciently long period so
that the capillary pore system in the specimen dries out.

' //..-
. ‘14” 1.1%,
1%”? '
ll
/

v/ A ’ y
I
I

f l.. ..
. I / J: ,5.
.4391}! .

I”
l

 

Figure 2.5 CBI Climate Box [31].

As a result of the above wetting-drying cycle, the capillary pore system of specimens is
both ﬁlled with and emptied of water during the conditioning cycle that lasts about 6
hours. This means that the ﬁbers embedded in specimens come in contact with the alkaline
pore water of the concrete during the moistening phase, and that any decomposition prod-
ucts which are formed as a result of the reaction between the ﬁber components and the
pore water can be transported away from the ﬁber during the drying phase. Figure 2.6 (a)
presents the reduction in ﬁexural strength of specimens incorporating sisal ﬁbers due to
ageing under repeated wetting-drying cycles. Considerable decrease in ﬁexural strength is
observed as the number of cycles increases. This can be attributed to the attack on natural

ﬁbers by the alkaline pore water of concrete.

The alkalinity of pore water in the matrix can be reduced and the weathering resistance of
the composite can be improved by replacing part of the cement with silica fume. The pH
value for the pore water reduces from 13.2 for a matrix with ordinary Portland cement to
12.0 for a matrix in which 33% of Portland cement by weight is substituted with silica
fume. Figure 2.6 (b) shows the ﬁexural strength of specimens reinforced with sisal ﬁbers
after 0 and 120 cycles of wetting and drying as a function of the percentage of cement by
weight substituted with silica fume. Specimens in which up to 20% of Portland cement

 

 

27

was replaced with silica fume show comparatively insigniﬁcant drops in strength after 120
wetting drying cycles when compared with specimens without silica fume. A dramatic
improvement is obtained when 30-50% of the cement is replaced with silica fume. Hence,
an effective approach to reducing the alkaline attack and thus enhancing the durability of
natural ﬁber-cement composites is through reducing the alkalinity of concrete pore water
by the use of silica fume [39, 40, 41].

2.2. 2.2 Processed Fibers

Earlier durability test results with processed cellulose ﬁbers (paper pulp) suggest that cel-
lulose ﬁber-cement composites may be prone to degradation in certain exposure condi-
tions. Four potential ageing mechanisms were investigated by Sharman (1983) [32] and
Sharman, et al. (1986) [32] for autoclaved cellulose ﬁber reinforced cement sheets. The
ageing mechanisms considered were carbonation, microbiological attack, moisture stress-

ing of wood ﬁbers, and increase in ﬁber-to-matrix bond, acting independently or together.

Carbonation is an important ageing mechanism in asbestos cement, causing embrittle-
ment. In the case of cellulose ﬁbers, carbonation may also increase susceptibility to micro-
biological attack by reducing the alkalinity of the cement pore water and/or increasing the
bonding of cementitious matrices to cellulose ﬁbers. The rate and extent of carbonation
depends on the physical conditions of the sheet (e.g. porosity) and local climatic condi-
tions (relative humidity and temperature). Microbiological attack on cellulose ﬁbers in
cement products was studied by Mansur and Aziz (1982) [42], who found it to be unlikely
in the highly alkaline environment of cementitious matrices. Long-term drop in the pH of
cement products after weathering and carbonation effects, however, may have adverse

effects on resistance to biological attack.

The behavior of cellulose ﬁber reinforced cement composites has been investigated under
various accelerated ageing effects, including (1) hot water soak, (2) moisture cycling, (3)
carbonation, and (4) fungal cellar exposure. Comprehensive ageing test data have been
reported by Shannan et al. (1986) [32] for autoclaved cement sheets consisting of 8% kraft
pulp (P. Radiata), 46% cement and 46% silica sand, and also by other investigators. A

brief review is presented below.

 

28

3° Flexural Strength (N/MM"2)

 

26

20"

15"

101

J;

 

 

 

 

O 20 4O 60 80 100 120 140
Maine Overs:
(a) Pure Cement

o Flexural Strength (N/mmnzl

5
I ‘ + o m. +120 Grated

‘0

 

 

 

30"

/\

20"

 

 

 

o 1: Y T T Y r
o 10 20 so 40 so so
Silica Fume-Binder Ratio ('5)

(b) With Silica Fume

Figure2.6 (cont’d) Flexural Strength of Composites Reinforced with Sisal Fibers
After Wetting-Drying Cycles [31].

Hot Water Soak

Hot water soak tests have been used to accelerate the ageing process and the subsequent
strength loss of glass ﬁber reinforced cement composites. In the case of cellulose ﬁber

reinforced cement, this test has been used to investigate any ageing effects on ﬁbers or

29

their bond to cementitious mauices. Effects of 350 days of immersion in 50 deg. C (122
deg. F) water on various aspects of the engineering properties of cellulose ﬁber reinforced
cement composites are presented in Table 2.2. It should be noted that changes in material
properties in warm water soak occurred dominantly in the ﬁrst 20 days of immersion;
thereafter, the material pr0perties stayed practically constant. These test results indicate
that warm water immersion, as an accelerated weathering condition, causes an increase in
elastic modulus of wood ﬁber reinforced cement composites; other eﬂ'ects of accelerated

ageing under warm water do not seem to be signiﬁcant.

Harper (1982) [43] has observed similar effects of warm water immersion on wood ﬁber
(kraft pulp) cement composites, implying that degradation of chemically processed wood
ﬁbers under weathering effects is unlikely. This can be explained by the fact that lignin,
which is the wood ﬁber constituent most susceptible to alkali attack, is totally removed
during the kraft pulping process (noting that kraft pulp is the main wood ﬁber type used in
cement composites). The fact that mechanical properties of wood ﬁber reinforced cement
sheets remained much the same throughout exposure indicate that there is no major
change in the nature of ﬁber-to-cement bond. Failure in mechanical testing was predomi-

nantly by ﬁber pullout.

Table 2.2 Wood Fiber-Cement Composites - Hot Water (50 deg. C, 122 deg. F) Soak
Test (see Appendix B for notation [32]

 

 

 

 

 

 

 

 

 

 

 

 

Number of Moisture Modulus of Modulus of
D Mechanical Movement (%) E1asticity(GPa) Rupture (MPa)
ays
Irn (1 Test Method
meme MD CD MD co MD CD
0 Mean 0.302 0.283 7.92 3.88 18.6 103—
SD. 0.008 0.005 0.33 0.70 0.6 0.2
350 Mean 0.235 0.224 9.79 8.15 19.2 12.1
SD. 0.009 0.012 0.33 0.35 0.9 1.2

 

 

 

30

Accelerated Wetting and Drying

Repeated wetting-drying tests are widely used for accelerated weathering of particle-
boards. Shannan et. al (1986) [32] have reported test results on wood ﬁber reinforced
cement sheets subjected to 10, 20, 35 and 50 cycles of wetting and drying. The test results
showed only a small increase in modulus of elasticity up to 10 cycles with no further
changes after 10 cycles. In this work, possible embrittlement of the composite under

repeated wetting-drying effects was not investigated.

Recent studies on repeated wetting and drying are reported by Akers et al. [34]. In this
study, the effects of natural versus accelerated weathering conditions on the ageing behav-
ior of wood ﬁber reinforced cement composites were also investigated. TWO products
manufactured on a standard Hatschek machine were used in this study. They were: ( 1) 8%
wood ﬁbers in a Portland cement-based mauix and cured at ambient temperature and rela-
tive humidity; and (2) 8% wood ﬁbers in a Portland cement and silica based matrix and
autoclaved. The dimensions of the ﬂat sheet products were 600 x 400 x 6 mm (24 x 16 in x
0.23 in) for natural weathering and 200 x 100 x 6 mm (8 x 4 x 0.23 in) for accelerated age-
ing tests. The Portland cement-based and Portland cement silica-based products were
exposed to natural weathering for 5 and 4 years, respectively. Similar products were also

subjected to accelerated tests.

The accelerated test involved repeated cycles of wetting and drying, with each cycle con-
sisting of the following steps over a 24-hour time period: (a) 9 hours submersion under
water at 20 deg. C (68 deg. F); (b) 3 hours in air at 20 deg. C (68 deg. F); (c) 9 hour of
infrared radiation at 80 deg. C (176 deg. F) in air; and (d) 3 hours cooling down to 20 deg.
C (68 deg. F) in air. The natural and accelerated ageing effects on wood ﬁber-cement com-

posites are discussed in the following.

The test results on specimens exposed to natural weathering (Figure 2.7) showed a general
increase in strength and stiffness. This increase in strength is associated with an increase
in density from 1750 to 1870 kg/m3 (109 to 117 lb/ft3) over a period of 5 years for Port-
land-based (air cured) Specimens and from 1610 to 1790 kg/m3 (100 112 lb/ft3) over a
period of 4 years for Portland cement/silica-based (autoclaved) specimens. The products

were found to be well carbonated at the end of the natural weathering test period.

The degree of polymerization of wood ﬁbers in the naturally aged products was found to

 

31

decrease with age by about 20% for air cured products and 35% for autoclaved products.
This is in contrast to the tendency in strength to increase with age. Thus, the weakening
effects which might have been expected upon reduction in the degree of polymerization
are apparently more than compensated for by other processes, including the improvement

in ﬁber—to-matrix bonding and overall densiﬁcation of the composite under ageing effects.

The naturally aged specimens were found to be brittle. The embrittlement of these com-
posites may have been due to the increase in ﬁber-to-matrix bond, which results in ﬁber
rupture rather than pullout dominating failure at fracture surfaces; this ageing effect
reduces the frictional energy absorption associated with ﬁber pullout. The specimens sub-
jected to accelerated ageing (up to 3 months) in ambient environment (with minimal car-
bonation effects) showed drops in ﬁexural strength (16.4 to 12.3 MPa, 2.38 to 1.78 ksi).
This is in contrast to the increase in ﬁexural strength shown by naturally aged composites.
The increase in elastic modulus and the drop in degree of polymerization were, however,

comparable.

3° Flexural Strength (N/mrn“2)

 

 

 

A Air-02nd None:
0 Autoclaved Mortar

 

 

 

 

 

 

10L . . g
o r 2 a 4 5 0
Years of Exposure to Natural Weathering

(a) Flexural Strength

Figure 2.7 Flexural Strength and Young’s Modulus of Air Cured and Autoclaved
Products Exposed to Natural Weathering [34].

32

25 Elastic Modulus (1(N/rnrn"2)

 

 

 

A arr-cud item:
0 Autoclaved lea-tar

 

 

 

 

 

 

5 J. 1 L 1
o r 2 a 4 a a
Year: of Exposure to Natural Weathering

(b) Elastic Modulus

Figure 2.7 (Cont’d) Flexural Strength and Young’s Modulus of Air Cured and Auto-
claved Products Exposed to Natural Weathering [34].

Observation of the fractured surfaces of composites after accelerated ageing in ambient
environment indicated invariably broken ﬁbers with their hollow nature clearly shown
(see Figure 2.8) [35]. In most cases the matrix around the ﬁbers was quite dense (Figure

2.8). It was typical to observe debonding between wood ﬁbers and the surrounding matrix.

Carbonation

Carbonation of the matrix by atmospheric carbon dioxide has been observed to cause loss
of strength in asbestos cement composites. Since wood ﬁber cement sheets will also
undergo carbonation in everyday use, it is desirable to account for this process in any

accelerated ageing tests on wood ﬁber-cement composites.

 

 

 

Figure 2.8 Scanning Electron Micrographs of Fractured Surfaces After Accelerated

Weathering in Ambient Environment [35].

Sharman et al. (1986) [32] have conducted carbonation tests on autoclaved wood ﬁber
reinforced cement composites. The specimens were stored in a tank (RH 65% and temper-
ature 20 deg. C, 68 deg. F) which was supplied with carbon dioxide at a rate of 40 cm3/
min (2.5 in3/min). The extent of carbonation was assessed by the measurement of CaCO3
content, which reached levels ranging from 36 to 41%, when the absorption of C02 virtu-
ally ceased. The signiﬁcant changes noted in the properties of carbonated wood ﬁber rein-
forced cement sheets were increased tensile strength, intemal bond. and moisture
movement (Figure 2.9). An increase in modulus of rupture was also observed, and the
ﬁexural stiffness showed a similar increase. It is worth mentioning that the increase in
moisture movement seems to be the only potentially deleterious effect of carbonation

observed in this investigation.

34

 

 

 

 

 

 

 

 

m Tensile Strenrth 00-) u Internal Bond Strength (MPa)
. F 0.5 » v
v V
O 0.4 r-
e r- o o 0
o
0.3 1- V
v
4 > A
A A
A A oz ~
2r
0.1 r-
o 4 4 t o L ‘ ‘
O 10 10 30 60 o 10 go 80 .0
Calcium Carbonate (2 by Veil“) Calcium Carbonate (X by weight)
(a) Tensrle Strength (b) Internal Bond Strength

‘ Ioiature Iovernenl (%)

 

 

 

 

0.6 >
A Load perpendicular
as r H 8 E to principle fiber
I direction
0.: ~ 8 0 Load parallel to
principle fiber
oz .. direction
V Load perpendicular
0,] L to plane of sheet
0 A
0 10 20 30 «1

Calcium Carbonate (Z by weight)

(c) Moisture Movement
Figure 2.9 Effects of Carbonation on Wood Fiber Reinforced Cement Sheets [32].

Akers et al. [34] in more recent studies on carbonation effects, subjected wood ﬁber-

35

cement composites to accelerated weathering conditions which also involved exposure to
a carbon dioxide rich environment. Two composites (air cured cement-based and auto-
claved Portland cement silica-based matrices) were subjected to the following accelerated

weathering cycles.

(a) 8 hours submersion under water at 20 deg. C (68 deg. F)

(b) 1 hour in oven at 80 deg. C (176 deg. F)

(c) 5 hours in oven at 20 deg. C (68 deg. F) in a saturated C02 environment
((1) 9 hour in oven at 80 deg. C (176 deg. F)

(e) 1 hour cooling down from 80 deg. C to 20 deg. C (176 to 68 deg. F)

The test cycles chosen were Optimized by trial and error experiments based on the degree
of carbonation and water penetration into the products. Table 2.3 shows the test results on
the ageing of composites subjected to natural weathering and accelerated ageing tests. The
results suggest that the development of mechanical properties of wood ﬁber reinforced
cement composites when exposed to C02 rich accelerated test simulates more closely the
behavior in natural weathering. Accelerated ageing in a C02 rich environment and natural
weathering both led to an increase in strength and elastic modulus. Also, the increase in
the degree of carbonation in a C02 environment compares favorably with the naturally
weathered products. The increase in density of the product with age may be associated

with carbonation of the matrix.

With respect to wood ﬁber prOperties, there is a breakdown in the molecular chain of the
ﬁbers with age, which may be directly correlated with the decrease in degree of polymer-
ization (see Table 2.3). The drop in degree of polymerization should logically result in a
drop in tensile strength of the wood ﬁbers; however, other factors such as carbonation,
which tend to increase the strength and elastic modulus of the composite overshadow any

negative effects of the damage to wood ﬁbers.

36

Table 2.3 Properties of Wood Fiber-Cement Products Exposed to Different Ageing

 

 

 

 

 

Conditions [34]

Type of Ageing Flexural Elastic Modulus Density
Strength (N/ kN/mm2 Kg/m3
mm )

Non Aged 16.4 i 0.9 10.9 $1.3 1770 :1: 20

Accelerated Aged 3 12.3 i 1.1 14.8 5; 1.2 1780 :1: 10

Months Ambient Environ-

ments

Accelerated Aged 3 23.9 3]; 2.2 18.9 1; 1.4 1800 1; 10

Months C02 Rich Environ-

ments

Natural Weathering 25.1 $1.6 18 3; 0.9 1830 :l: 40

 

 

 

 

 

 

The mode of fracture after ageing was brittle with most of the ﬁbers being broken at the
fractured plane (Figure 2.10 ). Very frequently, the circular cross section of the broken
ﬁber was ﬁlled up with dense hydration products (Figure 2.10 ) and there was no perimeter

debonding. This microstructure is referred to as “brittle petriﬁed.”

It is suggested that the increase in strength and rigidity of the petriﬁed ﬁbers, and the
increase in their bond strength (due to matrix densiﬁcation and elimination of shrinkage-
debonding from the surrounding matrix) can account for the increase in strength and elas-

tic modulus of the composite upon ageing.

Naturally aged composites produced microstructural features at failure surfaces similar to
those of composites subjected to accelerated ageing in a carbon dioxide rich environment.
This suggests that petriﬁcation takes place more readily under carbonating conditions,

probably due to the lower pH and greater solubility of the hydration products.

37

 

Figure 2.10 Brittle Fracture in a Composite After Accelerated Ageing in a C02 Rich

Environment [34].

2.3 THEORETICAL CONSIDERATIONS

Basics of Fracture Mechanics

When the tensile strength of a brittle material is reached in a structure, cracking will occur.
The study of the conditions around and in front of a crack tip is called fracture mechanics
[44]. Fracture mechanics was ﬁrst studied for brittle materials such as glass [45]. The ﬁrst
applications to concrete appear to have been made by Neville [46] and by Kaplan [47]. A
historical review and an annotated bibliography of the applications of fracture mechanics

to cement and concrete is given by Mindess [48].

The application of fracture mechanics to concrete structures has provided new ways of un-
derstanding and modeling phenomena which could only be treated empirically before.

There is a growing international interest in these questions and this is reﬂected in recent

38

published literature. Some works covering the main parts of the development have been
presented by Mttrnan [49], Shah [50], Carpinteri and Ingraffea [51], Sih and DiTommaso
[52], Reinhardt [53], and Ewalds [54].

Fracture mechanics refers to the analysis of the fracture of materials by the rapid growth
of pre-existing ﬂaws or cracks. Such rapid (or even catastr0phic) crack growth may occur
when a system requires sufﬁcient stored energy that, during crack extension, the system
releases more energy than it absorbs. Fracture of this type (often referred to as fast frac-

ture) can be predicted in terms of energy criterion [55.56.57].

If we consider an elastic system containing a crack and subjected to external loads, the to-

tal energy in the system, U, is

U: ('WL + Up) + Us .............. (Equation 2.1)

where

-WL=work due to the applied loads

UE: strain energy stored in the sytem

Us: surface energy absorbed for the creation of new crack surfaces

A crack will propagate when dU/dc < 0, where dc is the increase in the crack length. Using
this theory, one can derive the Grifﬁth equation, which gives the theoretical fracture

strength for brittle, linearly elastic materials:
ween/mm ..................... (Equation 2.2)
where

Of = Stress at ﬁrst crack strain

c = one half of crack length

75: surface energy of the material

This is the basic equation of linear elastic fracture mechanics (LEFM).

If we deﬁne a parameter Ge: 275 = critical strain energy release rate, then we may write

39

the criterion for catastrophic crack growth as

of Ute)"2 = (EGQU2 ....................... (Equation 2.3)

That is, fracture will occur when, in a stressed material, the crack reaches a critical size (or
when in a material containing a crack of some given size, the stress reaches a critical val-
ue).

1/2
)

Alternatively, we may deﬁne a parameter Kc: O'(rtc = critical stress intensity factor.

K3 = EGC ...................................... (Equation 2.4)

Kc has the units of Wm”, and is often referred to as the fracture toughness (not to be con-
fused with the term “toughness”, which is used to refer to the area under the load-deﬂec-

tion or sness-strain curve).

The LEFM parameters, Gc and Kc, are one-parameter descriptions of the stress and dis-
placement ﬁelds in the vicinity of a crack tip. In much of the early work on the applica-
tions Of fracture mechanics to cement and concrete, it was assumed that they provided an
adequate failure criterion. However, later research showed that even for these relatively
brittle materials, LEFM could only be applied to extremely large sections (e.g., mass con-
crete structures, such as large dams). For more ordinary cross-sectional dimensions, non-
linear fracture mechanics parameters provide a much better description of the fracture

process.

Fibers enhance the strength and, more particularly, the toughness of brittle matrices by
providing a crack arrest mechanism. Therefore, fracture mechanics concepts have also
been applied to model ﬁber reinforced cement composites. Mindess [48] has reviewed the
difﬁculties in modelling cement composites based on the fracture mechanics approach.
LEFM might be adequate to predict the effects of the ﬁbers on ﬁrst cracking. However, to
account for the post-cracking behavior (which is responsible for the enhanced toughness
of ﬁber-cement composites), it is essential to resort to elastic-plastic or nonlinear fracture
mechanics. A measure of toughness (i.e., the energy absorbed during fracture) can be Ob-
tained from the area under the stress-suain curve in tension. The fracture mechanics con-
cepts which could provide a more precise measure of toughness of ﬁber reinforced cement
composites include the crack mouth opening displacement (CMOD), R-curve analysis, the
ﬁctitious crack model (FCM), and various other treatments, all of which model (either im-

plicitly or explicitly) a zone of discontinuous cracking, or process zone, ahead of the ad-

4o

vancing crack. These approaches provide fracture parameters which are, at least,
dependent on the ﬁber content, whereas the LEFM parameters (Gc or Kc) are most often
insensitive to ﬁber content. It might be added here that, while the J integral has often been
used to describe these systems, theoretically it cannot be applied to composite systems
such as ﬁber reinforced concrete, where there is substantial stress relaxation in microc-

racked region in the vicinity of the crack tip.

In the investigation of the ﬁber-crack interactions using fracture mechanics concepts, three

distinct issues must be considered:

(1) Crack Suppression: This is the increase in stress, due to the presence of the ﬁbers, re-

quired for crack initiation (i.e., the increase in the ﬁrst-crack stress).

(2) Crack Stabilization: This refers to the arrest of the cracks already generated, which

have begun to propagate across the ﬁbers.

(3) Fiber-matrix Debonding: This process can be modelled as crack propagation along the
ﬁber-matrix interface.

Crack Suppression

Romualdi and Batson [57] were the ﬁrst to use LEFM concepts to analyze the mechanics
of crack suppression in a cement matrix induced by the presence of ﬁbers. In the absence
of any cracks, the extensions of both the matrix and the ﬁber under tensile loading are the
same. However, when cracks are present, the matrix tends to extend more than the ﬁbers,
because of the stress concenuations just ahead of the crack tip. The ﬁbers oppose this ten-
dency. Through interfacial shear bond stresses, they apply ‘pinching forces’ which effec-
tively reduce the stress intensity factor at the crack tip. As a result, higher applied stresses
are now required to produce a stress ﬁeld ahead of the crack tip such that the maximum
stress exceeds the critical stress intensity factor, Kc, of the cement matrix. The shear bond
stress distribution which causes this pinching effect is shown in Figure 2.11; this permits a
determination of the contribution of neighboring ﬁbers to the stress intensity factor. This
work led to the introduction of the concept of the spacing factor, S; the stress required to

cause matrix cracking was found to be inversely proportional to S.

Rt
do

(110th

41

 

e-—

N
O
x, —-—, ”Qt—- “‘

 

vecrons ngpnesur/

BOND STRESS
DISTRIBUTION

 

 

Figure 2.11 The ‘Pinching Effect’ and Interfacial Shear Stress Distribution Predicted
by Romualdi and Batson in the Arrest of Crack Propagation by Fibers [5‘7].

Romualdi and Mandel [58] calculated the effective spacing factor in a 3-dimensional ran-
dom short ﬁber reinforced concrete:

5 = 2.76r(1/v,)”2 .................. (Equation 2.5)

They used this relationship to demonstrate the validity of the spacing factor concept for
predicting the improvement in ﬁrst crack stress due to the presence of ﬁbers, as shown in
Figure 2.12. A number of other numerical expressions for the spacing factor have since
been developed, generally by considering the distances between ﬁbers crossing a given
plane in the composite [58]. Though the various spacing equations are similar in form,

they do not properly account for the chief geometrical factors which deﬁne a ﬁber, i.e.,
length and diameter.

There are a number of limitations to the application of spacing factor equations for pre-
dicting ﬁrst crack strains. For instance, some experimental results considered by Edginton

et al. [60] showed considerable deviation from the theoretical predictions of Romualdi et

42

a1 [59]. It appears that in calculations of S, one must account not only for length and diam-
eter, but also for the effects of ﬁber orientation and the nature of the ﬁber-matrix bond.
Bond has been generally assumed to be ‘perfect’ [59, 61], which is not realistic for short
ﬁber-cement composites. Swamy et al [61] have suggested the concept of ‘effective’ spac-
ing, which takes into account modiﬁcations due to both geometrical and bond

considerations.

An alternative approach to the calculation of the ﬁrst-crack stress was proposed by
Aveston et al [62], based on energy balance considerations. For a crack to form under con-

ditions of a ﬁxed tensile stress, the energy changes to be considered are:
(1) Work, AW, done by the applied stress, since the body length changes.
(2) Work, 'dev done in debonding the ﬁber from the matrix. This term can be calculated

through fracture mechanics analysis of ﬁber-mauix debonding assuming that the debond—
ing energy at the ﬁber-matrix interface, Gdbv is less than or equal to the surface energy, 7“,,

of the matrix.
(3) Work, Us, done by frictional slip after debonding.
(4) The reduction in elastic strain energy of the ﬁbers after matrix cracking, AUm.

(5) The increase in elastic strain energy of the ﬁbers after matrix cracking, AUf, due to

load transfer from the cracked mauix to the ﬁber.

Hence, if the work expended in creating a new crack (i.e., the surface energy of the crack

area) is m, then a crack will only form if [55]:
2'!me + You + Us + AUfS AW + Aunt ............ Equation 2.6

On this basis, the ﬁrst cracking strain of the matrix is:

an... = (12 tn, in, E; v?) /(Ec Em2 1’ vm) ....................... Equation 2.7

Equation 2.7 predicts an effective increase in the matrix cracking strain for a high ﬁber
volume content, a high interfacial frictional shear strength, and a small ﬁber diameter. In
essence, this is another way of predicting the degree of crack suppression (i.e., the in-
crease in ﬁrst—crack stress), in accordance with the spacing factor concept. A signiﬁcant
increase in the matrix cracking strain should occur with well-bonded, small-diameter ﬁla-
ments. This has been observed with asbestos and glass ﬁbers, which generally consist of

ﬁlaments (often in bundles) less than ~15 pm in diameter.

 

 

 

 

 

 

'- \
O r." t
_ ~g \\ ..
g \ o won-ctr 'CNSION
I ‘\ 0 one no snow:
I 2 0 ~. x -
T: \THEORETICAL
\a \
2 #— ° 3 ‘ °
.. \ E
I EXPERIMENTAL N
h. \
m 3 F “- «r
~¥ V
3 I 2

WIRE Spacwcmm

Figure 2.12 Effect of Fiber Spacing on First Crack Stress Ratio [61].

Crack Stabilization

Once ﬁrst cracking has taken place in the brittle matrix, ﬁbers serve to inhibit unstable
crack propagation. At this stage. the cracking patterns are complex, with discontinuous
microcracks present ahead of the principal crack. This can be deduced from various ana-
lytical models, and has also been Observed microsc0pically (by Bentur and Diamond
[63]in Figure 2.13. Thus, in cracked composite, it is difﬁcult to deﬁne the ‘true’ crack tip.
The simplistic deﬁnition of a traction-free crack (as assumed in LEFM) is not applicable
to FRC. Stress is transferred across the crack by a variety of mechanisms, as can be seen
from the idealization of a crack proposed by Wecharantana and Shah [64] in Figure 2.14.
Three distinct zones can be identiﬁed [55]:

(1) Traction Free Zone:
(2) Fiber bridging zone, in which stress is transferred by frictional slip of ﬁbers; and

(3) Matrix process zone, containing microcracks, but with enough continuity and aggre-

gate interlock to transfer some stress in the matrix itself.

 

Figure 2.13 Complex Crack Patterns at the Interaction of an Advancing Crack and a
Fiber Lying in its Path [63].

LOAD F

   
  

TRACTION FREE CRACK LENGTH

GGRE (SAT E

A
r_.BRIDGING

LOAD p he»: seems LENGTH

CRACK CLOSING FRESSURE
FIBRE BRIDGING

MATRIX PROCESS
/_ZONE

 

Figure 2.14 Idealized Representation of an Advancing Crack and the Stress Field

Around it, in a Fiber Reinforced Cement [64].

45

Many of the analytical treatments of these effects involve either the assumption of speciﬁc
stress ﬁeld around the apparent crack tip, or the consideration of a traction free crack sur-

face which is subjected to a closing pressure [65], as shown in Figure 2.15.

......

 

 

Figure 2.15 Schematic Description of a Traction Free Crack with a Closing Pressure,
to Model the Fracture Behavior of Fiber Reinforced Cement [60].

There are, in addition, a number of models which take different approaches to fracture me-
chanics modeling of cement composites. Bazant and his co-workers have developed a
smeared crack model, in which fracture is modelled as a blunt smeared crack band [66].
The fracture properties are characterized by three parameters: fracture energy, uniaxial
strength, and width of the crack band. This approach lends itself particularly well to com-
puter-based ﬁnite element modelling of cracks. For very large structures, this theory be-
comes equivalent to the LEFM approach. For smaller structures, however, the theory
predicts a lower critical strain energy release rate, because the fracture process zone can
not develop fully. This theory has been shown to provide a good ﬁt to experimental data

from the literature.

Hillerborg [67] has developed the ﬁctitous crack model to describe the fracture of both
plain and ﬁber reinforced concretes. In this model, the deformation of a specimen is given

in terms of two diagrams:
(l) The stress-strain (6 versus a) diagram (including the unloading branch);

(2) The stress deformation (0 versus w) diagram for the fracture zone itself.

Our
deb

0:“

Mler

45

The 6 versus w curve gives, the additional deformation in a test specimen due to the pres-
ence of a damage zone. Moreover, the area under the o-w curve (called the fracture ener-
gy) equals the energy absorbed per unit (projected) area during the fracture process due to

the additional deformation of the process zone.

However, none of the non-linear fracture mechanics approaches described above appears
to provide a truly fundamental fracture parameter, independent of specimen geometry and
loading conditions. It is also difﬁcult to interpret the physical signiﬁcance of the stress dis-
tributions ahead of the crack tip which are assumed (or implied) in various models. It is,
therefore, difﬁcult to disagree with Majamundar and Walton [68], who conclude that ‘for
composites which display the phenomenon of multiple cracking and rising stress-strain re-
lationships it would appear that fracture mechanics approaches currently under consider-
ation will have little prospect of success in producing parameters that will be useful in
design' [55].

Fiber-Matrix Debonding

A fracture mechanics approach has also been applied to the problems of ﬁber debonding
and pull-out, as an alternative to the treatment based on the analysis of elastic and friction-
al shear stresses. The object is to develop material parameters to account for debonding
which are more reliable and easier to evaluate experimentally than the interfacial shear
bond strength values. In this treatment, it is assumed that the debonded region is traction
free (If =0 in Figure 2.16), and this zone is treated as an interfacial crack of length b. Us-
ing the classical Grifﬁth theory (of LEFM), the conditions leading to the propagation of

this crack, and to spontaneous debonding, can be calculated.

Outwater and Murphy [69] calculated the ﬁber tensile stress (0) required for catastrophic

debonding as

o=[(8 Ef odb)1’2]/d ............... Equation 2.8

where Gdb is the energy required to debond a unit surface area of ﬁber, and Bf and d are ﬁ-

ber elastic modulus and diameter, respectively.

 

47

AXIS or /,
SYMMETRY I
1'”? —sI.p
U . h
J__-__va1

J', ZIN-
‘ 4 g | t d

I V ‘.
.o I.~[ DEBONDED’
. T,

 

I
I _L_.’%qp___T, 3
’ I

Y. 'I
”9 Vt; .BONDED I

 

 

Figure 2.16 Schematic Description of the Model Used to Consider the Pullout Prob-
lem in Terms of Fracture Mechanics Concepts, With a Propagating Debonding
Crack of Length b [65].

The above analysis considers only the energy balance in the ﬁber itself. Subsequently, a
solution was developed which takes into account the compliance of the entire pull-out sys-
tem. Morrison et al [70] further extended the analysis by taking into account also the fric-
tional resistance in the debonded zone, which is the more realistic case for ﬁber reinforced
cement composites. The critical strain energy release rates for debonding calculated by
Morrison et a1 [70], that is 2.5 N/m, is similar to those calculated by Mandel et al, which is
44.7 N/m. these values are, however, lower than typical values of the critical strain ener-
gy release rate of plain mortars, which are typically 5-12 N/m. Thus, for a crack to follow
the path of least resistance, it should propagate along the interface rather than through the

matrix.

This fracture mechanics approach tends to conﬁrm the conclusions reached previously.
That is, the interface in ﬁber-cement composites is relatively weak. This leads to preferen-

tial crack propagation along the ﬁber-matrix interface.

01

01'
VC
tic

[III

48

Theories Applied to Cellulose Fiber Reinforced Cement

Composite materials approach (rule of mixture) has been used by several researchers to
predict the tensile, ﬁexural strength and other characteristics of cellulose ﬁber reinforced
cement composites [71]. In this approach, strength is assumed to be the sum of the effects
of the matrix and the ﬁbers; the ﬁber contribution is governed by pull-out (assuming that
ﬁber-pull out dominates the behavior) and is a function of tl/d, while the matrix contribu-
tion is a function of the strength of a void free mauix, omo, multiplied by its solid content,
(l-Vo). Therefore,

For tensile strength:

ocu=<3mo(l-V0)Vm + ZnIVﬂ/d .............. (Equation 2.9)

For ﬂexural strength:

Ob=Ot/B (ob)mo (1-V0)Vm +2natVﬂ/d ........... (Equation 2.10)
Similarly, the modus of elasticity in tension, Et, and bending Eb, is:

r~:b,E,=E,,,o(1-vo)vnn + nEfvf. ....................... (Equation 2.11)

To use Equations 2.9 through 2.11 for predicting strength and E modulus of cellulose ﬁber
cement composites, it is necessary to determine the constants TI, Ot, [3, the interfacial bond

strength (I), the ﬁber aspect ratio (W) and the void free matrix properties, Emo and cum.

The efﬁciency factor (n) can be taken as 0.41 after Romualdi and Mandel[58]. The values
of or and [3, which are the ratios of bending strength to tensile strength of the composite
and matrix, respectively, have been found to be 2.96 and 2.81, respectively. The Properties
of the void free matrix have been determined in separate tests of the matrix only, with a
void content of 23%, from which tensile strength, bending strength and modulus of elas-
ticity for Vo=0 were calculated. Substituting these parameters in Equation (2.9-2.11) per-

mits the prediction of the mechanical properties as a function of ﬁber content.

As shown in Figure 2.17, these predictions agree satisfactorily with experimental results.
While these theories predict a continuous increase in tensile strength and bending strength
with increasing ﬁber mass fraction, the experimental results show no strength improve-
ment beyond 6% ﬁber mass fraction. This is probably a consequence of the relatively large

void fractions at larger ﬁber mass fraction, which cause further reductions in the interfa-

cial I
for II

b. Van'a

Fig“It 2.
lllloSe Flt]

49

cial bond strength below 0.35 MPa (50 psi). The large void fractions are also responsible

f or the low moduli in bending and in tension for ﬁber mass fractions greater than 0.08.

 

“that no enact, 5.1-«(nun a, o. 90"
_. 1..

 

 

 

 

“SS ”MCI“! I ‘3'53‘. '

a. Variation of Tensile and Bend Strengths With Mass Fraction of Fibers

I, .-_
au—xsx'z: mow aoIEI

KR: xv. I
20 w I
LCMER 51m

0 y
l e

 

 

 

o'

A
'0.

he
Cr

5,—5RE-(759 ‘93‘ 00151
0"“? 300w

I

|

I

4‘ T I .3»? 5::0 !
‘1

I

I
J

2

 

 

 

vutmos MOULMJS t, then.»
- 8

.1...
lo]

2 - 5 . 3 '0
”SS FRACTION CF FSRESJII. 1’43

1). Variation of Young’s Modulus in Bending and in Tension With Mass Fraction of
Fibers

Figure 2.17 Experimental Vs. Theoretical Predictions of Different Properties of Cel-
llllose Fiber Cement Composites [71].

Das I

her 0

and II

the C1

distn'l
matrix
of the
reasor
beyon
trix stI
dict th
conuit
high ﬁl
that, w]
the frat
failed b

50

Das Gupta et a1 [72] applied a somewhat different composite material model to natural ﬁ-
ber cement composites. They included in their calculation the effects of ﬁber orientation
and length, and developed different relations for ﬁbers with lengths greater or shorter than
the critical length. Their model was used to analyze pastes reinforced with short randomly
disuibuted coir ﬁbers. The average ﬁber diameter was 0.119 mm (0.004 in) and the ﬁber
mauix bond determined by pull-out tests was found to be 1.5 MPa (220 psi). Comparison
of their experimental and analytical results is provided in Figure 2.18. The agreement is
reasonably good, up to a ﬁber content of 5%, but the model does not predict the decline
beyond this ﬁber volume, which is probably associated with a reduction in bond and ma-
trix strength due to poor compaction. Andonian et a1 [71] have reported on efforts to pre—
dict these effects by considering the influence of an increase in void content on the
contribution of the mauix. Their data does not show a marked reduction in properties at
high ﬁber contents, probably due to a different method of preparation. It should be noted
that, while Andonian et al [71] assumed ﬁber pull out, dominating the behavior analysis of
the fractured surfaces in a later work suggested that a large portion of ﬁbers may have

failed by ﬁber fracture [73].

 

l cucuurto
o unamtarat

0'
O
I
l

20*-

IENSILE STRENGTH , MPO
N
u
I

 

 

 

 

I 2 3 4 5 6
FIBRE CONTENT , ’l. VOI

Figure 2.18 Relations Between Experimental and Calculated Tensile Strength in Coir
Fiber Reinforced Cement[[72].

Frz
ber
mo
ues
inse
ites.

OUS,

Ire 2.

110m 1

{01 the

51

Fracture mechanics concepts have also been applied to predict the behavior of cellulose ﬁ-
ber reinforced cement composites. Fracture parameters were found to be dependent on the
moisture content, with wet cellulose ﬁber cement composites having higher toughness val-
ues (Mai et a1 [73]; and Mindess et al [74]). Wet composites were also found to be notch
insensitive, suggesting that LEFM cannot adequately model the behavior of such compos-
ites. This was probably the result of the cracking mode, in which the crack path was tortu-

ous, with some ﬁbers failing by pull-out.

Hughes and Hannant [75] examined the effects of moisture on the ﬁrst crack stress, by
considering the reduction in the modulus of elasticity of the wet ﬁber, and using the con-
cepts of crack arrest when the ﬁber spacing is less than the critical ﬂaw size. Their results,
assuming a ﬁber modulus of 4 and 40 GPa (580 and 5800 ksi) for wet and dry conditions,
respectively, are presented in Table 2. 4. If the bond is assumed to remain constant once
wet composites are dried, at or about 0.5 MPa (70 psi), (as suggested by Andonian et a1
[75]), the matrix failure strain should increase upon drying. If the bond increases forrn0.5
MPa (70 psi) to an assumed value of 2.0 MPa, the matrix failure strain would be expected
to double. The increase in bond on drying may be a result of hydrogen bonding, which is
more readily generated in dry state. The transition of the fracture mode form ﬁber fracture
in the dry state to ﬁber pull out in the wet state may be associated with the effects calculat-
ed in Table 2.4, in which a large stress can be developed in the ﬁber prior to matrix crack-

ing, leading to the rupture of ﬁbers failure matrix fails.

Mia et a1 [73] have considered slow stable crack growth (based on LEFM) as a prominent
feature of the fracture behavior of cellulose ﬁber cement. Double cantilever beam speci-
men with side grooves were used to Obtain resistance curves. Fiber reinforced composites
COntaining 8% mass fraction of bleached ﬁbers were tested in wet and dry conditions. Fig-
ure 2.19 shows typical load-displacement curves during slow crack growth in a dry and a
Wet sample of cellulose ﬁber cement. The load at which the load-deﬂection curve deviates
from linearity is taken as the onset of crack growth and is used to calculate G and K values
fer the construction of slow crack growth resistance curve. It is apparent from Figure 2.19

t1lat, upon unloading to zero load after crack extension, there is a permanent deformation

(5r),

Table 1
Modul

Across

l—

Moist
State

I

5
V
22

 

IEIE’I

* Assur.

Figure 2.

1101] in C‘

52

Table 2.4 Effect of Moisture Content and the Resulting Assumed Change in the Fiber
Modulus of Elasticity on the Calculated Matrix Failure Strain and the Fiber Stress
Across a Stable Flaw [73]

 

 

 

 

Moist. Fiber Modulus Bond Strength Matrix Failure Max. Fiber Stress

State (GPa) "' (MPa) * Strain (%) *"‘ Across a Stable
Flaw (MPa) **

Wet 4 0.5 0.074 27

Dry 40 0.5 0.106 96

Dry 40 2.0 0.156 172

 

 

 

 

 

 

 

* Assumed ** Calculated; The unreinforced matrix strain is assumed to be 0.05%.

 

 

 

 

 

 

 

 

 

(a) Dry (b) Wet

Figure 2.19 Typical Load (P) Against Displacement (8) Records For Crack PrOpaga-
tion in Cellulose Fiber Cements[73].

Slow t

KRVM

are alsc
amea
the mat
other in
given In
in the FI
cal expr
Consider
Stills gin
Electing
HOWeve]
Calculati,
higher it

be in 1hr

53

Slow crack resistance curves for the dry and wet composites are shown in the Figure 2.20.

KR values predicted from the analytical expression

KR2 = [(12 P2a2)/(13BH3)] [1+1.32 (H/a) +).532 (H/a)2] ........... (Equation 2.12)

are also shown in Figure 2.20. For the dry samples we must have KR (Equation 2.12) the
same as those from KR: (EGR)l/2 since they both assume linear elasticity to be obeyed by
the material. Also as Sr is small, KR" =(KGR"')1’2 should be approximately equal to the
other two KR calculations. These results are borne out by the similar KR against Aa results
given in Figure 2.20, even though different equations are used. For the wet samples given
in the Figure 2.20, all these KR * curves are different. The KR curve based on the analyti-
cal expression by Equation 2.12 gives the worst results and the one based on GR without
considering residual displacement in KR=(EGR)m, still underestimates the true KR" re-
sults given by the upper most curve. It was concluded in this study that for dry samples ne-
glecting residual displacement does not signiﬁcantly alter the KR and GR curves.
However, for wet composites the residual displacements are large and must be included in
calculations of the true KR* and GR“ values. Crack growth resistance curve is much
higher for wet than for dry composites due to the different deformation behavior of the ﬁ-

ber in the two states (Figure 2.20).

“0%;

'5. ‘I a IEG;IW-KR‘
" a (56,)” -I(.
O KI (Equation 1)

 

be (an)

(a) Dry Samples

 

 

a 112 a
,L 0 I56.) «a
O K. - (Eouottonli
A IEGII‘R-KI

 

 

50 100 150
Animal

(b) Wet Samples

Figure 2.20 Crack Growth Resistance (K) Plotted Against Crack Extension (Aa) [73].

 

CHAPTER 3

DETERNIINATION OF INFLUENTIAL VARIABLES
IN THE PROCESSING OF RECYCLED CELLULOSE
FIBER-CEMENT

3.1 INTRODUCTION

Cellulose ﬁber reinforced cement composites are quite different in mix proportioning and
processing from normal concrete or mortar. The process includes beating and reﬁnement
of ﬁbers in a slurry, and mixing of all constituents in water. The slurry has a low solid con-
tent in order to uniformly disperse the ﬁbers; vacuum is therefore applied to extract the ex-
cess water. The composite is ﬁnally compacted under pressure, and curing is usually
reached under high-pressure steam effects for accelerated strength gain in prefabrication
facilities.

In a detailed break down of this production process, eleven proportioning / processing
variables were distinguished as potentially inﬂuential in determining the end product
(wastepaper ﬁber—cement composite) qualities. The inﬂuential variables detected at this
stage of research are to be Optimized in the next stage of the project.

A comprehensive program was designed in order to identify those variables with statisti-
cally signiﬁcant effects on the ﬂexural performance Of cellulose ﬁber reinforced cement
composites. A 1/64 fractional factorial design of experiments was used with eleven vari-
ables. This design consisted of a total of 32 diﬂ'erent combinations of the proportioning I
processing variables, each considered at two levels. The control composite with 100%
virgin cellulose ﬁbers was also considered in this phase of study. Three ﬂexural tests were
conducted on specimens prepared from each mix, and the average values were used in sta-

tistical analyses.

55

 

56

3.2 VARIABLES AND EXPERIMENTAL DESIGN

Recycled Fiber-cement composites are manufactured by: (1) Reﬁnement (beating) Of re-
cycled wastepaper ﬁbers to expose microﬁbrils on the ﬁber surfaces, and make the ﬁbers
discrete and compatible with cement. Beating is done using a slurry containing 5% by
weight of ﬁbers. A laboratory scale pulp disintegrator (TMI reﬁner) was used at a speed of
3000 rpm. The beating time ranged from 5 to 10 minutes; (2) Proportioning the ﬁbers (re-
cycled & virgin), sand, ﬂocculating agent and cement; (3) Mixing the ingredients in water
to produce a slurry of 20% solids; a high-speed mixer is used to achieve a uniform disper-
sion of recycled cellulose ﬁbers and other mix ingredients in this slurry. Flocculating agent
is the last solid constituent to be added, which improves the binding of cement particles to
cellulose ﬁbers and controls the escape of cement particles during vacuum dewatering.
The mixing time needed to uniformly disperse mix ingredients is approximately 10 min-
utes; (4) Vacuum dewatering of the slurry to extract the excess water; the extraction of wa-
ter is actually performed in two stages. First, the excess water on top of the settled slurry
is removed, and then the settled slurry is put onto a vacuum box 305 mm by 305 mm (12
in by 12 in) in planer dimensions. The slurry is evenly spread onto the screen of vacuum
box in layers of about 2.5 mm (0.1 in) thick, and the vacuum varied from 127 mm to 254
mm (5 in to 10 in) of mercury. It takes a build up of four to ﬁve layers during the vacuum
application process in order to make the panel (about 10 mm (0.4 in) thick); (5) Compac-
tion under pressure to produce a dense composite with even surface. A 110 KN (12 ton)
hydraulic press was used for the compaction of 305 mm x 305 mm x 10 mm (12 in x 12 in
x 0.4 in) panel. Compaction pressure of 0.7 to 1.4 MPa (100 to 200 psi) was used.; and (6)

Curing of the composite.

Different components Of this process are presented in Figure 3.1 for a laboratory scale pro-

duction facility.

A total of 11 key variables (factors) deﬁning this production process were selected to be
investigated in this experimental program. These variables are: (l) recycled ﬁber source;
(2) ﬁber mass fraction; (3) ﬁber beating level; (4) substitution level of virgin ﬁbers with re-
cycled ones; (5) sand/binder ratio; (6) maximum particle size of sand; (7) silica fume!
binder ratio; (8) ﬂocculating agent/binder ratio; (9) vacuum level; (10) compaction pres-
sure; and (11) curing condition. Each factor was considered at two levels in a (1/64) frac-

tional factorial design of experiments (Table 3.1).

57

 

(b) Mixing

Figure 3.1 Components of the Laboratory Scale Manufacturing Process.

58

 

(c) Vacuum Dewatering

” 7",“ir‘trwvrv' - .
'. -r ’ '

    

(d) Compaction

Figure 3.1 (Cont’d) Components of the Laboratory Scale Manufacturing Process.

59

 

(e) High-Pressure Steam Curing

Figure 3.1 (Cont’d) Components of the Laboratory Scale Manufacturing Process.

3.2.1 Recycled Fiber Source: The recycled cellulose ﬁbers used in this investigation are
made through dry mechanical processing of wastepaper. Magazine papers were selected
for use in this investigation because of their high level of non-cellulosic materials (e.g.,
clay, latex, etc.); these impurities cause difﬁculties in the conventional wet method of de-
riving cellulose ﬁbers from waste magazine paper. The wastepaper source and speciﬁcs of
the dry recycling process were thought to be possible sources of variations in wastepaper
ﬁber-cement composite qualities. A number of sources of dry processed wastepaper ﬁbers
were surveyed and ﬁnally two sources were selected which used magazine papers and dif~
ferent dry processing techniques. A detailed presentation of the properties Of ﬁbers ob-
tained from these sources is presented later in this chapter. Interﬁbe recycled ﬁbers are
manufactured through pulverization process by different sizes equipment whereas Ameri-
can Fillers ﬁbers are manufactured by hammer mill grinding process.

3.2.2 Fiber Mass Fraction. This is a key variable of the investigation. Fiber mass fraction
is deﬁned in this investigation as the mass ratio of ﬁbers to the total solid constituents of

the composite. Two ﬁber mass fractions, 5 and 8%, were used to study the effects on the

composite material qualities.

3.2.3 Fiber Beating Level. The beating of ﬁbers in a water slurry exposes microﬁbrils on
ﬁber surfaces. This helps improve the ﬁltering action of ﬁbers (to avoid cement loss dur-
ing dewatering), and also enhances the mechanical bonding of ﬁbers to cement-based ma-
uices. Fiber beating level is measured in Canadian Standard Freeness (CSF), which
decreases as ﬁber beating increases (Tappi 1207). The two levels of CSF presented in Ta-
ble 3.1 refer to unbeaten (as received) and beaten ﬁbers (beating was continued for 10
minutes in this case). Unbeaten ﬁbers from source “a” and “b” had CSF values of 620 and
580, respectively. After beating the corresponding CSF values were 520 and 480, respec-
tively.

3.2.4 Fiber Substitution Level.The next variable in Table 3.1 is ﬁber substitution level,
which has two levels Of 50% and 100%; at the 50% level, 50% of ﬁbers (by weight) were
virgin cellulose ﬁbers and the remaining 50% were recycled ﬁbers. At the 100% substitu-

tion level, the whole ﬁber content was recycled magazine ﬁbers.

3.2.5 Sand size. Gradation of the silica sands used is illustrated in Table 3.2. Maximum
particle sizes of the two silica sands were 50 |.l.m and 600 um (zero45 in to 24*1045 in).

3.2.6 Sand Content. Silica sand was used in this study at sand/binder ratios (by weight) of
0.25 and 0.75.

3.2.7. Silica Fume. Cement substitution with silica fume at 0 and 10% was also consid-
ered (see Table 3.3 for the physical and chemical properties of the Type I Portland cement
and Silica fume used in this investigation). Silica fume is a by-product from the reduction
of high-purity quartz with coal in electric arc furnaces in the production of silicon and sil-
icon alloys. The ﬁneness and pozzolanic reactivity of silica fume make it highly eﬂ‘ective
in enhancing the density and chemical stability of the bulk of cement paste and particular-
ly the interface zones. The consumption of calcium hydroxide (a relatively unstable ce-
ment hydration product), the reduction in the alkalinity of cement pore water, and
reducing permeability of the matrix are some key mechanisms through which silica fume

could positivelyimprove the long-term stability of cellulose ﬁber-cement composites.

3.2.8 Flocculating Agent. Flocculating agent was used to improve the binding of cement
particles to cellulose ﬁbers and to control escape of cement particles during the vacuum

dewatering process. The dosage rates of the ﬂocculent considered in this experimental

61

study were 0.2% and 0.6% of binder (by weight). NALCO Chemical Company’s product
“NALCLEAR 9798 Pulv Flocculent” was used. It is an acrylamide acrylate polymer. Its
speciﬁc gravity is 0.75, and its pH value at 0.2% is 8.2.

3.29 Vacuum Level. The vacuum level was also studied as a variable; vacuum was ap-
plied at 5 in and 10in (127 mm and 254 mm) of mercury to the composite to extract the

excess water from the slurry.

3.2.10. Compaction Pressure. The compaction pressure was another variable considered;

the thin cement sheets were pressed under 0.7 Mpa and 1.4 Mpa (100 psi and 200 psi) of

pressure.

3.2.11 Curing Condition. Two curing conditions were considered: moist curing (“a”) and
high pressure steam curing (“b”). In moist curing, the composites were stored in a moist
room for 7 days followed by 21 days of air drying in the laboratory. In high-pressure cur-
ing, an autoclave was used to cure the specimens at a pressure of 0.86 MPa (125 psi) for 8
hours, with specimens subsequently stored in the laboratory environment up to the test age
of 28 days.

High-pressure steam curing (autoclaving) has been developed for several reasons [76].
Firstly, it increases the rate of hydration reactions; curing in steam at 0.86 MPa (125 psi)
for 8 hours is roughly equivalent to 28 days of moist curing at normal temperature. This
permits a high rate of tum-over in the production of precast products. Secondly, products
can be made which are superior to those of normally cured concrete; chemical resistance
and dimensional stability can be improved. Lastly, there is a possibility of replacing the
cement partly or wholly by waste materials, which are unreactive at ordinary temperatures
but which possess cementing properties at higher temperatures. The adoption of steam
curing (high-pressure or atmospheric pressure) in commercial practice is determined pri-
marily by economic considerations. The development of strength in a high-pressure steam
cured product depends at least in part, on the reaction between cement and the ﬁne silica.
The initial setting reactions, and phases and subsequent behavior of the product in the case
of autoclaved cement-silica are more stable than those occurring at ordinary temperatures.
The primary reaction at 85-200o C is probably always the hydration of cement to give to-
bermorite gel and Calcium hydroxide [76]. Addition of small amounts of ground silica

contributes to suppress the formation of calcium hydroxide. So autoclaving has a number

 

62

of positive effects; namely reduction of calcium hydroxide, and formation of better crys-

tallized calcium silicate hydrate [76].
The 1/64 fractional factorial design of experiments in this phase of the study (Table 3.1)

was completely randomized. All main effects were clear of 2-factor interactions, but due
to the fractional nature of design, 2-factor interactions were compounded with one anoth-
er; in other words, this experimental design reveals the effects of all variables on the com-
posite material performance, but can not provide any information on possible interactions
between diﬂ'erent variables. The resulting specimens were tested for ﬂexural performance
(strength, toughness, and initial stiffness). The ﬂexural test set-up (Japanese Standard JCI

SF4 [77] and ASTM C1186 [78]) is presented in Figure 3.3.
Table 3.1: Fractional Factorial Design of Experiments

 

 

 

 

. Run
Experiment om“ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Variables Units
FrbetSourco m a a a b b a a a b a a a b b b b

 

Fiber Mm percent 5 5 5 8 5 8 5 8 5 5 8 8 8 8 8 5
Fractioa

 

Fiber Beating CSF b a a b b a a b b a b a b a a b
Level

FiberSubstitu- parcel! 100 50 100 100 50 50 100 50 50 50 100 100 50 100 100 100
tioalevel

Sand/Binder weight .25 .25 .25 .75 .75 .75 .75 .75 .25 .75 .75 .25 .75 .25 .75 .75

 

 

 

SandMaai-um 600506006005060050505050506006005050600

mum Size

 

Silica Fume! weight 0 0 .1 .1 .1 .1 .l 0 .1 0 .1 0 0 0 0 0
Binder ratio

FlocAgent/ weight .006 .002 .006 .006 .006 .002 .002 .002 .002 .006 .006 .002 .002 .002 .006 .002
binder ratio

 

 

Vacuum Level inchof S 5 10 10 5 10 S 5 10 10 10 10 5 10 5 10

 

 

mercury
Coupactioa psi 100 200 200 200 100 200 100 100 100 100 100 200 100 100 100 100
Pressure
CitingCondi- name a b a a b a b a a a b a b b a b

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

63

Table 3.1 (Cont’d): Fractional Factorial Design Of Experiments

 

 

 

 

 

 

 

. Run cont
Experiment Order 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 m1
Variables Units
Fiber Source name a a b b b b b b b a a b b b a a trek

 

Fiber Man power: 8 8 5 5 8 8 5 5 8 5 8 5 8 5 5 8 8
Fraction

 

Fiber Beating CSF b a a a b a b a b b b b a a b a a
Level

Fiber Substitu- percera 100 50 50 100 100 50 100 100 50 50 50 100 50 50 50 100 0
tion Level
Sand/binder weight .25 .25 .25 .75 .25 .25 .25 .25 .25 .25 .25 .75 .75 .75 .75 .75 .75

 

 

 

SandMaxi- tun 5060050600600505060060060050505060060060050
mumSize

SiiiaFulne/ weight .1 .1 O .1 .1 .1 0 .1 0 .1 0 O .1 0 .1 0 .1
Binder ratio
Floc.Ageatl weight .002 .002 .002 .002 .002 .006 .006 .006 .006 .002 .006 .002 .002 .006 .006 .006 .006
binder ratio

Vaarurnlaevel incbof 5 5 5 5 5 5 5 10 10 10 10 10 10 10 5 5 10

 

 

 

 

 

rmrcury
Compaction pn‘ 100 100 200 100 200 200 200 100 100 200 200 200 200 200 200 200 200
Pressure
ClaingCondi- name a b a a b a b b a b b a b b a b a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

64

Table 3.2: Sand Gradation (ASTM 0371: Percent Retained on Individual Sieves)

 

 

 

 

 

 

 

 

 

Sieve No. Silica Sand Ground Silica
(ASTM E11) Sand l
‘ao===TT—=
50 0.5
70 22.9
100 47.1
140 21.6
200 7.3 0.1
270 0.4 1.9
325 - 2.4
- 325 - 96.6

 

 

 

 

 

Table 3.3 Properties of the Binders

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Properties Binder Type

Cement Silica Fume
CaO 63.24 -
Si02 21.14 96.50
A1203 5.76 0.15
Fez03 2.93 0.15
803 2.46 -
MgO 2.06 0.20
K20 0.79 0.04
C - 1.40
NaZO - 0.20
Surface Area 0.16 m2/gm 20.25 m2/gm
Speciﬁc Gravity 3.15 2.3
Fineness (% retained in 10.7 -
#325 sieve)

 

 

 

65

Vacuum Chamber
with Specimen j

 

Specunen

Filter Inlet line

Screen

 

 

 

 

Outlet

[‘—

 

 

 

 

 

Vacuum Line /

 

 

 

 

 

 

 

 

 

Vacuum Pump

 

 

 

 

Moisture Trap 1

 

 

Moisture Trap 2

Figure 3.2 Manufacturing Process- Slurry Dewatering.

3.3 RECYLED WASTEPAPER FIBERS

Recycled ﬁbers are derived from recycling of wastepaper (magazine) by dry mechanical
processing. American Filler’s recycled ﬁbers (ﬁber “b" in Table 3.1) had 100% magazine
paper whereas the recycled ﬁbers from Interﬁbe (ﬁber “a” in Table 3.1) was made of
wastepaper with 90% magazine paper constituent. Both used dry mechanical processing to
disintegrate paper into ﬁbers with increased surface area. These ﬁbers are mainly cellulo-
sic ﬁbers and, being dry processed have a fraction of ﬁne fragments and ﬁne coating mate-
rials from the wastepaper (mainly glossy coated magazine paper). Cellulose ﬁbers are
derived from wood and in the process of paper making these ﬁbers undergo some modiﬁ-
cations. Magazine papers are mostly made from chemical pulp whereas newsprint is made
mainly from ground pulp. The ﬁbers used in magazine are a blend of softwood and hard-
wood to produce desirable prOperties at minimum cost. Besides ﬁbers, magazine paper
also constitutes 20% of its composition as coating materials (clay, processing chemicals,
Pigment particles, and latex) which impart desirable characteristic to magazine paper.
The recycled ﬁbers were analyzed through Scanning Electron Microscope observations,

measurement of length disuibution, and thermo-gravimeuic compositional analysis.

3.3.1 Fiber Length Distribution
Standard equipment used for cellulose ﬁber length distribution (Kajani F8100 machine)

was used to carry out this analysis. This equipment microscopically measures the lengths
of all ﬁbers in a sample and gives ﬁber length ranges in weight (see Table 3.4). The results
presented in Table 3.4 conﬁrm that the recycled ﬁbers are shorter than the virgin softwood
kraft pulp (Southern Softwood Kraft, SSK). The virgin kraft pulp (SSK) used in this inves-
tigation presents a high-quality ﬁber with desirable (3 mm) length and high reinforcement
efﬁciency for cement applications. The high ﬁne content of recycled ﬁbers may result
from the damage to ﬁbers during recycling. The presence of shorter hardwood ﬁbers in re-
cycled wastepaper, and the presence of impurities in wastepaper are also responsible for
the high ﬁne content of recycled ﬁbers. Shorter recycled ﬁbers are expected to have lower
reinforcement efﬁciencies; the ﬁnes in recycled ﬁbers would act more like ﬁllers than rein-
forcing ﬁbers. The virgin cellulose ﬁbers had an average length of 3.3 mm (0.13 in)
whereas the recycled ﬁbers from Interﬁbe (ﬁber “a" in Table 3.1) had an average length of
1.2 mm (0.05 in) and that for American Filler (ﬁber “b”) had an average length of 1.3 mm
(0.051 in). The average values and 95% conﬁdence intervals of the ﬁber lengths are pre-
sented in Frgure 3.4.

Table 3.4 Recycled Fiber Length Distribution (Weight%)

 

 

 

 

 

 

 

 

 

 

 

l Length Virgin Recycled Recycled
(mm) Cellulose Source 1 Source 2
<0.2 4.16 12.47 1.96
0.2-1.0 16.72 44.43 49.86
1.0-2.0 20.63 27.96 29.63
2.0-3.0 16.56 12.12 15.28
3.0-4.0 12.38 2.62 2.66
4.0-5.0 10.90 0.38 0.51
5.0-6.0 7.55 - -
6.0-7.0 4.11 - -
:70 6.95 - -

 

 

 

 

 

 

67

Length (mm)

 

 

 

 

 

 

 

 

 

 

 

 

. t- ' f. l ,f
Vir. Cell. (SSK) Recy. Cell. (AF) Recy. Cell. (I)
Fiber Type

 

Figure 3.3 Fiber Length Distribution: Average Values and 95% Conﬁdence Intervals.

3.3.2 Morphology

The morphology of recycled and virgin ﬁbers was studied under a Scanning Electron Mi-
croscope (SEM). The shape of the recycled ﬁbers was observed to be quite different from
that of virgin cellulose ﬁbers. Virgin ﬁbers are hollow and cylindrical in shape with a vary-
ing diameter. Recycled ﬁbers appear to be ﬂatter and twisted. The surface of the recycled
cellulose ﬁbers, as compared to the virgin counterpart, is relatively rough, with mi-
croﬁbrils exposed on the surfaces. Fine clay particles and ﬁber fragments can also be seen

in the micrographs (Figure 3.5)

 

(b) Recycled Fibers (American Fillers, “b”)

Figure 3.4 Scanning Electron Micrographs of Virgin and Recycled Cellulose Fibers.

ISKU X2'008

 

(c) Virgin Cellulose Fibers (SSK)

Figure 3.4 (Cont’d) Scanning Electmn Micrographs of Virgin and Recycled Cellulose
Fibers

3.3.3 Compositional Analysis

In order to investigate the composition of virgin and recycled ﬁbers, Thermo Gravimetric
Analysis (T GA) was carried out. The aim was to investigate and estimate the contents of
pure cellulose in these ﬁbers. The rate of heating used was 20° C (68° F) per minute and
the weight loss increasing with temperature changes was recorded. Cellulose decomposes
at 350° C (660° F). In all cases a sharp weight loss is observed around this temperature.
The trends in weight loss are shown in Figure 3.6. From these weight loss curves one may
estimate cellulose content in recycled ﬁbers around 67%; one may also approximate lignin
content of and 13% (lignin decomposes at 450° C, 840° F). The ash is the ﬁne clay which
was used in surface texture of the magazine paper and also the latex and other materials

used in the formation of paper. Vrrgin cellulose ﬁbers had about 90% of cellulose content;

70

their lignin content was small (probably because magazine paper uses chemical pulps with

low lignin content), and the ash content was also small, as expected, due to the low levels

of impurities in virgin ﬁbers

'0‘” ($1

 

1W

 

00*

 

 

40t-

 

 

20r-

 

 

 

0 too ‘°° 90° 030 1000
Temperature (deg. C)

(a) Recycled Fibers (Interﬁbe, “3”)

W! (S,

 

 

 

 

 

° '°° ‘°° '00 000 1000
Temperature (deg. C)

(b) Recycled Fibers (American Fillers, “b”)
Figure 3.5 Fiber Compositional Analysis by Thermogravimetry.

71

Weight (96)

 

100

 

 

 

 

 

O 1 l l l
O 200 400 600 800 1000

Temperature (deg. C)

(c) Virgin Cellulose Fibers

Figure 3.6 (Cont’d) Fiber Compositional Analysis by Thermogravimetry.

3.3.4 Canadian Standard Freeness and Fiber Reﬁnement. Freeness test is an empirical
process that gives an arbitrary measure of the rate at which a suspension of 3 gm (0.048
oz.) of pulp in one liter of water may be drained [79]. The result depends mainly upon the
quantity of debris present, the degree of ﬁbrillation of ﬁbers, their ﬂexibility and ﬁneness.
The higher the value the Canadian Standard Freeness (CSF) the lesser would be the reﬁne-
ment and vice versa. Virgin cellulose without reﬁnement had a CSF value of 700, whereas
the recycled ﬁbers without reﬁnement had a CSF value of 620 in the case of Interﬁbe (“a”)
and 580 in the case of the American Filler (b”) ﬁbers. After reﬁnement these values
dropped to 600, 520, and 480 respectively.

72

3.4 EXPERIMENTAL SET UP

The ﬂexural test procedures recommended by the Japanese Concrete Institute (JCI-SF)
[77] were followed for testing the thin-sheet samples. The ﬂexural test samples had a clear
span of 9 in (228 mm), width of 4.5 in (115 mm) and thickness of 0.25 in (6 mm). The test
set up is shown in Fig. 3.6. A displacement-controlled ﬂexural test was conducted at a dis-
placement rate of 1/3000 of span length per minute. A computer-controlled data acquisi-
tion system was used to record the test data and plot the ﬂexural load-deﬂection curves.
These load-deﬂection curves can be characterized by ﬂexural strength, ﬂexural toughness
(determined by the Japanese Concrete Institute [77] as the area under the load-deﬂection
curve up to a ﬁexural deflection equal to span length divided by 150), and initial ﬂexural
stiffness (deﬁned here as the stiffness obtained through linear regression analysis of the

load-deflection points for loads below 20% of maximum flexural load.

 

Figure 3.6. Flexural Test Setup.

73

3.5 TEST RESULTS AND STATISTICAL ANALYSIS

The test results obtained for the fractional factorial experimental design of Table 3.1 are
shown in the Figures 3.7 -3.9. Typical ﬂexural load-deﬂection curves are presented in Fig-
ure 3.10. The trends in the effects of different variables, as indicated by the fractional fac-
torial analysis of variance of test results, are shown in Figures 3.11 -3.21. One may
visually estimate the relative signiﬁcance of different variables through comparing the

sloes of the trend lines shown in these ﬁgures.

Table 3.5 Flexural Performance of Recycled Wastepaper Fiber-Cement Composites

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Experiment Flex. Str. Mean Flex. Tou. Mean Init Stif. Mean Init
(By run (MPa) Flex. Str N-mrn Tou. N/mrn Stif.
(95% (95% (95%

01' def ) Con. Int.) Con. Int.) Con. Int.)

1 7.372 7.429 13.234 13.710 211.0 214.5
7.865 (1.609) 13.897 (0.677) 220.7 (20.98)
7.05 13.675 211.9

2 10.930 10.936 34.890 32.865 198.2 206.8
11.32 (1.489) 30.143 (9.334) 204.6 (38.82)
10.56 34.560 217.7

3 8.5 8.571 45.112 45.253 69.44 70.40
8.889 (1.133) 45.456 (0.705) 71.54 (4.158)
8.324 48.345 70.23

4 8.198 7.999 2.900 2.686 290.5 276.6
7.678 (1.101) 2.334 (1.203) 278.7 (59.22)
8.123 2.789 260.5

5 6.876 6.888 65.123 63.498 245.6 256.6
6.559 (1.317) 63.234 (5.918) 270.5 (49.97)
7.231 62.675 253.7

6 8.543 8.732 53.023 54.281 171.0 172.2
8.667 (0.898) 56.345 (7.061) 178.2 (21.74)
8.987 52.760 167.3

 

 

74

 

 

 

 

 

 

 

 

 

 

 

 

 

7 6.785 7.075 32.876 33.755 119.9 116.2
6.987 (1.345) 34.498 (3.213) 123.5 (37.61)
7.454 31.992 105.3

8 14.012 14.556 16.765 17.455 313.8 303.9
14.567 (2.111) 18.654 (4.085) 301.8 (35.15)
15.089 15.167 296.2

9 7.871 7.806 78.491 81.465 202.1 201.3
7.897 (0.532) 83.432 (10.27) 211.3 (40.85)
7.650 81.880 190.5

10 4.773 4.927 77.453 77.878 87.54 82.89
4.987 (0.529) 77.098 (4.154) 77.22 (20.53)
5.023 77.087 83.92

11 6.123 6.058 10.543 10.174 59.53 63.89
6.030 (0.219) 9.81 1 (1.434) 68.43 ( 17.46)
6.022 10.290 63.70

12 5.456 5.666 8.765 8.471 155 158.7
5.555 (1.106) 8.176 (1.154) 162.3 (20.37)
5.987 8.432 169.2

13 7.024 7.180 38.777 39.509 108.7 98.92
7.453 (0.928) 40.209 (2.808) 96.66 (34.67)
7.065 39.430 91.43

14 5.667 5.789 79.337 81.468 107.8 103.1
5.713 (0.678) 85.987 (15.34) 90.54 (43.13)
5.986 79.235 111.0

15 4.603 4.815 13.875 13.694 169.9 178.3
4.967 (0.742) 13.650 (0.64) 180.3 (29.75)
4.815 15.776 184.7

16 7.345 6.974 42.342 44.215 41.47 41.94
7.123 (1.813) 45.112 (6.36) 45.55 (13.31)
6.456 45.666 38.81

17 5.495 5.653 5.678 5.853 31.22 32.92
5.567 (0.84) 6.234 (1.292) 34.2 (6.126)
5.890 5.556 33.23

18 4.786 4.783 37.231 38.402 140.2 150.20
4.667 (0.45) 38.998 (3.976) 160.0 (38.75)
4.897 39.223 150.4

 

 

 

 

 

 

 

 

75

 

 

 

 

 

 

 

 

 

 

 

 

 

19 4.320 4.180 26.456 27.704 141.1 141.8
4.234 (0.677) 29.543 (6.373) 151.2 (35.84)
3.987 28.345 133.0

20 4.123 4.409 24.345 23.568 270.2 253.2
4.674 (1.08) 23.765 (3.492) 245.7 (57.84)
4.430 22.333 243.8

21 5.732 5.324 21.560 22.232 220.1 231.10
4.564 (2.58) 22.543 (2.286) 230.4 (44.73)
5.678 23.767 242.9

22 7.213 7.282 75.977 75.843 142.3 140.9
7.512 (0.798) 76.988 (4.768) 130.2 (39.44)
7.123 71.440 150.2

23 6.267 6.396 13.300 12.565 74.54 78.81
6.378 (0.544) 10.432 (4.322) 80.23 (14.77)
6.543 13.101 81.66

24 5.987 6.179 32.987 33.252 190.2 194.4
6.320 (0.675) 32.876 (2.188) 194.7 (15.96)
6.230 35.032 198.3

25 7.467 7.419 53.890 55.943 450.2 460.6
7.231 (0.644) 57.452 (7.223) 471.2 (41.17)
7.560 58.543 460.3

26 8.234 8.625 11.222 11.499 183.3 192.5
8.987 (1.479) 11.977 (1.629) 190.5 (40.47)
8.654 10.897 203.7

27 5.754 5.58 5.654 5.623 130.5 133.6
5.456 (0.606) 5.567 (0.19) 147.5 (49.57)
5.532 5.719 122.8

28 10.378 10.356 10.888 11.278 251.3 262.7
10.689 (1.348) 11.651 (1.496) 275.6 (47.83)
10.002 11.312 , 261.2

29 6.567 6.629 12.765 13.102 282.4 291.9
6.345 (1.254) 13.543 (1.568) 297.8 (32.53)
6.976 13.444 295.4

30 5.543 5.385 69.665 67.191 341.2 354.5
5.125 (0.889) 64.120 (11.05) 353.0 (55.53)
5.487 67.430 369.4

 

 

 

 

 

 

 

 

.';'I(‘I)1)I’l (MI ’0)
as \ an a '7: 7‘ ‘7

A

Flexural Strength (MPO)

omed\m

dd
ON

76

 

 

 

 

31 8.345 8.330 14.638 14.667 110.5 118.6
8.123 (0.785) 14.678 (0.108) 128.4 (35.65)
8.523 17.334 116.7

32 4.234 4.307 16.876 16.404 160.5 161.3
4.675 (1.322) 15.955 (1.806) 151.9 (38.26)
4.012 16.290 171.4

33 (Control) 14.564 14.792 81.987 84.402 388.1 407.1
14.024 (3.545) 80.654 (21.08) 423 (69.18)
15.789 90.567 410.3

 

 

 

 

 

 

 

 

Figure 3.7. Flexural Strength Test Results.

hulnllmtllllhtlllt

.——u——:—~—-—~—

Experiment (By run order)

no
N

Control (33)

11 (t It. I! 111‘ ll‘ I.‘
e l E ll.‘ 4 ‘1‘ l‘ 1 5 4 ‘. 3

3 2 2 4| ”It
nizt 2v 2:0»...726— 2:3..2. ~2:t\2» «ualtttm ‘33}:

77

A3 3.28 as 3.28

 

 

mm mm.
3” Z”
on on
mm mm
mm mm
R R
mm mm
mm mm
x a
mm mm
a o a
E S
cm m cm
2 n 9
e m. e
t ( . t
S m m S
2 m m 2
z .m R z
e a M 2
S m—
: m :
9 .9 o—
m w m
w T w
R m a
o m o
m H m
e 8 v
m d m
N

TnLnnnnn.._ My Tonnon.1..._

mum mmmmmo mmmmmmmmmwo

Eta £233 ...53: €35 82.? .3:

Experiment (By run order)

Figure 3.9 Initial Stiffness Test Results.

78

 

 

 

 

 

 

 

 

 

Flenurer Stress (Mpa) Ftenurel Stress (Mpa)
0 2O
—- "be! men t (6‘) — FIDO! Beetlno l'e')
- 9100! use I (as) — - Floor 5...,” ('0')
15 15 ’-
'\
, '\
./ ‘ ~
13 r . ‘ .
\ /\
\ .' 4‘
\\ 5 b I \ \
\\
\
x A J o 1 1
2 3 O 1
Depeuon {mm} Deﬂection (mm)

 

 

 

 

 

 

Flexural Stress (MP3)
20
— Virgin (mi 8%)
""" virzrec. 50%( mi 8%)
15 r — Recycled(mf 8%)
10 r
5 )—
O I i
O l 2 8

Deflection (mm)

Figure 3.10 Typical Flexural Load-Deﬂection Curves.

79

 

 

Hand “to. M ‘00 Initial Stittneea (Warm)

 

 

 

 

 

 

 

 

 

 

100 ~
0 1 ‘ o 1 1
1 a - 1 2
Fiber Source Fiber Source
Flexural Strength (MPa)
15
1o - ME]
6 i-
o 1 L
1 2 '

Fiber Source

Figure 3.11 The ﬁends in Fiber Source Effects.

80

 

 

 

 

 

 

 

 

Fiend m (It-weal “)0 initial atmneae (ti/mm)
w .-
300 '7
200 P \
‘0 P
100 '-
o A 1 o 1 1
C I - I I
Fiber Maee Fraction (1.) Fiber Maae Fraction (1:)

Flexural Strength (MPa)

 

15

10"

 

 

 

0 1 l
5 8 -

Fiber Mass Fraction (96)

Figure 3.12 The Trends in Fiber Mass Fraction Effects.

81

 

 

 

 

 

 

 

 

Flannel Toudtneea (N-mrnl minitlal Stitfneaa (ll/mm)

” b
300 "

40 “ /
100 ~

0 1 ‘ o 1 1
e.e lb. - ... .b.
Fiber Beating Level Fiber Beating Level

5 Flexural Strength (MPa)

 

10*

 

 

 

o l 1
I.’ .b. _

Fiber Beating Level

Figure 3.13 The Trends in Fiber Reﬁnement Effects.

82

Normal Tm (ii-run) ‘00 initial Stiffneae (ll/mm)

 

 

/

 

 

 

 

 

 

200 e
w b \
100 >-
o A 1 o A l
60 DO - 60 00
Fiber Subetitution Level ('5) Fiber Subatltutlon Level (%)

5 Flexural Strength (MPa)

 

10*

 

 

 

o 1 1
50 100 -

Fiber Substitution Level (%)

Figure 3.14 The ﬁends in Fiber Substitution Level Effects.

83

 

 

 

 

 

 

 

 

Flamed m (IO-eel) 400 initial Btittneaa (ll/um)

” )-
80° p
200 /

‘0 /
too -

o 1 A o l l
.0 ”0 e 00 .00
Sand Maxinun Size (mic) Sand Maxim-n Sin (mic)

Flexural Strength (MPa)

 

1o... \

 

 

 

o 1 ‘ 1
50 600 -

Sand Maximum Size (mic.)

Figure 3.15 The Trends in Sand Maximum Size Effects.

 

 

 

 

 

 

 

 

Flamed m (Db-a) eoo initid Slim-ea (Warm)

” h
800 r
200 i- /

40 ~ \
100 '-

o J L o 1 l
025 0.76 - 0.26 0.75
Sandi Binder Ratio Band I Binder Ratio

Flexural Strength (MPa)

 

10" /

 

 

 

o 1 1
0.25 0.75 -

Sand/ Binder Ratio

Figure 3.16 The Trends in Sand/Binder Ratio Effects.

85

 

rum W (De-rem) minltiai Stillmaa (Ween)

 

20° "' /

 

 

 

 

 

 

‘0 , /
100 '-
o 1 I o L l
0 0.1 - 0 0.1
SF I Binder Ratio SF I Binder Ratio

Flexural Strength (MPa)

 

15

 

 

 

o 1 1
O 0.1 -

SF / Binder Ratio

Figure 3.17 The Trends in Silica Fume/Binder Ratio Effects.

86

MW“ minltlel 81th

 

 

 

 

 

 

 

 

‘0 i-
100 ’-
0 1 A o L 1
0.002 0.” - 0.002 0.”
FA I 81060! Ratio FA I Binder Bath

Flexural Strength (MPa)

 

 

 

 

 

15
10 ~ 4‘ —i—
5 e
o 1 1
0.002 0.006 '

FA I Binder Ratio

Figure 3.18 The Trends in Flocculating Agent/Binder Ratio Effects.

87

Mi m 00-.) mlnltlal “the. (ill/mi)

 

 

 

 

 

 

 

 

so » /
100 ~
0 A l o L A
e o - e 10
Vacuun Level (in of mercuy) Vacuun Level (in of merctty)

Flexural Strength (MPa)

 

 

 

 

o 1 1
5 10 -

Vacuum Level (in of mercury)

Figure 3.19 The Trends in Vacuum Level Effects.

88

 

 

 

 

 

 

 

 

Md m (ti—euro ‘00 initial Otiiinaaa (ll/en)

” F
300 b
200 P ./

40 e-””’ﬁﬁ*
100 >

o 1 l o l 1
too :00 - too 200
Convection Preeeure (pal) Compaction Preeeure (pol)

Flexural Strength (MPa)

 

10’- /

 

 

 

o 1 1
100 200 '-

Compaction Pressure (psi)

Figure 3.20 The Trends in Compaction Pressure Effects.

89

 

'M m ate-i 400m“ um... «I.»

 

300’-

 

 

 

 

/

i 4

 

nor-1 Autoclave
Curing Condition

 

 

 

 

200 >-
‘0 \.
100 »
o . 1 o
Helet Autoclave -
Owing Condition
Flexural Strength (MPa)
10 ~ *//*
5 .-
0 1 1
Molet Autoclave -

Curing Condition

Figure 3.21 The Trends in Curing Condition Effects.

 

90

Among the proportioning / processing variables investigated in the fractional factorial de-
sign of experiments (ﬁber source, ﬁber mass fraction, ﬁber beating level, ﬁber substitution
level, sand maximum size, sand/binder ratio, silica fume/binder ratio, ﬁocculating agent]
binder ratio, vacuum level, compaction pressure and curing conditions), ﬁber mass frac-
tion, ﬁber substitution level and ﬁber reﬁnement were found, through fractional factorial
analysis of variance, to have statistically signiﬁcant effects on the ﬁexural performance of
composite at 95% level of conﬁdence (the power of analysis was 0.91). A comprehensive
presentation of statistical analysis outcomes for ﬁexural strength, ﬁexural toughness and
ﬁexural stiffness is given in Table 3.6. ASTM C1186 does not specify precision limits for
this type of composite; however coefﬁcient of variation was calculated to be 3%, 7% and
9% for ﬁexural strength, toughness and stiffness (as obtained form the analysis of variance
of variables), which is with in acceptable limits as reported in the literature on cellulose ﬁ-

ber-cement composites.

Table 3.6 Results of the Analysis of Variance (Flexural Strength, Toughnees and Ini-
tial Stiffness)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hexural Strength

Sauce Sum-of oSquam DF Mean-Square F-Ratio P
Fiber Source 150442 702 1 150442.702 2.937 0.093
Fiber Man 1'. 366238.845 1 366238.485 7.149 0.010
Fiber Beating 1. 249820. 824 1 249820.824 4.877 0.032
Fiber Sub 1 1292251 .069 1 1292251069 25.225 0.000
Sand/Binder 362939 1 362.939 0.007 0.933
Sand Max. 8. 571550.788 1 57650.788 1.125 0.294
Silica Fl Bin. 144313.158 1 144313.158 2.817 0.099
Floc. AgJBin 3073 r .074 1 30731.074 0,600 0,442
Vacuum L 125095.997 1 125095.997 2,442 0.124
Compaction P 72959154 1 72959.154 1.424 0.238
Citing Con. 17725.932 1 17725.932 0.346 0559
Error 2663920399 52 51229248

 

 

 

 

 

 

 

91

Table 3.6 (Cont’d.) Results of the Analysis of Variance (Flexural Strength, Toughness
and Initial Stiffness

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Toughness
Soiree Sum—of Squares DF Mean-Square F-Ratio P
Fiber Sauce 44.020 1 44.020 0.978 0.327
Fiber Mass F. 45.166 1 45.166 1.004 0.321
Fiber Beating 1. 33.483 1 33.483 0.744 0.392
Fiber Sub. 1 47.193 1 47.193 1.049 0.311
Sandeinder 50.929 1 50.929 1.132 0.292
Sand Max. S. 39.176 1 39.176 0.870 0.355
Silica Fl Bin. 49.905 1 49.905 1.109 0.297
Floc. A;./Bin 54.118 1 54.118 1.202 0.278
Vaaium L 46.974 1 46.974 1.044 0.312
Companion P. 49.669 1 49.669 1.104 0.298
0:11.; Can. 40520 1 40520 0.900 0347
Error 2340.381 52 451117

Initial Stiffness
Sauce Sum-d -Squaree DF Mean-Square F-Ratio 1’
Fiber Source 40.591 1 405911 1.866 0.22
Fiber Mass P. 0.893 1 0.893 0.041 0.840
Fiber Beating 1. 19.840 1 19.840 0.912 0344
Fiber Sub. 1 159.915 1 159.915 7.353 0.009
Sand/Binder 20.242 1 20.242 0.931 0.339
Sand Max. S. 57.351 1 57.351 2.637 0.110
Silica Fl Bin. 1.224 1 1.224 0.056 0.813
Hoe. A;.lBin 22.788 1 22.788 1.048 0.311
Vaarum L 11.912 1 11.912 0.548 0.463
Compaction P. 27.326 1 27.326 1.252 0.268
Curing Con. 0.397 1 0.397 0.018 0.893
Em! 1130.839 52 21.747

 

 

 

 

 

 

 

 

92

3.6 DISCUSSION OF RESULTS

Among the eleven proportionin g / processing variables considered in this study, three (to-
tal ﬁber mass fraction, substitution level of virgin cellulose ﬁbers with recycled ones, and
ﬁber reﬁnement condition) proved to have statistically signiﬁcant effects, at 95% level of
conﬁdence, on the ﬁexural performance of wood ﬁber reinforced cement composites. In
order to Optimize the composites, it is thus necessary to determine the optimum combina-
tion of these variables which produce composites with highest performance-to-cost ratios.
In the optimization process, other variables with statistically insigniﬁcant effects on the
end product qualities may be ﬁxed. The speciﬁc levels for these ﬁxed variables (selected
based on the trends observed in fractional factorial analysis of variance shown in Figures
3.11-3.21, with due consideration given to the ease of processing and commercial produc-

tion conditions) are presented below:

Fiber source: American Fillers (“b")

Sand maximum size: 50 um (Refer to Table 3.2 for gradation)
Sand/binder ratio: 1.0

Silica fume/binder ratio: 0.1 (Refer to Table 3.3 for properties)
Flocculating agent/Binder ratio: 0.002

Vacuum level: 15 in (254 mm) of mercury

Compaction pressure: 200 psi (1.4 MPa)

Curing Condition: High pressure steam curing

3.7 SUMMARY AND CONCLUSIONS

In this ﬁrst phase of the experimental investigation, total of 11 key variables (factors) de-
ﬁning the production process of wastepaper ﬁber-cement composites were selected; the
main intent was to distinguish those factors with statistically signiﬁcant eﬁ‘ects on the

composite material performance characteristics. These variables were: (1) recycled ﬁber

93

source; (2) ﬁber mass fraction; (3) ﬁber beating level; (4) substitution level of virgin ﬁbers
with recycled ones; (5) sand/binder ratio; (6) maximum particle size of sand; (7) silica
fume/binder ratio; (8) ﬁocculating agent/binder ratio; (9) vacuum level; (10) compaction
pressure; and (11) curing condition. Each factor was considered at two levels in a (1/64)
fractional factorial design of experiments. This experimental design reveals the effects of
all variables on the composite material performance, but can not provide any information
on the possible interactions between different variables. The resulting composite, were

tested for ﬁexural performance (strength, toughness, and initial stiffness).

The ﬁexural test data was analde statistically by fractional factorial analysis of variance.
Among the eleven prOportioning / processing variables considered in this study, three
(total ﬁber mass fraction, substitution level of virgin cellulose ﬁbers with recycled ones,
and ﬁber reﬁnement condition) proved to have statistically signiﬁcant effects, at 95% level
of conﬁdence, on the ﬁexural performance of wood ﬁber reinforced cement composites. In
order to Optimize the composites, it is thus necessary to determine the optimum combina-
tion of these variables to produce composites with highest performance-to-cost ratios. In
the optimization process, other variables with statistically insigniﬁcant effects on the end
product qualities may be ﬁxed.

The recycled wastepaper ﬁbers were also analyzed and compared with virgin cellulose
ﬁbers. The recycled ﬁbers were found to be smaller in length than virgin cellulose ﬁbers.
The surface of the recycled ﬁbers was more roughened and ﬁbrillated by the recycling
process as compared to virgin cellulose ﬁbers. Cellulose content in recycled ﬁbers was
found to be lower than virgin cellulose ﬁbers. Recycled ﬁbers had a signiﬁcant amount
(close to 20%) of ﬁnes which are expected to play a ﬁlling role, rather than reinforcing

role, in cellulose ﬁber-cement composites.

CHAPTER 4

OPTIMIZATION OF INF LUENT IAL VARIABLES

4.1 INTRODUCTION

The three inﬂuential variables identiﬁed in the previous phase Of study (ﬁber mass frac-
tion, substitution level of virgin with recycled ﬁbers, and ﬁber reﬁnement level) were se-
lected to be optimized for producing composites with highest performance-to-cost ratios.
The optimization experimental design was formulated based on response surface analysis
techniques. The composites were optimized considering their ﬂexural performance

(strength, toughness, and initial stiffness) and cost.

“Design-Expert” a commercial software deve10ped for the purpose of statistical data anal-
ysis and presentation, was used for the optimization purposes. In general, this software
helps the user to understand relations between several variables (x1,x2,...) and one or more

response variables (R1,R2,...). “Design-Expert” can accomplish following types of goals:

(1) Response Surfaces: Response surfaces are plotted in the form of contour lines present-
ing the response as function Of two variables, while the remaining variables are held con-

stant.

(2) Optimization: The word Optimization, as used in this dissertation refers, to the process
of establishing the ranges of some variables (x1,x2,.) that allow some response variables

(R1,R2,..) to meet certain speciﬁcations (or constraints).

Optimization as done by “Design Expert” gives a graphical display Of the region of the ex-
planatory variables that simultaneously satisﬁes the constraints (or the speciﬁcation re-
quirements). Regions are shaded (or ﬁlled in) if one or more constraints fail within the
region. Blank region (non-shaded) do satisfy all constraints and they are referred to as the

optimum region (see Figure 4.2).

94

The Optimized composite identiﬁed in this phase of research were produced (at diﬁ'erent
total ﬁber volume fractions), and their mechanical and physical performance were com-

pared with those of conventional composites made fully with virgin ﬁbers.

4.2 OPTIMIZATION EXPERIMENTAL PROGRAM

The experimental program for optimization through response surface analysis (using “De-
sign Expert”) is presented in Table 4.1 Various combinations of the three statistically inﬂu-
ential variables are considered in this experimental program of the production of wood

ﬁber-cement composites.

95

Table 4.1 Optimization Experimental Program

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variables

Exper Fiber mass Fiber Fiber
iment Fraction Substitution Reﬁnement

# (%) Level (%) Level“
1. 12 75 12.5(500)
2. 8 50 20(480)
3. 16 50 20(480)
4. 8 100 5(490)
5. 16 100 20(410)
6. 12 75 12.5(500)
7. 12 75 12.5(500)
8. 16 100 5(490)
9. 12 75 12.5(500)
10. 8 50 5(570)
11. 16 50 5(570)
12. 8 100 20(410)

 

* min. of beating @3000 rpm (Canadian Standard Freeness)

 

96

For the remaining eight proportioning/processing variables (not statistically inﬂuential),
ﬁxed levels were used in the Optimization experimental program. The ﬁxed values were as
follows; ﬁber source: American Fillers (ﬁber “b”), sand maximum size: 50 um, sand!
binder ratio: 1, silica fume! binder ratio: 0.1, ﬂocculating agent lbinder ratio: 0.002, vacu-
um leve1215 in (254 mm) of mercury, compaction pressure: 200 psi (1.4 MPa) and curing

condition as high pressure steam curing.

4.3 TEST RESULTS AND ANALYSIS

Typical ﬂexural load-deﬂection curves produced for the composites of Table 4.1 are pre-
sented in Figure 4.1. Flexural strength, toughness and initial stiﬁ‘ness test results are given

in Table 4.2. Figure 4.2 presents some key analytical results produced in the optimization

process.

Flexural Stress (M Pa)

 

 

 

 

18
*‘Virgin Cel. Composite
“50% Flee.(expt. 2)
15 ”' /,r‘\..\ "'100% Rec.(expt. 5)
/ \\
J \
12 - ,’ "k
I \
I \
I \
I \
9 ‘— l. \\
, t
/ \
I \
6 _ I, ‘ \\
/.= \
.' \
a — ‘
/_ h
f \
/’ \\
l .
o I l 1 " l l \\:
0 0.5 1 1.5 2 2.5 3

Deflection (mm)

Figure 4.1 Typical Load-Deﬂection Curves.

Table 4.2 Flexural Performance

97

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Experiment Flex. Str. Mean Flex. Tou. Mean Init. Stif. Mean Init
(B R1111 (MPa) Flex. Str N-mm Tou. N/mm 8111'.
y (95% (95% (95%
Order) Con. Int.) Con. Int.) Con. Int.)
1 12.99 13.10 87.50 85.985 278.30 286.84
13.23 (0.24) 89.23 ($16.5) 294.23 ($31.4)
13.09 81.22 287.99

2 13.92 13.99 103.23 97.971 171.11 159.57
14.11 (0.18) 99.00 ($22.9) 158.34 ($43.0)
13.95 91.67 149.26

3 7.87 7.71 31.50 31.83 24.87 28.808
7.66 (0.30) 31.22 ($3.23) 29.432 ($14.3)
7.71 32.77 32.123

4 7.78 7.61 26.25 26.119 247.23 243.62
7.39 (0.4) 25.99 ($0.66) 233.67 ($34.2)
7.65 29.09 249.98

5 6.23 6.00 8.75 8.901 90.381 92.172
5.92 (0.38) 9.24 ($1.15) 99.366 ($25.4)
5.87 8.72 86.77

6 10.22 10.24 87.50 84.657 250.44 251.71
10.33 (0.16) 88.24 ($21.8) 261.49 ($36.0)
10.17 78.23 243.21

7 12.16 12.17 92.66 88.429 260.32 254.04
12.28 (0.21) 87.45 ($15.1) 255.88 ($28.9)
12.06 85.12 245.91

8 6.37 6.284 15.75 15.655 66.88 66.597
6.27 (0.17) 15.55 ($0.38) 69.70 ($12.7)
6.20 15.66 63.21

9 11.44 11.437 105.01 101.11 131.29 140.10
11.51 (0.16) 97.34 ($15.0) 143.23 ($30.3)
11.35 101.00 145.78

10 13.23 13.219 91.117 88.984 167.34 159.28
13.08 (0.26) 88.33 ($7.42) 161.39 ($36.4)
13.34 87.50 149.12

 

 

X3: lib bl

98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11 7.13 6.738 89.23 88.689 31.29 31.408
6.89 (0.35) 89.33 (4.04) 29.92 (6.06)
6.19 87.45 33.01
12 10.02 10.159 115.23 116.58 69.93 69.741
10.28 (0.49) 110.55 (5.49) 73.67 (15.75)
10.18 123.98 65.62
100.0 \ DESlGN-EXPERT Aneiyala
Model: \\
Linear 91.7.» \\
:::::.:'i \ ...
3’
Variables: g \ 1433
7 - 2: anL i 7501 \\
one an a; X 88.7 i ‘ 1 l
:0 015-.1250 \72
\.
00.3 \
“:00 030 10.07 12.00 13.83: 14.07 10.00
Xt; fib mi
20.00 DEMON-EXPERT Aneiyaia 20.00 DEan-IXPEIT Anliylio
17.00.. ".50“
”'00” / 10.00..
/ a
12.00 » 1" 1001 14:: “0° 8 12.00. 1433
/ ,‘3 ”‘4
10.00.. 10.00 ..
7.00.. , 7.00..
[I
6.00 1 - /
00.0 00.0 00.7 70.0 03.: 01:7 100.0 IWaco 0.0: 1200 10.3: 14.07 10.00
xz (lb .1101. X1: fib mt

a. Flexural Strength

Figure 4.2 Optimization: Response Surface Analysis.

Model:
Linear

Reaper-tee:
Toughneee

Vertebiee:
X . (lb mt
Y - lib eubL

Conetanta:
05 hi - 12.60

DESDN-IXPEIT Analyail

 

 

 

 

DESDN-IX'IIT Moiyale

100.0 \\
01.7..
00.0.»
8'
I
9 70.0. 9”"
N
x 007
' ‘ 0.020
00.0.
56.6 - -\ - - \\.
0.00 0.03 10.07 12.00 10.00 14.071000
thfib mi
20.00

oesnmurear Analyaie

 

20.00

"0.0 0

10.00 0

7.00 o

 

0.00

10.00 ..
3
I 12.00 4
a
X

2/

/....

 

0.35

   

 

 

0.00

b. Flexural Toughness

0.30

10.07 12.00
X1; fib mi

13.0: 14707 10.00

Xlzibbi

 

17.00 0

10.00 4

12.00 4

10.001

7.504»

 

0.00

 

 

 

50.0

00.0 00.7 70.0 00.: 01.7 100.0
xz fib aubl.

Figure 4.2 (Cont’d) Optimization: Response Surface Analysis.

100

DESDN-EXPIIT AnalyIIO

 

 

 

 

 

 

   

 

 

 

 

 

 

100.0
M0601;
Lu" 01.7..
00090000:
0111 0111 03.3..
.1
0
Veﬂoblu: 3
x u "b N" D 78.01» .3.‘
8
Y - 110 000L
N
x 00 71
Cami-r110: ° '
ﬂb bl - 12.00
6.-30
00.0 0 J - - - t
0.00 0.33 10.07 12.00 13.33 14.07 10.00
X1; lib r111
”SIGN-EXPERT Anulyele
DESDN-IXPERT Analysle 209°
20.00
17.00.
"0.01? r 0"
.00..
10.00.» 10
3
3 e
e 12.001» .. 12.00»
0)
it
x x 1000
10.00.. ' "
7. 0
7.00.. 0 "
0.00 . . - ‘ =
8-00 ‘ - - - g 0.00 0.00 10.07 1200 10.00 14.07 10.00
00.0 00.3 00.7 70.0 03.3 01.7 100.0
X1: 110 m1
X2. 110 lubL

1:. Initial Stiffness

Figure 4.2 (Cont’d) Optimization: Response Surface Analysis.

Lheer

Remains;
Cast

Venebhm:
x-lbsubL
Y-DN

Cunt-Ms;
lb n11 - 1200

d. Cost

Response

R1: Flex str
R2: Init Stif
R3: Toughness
R4: Cost

It Flex 01!
lazhu 001
00:10mmvwms
I4: Cost

Constants:
00 01- 1200

c. Optimization

X3: lib bl

101

DESEN-IXPIIT Analyst!

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20.00
1100.
1000»
1”” 0.010 0.40
1000»
7000
000 -. - - ‘ ‘
00.0 00.3 00.7 70.0 03.3 01.7 100.0
xz:lnasubL
Lo Hi Transfer:
1900 . 000 .
1 . 500 .
0.390 .
O . 5 4 0 .
Demon-0x020? Anelysle
1000
0t7u
013w
110_L0
760°‘* 2-L°
00.7.»
00.3.»
m~L°/ I4_Lo
000 \> L - - - - ’
0.00 0.33 10.07 12.00 13.33 14.07 10.00
X1; flb ml

Figure 4.2 (Cont’d) Optimization: Response Surface Analysis.

Linear
Linear

102

The response surface contours of Figure 4.2 can be interpreted to understand eﬂ'ects of
several variables on response. Looking at the response surfaces for ﬁexural strength, we
ﬁnd that ﬂexural strength (Figure 4.2 a) increases as the ﬁber substitution level is de-
creased from 100% to 50%.; also we observe ﬂexural strength increases as ﬁber mass frac-
tion is reduced from 16% to 8%. We also observe that the ﬁexural strength increases as the
ﬁber heating is increased from O to 15 minutes of beating.

In the analysis of ﬂexural toughness response surfaces (Figure 4.2 b), it can be observed
that toughness decreases as the ﬁber mass fraction is reduced from 16% to 8%. Toughness
is also reduced the ﬁber substitution is increased from 50% to 100%. Toughness increases
as the beating level increases from 0 to 20 minutes of beating.

Analyzing the response surfaces contours for initial stiﬁ‘ness (Figure 4.2 c), stiffness is ob-
served to increase as the ﬁber mass fraction is decreased from 16% to 8%. Stiffness in-
creases also with the increase of ﬁber substitution level from 50% to 100%. As the beating

time is increased the initial stiﬁ'ness also increases.

The increase in the ﬁber substitution level is observed in Figure 4.2 d to reduce the cost of

raw materials in the composite.

Optimization plots were then generated (see Figure 4.2 e) for achieving minimum strength
of ﬂexural strength (1900 psi, 13.2 MPa), ﬂexural toughness of (0.39 lb in, 44 N-mm) and
initial stiffness of (1.5 Klin, 175 N/mm) at the lowest possible cost. Figure 4.2 0 shows the
discarded (shaded) regions and the optimum (clear) region for optimum performance.

The optimum levels of the statistically inﬂuential variables derived form the above pro-

cess are as follows:
Fiber Mass Fraction: 8%
Substitution Level of Virgin with Recycled Frbers: 50%

Fiber Reﬁnement Level: Canadian Standard Freeness of 540 (12.5 minutes of beating
@3000 rpm)

The ﬁxed levels of other prOportioning / processing variables in the optimized composite

were as follows:

Sand/Binder Ratio =1; Silica Fume/Binder Ratio = 0.1; Vacuum Level = 15 in (254 mm)

103

of mercury; Compaction Pressure = 200 psi (1.4 MPa); Curing Condition = High Pressure
Steam Curing.

The ﬁexural performance of the optimized composite is compared in Figure 4.3 and Table
4.3 with that of the control composite made fully with virgin softwood kraft cellulose ﬁ-
bers. The optimized recycled composite is observed to produce ﬂexural performance char-
acteristics comparable to those obtained when the composite is made fully with high-
quality virgin ﬁbers (softwood kraft pulp). Analysis of variance of ﬂexural strength,
toughness and stiffness test results (Table 4.3) indicated that, at 95% level of conﬁdence,
the optimized recycled and virgin composites has statistically comparable ﬂexural
strength, and the difference in ﬁexural toughness and stiﬁ'ness was statistically signiﬁcant.
The recycled composites had an average ﬂexural strength which was only 1.6% less than
that of virgin composites.

Table 4.3 Results of the Analysis of Variance (Flexural Strength, Tbughness and Ini-
tial Stiffness)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Flexural Strength
Sauce Sum-0f -Squsm DF Mean-Square F-Rstio 1’
Type «cm... 0.061 1 0.061 1.837 0.191
an: 0.635 18 366238.485 0.033
7°"th
Type dCormosite I636079 1 I 636.079 F 26331 I 0.00
am I434.825 18 I 24.157 I I
1611105111160.
typedepodrc It306.051 I 1 [1306051 I 97.559 I 0.00
Error I24o.97 I 18 I 13.387 I I

 

0 Flexural Stress (MPO)

 

— Cont. (100‘ 1111.) ‘ ' ' Opt. (60$ rec.)

‘5 " I if

: '. ‘\
' '~ ‘\
IA I. K
‘0 I. “ “\

 

 

 

 

Type of Composite

Flexural Toughness (N-mm)

 

100

 

 

0011111000 vir.) 091.1000 rec.)

Type of Composite

50*
25'
at l l

 

l()4

 

O Flexural Strength (MPO)
2

15"

10r'

 

 

0011111001. vlr.) 0011005 rec.) -
Type of Composite

0 Flexural Stiiiness (N/mm)

160

100

60

 

 

 

r-
p
1 l

Cont. (100$ vir.) 0914000 rec.) -
Type 01 Composite

 

 

Figure 4.3 Flexural Performance of the Optimized Recycled Composite Vs. Virgin

Composite.

105

4.4 EVALUATION OF THE OPTIMIZED COMPOSITE

Density, water absorption and dimensional stability tests were carried out on the optimized
composites following ASTM C1186 [78] procedures. Density is deﬁned in this investiga-
tion as the mass per unit volume of the composite expressed in gm/cm3. Water absorption
is the increase in mass of the test specimen expressed as a percentage of its dry mass after
immersion in water for a speciﬁed period of time as prescribed in ASTM C1185 and
ASTM C1186. Dimensional stability is measured in terms of the linear variation in length
of test specimen. with change in humidity from 90% to 30% as per ASTM C1185. An
environmental chamber (see Figure 4.4) was used to produce the humidity and tempera-

ture conditions required for different tests.

 

Figure 4.4 Environmental Chamber For Conditioning of Test Specimens.

106

Optimized composites were produced at different ﬁber made fraction of 4%, 8%, and
12%, noting that 8% is the optimum ﬁber content These composites were evaluated (ver-
sus the control composite) based on ﬁexural performance, density, water absorption, di-

mensional stability and moisture content (ASTMC1186).
4.4.1 FLEXURAL PERFORMANCE

The ﬂexural load-deﬂection curves for the optimized composites with different total ﬁber
contents are compared with that of control composite made fully with virgin softwood
kraft pulp (see Figure 4.5). The recycled composites all had 50% virgin ﬁbers replaced

with recycled magazine paper ﬁbers.

Flexural Strength (MPa)

 

10

" 'Virgin Cel. Composne
h'Opt. (mi 8%) —"
15 _"Opt. (m14%) ’ ——t ‘

”'Opt. (mi 12%) ,’ x
12 _ ,l \

 

 

 

 

0 .I 1 1 1 r l
0 0.5 1 1.5 2 2.5 3

Deflection (mm)

Figure 4.5 Flexural Load-Deﬂection Curves.

107

Figures 4.6 through 4.8 compare the ﬂexural strength, toughness, and initial stiffness of
optimized composite vs. those of the control composites. While ﬁexural strength and
toughness at 8% ﬁber mass fraction is observed to drop with the substitution of 50% of
virgin ﬁbers with recycled ones, the initial ﬂexural stiffness is observed to increase with
the use of recycled ﬁbers. This may be illustrated by the fact that the ﬁne fraction of recy-
cled ﬁbers acts more as ﬁllers than reinforcing ﬁbers in the composite. The higher rein-
forcement efﬁciency of virgin ﬁbers reﬂect in higher ﬂexural strength and toughness
qualities of the composite, while the ﬁller action of ﬁnes in recycled ﬁbers leads to a dens-
er structure of the matrix which reﬂects in a higher initial stiffness. It is important, howev-
er, to note that the differences between the qualities, obtained with virgin and optimized
recycled ﬁbers (at 8% ﬁber mass fraction), are relatively small. Figures 4.6 through 4.8
also indicate that the increase in ﬁber content from the optimum levels of 8% to 12% has

adverse effects on ﬂexural strength and stiffness, but tends to improve toughness charac-

teristics.

Flexural Strength (MP0)
16

 

l4—

lO~

8 __ -60: Recycled-Optimized
+100: Virgin

 

O l 1 1 1
2 4 6 8 1 O 1

Fiber Mass Fraction (%)

 

 

14

ref

Figure 4.6 Flexural Strength Vs. Fiber Mass Fraction.

108

FlexuraL Toughness (N—mni)

 

140

120—

100*-

80*-

60~

40»

 

+

9 007. Recycled -Optt mixed

+100: Virgin

l l

 

 

6 8 10 12
Fiber Mass Fraction (7.)

Figure 4.7 Flexural Toughness Vs. Fiber Mass Fraction.

Initial Stiffness (N/mm)

l4

 

400

350 —

300 -

250

l

l

200

T

150
100—

50

l

 

0 so: Recycled-Optimised
+100: Virgin

l l J l

 

 

f0

6 8 10 12
Fiber Mass Fraction (%)

Figure 4.8 Initial Stiffness Vs. Fiber Mass Fraction.

l4

10‘)

4.4.2 DENSITY

The measured density of recycled and virgin ﬁber-cement composites are presented in
Figure 4.9. At 8% ﬁber mass fraction. substitution of half of the virgin ﬁbers with recycled
ﬁbers is observed to increase the density of the composite. This further conﬁrms the ﬁlling
action of the ﬁne fraction of recycled ﬁbers which leads to a denser structure of composite.
The increase in the total ﬁber mass fraction is observed in Figure 4.9 to consistently re-

duce the density of the composite material.

Density tgrii/cu cm)

 

L8

L6r—

L4e-

.302 Recycled Optimised
+100: Virgin

0.8 —

0.4 *-

(LZe

 

 

0 l l l .l l

2 4 6 8 10 12 14

Fiber Mass Fraction (%)

 

Figure 4.9 Density Vs. Fiber Mass Fraction.

4.4.3 WATER ABSORPTION

Optimized composites with recycled ﬁbers are observed in Figure 4.10 to show reduced
water absorption when compared with control composites made fully with virgin ﬁbers.
The denser structure resulting from the ﬁlling action of the ﬁnes in recycled ﬁbers could
be responsible for this phenomenon. Figure 4.] I shows strong correlation between water
absorption and density of recycled composites with different total ﬁber mass fraction (cor-

relation coefﬁcient 0.81).

110

Water Absorption (%)
35

 

30~ x

25r-

l

20

. 50! Recycled-Optimized
X ioox Virgin

15*-

10r-

 

 

 

O l 1 i 1 l
2 4 6 8 10 12 14

Fiber Mass Fraction (%)

Figure 4.10 Water Absorption Vs. Fiber Content.

Inter Absorption (%)

 

50

40»-

30’— \

10—

 

 

l L

l 1.2 1.4 1.6 1.8 2
Density (g/cu. cm)

 

Figure 4.11 Correlation Between Density and Water Absorption in Recycled Com-

posite.

111

4.4.4 DIMENSIONAL STABILITY

It is important for the thin-sheet cement products to have an acceptable dimensional sta-
bility at varying moisture content Dimensional stability is measured (ASTM C1186) in
terms of moisture (dimensional) movements expressed as the percentage changes in

length as relative humidity is increased form 30% to 90%.

The dimensional (moisture) movement of the optimized recycled composite is compared
with that of the control composite (both at 8% total ﬁber mass fraction) in Figure 4.12.
The optimized composite is observed to possess a better dimensional stability than the vir-
gin ﬁber—cement composite. The denser structure of recycled composite and its reduced
water absorption at least partly describe the desirable aspects of the recycled composite

performance.

Dirnomlonal Change (96)
0.1

 

0.08

0.06

0.04

0.02

 

 

 

 

 

Type Of Composite

Figure 4.12 Dimensional Stability Test Results.

112

4.7 MOISTURE CONTENT

The moisture contents (at 50% Relative Humidity in the environment) of recycled com-
posites with different mass fraction are compared in Figure 4.13 with that of the control
composites.The increase in total ﬁber content of recycled composite is observed in Figure
4.13 to increase the composite moisture content. This could be illustrated by the increase
in porosity of matrices incorporating higher ﬁber contents, and also by the afﬁnity of wood
ﬁber for moisture. Recycled composites, with a dense structure resulting from the ﬁlling

action of ﬁne constituents in recycled ﬁbers, are observed to have lower moisture contents

than virgin composites.

Moisture Content (%)

 

0.7

0.6

0.5

 

0.4

+..... u...

 

0.3

 

0.2

0.1

 

 

 

° 4 8 12 16
Fiber Mass Fraction (%)

Figure 4.13 Moisture Content Vs. Fiber Mass Fraction.

113

4.5 TECHNICAL EVALUATION OF RECYCLED COMPOSITES

ASTM speciﬁcations and two leading commercial products provided the criteria for tech-
nical evaluation of the optimized recycled wastepaper ﬁber-cement composites. Table 4.4
presents a comparison of some key technical qualities of the recycled composite produced
in this research versus ASTM limits and also those of commercial products.

 

 

 

 

Table 4.4 Technical Evaluation
Properties ASTM C l 185 Commercial Commercial Optimized
Product 1 “‘ Product 2 ** Composite
Flexural Strength Grade I: 580 2000-3000 2000 1900-2100
(psi) Grade II: 1450
Grade 111: 2320
Density (gm/cm3) does not specify 1.21 1.35 1.4
Dimensional Stabil- does not specify 0.06 0.05 0.06
ity (%)

 

 

 

 

 

 

 

"' manufactured by James Hardie; ‘N manufactured by Eternit

Table 4.3 indicates that the optimized composite meets the minimum requirements for
ﬂexural strength of ASTM Grades I and II thin-sheet cement products. ASTM C1186-91
speciﬁes two types (“A” and “B”) of ﬂat sheets, according to their intended applications.
Type “A” sheets are intended for exterior applications, where they may be subjected to the
direct action of sun, rain, or snow. Type “B” sheets are intended for interior applications,
and for exterior applications, where they will not be subjected to the direction action of
sun, rain, or snow. The sheets are further classiﬁed into four grades according to their ﬂex-
ural strengths. ASTMC1186 does not specify limits on density or dimensional stability.
When compared with commercial products, the recycled composites are observed in Table

4.4 to possess acceptable density and dimensional stability.

The recycled composites produced in this research seem to present technically viable and

economically! environmentally superior alternatives to conventional wood ﬁber reinforced

114

thin-sheet cement products.

4.6 SUMNIARY AND CONCLUSIONS

The inﬂuential variables in the processing of recycled wood ﬁber-cement composites were
optimized based on response surface analysis techniques. The variables optimized here
were: total ﬁber mass fraction, level of substitution of virgin ﬁbers with recycled ﬁbers,
and the beating (reﬁnement) level of ﬁbers. Optimization was based on ﬂexural strength,
initial stiffness and toughness of the composites. Due consideration was also given in the
optimization process to the cost of raw materials. The optimized composites were then
technically evaluated versus virgin composites, ASTM speciﬁcations, and commercial

products. The conclusions derived are summarized below.

(1) Analysis of results indicated that optimum composites are obtained using 8% ﬁber
mass fraction, 50% substitution level of virgin with recycled ﬁbers, and reﬁnement (heat-
ing) of ﬁbers to a Canadian Standard Freeness (CSF) of 540.

(2) The optimized recycled wood ﬁber-cement composites were shown to possess ﬂexural
strength, density and dimensional stability characteristics satisfying ASTM speciﬁcations
and comparable to those of commercially available virgin wood ﬁber reinforced thin-sheet

cement products.

(3) The optimized recycled composites produced ﬂexural strength, stiﬂ'ness and toughness
characteristics comparable to those of virgin composites. Compared to virgin wood ﬁber-
cement composites, the optimized recycled composites possessed somewhat lower ﬂexur-
al strength and toughness but higher initial ﬂexural stiffness. The difference in ﬂexural
toughness and toughness were statistically signiﬁcant. Recycled composites also showed
reduced moisture (dimensional) movements, lower water absorption and moisture content,

and higher density when compared with virgin wood ﬁber-cement composites.

(4) The ﬁne content of recycled ﬁbers seem to play more of a ﬁlling role than a reinforcing
role. Hence, recycled composites present a denser microstructure which reﬂects in higher
stiffness, lower water absorption and moisture content and reduced dimensional (mois-
ture) movements of recycled composites. Reduced reinforcing action of ﬁnes in recycled
ﬁbers, however, reﬂects in somewhat reduced ﬂexural strength and toughness of recycled

composites when compared with virgin composites.

CHAPTER 5

DURABILITY AND MOISTURE SENSITIVITY

5.1 INTRODUCTION

The service life of construction materials is expected to be several decades, and therefore
there is a need to evaluate the long-term performance of new construction materials in the
environment they will be exposed to. Potential problems with the durability of wood in ce-
ment-based matrices further underline the critical need for durability studies of any wood-
cement composite. Furthermore, the sensitivity of wood to moisture makes it necessary to

assess the composite material performance at variable moisture conditions.

The main thrust of this phase of research was to assess and (if necessary) improve the
long-term durability and moisture-sensitivity of the optimized wastepaper ﬁber-cement
composites developed in this investigation. For the assessment of long-term durability, ac-
celerated ageing tests in laboratory were adopted in order to simulate long-term ﬁeld ex-

posure conditions.

With the development of new ﬁbers, and their application in composites in different cli-
mates, a need often arises to devise tests for a particular material and application, and such
tests are seldom detailed in the standards or speciﬁcations of the various agencies. In de-
veloping accelerated ageing tests, two stages must be considered. First, the potential age-
ing mechanisms should be identiﬁed in order to choose an appropriate means of
accelerating them, using variables such as temperature, radiation or moisture eﬂ‘ects. Sec-
ond, the duration or number of cycles in the accelerated test should be translated into time
in natural weathering conditions. The correlation between the time in accelerated and nat-

ural ageing is not a unique function, since it depends on the climatic conditions in different

115

116

zones, and even within the same zone there may be differences in the micro-climate, for

instance, the direction in which the component faces.

The accelerated ageing tests adopted in this study cover the ASTMC1186 [78] methods as
well as those selected to pronounce certain aspects of the weathering eﬂ'ects on wood ﬁ-
ber-cement composites. The ageing effects on engineering properties and the microstruc-
ture of composites were investigated. Microstructural studies utilized the Scanning
Electron Microscopy, therrnogravirnetry and X-ray diﬂ'raction techniques. Appropriate
measures were adopted and evaluated for controlling the ageing and moisture eﬂ‘ects on

wastepaper ﬁber—cement composites.

5.2 EXPERINIENTAL NIETHODS

The effects of moisture and accelerated ageing on the ﬂexural performance and micro-
structural characteristics of the optimized wastepaper ﬁber-cement composites and control
composites (made fully with virgin softwood kraft ﬁbers) were investigated. This section
introduces the moisture sensitivity, accelerated ageing and microstructural evaluation pro-

cedures used in this investigation.

5.2. 1 Moisture Sensitivity

In order to assess moisture eﬂ‘ects on the composite material performance, ﬂexural test
specimens were conditioned as follows and then subjected to ﬂexural loading [78]:

Air drying: Place the test specimens, for 4 days in a controlled atmosphere of 73 ° F (23
°C) and 50% relative humidity and in such a manner that all faces are adequately ventilat-
ed.

Oven-drying: Dry out the test specimen in an oven at 216 °C (102 °C) until the difference
between two consecutive weighing, at intervals not less than two hours, does not exceed
0.1% by mass.

Saturation: Immerse specimens to be tested in wet conditions in water at a temperature of
73 °F (23 ° C) for a period of 48 hours. Test the specimen immediately upon removal from

water.

Previous studies on virgin wood ﬁber-cement composites indicate that moisture effects on

117

the ﬁbers and their bond to cementitious matrices fully govern the moisture effects on
composite material properties. There are, however, some eﬂ'ects of moisture on cement-
based matrices (though overshadowed by moisture effects on the ﬁbers and interface

zones) which are described in the following.

Rapid drying of cement based materials may induce tensile cracks due to non-uniform
drying (and hence diﬂ‘erences in drying shrinkage) of the specimen. These cracks do not
have much eﬂ'ect on compressive strength but will lower the ﬂexural and tensile strengths
[70,71]. If drying takes place very slowly, the internal stresses can be redistributed and
alleviated by creep, an increase in strength may result from drying.

Wetting of cement-based materials may lead to losses in compressive strength as a result
of the dilation of cement gel by adsorbed water and also breaking of Si-O-Si bonds, which
lead to reduction of the cohesion between solid particles. Conversely, when the wedge-ac-
tion of water upon drying ceases, an apparent increase in strength of the specimen is re-
corded. Resoaking of oven dried specimens in water reduces their strength to the value of
continuously wet-cured specimens, provided they have been hydrated to the same degree.

The variation in strength due to drying is thus a reversible phenomenon.

5.2.2 Accelerated Ageing

The accelerated ageing conditions used in this investigation are adopted to pronounce the
physical and chemical causes of deterioration in wood ﬁber reinforced cement composites.
These methods and mechanisms through which they accelerate the ageing process are de-

scribed in this section.

Repeated Wetting and Drying

Repeated wetting and drying cycles simulating repeated rain-heat conditions in natural
weathering promote some key chemical and physical mechanisms of deterioration in
wood ﬁber-cement composites. These conditions accelerate any potential attack by the al-
kaline pore water of cement-based matrices on certain wood ﬁber constituents; they also
promote migration (through dissolution and re-precipitation) of some cement hydration

products from the matrix into the ﬁber cores and their interface zones. These microstruc-

118

rural changes would reﬂect the engineering qualities of aged composites.

The repeated wetting-drying test adopted in this investigation (see Figure 5.1 for test set
up) follows the ASTM C1185 procedure. A total of 25 cycles of wetting/drying were used.
In each cycle. specimens were moistened by spraying water for three hours at 30°C (86
°F). and then dried for three hours at 60° C (140 °F).

   

    

:d. 'f‘"

    
 

’0

. -....._o-
u a I

    

Figure 5.1 Wetting/Drying Experimental Set Up.

Freeze/thaw

This test investigates the possible degradation of cement-based materials exposed to re-
peated freeze- thaw cycles. Freezing of water in the cement paste capillary pores, due to
the volume increase of water upon Turing to ice, would cause internal pressures which
lead to cracking and deterioration of concrete. A total of 50 cycles are applied as required
by ASTM C1185. Each cycle lasts 4 hours, and consists of cooling the specimen to -20°C
(-4 °F) over a period of one hour, holding the specimen at -20°C (-4 0F) for one hour,

119

thawing it to 20° C (68 °F) over a period of one hour, and maintaining the specimen for
one hour at 20°C (68 °F) before proceeding to freezing. The experimental set up is shown

in Figure 5.2.

 

Figure 5.2 Freeze-Thaw Test Apparatus.

Wetting/Drying and Carbonation

Carbonation plays a key role in natural ageing of cellulose ﬁber-cement composites. Dis-
solution of calcium hydroxide in pore water and its precipitation within wood ﬁber cores
and at the interface zones would be accompanied with carbonation which turns calcium
hydroxide into calcium carbonate and pronounces the weathering effects. Each cycle in
this accelerated ageing test consists of 8 hours of saturation under water, heating in oven
for one hour at 80° C (176 ° F), carbonation for a period of 5 hours in a rich carbon diox-
ide environments, heating for 9 hours in oven, followed by cooling for one hour at room

temperature

A constant ﬂow carbon dioxide incubator was used to produce the rich carbon dioxide en-

120

vironment. The carbon diOXidC (supply) cylinder regulator was adjusted to provide 0.14
MPa (20 psi) to the unit. The carbon dioxide ﬂowmeter was adjusted to 0.4 liters per
minute by the adjusting valve on front of the chamber. The air ﬂowmeter was adjusted to
4.0 liters per minute to obtain l0‘7c C02 concentration (rich carbon dioxide environment)

in the chamber. The carbonation chamber used in this study is shown in Figure 5.3.

The extent of carbonation was determined using thermogravimetric analysis and x-ray dif-
fraction. Density of the optimized composite was also determined to investigate any phys—
ical Changes accompanied with chemical changes before and after the wetting-drying

cycles and carbonation.

 

 

 

Figure 5.3 Carbonation Chamber Producing Rich Carbon Dioxide Environment.

Hot Water Bath

Any deleterious chemical reactions (e.g. alkali attack of pore water on some wood ﬁber

constituents) taking place under natural ageing would be accelerated upon immersion in

121

hot water. ASTM C1185 speciﬁcations were followed; the temperature of water was 60° C
(140 °F), and the immersion period was 55 days.

Ten replicated panel specimens (12 x 4.5 x 0.25 in) were soaked in hot water. Flexural
tests were performed according to the Japanese Specification JCI-SF [77]. The test results
on aged specimens were compared with those of saturated unaged specimens. The hot wa-

ter bath used in this investigation is shown in Figure 5.4

 

Figure 5.4 Hot Water Bath.

Permeability

Water permeation through capillary pores into cement-based materials is fundamental to
their long-term durability characteristics. In hydrated cement paste, the size and continuity
of the pores at any point during the hydration process would control the coefﬁcient of per-
meability. A direct water permeability test set up was designed (see Figure 5.5). The water
permeated was measured and the coefﬁcient of permeability was determined using Dar-

cy’s expression.

122

The test specimens used for this test was cylindrical with 100 mm (4 in) diameter and 6
mm (0.25 in) thickness. The saturated test specimen is placed in between plexiglass rings
and clamped by four threaded bars. The specimen perimeter surface is glued in order to
ensure all permeation is across the thickness of the specimen. Top portion of the ring is
ﬁlled with care to ensure all air bubbles are eliminated during ﬁlling and a head of water
about 300 mm (12 in) is formed, in the upper pipet. The drop in head is noted after 24
hours. The observations are continued until a steady state ﬂow is achieved, and penne-

ability coefﬁcient is then calculated.

 

Figure 5.5 Direct Water Permeability Test Set Up.

5.2.3 Microstructural Investigations

Changes in the microstructure of composites upon ageing were investigated using the
Scanning Electron Microscopy and thermogravimetry techniques. X-ray diffraction was
also used to conﬁrm the results of thermogravimetric analysis. These microstructural test

procedures are brieﬂy described in this section.

Scanning Electmn Microscopy: In order to investigate the ageing mechanism and the

123

corresponding microstructural changes, fractured surfaces of the composites were ob-
served under SEM. The focus was to observe changes in ﬁber morphology and failure
modes, ﬁber-matrix interface zones and the matrix upon ageing.

Thermogravimetry: Thermogravimetric analysis was used to determine compositional
changes (in calcium hydroxide and calcium carbonate content) in wastepaper ﬁber-cement

composites under various accelerated ageing conditions.

X-Ray Diﬂraction: In order to supplement and support the thermogravimetric observa-
tions, x-ray diffraction was used to determine the compositional changes (in calcium hy-

droxide and calcium carbonate content) upon ageing.

5.3 MOISTURE AND AGEING EFFECTS ON ENGINEERING PROPERTIES

5.3.1 Moisture Sensitivity

The ﬂexural load-deﬂection behavior of recycled and virgin wood ﬁber-cement compos-
ites at different moisture conditions are shown in Figure 5.6a. It appears that the increase
in moisture to saturated level adversely affects the ﬂexural strength and stiffness of both
virgin and recycled composites but causes an increase in ﬂexural toughness (see Table 5.1
and Figures 5.6 b through 5.6 d).

Results of the analysis of variance of test data is presented in Table 5.2. One way analysis
of variance of ﬂexural strength, stiffness and toughness test results at different moisture

contents conﬁrmed the moisture sensitivity of recycled and virgin composites.

There was a signiﬁcant difference in ﬂexural strength for various conditions of moisture
content in virgin and recycled composite. However, air-dried and-oven dried recycled
composites proved to be statistically equivalent in ﬂexural strength but diﬂ'erent in ﬂexural
toughness and stiffness at 95% conﬁdence level (Using multiple comparison by contrast).
Two-way analysis of variance of results was also conducted for different composites (re-
cycled versus virgin) and moisture conditions. The ﬁber-moisture interaction proved to be
statistically signiﬁcant at 95% level of conﬁdence, indicating that moisture eﬂ'ects on re-
cycled composites differ from those on virgin composites. Damaging effects of saturation

are less pronounced in the case of recycled ﬁbers. Upon saturation recycled composites,

124

when compared with air-dried condition, has a drop of 37% in ﬂexural strength and 16%
in initial stiffness, while the corresponding drops in virgin composites were 47% and 28%,
respectively. Both recycled and virgin composites showed an increase of 32% in ﬂexural

toughness upon saturation.

Table 5.1 Flexural Performance of Recycled Wastepaper Fiber-Cement Com-

posite at Diﬂ’erent Moisture Conditions

 

 

 

 

 

 

 

 

 

 

Experiment Flex. Str. Mean Flex. Tou. Mean Init. Stif. Mean Init

(MPa) Flex. Str N-mrn Tou. N/mm Stif.

(95% (95% (95%
Con. Int.I Con. Int. Con. Int.)

Control 8.2 8.06 91.234 90.72 119.90 118.6
Saturated 7.993 (3.28)) 89.887 (111.4 121.29 (119.9)

8.114 87.81 130.32

8.003 92.431 115.87

8.117 91.098 117.32

8.007 85.125 110.43

7.996 90.543 115.55

8.01 89.67 117.98

8.12 95.098 119.56

8.09 94.333 118.54
Control 15.321 15.09 70.39 67.88 171.17 164.35
Air Dried 15.003 ($.63) 69.342 (18) 170.77 (i181)

14.894 66.455 166.33

14.998 63.22 167.45

14.888 67.87 163.38

15.226 66.43 160.23

15.115 68.23 159.98

15.301 69.09 157.78

15.227 69.02 160.98

15.02 68.78 165.43

 

 

125

 

 

 

 

 

 

 

 

 

 

 

Control 15.72 15.38 53.234 51.99 189.32 151.66

15.12 (10.88) 54.123 (17.84) 180.87 (110.2)
Oven dried 15.439 50.669 183.66

15.445 52.144 182.87

15.056 50.1 179.95

15.554 48.21 183.32

15.221 48.321 185.79

15.666 53.872 184.45

15.221 50.989 183.24

15.432 51.996 185
Recycled 9.554 9.392 80.003 77.512 155.32 151.66
Saturated 9.1 12 (10.5) 78.21 1 (17.0) 154.22 (18.7)

9.443 77.439 153.87

9.234 77.55 150.37

9.501 80.891 152.12

9.456 76.8 153.87

9.342 75.124 149.96

9.37 76 150.27

9.531 75.898 150.44

9.493 77.205 148.99
Recycled 15 14.91 59.999 58.65 183.98 180.51
air dried 14.891 (10.34) 60.21 (4.72) 182.34 (7.52)

14.995 58.367 182.34

14.992 59.381 180.54

14.89 58.002 180.65

14.95 57.10 180.22

14.96 57.98 178.43

14.765 56.993 177.87

14.779 58.554 178.87

14.914 60.013 179.87
Recycled 15.31 15.03 54.11 53.23 234.46 228.53
Oven dried 15.029 (10.42) 54.123 (15.75) 230.00 (121.5)

15.002 54.987 221.98

15.01 53.1 225.55

15.08 54.876 228.11

14.997 53.223 218.89

15.026 54.775 232.32

14.96 52.092 236.33

14.992 51.796 230.34

14.9 51.285 227.34

 

 

126

Table 5.2 Results of the Analysis of Variance (Flexural Strength, 'Ibughnese and Ini-
tial Stiffness)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Heat-1.1 Strength
Virgin Recycled
Source Sum-of DF Mean- F- P Sum-0f DF Mean- F- P
' Square Ratio ' Square Ratio
Squares Squares
Ween-144.211 2 72.105 1726.1 0.00 1053 2 5267 256.3 0.00
ditioa 36
Farm 0459 11 0.042 0.247 12 0.021
70081100"
Virgin Recycled
moia.cond. 3679.5I 2 I1339.7I 691.0I 0.00 1725.3I 2 I362.9 I565.35I 0.00 I
Error 31.949I 12 I2.662I I 13316] 12 I1526I I I
Initial Stiﬂ'neaa
Virgin Recycled
Moon. 6633.9 2 [3344.4 I 3.76 0.054 14539.I 2 I 7269 I573.93I 0.00 I
Error 10651.9 12 [887.66I 150.67I 12 112.556I I I
‘I‘wewayAnalyabeIVari-aeeefRe-h
Flexural Strength Toughness
moist 301.606 2 150.303 1664.6 0.00 430.24 1 430.2 143.4 0.00
1
ﬁber 0.762 1 0.762 3.41 0.007 2239.4 1 2239.4 663.69 0.00
moist‘ﬁber 3.719 2 1.359 20524 0.00 20.21 1 20.21 6.03 0.026
aror 2.537 23 0.091 53.53 16 3349
11111181511306."
moist 2397.3 1 2397.3 139.93 0.00
riser 7095.1 1 7095.1 342.3 0.00
moist‘ﬁber 463.39 16 468.89 22.65 0.00
FJra 331.15 16 20.697

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

127

 

 

 

 

 

 

 

 

 

Pleural emu (MM)
20 — Air-Dried "Oven Dried "'Saturated
,T‘\ I 95% Confidence

1 5 *- " — Interval
‘ o P- I,
5 r 1'
0 i L
0 1 4
Deflection (mm)
Virgin
Flexural Streee (MPa)
20 - Alt-Dried ' " Oven Dried "' Saturated
95% Conﬁdence
‘5 r- I, _L_ IMONII
10 l' I” I \\
s i- ll :1
III “t.
o , 1“ ‘
0 1 2 3 4
Deﬂection (mm)
Recycled

(a) Load-Deﬂection Behavior
Figure 5.6 Moisture Effects on Flexural Behavior of Recycled and Virgin Wood Fi-

ber-Cement Composites.

ﬂea-rel mm!”
Flam EA» om Elam om

 

 

“‘ "I“ .....

......

.......

10"

(b) Flexural Strength

.......

 

 

 

 

 

 

 

 

“WW Virgil
Type of Composite

WWW")

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Neural manure-eel
120
IWEqumw m'ﬂmwgﬂbrhdljmw
100- I momma arm I “limb“
5‘57:
”_ 200~ ......
I ..: :::
”I. ‘”b .......
40- 100» ’:::i_.
20. “L ...... .,
75.: E
o . .. o
Recycled my. Recycled Vlrgln
Type of Compoalle Wpe oi Composite

c Toughness (1 Initial Stiffness

Figure 5.6 (Cont’d.) Moisture Effects on Flexural Behavior of Recycled and Virgin
Wood F iber-Cement Composites.

129

5.3.2 Repeated Wetting and Drying

Figure 5.7a presents the effects of repeated wetting-drying cycles on the ﬂexural load-de-
ﬂection behavior of recycled and virgin cellulose ﬁber-cement composites. The effects of
repeated wet-dry cycles on ﬂexural strength, stiffness and toughness of recycled and virgin
composites are presented in Table 5.3 and Figures 5.2 b-d. In both cases of virgin and re-
cycled composites it is observed that repeated wetting and drying leads to increase in ﬂex-
ural strength and initial stiffness whereas the toughness is reduced substantially and failure

occurs in a brittle mode.

Results of the statistical analysis of the test data Table 5.3 are presented in Table 5.4. Re-
peated wetting~drying cycles had statistically signiﬁcant effects on all ﬂexural properties
(strength, toughness and stiffness) of recycled composites. Two-way analysis of variance
(with two factors of recycled versus virgin ﬁbers and aged versus unaged composites) re-
vealed that there was a statistically signiﬁcant difference, at 95% level of conﬁdence, be-
tween ageing effects on recycled and virgin composites as far as ﬂexural strength and
toughness are concerned; the initial stiﬂ‘ness of virgin and recycled composites, however,
effected similarly by ageing under repeated wetting-drying cycles. Damaging effects of re-
peated wet-dry cycles on recycled composites (which lost 32% of toughness upon ageing

was less than that on virgin composites (which lost 45% of toughness upon ageing).

 

 

Table 5.3 Repeated Wetting-Drying Effects on the Flexural Performance of
Recycled and Virgin Fiber-Cement Composites
Experiment Flex. Str. Mean Flex. Tou. Mean Init. Stif. Mean Init
(MPa) Flex. Str N-mm Tou. N/mm Stif.
(9593 (95% (95%
Conf. Int.) Conf. Int.) Conf. Int.)
Control 15.70 15.71 39.99 37.32 228.3 229.21
Aged 15.66 (10.5) 40.11 (17.39) 235.4 (119.6)
15.55 35.76 227.9
15.60 36.77 228.2
15.67 37.87 220.0
15.67 37.99 226.5
15.63 35.87 233.5
15.87 36.44 235.1
15.98 40.10 223.9
15.76 35.32 232.8

 

 

 

 

 

 

 

 

 

130

 

 

 

 

Recycled 15.101 15.04 38.11 40.241 249.09 245.9
Aged 14.992 (10.22) 39.02 (16.6) 240.98 (115.4)

15.002 37.97 240.65

14.987 40.23 240.98

14.987 43.23 245.67

14.994 41.23 243.98

14.998 40.23 249.87

15.105 39.19 248.88

15.112 41.22 249.92

15.006 41.98 248.99
Control 15.1 15.09 67 67.3 164.4 161.9

15.02 (10.51) 66.39 (18.3) 154.3 (123.3)
Unaged 15.43 63.12 171.6

15.1 69.34 167.4

14.99 70.74 152.2

14.98 69.43 162.4

15. 11 66.98 159.9

15.09 66.12 157.7

14.96 67.38 162.9

15.11 66.55 165.7
Recycled 14.9 14.85 58 59. 1 1 180.2 184.1
Unaged 14.88 (10.23) 57.23 (14.0) 178.2 (117.1)

14.77 59.98 180.9

14.92 60.11 189.2

14.89 58.23 187.7

14.81 59.12 185.9

14.75 58.87 186.7

14.92 59.98 186.6

14.81 60.34 187.9

14.86 59.22 177.7

 

 

 

 

 

 

 

 

131

Table 5.4 Results of Analysis of Variance (ﬂexural strength, toughness and stiffness)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source I Sum-of Sq. I DF I Mean-Sq. I F-Ratio I P
One Way Analysis of Variance

Flexural Strength

Ageing(wd) I 0.072 1 I 0.072 I 11.59 I0.009

Error I 0.05 3 I 0.006 I

Toughness

Ageing(wd) I 973.13 I 1 I 973.13 I 659.53 I 0.00

Error I 11.8 I 8 I 1.475 I

Initial Stiffness

Ageing(wd) I 9773.1 I 1 I 9773.1 R901 I 0.00

Error I 37.33 I3 I 10.93 I

Two way Analysis of Variance

Flexural Strength

Fiber 1.10 1 -1.10 91.92 0.00

Ageing(wd) 0.77 1 0.77 64.55 0.00

Fiber *Age. 0.20 1 0.20 17.17 0.00

Error 0.19 16 0.01

Toughness

Fiber 44.01 1 44.01 10.68 0.01

Ageing(wd) 3141.27 1 3141.27 762.25 0.00

Fiber*Age 170.47 1 170.47 0.00

Error 65.93 16 4.12

Initial Stiffness

Fiber 959.39 1 959.39 64.02 0.00

Ageing(wd) 20147.2 1 20147.2 1344.52 0.00

Fiber‘Age. 0.51 1 0.51 0.03 0.85

Error 239.756 16 14.98

 

 

 

132

Heart-l Str-ea (MPa)

 

 

 

 

 

I 331cm...
“mi
15 1‘
I —m
‘0 “imaged
e
A 1 ‘\
° 4455 '.0511522533.544.63
o 05 l is a 23 :1 as mnm'onmm)
Tv‘ndontmm‘
Recycled a Load Deﬂection Virgin

'Ie-l ”an“ m.)
2° 'Tmigro

I I“. Conﬁdence int-war
l w-

 
   

WPO 0! Compoilte

b Flexural Strength

 

typ- or c M» of Compoelte
empoeite

c Toughncss (.1 Initial SIIITIICSS

Figure 5.7. Effect of Repeated Wetting and Drying Cycles on Flexural Behavior.

133

5.3.3 Repeated Freeze-Thaw

The recycled and virgin cellulose ﬁber-cement composites were tested for ﬂexural perfor-
mance before and after exposure to freeze-thaw cycles. The eﬂ'ects of repeated freeze-
thaw cycles on ﬂexural load-deﬂection behavior are presented in Figure 5.10a. Table 5.5
and Figures 5.8 bod show the freeze-thaw eﬂ'ects on ﬂexural strength, stiffness and tough-
ness of virgin and recycled composites. No delamination was observed in the specimens
subjected to repeated freeze-thaw cycles; the aged specimens were intact and marginally
stiffer than unaged optimized composites.

Results of the analysis of variance of test data are presented in Table 5.6. It was concluded
that freeze thaw ageing did not have statistically signiﬁcant eﬂ'ect on ﬂexural strength and
toughness of recycled composite; however, the effect on initial stiffness was statistically
signiﬁcant at 95% level of conﬁdence. Two way analysis of variance (taking into account
the virgin composite) revealed that freeze-thaw cycles aﬂ'ect recycled composites in a way
different from virgin composites (similar effects were observed on ﬂexural strength and
toughness irrespective of the composite type). While recycled composites showed a 1%
increase in stiffness under repeated freeze-thaw cycles, virgin composites showed a drop

of 9% in stiffness under this ageing condition.

 

 

Table 5.5 Repeated Freeze-Thaw Effects on the Flexural Performance of
Recycled and Virgin Fiber-Cement Composites.
Experiment Flex. Str. Mean Flex. Tou. Mean Init. Stif. Mean Init
(MPa) Flex. Str N-rnrn Tou. N/mm Stif.
(95% (95% (95%
CoannLI Conf. Int.) Conf. Int.)
Control 15.00 14.97 67.334 66.88 170.32 168.68
Aged 14.99 (10.29) 67.34 (13.27) 170.11 (14.71)
15.11 68.21 168.84
15.01 68.43 167.32
14.95 65.98 166.43
14.99 66.32 167.79
14.93 66.87 169.32
14.82 65.54 168.76
14.99 66.93 169
14.91 67.87 168.87

 

 

 

 

 

 

 

 

 

 

 

 

 

Control 15.1 15.09 67 67.3 164.4 161.9
Unaged 15.02 (11.1) 66.39 (13.2 154.3 (12.9)

15.43 63.12 171.6

15.1 69.34 167.4

14.99 70.74 152.2

14.98 69.43 162.4

15. l 1 66.98 159.9

15.09 66.12 157.7

14.96 67.38 162.9

15.11 66.55 165.7
Recycled 14.80 14.79 59.08 58. 177 187.56 187.04
Aged 14.79 @006) 59.10 (13.74) 186.65 (13.89)

14.7 58.09 187.54

14.79 57.09 187.43

14.78 56.12 185.55

14.80 57.88 185.39

14.79 58.5 187.6

14.79 58.39 186.53

14.83 59.07 188.54

14.81 58.34 187.54
Recycled 14.9 14.85 58 59.11 180.2 184.1
Unaged 14.88 (10.23) 57.23 (14.0) 178.2 (117.1)

14.77 59.98 180.9

14.92 60.11 189.2

14. 89 58.23 187.7

14.81 59.12 185.9

14.75 58.87 186.7

14.92 59.98 186.6

14.81 60.34 187.9

14.86 59.22 177.7

 

 

 

 

 

 

 

 

135

Table 5.6 Results of Analysis of Variance (ﬂexural strength, toughness and stiffness)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source I Sum-of Sq. IDF Mean-Sq. I F-Ratio I P

One Way Analysis of Variance
Flexural Strength
Ageing(F'I') I 0.01 I 1 0.01 I 10.5 I 0.012
Error I 0.003 I 3 0.001 I I
Toughness
Ageing(FI') 2.246 I 1 [2.246 I 1.13 I 0.317
Error 15.733 I 3 I 1.974 I I
Initial Stiffness
Ageing(FI') I 3.31 I1 I 3.31 I 7.66 I 0.02
Error I 3.63 I 3 I 1.03 I I

Two way Analysis of Variance

Flexural Strength
Fiber 0.18 1 0.18 56.32 0.00
Ageing(FI') 0.01 1 0.01 2.31 0.11
Fiber ‘Age. 0.00 1 0.00 0.023 0.37
Error 0.051 16 0.003
Toughness
Fiber 308.03 1 308.03 278.7 0.00
Ageing(FT) 3.72 1 3.72 3.33 0.03
Fiber*Age 0.06 1 0.06 0.054 0.319
Error 17.62 16 1.101
Initial Stiffness
Fiber 2296.3 1 2296.3 1053.89 0.00
Ageing(PT) 119.786 1 119.78 54.97 0.00
Fiber'Age. 30.71 1 30.71 14.09 0.002
Error 34.863 16 2. 179

 

 

 

136

Rental Sven (MPa) Mel emu (MPa)

 

 

3 Load Deﬂection 5

  

 

 

 

 

o ‘ , ‘
0 0.5 I 1.5 2 2.5 3 3 5 0 I 5 5 n E ‘ I 15 2 2.5 3 3.5 0 4.5 5
Deﬂection (mm) Oefledron ("1'")

ﬂea-d "so, (Illa)
20 he _

A'Unaqed -_.‘ Iged

 

Npe of Composite

b Flexural Strength

      

muesli We. (ll-n)
300
Unaged Aged
35° I use Comm-mi I
200
m
100
ID
A -
lecycled "I."
‘Iype of Composite WP. 0' Composite
c Toughness d Initial Stiffness

Figure 5.8 Effects of Repeated F reeze-thaw Cycles on Flexural Behavior.

137

5.3.4 Repeated Wetting-Drying and Carbonation

This accelerated condition seems to best simulate natural weathering eﬂ'ects on cellulose
ﬁber-cement composites [32]. The eﬂ’ect of repeated wetting-drying and carbonation on
ﬂexural load-deﬂection behavior is presented in Figure 5.9a. This ageing process is ob-
served to increase ﬂexural strength and initial stiﬂ'ness while causing embrittlement of the

composite.

The test data of Table 5.7 was analyzed statistically (Table 5.8). Wetting-drying and car-
bonation ageing had (in one way analysis of variance) statistically signiﬁcant effect on
ﬂexural strength, toughness and stiffness. In addition, the ﬁber‘ageing interactions were
statistically signiﬁcant at 95% level of conﬁdence for all the three ﬂexural responses
(strength, toughness and stiﬂ'ness), indicating that this ageing process aﬂ'ects recycled and
virgin composites differently. The damaging effects of repeated wetting-drying and car-
bonation on recycled composites (47% loss) was less pronounced than for virgin compos-
ites (54% drop).

 

 

Table 5.7 Effects of Repeated Wetting-Drying and Carbonation on the
Flexural Performance of Recycled and Virgin Fiber-Cement
Composites
Experiment Flex. Str. Mean Flex. Tou. Mean Init. Stif. Mean Init
(MPa) Flex. Str (N-rnm) Tou. (N/mm) Stif.
(95% (9596 (9593
Conf. Int.) Conf. Int) Conf.
1111.).
Control 16.11 16.13 32.12 31.05 273.23 270.40
Aged 16.20 (10.93) 27.45 (18.12) 260.56 (123.2)
16.19 28.54 280.54
16.20 31.99 268.67
16.15 29.09 267.87
16.2 33.34 270.98
16.25 32.87 278.78
16.54 31.22 265.54
15.77 30.00 267.31
15.712 33.87 270.54

 

 

 

 

 

 

 

 

 

138

 

 

 

 

Recycled. 15.62 15.65 33.19 31.322 270.12 270.91
Aged 15.55 (10.29) 32.76 (19.48) 265.87 (117.1)

15.76 29.67 275.98

15.55 34.76 272.65

15.75 32.98 269.98

15.66 31.87 267.77

15.61 32.87 263.87

15.63 28.00 275.65

15.71 27.98 276.76

15.67 29.1 1 270.41
Control 15.1 15.09 67 67.3 164.4 161.9
Unaged 15.02 (10.51) 66.39 (18.3) 154.3 (123.3)

15.43 63.12 171.6

15.1 69.34 167.4

14.99 70.74 152.2

14.98 69.43 162.4

15. 1 1 66.98 159.9

15.09 66.12 157.7

14.96 67.38 162.9

15.11 66.55 165.7
Recycled 14.9 14.85 58 59. 1 1 180.2 184.1
Unaged 14.88 (10.23) 57.23 (14.0) 178.2 (117.1)

14.77 59.98 ' 180.9

14.92 60.11 189.2

14. 89 58.23 187.7

14.81 59.12 185.9

14.75 58.87 186.7

14.92 59.98 186.6

14.81 60.34 187.9

14. 86 59.22 177.7

 

 

 

 

 

 

 

 

139

Table 5.3 Results of Analysis of Variance (ﬂexural strength, toughness and stiffness)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source I Sum-of Sq. I DF I Mean-Sq. I F-Ratio I P

One Way Analysis of Variance
Flexural Strength
Age.(WDC) [2.134 I 1 [2.134 I33013 I000
Error I 0.045 I 3 I 0.006 I
Toughness
Age.(WDC) 2013.19 I 1 I 2013.19 I 524.99 I 0.00
Error 30.75 I 3 I 3.34 I
Initial Stiffness
Age.(WDC) [19330 I 1 I 19330 I615 I000
Error I 252.03 I 3 I 31.51 I

Two way Analysis of Variance

Flexural Strength
Fiber 0.702 1 0.702 229.72 0.00
Age.(WDC) 4.5 1 4.5 1476 0.00
Fiber *Age. 0.131 1 0.131 42.75 0.00
Error 0.049 16 0.003
Toughness
Fiber 53.694 1 53.694 13.153 0.002
Age.(WDC) 5043 1 5043 1237 0
Fiber*Age 73.23 1 73.23 19.13 0.00
Error 65.23 16 4.031
Initial Stiﬂ‘ness
Fiber 265.64 1 265.64 10.03 0.006
Age.(WDC) 47321 1 47321 1315 0.00
Fiber*Age. 336.1 1 336.1 12.76 0.003
Error 421.375 16 26.336'

 

 

 

140

Flaunt Streaa (MPa)
20

a Load Deﬂection

      

o 1
00.51152258154455

0 0.5 l 1.5 2 2.5 3 3.5 4 4.5 5
Dela-enliven) Mimi
Virgin Recycled

lie-vat been (M's)
zo-r ~ . . 7~ ~ 7
'Unaged :Aged

— S's Connoeeca menial

'5‘ ' Tri I I
i i l
b Flexural Strength ,0, ,1, I j. I
I “'1 1
. 4,1 1
. I 71 i I
:34 ’1 ‘
,A ;j , AA
leeyeled

 

 

 

 

 

Recycled min 1.
Type of Composlte 0' “”90""

c Toughness d Initial Stiffness

Figure 5.9 Effects of Repeated Wetting-Drying and Carbonation on Flexural Behav-
ior.

141

5.3.5 Extended Immersion in Hot Water Bath

Long-term immersion in hot water bath had relatively small eﬂ'ects on the ﬂexural perfor-
mance of virgin and recycled ﬁber-cement composites; this ageing process resulted in
some embrittlement of the ﬂexural behavior (Figure 5.10 and Table 5.9).

Results of the analysis of variance of test data are presented in Table 5.10. The effect of
long-term immersion in hot water on ﬂexural strength, stiﬂ'ness and toughness of recycled
composites was not statistically signiﬁcant at 95% level of conﬁdence. Two-way analysis
of variance of results (virgin and recycled ﬁbers) concluded that only ﬂexural stiffness of
virgin and recycled composites was affected differently by this ageing process. While hot
water immersion slightly reduced the initial stiffness of recycled composites (by 3%) it
caused a small increase of 3% in the stiffness of virgin composites.

 

 

Table 5.9 Effects of Hot Water Immersion on the Flexural Performance of
Recycled and Virgin Fiber-Cement Composites
Experiment Flex. Str. Mean Flex. Tou. Mean Init. Stif. Mean Init
(MPa) Flex. Str (bl—mm) Tou. (Nlmm) Stif.
(9593 - (9596 (95%
Conf. Con. Int.) Con. Int.)
Int.)I
Control 15.07 15.09 61.13 63.43 165.43 168.3
100% virgin 14.98 (10.28) 62.45 (16.93) 166.76 (112.3)
Aged 15.06 67.32 170.54
15.12 63.33 162.34
15.11 62.98 171.11
15.09 63.54 169.87
15.1 65.45 168.43
15.09 63.22 165.67
15.16 61.98 172.09
15.23 62.87 170.72

 

 

 

 

 

 

 

 

 

 

 

 

 

Control 15.1 15.09 67 67.3 164.4 161.9

15.02 (10.51) 66.39 (13.3) 154.3 (123.3)
Unaged 15.43 63.12 171.6

15.1 69.34 167.4

14.99 70.74 152.2

14.93 69.43 162.4

15.11 66.98 159.9

15.09 66.12 157.7

14.96 67.33 162.9

15.11 66.55 165.7
Recycled 14.30 14.31 63.32 59.59 134.32 173.93
Aged 14.77 (1015) 54.43 (114) 133.33 (113.57

14.79 55.54 173.93 )

14.30 57.12 130.56

14.32 55.23 175.43

14.79 61.09 172.37

14.31 62.23 176.37

14.30 63.32 173.45

14.92 62.54 177.99

14.73 62.21 130.54
Recycled 14.9 14.35 53 59.11 130.2 134.1
Unaged 14.33 (10.23) 57.23 (14.0) 173.2 (117.1)

14.77 59.93 130.9

14.92 60.11 139.2

14.39 53.23 137.7

14.31 59.12 135.9

14.75 53.37 136.7

14.92 59.93 136.6

14.31 60.34 137.9

14.36 59.22 177.7

 

 

 

 

 

 

 

 

143

Table 5.10 Results of Analysis of Variance (ﬂexural strength, toughness and stiﬂneas)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source I Sumcf Sq. DF I Mean-Sq. IF-Ratio I P

One Way Analysis of Variance
Flexural Strength
Age.(HW) I 0.00 1 I 0.00 I 0.097 I 0.764
Error I 0.021 8 I 0.003
Toughness
Age.(HW) I 1.421 I 1 I 1.421 I 0.103 I 0.751
Error I 105.51 I 3 I 13.139
Initial Stiffness
Age.(HW) I 1.133 I 1 I 1.133 I 0.11 I 0.743
Error I 35.712 I 3 I 10.714

Two way Analysis of Variance

Flexural Strength
Fiber 0.27 1 0.27 112.746 0.00
Age.(HW) 0.00 1 0.00 0.042 0.84
Fiber *Age. 0.00 l 0.00 0.117 0.736
Error 0.038 16 0.002
Toughness
Fiber 171.4 1 171.4 27.15 0.00
Age.(HW) 34.03 1 34.03 5.39 0.034
Fiber‘Age 0.286 1 0.286 0.045 0.834
Error 100.99 12 6.312
Initial Stiffness
Fiber 1540.89 1 1540.89 91.45 0.00
Age.(HW) 86.32 1 86.32 5.12 0.038
Fiber’Age. 128.37 1 128.37 7.619 0.014
Error 269.579 16 16.849

 

 

 

 

     

 

 

   

 

Flexural $reas (MPa)
”m l
I an cum I
I.“ ‘5 [I
is I ‘.
to [32“ a Load Deﬂection ‘° :
5 .
5
i n -
o ' "00511522533544.5151
o n: r 15 2 25 3 35 4 95 5
_ Deﬂection (mm)
ce‘bU-cr 01'
Vir '
Recycled gin
Iran-at helium (Illa)
1° 'IJnaped E“. Aged
I I In Cam 1mm
'5' J7 :— 1
- 591
. l :4:
to» f :' '1 5:
/1 3?:
a- I 1 ‘
(b) ﬁexural Strength I . I
.97 ’1 :i‘
7 ,1 ' (1
o A i A ‘ .17,
Recycled Virgil
Type oi Composite
Flannel heaviness Ola-l Flea-d sen-r.“ (ll/n.)
.. 1 see
""1 .unaged E34994 . Uri-god Aged
iv 2» r
.0 1. HS Coaloanca M 8‘ Most-ca Interval
see
so -
150
on
I toeI
.. . ”I
l I
I I .
o Recycled Virgin Recycled —'
Type of Composite

Virgil
Typo oi Composite

c Toughness d Initial Stiffness

Figure 5.10 Effects of Hot Water Bath Immersion on Flexural Behavior.

145

5.3.6 Comparison of Different Ageing Effects

In order to compare the effects of different accelerated ageing conditions on the ﬂexural
performance of recycled and virgin ﬁber-cement composites (see Figure 5.11), two-way
analysis of variance of ratios of aged to unaged ﬂexural strength, stiﬁ'ness and toughness
test results were conducted. The two factors in these analyses were: composite type (recy-
cled and virgin), and accelerated ageing condition (wet-dry, wet-dry and carbonation,
freeze-thaw, hot water immersion). Results of the analysis of variance are presented in Ta-
ble 5.11. the global analysis of variance suggested that, at 95% level of conﬁdence, all the
ageing effects have comparable effects on flexural strength but there are statistically sig-
niﬁcant differences between the effects of different accelerated ageing conditions on ﬂex-
ural stiffness and toughness. There was no statistically signiﬁcant diﬁ‘erence between
virgin and recycled composites as far as ageing effects on ﬂexural strength and stiffness
are concerned; the two composites, however, behaved differently in ageing effects on ﬂex-

ural toughness.

Multiple comparison of results indicated, at 95% level of conﬁdence, that each of the ac-
celerated conditions have distinctly different effect on ﬂexural toughness and stiffness, ex-
cept for the hot water immersion and freeze-thaw condition which had statistically

comparable effects on flexural stiffness.

As far as the overall ageing effects on ﬂexural performance is concerned, repeated wet-
ting-drying and carbonation cycles produced the most pronounced effects whereas hot wa-
ter immersion caused the least effects. The trends in the response of virgin and recycled

composites to ageing were generally comparable.

146

Table 5.11 Results of Analysis of Variance of Different Ageing Methods
(see Appendix III for notations)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source I Sum.of Sq. 7 DP Mean-Sq. F—Ratio I P
One Way Analysis of Variance

Flexural Strength

Age. meth- 0.009 3 0.003 2.189 0.167
ods

Error 0.01 l 8 0.001

WD VS 0.002 1 0.002 1.494 0.256
WDC

Error 0.011 8 0.011

WDC VS 0.002 1 0.002 1.235 0.299
HW

Error 0.011 8 0.001

Toughness

Age. meth- 0.48 3 0.16 548.52 0.00
ods

Error 0.002 8 0.00

WD Vs FI‘ 0.144 I 1 I 0.144 I 494.22 I000
Error 0.002 I 8 I 0.00 I I
WD WDC 0.035 1 0.035 120.914 0.00
Error 0.002 8 0.00
WD Vs HW 0.141 1 0.141 483.657 I 0.00

 

 

 

 

 

147

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Error 0.002 8 0.00
Stiffness
Age. Meth- 0.534 3 0.178 296.83 0.00
ods
Error 0.005 8 0.001
WDC Vs 0.375 1 0.375 625 0.00
HW
Error 8 0.001
FT Vs HW 0.00 1 0.00 0.25 0.63
Error 0.005 8 0.001
Aged/Unaged Ratio (Flexural Strength)
2
.Rec.
I 95% Confidence Interval
[ZJvm
1.5
j .—
0.5 ~ - %
o // //
WD FT WDC HW

Flexural Strength

Ageing Methods

Figure 5.1] Comparison of Different Ageing Methods

 

148

2 Aged/Unaged Ratio (Flexural Toughness)

 

I 95% Confidence Interval

 

., “/M’ 4.

WD FT
Ageing Methods

\

 

 

 

 

 

 

Toughness

Aged/Unaged Ratio (Initial Stiffness)

 

 

 

  

 

 

 

 

 

 

537' ...... . mm]
1 _ é /
o-s- . Z / .........

Ageing Methods

Stiffness

Figure 5.1] (Cont’d.) Comparison of Different Ageing Methods

149

5.3.7 Permeability

The permeability coefﬁcient was measured in three replications for the optimized recycled
and virgin cellulose ﬁber-cement composites, and also for the cement-based matrix with
no ﬁbers all in unaged and aged (after repeated wetting-drying and carbonation) condi-
tions. The permeability test results are presented in Table 5.12 and Figure 5.12. The recy-
cled composite is observed to have a lower permeability coefﬁcient (24* 10 '11 cm/sec, 9 *
10 '11 in/sec) than virgin composites (36"10‘11 cm/sec, l4 * 10 '11 in/sec) in unaged con-
dition. This further conﬁrms the better consolidation of recycled composites (where the
ﬁnes in ﬁber play a ﬁller role and, unlike ﬁbers, facilitate consolidation) reﬂected in the
higher density and lower water absorption of recycled composites (when compared with
virgin ones). Fibrous composites, however, had permeability coefﬁcient higher than the

plain cement-based matrix (6*10‘“ cm/sec, 2 4 10 '11 ill/sec).

Table 5.12 Permeability Coefﬁcient of Virgin and Recycled Composites

 

 

 

 

 

Type of Composite Permeability Coef- Mean (95% Conﬁ-
ﬁcient deuce Interval)
( 10*"11 cm/sec)
Control (Unaged) 366,387,353 36.86 Q6.72)
Recycled (Unaged) 22.0,24.5,25.0 23.83 ($6.30)
Control (Aged) 330,267,400 33.23 (6.65)
Recycled (Aged) 19.2,18.7,19 18.96 (10.98)

 

 

 

 

 

One way analysis of variance was canied out on permeability test results (Table 5.13).
Permeability coefﬁcients of different composites found to be statistically diﬂ'erent at 95%
level of conﬁdence; recycled composites had a lower permeability than virgin composites.
The effect of ageing on reducing the permeability coefﬁcient was also statistically signiﬁ-
cant. This observation could be attributed to the densiﬁcation eﬁ'ects of ageing of the ﬁber

cores and interfaces.

Two-way analysis of variance with composite type (virgin versus recycled) and age (un-

aged versus aged) the two factors indicated that while recycled and control composite had

 

 

 

150

different permeability at 99% level of conﬁdence, ageing effects on permeability was sig-
niﬁcant at 95% level of conﬁdence. The recycled and virgin composites were also ob-

served to perform similarly as far as the ageing effects on permeability is concerned.

Table 5.13 Results of the Analysis of Variance of Permeability Coefﬁcients

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source Sumoof-Sq. DF I Mean-Sq. I F-Ratio P
Effect of Type (recycled, control)

type of 1463.48 2 731.74 366.48 0.00
Composite

Error 1 1.98 6 1.997

Effect of Ageing

type of . 1907.9 5 381.5 45.2 0.00
composrte

Error 12 101.24 12 8.43

Two way analysis of variance

Type 557.849 1 P 557.849 44.7 0.00
Age 54.53 1 54.53 4.37 0.07
Type*Age 1.19 l 1.19 0.09 0.765
Error 99.82 8 12.47

 

 

151

Coefficient of Permeability (10“-11 cm/sec)
40

 

l 95% Confidence interval

 

 

 

Vir.(Unag.) Rec.(Unag.) Vir.(Ag.) Rec.(Ag.) .
Type of Composite

Figure 5.12 Water Permeability Coefﬁcient Test Results.

5.4 MICROSTRUCTURAL AND COMPOSITIONAL CHANGES UNDER MOIS-
TURE AND AGEING EFFECTS

5.4.1 SEM Observations:

The Scanning Electron Micrographs of the fracture surfaces of unaged and aged compos-
ites are presented in Figure 5.12 a through 5.12 e. For the unaged composite the dominant
mode of failure is observed to be ﬁber pull out (see Figure 5.12a). In the case of specimens
subjected to repeated cycles of freeze-thaw, the fracture surface is observed to have a com-
bination of ﬁber pull out and ﬁber fracture (see Figure 5.12b). In the case of hot water im-
mersed composites (Figure 12c) also the mode of failure is observed as ﬁber pull out
accompanied with ﬁber fracture. For the repeated wetting-drying test the mode of failure
at the fracture surface is dominated by ﬁber fracture (see Figure 5.12d). Also, for repeated
wetting-drying and carbonation ﬁber rupture is observed to dominate the failure mode (see

Figure 5.126).

 

152

In addition. the appearance of ﬁbers and their interface is also affected by ageing effects.
In the case of wetting-drying and wetting drying and carbonation, a dense ﬁber matrix in-
terface is observed. the ﬁbers in case of wetting-drying with carbonation appear to be
ﬁlled with some hydration and carbonation products. In the case of hot water bath and
freeze-thaw ageing tests the densiﬁcation of the interface and ﬁlling of ﬁber core is not
pronounced. It appears that the ageing process is most pronounced under repeated wet-

ting-drying and carbonation.

 

a. Unaged

Figure 5.13 Scanning Electmn Micrographs of Fracture Surfaces Under Various
Accelerated ageing Conditions.

 

c. After Repeated Freeze—Thaw Cycles

Figure 5.13 Scanning Electmn Micrographs of Fracture Surfaces Under Various
Accelerated ageing Conditions.

, .115
/’ .;- w"-

l":-

' ﬂ 04
ft‘sxu,v'/ 141,59 4" 1 18°F“! eeee

 

C. After Hot Water Bath

Figure 5.13 (Cont’d) Scanning Electron Micrographs of Fracture Surfaces Under
Various Accelerated ageing Conditions.

155

5.4.2 Thermogravimetric Analysis

Thermogravimetry is a technique in which the mass of a substance is monitored as a func-
tion of temperature or time as the sample specimen is subjected to a controlled tempera-
ture history [80]. This relationship was used in this investigation to determine the relative
quantities of calcium hydroxide and calcium carbonate in the aged and unaged composite,
which are expected to correlate to the ageing process of wood ﬁber-cement composites.
Mixing of cement with water results in a complex hydration process. The hardening pro-
cess over a long period is principally associated with the hydration of the silicate phases.
Although the reaction rates of these phases are very diﬁ'erent, the ﬁnal products are the
same. In a mature paste, approximately 70% by volume is taken by a colloidal silicate hy-
drate (CSH). Calcium hydroxide (CH) comprises about 20% of volume [81].

TGA weight loss curves of ﬁber-cement composites can be interpreted for compositional
analysis. Dehydration occurs over the range 105 to 440° C (221 °F to 824 °F), followed by
dehydroxylation effecting calcium hydroxide in the range 440 to 580° C (824 ° F to 1076
0 F) with calcium carbonate dissociation occurring in the region 580 to 10000 C (1076 o F
to 1832 °F). Amounts of calcium hydroxide and calcium carbonate can thus be computed
based on weight changes at these temperature changes. Free calcium hydroxide can be cal-
culated as follows [81].

Free calcium hydroxide = 4.11*(de) + 1.68“' (de)

where

de:% weight loss within 440 °C to 580 ° C (824 ° F to 1076 0 F)
de:% weight loss within 580 °C to 1000 °C (1076 0 F to 1832 °F)
The calcium carbonate content can be calculated as follows:

Calcium carbonate = Weight loss from 580 °C to 1000 °C (1076 0 F to 1832 °F).

Typical weight loss curves are presented in Figure 5.18; the calcium hydroxide and calci-

um carbonate contents may be calculated from these curves.

The amounts of free calcium hydroxide and calcium carbonate for unaged and various

156

aged composites are given in Table 5.14.

Table 5.14: Thermogravimetric Compositional Analysis

 

 

 

 

 

 

 

 

 

 

 

 

Type of
Ageing . 4 Mean (95% . (. Mean (95%
Treatment CH (Werght‘ic) Con. Int.) CC (Weight/o Con. Int.)

Unaged 22.21.229.219 22.29 ($2.33) 3.12, 3.05, 3.37 3.18 ($0.53)
Aged (WD) 17.7. 18.23.17.8 17.94 ($1.05) 7.94,8.04,7.78 7.92 ($0.42)
Aged (WDC) 18.9,18.45,l8.11 18.49 ($1.55) 838,812,847 8.32 ($0.58)
Aged (FT) 21.03.20.96.21.l 21.04 ($0.35) 4.33, 4.23, 4.41 4.32 ($0.28)
Aged (HW) 17.03.16.67,l6.9 16.89 ($0.76) 5.47, 5.39, 5.54 5.46 ($0.24)

 

CC: Calcium Carbonate; CH Calcium Hydroxide

Wbmmtnu

100

 

90*

80-

 

— Aged Rec.

'“ Unaged Rec.

L

 

 

 

7O
0

200 400

Temperature (deg. C)

600 000

1000 1200

Figure 5.14 Thermogravimetric Analysis: Typical Weight Loss Curves.

 

157

Different ageing processes are observed in Table 5.14 to cause a drop in calcium hydrox-
ide content and an increase in the amount of calcium carbonate. The increase in calcium
carbonate under repeated wetting-drying and wetting-drying with carbonation is particu-
larly pronounced. The results suggest that recycled ﬁber reinforced cement composites
tend to carbonate under weathering eﬁ'ects. The dominant mode of failure in the unaged
composite was ﬁber pull-out when the calcium carbonate content was low. After accelerat-
ed ageing, the mode of failure was brittle with ﬁber rupture dominating the behavior, and
the content of calcium carbonate was also increased. It is suggested that the increase in
strength and rigidity of the petriﬁed ﬁbers, and the increase in their bond strength due to
matrix densiﬁcation and to elimination of shrinkage debonding at the interface zones
(these phenomena involve dissolution, migration and are pronounced under carbonation

effects), can account for the increase in strength and stiﬁ‘ness of the composite.

In order to establish the correlation of calcium carbonate and calcium hydroxide contents
with engineering properties of composites, some correlation analyses were conducted the
results of which are presented in Figure 5.15. These analyses were conducted in order to
conﬁrm the key role calcium hydroxide and formation of calcium carbonate play in the
ageing of composites. Both unaged and aged (under diﬁerent accelerated tests) compos-
ites were used in the development of Figure 5.15.

Figure 5.15 a shows a positive correlation between calcium carbonate and ﬁexural
strength; however, this correlation is not signiﬁcant at 95% conﬁdence level. But the cor-
relation of calcium carbonate content with ﬁexural toughness is strongly negative (Figure
5.15 b) and is statistically signiﬁcant at 95% conﬁdence level (correlation coefﬁcient: -
0.93) Calcium carbonate content shows strong positive conelation with ﬁexural stiffness
(Figure 5.15 c) at 95% conﬁdence level (correlation coefﬁcient 0.914). This tendency of
increased stiffness with increased calcium carbonate content was observed consistently in

aged composites.

The correlations of calcium hydroxide content with engineering properties (strength, stiﬂ-
ness and toughness) were not statistically signiﬁcant (Figure 5.15 d-f). A strong positive
correlation was observed between calcium carbonate content and density (correlation co-
efﬁcient: 0.941) which was statistically signiﬁcant at 99% conﬁdence level (see Figure
5.15 g). Correlation of calcium hydroxide content with density is presented in Figure 5.15
h.

158

Calcium carbonate content seems to have a strong correlation than calcium hydroxide
content with engineering properties. This may be attributed to the fact that ageing causes
dominantly a migration of calcium hydroxide (and only a small change is its content as a
result of carbonation) while it actually forms calcium carbonate (which reﬂects in in-

creased calcium carbonate content.

Math-an“

 

Caner-ammo”!
magma-mama:

 

 

to 1 1
° ‘ 0 1:

Calcium Carbonate (height $)

 

a. Correlation of Calcium Carbonate Content with Flexural Strength

MW”

 

 

 

 

 

100
MGM-4m
Moravian-“ICU

00%

go»

40>

20 ' 1

O 4 I 12

Calcium Carbonate (welght %)

b. Correlation of Calcium Carbonate Content with Flexural Toughness

~

Figure 5.15 Correlation of Composition with Engineering Properties.

159

 

 

 

 

 

 

 

c. Correlation of Calcium Carbonate Content with Flexural Stiffness

 

 

 

l
10‘ ‘
1' 3o 29

cucm Hydmm (wow *1

(1. Correlation of Calcium Hydroxide Content with Flexural Strength

Figure 5.15 (Cont’d.) Correlation of Composition with Engineering Properties.

160

 

 

 

 

”I
Camden Coulee-n - one
Meet-glean)
.01-
“lh '
2° .
t! 20 28

Caicium Hydroxide (mm as)

e. Correlation of Calcium Hydroxide Content with Flexural toughness

 

mm 48
macaw-mega,

250'-

'\

 

 

 

130 1
is 20 23

Calcium Hydroxide (weight %)

f. Correlation of Calcium Hydroxide Content with Flexural Stiffness

Figure 5.15 (Cont’d) Correlation of Composition with Ageing Properties.

161

     

1’30 us I.“ 1.0

0”“‘7 (out Icu can)

I.“

g. Correlation of Calcium Carbonate Content with Density

 

 

 

“mm”

2.

so mun-an

23
n a
ti
'0‘ 1 r r
I.” ‘,“ ‘.“ ‘.‘ I.“

0‘0“" (9M letr cm)

h. Correlation of Calcium Hydroxide Content with Density

Figure 5.15 (Cont’d) Correlation of Composition with Engineering Properties.

162

The process of ageing seems to increase density. ﬂexural stiffness and strength while re-
ducing toughness and permeability. The fact that these changes in engineering properties
all have their roots in structural changes of composite upon ageing is reﬂected in strong
(negative or positive) correlations between these qualities (see Figure 5.16) It should be

noted that unaged and various aged composites were used to develop Figure 5.16.

hue-pm

17
1' emu-acme... l
I MWImCLJ

 

101'

 

 

‘3 § 1 A r ‘1
2° ‘° 00 so too
Hem Toughneaa (ll-M)

a Correlation of Flexural Strength with Flexural Toughness

 

 

 

 

 

I 1

msmm

b Correlation of Flexural Strength with Flexural Stiffness

Figure 5.16 Correlation Between Various Engineering Properties

163

 

2301'
r.

 

 

c. Correlation of Flexural Toughness with Flexural Stiffness

“I” m
"I

 

“Wag”,
Mme-ecu

14>

 

 

 

1.3.
00mm (gm Icu cm)

(1. Correlation of Density with Flexural Strength

Figure 5.16 (Cont’d) Correlation Between Various Engineering Properties

164

Permeability (10"-11 crnlaec)

 

 

 

 

30.
1 Correlation coefficient : -0.99 T
(statistically significant at 99% CL) 3
20 r-
1 o g i l l
1.38 1.42 1.46 1.5 1.54

Density (gm [cu cm)

e. Correlation of Density with Permeability

Figure 5.16 Correlation Between Various Engineering Properties

5.4.3 X-ray Diffraction (XRD). XRD technique offers a convenient way to determine the
mineralogical analysis of crystalline solids. If a crystalline mineral is exposed to X-rays of
a particular wavelengths, the layers of atoms diffract the rays and produce a pattern of
peaks which is characteristic of the mineral [83]. The horizontal scale (diffraction angle)
of a typical XRD pattern gives the crystal lattice spacing, and the vertical scale (peak
height) gives the intensity of the diffracted ray. When the specimen being X-rayed con-
tains more than one mineral, the intensity of characteristic peaks from the individual min-
erals are proportional to their amount. For the recycled wastepaper ﬁber-cement
composites, X-ray diffraction patterns (Figure 5.17) were studied for composition of calci-
um hydroxide and calcium carbonate in order to conﬁrm the results of thennogravimetry
analysis. The x-ray diffraction pattern (typically shown in Figure 5.17) support that calci-
um hydroxide decreases and calcium carbonate increases with ageing. which seems to be
the compositional change reﬂecting the mechanism of ageing in wastepaper ﬁber-cement

composites.

165

 

 

 

Intensity
1000 I ﬂ
——Una9ed I
800 -—Aged '
I l
r
000
1| 1
i
400 ' i
. !
§
200 i
i
l

 

Diffraction Angie

Figure 5.17 X-Ray Diﬁraction Pattern.

5.5 AGEING MECHANISMS

The following modes of composite failure were observed in aged and unaged composites:

(1) Ductile: in which the ﬁber pulled out of the matrix to a considerable extent and the ma-
trix around the ﬁbers was relatively porous. This mode was typical of the unaged compos-
ite (Figure 5.12a).

(2) Semi-ductile: 'in which some ﬁbers broke while others pulled out close to the fractured
surface. revealing the hollow nature of the ﬁber. Fiber-mauix separation or debonding at
the interface could be seen, but the matrix at the interface was quite dense.This mode was
typical of composites subjected to accelerated ageing in hot water bath or under repeated
cycles of freeze/thaw (Figures 5.12c and 5.12d).

(3) Brittle petriﬁed: in which the ﬁbers broke at the fractured surface, revealing circular

cross sections ﬁlled with dense reaction products. Thematrix was as dense as in the brit-

tle-hollow case. but no interfacial separation or debonding could be seen. This mode was

166

typical of composites exposed to ageing in conditions which promoted carbonation. This
mode was observed in wetting/drying with carbonation and to some extent in wetting/dry-
in g accelerated ageing tests (Figure 5.12d and 5.12b). The content of calcium carbonate in
composites were higher for those underwent wetting-drying and carbonation than com-

posites exposed to wetting-drying.

In unaged composite failure was characterized by pull~out of the ﬁbers. This resulted in a
ductile failure mode. In the aged composites, the prevailing mode of failure was a brittle
one, with most ﬁbers breaking at or close to the fractured surface. The densening of the
mauix around the ﬁbers and within ﬁber cores, which was observed after some accelerated
ageing treatments, could be a factor contributing to the reduction in toughness. These phe-
nomena tend to reduce the ﬂexibility and deformation capacity of ﬁbers and increase the
chance of their rupture prior to pull-out. A change of this kind can be accompanied by a
reduction in strength, or an increase in strength, depending on the nature of the ﬁbers and
the effect of the ageing conditions on the properties of the ﬁbers themselves. Petriﬁcation
of ﬁbers and increased bonding are physical phenomena that tend to increase the strength
of the composite materials; the ﬁnal strength of the composite would, however, be in-
creased if these physical phenomena overshadow any attack of ﬁbers upon ageing in the
alkaline environment. It was observed that after accelerated ageing under hot water im-
mersion or repeated freeze-thaw cycles the composite became brittle, and the mode of fail-
ure of the ﬁbers was brittle-hollow. After accelerated ageing under repeated wetting-
drying and carbonation the composite showed a reduction in toughness but its strength and
elastic modulus increased. In this case, the more typical mode of ﬁber failure was brittle-
petriﬁed, and brittle-hollow failure was less frequently observed. Thus there appears to be
a correlation between the changes in the properties of the composite after ageing and the
ﬁber failure mode: semi-ductile failure occurred in composite which did not undergo wet-
ting-drying (and carbonation), say under hot water immersion or repeated freeze-thaw cy-
cles, whereas brittle-petriﬁed failure mode was more pronounced in the composite which
underwent wetting-drying (repeated wetting-drying, repeated wetting-drying and carbon-
ation) which also gained strength. In any case; the fact that ﬂexural strength was either on
start or increased under different accelerated ageing effects indicates that structural chang-
es in the composite which tend to favor strength increase overshadowed any attack on 1i-

gno-cellulosic ﬁbers in the alkaline environment of cement.

167

5.6 MPROVEMENT OF DURABILITY

The petriﬁcation of ﬁbers upon ageing (as observed in accelerated wetting-drying with
carbonation tests) was identiﬁed as the main concern for the longevity of composites. Four
alternative methods were evaluated in this investigation in order to physically or chemi-
cally control the process of ageing and petriﬁcation in wastepaper ﬁber-cement compos-
ites. These methods are described below.

5.6.1 Use of Polymer Dispersion

Efforts were made to modify the cement-based matrix using styrene butadiene polymer
dispersion. The polymer particles are spherical and very small (0.01 to 1 ttm in diame-

ter), and are held in suspension in water by surface active agents [83].

Voids are responsible for low strength as well as poor durability of cement-based compos-
ites in severe environments. Eliminating voids by ﬁlling them with polymer should iru—
prove the characteristics of the material. Polymers have low viscosity, high boiling point
and low cost. Polymer modiﬁcation has been used to improve the water-tightness, durabil-
ity and adhesion capacity of concrete materials. They block the capillary pores and iru-
prove ﬁber-matrix bond.

In our study the latex polymer was added to the slurry. Once the slurry was subjected to
vacuum dewatering, however, the dewatering process was disrupted in the presence of the
polymer dispersion. The screens used to prevent the loss of solids from the fresh compos-
ite during dewatering were blocked by the ﬁne latex particles and the excess water could
not be extracted. The use of polymer dispersions in slurry-dewatered wood ﬁber-cement

composites was thus unsuccessful.

5.6.2 Higher Silica Fume Substitution

Silica fume is a by-product from the reduction of high-purity quartz with coal in electric
arc furnaces in the production of silicon and silicon alloys. The ﬁneness and pozzolanic re-
activity of silica fume make it highly effective in enhancing the density and chemical sta-

bility of the bulk of cement paste and particularly at the interface zones. The consumption

 

168

of calcium hydroxide (a cement hydration product), the reduction of the alkalinity of ce-
ment pore water, and reduced permeability of the matrix are some key mechanisms
through which silica fume could positively improve the long-term stability of cellulose ﬁ-
ber-cement composites. Potential advantages of using silica-fume contents in cellulose ﬁ-

ber-cement composites are discussed below.

(1) Pozzolanic reaction of silica fume with the calcium hydroxide produced by the hydra-
tion of cement leads to the formation of products which are more stable than calcium hy-
droxide; this may reduce the tendencies toward petriﬁcation which is at least partly
responsible for the embrittlement of composites upon ageing. The presence of silica fume
also helps further enhance the chemical stability of the matrix when high-pressure steam
curing is used.

(2) Fineness as well as the pozzolanic activity of silica fume produce a dense microstruc-
ture at the ﬁber-matrix interface zones, which is expected to reduce moisture-sensitivity of

the composite.

(3) Reduced permeability and porosity of the matrix in the presence of silica fume is ex-
pected to protect ﬁbers and thus interface zones against moisture and ageing effects.

(4) Reduced alkalinity of the cement pore water in the presence of silica fume helps con-
trol any alkali attack on the ﬁbers.

In this investigation, 30% of cement by weight was substituted with silica fume (in control
mixtures presented so far only 10% of cement was substituted with silica fume). Other
proportioning and processing variables were kept constant at the optimum levels used in
the control composites (e.g. 50% of cellulose ﬁber was replaced with recycled magazine
Paper).

The ﬂexural load deﬂection curves of the high-silica fume recycled composite (30% silica
fume content), the conventional recycled composite and the control composite (100% vir-
gin ﬁbers) in unaged, aged (subjected to repeated wetting-drying and carbonation) and
aged saturated conditions are presented in Figure 5.18 a. The corresponding ﬂexural

strength, stiffness and toughness test results are presented in Table 5.15 and Figures 5.18
b-d.

Results of the analysis of variance of the data in Table 5.15 are presented in Table 5.16.
Ageing (repeated wetting-drying) and carbonation had statistically signiﬁcant effects on

169

ﬂexural strength and toughness at 99% level of conﬁdence, but not on ﬂexural stiffness.
Two-way analysis of variance with composite type (low silica fume versus high-silica
fume) and age (unaged versus aged) as the two factors indicated that ageing effects on
low-silica fume and high-silica fume recycled composite were statistically diﬁ'erent at
99% level of conﬁdence, further conﬁrming the eﬁ‘ectiveness of high-silica fume contents

in controlling the ageing mechanisms.

A thermogravimetric analysis of the aged composite was canied out for the high silica
fume composite. The calcium carbonate content was observed to drop signiﬁcantly when

compared to the aged low silica fume (optimized) recycled composite considered earlier.

The results presented above indicate that 30% replacement of cement with silica fume in
recycled ﬁber-cement composites is highly effective in controlling the ageing mechanisms
and moisture eﬂ'ects; this presents a practical, economical and efﬁcient approach for en-

hancing the durability and moisture resistance of wastepaper ﬁber-cement composites.

Two-way analysis of variance with silica fume content (high versus low) and moisture
conditions (aged versus aged—saturated) as factors indicated that while saturation has sta-
tistically signiﬁcant effects on ﬂexural performance at 99% level of conﬁdence, the satura-
tion effects on the ﬂexural performance at 99% level of conﬁdence, the saturation eﬁ'ects
on the ﬁexural strength and stiffness (but not toughness) of aged high-silica fume compos-
ites were less pronounced than the corresponding effects on aged low-silica fume compos-
ites.

Table 5.15

 

 

Flexural Performance of Recycled Wastepaper Fiber-Cement
Composite After Higher Silica Fume Substitution.
Experiment Flex. Str. Mean Flex. Too. Mean Init. Stif. Mean Init
(MPa) Flex. Str (N -m) Tou. (Nlmm) Stif.
(95% (95% (95%
Con. InLI Con. Int.) Con. Int.)
Hi gh-Silica 15.28 15.475 79.98 81.83 170.0 166.48
Fume 15.32 ($0.56) 80.23 ($13.5) 169.8 ($30.1)
Recycled 15.41 81.43 176.5
Unaged 15.321 81.09 160
15.67 76.34 159.6
15.55 85.43 156
15.43 83.32 161.6
15.55 89.11 180.4
15.54 81.22 167
15.67 80.19 165.65

 

 

 

 

 

 

 

 

 

170

 

 

 

 

 

High-Silica 16.1 15.963 72.234 71.721 180.87 170.85
Fume 16.23 (:05) 73.657 ($6.1) 176.5 (139.9)
Recycled 15.99 73.11 190.6

15.876 71.009 160.6
Aged 15.834 76.543 174.0

15.811 75.211 175.0

15.892 76.543 163.8

15.912 70.006 165.0

15.997 71.721 162.1

15.712 73.33 159.9
High-Silica 11.98 11.68 68.543 60.991 160.23 156.44
Fume 11.96 ($1.4) 60.765 (123.5) 164.09 (139.1)
Recycled 12.11 59.328 158.54
Aged and 11.89 58.123 165.0
Saturated 11.22 62.985 176.0

11.01 61.432 147.5

11.43 57.543 145.1

11.88 59.876 149.0

11.55 60.887 149.9

11.76 62.311 152.2
Recycled 14.9 14.85 58 59.11 180.2 184.1
(10% Silica 14.88 (10.23) 57.23 ($4.0) 178.2 (117.1)
Fume) 14.77 59.98 180.9
Unaged 14.92 60.11 189.2

14.89 58.23 187.7

14.81 59.12 185.9

14.75 58.87 186.7

14.92 59.98 186.6

14.81 60.34 187.9

14.86 59.22 177.7
Recycled 15.62 15.65 33.19 31.322 270.12 270.91
(10% Silica 15.55 @029) 32.76 ($9.48) 265.87 (117.1)
Fume) 15.76 29.67 275.98
Aged 15.55 34.76 272.65

15.75 32.98 269.98

15.66 31.87 267.77

15.61 32.87 263.87

15.63 28.00 275.65

15.71 27.98 276.76

15.67 29.11 270.41

 

 

 

 

 

 

 

 

 

171

 

 

 

 

 

 

 

 

 

 

 

Recycled 10.2 10.22 55.3 54.41 206 208.66
(10% Silica 10.5 ($0.73) 50.12 ($9.31) 210 ($33.2)
Fume) 10.11 57.65 211.
Aged and satu- 10.23 54.4 195.4
rated 10.37 51.08 213

10.01 53 204.5

9.99 55.2 225

10.43 56.69 216.8

10.18 55.8 204.8

10.33 54.91 200.1
Control 15.1 15.09 67 67.3 164.4 161.9
Unaged 15.02 ($0.51) 66.39 ($8.3) 154.3 ($23.3)

15.43 63.12 171.6

15.1 69.34 167.4

14.99 70.74 152.2

14.98 69.43 162.4

15.1 1 66.98 159.9

15.09 66.12 157.7

14.96 67.38 162.9

15.11 66.55 165.7
Control 16.1 1 16.13 32. 12 31.05 273.23 270.40
Aged 16.20 ($0.93) 27.45 ($8.12) 260.56 ($23.2)

16.19 28.54 280.54

16.20 31.99 268.67

16.15 29.09 267.87

16.2 33.34 270.98

16.25 32.87 278.78

16.54 31.22 265.54

15.77 30.00 267.31

15.712 33.87 270.54
Control 10.43 10.703 50.234 49.002 160.77 160.79
Aged and Satu- 11.12 ($0.29) 49.543 ($5.56) 165.87 ($18.2)
rated 10.54 45.876 138.54

10.32 51.987 128.43

10.77 55.654 170.34

10.34 39.998 144.51

11.009 51.987 187.65

10.987 42. 123 166.43

10.883 57.876 177.76

10.626 50.002 167,54

 

 

172

Table 5.16 Results of Analysis of Variance (ﬁexural strength, toughness andstiffness)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source Sum-of Sq. OF I Mean-Sq. I F-Ratio I P

One Way Analysis of Variance
Flexural Strength
Age. I 0.529 I 1 I 0.529 I 30.446 I 0.001
Error I 0.139 I 8 I 0.017 I I
Toughness
Age. 163.21 1 I 163.21 I 29.8 I 0.001
Error 53.8 8 I 5.47 I I
Initial Stiffness
Age. I 116.07 1 I 116.07 I273 I 0.137
Error I 340.13 8 I 42.51 I I

Two way Analysis of Variance

Flexural Strength
Silica F. 1.049 1 1.049 109.42 0.00
Age. 2.549 1 2.549 265.93 0.00
SF *Age. 0.125 1 0.125 13.02 0.002
Error 0.153 16 0.01
Toughness
SF 4339 1 4339 972.04 0.00
Age. 1741.7 1 1741.7 390.15 0.00
SF*Age 435.8 1 435.8 97.62 0.00
Error 7 1.42 16 4.464
Initial Stiffness
SF 4178.218 1 4178.2 84.132 0.00
Age. 7150.43 1 7150.43 143.98 0.00
SF*Age. 12117 1 12117 244 0.00
Error 794.6 16 49.66

 

 

 

173

 

Source I Sum-of Sq. I DF I Mean-Sq. I F—Ratio I P

 

One Way Analysis of Variance (Aged Vs aged Sat.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Flexural Strength

Sat. Age. I 33.74 I 1 I 33.74 [469.08 I 0.00
Error I 0.576 I 8 I 0.072 I I
Toughness

Sat. Age. I 894.916 I 1 I 894.916 [63.379 I 0.00
Error I 112.96 I 8 I 14.12 I I
Initial Stiffness

Sat. Age. I 1086.05 I 1 [1086.05 I 15.745 I 0.004
Error [551.83 I 8 I 68.97 I I
Two way Analysis of Variance (aged versus aged sat. and silica fume content)
Flexural Strength

Silica F. 7.07 1 7.077 155.17 0.00
Sat. Age. 7.80 1 7.8 171.2 0.00
SF *Sat. 1.438 1 1.438 31.53 0.00
Age.

Error 0.73 16 0.046

Toughness

SF 4373 1 4373 585 0.00
Sat. Age. 812 1 812 108.767 0.00
SF*ASat. 21.37 1 21.37 2.86 0.11
ge

Error 119.45 16 7.466

Initial Stiffness

SF 28146 1 28146 2254 0.00
Sat 16229 1 16229 1299 0.00
Age.(W DC)

SF‘SatAge 3707 1 3707 297 0.00
Error 199.74 16 12.48

 

 

 

 

 

 

 

10*

Figure 5.18 Effects of High Silica Fume Contents on Flexural Performance, Durabil-

"Induces“
-mee "Aged 'Ageetd.

 

I74

 

 

 

 

 

 

Deﬂection (mm)

High Silica Fume

L ,I IW‘
I '-.
o - . s
O 1 2 2 4
Deflectionhun)
Recycled

3 Load Deﬂection

ity and Moisture Sensitivity.

Nude-mm

 

10"

 

—W com H..'I'.'

 

 

,I\ I 00% Could-tea
: : m
1 ; 3 ‘
Deflecdon (tern)
Virgin

I75

Flexural Strength (MPa)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2° I'Elaged aAged Daged and saturag;
I 95% Confidence interval I
15 *- I
“E E
to» E
5 r-
o ' ‘
Recycled Virgin High-SF Rec.
Type of Composite
b Flexural Strength
Flexural Stiffness (ll/mm)
Flexural Toughness (ll-mm) ‘00

 

 

100?“ BE“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3:.- aea—ee I 95% confidence Interval 3*“ m I 95% Confidence Interval
m-
80 " E 300 c
i
60 h .
E / 200 *-
40 ~ If E
100 r
20 '-
l O ; grj ~.,
0 atattered Virgin Hlell 8F Rec J "”th "'9'" "W's" 9°C-
Type of Composite WP‘ 01 Commsite
c Toughness (1 Initial Stiffness

Figure 5.18 (Cont’d) Effects of High Silica Fume Contents on Flexural Performance,

Durability and Moisture Sensitivity.

176

‘Wbmhtnu
100 -

 

90‘-

 

80*.

' Uneged Rec.

_M.d Rec.

70 ’ 'Hlnhor-Sf (Mun i 1 i .
0 200 400 600 800 'l 000 1 200

Tempereture (deg. C)

 

 

 

Figure 5.19 TGA Results Silica Fume Modiﬁed

5.6.3 Using Carbonation in Processing

Carbonation of the fresh composite at appropriate moisture content was hypothesized to
stabilize the composite by controlled conversion of calcium hydroxide into calcium car-
bonate. While this concept seems to warrant further investigation, preliminary efforts to
establish a moisture content in fresh composite at which carbonation can effectively take

place were not successful.

5.6.4 Using Special (low calcium hydroxide) Cement

Calcium hydroxide is the cement hydration product which seems to be responsible for the
petn'ﬁcation and thus embrittlement of wood ﬁber-cement composites upon ageing. Ad-

verse effects of calcium hydroxide on ageing are also observed in glass ﬁber reinforced

177

concrete. In order to control ageing effects, the glass ﬁber-cement industry, in cooperation
with cement industry, has developed cements the hydration of which produces minimal
calcium hydroxide. The Special cement considered in this investigation was from Molloy
Company [84]. The cement in general has 8 parts of ordinary Portland cement, 3 parts are
calcium sulphoaluminate and one part of a synthetic powder. The exact chemical compo-
sition is not revealed by the company. It is called a low calcium hydroxide cement. It has
been used successfully in glass ﬁber reinforced concrete products. Its cost is ﬁve to six

times that of ordinary Portland cement.

The ﬁexural load-deﬂection curves for unaged, aged (under repeated wetting—drying and
carbonation) and aged-saturated composites made with special cement and recycled ﬁbers,
regular cement and recycled ﬁbers, and virgin ﬁbers and regular cement are presented in
Figure 5.20 a. The corresponding ﬂexural strength, sﬁﬂness and toughness test results are
presented in Table 5.17 and Figures 5.20 b-d.

Analysis of variance was carried out on the test data of Table 5.17 (see Table 5.18). The
ageing effects on the ﬂexural toughness and stiffness (but not strength) of composites
made with special cement were statistically signiﬁcant at 95% level of conﬁdence. 'I\vo-
way analysis of variance suggested, at 99% level of conﬁdence, that ageing eﬂ‘ects on ﬂex-
ural stiffness and toughness were less pronounced when special cement was used to re-

place regular Portland cement.

One -way analysis of variance (aged versus aged-saturated recycled composites with spe-
cial cement) followed by two~way analysis of variance with composite type (special ce-
ment versus regular cement) and saturation condition (aged versus aged-saturated)
indicated, at 99% level of conﬁdence, that saturation inﬂuences the ﬂexural strength, stiﬂ-
ness and toughness of recycled composites with special cement; while moisture eﬁ'ects on
ﬂexural strength were comparable in composites with special cement and regular cement,
the corresponding effects on toughness and stiffness were inﬂuenced by the type of ce-
ment. With special cement moisture effects were less pronounced when compared with

regular cement.

Thermogravimetric compositional analysis (Figure 5.21 and Table 5.19) indicate that the
special cement reduces the calcium carbonate content of aged composites. This can be at-

tributed to the reduced calcium hydroxide content of the special cement hydration prod-

178

“cm.

A comparison of the calcium hydroxide and calcium carbonate contents of aged compos-
ites with regular cement, hi gh-silica fume binder and special cement is presented in Table
5.19. Analysis of variance of the results presented in Table 5.19 (see Table 5.20) suggested
that ageing effects on calcium carbonate content were signiﬁcant at 99% level of conﬁ-
dence, and different composites (high-silica fume, special cement, and optimum with low
silica fume content) produced different calcium carbonate contents upon ageing. The least
calcium carbonate content (and thus conceivably the most stable performance under age-

ing eﬁ'ects) was obtained in the high-silica fume composite.

 

 

 

Table 5.17 Flexural Performance of Virgin and Recycled Wastepaper
Fiber-Cement Composite Using Special and Regular Cement

Experiment Flex. Str. Mcsn Flex. Tou. Mesa Init. Stif. Mesa Init

(MPs) Flex. Str (N -m) Tou. (Nlmm) Stif.

(95% cm (95% CI) (95% C1)

Recycled 14.88 14.78 60.556 61.32 160.65 159.61
Using Special 14.78 (i134) 61.987 (17.88) 161 (131.0)
Cement 14.98 62.334 155.87

15.1 59.113 167.43
Unaged 14.854 60.543 150.54

14.657 61.321 170.33

14.765 60.876 152.98

15.32 58.123 148.25

14.23 63.129 170.43

14.24 65.222 158.65
Recycled 15.1 15.215 55.665 54.772 190.54 188.67
Using Spec. 14.887 (10.94) 56.321 (8.52) 185.34 (125.1)
Cem. 15.54 55.449 187.34

15.342 58.21 177.45
Aged 15.119 53.123 195.33

14.995 54.567 199.22

14.987 52.32 190.44

15.236 51.987 193.33

15.339 57.32 183.25

15.612 52.765 184.44

 

 

 

 

 

 

 

 

 

179

 

 

 

 

 

Recycled 11 11.07 65.76 65.32 170.43 171.97
Using Spec. 10.99 (11.13) 65.44 (110.9) 175 (127.1)
Cem. 11.02 64.98 180.65

10.98 69.43 169.32
Aged and 11.06 68.23 161
Saturated 10.57 61.22 171

11.11 61.87 168.8

11.08 68.54 185.44

11.15 64.32 167.2

11.76 63.33 170.9
Recycled 15.62 15.65 33.19 31.322 270.12 270.91
Using Regular 15.55 (10.29) 32.76 (19.48) 265.87 (117.1)
cement 15.76 29.67 275.98
Aged 15.55 34.76 272.65

15.75 32.98 269.98

15.66 31.87 267.77

15.61 32.87 263.87 ‘

15.63 28.00 275.65

15.71 27.98 276.76

15.67 29.11 270.41
Recycled 10.2 10.22 55.3 54.41 206 208.66
Using Regular 10.5 (10.73) 50.12 (19.31) 210 (133.2)
cement 10.11 57.65 211.
Aged and Satu- 10.23 54.4- 195.4
rated 10.37 51.08 213

10.01 53 204.5

9.99 55.2 225

10.43 56.69 216.8

10.18 55.8 204.8

10.33 54.91 200.1
Recycled 14.9 14.85 58 59.11 180.2 184.1
Using Regular 14.88 (10.23) 57.23 (14.0) 178.2 (117.1)
cement 14.77 59.98 180.9
Unaged 14.92 60. 1 1 189.2

14.89 58.23 187.7

14.81 59. 12 185.9

14.75 58.87 186.7

14.92 59.98 186.6

14.81 60.34 187.9

14.86 59.22 177.7

 

 

 

 

 

 

 

 

 

180

 

 

 

 

Control 15.1 15.09 67 67.3 164.4 161.9
Unaged 15.02 (10.51) 66.39 @83) 154.3 (123.3)

15.43 63.12 171.6

15.1 69.34 167.4

14.99 70.74 152.2

14.98 69.43 162.4

15.1 1 66.98 159.9

15.09 66.12 157.7

14.96 67.38 162.9

15.1 1 66.55 165.7
Control 16.1 1 16.13 32. 12 31.05 273.23 270.40
Aged 16.20 (10.93) 27.45 (18.12) 260.56 (123.3)

16.19 28.54 280.54

16.20 31.99 268.67

16.15 29.09 267.87

16.2 33.34 270.98

16.25 32.87 278.78

16.54 31.22 265.54

15.77 30.00 267.31

15.712 33.87 270.54
Control 10.43 10.703 50.234 49.002 160.77 160.79
Aged and Satu- 11.12 (10.29) 49.543 (15.56) 165.87 (118.2)
rated 10.54 45.876 138.54

10.32 51.987 128.43

10.77 55.654 170.34

10.34 39.998 144.51

11.009 51.987 187.65

10.987 42.123 166.43

10.883 57.876 177.76

10.626 50.002 167.54

 

 

 

 

 

 

 

 

 

181

Table 5.18 Results of Analysis of Variance (ﬂexural strength, toughness and stiffness)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source I Sum-of so. I or: I Mean-Sq. I F-Ratio I P

One Way Analysis of Variance
Flexural Strength
Age. I 0.161 I 1 I 0.161 I 3.72 I 0.09
Error I 0.345 I 8 I 0.043 I I
Toughness
Age. I 144.4 I 1 I 144.4 I 32.08 [000
Error I 36 I 8 I 4.5 I I
Initial Stiffness
Age. I 200052 I 1 [200052 I 84.05 I 0.00
Error I 190.4 I 8 I 23.8 I I

Two way Analysis of Variance

Flexural Strength
Spec. Cem. 232.49 1 232.49 0.946 0.345
Age. 207.56 1 207.56 0.844 0.372
SC *Age. 264.192 1 264.19 1.075 0.315
Error 3933.8 16 245.86
Toughness
SF 731.8 1 731.8 276.25 0.00
Age. 1295.7 1 1295.7 489.1 0.00
SC*Age 684.21 1 684.2 258.29 0.00
Error 42.38 16 2.649
Initial Stiffness
SC 13520 1 13520 515.5 0.00
Age. 17169 1 17160 654.71 0.00
SC*Age. 4620 l 4620 176. 19 0.00
Error 419.6 16 26.22

 

 

 

182

 

Source ISum-of Sq. IDF IMean-Sq. IF-Ratio IP

 

One Way Analysis of Variance (aged vs. sat. aged

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Flexural Strength

Sat. Age. I 1.832 I 1 I 1.832 I 33.207 I 0.00
Error I 0.44 I 8 I 0.055 I
Toughness

Sat. Age. I 1570 I1 I 1570 [210.9 I 0.00
Error I 59.552 I 8 [7.444 I
Initial Stiffness

Sat. Age. 2992 I 1 I 2992 I 72.11 I 0.00
Error 332 I 8 I 41.5 I

Two way Analysis of Variance (aged vs aged sat, special cement vs. regular cement)
Flexural Strength

Spec. Cem. 8.32 1 8.32 93.107 0.00
Sat. Age. 0.8 1 0.8 8.952 0.009
SC ‘Sat. 0.003 1 0.003 0.038 0.848
Age.

Error 1.43 16 0.089

Toughness

SC 1277.29 1 1277 48.68 0.00
Sat. Age. 2060 l 2060 78.53 0.00
SC‘ASat. 463.55 1 463.55 17.66 0.001
ge

Error 393.55 15 26.237

Initial Stiffness

SC 5985 l 5985 28.54 0.00
Sat. Age 1377.8 1 1377.8 6.57 0.021
SC*Sat. 1805 l 1805 8.607 0.01
Age.

Error 3355 16 209.72

 

 

 

 

 

 

 

 

183

Fienrrei Stress (IPe)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2° -Aged " Unaged "'Aged-Sst.
95% Confidence
15 )— 7i. — interval
10 r-
s r
o r l
O 1 3 4
Deﬂection (mm)
Special Cement
Flexursl s p
Fienrrei Stress (lute) 20 trees 1" ')
2° — Une ed ° ’A .
—Unsged - ’Aged “Aged 3.1. 9 9'" A9“ su.
~ I T 95% Confidence
95% Confidence — I \ —.
15 " ,I- .1. -— mun/.1 15 1' 1| interval
10 ~ I .| 10 - ,’ ‘
,I I. I, i.
5 C II 41 5 h I I -
’I i. I, I
o I " 1 I . I 1 I 21
O 1 2 3 4 0 1 2 3 4
Deﬂection (mm)

Recycled

and Moisture Sensitivity

a Load Deflection

Deflection (mm)

Vrrgin

Figure 5.20 Effects of Using Special Cement on Flexural Performance, Durability

184

Flexural Strength (HPa)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2° Wged QAged Uaged and saturated
I 95% Confidence interval
15 '-
10, "r it
5 )-
Recycled Virgin Special Cem. Rec.
Type of Composite
b Flexural Strength
Flexural Toughness (Ii-mm) Flexural Stittness (N/mm)
1 00 I; ‘00 EM
She-O -— a...
Duo- . a... 95% Confidence mum" Zinc- .- W I 95% confidence interval
80 -
300 -
.9
LIE 200 - If
40 - {I EI
100 ~ ”"-
20 -
: ‘ , . o . i : . 1
Recycled Virgin Special Cem. Rec. Recycled Virgin Special Cem. Rec.
Type of Composite Type of Composite
c Toughness (1 Initial Stiffness

Figure 5.20 (Cont’d.) Effects of Using Special Cement on F lexural Performance, Du-

rability and Moisture Sensitivity

 

185

 

70

' Unaged Rec.
‘9“ Rec.

‘ sPeciel ement

200 400

600 800

  

Temperature (deg. c)

Figure 5.21 TGA Special Cement Composite

1000

  
  
   

1200

Table 5.19 Thermogravimetric Compositional Analysis

 

Type of

 

 

 

 

 

 

 

 

 

 

. . Mean (95% . 1 Mean (95%
r ( r
Reﬁnement CH (Weight 7c) Con. Int.) CC (Wclght‘7o Con. Int.)
Unaged 22.21.229.219 22.29 (12.33) 3.12, 3.05, 3.37 3.18 (10.53)
Aged 17.7, 1823.17.11 1794 (11.05) 794,804,778 7.92 (10.42)
SF Aged 4.68, 4.65, 4.50 4.61 (10.37) 1.3, 1.26, 1.32 1.29 (10.09)
SC Aged 9.91. 9.16, 10.21 9.76 (12.12) 2.65, 2.32, 2.1 2.35 (10.88)

 

 

 

186

Table 5.20 Results of the Analysis of Variance (TGA)

 

 

 

 

 

 

 

 

 

 

 

 

Source Sum-of Sq. DF Mean-Sq. I F-Ratio I P
One Way Analysis of Variance

Calcium Carbonate Content

Ageing 60.02 I 4 15.00 I 812.916 I 0.00

Error 0.185 I 10 0.018 I I

Type of 26.58 1 26.58 1440.23 0.00

Composite

Error 0.185 10 0.018

 

 

 

 

 

 

 

 

5.6.5 Comparison of Special Cement with High Silica Fume Contents

The ﬂexural performance of aged-saturated composites with special cement and with
high-silica fume binder were composed in this section. Analysis of variance followed by
comparison of means (see Table 5.21 for special cement versus high-silica fume compos-
ites) indicated that, at 95% level of conﬁdence, silica fume and special cement produced
higher strengths in aged saturated condition than optimized (low silica fume with regular
cements) recylced composite; high silica fume composites were found to be superior to
special cement composites as far as aged-saturated ﬂexural strength, stiffness and tough-
ness are concerned. ’

Flexural strength of the silica fume was sadistically signiﬁcantly superior at 95% level of
conﬁdence. On ageing the increase(3%) in ﬂexural strength was observed in both compos-
ites.

Two-way analyses of variance were also conducted to study the effects of ageing on high-
silica fume and special cement composites. Upon ageing both composites exhibited com-
parable (about 5%) and statistically signiﬁcant, at 95% level of conﬁdence, increases in
ﬂexural strength. Ageing also has statistically signiﬁcant effects on toughness, and high-
silica fume and special cement composites showed comparable drops (about 12%) in
toughness upon ageing (when compared with 40% drop in low silica fume composites
with regular cement and the increase in stiffness in high-silica fume composite upon age-
ing (about 5%) was less than that in special cement composite (about 18%).

 

187

Table 5.21 Results of Analysis of Variance

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Source ISum-of Sq. DF Mean-Sq. IF-Ratio IP
One Way Analysis of Variance

Flexural Strength

cem vs sf I 4.598 I 1 I 4.598 I 76.643 I 0.00

Error I 0.96 I 16 I 0.06 I I

Toughness

cem vs sf I 4902.2 1 4902.2 I 1479.6 I 0.00

Error I 53.0 16 3.31 I I

Stiffness

cem vs sf I 48381.3 1 48381 I 131.4 I 0.00

Error I 5889 16 368 I I

 

 

 

 

 

5.7 SUMMARY AND CONCLUSIONS

The effects of moisture and accelerated ageing on the flexural performance and micro-
structural characteristics of the Optimized wastepaper ﬁber-cement composites and control
composites (made fully with virgin softwood kraft ﬁbers) were investigated. Microstruc-
tural studies utilized the Scanning Electron Microscopy, thermogravimetry and X-ray dif-
fraction techniques. Appropriate measures were adopted and evaluated for controlling the
ageing and moisture effects on wastepaper ﬁber-cement composites. It was concluded
that:

(1) The increase in moisture content of virgin and recycled composites reduced ﬂexural
strength and stiffness, and increased toughness of the composites One way analysis of
variance of ﬂexural strength, stiffness and toughness test results at different moisture con-

tents conﬁrmed the moisture sensitivity of recycled and virgin composites.

Among the moisture conditions considered (oven-dried, air-dried and saturated), satura-
tion produced a distinct ﬁexural behavior. Air-dried and-oven dried recycled composites

 

 

a-

188

proved to be statistically equivalent in ﬂexural strength but different in ﬂexural toughness
and stiffness at 95% conﬁdence level (using multiple comparison by contrast). Two-way
analysis of variance of results was also conducted for different composites (recycled ver-
sus virgin) and moisture conditions. The ﬁber-moisture interaction proved to be statistical-
ly signiﬁcant at 95% level of conﬁdence, indicating that moisture effects on recycled
composites differ from those on virgin composites. Damaging effects of saturation are less
pronounced in the case of recycled ﬁbers. Upon saturation, recycled composites exhibited
a drop of 37% in ﬂexural strength and 16% in initial stiffness when compared with air-
dried composites; the corresponding drops in virgin composites were 47% and 28%, re-
spectively. Both recycled and virgin composites showed an increase of 32% in ﬂexural
toughness upon saturation.

(2) Repeated wetting-drying and particularly wetting-drying and carbonation cycles
caused an increase in ﬁexural stiffness and strength of virgin and recycled composites but

led to reduced toughness and embrittlement of the materials.

Repeated wetting—drying cycles had statistically signiﬁcant effects on all flexural proper-
ties (strength, toughness and stiffness) of recycled composites. Two-way analysis of vari-
ance (with two factors of recycled versus virgin ﬁbers and aged versus unaged
composites) revealed that there was a statistically signiﬁcant diﬁ'erence, at 95% level of
conﬁdence, between ageing effects on recycled and virgin composites as far as ﬁexural
strength and toughness are concerned; the initial stiffness of virgin and recycled compos-
ites, however, was affected similarly by ageing under repeated wetting-drying cycles.
Damaging effects of repeated wet-dry cycles on recycled composites (which lost 32% of
toughness upon ageing) was less than that on virgin composites (which lost 45% of tough-

ness upon ageing).

(3) Statistical analyses indicated, at 95% level of conﬁdence, that the addition of carbon-

ation to wet-dry cycles leads to pronounced effects of ageing on composites

(4) Repeated freeze thaw cycles did not have statistically signiﬁcant effect on ﬂexural
strength and toughness of recycled composite; however, the effects on increasing the ini-
tial stiffness was statistically signiﬁcant at 95% level of conﬁdence. Two way analysis of
variance (taking into account the virgin composite) revealed that freeze-thaw cycles affect
the initial stiffness recycled composites in a way different from virgin composites (similar

 

 

189

effects were observed on ﬂexural strength and toughness irrespective of the composite
type). While recycled composites showed a 1% increase in stiffness under repeated freeze-
thaw cycles, virgin composites showed a drop of 9% in stiffness under this ageing condi-
tion.

(5) The effects of long-term immersion in hot water on ﬂexural strength, Stiffness and
toughness of recycled composites were not statistically signiﬁcant at 95% level of conﬁ-
dence. Two-way analysis of variance of results (virgin versus recycled ﬁbers) concluded
that only ﬁexural stiffness of virgin and recycled composites was affected differently by
this ageing process. While hot water immersion slightly reduced the initial stiﬁness of re-
cycled composites (by 3%) it caused a small increase of 3% in the stiffness of virgin com-

posites.

(6) Comparing the effects of different accelerated ageing tests based on statistical analy-
ses, it appears that wetting-drying and carbonation is the most effective method to bring
about changes in physical and mechanical properties of the wood ﬁber-cement compos-
ites. In order to compare the effects of different accelerated ageing conditions on the ﬂex-
ural performance of recycled and virgin ﬁber-cement composites, two-way analysis of
variance of ratios of aged to unaged ﬂexural strength, stiffness and toughness test results
were conducted. The two factors in these analyses were: composite type (recycled and vir-
gin), and accelerated ageing condition (wet-dry, wet-dry and carbonation, freeze-thaw, hot
water immersion). Results of the analysis of variance indicated that, at 95% level of conﬁ-
dence, all the ageing conditions have comparable effects on ﬁexural strength but there are
statistically signiﬁcant diﬁ‘erences between the effects of different accelerated ageing con-
ditions on ﬂexural Stiffness and toughness. There was no statistically signiﬁcant difference
between virgin and recycled composites as far as ageing effects on ﬁexural strength and
stiffness are concerned; the two composites, however, behaved differently in ageing ef-

fects on ﬁexural toughness.

Multiple comparison of results indicated that, at 95% level of conﬁdence, each of the ac-
celerated conditions have distinctly different effects on ﬂexural toughness and Stiffness,
except for the hot water immersion and freeze-thaw condition which had statistically com-
parable effects on ﬂexural Stiffness. As far as the overall ageing effects on ﬂexural perfor-
mance is concerned, repeated wetting-drying and carbonation cycles produced the most

pronounced effects whereas hot water immersion caused the least effects. The trends in the

 

190

response of virgin and recycled composites to ageing were generally comparable.

(7) Ageing effects on the morphology of ﬁbers and failure mode as observed in SEM
(comparing different accelerated ageing conditions) were studied and it was concluded
that ﬁbers were not unaffected by the freeze-thaw and hot water immersion ageing condi-
tions. In the case of wetting-drying and carbonation, however, tendencies towards ﬁlling
of ﬁbers, densiﬁcation of interfaces and dominance of ﬁber rupture in failure mode were
observed. For the unaged composite, the dominant mode of failure was observed to be ﬁ-
ber pull out. In the case of specimens subjected to repeated cycles of freeze-thaw, the frac-
ture surface was observed to Show a combination of ﬁber pull out and ﬁber fracture. In the
case of hot water immersed composites also the mode of failure was observed to be ﬁber
pull out accompanied with ﬁber fracture. For the repeated wetting-drying ageing condition
the mode of failure at the fracture surface was dominated by ﬁber fracture. Also, for re-
peated wettin g-dryin g and carbonation ﬁber rupture was observed to dominate the failure
mode.

In addition, the appearance of ﬁbers and their interfaces was also affected by the ageing
processes. In the case of wettin g-dryin g and wetting drying with carbonation, a dense ﬁber
mauix interface was observed. Fibers in case of wetting-drying with carbonation appeared
to be ﬁlled with hydration and carbonation products. In the case of hot water immersion
and freeze-thaw ageing conditions the dcnsiﬁcation of the interface and ﬁlling of ﬁber
core was not pronounced. It appears that the ageing process is most pronounced under re-

peated wetting-drying and carbonation condition.

(8) Results of thermogravimetric analysis suggested that compositional changes occur in
wood ﬁber-cement composites upon ageing. The calcium carbonate content increases and
calcium hydroxide content decreases under accelerated ageing conditions. These trends
were conﬁrmed through X-ray diffraction analysis. Wetting-Drying and carbonation fol-
lowed by wetting-drying were observed to result in most pronounced compositional
changes. These changes partly explain the brittle behavior of the composite after wetting-
drying and carbonation. Under ageing effects there seems to be a tendency in the calcium
hydroxide constituent of cement hydration products to dissolve in cement pore water and
this process is accompanied with the carbonation of calcium hydroxide which produces
calcium carbonate. The “petriﬁed” ﬁbers with strong bonding to matrix tend to be Strong
but brittle. It may be hypothesized that the increase in Strength and rigidity of the petriﬁed

191

ﬁbers, and the increase in their bond strength due to the densiﬁcation and also elimination
of shrinkage debondin g at the interface zones account for the increase in Strength and stiﬂ-
ness of the composite. Petriﬁcation and well-bonded ﬁbers, however, tend to fracture prior
to pulling out of the matrix; this eliminate the energy absorption associated with ﬁber pull

out and thus causes embrittlement of the composite.

(9) A Statistically Signiﬁcant negative correlation was observed between calcium carbon-
ate content and toughness, in unaged and aged composites, the positive correlation be-
tween calcium carbonate content and Stiffness were also statistically signiﬁcant 11
statically signiﬁcant positive correlation was also observed between density and calcium
carbonate content. These correlations conﬁrmed the key role the formation of calcium car-
bonate plays in the ageing of the composites. The correlations between calcium hydroxide
content and various engineering properties of unaged and aged composites were not statis-
tically Signiﬁcant. This further underlines the signiﬁcant effects of carbonation in the pro-
cess of petriﬁcation of ﬁbers upon ageing.

The correlations of calcium hydroxide content with engineering properties (Strength, stiff-
ness and toughness) were not statistically Signiﬁcant. A strong positive correlation was
observed between calcium carbonate content and density (correlation coefﬁcient: 0.941)
which was statistically signiﬁcant at 99% conﬁdence level. Calcium carbonate content
seemed to have a strong correlation than calcium hydroxide content with engineering
properties. This may be attributed to the fact that ageing causes dominantly a migration of
calcium hydroxide (and only a small change is its content as a result of carbonation) while
it actually forms calcium carbonate (which reﬂects in increased calcium carbonate con-

tent

(10) Since formation of calcium carbonate through carbonation of calcium hydroxide as
well as the transport of calcium hydroxide to ﬁber cores and interfaces play critical roles
in ageing effects on the composites, reduction of calcium hydroxide seems to provide for a
more stable composite. Replacement of relatively high levels of cement with silica fume
presents an approach to the reduction of calcium hydroxide content. Silica fume also helps
reduce the permeability of composites and the alkalinity of pore water. While the opti-
mized recycled composites has 10% Silica fume content, the reﬁned high-silica fume com-
posite considered had 30% silica fume content.

ﬁx:
{

 

192

Repeated wetting-drying and carbonation Still had Statistically Signiﬁcant effects, at 95%
level of conﬁdence on ﬂexural strength and toughness (but not sﬁﬁness) of high-silica
fume composites. Statistical analyses including results for low silica fume composites,
however,; indicated that ageing effects on high-Silica fume composites were much less
pronounced; the calcium carbonate content of aged high-Silica fume recycled composites
was also less than that of low silica fume composites. While saturation of aged high-silica
fume composites led to statistically Signiﬁcant effects on ﬂexural performance, statistical
analyses indicated, at 99% level of conﬁdence, that the damaging effects of moisture on
ﬂexural Strength and Stiffness were less pronounced in aged high-silica fume recycled

composites when compared with corresponding low-Silica fume composites.

In short, 30% replacement of cement with silica fume in recycled ﬁber-cement composites
was found to be highly effective in controlling the ageing mechanisms and moisture ef-
fects; this approach presents a practical, economical and efﬁcient approach for enhancing

the durability and moisture resistance of wastepaper ﬁber-cement composites.

(11) In an alternative approach to reduce calcium hydroxide content of cement hydration
products a Special cement was considered (consisting of 8 parts of Portland cement, 3
parts of calcium sulphoaluminate, and one part of synthetic powder) the hydration of

which does not produce much calcium hydroxide.

While repeated wetting—drying and carbonation effects on the ﬂexural toughness and stiff-
ness (but not strength) of composites made with special cement were still statistically Sig-
niﬁcant at 95% level of conﬁdence, it was concluded at 99% level of conﬁdence, that
ageing effects on ﬂexural Stiffness and toughness were less pronounced when special ce-

ment was used to replace regular Portland cement in recycled composites.

One -way analysis of variance (aged versus aged-saturated recycled composites with spe-
cial cement) followed by two-way analysis of variance with composite type (special ce-
ment versus regular cement) and saturation condition (aged versus aged-saturated)
indicated, at 99% level of conﬁdence, that while moisture effects on ﬂexural strength were
comparable in composites with special cement and regular cement, the corresponding ef-
fects on toughness and Stiffness were inﬂuenced by the type of cement. With Special ce-

ment, moisture effects were less pronounced when compared with regular cement.

Thermogravimetric compositional analysis indicated that the special cement reduces the

193

calcium carbonate content of aged composites. This can be attributed to the reduced calci-
um hydroxide content of the special cement hydration products.

(12) Tire effectiveness of special cement and Silica fume in recycled composites was com-
pared. It was found that silica fume is superior in performance when compared with spe-
cial cement. analysis of variance of the results conﬁrmed that different composites (high-
Silica fume, Special cement, and optimum with low silica fume content) produced diﬂerent
calcium carbonate contents upon ageing. The least calcium carbonate content (and thus
conceivably the most Stable performance under ageing effects) was obtained in the high-

Silica fume composite.

(13) Effects of polymer dispersion and carbonation as partly of the process were unsuc-
cessfully investigated for improving the durability and moisture sensitivity of the wood ﬁ-

ber-cement composites.

 

CHAPTER 6

COST ANALYSIS

Recycled wastepaper utilization as reinforcement in thin ﬁber-cement Sheets is a potential
market area for beneﬁcial use of wastepaper. It is not only a technically feasible and envi-
ronmentally beneﬁcial alternative, but also an economically attractive option for the thin
cement products industry. This section evaluates the economic advantages of recycling

wastepaper ﬁbers in cement composites for Siding applications.
6.1 Cost of Recycled Wastepaper Fiber-Cement Composites
Cost of various constituents of cellulose ﬁber-cement composites are presented below.

Recycled Wastepaper Fibers $0.32 / lb

Virgin Cellulose Fibers $0.66 / lb

Portland Cement Type 130.07 / lb

Silica Fume $0.16 / lb

Ground Silica sand $0.02 / lb

Flocculating agent $2.00 /lb

Assumed cost of labor and equipment: $0.07 per sq.ft. of panel

For the optimum mix composition of recycled ﬁber-cement composites considered in this
investigation, the cost of 6 m (1/4 in) panel with the above material costs can be esti-

mated at $0.37 per square ft.

194

 

195

6.2 Comparative Cost Analysis

The market was surveyed for alternative materials available for the production of Siding
panels The mean costs of alternative siding panels (ﬁnished products) in the market are
compared in Table 6.1 with that of the optimized recycled ﬁber-cement composites.
Table 6.1 Cost Comparison with Alternate Siding Materials in Market

 

 

 

 

 

 

 

 

Type of Siding C0533 f“ :I

aluminium Siding 0.88
Vinyl Siding 0.55
Solid Wood 0.75
Plywood 0.58
Virgin Cellulose-Cement 0.56
(Hardie)

Wastepaper ﬁber-cement 0. 37

 

 

 

 

The cost comparison of Table 6.1 reﬂects the economic beneﬁts of replacing virgin cellu-
lose ﬁbers with wastepaper ﬁbers in thin cement products. Recycled ﬁber-cement panels
with desirable technical characteristics present an economically superior alternative to
various siding materials such as vinyl, plywood, solid wood, virgin cellulose-cements, and

aluminium.

6.3 Life Cycle Costs

One major concern in cost analysis for construction materials is the life-cycle costs which
cover, besides the initial expenses, the maintenance costs. In short, life-cycle costs include
all costs anticipated over the life of the product. Table 6.2 presents a comparative life—
cycle cost analysis of wastepaper ﬁber cement composites and other Siding materials.
Desirable durability characteristics and low maintenance expenses of wastepaper ﬁber—

cement boards make their life-cycle cost even more attractive than their initial cost.

 

Table 6.2 Life Cycle Cost Analysis

196

 

 

 

 

 

 

 

 

 

 

 

 

 

Initial Design Maintenance Annual cost ($

Type of Siding Cost Life Cost/yr. per unit area sq
5 Years $ ft.)

Aluminium 0.88 50 — 0.0176
Vinyl 0.55 50 - 0.011
Solid Wood 0.75 50 $005 0.02
Plywood 0.58 50 $0.005 0.0166
Hardie (Virgin 0.56 50 - 0.011
Cellulose—
Cement)
Wastepaper (1.37 50 - 0.007
ﬁber-cement

 

Comparative studies of loss in product value (Table 6.3) and net product value after 50

years (Table 6.4) further conﬁrm the economic superiority of wastepaper ﬁber-cement

composites when compared with alternative siding materials.

Table 6.3. Loss in Product Value in Design Life

 

Product Value (%)

 

 

 

 

 

 

 

 

 

 

 

Type of Siding 0-5 Years 6-14 Years 15-50 years
Value Loss in value value

Aluminium 100 10% / year 10%

Vinyl 100 10% /year 10%

Solid Wood 95 10% 5%

Plywood 95 10% 5%

Hardie (Virgin 100 2.2%/year losing @ 2.2%

Cellulose-

Cement)

Wastepaper 100 2.2% /year losing @ 2.2%

ﬁber-cement

 

 

 

 

 

197

Table 6.4 Net Product Value After 50 years

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'Iype of Sid- Initial Inﬂation Dcpr./yr. Depr./yr/ Depr. Net
ing Cost 1-5 years 6-14 15-50 Value
* 100 / years years After 50
sq.ft. years
Aluminium 88 3% - 10% 10% 3.88
Vinyl 55 3% - 10% 10% 2.43
Solid Wood 75 3% 1% 5% 5% 20.82
Ply wood 58 3% 1% 5% 5% 16.1
JH(vir. cell) 56 3% - 2.2% 2.2% ‘ 92.9
WPFC 37 3% - 2.2% 2.2% 61.39

 

 

 

A market survey revealed that aluminium siding was the most popular siding in the seven-
ties and had almost 90% of market share. Presently vinyl has become the dominating Sid-
ing material. Some concerns were expressed during the market survey in relation to
aluminium and vinyl sidings. Aluminium sidings fade away with time. Aluminium and
vinyl sidings are Sheets with low load resistance when compared with ﬁber-cement or
wood-based Sidin gs. It is also important to note that wood sidings have to be periodically
maintained throughout their life. Wood sidings also have a relatively low ﬁre resistance;

cellulose ﬁber cement composites, on the other hand, possess desirable ﬁre resistance.

The savings associated with the use of wastepaper ﬁbers in cement products are not only
in the product manufacture; a comprehensive economical study, can not neglect the
avoided landﬁll costs (estimated at $50 per ton) associated with diverting market-limited
wastepaper from landﬁlls.

The world of building materials is very dynamic. Changes in Siding materials from solid
wood to aluminium and then to vinyl reﬂect the fact that technological progress, economi-

cal factors and environmental issues had to change the siding materials dominating at dif-

 

198

ferent times. There are good economic and environmental reasons to support successful
implementation of technology developed in this research for the utilization of wastepaper
ﬁbers in thin cement products.

6.4 SUMMARY AND CONCLUSIONS

A comprehensive cost analysis was conducted on the developed high-Silica fume recycled
composite versus alternative Siding materials. It was concluded that:

(1) Recycled ﬁber-cement composites present technically desirable qualities together with
lowest initial and life-cycle costs when compared with alternative Siding materials.

(2) Recycled wastepaper ﬁber-cement composites present a number of positive qualities
which make them further competitive. Their ﬁre resistance is superior to wood or vinyl
sidings. They do not fade away like aluminium sidings with time. Aluminium and vinyl
sidings are sheets with low load resistance when compared with ﬁber-cement or wood-
based sidings.

(3) The environmental beneﬁts and avoided landﬁll costs associated with the use of the
deve10ped recycled composites further add their marketability.

 

 

CHAPTER 7

SUMMARY AND CONCLUSIONS

Recycling in construction presents the p0tentials for high-volume use of waste materials
in products with long service life, while avoiding costly separation and puriﬁcation steps.
This research focussed on the use of ﬁbers obtained through dry processing of market-lim-
ited magazine paper (with relatively high non-cellulosic constituents) as reinforcement in
thin-Sheet cement products. Virgin cellulose ﬁber-cement composites have found applica-
tions in sidings, and sofﬁts, tile backerboard, roof tiles, fencing and a variety of commer-
cial ﬁelds where durability and ﬁre resistance of thin panels are of concern. This research
evaluated the technical feasibility and economical viability of partly replacing virgin cel-
lulose ﬁbers with wastepaper ﬁbers in thin cement products produced by the slurry-dewa-
tering technique.

The research was conducted in four phases concerned with: ( 1) identiﬁcation of the key
proportioning/processing variables in the production of wastepaper ﬁber-cement compos-
ites by the Slurry-dewatering methods; (2) optimization of the inﬂuential variables and
determination of the physical and mechanical properties of the optimized recycled waste-
paper ﬁber-cement composites; (3) assessment and improvement of the long-term durabil-
ity and moisture sensitivity of the optimized recycled wastepaper ﬁber-cement
composites; and (4) assessment of the cost-competitiveness of the optimized recycled
ﬁber-cement composites versus alternative siding materials.

Comprehensive sets of replicated experimental data were generated in this Study and were
analyzed statistically using the analysis of variance and response surface analysis tech-
niques in order to derive statistically reliable conclusions. The observations and trends
were further investigated by micro-structural studies including scanning electron micros-

copy, thcrmogravimetric analysis and x-ray diffraction techniques.

199

 

 

200

DETERMINATION OF INFLUENTIAL VARIAB LES IN THE PROCESSING OF
RECYCLED CELLULOSE FIBER-CEMENT

In the ﬁrst phase of the experimental investigation, total of 11 key variables (factors) de-

ﬁning the production process of wastepaper ﬁber-cement composites were selected; the
main intent was to distinguish those factors with statistically significant effects on the
composite material performance characteristics. These variables were: (1) recycled ﬁber
source; (2) ﬁber mass fraction; (3) ﬁber beating level; (4) substitution level of virgin ﬁbers
with recycled ones; (5) sand/binder ratio; (6) maximum particle size of sand; (7) Silica
fume/binder ratio; (8) ﬂocculating agent/binder ratio; (9) vacuum level; (10) compaction
pressure; and (11) curing condition. Each factor was considered at two levels in a (1/64)
fractional factorial design of experiments. This experimental design reveals the effects of
all variables on the composite material performance, but can not provide any information
on the possible interactions between different variables. The resulting composite, were
tested for ﬂexural performance (Strength, toughness, and initial stiﬂ'ness).

The ﬂexural test data was analyzed statistically by fractional factorial analysis of variance.
Among the eleven proportioning / processing variables considered in this study, three
(total ﬁber mass fraction, substitution level of virgin cellulose ﬁbers with recycled ones,
and ﬁber reﬁnement condition) proved to have Statistically Signiﬁcant effects, at 95% level
of conﬁdence, on the ﬂexural performance of wood ﬁber reinforced cement composites. In
order to optimize the composites, it is thus necessary to detemrine the optimum combina-
tion of these variables to produce composites with highest performance-to—cost ratios. In
the optimization process, other variables with Statistically insigniﬁcant eﬂ'ects on the end
product qualities may be ﬁxed.

The recycled wastepaper ﬁbers were also analyzed and compared with virgin cellulose
ﬁbers. The recycled ﬁbers were found to be smaller in length than virgin cellulose ﬁbers.
The surface of the recycled ﬁbers was more roughened and ﬁbrillated by the recycling
process as compared to virgin cellulose ﬁbers. Cellulose content in recycled ﬁbers was
found to be lower than virgin cellulose ﬁbers. Recycled ﬁbers had a Signiﬁcant amount
(close to 20%) of ﬁnes which are expected to play a ﬁlling role, rather than reinforcing

role, in cellulose ﬁber-cement composites.

 

 

201
OPTIMIZATION OF INFLUENTIAL VARIABLES

The inﬂuential variables in the processing of recycled wood ﬁber-cement composites were
optimized based on response surface analysis techniques. The variables optimized here
were: total ﬁber mass fraction, level of substitution of virgin ﬁbers with recycled ﬁbers,
and the beating (reﬁnement) level of ﬁbers. Optimization was based on ﬂexural strength,
initial Stiffness and toughness of the composites. Due consideration was also given in the
optimization process to the cost of raw materials. The optimized composites were then
technically evaluated versus virgin composites, ASTM speciﬁcations, and commercial

products. The conclusions derived are summarized below.

(1) Analysis of results indicated that optimum composites are obtained using 8% ﬁber
mass fraction, 50% substitution level of virgin with recycled ﬁbers, and reﬁnement (beat-
ing) of ﬁbers to a Canadian Standard Freeness (CSF) of 540.

(2) The optimized recycled wood ﬁber-cement composites were Shown to possess ﬂexural
strength, density and dimensional Stability characteristics satisfying ASTM speciﬁcations
and comparable to those of commercially available virgin wood ﬁber reinforced thin-Sheet
cement products.

(3) The optimized recycled composites produced ﬂexural strength, Stiﬂ‘ness and toughness
characteristics comparable to those of virgin composites. Compared to virgin wood ﬁber-
cement composites, the optimized recycled composites possessed somewhat lower ﬂexur—
al strength and toughness but higher initial ﬂexural Stiffness. The difference in ﬂexural
toughness and toughness were statistically signiﬁcant. Recycled composites also showed
reduced moisture (dimensional) movements, lower water absorption and moisture content,

and higher density when compared with virgin wood ﬁber-cement composites.

(4) The ﬁne content of recycled ﬁbers seem to play more of a ﬁlling role than a reinforcing
role. Hence, recycled composites present a denser microstructure which reﬂects in higher
Stiffness, lower water absorption and moisture content and reduced dimensional (mois-
ture) movements of recycled composites. Reduced reinforcing action of ﬁnes in recycled
ﬁbers, however, reﬂects in somewhat reduced ﬂexural strength and toughness of recycled

composites when compared with virgin composites.

202

DURABILITY AND MOISTURE-SENSITIVITY

The effects of moisture and accelerated ageing on the ﬂexural performance and micro-
structural characteristics of the optimized wastepaper ﬁber-cement composites and control
composites (made fully with virgin softwood kraft ﬁbers) were investigated. Microstruc-
tural studies utilized the Scanning Electron Microscopy, thermogravimetry and X-ray dif-
fraction techniques. Appropriate measures were adopted and evaluated for controlling the
ageing and moisture effects on wastepaper ﬁber-cement composites. It was concluded
that:

(l) The increase in moisture content of virgin and recycled composites reduced ﬂexural
Strength and stiffness, and increased toughness of the composites One way analysis of
variance of ﬂexural strength, Stiffness and toughness test results at different moisture con-

tents conﬁrmed the moisture sensitivity of recycled and virgin composites.

Among the moisture conditions considered (oven-dried, air-dried and saturated), satura- -
tion produced a distinct ﬂexural behavior. Air-dried and-oven dried recycled composites
proved to be statistically equivalent in ﬂexural strength but different in ﬂexural toughness
and Stiffness at 95% conﬁdence level (using multiple comparison by contrast). TWO-way
analysis of variance of results was also conducted for different composites (recycled ver-
sus virgin) and moisture conditions. The ﬁber-moisture interaction proved to be statistical-
ly signiﬁcant at 95% level of conﬁdence, indicating that moisture effects on recycled
composites differ from those on virgin composites. Damaging effects of saturation are less
pronounced in the case of recycled ﬁbers. Upon saturation, recycled composites exhibited
a drop of 37% in ﬂexural Strength and 16% in initial stiﬂ‘ness when compared with air-
dried composites; the corresponding drops in virgin composites were 47% and 28%, re-
spectively. Both recycled and virgin composites showed an increase of 32% in ﬂexural
toughness upon saturation.

(2) Repeated wetting-drying and particularly wetting-drying and carbonation cycles
caused an increase in ﬂexural Stiffness and strength of virgin and recycled composites but

led to reduced toughness and embrittlement of the materials.

Repeated wetting-drying cycles had Statistically Signiﬁcant effects on all ﬂexural proper-

ties (strength, toughness and stiffness) of recycled composites. Two-way analysis of vari-

 

 

203

ance (with two factors of recycled versus virgin ﬁbers and aged versus unaged
composites) revealed that there was a statistically signiﬁcant difference, at 95% level of
conﬁdence, between ageing effects on recycled and virgin composites as far as ﬁexural
strength and toughness are concerned; the initial Stiffness of virgin and recycled compos-
ites, however, was affected Similarly by ageing under repeated wetting-drying cycles.
Damaging effects of repeated wet-dry cycles on recycled composites (which lost 32% of
toughness upon ageing) was less than that on virgin composites (which lost 45% of tough-

ness upon ageing).

(3) Statistical analyses indicated, at 95% level of conﬁdence, that the addition of carbon-

ation to wet-dry cycles leads to pronounced effects of ageing on composites

(4) Repeated freeze thaw cycles did not have statistically signiﬁcant eﬂ'ect on ﬂexural
Strength and toughness of recycled composite; however, the effects on increasing the ini-
tial stiffness was Statistically signiﬁcant at 95% level of conﬁdence. Two way analysis of
variance (taking into account the virgin composite) revealed that freeze-thaw cycles affect
the initial Stiffness recycled composites in a way different from virgin composites (similar
effects were observed on ﬂexural strength and toughness irrespective of the composite
type). While recycled composites showed a 1% increase in stiffness under repeated freeze-
thaw cycles, virgin composites showed a drop of 9% in Stiffness under this ageing condi-
tion.

(5) The effects of long-term immersion in hot water on ﬂexural Strength, Stiffness and
toughness of recycled composites were not statistically signiﬁcant at 95% level of conﬁ-
dence. Two-way analysis of variance of results (virgin versus recycled ﬁbers) concluded
that only ﬂexural stiffness of virgin and recycled composites was affected differently by
this ageing process. While hot water immersion Slightly reduced the initial Stiffness of re-
cycled composites (by 3%) it caused a small increase of 3% in the stiffness of virgin com-

posites.

(6) Comparing the effects of different accelerated ageing tests based on statistical analy-
ses, it appears that wetting-drying and carbonation is the most effective method to bring
about changes in physical and mechanical properties of the wood ﬁber-cement compos-
ites. In order to compare the effects of different accelerated ageing conditions on the ﬂex-

ural performance of recycled and virgin ﬁber-cement composites, two-way analysis of

 

 

204

variance of ratios of aged to unaged ﬂexural Strength, Stiffness and toughness test results
were conducted. The two factors in these analyses were: composite type (recycled and vir-
gin), and accelerated ageing condition (wet-dry, wet-dry and carbonation, freeze-thaw, hot
water immersion). Results of the analysis of variance indicated that. at 95% level of conﬁ-
dence, all the ageing conditions have comparable effects on ﬂexural strength but there are
statistically signiﬁcant differences between the effects of different accelerated ageing con-
ditions on ﬂexural Stiffness and toughness. There was no statistically signiﬁcant diﬂ'erence
between virgin and recycled composites as far as ageing effects on ﬂexural Strength and
stiffness are concerned; the two composites, however, behaved differently in ageing ef-

fects on ﬂexural toughness.

Multiple comparison of results indicated that, at 95% level of conﬁdence, each of the ac-
celerated conditions have distinctly different effects on ﬂexural toughness and Stiffness,
except for the hot water immersion and freeze-thaw condition which had statistically com-
parable effects on ﬂexural Stiffness. As far as the overall ageing effects on ﬂexural perfor-
mance is concerned, repeated wetting~drying and carbonation cycles produced the most
pronounced effects whereas hot water immersion caused the least effects. The trends in the
response of virgin and recycled composites to ageing were generally comparable.

(7) Ageing effects on the morphology of ﬁbers and failure mode as observed in SEM
(comparing different accelerated ageing conditions) were Studied and it was concluded
that ﬁbers were not unaffected by the ﬁeeze-thaw and hot water immersion ageing condi-
tions. In the case of wetting-drying and carbonation, however, tendencies towards ﬁlling
of ﬁbers, densiﬁcation of interfaces and dominance of ﬁber rupture in failure mode were
observed. For the unaged composite, the dominant mode of failure was observed to be ﬁ-
ber pull out. In the case of specimens subjected to repeated cycles of freeze-thaw, the frac-
ture surface was observed to Show a combination of ﬁber pull out and ﬁber fracture. In the
case of hot water immersed composites also the mode of failure was observed to be ﬁber
pull out accompanied with ﬁber fracture. For the repeated wetting-drying ageing condition
the mode of failure at the fracture surface was dominated by ﬁber fracture. Also, for re-
peated wetting-drying and carbonation ﬁber rupture was observed to dominate the failure
mode.

In addition, the appearance of ﬁbers and their interfaces was also affected by the ageing
processes. In the case of wetting-dryin g and wetting drying with carbonation, a dense ﬁber

 

 

205

matrix interface was observed. Fibers in case of wetting-drying with carbonation appeared
to be ﬁlled with hydration and carbonation products. In the case of hot water immersion
and freeze-thaw ageing conditions the densiﬁcation of the interface and ﬁlling of ﬁber
core was not pronounced. It appears that the ageing process is most pronounced under re-

peated wetting-drying and carbonation condition.

(8) Results of thermogravimetric analysis suggested that compositional changes occur in
wood ﬁber-cement composites upon ageing. The calcium carbonate content increases and
calcium hydroxide content decreases under accelerated ageing conditions. These trends
were conﬁrmed through X-ray diffraction analysis. Wetting-Drying and carbonation fol-
lowed by wetting-drying were observed to result in most pronounced compositional
changes. These changes partly explain the brittle behavior of the composite after wetting-
drying and carbonation. Under ageing effects there seems to be a tendency in the calcium
hydroxide constituent of cement hydration products to dissolve in cement pore water and
this process is accompanied with the carbonation of calcium hydroxide which produces
calcium carbonate. The “petriﬁed" ﬁbers with strong bonding to mauix tend to be strong
but brittle. It may be hypothesized that the increase in Strength and rigidity of the petriﬁed
ﬁbers, and the increase in their bond strength due to the densiﬁcation and also elimination
of Shrinkage debonding at the interface zones account for the increase in strength and stiff-
ness of the composite. Petriﬁcation and well-bonded ﬁbers, however, tend to fracture prior
to pulling out of the matrix; this eliminate the energy absorption associated with ﬁber pull

out and thus causes embrittlement of the composite.

(9) A statistically signiﬁcant negative correlation was observed between calcium carbon-
ate content and toughness, in unaged and aged composites, the positive correlation be-
tween calcium carbonate content and stiffness were also statistically Signiﬁcant a
statically Signiﬁcant positive correlation was also observed between density and calcium
carbonate content. These correlations conﬁrmed the key role the formation of calcium car-

bonate plays in the ageing of the composites. The correlations between calcium hydroxide

content and various engineering properties of unaged and aged composites were not statis-

tically signiﬁcant. This further underlines the signiﬁcant effects of carbonation in the pro-
cess of petriﬁcation of ﬁbers upon ageing.

The correlations of calcium hydroxide content with engineering properties (Strength, stiff-

ness and toughness) were not statistically signiﬁcant. A Strong positive correlation was

 

206

observed between calcium carbonate content and density (correlation coeﬁicient: 0.941)
which was statistically signiﬁcant at 99% conﬁdence level. Calcium carbonate content
seemed to have a strong correlation than calcium hydroxide content with engineering
properties. This may be attributed to the fact that ageing causes dominantly a migration of
calcium hydroxide (and only a small change is its content as a result of carbonation) while
it actually forms calcium carbonate (which reﬂects in increased calcium carbonate con-

tent.

(10) Since formation of calcium carbonate through carbonation of calcium hydroxide as
well as the transport of calcium hydroxide to ﬁber cores and interfaces play critical roles
in ageing effects on the composites, reduction of calcium hydroxide seems to provide for a
more stable composite. Replacement of relatively high levels of cement with silica fume
presents an approach to the reduction of calcium hydroxide content. Silica fume also helps
reduce the pemreability of composites and the alkalinity of pore water. While the opti-
mized recycled composites has 10% silica fume content, the reﬁned high-silica fume com-
posite considered had 30% silica fume content.

Repeated wetting-drying and carbonation still had Statistically Signiﬁcant effects, at 95%
level of conﬁdence on ﬂexural strength and toughness (but not Stiffness) of high-silica
fume composites. Statistical analyses including results for low Silica fume composites,
however,; indicated that ageing effects on high-silica fume composites were much less
pronounced; the calcium carbonate content of aged high-Silica fume recycled composites
was also less than that of low silica fume composites. While saturation of aged high-silica
fume composites led to statistically signiﬁcant effects on ﬂexural performance, statistical
analyses indicated. at 99% level of conﬁdence, that the damaging effects of moisture on
ﬂexural strength and Stiffness were less pronounced in aged high-silica fume recycled

composites when compared with corresponding low-Silica fume composites.

In short, 30% replacement of cement with silica fume in recycled ﬁber-cement composites
was found to be highly effective in controlling the ageing mechanisms and moisture ef-
fects; this approach presents a practical, economical and efficient approach for enhancing

the durability and moisture resistance of wastepaper ﬁber-cement composites.

(11) In an alternative approach to reduce calcium hydroxide content of cement hydration

products a Special cement was considered (consisting of 8 parts of Portland cement, 3

 

 

207

parts of calcium sulphoaluminate, and one part of synthetic powder) the hydration of
which does not produce much calcium hydroxide.

While repeated wetting-drying and carbonation effects on the ﬂexural toughness and stiff-
ness (but not Strength) of composites made with Special cement were still statistically sig-
niﬁcant at 95% level of conﬁdence, it was concluded at 99% level of conﬁdence, that
ageing effects on ﬂexural Stiffness and toughness were less pronounced when special ce-

ment was used to replace regular Portland cement in recycled composites.

One -way analysis of variance (aged versus aged-saturated recycled composites with spe-
cial cement) followed by two-way analysis of variance with composite type (special ce-
ment versus regular cement) and saturation condition (aged versus aged-saturated)
indicated, at 99% level of conﬁdence, that while moisture effects on ﬂexural strength were
comparable in composites with special cement and regular cement, the corresponding ef-
fects on toughness and Stiffness were inﬂuenced by the type of cement. With special ce-

ment, moisture eﬂ'ects were less pronounced when compared with regular cement.

Thermogravimetric compositional analysis indicated that the special cement reduces the
calcium carbonate content of aged composites. This can be attributed to the reduced calci-
um hydroxide content of the special cement hydration products.

(12) The effectiveness of special cement and silica fume in recycled composites was com-
pared. It was found that silica fume is superior in performance when compared with Spe-
cial cement. analysis of variance of the results conﬁrmed that different composites (high-
silica fume, special cement, and optimum with low silica fume content) produced diﬂ‘erent
calcium carbonate contents upon ageing. The least calcium carbonate content (and thus
conceivably the most Stable performance under ageing effects) was obtained in the high-

silica fume composite.

(13) Effects of polymer dispersion and carbonation as partly of the process were unsuc-
cessfully investigated for improving the durability and moisture sensitivity of the wood ﬁ-

ber-cement composites.

COST ANALYSIS
A comprehensive cost analysis was conducted on the deve10ped high-silica fume recycled

composite versus alternative siding materials. It was concluded that:

 

 

208

(1) Recycled ﬁber-cement composites present technically desirable qualities together with
lowest initial and life-cycle costs when compared with alternative siding materials.

(2) Recycled wastepaper ﬁber-cement composites present a number of positive qualities
which make them further competitive. Their ﬁre resistance is superior to wood or vinyl
sidings. They do not fade away like aluminium sidings with time. Aluminium and vinyl
sidings are sheets with low load resistance when compared with ﬁber-cement or wood-
based sidings.

(3) The environmental beneﬁts and avoided landﬁll costs associated with the use of the
developed recycled composites further add their marketability.

APPENDIX I
THEORETICAL CONSIDERATIONS

1.1 INTRODUCTION

Theoretical illustrations of the mechanisms through which ﬁbers enhance the mechanical
properties of concrete are generally based on two concepts [85]. According to the ﬁrst one
(spacing concept) ﬁbers enhance the concrete performance mainly through limiting the
Size and preventing the propagation of the internal ﬂaws in concrete. The other one (com-
posite material concept) suggests that ﬁbers contribute to the concrete load and deforma-
tion capacities through mobilizing their pull-out resistance. More fundamental theoretical

approaches to ﬁber-cement composites based on fracture mechanics concepts.

A brief review of the spacing concept as applied to ﬁbers in cementitious matrices is pre-
sented here. The virgin cellulose ﬁbers and recycled wastepaper ﬁbers are compared based
on spatial distribution. The theoretical spacing calculated is compared with the actual
SEM observations. A brief introduction to the application of fracture mechanics to ﬁber-

cement composites is also presented.

1.2 Spacing Concept As Applied To Thin Fiber-Cement Sheets

Concrete and mortar have an inherent internally ﬂawed structure. The Strength of such a
material can be increased by increasing the fracture toughness, decreasing the Size of the
ﬂaws, and decreasing the Stress intensity factor at the tip of the internal cracks. These ob-
jectives, according to the spacing concept, can be achieved using more closely spaced ﬁe
bers in the matrix. The ﬁber Spacing in this approach is a statistical description of the
distance between centroids of ﬁbers. A popular expression for ﬁber Spacing is derived be-
low [70].

If the ﬁbers were aligned and uniformly spaced without overlapping (Figure 1.1a), then the
average Spacing of the ﬁber centroids (Sc) would be:

209

 

210

SC: (v/N)I/3 ............ Equation 1.1

where V: total volume of ﬁber reinforced composite
N: total number of ﬁbers

= Vf/V l
Vf= volume fraction of ﬁbers
V1: volume of Single ﬁber
For unit V
Sc: (1/N)"3

= (V 1/V 01/3

In actual condiu'ons, however, ﬁbers are not aligned, and they also overlap. One would as-
sume that the ratio of the average of the projected length in one direction to the total length
(119) is a proper correction for the random orientation of ﬁbers (Frgure 1. lb) [58].

119: Average of projected length/ Total length
= 0.41

Hence, the effective volume fraction of ﬁbers is only 41% that of the actual volume frac-
tion, and the average ﬁber spacing after correction for random orientation of ﬁbers (Sm)

becomes:

scc = (V/0.41N)V3' ....................... Equation 1.3

= (v,/0.41 val/3

 

211

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l l I l l
__ ii
.4“ L F791 I ‘l. I
l
Llii-i I l l r— 1 I l
r T '3 f
I I I I I I I isc Piojecged 1fI I
A . engt L f
1 i l l l I l 8 l
Li I i I l I l i
- - ‘9 X ST— 1 I
L l l
. a... . I :'
e'- ‘ E 1
SC. I e e—IiLsg e
e e I e e
A-A
(a) Aligned. Uniformly Spread (b) Projecuon of Randomly (c) Overlapping
fibers Uithout Overlapping Oriented Fibers

Figure 1.1 Spacing of Fibers.

 

 

 

 

 

 

 

212

Overlapping of ﬁbers is another factor that modiﬁes the simple spacing expressions de-
rived above. This factor is especially important for ﬁbers with lengths greater than the av-
erage ﬁber spacing (If > S). Such ﬁbers tend to overlap with other ﬁbers, thus increasing
the number of ﬁbers crossing any section by a factor lI/S:

Ne: N (lp’S)

By substituting N in Eq. 1.3 with Ne, one gets [56]:
8= (V/<o.41*N*aI/S)» 13

Hence,

3: (V/(o.41*N*1f) 1(7- ............................. Equation 1.4
where: Ne: effective number of ﬁbers,

8: average spacing of ﬁber centroids in the composite reinforced with short, randomly
dispersed ﬁbers.

As the above measure of ﬁber spacing (S) decreases, the ﬂaw size and the stress intensity
factor at the tip of internal cracks decrease and the fracture toughness increases. Hence,
the ﬁrst crack and ultimate ﬁexural strengths of ﬁber reinforced concrete are expected to
be decreasing functions of ﬁber spacing. According to the spacing concept, there is a
unique relationship between tensile strength of ﬁber reinforced concrete and ﬁber spacing
irrespective of ﬁber geometry, interfacial bond properties and tensile strength. This func-
tion has been derived empirically by number of researchers, as typically shown in Figure
1.2 [56, 79, 86].

 

 

213

 

M IAIIU

SIIDNLIN IATIO

Sillﬂhl" IAIIU

 

S!
l
c
f.
D
D
o
I

 

. C 5 ..G 0 3.) ..O C 0.1 0.; C.J
Fill! SPACI'C. ll. FlIEI SPACINC. 1!. FIIEI SPACISC. KS.

(a) Ultimate Tensile Strength (b) first Creel Strength (cl First Crack Strength

Figure 1.2 Empirical Relationship Between the Tersile Strength of Steel Fiber Rein-
forced Concrete and the Spacing of the Fibers [89]

L3 Virgin Cellulose Vs. Recycled Wastepaper Fibers

Virgin cellulose ﬁbers are on the average 3 mm (0. 12 in) long with a negligible amount of
ﬁne fragments as shown in Table 3.4 of Chapter 3. Recycled cellulose ﬁbers are shorter in
length (average length of 1 mm, 0.04 in, for the dry—processed wastepaper ﬁbers consid-
ered in this investigation) due to the recycling process but are larger in number. This re-

sults in a situation of ﬁber reinforcement conditions shown in Figure 1.3.

The shorter recycled ﬁbers are more closely spaced than the longer virgin ﬁbers.

 

 

214

Vigil

A VIII! Cellula- ﬁber
BrlhnaCnd

 

 

 

 

 

 

Figure [.3 Virgin Vs. Recycled Cellulose Fibers Arresting Cracks.

The ﬁber-cement production technique tends to orient ﬁbers in the plane of the thin-sheet
cement product. The two dimensional orientation of ﬁbers requires the modiﬁcation of

Equation 1.4 as follows

8: (V/(0.63 *N*l{)m ............................ Equation 1.5

On average there are three times more recycled ﬁbers (N=Vf/Vl) when compared with
virgin ﬁbers for the same ﬁber volume fraction (see Figure 1.3). Mth an average diameter
of 20 pm, the average spacing of recycled and virgin ﬁbers were 0.07 mm and 0.21 mm
(0.003 in to 0.008 in) respectively, as obtained from Equation 1.5.

SEM observations generally show that the inter ﬁber spacing varies from about 0.05 mm
to 0.3 mm (0.002 in to 0.012 in).

Obviously the trends observed in experiments on virgin and recycled ﬁber-cement com-

 

 

215

posites can not be illustrated with the spacing concept. Broken recycled ﬁbers containing
ﬁnes are more closely spaced than virgin cellulose ﬁbers; this closer spacing, however,
does not lead to increased strength of the composites with recycled ﬁbers. This could be il-
lustrated as follows: (I) the recycling process not only shortens the ﬁbers but also damage
their tensile strength and overall qualities; (2) ﬁnes in recycled ﬁbers do not have the slen-
derness needed to act as ﬁbers and they play more of ﬁlling role; (3) the ﬁbers used in
magazine paper even in the original form may not have been as high-quality as the virgin
kraft cellulose ﬁbers used in this investigation; and (4) the spacing factor which disregards
the effects of ﬁber-to-matrix bond and the tensile strength of ﬁbers can not present a com-
prehensive approach for predicting the strength characteristics of wood ﬁber-cement com-

posues.

1.4 Fracture Mechanics

Fiber reinforced cementitious matrices form a class of composites with the following dis-
tinct characteristics: the ultimate tensile strain of the matrix is much lower than the ulti-
mate tensile strain of the ﬁber, and the bond strength at the ﬁber-matrix interface is
relatively small; this implies that with increasing tensile stresses on the composite, the ma-
trix will crack at some level of stress long before the ﬁbers reach their failure strength.
Furthermore, the composite response after matrix cracking will greatly depend on the in-

teraction of ﬁbers with cracks and the pullout resistance of ﬁbers [87].

The classical fracture mechanics concept of brittle crack propagation in materials has been
applied only relatively recently to cementitious matrices such as paste, mortar and con-
crete. In 1963 Romualdi and Batson[57] proposed a model to explain how closely spaced
discontinuous stiff ﬁbers can increase the stress causing rapid crack propagation. They es-
sentially applied to ﬁber reinforced concrete, the classical Grifﬁth approach as modiﬁed

by Irwin and described by
0c: KIC/(anaIVZ ...................................... Equation 1.6
where

0C: stress on onset of rapid crack propagation

216

ch= critical plane strain fracture toughness (opening mode I)

a: parameter depending on the specimen and crack geometry, loading rate, crack tip

sharpness, etc., and

a= half-length of a penny-shaped internal crack or length of external crack.

The expression proposed by Romualdi and Batson [57] for ﬁber reinforced concrete is
given by

 

6C: K/(S)lﬂ ........... - --- Fquatron 1.7

where

K: constant related to the fracture toughness and

S: average ﬁber spacing

Applying of this equation to virgin cellulose ﬁber-cement composites gives
oc= K/(S)m

= K/(.21)m

=0.91K

For virgin cellulose ﬁbers the stress to start crack propagation is 0.577 of critical stress in-

tensity factor related constant.

For recycled cellulose ﬁbers
0c: K/(s)ln

= K/(omWZ

=3.78K

So in case of recycled cellulose ﬁbers, this approach indicates that a higher value of stress
is required to start rapid crack propagation. Again, the disregard for the bond strength and
ﬁber tensile strength effects in this approach leads to the erroneous conclusion and that

shorter ﬁbers are more effective in increasing the tensile strength of composites than that

 

 

217

of longer virgin cellulose ﬁbers.

The effect of ﬁber debonding in the classical fracture mechanics relation has been tackled
by Naarnan [87]. it was shown, that due to the process of debonding in ﬁber reinforced
concrete, a pseudo-plastic zone of radius R (Figure 1.4) develops at the crack tip. The
pseudo-plastic zone represents an area where the matrix is cracked but where the ﬁbers
bridging the cracked surfaces are still debondin g, thus providing a pullout resistance. This
increases the assumed critical internal ﬂaw size for the composite by 2R. Consequently,
they proposed the following relation which applies in the three-dimensional case of a ten-

sile prism

0c: KIc/(4*(d+R)/1t)1/2

where
ch= fracture toughness at critical propagation (opening Mode 1)

d= crack length; or diameter of largest weak area deﬁned as that area having no or little

amount of ﬁber crossings by probability analysis, and

R: length of radius of pseudo-plastic zone from crack tip.

So the spacing concept and fracture concept applications were not successful because the

composite did not satisfy the assumptions made in the both approaches.

218

 

 

 

 

00

It
‘PSIUOOPLASYIC long)

"I
l
l

 

 

   

I I I I I‘ l I
(b) I I I
a I. I l x:
I I I. i do}!
I q" I III
: 91.1/I
u I / '
I 1""?! II
. I :II’ I g,

 

 

Figure 1.4 Assumed Critical Crack Model Controlling Fracture of Fiber-Reinforced

Concrete Composite.

 

APPENDIX II

NOTATIONS AND SYMBOLS

The following symbols were used in this study:

AD Air Dried
OD Oven Dried

Sat Saturated

WD Wetting-Drying

WDC Wetting-Drying and Carbonation
FT Freeze-thaw

HW hot water

SF Silica fume

CSF Canadian Standard Freeness

JCI-SF Japanese Concrete Institute Standards for Test Methods of Fiber Reinforced Con-
crete

RH Relative Humidity

SEM Scanning Electron Micrographs

FA Flocculating Agent

SSK Southern Softwood Kraft

SD Standard Deviation

TAPPI Technical Association of the Pulp and Paper Industry
TGA Thermogravimetric Analysis

EPA Environmental Protection Agency (USA)
psi Pounds per square inch

lb in Pound inch

K/rn Kips per inch

mm millimeter

N Newton

CD Cross Direction

MD Machine Direction

0 stress

5 strain

A displacement

I 95% Conﬁdence Interval (CI)

SC Special cement

219

APPENDIX III

STANDARD SPECIFICATIONS

The following standard speciﬁcations were used in this study:

1. ASTM C 1185 Standard Test Methods for Sampling and Testing Non-Asbestos Fiber-
Cement Flat Sheet, Rooﬁng and Siding Shingles, and Clapboards (American Society of
Testing Materials)

2. ASTM C1186-91, “Standard Speciﬁcation for Flat Non-Asbestos Fiber-Cement Sheets”

3. JCI SF 4 Method of Tests for Flexural Strength and Flexural Toughness of Fiber-Rein-
forced Concrete (Japan Concrete Institute)

4. TAPP1207 Technical Solubility of Pulp (Technical Association of the Pulp and Paper
Industry)

5. ASTM C670-91a, “Standard Practice for Preparing Precision and Bias Statement for
Test Methods for Construction Materials,” (American Society of Testing Materials)

220

BIBLIOGRAPHY

1. Congress of the United States (1989), “Facing America’s Trash: What Next for Munici-
pal Solid Waste”, Report OTA-0-424, Ofﬁce of Technology Assessment, Washington,
DC, October 1989, pp. 377.

2. Michigan Department of Natural Resources (1986), “Waste stream assessment”, Waste
Management Division, Resource Recovery Section, June.

3. NSWMA (National Solid Waste Management Association) 1989, “Landﬁll Capacity in
the Year 2000”, Special Report. Washington, DC, pp. 5.

4. Michigan Department of Natural Resources (1988), “Michigan Solid Waste Policy”,
Waste Management Division, Resource Recovery Section, June, 38 pp.

5. Tappi Journal (1990), “API survey shows U.S. paper industry plans expansion and in-
creased wastepaper use“, Vol. 73, No. 3, March, p. 20.

6. Michigan DNR (1987), ”State Mde Market Study for Recyclable Paper, Glass
and Metals”, Feb.

7. American Paper Institute (1990), “Facts About Wastepaper Recycling,” 4 pp.

8. Lepoutre, P. (1989), “The Structure of Paper Coatings: An Update,” Progress in Organic
Coatings, Vol 17, pp. 89-106.

9. Sahani, Chandra .l. & Mlson, William K. (1987), “Preservation of Libraries and Ar-
chives,” American Scientist, Vol. 75, No. 3, May-June, pp. 240-51.

10. Franklin Associates, Ltd. (1987), “Statewide market study for recyclable paper, glass
and metals”, Background Report presented to Michigan Department of Natural Resources,

Community Assistance Division, Resource Recovery Section, February.

I I. Goldgerg, D. (1989), “Markets beyond the paper mill”, Waste Age, December, pp. 37-
40.

12. Apotheker, S. (1990), “Mixed Reviews for Mixed Paper”, Resource Recycling, Janu-
ary, pp. 26—30.

13. Thomas. C. 0., Thomas. R. C. and Hover, K. C. (1987), “Wastepaper Fibers in Ce-

mentitious Composites". Journal of Environmental Engineering, American Society of

221

222

Civil Engineers, Vol. 113. No. 1, February, pp. 16-31.

14. Darcey. S. (1988), “Open Markets Welcome Recyclables”, World Wastes, August, pp.
4()-44.

15. Tye. R.P., Ashare, E., Guyer, EC. and Sharon, AC. (1978), “An Assessment of Ther-
mal Insulation Materials for Building Applications”, ASTM Special Technical Publication
718, pp. 9-26.

16. Coutts, R.S.P., “Sticks and Stones”, Forest Products Newsletter, SCSIRO Division of
Chemical and Wood Technology, Vol. 2, No. 1, January 1988, pp. 1-4.

17. Coutts, R.S.P. and Warden, PG. (1990), “Effect of Compaction on the Properties of
Air-Cured Wood Fiber Reinforced Cement,” Cement and Concrete Composites 12, pp.
151-156.

18. Vinson. Kenneth D., and Daniel, James I. (1989), “Advances in the Development of
Speciality Cellulose Fibers Speciﬁcally Designed for the Reinforcement of Cement Matri-
ces, Early Strength of Composites Based upon Ordinary Portland Cement”, Paper present-
ed at the 1989 Annual Convention, American Concrete Institute, Atlanta, Georgia,

February 19-24.
19. Soroushian, P., Suchsland, O. and Murikunte, S. (1990), “Cellulose Fiber Reinforced

Cement Composites”, Proceedings, MSU Concrete Technology Seminar 111, Engineering

Lifelong Education Program, Michigan State University, February, pp. 11.1-1 1.46.

20. Eterline, High Performance Fiber Reinforced Cement Panels, Eternit, Inc., Blandon,
Pennsylvania.
21. Matsuo, R., Kobayashi, Y., Kitamura, T., Kato, S., Hiragima, M., and Kago, M.

(1982), “Research on Manufacture of Pulp Cement Board Using Recovered Pulp From
Municipal Solid Waste for Reinforcement (Part I): Continuous Formation of Pulp Cement
Boards by Cylinder Machine”, Japan Tappi, Vol. 36, No. 7, July pp. 733-737.

22. Hiragima, M., Kago, M., Kobayashi, Y and Matsuo, R. (1982), “Research on Manu-
facture of Pulp Cement Board Using Recovered Pulp From Municipal Solid Waste for Re-
inforcement (Part II): Batchwise Formation of Pulp Cement Boards”, Japan Tappi, Vol.

36, No. 5, May, pp. 557-564.

23. Coutts, R.S.P. (1983), “Wood Fibers In Inorganic Matrices,” Chemistry In Australia,

223

Vol. 50, No. 5. May. pp. 143-148.
24. Kocurek, M.J. and Stevens, C.F.B. (1983), “Pulp and Paper Manufacture, Vol. 1; Prop-

erties of Fibrous Raw Materials and their Preparation for Pulping,” Joint Text Book Com-
mittee of the Paper Industry, Atlanta, Georgia, pp. 182.

25. lnterﬁbe Speciﬁcations (1990), “Modiﬁed Cellulose Fibers - Spec“. 185,” lnterﬁbe Cor-
poration, Portage, Michigan, pp. 1—4.

26. Coutts, R. S. P. (1989), “Wastepaper ﬁbers in cement products”, The International
Journal of Cement Composites and Lightweight Concrete, Vol. 11, No. 3, August, pp. 143-
147.

27. Coutts, R.S.P. J. Composites (1984), Vol. 15, No. 2, April, pp. 139-43

28. Coutts, R.S.P. and PG. Warden, J.(l984), The International Journal of Cement Com-
posites and Lightweight Concrete, Vol. 9, Number 2, pp. 69-73.

29. Coutts. R.S.P. (1987), Journal of Materials Science Letters, Vol. 6, No. 8, August, pp.
955-7.

30. Coutts, R.S.P. (1986), “High Yield Wood Pulps as Reinforcement for Cement Prod-
ucts”, Appita, Vol. 39, No. I, January,pp. 31-5.

31. Gram, HE, (1983) “Durability of Natural Fibers In Concrete,” Swedish Cement and

Concrete Research Institute, Stockholm, 255 pp.

32. Sharman, W.R. and Vautier, BR (1986), “Accelerated Durability Testing of Autoclaved
Wood-Fiber-Reinforced Cement-Sheet Composites,” Durability of Building Materi-
als, Vol. 3,pp. 255-275.

33. Sharman, W.R. (1983), “Durability of Fiber-Concrete Sheet Claddings,” New Zealand

Concrete Construction, Aug, pp. 3-7.

34. Akers, S.A.S and Studinka, J .B. (1989), “Ageing Behavior of Cellulose Fiber Cement
Composites in Natural Weathering and Accelerated Tests,” The International Journal

of Cement Composites and Light Weight Concrete, Vol. 11, No. 2, May, pp. 93-97.

35. Bentur, A. and Akers, S.A.S. (1989), “The Microstructure and Ageing of Cellulose
Fiber Reinforced Cement Composites Cured in a Normal Environment,” The Interna-

tional Journal of Cement Composites and Light Weight Concrete, Vol. 11, No. 2, May

 

36.

39.

40.

41.

43.

44

224

, pp. 99-109.

Bentur, A. and Akers, S.A.S. (1989), “The Microstructure and Ageing of Cellulose
Fiber Reinforced Autoclaved Cement Composites Cured in a Normal Environment,”
The International Journal of Cement Composites and Light Weight Concrete, Vol. 11,
No.2, May, pp. Ill-115.

Akers, S.A.S, Crawford, D., Schultest, K. and Gemeka, DA. (1989), “Micromechan-
ical Studies of Fresh and Weathered Fiber Cement Composites, Part 1: Dry Testing,”
The International Journal of Cement Composites and Lightweight Concrete, Vol. 11,
No. 2, May, pp. 117-124.

. Tait, RB. and Akers, S.A.S. (1989) , “Micromechanical Studies of Fresh and Weath-

ered Fiber Cement Composites, Part 2: Wet Testing,” The International Journal of
Cement Composites and Lightweight Concrete, Vol. 11, No. 2, May, pp. 125-131.

Pririe B.J., Glasser, F.P., Schmitt-Henco, C. and Akers, S.A.S. (1990), “Durability
Studies and Characterization of the Matrix and Fiber-Cement Interface of Asbestos-
Free Fiber-Cement Products, Cement and Concrete Composites, Vol. 12, pp. 233-244.

Radjy, F.F., Sellevold, E.J., Moell, SM. and Danielssen, T. (1986), “Use of Microsilica
Additives in Asbestos Free Fiber Reinforced Cements,” Proceedings, Third RILEM
International Symposium on Developments in Fiber Reinforced Cement and Concrete,

Vol. 1, Jul.

Bentur, A. (1989), “Silica Fume Treatments as Means for Improving Durability of
Glass Fiber Reinforced Cements,” Journals of Materials in Civil Engineering, Vol. I,
No. 3, Aug., pp. 167-183.

. Mansur, MA. and Aziz, M.A. (1982)“A Study ofJute Fiber Reinforced Cement Com—

posites,” International Journal of Cement Composites and Lightweight Concrete, Vol.

4. No. 2, pp. 75-82.

Harper, 3., (1982)“Developing Asbestos-Free Calcium Silicate Building Boards,”
Composites, Apr., pp. 123-128.

. Elfgren. L. (1989). “Fracture Mechanics of Concrete Structures”, Report of the Tech

nical Committee 90-FMA Fracture Mechanics to Concrete - Applications. pp. 1-15.

 

45. Grifﬁth, A.A., “The Phenomenon of Rupture and Flow in Solids,” Philodophical
Transactions, A, Royal Society of London, 22, pp. 163-198.

46. Neville, AM. (1959), “Some Aspects of the Strength of Concrete,” Civil Engineering,
London, vol.54, pp. 1153-1156.

47. Kaplan, F.M., “Crack Propagation and the Fracture of Concrete,” J. American Con-

crete Institute, 58, pp. 591-610.

48. Mindess, S., “The Application of Fracture Mechanics to Cement and Concrete: A His-
torical Review,” Fracture Mechanics of Concrete, (ed. F.H. Wittman), elsevier, Amster-

dam, pp. 1-30.

49. Vmean, F.H., editor (1983), “Fracture Mechanics of Concrete,” Elsevier, Amster-
dam, 8+680, (ISBN 0-444-42199-8).

50. Shah, S.P. editor (1985), “Application of Fracture Mechanics to Cementitious Com-
posites,” NATO ASI Series, Serie E, 714, (ISBN 90-247-3176-3).

51. Carpinteri, A. and lngraffea A.R. editors, (1984), “Fracture Mechanics of Concrete:
Material Characterization and Testing,” Martinus Nijhoff, The Hogue, 202, (ISBN 90-
247-2959-9).

52. Sih, GO and DiTommaso, A., editors, “Fracture Mechanics of Concrete: Structural
Applications Stang, H., and Shah, SP. (1985), “Fracture Mechanical Interpretation of the
Fiber/Matrix Debonding Process in Cementitious Composites,” Proc. International Con-

ference on Fracture Mechanics of Concrete, Laussane, Switzerland, Oct.

53. Reinhardt, H.W. (1986), “The Role of Fracture Mechanics in Rational Rules for Con-
crete Desi gn,” IABSE PERIODICA, No. 1, IABSE Surveys, s-34/86, Feb. 15,.

54. Ewalds. H.L., smf Wanhill, R.J.H. (I984), “Fracture Mechanics,”, Edward Arnold,

London.

55. Bazant, Z. P. (1985), “Mechanics of Fracture and Progressive Cracking in Concrete
Structures”, Fracture Mechanics of Concrete: Structural Application and Numerical Cal-

culation. pp. 1-5.

56. Bentur, Arnon and Mindess, Sidney (1990), “Fiber Reinforced Cementitious Compos-
ites.” Elsevier Science Publishers Limited. pp. 99-109.

 

”>06

A—‘-

57. Romualdi, JP, and Batson, B.V. (1963), “Fiber Reinforced Concrete Properties”,
Journal of Engineering Mechanics division, ASCE, Vol. 89, No. EM3, June, pp. 147-168.

58. Romualdi, JP, and Mandel, J.A. (1964), “Tensile Strength of Concrete Affected by
Uniformly Distributed and Closely Spaced Short Lengths of Wire Reinforcement,” ACI
Journal, Vol. 61, June, pp. 657-670.

59. Krenchel, H., (1975), “Fiber Spacing and Speciﬁc Fiber Surface,”. In Fiber Reinforced
Cement and Concrete, ed. A. Neville. Proc. RILEM Symp.. The Construction Press, UK,
pp.69-79.

60. Edginton, J., Hannant, DJ. & \Vrlliams, R.l.T. (1974), “Steel Fiber Reinforced Con-
crete.” Current Paper CP69/‘74, Building Research Establishment. UK, 17 pp.

61. Swamy, R.N., Mangat, P.S. & Rao, C.V.S.K. (1974), “The Mechanics of Fiber Rein-
forcement of Cement Matrices.” Fiber Reinforced Concrete, ACI SP-44, ACI, Detroit, pp.
1-28.

62. Aveston, A., Cooper, G.A. & Kelly, A. (1971), “Single and Multiple Fracture,” In The
Properties of Fiber Composites, Proc. Conf. National Physical Laboratories, IPC, science
and technology Press, UK, 1971, pp. 15-24.

63. Bentur, A. & Diamond, S., (1985), “Crack Patterns in Steel Fiber Reinforced Cement
Paste,” Materials and Structures, Vol 18, pp. 49-56.

64. Wecharanta, M. and Shah, SP. (1983), “A Model for Predicting Fracture Resistance of
Fiber Reinforced Concrete,” Cement and Concrete Research, Vol. 13, No. 6, Nov., pp.
819-829.

65. Jenq, Y.S., and Shah, SP. (1986), “Crack Propagation in Fiber Reinforced Concrete,”

Journal of Structural Division, ASCE, Jan.

66. Bazant, Z.P. (I985), “Mechanics of Fracture and Progressive Cracking in Concrete
Structures,” Fracture Mechanics of Concrete, ed. G.C. Sih & A. DiTommaso. Martinus
Nijhoff Publishers, The Netherlands, 1985, pp. 1-94.

67. Hillerborg, A. (1980), “Analysis of Fracture Mechanics by Means of the Fictitious
Crack Model. Particularly of Fiber Reinforced Concrete,” International Journal of Cement

Composites. Vol. 2. N0. 4, Nov., pp. 177-184.

68. Majumdar. A.J., Shah. S.P. & Jenq. Y.S. (I988), “Fracture Processes in Fiber Rein-

227

forced Cement Sheets.” Applications of Fracture Mechanics to Cementitious Composites,

ed. S.P. Shah. Martinus Nijhoff Publishers, pp. 157-85.

69. Outwater, J.O. & Murphy, M.H. (1969), “On the Fracture Energy of Unidirectional
Laminates,” Proc. 26th Annual conf. on Reinforced Plastics, Composites Division of So-

ciety of Plastic Industry, Paper ll-c-l, pp. 1-8.

70.Morrison, J.K.. Shah, S.P. & Jenq, Y.S. (1988), “Analysis of Fiber Debonding And Pull
out In Composites,” J. Eng. Mech. Div., ASCE, 114, pp. 277-94.

71. Andonian, R., Mai, Y.W., and Cotterell, B. (1979), “Strength and Fracture Properties
of Cellulose Fiber Reinforced Cement Composites,” International Journal of Cement

Composites, Vol. 1, No. 3, pp. 151-158.
72. Das Gupta. N.C., Paramasivam, P. and Lee, S.L., (1978), “Mechanical Properties of

Coir Reinforced Cement Paste Composites,” Int. J. Housing Science and Its Applications,

2. pp. 391-406.

73. Mai, Y.W. and Hakeem. MI. (1984), “Slow Crack Growth in Bleached Cellulose Fiber
Cements,” J. Mat. Sci.Letters 3, pp. 127-30.

74. Mindess, S. and Bentur, A. (1982), “The Fracture of Wood Fiber Reinforced Cement,”
Int. J. Cem. Comp. and Ltwt. Concr., pp. 245-50.

75. Hughes, DC. and Hannant, DJ. (1985), “Reinforcement of Grifﬁth Flaws in Cellulose
Reinforced Cement Composites,” J. Mat. Sci. Letters, 4 pp. 101-2.

76. Taylor, H.F.W. (I964), “The Chemistry Of Cements,” Published by Academic Press
Inc. (London) Ltd, pp. 423-429.

77. JCI SF 4 Method of Tests for Flexural Strength and Flexural Toughness of Fiber-Rein-
forced Concrete (Japan Concrete Institute) 9

78. ASTM C1186-91, “Standard Speciﬁcation for Flat Non-Asbestos Fiber-Cement
Sheets”

79. TAPPI207 Technical Solubility of Pulp (Technical Association of the Pulp and Paper
Industry)

80. Eamst, Charles M (1987), “The Modern Thermogravimetric Approach to the Compo-

sitional Analysis of Materials.” ASTM STP 997, Compositional Analysis by Thermo-
gravimetry. pp. 1-18.

81. Bhatty. Javed I. Reid, K.J., Dellimore, D., Gamlen, G.A.G., Mangabhai, R.J., Rogers,

 

PF. and Shah, Tahir H. (1987), “The Derivation of Kinetic Parameters in Analysis of Port-
Iand Cement for Portlandite and Carbonate by Thermogravimetry,” ASTM STP997, pp.
204-215.

82. Papadakis, V.G., Fardis. MN. and Vayenas, CG. (1992), “Hyaration and Carbonation
of Pozzolanic Cements,” ACI Materials Journal, v.89, N0. 2, Masr-April, pp. 119-130.

83. Mehta, P. Kumar (1986), “Concrete: Structure, Properties, and Materials”.

84. Henry J. Molloy & Associates, Inc., GRFC Products and Services, PO. Box 515,
Hutchins, Texas 75141

85. Soroushian. R, Lee, C. and Bayasi, Z. (1989), “Fiber Reinforced Concrete: Theoretical
Concepts and Structural Design,” Concrete Technology Seminar 1989, Michigan State
University.

86 Shah, SP. ‘and Rangan, B.V. (197]), “Fiber Reinforced Concrete Properties,” ACI
Journal, February, pp. 126-135.

87. Naaman, A.E. and Shah, SP. 1978), “Fracture and Multiple Cracking of Cementitious
Composites,” ASTM Technical Publication 678, Fracture Mechanics Applied to Brittle

Materials.