Citation
Sustainable and Resilient Earthen Masonry Systems

Material Information

Title:
Sustainable and Resilient Earthen Masonry Systems Enhancing Strength Properties and Flexural Performance using Polypropylene Fibers
Creator:
Donkor, Peter
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (167 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Design, Construction, and Planning
Design, Construction and Planning
Committee Chair:
OBONYO,ESTHER ADHIAMBO
Committee Co-Chair:
KIBERT,CHARLES JOSEPH
Committee Members:
SULLIVAN,JAMES G
FERRARO,CHRISTOPHER CHARLES
Graduation Date:
12/19/2014

Subjects

Subjects / Keywords:
Adobe ( jstor )
Cements ( jstor )
Compressive strength ( jstor )
Construction materials ( jstor )
Flexural strength ( jstor )
Masonry ( jstor )
Matrices ( jstor )
Mortars ( jstor )
Soils ( jstor )
Structural deflection ( jstor )
Design, Construction and Planning -- Dissertations, Academic -- UF
fibers
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Design, Construction, and Planning thesis, Ph.D.

Notes

Abstract:
The use of affordable, sustainable, locally available, and damage resilient construction materials are key to addressing the global need for adequate housing. Compressed earth blocks (CEB); a modern form of the adobe brick are gaining popularity as a construction material globally because they are stronger and more dimensionally stable compared to earlier forms of earthen construction materials. Despite the performance improvement achieved through using CEBs, they are still very brittle and of lower strength in comparison to mainstream walling materials like concrete masonry units (CMU) and fired bricks. CEBs and other earthen masonry systems are weak in tension and exhibit low resistance to bending. This research investigated the potential of addressing some of the shortcomings of earthen masonry systems by assessing the influence of cross-linked and embossed macro polypropylene fibers on the strength, deformability, and general flexural performance of CEBs. Unreinforced and fiber-reinforced CEBs and short flexural beams were produced with fiber mass dosages of 0.2, 0.4, 0.6, 0.8, and 1.0%. Laboratory tests of material properties were determined. Properties determined included compressive strength, 3-point bending strength, first-peak strength, peak strength, residual strength, and flexural toughness. Earthen mortar was produced and used to cast two-block and seven-block CEB masonry prisms. Compressive strength and flexural bond strength of the masonry prisms were determined to evaluate compatibility of the earthen mortar with the fiber-reinforced CEBs. There was a general improvement in flexural performance and ductility of the fiber-reinforced matrices as evidenced by the load-deflection behavior, equivalent flexural strength, residual strength, and flexural toughness. Relationships between fiber quantity and enhancements in tested mechanical properties were observed and predictive models for compressive strength and equivalent flexural strength proposed. An observation of fractured surfaces after flexural strength testing using scanning electron microscopy (SEM) showed both fiber fracture and pullout; an indication of good fiber-matrix bonding. The earthen mortar was deemed compatible with the fiber-reinforced CEBs based on prism compressive strength, flexural bond strength, and failure mode. The results of this research show that when carefully designed and produced, polypropylene fiber-reinforced CEBs can be used to construct CEB masonry with improved ductility, deformability, and flexural performance. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2014.
Local:
Adviser: OBONYO,ESTHER ADHIAMBO.
Local:
Co-adviser: KIBERT,CHARLES JOSEPH.
Statement of Responsibility:
by Peter Donkor.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Donkor, Peter. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
974007339 ( OCLC )
Classification:
LD1780 2014 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

SUSTAINABLE AND RESILIENT EARTHEN MASONRY SYSTEMS : ENHANCING STRENGTH PROPERTIES AND FLEXURAL PERFORMANCE USING POLYPROPYLENE FIBERS By PETER DONKOR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLO RIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014

PAGE 2

© 2014 Peter Donkor

PAGE 3

To my father , Lt. Col. J. K. Donkor (Ret.) , for his unflin ching suppor t and belief in me

PAGE 4

4 ACKNOWLEDGMENTS I am eternally grateful to my Lord and Savior Jesus Christ without whose love and grace none of this would be possible. I would like to extend my profound gratitude to my advisor and committee chair, Dr. Esther Obonyo for her support, advice, and direction through this research process. As my mentor, Dr. Obonyo has provided academic and professional guidance throughout my doctoral studies. Most of all, I am grateful to Dr. Obonyo for giving me this opportunity. I am a lso grateful to Dr. Charles Kibert, my Co Chair, for his support. I always came out of my short meetings with Dr. Kibert with new and exciting ideas to explore. These short meetings were filled with nuggets of wisdom and I am grateful for that. I am gratef ul to Dr. Christopher Ferraro for his support during this research process. Dr. time out of his busy schedule to accommodate me, listen to my ideas, and offer insightful di rections. I thank Dr. James Sullivan for his guidance and support. Dr. replicability of the research process helped to refine the research. Special thanks to Dr. Abigai l Osei Asamoah for her support and critique. Dr. Osei Asamoah served as a second pair of eyes with the statistical analysis done in this research. Thanks to Dr. Charles Nmai of BASF Corporation for the supply of the fibers and insightful suggestions on imp roving upon this research. I will like to acknowledge the help and support of my co lleagues; Felicity Tackie Otoo, Malarvizhi Baskaran , Joel Wao, and Shirley Morque who assisted with this research in numerous ways. My sincere thanks go to Richard DeLorenz o, Patrick Carlton, and other employees of the Florida Department of Transportation (FDOT) Materials Lab for their help in specimen

PAGE 5

5 preparation and testing. The National Science Foundation (NSF) prov ided funding for this research [ NSF p roject No . CMMI 1131 175 (I nvestigators: E. Obonyo , University of Florida ; F. Matta, University of South Ca rolina; E. Erdogmus and A. Schwer , University of Nebraska Lincoln ) ] . Finally, a special appreciation goes to my mother Theresa Donkor and my sisters; Rita, Tina, and Beat rice who have always supported and encouraged me.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 Background ................................ ................................ ................................ ............. 16 Objectives, Scope, Limitations, and Hypotheses ................................ .................... 22 Objectives ................................ ................................ ................................ ......... 22 Scope ................................ ................................ ................................ ............... 22 Limitations ................................ ................................ ................................ ........ 23 Hypotheses ................................ ................................ ................................ ...... 24 Research Approach ................................ ................................ ................................ 24 Outline of Dissertation ................................ ................................ ............................. 26 2 LIT ERATURE REVIEW ................................ ................................ .......................... 27 Overview ................................ ................................ ................................ ................. 27 Existing Earthen Construction Systems/Techniques Reviewed .............................. 28 Adobe ................................ ................................ ................................ ............... 29 Cob ................................ ................................ ................................ ................... 30 Rammed Earth ................................ ................................ ................................ . 31 Compressed Earth Blocks (CEB) ................................ ................................ ............ 32 Advantages of CEBs ................................ ................................ ........................ 34 Disadvantages of CEBs ................................ ................................ .................... 35 CEB Production ................................ ................................ ................................ ...... 37 Stabilization ................................ ................................ ................................ ............ 39 Cement Stabilization ................................ ................................ ........................ 41 Lime S tabilization ................................ ................................ ............................. 44 Fiber Reinforced CEBs ................................ ................................ ........................... 46 Fiber Effects on Mechanical Properties ................................ ................................ .. 48 Compressive Strength ................................ ................................ ...................... 49 Flexural Strength ................................ ................................ .............................. 55 Mortar for CEB Masonry Systems ................................ ................................ .......... 57 CEB Masonry Systems ................................ ................................ ........................... 60 Chapter Summary ................................ ................................ ................................ ... 64 3 METHODOLOGY ................................ ................................ ................................ ... 66 Optimization of Polypropylene Fiber Reinforced CEBs ................................ .......... 66

PAGE 7

7 Microstructure Analysis ................................ ................................ ........................... 68 Characterization of Constitu ent Materials ................................ ............................... 68 Cement ................................ ................................ ................................ ............. 68 Soil ................................ ................................ ................................ ................... 68 Liquid limit ................................ ................................ ................................ .. 68 Plastic limit ................................ ................................ ................................ . 70 Proctor Compaction test ................................ ................................ ............ 70 Fibers ................................ ................................ ................................ ............... 72 Specimen Preparation ................................ ................................ ............................ 73 Flexural Beams for Fiber Length Determination ................................ ............... 74 F ield Produced Blocks ................................ ................................ ...................... 74 L aboratory Produced Short Flexural Beams ................................ ..................... 75 E arthen M ortar ................................ ................................ ................................ . 78 M as onry P risms ................................ ................................ ................................ 79 Specimen Testing ................................ ................................ ................................ ... 79 Preliminary Tests for Fiber Length Selection ................................ .................... 79 Block Compressive Strength ................................ ................................ ............ 83 Block 3 Point Bending Test ................................ ................................ .............. 84 Beam Flexural Strength Test ................................ ................................ ............ 85 Prism Compressive Strength Test ................................ ................................ .... 87 Prism Flexural Bond Strength Test ................................ ................................ ... 88 Microstructure Ana lysis Scanning Electron Microscopy (SEM) ............................ 89 Quality Control/Assurance Procedures ................................ ................................ ... 91 Soil Preparation ................................ ................................ ................................ 92 Fiber Addition ................................ ................................ ................................ ... 94 Moisture Content ................................ ................................ .............................. 96 Compression and Curing ................................ ................................ .................. 96 4 TEST RESULTS AND DISCUSSION ................................ ................................ ..... 98 Preliminary Results for Fiber Length Selection ................................ ....................... 98 Flexural St rength ................................ ................................ .............................. 98 Compressive Strength of Cubes Obtained from Failed Beams ...................... 101 Influence of Fiber Length on Flexural and Compressive Stren gth .................. 103 Compressive Strength of CEBs ................................ ................................ ............ 103 Adequacy of Sample Size ................................ ................................ ..................... 104 3 Point Bending Strength ................................ ................................ ...................... 111 Influence of Fibers on 3 Point Bending Strength ................................ .................. 111 Influence of Fibers on CEB Deformation ................................ ............................... 114 Relationship between Compressive Strength and 3 Point Bending Strength ....... 117 Load Deflection Response of Short Flexural Beams ................................ ............ 118 Flexural performance of Short Flexural Beams ................................ ..................... 121 Relationship between Fiber Content and Flexural Performance ........................... 124 Crack Patterns ................................ ................................ ................................ ...... 127 Scanning Electron Microscopy (SEM) Analysi s ................................ ..................... 128 Mortar Compressive Stre ngth ................................ ................................ ............... 130 Prism Compressive Strength ................................ ................................ ................ 132

PAGE 8

8 Prism Flexural Bond Strength ................................ ................................ ............... 133 Failure Mode ................................ ................................ ................................ ......... 135 5 CONCLUSIONS AND RECOMMENDATIONS ................................ ..................... 137 Overview ................................ ................................ ................................ ............... 137 Conclusions ................................ ................................ ................................ .......... 137 Flexural Performance, Compressive/3 Point Bending and Prism Bond Strength ................................ ................................ ................................ ....... 138 SEM Analysis ................................ ................................ ................................ . 138 Recommendations ................................ ................................ ................................ 139 Suggested Mix Proportions/Procedure ................................ ........................... 139 Recommendations for Future Research ................................ ......................... 140 APPENDIX A FLEXURAL PERFORMANCE RESULTS ................................ ............................. 142 B REGRESSION MODEL DIAGNOSTICS ................................ ............................... 144 C ADDITION AL SEM IMAGES ................................ ................................ ................. 154 LIST OF REFERENCES ................................ ................................ ............................. 157 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 167

PAGE 9

9 LIST OF TABLES Table page 2 1 Recommended proportions of soil constituents for best results ......................... 39 2 2 Strength of CEB/soil cement matrices ................................ ................................ 58 2 3 Influence of fibers on strength of CEB matrices ................................ .................. 58 2 4 Typical mortar types in use ................................ ................................ ................. 59 3 1 Chemical and mineralogical composition of cement ................................ ........... 69 3 2 Physical properties of soil ................................ ................................ ................... 71 3 3 Physical properties of MasterFiber MAC Matrix (polypropylene fibers) as provided by the manufacturer. ................................ ................................ ............ 73 3 4 Matrix mix proportions ................................ ................................ ........................ 74 3 5 Block mix proportions ................................ ................................ ......................... 75 3 6 Short flexural beam mix proportions ................................ ................................ ... 78 3 7 Mortar mix proportions ................................ ................................ ........................ 81 4 1 Modulus of rupture of tested beams (MPa) ................................ ...................... 101 4 2 Cube compressive strength (MPa) ................................ ................................ ... 102 4 3 Average flexural and compressive strength t test results for 54 mm and 27 mm fibers ................................ ................................ ................................ .......... 103 4 4 Compressive strength of CEBs ................................ ................................ ......... 104 4 5 Mean CEB compressive strength ................................ ................................ ..... 105 4 6 Model summary ................................ ................................ ................................ 109 4 7 Comparison of predicted and measured average co mpressive strength values (MPa) ................................ ................................ ................................ .... 111 4 8 3 point bending strength results of field produced blocks ................................ . 112 4 9 Results of compressive s trength and strain ................................ ...................... 115 4 10 Flexural properties (average of five specimens for each mix design) ............... 122

PAGE 10

10 4 11 Model summary ................................ ................................ ................................ 126 4 12 Mortar compressive strength (MPa) ................................ ................................ . 131 4 13 Prism compressive strength ................................ ................................ ............. 132 4 14 Flexural bond strength ................................ ................................ ...................... 134 A 1 Flexural performance of beams ................................ ................................ ........ 142 B 1 Model s ummary ................................ ................................ ................................ 144 B 2 ANOVA ................................ ................................ ................................ ............. 144 B 3 Coefficients ................................ ................................ ................................ ....... 145 B 4 Residuals s tatistics ................................ ................................ ........................... 146 B 5 Tests of n ormality ................................ ................................ ............................. 146 B 6 Model s ummary ................................ ................................ ................................ 149 B 7 ANOVA ................................ ................................ ................................ ............. 149 B 8 Coefficients ................................ ................................ ................................ ....... 149 B 9 Residual s tatistics ................................ ................................ ............................. 150 B 10 Tests of n ormality ................................ ................................ ............................. 151

PAGE 11

11 LIST OF FIGURES Figure page 1 1 Schematic diagram of research approach. ................................ ......................... 25 2 1 OPC hydration within CEB matrices. ................................ ................................ .. 43 2 2 Auram 300 block variations ................................ ................................ ................ 62 2 3 Hydraform dry staking system ................................ ................................ ............ 62 2 4 Sanford Winery, California. Built in steel reinforcement, vertically sleeved to attach to a wallhead bond beam ................................ ................................ ......... 63 3 1 Flow chart for specimen production and testing ................................ ................. 67 3 2 MasterFiber MAC Matrix fibers ................................ ................................ ........... 72 3 3 CEB molding process. ................................ ................................ ........................ 76 3 4 Soil drying in oven set at 93 o C.. ................................ ................................ ......... 77 3 5 Beam production ................................ ................................ ................................ 80 3 6 Freshly cast cubes. ................................ ................................ ............................. 81 3 7 CEB prisms. ................................ ................................ ................................ ........ 81 3 8 Flexural strength test setup ................................ ................................ ................ 82 3 9 Compressive strength testing ................................ ................................ ............. 83 3 10 Compressive strength test setup. ................................ ................................ ....... 84 3 11 3 point bending strength test.setup ................................ ................................ .... 84 3 12 Flexural strength test set up ................................ ................................ ................ 85 3 13 Schematic diagram of load deflection curve for calculating first peak strength, residual strength, and toughness where first peak load is equal to peak load. ... 86 3 14 Schematic diagram of load deflection curve for calculating first peak strength, residual strength, and toughness where peak load is greater than first peal load ................................ ................................ ................................ ..................... 86 3 15 Two block prism compressive strength test setup ................................ .............. 88 3 16 Flexural bond strength testing. ................................ ................................ ........... 88

PAGE 12

12 3 17 SEM JSM 6400 microscope for capturing micrographs of fractured surfaces. ... 91 3 18 SEM sample preparation ................................ ................................ .................... 92 3 19 Visual inspe ction of soil to be used for CEB production ................................ ..... 93 3 20 Soil sifting for CEB production. ................................ ................................ ........... 93 3 21 Cured CEBs showing porous surfaces wi th fibers sticking out ........................... 95 3 22 Short flexural beams showing with porous surfaces and fibers sticking out. ...... 95 3 23 The drop test ................................ ................................ ................................ ...... 9 6 3 24 CEB production and testing sequence ................................ ............................... 97 4 1 Typical load deflection curves for blocks. ................................ ........................... 99 4 2 Failure of blocks under compression. ................................ ............................... 102 4 3 Average compressive strength versus fiber content of blocks (n=30) .............. 108 4 4 Influence of fiber mass content on measured and predicted average compressive strength. ................................ ................................ ...................... 110 4 5 Average 3 point bending results of CEs versus fiber content of blocks (n=30 ) . 113 4 6 Typical failure mode of blocks ................................ ................................ .......... 116 4 7 Fibers bridging a crack during 3 point bending test ................................ .......... 116 4 8 Compressive strength and 3 point bending strength of blocks at different fiber mass fractions ................................ ................................ ................................ .. 118 4 9 ................................ ............... 120 4 10 Failed beams. ................................ ................................ ................................ ... 120 4 11 Predicted vs. experimentally recorded equivalent flexural strength .................. 126 4 12 Typical crack patterns. ................................ ................................ ...................... 128 4 13 Close up of fibers bridging crack . ................................ ................................ ... 128 4 14 SEM microgra ph of fractured beam surfaces ................................ ................... 129 4 15 SEM Micrograph of polypropylene fibers. ................................ ......................... 130 4 16 Typical hour glass type failure for tested specimens ................................ ........ 131

PAGE 13

1 3 4 17 Typical failure mode of blocks ................................ ................................ .......... 133 4 18 Failure between block unit and mortar joint. ................................ ..................... 135 4 19 Opposite sides of the first form of the second failure mode observed (shallow i ndentations/fractures ) ................................ ................................ ...................... 136 4 20 Opposite sides of the failure mode showing deep block fractures and block fibers sticking out. . ................................ ................................ ........................... 136 B 1 Residual plots ................................ ................................ ................................ ... 145 B 2 Residual diagnostics ................................ ................................ ......................... 147 B 3 Variance of residuals. ................................ ................................ ....................... 148 B 4 Residual plots ................................ ................................ ................................ ... 151 B 5 Residual diagnostics ................................ ................................ ......................... 152 B 6 Variance of residuals: Scatter plot of residuals vs predicted values ................. 153 C 1 Fiber showing cross linking and surface defor mation (embossment) ............... 154 C 2 Fractured CEB surface showing fiber pullout ................................ ................... 154 C 3 Fractured fiber after specimen failure ................................ ............................... 155 C 4 Fiber embedded in matrix after specimen failure ................................ .............. 155 C 5 Magnified image of fractured CEB surface ................................ ....................... 156 C 6 Zoomed in image of fractured CEB surface. ................................ ..................... 156

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requir ements for the Degree of Doctor of Philosophy SUSTAINABLE AND RESILIENT EARTHEN MASONRY SYSTEMS: ENHANCING STRENGTH PROPERTIES AND FLEXURAL PERFORMANCE USING POLYPROPYLENE FIBERS By P eter D onkor December 2014 Chair: Esther Obonyo Major: Design, Constru ction, and Planning The use of affordable , sustainable , locally available, and damage resilient construction materials are key to addressing the global need for adequate housing. Compressed earth blocks (CEB); a modern form of the adobe brick are gaining popularity as a construction material globally because they are stronger and more dimensionally stable compared to earlier forms of earthen construction materials. Despite the performance improvement achieved through using CEBs, they are still very brittl e and of lower strength in comparison to mainstream walling materials like concrete masonry units (CMU) and fired bricks. CEBs and other earthen masonry systems are weak in tension and exhibit low resistance to bending. This research investigated the poten tial of addressing some of the shortcomings of earthen masonry systems by assessing the influence of cross linked and embossed macro polypropylene fibers on the strength, deformability, and general flexural performance of CEBs. Unreinforced and fiber rein forced CEBs and short flexural beams were produced with fiber mass dosages of 0.2, 0.4, 0.6, 0.8, and 1.0%. Laboratory tests of m aterial properties were determined . Properties determined included compressive strength, 3 -

PAGE 15

15 point bending strength, first peak s trength, peak strength, residual strength, and flexural toughness. Earthen mortar was produced and used to cast two block and seven block CEB masonry prisms. Compressive strength and flexural bond strength of the masonry prisms were dete r mined to evaluate compatibility of the earthen mortar with the fiber reinforced CEBs. There was a general improvement in flexural performance and ductility of the fiber reinforced matrices as evidenced by the load de flection behavior , equivalent flexural strength, residual strength, and flexural toughness. Relationships between fiber quantity and enhancements in tested mechanical properties were observed and predictive models for compressive strength and equivalent flexural strength proposed. An observation of fractured sur faces after flexural strength testing using scanning electron microscopy (SEM) showed both fiber fracture and pullout; an indication of good fiber matrix bonding . The earthen mortar was deemed compatible with the fiber reinforced CEB s based on prism compre ssive strength , flexural bond strength , and failure mode. The results of this research show that when carefully designed and produced, polypropylene fiber reinforced CEBs can be used to construct CEB masonry with improved ductility, deformability, and fle xural performa nce.

PAGE 16

16 CHAPTER 1 INTRODUCTION Background The use of earth as a construction material and earthen construction techniques has existed for centuries. Earth was used as a building material in all ancient cultures mostly for residential and reli gious buildings (Adam and Agib, 2001; Minke, 2009). About 30% of all t he United Nations Educational, Scientific and Cultural Organization ( UNESCO ) World Heritage Sites are constructed with earth; a testament to the durability of earthen structures when wel l constructed and maintained (Jaquin and Augarade, 2012). Adobe, molded earth, wattle and daub, and cob are different forms of earthen construction that have been used for centuries. Globally, about one third of the human population resides in earthen shel ters. In developing countries, the number is estimated to be as high as 50% (Minke, 2009). It is generally accepted that earthen construction is sustainable considering that it uses indigenous soils and therefore reduces the use of manufactured materials ( UN Habitat, 2009). When used as a building material, earth has good strength in compression but is very weak in tension resulting in poor resistance to bending (UN Habitat, 1992). Historically, fibers have been used as reinforcement in earthen construction methods and techniques. Straw and horsehair were used to provide tensile reinforcement for sunbaked bricks and masonry mortar , and plaster respectively (ACI 544.1R, 1996). Aside using fibers to improve the brittleness of earthen masonry materials, fibers are also an effective way of reducing desiccation cracks associated with wet dry cycles and mitigating potential cracking induced by differential structural settlements. These performance enhancements are attributed to improvement in ductility due to fiber

PAGE 17

17 reinforcement (Li, 2005). Generally, earthen materials reinforced with fibers show an improved performance in resisting cracks and crack propagation, increase in compressive strength (depending on soil and fiber type) and increase in tensile strength (UN Habitat, 1992). The evolution of earthen construction techniques and systems over the years has been influenced by factors such as local culture, climate, knowledge, availability of materials and tools, and the advancement of architectural and constructio n technique (UN Habitat, 2009). Compressed earth blocks (CEB), also referred to as compressed stabilized earth blocks (CSEB) depending on the stabilization process used, are gradually gaining popularity over earlier forms of earthen masonry they are deem ed stronger and more dimensionally stable. For example, compacting moist soil mixed with between 4 10% of cement as a stabilizing agent can result in a significant increase in compressive strength, water resistance, and dimensional stability of blocks comp ared to traditional adobe (Morel et al. 2007). Generally, some of the benefits of using CEBs include improved strength and durability compared to adobe, and lower embodied energy levels compared to more conventional walling materials (Mesbah et al. 2004). CEB production is small scale, labor intensive, sustainable, and requires simple operating technology and maintenance. This makes CEBs well suited to help meet the projected global demand of about 96,000 new affordable housing units daily by the year 2030 (Hazeltine and Bull, 1999; Akubue, 2000; Adam and Agib, 2001; UN Habitat, 2005; UNESCO, 2010 ). Despite the strength improvement achieved through using CEBs over other traditional

PAGE 18

18 forms of ea rthen construction, they are still more brittle and weaker in bending and compression in comparison to CMU and fired bricks. The use of fibers as reinforcement in traditional earthen masonry has carried over into CEB production. Both natural (obtained fro m plants and animals) and synthetic fibers are used to reinforce soils for CEB production (Rigassi, 1995). The inclusion of fibers into CEB matrices creates a network of fibers, which improves tensile and shearing strengths, and also helps reduce shrinkage (UN Habitat, 1992; Rigassi, 1995; Mesbah et al. 2004; Elenga et al. 2011). Fiber reinforced blocks can withstand higher stresses by absorbing high amounts of energy making them particularly important in earthquake prone regions (UN Habitat, 1992). It has been demonstrated that sisal (Namango 2006; Manjunath, et al. 2013), coconut fiber (Khedari et al. 2005; Obonyo et al. 2010; Obonyo, 2011), straw (Binici et al. 2005), polyethylene (Elenga et al. 2011), jute (Islam and Iwashita, 2010; Harshita et al. 2014) , are all feasible options for CEB reinforcement. Improvement in ductility is widely accepted as a key benefit of fiber reinforcement i n soil cemen t composite s. An improvement in ductility prevent s catastrophic failure of earthen structures during events such as high winds and earthquakes. The delay in collapse as a result of improved ductility can be the factor determining if people get out alive or remain trapped inside a collapsing structure (Segetin et al. 2007; Subramaniaprasad et al. 2014). The inclusion of fibers into CEB matrices warrants further investigation as previous studies have suggested that they at times have some undesirable effects on CEBs such as reducing density, creating voids, causing micro fr actures at fiber soil

PAGE 19

19 interfaces, and reducing compressive strength (Rigassi, 1995; Khedari et al . 2005; Namango, 2006; Morton, 2008). Generally, synthetic fibers like polypropylene have been found to increase concrete mix cohesion, freeze thaw resistance , impact resistance, and plastic shrinkage. However, research findings on the influence of such fibers on concrete structural strength has been contradictory. For instance, contrasting results of increase and decrease in structural strength of concrete aft er synthetic fiber reinforcement have both been reported in literature (Wilson and Abolmaali, 2013). In cement stabilized soils used in geotechnical application s, the inclusion of randomly distributed polypropylene fibers have been found to increase unconf ined compressive strength, residual strength, absorbed energy, ductility, splitting tensile strength, and flexural toughness of soil mixtures (Maher and Ho, 1994; Consoli, et al. 1998; Tang et al. 2007; Jadhao and Nagarnaik, 2008; Consoli et al. 2009). Com and go through different compaction and curing processes. Because of these production differences, there is the need for an ev aluation of the effect of polypropylene The strength of reinforced earthen blocks is influenced by fiber type (tensile strength) and quantity (UN Habitat, 1992). Soil type, stabilizer type and quantity, and level of compaction also affect the strength properties and durability of CEBs. Mechanical properties such compressive strength, flexural strength and shear strength can therefore be enhanced if an optimal fiber reinforci ng ratio is identified and used (UN Habitat, 1992; Binici et al. 2005).

PAGE 20

20 Based on solely ecological metrics, natural fibers would be the preferred choice for CEB reinforcement because they are derived from renewable sources that are generally readily avail able at affordable costs. However, when untreated, natural fibers may negatively impact mechanical properties and durability of CEBs (Elenga et al . 2011). Untreated natural fibers can react chemically with ordinary Portland cement ( OPC ) in cementitious mat rices. The highly alkaline environment created through hydration causes fibers that are not alkali resistant to degrade, negatively affecting durability (ACI 544.1R, 1996). The net impact of such reactions on the strength properties of CEBs, especially dur ability due to the effect of the alkaline environment present in OPC, is a subject that needs to be further investigated before scaling up the use of natural fibers in real life applications (Obonyo, 2011). Such concerns have resulted in a growing interest in the use of synthetic fibers as reinforcement for CEBs. Where synthetic fibers have been used, it has mainly been fibers derived from post consumer plastic waste products. T his research used commercially available polypropylene fibers used for concrete production. Material properties of polypropylene fiber reinforced composites are influence d by fiber surface conditions , fiber geometry and length (aspect ratio) , fiber volume, and method of production / composition of matrices (Banthia and Gupta, 2006; ACI 544.1R, 2010). These identified factors influenced the selection of macro polypropylene fibers that are made of two cross linked filaments with an embossed surface. Structural performance of earthen masonry systems during hazards and load scenarios that ca n cause out of plane bending can be improved in two main ways a system level approach that uses structural reinforcement (steel rebar), reinforced

PAGE 21

21 concrete bond beams, roof to wall connections, and other established guidelines for unreinforced masonry pr actice , and im proving block and mortar properties such as (Elizabeth and Adams, 2005; Kestner et al. 2010 ; Islam and Iwashita, 2010; NIBS, 2010 ). Often, a combination of both approaches yields the best results. Th is research adopted the latter approach of improving block mortar properties by focusing on block strength, general flexural performance, and ductility using fiber reinforcement. Due to the concerns that exist with the use of fibers in CEB production, the focus of this research was to investigate the technical feasibility of incorporating a synthetic fiber (polypropylene) into CEBs to enhance general performance. The influence of polypropylene fibers on strength properties, and the pre and post crack behavi or of CEBs was evaluated in this study. Masonry i s generally defined as a built up construction or combination of building (masonry) units bonded together with or without mortar or other accepted methods of joining (ICC, 2012). In masonry systems that use mortar, a monolithic behavior is ensured by mortars that transmit loads between the masonry unit s (Azeredo et al. 2007). The bond strength of masonry is therefore essential to creating walls that can withstand lateral loading (Lawrence, 2008). E arthen (so il:cement) mortar is generally considered the most suitable for CEB masonry. However, contrary to general masonry practice, where weaker mortars are used, it is suggested using earthen mortars (soil:cement) with similar strength to blocks for CEB masonry ( Guillaud and Joffroy, 1995; Walker and Stace, 1997 ; (Venkatarama and Gupta, 2006). For this reason, a n earthen mortar mix was produced and used for the fabrication of CEB masonry prisms to evaluate its compatibility with the fiber reinforced CEBs.

PAGE 22

22 Objecti ves, Scope, Limitations, and Hypotheses Objectives The goal of this research was to develop fiber reinforced CEBs that can be used to construct CEB masonry systems resilient enough to withstand hazards such as high winds. This is to help improve the resil ience of structures constructed with the fiber reinforced CEBs by reducing failure probabilities and the associated consequences of such failures such as the loss of lives and property. To achieve this goal, this study was focused on theoretically and expe rimentally developing a fiber reinforced CEB matrix with improved ductility and flexural performance. The specific objectives were as follows: Perform a comprehensive literature review on fiber reinforced cement soil matrices. Investigate the effects of polypropylene fibers on the strength properties of cement stabilized CEBs. Design and experimentally validate an optimal fiber reinforced CEB matrix using polypropylene fibers. Visually assess composite failure mechanism of polypropylene fiber reinforced CEBs by evaluating the interfacial microstructure of fractured surfaces using scanning electron microscopy (SEM). Design and experimentally validate an optimal mix design for earthen mortar compatible with polypropylene fiber reinforced CEBs. Make recomm endations/propose specific guidelines for the inclusion of polypropylene fibers in to CEB matrices and appropriate test methods for assessing the effect of the fibers on flexural performance. Scope Fiber reinforcement is known to resist crack propagation, improve post crack performance, ductility, and fracture toughness of cementitious matrices (ACI 544.1R, 1996) . Despite these enhancements achieved through fiber reinforcement, there is a

PAGE 23

23 lack of information on the performance of cement stabilized CEBs rei nforced with engineered synthetic fibers. Research into synthetic fiber reinforcement of CEBs has been mainly focused on post consumer plastic waste products. The use of such fibers introduces the possibility of variations in fiber quality ; s uch potential variations, coupled with the inherent variation in soil properties make it difficult to isolate CEB performance properties that can be directly linked to specific fiber properties/attributes. Effects of both natural and synthetic fiber reinforcement on mec hanical properties such as compressive and flexural strength of CEBs have been varied. The scope of this research was limited to the optimization of CEB matrices through the production, testing of strength properties (compressive, 3 point bending, and flex ural) of unreinforced and polypropylene fiber reinforced CEB s, and designing an appropriate mortar compatible with the optimized blocks. Limitations The results of this research are limited to the soil type, fiber type, compression/compaction pressure, a nd curing regime used in the experimental work. The regression models proposed are specific to the mix designs used in this research. Mechanical properties of CEBs are influenced by the different variables/components used in production. Soil, stabilizer type and quantity, and fiber type, as well as level of compaction influence the strength, density, durability, and other properties of CEBs (UN Habitat, 1992; Rigassi, 1995; Walker, 1995; Montgomery, 2002). Where applicable, the physical and/or chemical pr operties of the materials used in the experimental work are presented to enable replication.

PAGE 24

24 Hypotheses The hypotheses of this research were: Polypropylene fibers can be used as reinforcement in cement stabilized CEB matrices to improve the ductility, f lexural performance, and deformability of CEBs. The addition of polypropylene fibers, as well as fiber characteristics, can affect CEB mechanical properties such compressive strength and flexural strength. Research Approach A schematic diagram of the rese arch approach adopted is presented in Figure 1 1. The methodology to evaluate the effects of polypropylene fibers on both fresh and cured properties of CEBs was developed based on an extensive literature review of the state of the art of fiber use in earth en masonry, concrete, earth stabilization (geotechnical application s), soil cement matrices, and laboratory experiments. In order to examine the influence of polypropylene fibers on compressive and flexural strength, quantitative relationships between mech anical properties of blocks and fibers were developed using statistical predictive models. The models expressed both compressive stren gth and flexural performance as function s of fiber content. An exploration of statistical relationships between mechanical properties of fiber re inforced CEBS and fiber content facilitated the identification of maximum cut offs (dosage) for fiber content beyond which the addition of fibers resulted in the degradat ion of mechanical performance. A n evaluation of different flexu ral testing protocols was undertaken to identify the best protocol to evaluate the influence of the fibers on flexural performance. Lastly, an earthen mortar mix was designed and used to produce CEB masonry prisms to evaluate compatibility with the fiber reinforced CEBs.

PAGE 25

25 Figure 1 1 . Schematic diagram of research approach.

PAGE 26

26 Outline of Dissertation Chapter 2 presents a literature review on earthen construction materials/ techniques and fiber reinforcement in soil cement matric es. The experimental design, materials, material properties, equipment, and test protocols used in the experimental program is presented in Chapter 3. Test results, analyses of results, and recommendations for quality assurance are presented in Chapter 4. Chapter 4 also includes the results of statistical analyses performed to determine the relationships between various physical properties tested and the polypropylene fibers used. Conclusions and recommendations for future studies are presented in Chapter 5 .

PAGE 27

27 CHAPTER 2 LITERATURE REVIEW Overview This chapter presents a literature review undertaken to develop a better understanding of earthen construction systems and techniques, CEBs, the effect of fibers on mechanical properties of soil cement matrices and CEBs, and the selection and design of appropriate mortars for CEB masonry. The literature review was also used to identify potentially suitable fibers for CEB reinforcement by looking at the history of fiber use in earthen masonry and best modern practices of fiber use in both the concrete and geotechnical industries. When used as a building material, earth is good in compression but weak in tension and therefore exhibits low resistance to bending. The weakness in tension means that for earth to be used a s a load bearing material, wall thickness serves as a conduit through which all forces must be transferred to the ground. This is the reason for the thick and massive nature of unstabilized earthen walls (Norton, 1997). Building codes and other standards s pecify minimum acceptable standards and criteria for constructing with earth. Most state and local building codes in the US are modeled after the International Building Code (IBC) with adoptions, variations, and specifications made to suit local conditions . As a result of their weakness, earthen masonry materials like adobe ( stabilized and unstabilized ) , and CEBs are restricted by US building codes to a maximum of two stories high and a minimum thickness of 10 inches for load bearing exterior walls in one story buildings (NMCB, 2009; ICC, 2012). These codes also specify methods, material selection criteria, and where applicable, quantification of physical properties of materials for earthen masonry construction.

PAGE 28

28 The literature review was also used to ident ify what the typical parameters of raw material selection for CEBs are, and minimum acceptable quantifiable properties. The review provided a fundamental understanding of strength development of fiber reinforced cement stabilized CEBs and general advantage s and disadvantages of CEBs. Existing Earthen Construction Systems/Techniques Reviewed This section of the dissertation is not to provide an exhaustive list of earthen masonry materials but to provide a general overview of some of the popular materials an d systems in use. Typically, the materials required to produce earthen masonry can be obtained locally, if not directly from building sites, thereby reducing transportation costs and associated emissions (Minke, 2009). Masonry systems made up of walls of e arth reduce the use of construction materials with high embodied energy, reduce construction waste, and conserves energy associated with building operation. The energy conservation achieved during building operation is the result of the heat storage capac ity associated with the thermal mass of earthen structures. When the principles of passive solar design are utilized in the design and construction of earthen structures, buildings can gain heat from the winter sun and store that heat in the wall mass to b e released during nighttime when temperatures drop. On the other hand, during the summer, walls can absorb excess heat from the living spaces thereby helping to maintain cool indoor temperatures (Elizabeth and Adams, 2005). Walls constructed with earthen m e thereby eliminating the potential for fungal growth; providing good indoor environments in the process (Norton, 1997, Elizabeth and Adams, 2005; Minke, 2009).The materials/con struction techniques reviewed here are adobe, cob, and rammed earth.

PAGE 29

29 Adobe Also referred to as sun dried or sun baked mud blocks, adobe are building units commonly used in arid regions all over the world. They are usually made of a compacted clay and stra w mixture (UN Habitat, 1992; Elizabeth and Adams, 2005). The addition of straw is typically to prevent cracking during the curing period, reduce shrinkage, and improve tensile strength (Elizabeth and Adams, 2005; Morton, 2008; Minke, 2009). The different t ypes of adobe in use include traditional adobe, stabilized adobe, pressed adobe, and burned adobe also known as Quemado (Smith, 1981; ICC, 2012). Traditional adobe also referred to as un stabilized adobe is typically made of a uniform mixture of clay, san d, silt, and most of the time straw. Stabilized adobe on the other hand is manufactured in a similar fashion as traditional adobe but is mixed with additives during the manufacturing process in order to limit water absorption into the cured adobe (Smith, 1 981; NMCB, 2009). Both pressed and burned adobes are manufactured in a similar manner as traditional or stabilized adobe. In the case of pressed adobe, the earthen mix is pressed with either hand or mechanical presses into dense bricks while with fired ado be, the mix is cured by low temperature kiln firing. Pressed adobe typically record high compressive strength and modulus of rupture numbers. Burned adobe is also typically much stronger than unfired adobe but generally not dense enough and may deteriorate with seasonal freeze thaw cycles (Smith, 1981). Typical minimum compressive strength for adobe in building codes is 2.07 MPa ( 300 psi ) (Kestner et al. 2010). Deterioration and maintenance of adobe structures tend to be problematic especially in high moist ure environments. Adobe bricks can be dimensionally unstable

PAGE 30

30 especially when unstabilized and are therefore prone to shrinking, swelling, warping and cracking in high moisture environments. Their resistance to elements of the weather is generally classifie d as weak to medium (UN Habitat, 2009). Due to the susceptibility of adobe structures to moisture damage, there is often the need to plaster adobe structures to help extend their life when used in humid areas. Plasters for exterior adobe walls however have to be compatible with the walling material. For example, plasters e the form of water vapor and may eventually lead to catastrophic failure of adobe walls (Elizabeth a nd Adams, 2005). Adobe generally has low tensile strength just like other earthen masonry material/systems and therefore designs using adobe should minimize tensile stresses (UN Habitat , 1992; Morton 2008; Minke, 2009). Cob Cob construction also referred to as stacked earth does not involve the use of individual dried masonry units (blocks) . Instead, a mixture of clay, sand and straw is produced, molded, and stacked one on each other to make walls and roofs . This is done while the mixture is still wet (Eli zabeth and Adams, 2005). Cob has been used in several places including the United Kingdom (UK), Iran, West Africa, and the American Southwest. Cob houses are known to have existed in England by the thirteenth century. It was the main method of building in the UK by the fifteenth century until the industrialization of the mid 1800s when cheap transportation made bricks the preferred construction material (Elizabeth and Adams, 2005; Kestner et al. 2010). A cob wall is usually constructed in layers of between 300 600 mm high by compacting wet cob mix at the head of the wall and tapering it down as the wall goes up. Each layer of cob is allowed to dry slightly before the next layer is placed; the dry

PAGE 31

31 cob providing a platform for workers to stand on (Weismann an d Bryce, 2006; Jaquin and Augarade, 2012). Cob walls typically tapper as they go up. It is commonplace to have walls 450 mm wide at the top (Weismann and Bryce, 2006). Construction with cob is labor intensive and needs no mechanization or formwork. Cob wal ls have been successfully used in the construction of buildings up to 7010 mm in height. An example of how high cob structures can go is the town of Shibam in Southern Yemen often Manhattan of the Deser a combination of c ob and adobe , Shibam has structures that are 10 stories high (Jerome et al. 1999; Kestner et al. 2010). Formal guidelines and testing protocols for cob are not as advanced as other forms of earthen construction (Kestner et al. 2010). Most modern load bear ing cob walls are built no longer than 3048 mm high. As a result of the manufacturing and construction processes involved in using cob, compressive strength of the material can vary widely but is generally similar to that of other earthen construction mate rials/methods which is 2.07 MPa ( 300 psi ) (Kestner et al. 2010). Rammed Earth Rammed earth is another earthen construction technique that has been used for centuries (Norton, 1997). This is evident by rammed earth foundations found in Assyria dating back to as far as 5000 BC (Minke, 2009). Examples of rammed earth structures that are hundreds of years old survive today in parts of North Africa, southern Europe, and the Middle East (Elizabeth and Adams, 2005).Rammed earth walls are built by making a mold in to which moist soil is poured, compacted, and then left to dry. Subsequently, the mold is released and the earthen form remains (Minke, 2009). Constructing with rammed earth has successfully continued into modern construction.

PAGE 32

32 A new form of rammed earth c onstruction also known as Pisé de terre was first used in Lyon, France in 1562. P isé de terre involves compress ing a moist mixture of earth between parallel wooden /metal plates that are later moved to other section s creating a monolithic wall in the proces s. Due to the absence of mortar joints, rammed earth walls are less susceptible to water damage (Norton, 1997; Elizabeth and Adams, 2005). In recent times, advances in soil stabilization and mechanization have led to the use of pneumatic rammers and light composite formwork instead of the traditional wooden rammer and heavy wooden formwork. The use of stabilizers like lime, OPC, and asphalt are incorporated into rammed earth soil mixtures to improve the strength (Burroughs, 2001). These advances have made i t possible to build faster, get a better quality finish, and smaller wall thicknesses (Heathcote, 2002). Typically, stabilized load bearing rammed earth walls have compressive strengths in excess of 2.0 MPa (290 psi) . The compaction during construction of rammed earth walls increases the density and strength of walls (Norton, 1997; Burroughs, 2001). Compressed Earth Blocks (CEB) The first CEBs were produced using wooden tamps to compress molded earth blocks (adobe) to improve their quality and performanc e. The earliest recorded use of presses for CEB production dates back to the 18 th century (Rigassi, 1995). Earthen construction techniques in developed countries dwindled during the industrial revolution but regained some interest after the Second World Wa r because of the shortage of industrial materials. Techniques used in massive road construction such as cement stabilization of earth roads were transferred to earthen construction to produce permanent structures. Hitherto, most earthen construction in dev eloped countries were limited to temporary structures.

PAGE 33

33 The first mechanical presses for CEB production, some of which were motor driven devices were introduced at the beginning of the 20 th century. CEB production was revolutionized in 1952 after the intro duction of the ClNVA RAM press designed by Raul Ramirez in Bogota, Columbia. The ClNVA RAM press allowed for the production of stronger block s at a faster rate (UN Habitat, 1992; Rigassi, 1995). Until this time, CEB use was predominantly in developing coun brought about an interest in earthen buildings in a few developed countries such as Australia. This was due to the low embodied energy and the thermal mass earthen structures provided. A new generation of presses; man ual, mechanical, and motor market for CEBs globally (Rigassi, 1995; Heathcote, 2002; Exelbirt, 2010). Developing countries until this period were mostly relying on tradi tional forms of earthen construction. The use of CEBs has been a success story in some sustainable construction efforts in developing countries. CEBs are therefore attracting attention from sustainable building movements within developed countries in an a ttempt to find more natural and ecologically friendly building materials and methods (Kibert, 2003). CEBs are produced by mixing soil with a stabilizer, slightly moistening the mixture, and compressing the mixture either manually or mechanically in a mold usually made of steel. Soil type, stabilizer type and quantity, and level of compaction affect the strength properties and durability of CEBs (UN Habitat, 1992). There are currently several types of stabilizing agents used in CEB production. These include cement, lime, pozzolanas, bitumen, gypsum, etc. The use of chemical stabilizers and compaction is to promote the use of CEBs in applications that are conventionally limited to fired clay

PAGE 34

34 bricks and CMU (Morel et al. 2007). CEBs are particularly suited for low cost construction and this typically places a limit on the extent to which chemical stabilizers can be used. For example, when used in volumes greater than 10%, OPC stabilization generally becomes uneconomical (Walker, 1995). Advantages of CEBs Typic ally, most of the materials required to produce CEBs can be obtained locally, if not directly from building sites, reducing transportation costs and associated emissions. CEB production and use does not require specialized equipment and training. The fire resistance properties of CEBs are also excellent (Adam and Agib, 2001; Minke, 2009). Locally manufactured CEBs are also an effective strategy for generating positive local economies their production and use in construction provides employment opportunities for indigenes (Murphy and Taub, 2010). Due to the wide range of soils that can be used for CEB production, environmentally unsustainable practices such as dredging river sand to get particular types of soils as is mostly the case with CMU production is av oided. CEB use also eliminates the need for wood fuel resources, a major advantage over fired bricks (Agevi, 1999; Mbumbia et al. 2000; Kerali, 2001; Montgomery, 2002). Some other benefits of using CEBs include lower embodied energy levels compared to conv entional walling materials (Mesbah et al. 2004; Minke, 2009; Kestner et al. 2010). For example, compared to a ton of fired clay bricks, a ton of 12% OPC stabilized CEBs use only 42% of the energy input needed and generates 62% of the equivalent CO 2 emissio ns generated by the fired bricks (Oti and Kinuthia, 2012). CEB production is also estimated to use about 1% of the energy required to produce the same volume of OPC concrete (Adam and Agib, 2001). However, the embodied energy of CEBs largely depends on the type and quantity of

PAGE 35

35 chemical stabilizers, and type of presses used. CEBs produced with hydraulic presses typically have higher levels of embodied energy compared to those produced with hand operated presses (Morton, 2008). CEB use can promote energy con servation during building operation because of the heat storage capacity associated with the thermal mass of earthen structures. During winter, structures store any heat gained from the sun to be released into the interior at night. On the other hand, dur ing summer, such walls through absorbing excess heat from living spaces can contribute to cool indoor temperatures (Elizabeth a nd Adams, 2005). The cool indoor temperature reported in CEB structures has also aterial. This means they either gain or lose heat due to the latent heat phenomena depending on local atmospheric conditions (Morony, 2004). In some regions, energy savings of as much as 80% over conventionally built and insulated frame houses have been re ported (Elizabeth and Adams, 2005). Disadvantages of CEBs Earth building materials in general have their peculiar disadvantages with CEBs being no different. A major disadvantage of CEBs and CEB technology is the lack of standardized procedures for produc tion and construction. Where codes and standards exist, they are very limited in scope (Adam and Agib 2001; Morton, 2008). On the other hand, variability in soils from site to site means optimum mix proportions differ thus complicating the issue of develop ing standards of practice and codes (Minke, 2009). Manual extraction of soil for CEB production using basic hand tools such as shovels and picks can be very labor intensive resulting in low output (Kerali, 2001). Fatigue

PAGE 36

36 becomes a factor with the use of ha nd operated presses and can lead to variations in block density and strength. Resistance to moisture can become an issue for CEB masonry when not properly constructed. CEBs like other cement based materials are affected by hydration triggered chemical an d structural changes that occur in the matrices (Obonyo, 2011). Generally, un rendered low cement, low density CEBs will deteriorate in less than 10 years under humid conditions. The deterioration is typically characterized by spalling of the exterior surf ace. This is especially so in areas that experience high precipitation (Adam and Agib 2001; Montgomery, 2002). Factors that cause deterioration in CEBs have been broadly categorized as physical, chemical, and biological. These identified factors mostly wor k together. Exposure to elements of the weather and other destructive agents that a building material may be exposed to during service can all lead to deterioration (Obonyo, 2010). Heathcote (2002) identifies wind driven rain as the predominant cause of lo ss of functionality of earthen walls. Wind driven rain causes erosion leading to the loss of surface area material. Earthen masonry including CEB masonry is typically protected from some of the agents of physical deterioration by the use of a suitable plas ter on exterior walls and the use of roof overhangs to prevent or reduce the pounding effect of wind driven rain. Water rising up through walls through capillary action is also a cause of deterioration of CEB structures. For this reason, building codes do not allow CEBs for subgrade construction (NMCB, 2009). The strength of CEBs are typically lower compared to mainstream masonry materials like fired bricks and CMU. While codes typically call out a compressive

PAGE 37

37 strength of 2.07 MPa ( 300 psi ) for CEB or stab ilized adobe, the compressive strength for CMU and fired bricks are typically 13.10 MPa ( 1900 psi ) and 17.24 MPa (2500 psi ) respectively. Due to their weakness, CEBs are restricted by codes to a maximum of two stories high. If not adequately reinforced/pro tected, CEBs exhibit low resistance to abrasion and impact loads (Adam and Agib 2001). Like other earthen construction materials, CEBs are weak in tension and therefore exhibit low resistance to bending moments. This weakness limits the structural applicat ion of CEBs (UN Habitat, 1992; Norton, 1997; Adam and Agib 2001). These drawbacks have resulted in the lack of institutional acceptance of earthen masonry in most countries leading to the underdevelopment of building codes and performance standards. The l ack of adequate test data to allow for the accurate prediction of structural performance of earthen masonry has not helped with the structural advancement of CEB masonry (Adam and Agib, 2001; Rigassi, 1995; Morton, 2008). However, with an adequate understa nding of earthen construction systems, the negative effects of the identified limitations or disadvantages of CEBs can be minimized. CEB Production CEB fabrication begins with carefully considering factors such as the intended use of the blocks load bear ing or non load bearing; location within structure internal or external; climatic conditions; local building codes; available skill level of construction workers, available soil type, etc. The listed factors ultimately influences the type of stabilizer use d if any, level of compaction, and the appropriateness of external render to CEB masonry. Small scale or on site production of CEBs typically begins with the extraction of appropriate soil, ensuring the soil is dry, screening the soil to rid it of unwanted elements, and selection of an appropriate stabilizer. This is followed by

PAGE 38

38 proportioning and mixing the soil, stabilizer and water, compacting the mix with an appropriate press, de molding and curing the blocks, and stockpiling cured blocks. After selectin g soil for CEB production, it is essential to evaluate some critical characteristics of the soil to determine the correct stabilization technique to use. Tests on soil for CEB production to determine parameters such as Liquid Limit, Plastic Limit, and Opt imum Moisture Content (Proctor Compaction) can be performed where the technology is available. These properties will be discussed in the Materials and Test Methods section of this dissertation . In the absence of apparatus for such tests, locally developed field assessment techniques are employed to help evaluate the different properties of soil necessary in developing an appropriate stabilization regime. Typical ranges of soil constituents recommended for CEB production is presented in Table 2 1. The data p rovided in Table 2 1 does not preclude the use of soils that fall outside the recommended ranges. Some studies using soils outside the recommended ranges that have yielded promising results include 87% sand (Arumala and Gondal, 2007; Obonyo et al. 2010), 4 8% clay (Elenga et al. 2011) and 44.7% clay (Chan, 2011). Using soil as a building material will involve either using soil from the project site, suitable soil brought from another site, or modifying local (site) soil to make it more suitable. Modifying s oil to make it more suitable for CEB production is generally referred to as stabilization (UN Habitat, 1992). Although stabilization is not always required, it is often deployed as a means to give CEBs properties that result in comparable performance and a ppearance to materials like fired bricks and CMU. Stabilization also lends CEBs pr operties that enable adherence to modern building code requirements (Morel et al, 2007; Jayasinghe and Mallawaarachchi, 2009).

PAGE 39

39 Table 2 1 . Recommende d proportions of soil constituents for best results (Rigassi, 1995; Norton, 1997) Soil Property Percentage Gravel 0 40% Sand/fine gravel 45 75% Sands 25 80% Slits 10 30% Clays 10 30% Stabilization There are different definitions of stabilization wh en it comes to CEB production but they all result in processes that yield the same product. Rigassi (1995) defines a process aimed mainly at preventing or minimizing water erosion and to a much lesser degree, increasing together particles typically by the use a chemical binder, such as c ement or lime. The process according to Morton increases durability and strength, but greatly increases embodied energy, waste, and life cycle cost. Norton (1997) states that the purpose of stabilizing soil is to improve the soil by either increasing its s trength or reducing variations in cohesion and size as a result of changes in moisture content, by reducing the erosive effect of water on the surface , or by a combination of the processes the properties of a soil water air system in order to obtain lasting properties which are compatible with a Habitat, 1992). The following conclusion can be drawn from the definitions provided: Stabilization improves targeted properties of soils and CEBs like density, compressive strength and durability.

PAGE 40

40 The stabilization regime used is influenced by soil type, technology available, and intended use of blocks. Stabilization increases the embodied energy of CEBs. There are a number of different ways of stabilizing soil. The use of a specific procedure does not necessarily preclude the use of another. Stabilizati on can be generally classified into three basic procedures: Mechanical stabilization: This process involves the modification of the properties of soil by treating its structure. This is achieved through the compaction of soil resulting in a modification o f its density, mechanical strength, compressibility, permeability, and porosity. Physical stabilization: This involves the modification of the properties of soil by treating its texture. This is achieved through controlling the mix of the various particle fractions of the soil by either eliminating inert particles and/or introducing desirable ones. Other techniques to use in modifying soil texture include heating, drying/freezing, electro osmosis, etc. Chemical stabilization: This is achieved by modifying the properties of soil through the addition of other materials or chemical products. The chemicals either helps in binding the grains of the soil together or results in a physicochemical reaction between the soil grains and the materials or the added produ ct. The physicochemical reaction may lead to the formation of new materials such as pozzolans when clay and lime react (UN Habitat, 1992; Rigassi, 1995; Auroville Earth Institute, n.d.). CEB production has been the subject of research especially in the ar ea of stabilizers and equipment. Further studies are required if the technology is to remain a viable and competitive alternative to mainstream modern building materials (Obonyo et al. 2010). A hybrid approach using both chemical and mechanical stabilizati on is the most common stabilization procedure used for CEB production. This procedure has resulted in a significant improvement in the physical properties of CEBs. There are currently several types of chemical stabilizing agents used in CEB production. The se include cement, lime, pozzolanas, bitumen, gypsum, soda waterglass, animal products, plant products, etc. (Rigassi, 1995; Minke, 2009; Deboucha and Hashim, 2011).

PAGE 41

41 Cement Stabilization The inclusion of cement into CEB matrices before compaction improves strength properties of CEB matrices and resistance to water a key threat to the durability of blocks (Morton, 2008; Minke 2009). The quantity of cement required for adequate soil stabilization depends on factors such as the soil type, required compressi ve strength, environmental conditions and levels of quality control. It is unnecessary and probably detrimental to mix cement with soils with clay contents greater than 20% (Rigassi, 1995). Depending on the exposure conditions of CEB masonry, the quantity of cement used for stabilization can be reduced. If the aim of stabilization is primarily to protect masonry against moisture damage or water erosion, the use of roof overhangs and plasters to prevent moisture intrusion can be used to achieve same, thus re ducing or totally eliminating the quantity of cement required (Minke 2009). Controlling the moisture content, level of compaction and the curing regime of CEBs is critical to getting the most from the added cement (Montgomery, 2002). The main effect of cem ent on sands and gravels is similar to the hydration in concrete or in sand cement mortar. The properties of hydraulic cement determine the way the constituent compounds react with water. When water is added to OPC, it goes through hydration, which causes OPC and hence the matrix made of it to harden and develop strength. When cement hydrates, tricalcium silicates (C S), dicalcium silicates (C S), tricalcium aluminate (C A), and tetracalcium Aluminoferrite (C AF) are formed. The C S formed gains most of the strength developed in the first 2 to 3 weeks; C S contributes to long term strength, while C A and C AF primarily contribute to early strength gain. The C A formed when OPC is initially mixed with water is much more

PAGE 42

42 reactive than the C S formed and theref ore has the possibility of leading to premature stiffening known as flash set (ACI E 701.E3, 2001). It is therefore recommended to compact CEB matrices soon after water has been added to the dry mix and a homogenous matrix has been formed (Rigassi, 1995). However when C A hydration is properly controlled, setting of hydraulic cement such as OPC is determined by the hydration behavior of C S. The period during which OPC matrices remain workable is due to the fact that C S only starts to react rapidly after a few hours of contact with water (ACI E 701.E3, 2001). According to Adam and Agib (2001), the cementitious gel calcium silicate hydrates, calcium aluminate hydrates and hydrated lime formed when cement hydrates is independent of the soil in CEB matrices. The calcium s ilicate hydrates and calcium aluminate hydrates form the main bulk of the cementitious gel, whereas the lime is deposited as a separate crystalline solid phase. The cementation process therefore results in the deposition of an insoluble binder capable of e mbedding soil particles in a matrix of cementitious gel ( Figure 2 1). The extent to which the hydration process can occur throughout the soil cement mix is dependent on time, temperature and type of cement. The lime released during hydration of the cem ent reacts with the clay content in the soil to form additional cementitious bonds. Compacting the wet matrix immediately after mixing helps lessen the possibility of the break down the newly created gel. In determining the quantity of cement to use for a given soil, Norton, (1997) suggests producing test blocks with varying cement content of between 5% and 10% and testing the blocks for their suitability for the intended use. The results of a shrink box test can also be used as a good preliminary guide (N orton, 1997; Adam and Agib,

PAGE 43

43 2001). Cement is generally a good stabilizer for granular soils but unsatisfactory for clays. With clayey soils, it is uneconomical to use cement because more cement is required to achieve adequate physical characteristics . Adam and Agib (2001) suggest between 3% and 18% of cement content by weight to be generally good for stabilization depending on soil type. Rigassi (1995) on the other hand suggests that with low proportions of cement stabilization of between 2 3%, the compres sive strength of certain soils is better when left unstabilized and that generally, at least 5 6% cement stabilization is needed to obtain satisfactory results. Compacting moist soil combined with between 4 10% of cement stabilization has been found to res ult in a significant increase in compressive strength, water resistance, and dimensional stability of blocks compared to traditional adobe (Morel et al, 2007). Figure 2 1. OPC hydration within CEB matrices (Montgomery, 2002) . The development of a bette r criteria to assess soil suitability for stabilization and accurately quantifying stabilization treatments based on soil properties will help reduce the need for extensive stabilization trials as is currently the practice for modern day earthen constructi on materials (Burroughs, 2001). The correct percentage of cement for each use depends primarily on the composition of the soil and the purpose for which the blocks are produced (Minke, 2009). The compaction effort to be used for block

PAGE 44

44 production will also determine cementation level the higher the compaction pressure, the greater the compressive strength of blocks produced (Gooding, 1994). Given the inherent variability in soil properties, producing trial blocks even though time and resource consuming seem s to be the best option. Other options might be more prescriptive than performance based thus disqualifying the use of soils that would have otherwise yielded blocks with desirable properties. Lime Stabilization Non hydraulic lime (quick lime or slaked li me) is best suited for clayey soils since it reacts with some of the clay and calcium minerals and water to produce a cement itious product . This increases the soils strength and reduces its susceptibility to water through a process known as the pozzolanic effect (Norton, 1997). Adam and Agib (2001) identify four basic reactions that occur when lime is added to soil; cation exchange, flocculation and agglomeration, carbonation, and pozzolanic reactions. Cation exchange is the exchange of positively and negat ively charged ions between lime and moist soil that results in the formation of bridges that hold particles together. This process results in hardening of the matrix and is mainly influenced by the exchange capacity of the soil. Flocculation and agglomera tion are processes that result in the collection of tiny suspended particles or the collision of particles to form larger ones. These processes result in the rearrangement of the face to face orientation of clay particles to a more compact edge face orien tation; improving strength, stiffness and workability in the process. Carbonation is an undesirable reaction, which results in the formation of relatively insoluble carbonates (Adam and Agib, 2001; Dhakal, 2012). The pozzolanic reaction; believed to be th e most important of the reactions described occurs between lime and certain clay minerals. The reaction is a long term process that

PAGE 45

45 produces cementitious compounds that bind the soil particles together. The pH environment created in the matrix results in f urther reactions between the silica and alumina in the lime with the clay particles thus providing extra strength (Norton, 1997; Adam and Agib, 2001; Dhakal, 2012). Lime reduces the degree to which clay absorbs water, thus making the soil less sensitive to changes in moisture content while also improving its workability. Given that the reaction of lime with clay minerals is a slow process, it may be appropriate to include some cement in the mix when using lime. Cement would in this case result in attaining some stability quickly, which in turn facilitates the molding and handling of blocks in the fresh state as a prelude to the slower lime strengthening process. Kerali (2001) and Obonyo (2011) have reported promising results using both cement and lime as sta bilizers in the same matrix. Since lime forms links with the clays present in soil, and hardly forms any with the sands, the use of lime as a stabilizer for CEB production should only be considered when cement stabilization is impossible (Rigassi, 1995). T his is often the case when the clay content in the soils being considered is greater than 30% (Kerali, 2001). Regarding the quantity of lime to use, Rigassi (1995) recommends between 6 12%, and Adam and Agib (2001) recommend between 5 10% proportions of l ime. While the recommendation by Adam and Agib (2001) is stated in weight relative to the weight of sand, the recommendation by Rigassi (1995) is assumed to be by volume even though not explicitly stated. Norton (1997) on the other hand suggests the produc tion of test samples as a means of determining the correct quantity of lime to use. As in the case of cement, the correct percentage of lime to use depends primarily on the composition of

PAGE 46

46 the soil and the purpose for which the blocks are produced. For exam ple, lime reacts more quickly with montmorillonite clays than with kaolinites (UN Habitat, 1992). Even though cement and lime are the most widely used chemical stabilizers (Montgomery, 2002; Deboucha and Hashim, 2011), pozzolans; both natural occurring and those obtained from industrial waste have been successfully used as supplementary stabilizers in CEB production although not on a large scale. Fiber Reinforced CEBs Fibers are widely used to stabilize earth for construction. The use of fibers such as st raw as stabilizing agents have found place in modern applications and are used in combination with other stabilization agents like cement, lime, and bitumen (UN Habitat, 1992). With CEBs, the addition of organic fibers like animal hair, coir, sisal, bambo o, straw, and mineral or synthetic fibers creates a network of omni directional fibers that improves tensile and shearing strengths, and also help to reduce shrinkage (Rigassi, 1995, Namango, 2006; Elenga et al. 2011). In addition to reducing shrinkage, im proving tensile strength, and increasing thermal resistance, fibers may lower density resulting in reduced weight and compressive strength (Morton, 2008). Rigassi (1995) states that although the use of fibers as reinforcement is common in adobe, they are i ncompatible with the CEB compression process as they render the mix too elastic but does not ssertion is however, contradicted by results of studies by Bouchi et al. (2005), Namango (2006), and Elenga et al. (2011). These cited studies successfully produced CEBs with fibers without using cement or any other chemical stabilizing agent. A major adv antage of adding fibers to matrices is that after initial matrix c racking, fibers bridge and restrain

PAGE 47

47 cracks from propagating. The additional forces required to pull out or fracture the fibers either protects the integrity of matrices or improves the post crack load carrying capacity (Suko ntasukkul and Jamsawang, 2012). Sisal fibers (gotten from the leaves of the Agave sisalana plant) for instance have been found to cause an improvement in ductility, tension, and inhibit tensile crack growth propagation aft er initial crack formation in CEBs (Mesbah et al. 2004). After exposure of a set of CEBs to elevated temperature and moisture conditions, Obonyo (2011) reported that soil cement fiber bricks had a residual strength of 87% compared to 48.19% and 46.20% for soil cement and soil cement lime blocks respectively. The author observed that the inclusion of fiber in the mix reduced Carbon content by over 1/3 while significantly increasing the Oxygen, Silica and Calcium contents. The findings while drawing light on the chemical changes that occur within CEBs with the addition of the fibers also raises the issue of chemical reactions in CEB matrices that are potentially triggered through the addition of fibers. In another study, Obonyo et al. (2010) used the modified Bulletin 5 Spray Test to assess the durability of a set of both coconut fiber reinforced and unreinforced CEBs. The blocks were sprayed with water at a pressure of 2.07 MPa (300 psi) and 4.14 MPa (600 psi) over a one hour period. The fiber reinforced block s had a maximum erosion depth of 40 mm at 2.07 MPa and 55 mm at 4.14 MPa compared to a maximum erosion depth 0.5 mm at 2.07 MPa water pressure and 0.8 mm at 4.14 MPa recorded for the unreinforced blocks. The erodability of the fiber reinforced blocks raise s some questions on how well fibers help in bonding the other constituent materials of CEBs. A set of compressive strength results provided in the study is not clarified to be either pre or post the erosion tests. Such detail would

PAGE 48

48 have contributed to bett er understanding the correlation between the level of erosion and compressive strength reduction if any since previous studies, (Khedari et al. 2005; Namango, 2006; Elenga et al. 2011) have attributed strength properties of fiber cement soil matrices to be in part a function of fiber matrix bonds. Khedari et al. (2005) studied coconut fiber reinforced CEBs at various mix ratios and concluded that coconut fiber reinforcement can reduce the thermal conductivity and weight of CEBs. Given the focus of differen t researchers on different aspects of fiber reinforced CEBs, there is the need to study the influence of fibers on key mechanical properties of CEBs. Without adequate knowledge of fibers on key mechanical properties, widespread acceptability of fiber reinf orced CEBs by building professionals would be hindered. Fiber Effects on Mechanical Properties Compressive and flexural strength are important parameters for satisfactory performance of load bearing walls especially when constructed with materials conside red to be unconventional (Jayasinghe and Mallawaarachchi, 2009). There is no general consensus on testing procedures for CEBs. CEB testing usually adopts methods, instrumentations and procedures used for CMU and fired clay. For example, block geometry and aspect ratio are factors that influence test results of CEBs but have not been standardized. Neither has the extent of influence of such factors been agreed on (Morel et al. 2007). The scenario outlined complicates direct comparison of the work of differen t researchers. Variability of soil and quality control measures during production also compounds the problem of comparing different research findings. However, relative changes within studies after fiber inclusion in matrices can be compared.

PAGE 49

49 One of the m ain problems for the development of new earth building has been t he lack of standard criteria that could inform an accurate evaluation of the finished material. This absence has a negative influence on the owners, decision makers and financial backers beca use, when considering investment in earth structures, they have no guarantee of the technical quality of the buildings, particularly with respect to their durability beyond the period of the loan (Houben and Guillaud, 1994). In the view of Morton (2008), t he situation described above has significantly improved since 1994 in some countries but same cannot be said for other places that still rely on standards written for other materials or other countries. The strength of fiber reinforced earthen blocks is i nfluenced by fiber type (tensile strength) and quantity (UN Habitat, 1992). Soil type, stabilizer type and quantity, and level of compaction also affect the strength properties and durability of CEBs. Mechanical properties such compressive strength, flexur al strength and shear strength can therefore be enhanced if an optimal fiber reinforcing ratio is identified and used (UN Habitat, 1992; Binici et al. 2005). The main mechanical properties of fiber reinforced CEBs to be reviewed in the subsequent sections of this chapter are compressive and flexural strength. Compressiv e S trength Compressive strength is an important parameter for load bearing masonry for residential construction (Jayasinghe and Mallawaarachchi, 2009). Compressive strength is a basic meas ure of quality and also the most commonly tested mechanical property of CEBs (Morel et al. 2007). There is no consensus on testing procedures for CEBs such as testing blocks in the dry or wet state (soaked in water) (Morel et al. 2007). The use of manually operated presses typically produces blocks with compressive strength between 1.7 3.0 MPa (250 435 psi) while mechanically operated presses have been known to produce blocks with compressive strengths in excess of 7 MPa (1000 psi) (Morel et al.2007; Morton , 2008). Soil type, stabilizer type and quantity, curing

PAGE 50

50 conditions, and testing procedures used have an influence on recorded compressive strength values. With fiber reinforced earthen materials, compressive strength largely depends on the quantity of fib ers used, initial compressive strength of the soil, tensile strength of fibers, and the internal friction between fibers and soil (UN Habitat, 1992). In a study to assess the influence of soil characteristics and cement content on CEB physical properties, Walker (1995) determined the dry compressive strength of CEBs to be greater than their corresponding saturated blocks (blocks soaked in water). The author found that increasing cement content improved the saturated compressive strength of the studied CEBs . Conversely, increasing clay content and using soils with a plasticity index above 20 negatively impacted saturated compressive strength. Increasing clay content was found to cause an increase in compressive strength, which was misleading. The author pro posed that CEBs should be tested in their saturated state instead of dry. However, this proposal might not always be the best option especially when the soil used has low clay content and if the CEBs are mostly going to be dry during their service life. T he saturated compressive strength recorded in the study fell in the range of 0.66 4.58 MPa (95 665 psi). The corresponding dry compressive strength was in the range of 2.53 7.86 MPa (370 1140 psi); an average of 72% to 283% higher compared to the s aturated compressive strength. Similar to the findings of Walker (1995), Venkatarama Reddy and Gupta (2006) found that increasing cement content positively influenced compressive strength. The saturated compressive strength values reported by Venkatarama R eddy and Gupta (2006) were in the range of 3.13 7.19 MPa (455 1043 psi) which were typically higher than those reported by Walker (1995). It is worth noting that the soil type and compaction effort used in both

PAGE 51

51 studies were different. In addition, while Walker (1995) used between 5 10% cement by volume, Venkatarama Reddy and Gupta (2006) used between 6 12% cement by weight. Eko et al. (2006) also found the compressive strength of CEBs produced with lateritic soils and stabilized with 6, 8, and 10% OPC (by weight) to be 2.2 MPa (320 psi), 2.7 MPa (390 psi), and 3.48 MPa (505 psi) respectively. Increasing cement content resulted in an increase in compressive strength in the study. While it is generally accepted that increasing OPC content results in an i ncrease in compressive strength, there is no direct correlation between the two. For example, doubling OPC content would not necessarily result in a doubling of compressive strength (Venkatarama Reddy and Gupta, 2006). Fibers are used together with stabili zers like OPC to improve CEB performance. Namango (2006) studied the effects of sisal fibers and cement on CEBs. The results showed an increase in compressive with a corresponding increase in sisal levels from 0.25% to 0.75%. Compressive strength values r ecorded were from 4.16 MPa (605 psi) at 0.25% sisal content to 9.14 MPa (1325 psi) at 0.75% sisal content. The 9.14 MPa value recorded at 0.75% sisal content represented a 90.5% increase in compressive strength compared to the unreinforced samples. However , at 1.0% sisal content, compressive strength dipped slightly from 90.5% to 84.8% when compared to the unreinforced samples. The inclusion of cement into the matrix yielded mixed results. In comparison to the blocks without cement, t he rate of increase in compressive strength wa s lower at 1.0% and 1 .25% sisal content; the higher amounts of sisal in combination with cement having a negative impact on compressive strength. The increase in strength recorded after the addition of the fibers was attributed to th e creation of

PAGE 52

52 isotropic matrices between the clay structure of the soil and fiber network the matrices were seen as opposing the movement of particles and cr eating stability mainly because the fibers appeared to distribute tension throughout the bulk of the material. In addition, fibers oppose crack formation and propagation in step with incr easing stress (UN Habitat, 1992; Houben and Guillaud, 1994). The decrease in strength reported when fiber content was greater than 1% was attributed to the development of micro fractures at sisal soil interfaces when sisal content was increased and this r e sulted in the reduction of compressive strength to 4.16 N/mm² at 1.25% sisal content. The study found the optimal sisal cement soil mixtures to be 0.5% sisal with 12% cement content yielding a 40.7% increase in compressive strength compared to the unreinfo rced samples. Elenga et al. (2011) produced CEBs using two clayey soils reinforced with polyethylene (PE) fibers derived from plastic waste for comparison to unreinforced blocks. Two types of polyethylene (PE) reinforcement derived from the plastic waste w as used in the study; a net and film. CEBs were produced with a compaction effort of 2.8 MPa (405psi) and 6.0 MPa (870 psi) with 0.3% fibers. Compared to the unreinforced CEBs, an increase in compressive strength of about 10% for the blocks produced at a c ompaction of 2.8 MPa (405 psi), and an increase in the range of 20% 30% for the blocks produced at a compaction effort of 6.0 MPa (870 psi) were observed. The increase in compressive strength after fiber reinforcement was attributed to the frictional for ces generated between the fibers and soil during the lateral extension of the blocks in testing. The results showed that the higher the compaction effort used in the production of the CEBs, the more pronounced the influence of the fibers on compressive str ength.

PAGE 53

53 This is because fiber matrix bonding is enhanced through compaction. A good fiber matrix bond increases the interfacial and frictional bonds activated when fibers are pulled out of matrices (Sukontasukkul and Jamsawang, 2012). Fiber surface conditio ns also have the potential of improving fiber matrix bonding (Khedari et al. 2005; Prasad et al. 2012). In a similar study to Elenga et al. (2011), Prasad et al. (2012) studied the effect of plastic waste fibers obtained from carry bags and water bottles on the performance of CEBs. Blocks were produced using cement, lime, and combinations of both as stabilizers. The fibers added to the mix made up 0.1% and 0.2% of the weight of the soil. The blocks reinforced with 0.1% of the plastic fibers showed an incre ase in compressive strength of between 3 to 10%. The blocks produced with the fibers obtained from the water bottles however did not show any improvement in compressive strength. The reason given for the poor performance of the blocks produced with the fib ers obtained from the water bottles was that there was not enough friction between the fibers and matrix. Binici et al. (2005) undertook a study aimed at engineering an earthquake resistant material with high compressive strength by investigating the mech anical properties of certain combinations of clay, cement, basaltic pumice (ground volcanic rock), lime, gypsum, straw, plastic fibers and polystyrene fabric. CEB s were produced with these materials and tested for their compressive strength. It was observ ed that the polystyrene fabric increased the compressive strength of the blocks. The plastic fiber reinforced blocks had the highest average compressive strength of 5.1 M Pa (740 psi)

PAGE 54

54 which was 21% higher than the blocks reinforced with straw. The study showed that fiber type and fiber distribution/orientation in matrices has an influence on CEB compressive strength. Khedari et al. (2005) studied coconut fiber reinforced CEBs at various mix proportions. A decrease in compressive strength with a corresponding increase in fiber content was observed during the study. The explanation given for the observed trend was that the development of strength properties of soil cement fi ber mixes mostly matrix, matrix identified bonds can be a ected by dimensions, surface conditions and the quantity of in a decrease in bond strength of the specimens, leading to a lower compressive strength of the fiber reinforced CEBs. The introduction of fibers into the mix was also seen as creating voids leading to a decrease in density. As a result, the compressive st rength of specimens decreased accordingly. The optimum mix ratio was 5.75:1.25:2.0 (soil:cement:sand by volume) and 0.8 kg of fibers per block. The findings provide valuable insight into fiber reinforced CEBs especially in the area of optimizing fiber cont ent. Galán Marín et al. (2010) studied the stabilization of soils with alginate (a natural on CEBs produced showed that the addition of wool fiber increased compressive s trength from 2.23 MPa (325 psi) to 3.05 MPa (440 psi) (37% increase). The addition of both alginate and wool fiber increased compressive strength from 2.23 MPa (325 psi) 4.44 MPa (640 psi) (99% increase). Combinations of both stabilizer and wool fiber were

PAGE 55

55 determined to be the best way of enhancing strength. Optimum wool/soil ratio needed to produce high strength blocks was determined to be 0.25%. Bouchi et al. (2005) investigated the physical and mechanical performance of different composite soils reinfor ced with chopped barely straw. Increasing barley straw reinforcement content up to 1.5% enhanced compressive strength by 10 20% depending on the type of the soil used. On the other hand, increasing fiber content above 1.5% resulted in a reduction in compre ssive strength. At 3.5% fiber reinforcement, reduction in compressive strength was about 45%. The effectiveness of the reinforcement was more pronounced with the clayey soils compared to the sandy soils. The study did not use a chemical stabilizer. The eff ect of increasing fiber length on the compressive strength was negligible. Flexural S trength Flexural strength of earthen masonry materials deserves special attention since adequate knowledge of it can allow structural design engineers to check its adequ acy with suitable design methods. Flexural strength data is becoming more important with the new trends of designing houses to be more hazard resilient during catastrophic events (Jayasinghe and Mallawaarachchi, 2009). Flexural strength is also referred to as modulus of rupture (MOR) or 3 point bending strength. Testing of CEBs is typically done by subjecting blocks to a single (center point) loading point with simple supports. Strength is calculated assuming pure bending and ignoring other factors such as shear and the arching action typically observed during failure. Center point loading may therefore result in an overestimation of the modulus of rupture since the values are evaluated using peak load (Walker, 1995; Morel et al. 2002; Morel et al. 2007).

PAGE 56

56 Namango (2006) reported an increase in flexural strength with a corresponding increase in sisal fiber content of CEBs from 0.25% to 0.75%. Flexural strength values recorded were from 0.85 MPa (125 psi) at 0.25% sisal to 1.63 MPa (235 psi) at 0.75% sisal co ntent. The value of 1.63 MPa recorded at 0.75% sisal content represented a 64.3% increase in flexural strength compared to the unreinforced blocks. However, at 1.0% sisal content, flexural strength reduced by 48.5% when compared to the unreinforced blocks. Similar to observations made for compressive strength, the study found the optimal sisal cement soil mixtures to be 0.5% sisal with 12% cement content yielding a 37.1% increase in flexural strength compared to the unreinforced blocks. The average compress ive strength in the study was in the range of 4.8 to 5.5 times higher than the average flexural strength of blocks. The ratio of compressive to flexural strength was close to the one sixth relationship proposed by Walker (1995) as a basis for rapid and ine xpensive field assessment of CEBs. Walker (1995) however cautions that test results based on the one sixth relationship can be conservative . Eko et al. (2006) on the other reported compressive strength values that were between 8.8 to 9.6 times more than th e flexural strength values. Elenga et al. (2011) reported the average flexural strength of unreinforced CEBs produced with a compaction effort of 6.0 MPa (870 psi) for the two different soil types used was about 1.5 MPa (215 psi). On the other hand, the f lexural strength of blocks reinforced with both film and net PE fibers and produced with a compaction effort of 6.0 MPa (870 psi) was about 1.6 MPa (230 psi) representing an increase of between 7 and 10% compared to the unreinforced blocks.

PAGE 57

57 It is generall y difficult to compare results from different CEB studies due to the widely varying experimental procedures used. Compaction pressure, soil and fiber type, and curing regime influence test results. Not all authors provide details on the variables in the ir studies to allow for easy comparison. Notwithstanding th is difficulty, Table s 2 2 and 2 3 sum up the findings in different studies of CEBs and soil cement matrices using relative differences . Mortar for CEB Masonry Systems The two main parameters of durab ility in masonry are a dimensionally stable unit and a mortar that forms a complete and permanent bond (Boynton and Gutschick, 1964). The response of masonry to tensile stresses imposed by high winds, floods, or other loads that can cause out of plane bend ing and failure is largely influenced by the flexural strength of the masonry, which in turn is influenced, by bond strength between mortar and masonry for unreinforced masonry (NCMA, 1994; Jayasinghe and Mallawaarachchi, 2009). The selection of an appropr iate compatible mortar for masonry construction is therefore critical to the performance of the masonry system. In all masonry systems, mortar plays an important role in the bonding of masonry elements along vertical and horizontal joints. According to Gu illaud and Joffroy (1995), in CEB masonry construction, mortar plays a threefold role: It bonds the masonry elements together in all directions (vertical and horizontal joints). It allows forces to be transmitted between the elements and notably vertical f orces (i.e. the weight of the elements themselves, or applied forces). It thus enables these forces to be distributed across the whole surface of the masonry elements. It compensates for any defects in horizontality in the execution of the masonry work.

PAGE 58

58 T able 2 2. Strength of CEB/soil cement m atrices Table 2 3 . Influence of fibers on s trength of C EB matrices * Geote chnical application s Author Optimal Mix Design ( Fiber and Stabilizer Type and Content ) Change in Mechanical Property Compressive Strength Flexural Strength Binici et al. (2005) 0.10% plastic fiber and 10.00% OPC 36.00% increase Bouhicha et al. 2005 1.50% barley straw (compressive strength) and 3.5% barley straw (flexural strength) 20.00% increase 16.00% increase Khedari et al. (2005) 0.25% coconut fiber and 1.2 5% OPC by volume 31.00% reduction Namango (2006) 0.50% sisal fiber and 12.00% OPC by mass 40.70% increase 37.10% increase Tang et al. (2007)* 0.25% polypropylene fiber and 8.00% OPC 132.00% increase Consoli et al. (2009)* 0.50% polypropylene fiber a nd 7.00% OPC 57.00% increase Galán Marín et al. 2010 0.25% wool, 19.50% alginate and 0.50% lignum 99.00% increase 29.00% increase Aguwu, 2013 0.25% coconut fiber 10.00% increase Author Stabilizer Type and Content (%) Average Dry Compressive Strength @ 28 days (MPa) Average Dry 3 Point Bending Strength / Modulus of Rupture (MPa) Fiber Walker, 1995 Walker, 2004 1 Binici et al. 2005 10 (OPC) Plastic fiber/ straw /polystyrene Bouhicha et al. 2005 1.0 3.0 Chopped barley straw Khedari et al. 2005 Coconut fiber (coir) Namango, 2006 1.63 Sisal Segetin et al. 2007 10 (OPC) Harakeke (flax plant) Galán Marín et al. 2010 0 19.5% (alignate) 0.5% (lignum) Wool Elenga et al. 2011 Aguwu, 2013 Coconut fiber (coir)

PAGE 59

59 Typical mortars used in conventional masonry and their strengths are presented in Table 2 4. Table 2 4. Typical mortar types in use (Klingner, 2010; ASTM C270 12a) It is generally recommended to use mo rtar of similar strength and similar materials to CEBs for CEB masonry systems . Mortar with strength values that are much lower cause the blocks to be highly susceptible to erosion and water infiltration, thus contributing to the deterioration of the block s. When mortar of very high strength compared to CEBs is used, erosion becomes uneven causing portions of the mortar to protrude from the surface of the masonry. This bears the risk of water stagnating on the exposed surface (Guillaud and Joffroy, 1995). T he stagnated water can erode the Earthen mortar is therefore considered the most suitable for CEB masonry (Guillaud and Joffroy, 1995; Walker, 1999; Venkatarama and Gupta, 2006; Morton, 2008). Venkatarama and Gupta (2006) reported earthen mortar to provide between 15 50% Type Compressive S trength MPa(psi) Mix Proportions (OPC/Hydrated Lime/Masons S and) Characteristics Typical A pplications M 17.23 (2500) 1/ 1/ 4 /3 High compressive and tensile bond strength Exterior walls, at or below grade S 12.41 (1800) 1/ ½/ 4 ½ Moderate compressive and tensile bond strength Masonry foundations, manholes, sewers, and brick pavement N 5.17 (750) 1/1/6 Low compressive and ten sile bond strength Exterior walls, above grade O 2.41 (350) 1/2/9 Very low compressive and tensile bond strength Non load bearing/interior walls

PAGE 60

60 more bond strength compared to cement or cement lime mortar in CEB masonry. After studying the influence of different mortar types on the bond strength of CEB masonry, Walker (1999) concluded that it was more beneficial using soil:cement (earthen mortars) rather than cement:lime:sand mortars especially with weaker blocks. CEB Masonry Systems CEB masonry systems can be constructed using blocks and mortar o r using interlocking (mortarless blocks) (Jayasinghe and Mallawaarachchi, 2009). When CEB masonry construction using mortar is unreinforced, bond strength and ultimately composite strength of the masonry system is influenced by both block and mortar prope rties. Factors such as the porosity, pore size distribution, absorption, and moisture content of the blocks, and the water retention, water content and workability of mortar during the construction process (Venkatarama Reddy and Gupta, 2006). Very little i s currently known about the structural behavior of CEB masonry systems compared to conventional masonry. The general assumption is that CEB masonry behaves in similar manner to conventional masonry (Tennant et al. 2013). The use of interlocking mortarless bricks/blocks for house construction made from CEBs and fired bricks, was pioneered in Africa, Canada, the Middle East and India masonry construction used as a low cost bu ilding material. The manufacturing and construction process requires no special skills and can be performed by inexperienced labor making it an attractive building material for developing countries (Stirling, 2011). Interlocking CEBs are typically produced using machines that guarantee good face texture therefore reducing or eliminating the need for rendering. Often, only joint sealing for protection from the weather is needed. The reduction and/or omission of joint mortar

PAGE 61

61 and rendering reduces cost and ac celerates the construction process but can also introduce some structural weaknesses ( Ramamurthy and Nambiar, 2004; Kitingu, 2009 ). Interlocking or mortarless CEBs come in different sizes and shapes depending on the types of presses used in their producti on. The Aurum 3000 press for example is a manual vertical block press manufactured by Aureka in Tamil Nadu, India. It can be used with 15 different molds that can be configured to produce about 75 different regular and interlocking CEBs some of which can b e seen in Figure 2 2. The Aurum block press is designed to allow for installation of reinforcing steel where necessary. The blocks typically weigh between 9 Kg and 10 Kg ( Kitingu, 2009; Bland 2011). Another type of interlocking blocks are the hydraform blo cks (Figure 2 3). The Hydraform series of block presses are manufactured by Hydraform in Johannesburg, South Africa. As is the case with the Arum block presses, the hydraform press comes in different models. The Hydraform blocks are designed to be dry stac ked but can also be constructed for use with vertical and horizontal reinforcing by means of a special press insert. Some CEBs come with grooves and holes making reinforcement embedding less cumbersome (Rai, 2007). Systems for reinforcing CEB walls have b een developed in order to improve the structural performance. Most of the regions exposed to high seismic activities have imposed norms and developed codes that require the use of vertical and horizontal reinforcement for all masonry. Examples can be found in Peru, Turkey, and the United States (Morton, 2008; Kestner et al. 2010). Systems typically used in these countries use reinforced concrete, wooden or steel ring (bond) beams, and reinforcement of the

PAGE 62

62 corners of walls and opening frames ( Figure 2 4) (Elizabeth and Adams, 2005; Morton, 2008; Kestner et al. 2010). Figure 2 2. Aura m 300 b lock variations (AEI, 2010 ) Figure 2 3. Hydraform dry staking system ( Bland, 2011 )

PAGE 63

63 The use of reinforcement considerably improves the tensile and bending strength o f masonry (Guillaud & Joffroy, 1995). Reinforcement is critical even in regions, which are not exposed to seismic risk particularly for thin wall construction. Reinforcement reduces the danger of cracking, which is typically the effect of differential sett ling, shrinkage; swelling, thermal expansion, rotation or shearing stress (at openings and walls junctions), the lateral force of the wind, sloping roofs, arches or vaults (Guillaud& Joffroy, 1995; Morton 2008; MSJC, 2008). Reinforcement of earthen masonry can sometimes be localized (placed in the weakest parts of the masonry structure) either at the corners or at the reveals of openings (Morton, 2008). System level evaluation of CEB masonry is outside the scope of the present research. Studies that have ex amined CEB masonry systems include Jayasinghe and Mallawaarachchi (2009), Kitingu, 2009; Bland (2011), Stirling (2011), Herskedal (2012), Tennant et al. (2013). Figure 2 4. Sanford Winery, California. Built in steel reinforcement, vertically sleeved to a ttach to a wallhead bond beam (Morton, 2008)

PAGE 64

64 Chapter Summary There are ongoing efforts directed at improving the performance of earthen construction methods and techniques including CEBs. The literature reviewed showed some of the efforts at improving the strength of CEBs to allow for their use in modern day applications. These efforts have been in both developed and developing countries. Earth, in common use for architectural construction for thousands of years and around the world, has in recent years at tracted renewed attention, including in industrialized countries, as a healthy, environmental friendly and economical building material. As a result, an increasing number of buildings constructed of rammed earth, soil blocks, mud bricks or adobes are being erected not just in the hot dry and temperate zones but also in the colder climates of Europe, North America, China and Japan (Minke, 2009). S ome drawbacks to the use of CEB masonry systems are the lack of applicable codes and standards, the lack of adeq uate data to help predict performance, their very brittle nature, and low strength (compressive and flexural). Reinforcing earthen construction/masonry materials with natural fibers has been traditionally used to prevent desiccation cracks and improve tens ile strength. Plant fibers have been the preferred fibers used for CEB reinforcement in studies because they are abundant, cheap, and renewable. However, their effects on CEB properties have been variable. They have either increased or reduced compressive and flexural strength and such variability in strength properties may be linked to the adhesion of fibers to the soil matrix, the hydrophilic character of the fibers, and the dispersion of fibers within matrices (Elenga et al. 2011). Where synthetic fibers have been used, they have been mainly focused on fibers derived from post consumer plastic waste products thus introducing the possibility of variations in fiber quality. Results of studies using these synthetic fibers deri ved from post consumer plastic waste has also been variable. There is therefore the need for a

PAGE 65

65 thorough investigation into fiber reinforced CEBs that can be easily replicated to help identify key fiber characteristics that can help improve CEB performance.

PAGE 66

66 CHAPTER 3 METHODOLOGY This chapter presents materials and material properties, instrumentation, testing procedures, and the scope of the experimental testing program used to evaluate the effects of polypropylene fibers on the mechanical properties of CEBs. The experimental testing p rogram employed in this study was primarily to evaluate the strength properties (compressive, 3 point bending, and flexural) of polypropylene fiber reinforced CEBs. The experimental work included designing suitable mortar compatible with the polypropylene fiber reinforced CEBs and SEM analysis of fractured fiber reinforced CEB surfaces. A flow chart of CEB production and testing is provided in Figure 3 1. The work was divided into three phases a s follows: Optimization of P oly propylene Fiber R einforced CEBs In this phase, readily available soil was classified, evaluated, and a suitable stabilizer for the soil identified. Best practices of fiber reinforcement in the concrete and geotechnical industries and earthen construction techniques were adopted to prod uce polypropylene fiber reinforced CEBs and beams. CEBs/beams were produced at different fiber mass fractions for compressive, 3 point bending, and flexural strength tests. A preliminary set of experimental work was undertaken to determine an appropriate f iber length , and the number of specimens to produce for each mix design. Fabrication and testing of cured specimen was done using the following recommendations and standards: Specimen production, [ACI 544.1R, 1996; ASTM C1116/C 1116M; Rigassi, 1995; Maher and Ho, 1994; Li, 2005] Flexural strength test, [ ASTM C293 10 ; ASTM C1609 12 ] Compressive strength test, [ASTM C140 12; ASTM D1634 00]

PAGE 67

67 Earthen mortar s pecimen production, [ACI 544.1R, 1996; ASTM C 270; ASTM C 780; Morton, 2009; Walker, 1999] Compressive strength test, [ASTM C109 13] Prism Compressive Strength test, [ASTM 1314 12] Prism flexural bond strength test, [ASTM E518 12] Figure 3 1. Flow chart for specimen production and testing

PAGE 68

68 Microstructure A nalysis This phase involved the analysis of fract ured surfaces during flexural strength testing using SEM. This was to understand the failure mechanism of the fiber reinforced CEBs and how fibers help prevent catastrophic failure. Characterization of Constituent Materials This section gives details of th e sand properties, cement, and fibers used for CEB/beam , and earthen mortar production. Cement Ordinary Portland cement was used for the production of all specimens. The chemical composition of the cement is provided in Table 3 1. Soil The soil used for block production was obtained locally in the Gainesville/Newberry, Florida area. The grain size distribution of the soil was determined using the American Association of State Highway and Transportation Officials (AASHTO) soil classification system M 145/ ASTM D3282. Liquid limit, plastic limit, and proctor compaction tests were run on the soil. Liquid l imit The liquid limit (LL) of soil usually expressed as a percentage , is defined as the water content at the boundary between the liquid and plastic state s of the soil (Minke, 2009). Soil typically begins to show some resistance to shearing at the LL. Using a Casagrande, the LL of the soil used was determined using ASTM D4318 05. The following steps were used in determining the LL of the soil: 1. A soil mixtu re was made by mixing the soil with distilled water to form a paste.

PAGE 69

69 2. A groove was formed in the soil by using a device held perpendicular to the surface of the bowl. 3. The crank of the Casagrande was then turned at a rate of 1.9 to 2.1 drops per second. The cranking was continued until the two halves of the soil specimen came together at the bottom of the groove along a distance of 13 mm (1/2 in). 4. The number of drops it took to close the groove was recorded and a slice of the soil mixture was removed to dete rmine its water content. The preceding steps were repeated with soil mixtures at slightly higher and lower water contents. The LL was determined to be the water content at which it took 25 blows to close the groove over a distance of 13 mm. The results o f the test are presented in Table 3 2. Table 3 1 . Chemical and mineralogical composition of cemen t Chemical Composition Value (%) SiO 20.47 Al O 5.19 Fe O 4.49 CaO 63.49 MgO 1.10 SO 2.55 Na 2 O 0.05 K 2 O 0.28 Mineralogical Composition Value (%) C S 54.38 C 2 S 17.98 C A 6.16 C 4 AF 13.66

PAGE 70

70 Plastic l imit The plastic limit (PL) of soil is the water content at the boundary between the plastic and semi solid states of the soil , typically expressed as a percentage (Minke, 2009). Plasticity defines the extent to which a soil can be distorted without any significant elastic reaction; typically cracking or crumbling occurring (Rigassi, 1995). At the PL, soil ceases to be plastic and becomes crumbly. The PL limit of the soil used for block production was de termined using ASTM D4318 05 in following the steps below: 1. About 2.0 g specimen was taken out of the specimen used for the LL test. 2. Typically, the specimen is rolled between the palms on a ground glass plate to form a thread of uniform diameter of about 3 However, in this case, the thread could not be rolled down to 3.2 mm without crumbling . The PL was therefore not determined. This was because of the high sand content of the soil (Table 3 2). Plasticity index (PI) is the difference between t he liquid limit (LL) and the plastic limit (PL). It describes the range of moisture content over which the soil is in a plastic condition (Burroughs, 2001). Since the PL for the soil could not be determined, PI could also not be determined. Proctor Compac tion t est The proctor compaction test is a laboratory compaction method used in determining the relationship between molding water content and dry unit weight of soils. The factors that affect the extent of compaction are compaction effort, soil type and g radation, moisture content, and dry density. The Proctor compaction test is the most widely used test for assessing the compactibility of a soil. Different soils have different optimum moisture contents meaning the conditions required to achieve optimal co mpaction would differ with different soils (Burroughs, 2001). The maximum dry density of the soil used for CEB production was determined using the proctor

PAGE 71

71 compaction test following the procedure outlined in ASTM D698 07. The procedure that was used in dete rmining the maximum compaction at optimum moisture content of soil is outlined below: 1. A soil sample 46 kg (10 lbs) passing No.4 sieve was prepared and the Proctor mold (1/30 without the base and extension weighed. 2. sample was then placed in the Proctor mold in 3 layers and compacted using 25 well distributed blows of the Proctor hammer . 3. The collar of the apparatus was detached without disturbing the soil inside the mold and the base removed to determine the weight of the mold and compacted soil. 4. The compacted soil was removed from the mold and a 20 30 g sample of the soil taken to fin d the moisture content. 5. The process was repeated for various moisture contents (7, 9, 11, and 13%) and the corresponding dry densities determined. The next step was plotting a compaction curve using the dry density and moisture content. The peak of the cur ve represents the maximum compaction at optimum moisture content. The physical properties of the soil are presented in Table 3 2. Table 3 2 . Physical properties of soil Property Composition Liquid Limit (%) Plastic Limit (%) Plasticity Index (%) Sand (%) Clay (%) Silt (%) Optimum Moisture Content Maximum Dry Density 33% (non plastic) 87.3% 12.2% 1.5% 9% 1784.5 kg/m 3

PAGE 72

72 Fibers The fibers used are synthetic polypropylene fibers obtained from the BASF Co rporation. The fibers (Figure 3 1) are composed of two circular filaments that are cross fiber with an embossed surface with depths from peak to valley of about 0.005 to 0.006 mm. The deformations provide mechanical anchor age between the fiber and matrices. The fibers have been successfully used in shotcrete and slab on grade applications to improve flexural toughness, impact resistance, residual strength, and durability (Wilson and Abolmaali, 2014). The physical properties of the fibers as provided by the manufacturer is presented in Table 3 3. Figure 3 2 . MasterFiber MAC Matrix fibers. A) Photograph of fibers. B ) SEM image of fiber showing cross linked filaments and grooves (Photo courtesy of author) .

PAGE 73

73 Table 3 3 . Physical properties of MasterFiber MAC Matrix (polypropylene fibers) as provided by the manufacturer . Specimen Preparation M aterial p roperties of polypropylene fiber reinforced composites are affected by fiber volume, fiber geometry and length (aspect ratio), fiber surface conditions, method of production, and composition of matr ices (Banthia and Gupta, 2006; ACI 544.1R, 2010). The init ial set of experiments focused on the influence of fiber length on the flexural and compressive strength of tested specimens. Determining the fiber length to use was part of the fiber reinforced CEB optimization process. All the other fiber parameters were kept constant at this stage . The fiber length that yielded the best results were used for follow up experimental work. Subsequently, t wo sets of specimens were produced for this research. One set was produced in a lab setting and another set in the field. The specimens produced in the field setting were CEBs for compressive strength and 3 point bending strength testing . Specimens produced in the lab were short flexural beam s fabricated with CEB matrices. The short flexural beams were tested to evaluate th e flexural performance of the fiber reinforced CEB matrices. Property Des cription Configuration Fiber Type Material Specific Gravity Melting Point Ignition Point Length Water Absorption Tensile Strength Alkali Resistance Chemical Resistance Color Electrical Conductivity Stick like fiber Embossed 100% virgin polypropylene 0.91 320 F (160 C) 1094 F (590 C) 54 mm (2.1 in.) Nil 85ksi (585MPa) min Excellent Excellent White, translucent Low

PAGE 74

74 F lexural Beams for Fiber Length Determination The fibers were used as both the commercially available length of 54 mm fibers and lengths of 27 mm cut from full length fibers (Prochazka et al., 2 010). The mix proportion is presented in Table 3 4. Details of specimen production are presented in the section on Laboratory Produced Blocks. Table 3 4 . Matri x mix proportions Mix Type OPC Content * (%) Fiber Content by Mass* (%) (54 mm) (27 mm) #1 8 # 2 8 0.2% # 3 8 0.2% # 4 8 0.14% 0.06% # 5 8 0.06% 0.14% *Percentage composition in relation to mass of dry soil F ield Produced Blocks Two sets of blocks were produced one set for compressive strength testing and the other for the 3 point be nding test. In order to remove lumps from the dry soil that was to be used for block production, the soil was run through a manual sifter with a 3.40 mm² mesh size. With the fiber reinforced matrices, the fibers were gradually introduced into the mix after the initial hand dry mix of sand and OPC had been observed to be thorough. After an additional 20 minutes, the matrix appeared uniform and thoroughly mixed with the fibers well dispersed. The dry mix was watered gradually in a uniform manner during mixin g. Approximately 0.6 kg of water was required for every 45.36 kg of matrix. The mixing process continued for 10 to 20 more minutes depending on the quantity of fibers present. The fiber reinforced matrices were produced with fibers at 0.2, 0.4, 0.6, 0.8, a nd 1.0 mass fractions in relation to mass of dry soil . The mix design

PAGE 75

75 is presented in Table 3 5 . When the fiber content exceeded 0.6% (by mass ), it took a longer time to attain good fiber dispersion and homogeneity. A hydraulic operated block making press was used to produce the blocks. The press exerted a maximum load of approximately 4355 kg generating a pressure of 1.6 MPa on the matrices in a 30 second process. After compression, the blocks were ejected from the mold (Figure 3 3), moved, and placed on p allets outdoors. The nominal dimensions of blocks produced were 191 mm x 203 mm x 121 mm for the compressive strength testing and 229 mm x 203 mm x 121 mm for 3 point bending testing. These weighed 9.07 kg and 11.34 kg respectively. Slightly longer blocks were required for the 3 point bending test. The blocks were moist cured (sprayed with water) under a tarp for the first 7 days. The blocks were kept under the tarp for the next 21 days without further curing. Testing was done 28 days after production. Tab le 3 5 . Block mix proportions Mix OPC Content* (%) Fiber Content* (%) Water Cement Ratio PP 0.0 8 0.00 0.17 PP 0.2 8 0.20 0.17 PP 0.4 8 0.40 0.17 PP 0.6 8 0.60 0.17 PP 0.8 8 0.80 0.17 PP 1.0 8 1.00 0.17 *Percentage composition in relation to mass of dry soil . L aboratory Produced Short Flexural Beams Dry soil to be used for beam production was passed through a manual sifter with a 3.40 mm² mesh size to remove lumps. A visual inspection was also conducted to ensure that the soil was free from organic m atter. The sifted soil was kept in an oven set

PAGE 76

76 at 93 o C ( Figure 3 4) for 24 hours, after which it was taken out for cooling before production commenced. Matrix mixing was done using a concrete mixer starting with a dry mix of soil and OPC. For the fibe r reinforced samples, the polypropylene fibers were gradually introduced into the dry mix in batches as mixing continued. Fibers were introduced in batches of 0.045 kg every one minute. Mixing continued for an additional 3 minutes after the last batch of f ibers were introduced. Figure 3 3. CEB molding process. A ) Wet mix . B ) Block press. C) Hydraulic pump connected to b lock press . D ) Block being ejected from block press (Photo courtesy of author).

PAGE 77

77 Figure 3 4. Soil drying in oven set at 93 o C . (Photo court esy of author). After the dry mix had been observed to be uniform and consistent with well dispersed fibers, it was watered gradually in a uniform manner while mixing continued for an additional 10 minutes. At this point, the wet mix was visually deemed t o be uniform and thoroughly mixed with the fibers well dispersed. Approximately 0.6 kg of water was required for every 45.36 kg of matrix. The fiber reinforced matrices were produced with fibers at 0.2, 0.4, 0.6, 0.8, and 1.0 mass fractions in relation to mass of dry soil ( Table 3 6 ).

PAGE 78

78 Table 3 6 . Short flexural beam mix proportions Mix OPC Content* (%) Fiber Content* (%) FB 0.0 8 0.00 FB 0.2 8 0.20 FB 0.4 8 0.40 FB 0.6 8 0.60 FB 0.8 8 0.80 FB 1.0 8 1.00 *Percentage composition in relation to ma ss of dry soil A heavy duty steel mold lined with form board was filled with matrix, covered with a form board lid, and compressed 51 mm down at a rate of 223 N/min. Compression was done using a Test Mark CM 500 series compression machine with a maximum c ompression capacity of 2,224 kN. All beams were produced using 8.62 kg of matrix as part of an effort to minimize variations in beam densities. The maximum compression pressure for each beam was 1.6 MPa. The nominal dimensions of beams produced were 413 mm (length) x 102 mm (width) x 102 mm (height). After de molding, the fresh samples were left in situ for 24 hours prior to being moved and stacked. The beams were moist cured under plastic sheets for 7 days and then air cured under the plastic sheets for th e next 21 days. Testing was done28 days after fabrication. Pictures of the production process is presented in Figure 3 5. E arthen M ortar The same sifted oven dried soil used for beam production was used for mortar production. Mixing was done by hand. Cube molds of 50 mm were cast for compressive strength testing ( Figure 3 6 ). Three mix proportions with OPC content that was the same, one and a half, and twice the OPC content used for block production were

PAGE 79

79 produced ( Guillaud and Joffroy, 1995). The mix de sign is presented in Table 3 7 . M olding the specimens was done within 2 min after completion of mixing of the mortar. Mortar was placed in the cubes in layers of about 25 mm and tamped 32 times in 4 rounds. The fresh samples were covered with polythene she ets and moist cured for 7 days. Testing was done 28 days after casting. M asonry P risms Two sets of prisms were cast; two block prisms for compressive strength testing and seven block prisms for flexural bond strength testing. The two block prisms were ca st by assembling two CEBs, one on top of the other, using earthen mortar as the bonding material o n the contact surface of the CEBs. A mortar joint thickness of about 10 mm was used in all cases. Two 203 mm x 203 mm CEBs were cast, bonding one on top of th e other using the earthen mortar. The CEBs used for casting prisms were prewetted before laying. Prisms for the flexural bond strength test were cast using seven half CEBs with mortar joints of about 10 mm thick. The prisms were cured under plastic sheets and tested 28 days after fabrication ( Figure 3 7 ). Specimen Testing Preliminary Tests for Fiber Length Selection Flexural strength testing was done using ASTM C293/C293M 10. A total of 4 samples were tested for each mix design. A Tinius Olsen compress ion machine with a maximum load capacity of 400 KN was used. The machine was set up with a mounting jig (yoke) to record mid span deflection. Two linear variable displacement transducers (LVDTs) with a stroke of ±17.8 mm were mounted on either side of the centerline of samples to record mid span deflection ( Figure 3 8 ). The rate of compression was set at

PAGE 80

80 267 N/min until failure for the un reinforced samples, and until a displacement of 10 mm was reached for the fiber reinforced samples. The values for f lexural strength w ere computed using Equation 3 1. Figure 3 5. Beam production. A) Matrix in concrete mixer. B) Matrix in steel mold. C) Matrix under compression. D ) Compres sed matrix in mold. E ) St ripped beams ready for curing. F ) Beams being cured under plastic sheet (Photo courtesy of author). .

PAGE 81

81 Figure 3 6 . Freshly cast cubes (Photo courtesy of author). Table 3 7 . Mortar mix proportions Mix Type Mortar Proportion by Weight (C:S)* Water Cement Ratio EM # 1 1:11.5 2.70 EM # 2 1:8.0 1.80 EM # 3 1:6.0 1.38 *C=cement; S=soil Figure 3 7. CEB prisms . A ) Two block prisms freshly cast . B ) Seven block prism being cured under plastic sheet (Photo courtesy of author).

PAGE 82

82 Figure 3 8 . Flexural strength test setup . A ) Photograph . B ) Schematic drawing (Photo courtesy of author). ( 3 1) Where; R = modulus of rupture, [N/mm²]; P = maximum applied load indicated by the testing machine, [N]; L = span length, [305 mm]; b = average width of specimen at the fracture [mm]; d = average depth of specime n at the fracture [mm]. Portions of blocks broken in fracture were used for compressive strength testing in accordance with ASTM D1634 [ Standard Test Method for Compressive Strength of Soil Cement Using Portions of Beams Broken in Flexure (Modified Cube Me thod) ] . Cubes of size 102 mm x 102 mm were obtained from tested samples with a band saw. These were capped with a thin layer of gypsum plaster prior to testing to create an even surface to facilitate the uniform transfer of stresses between platens and spe cimens ( Figure 3 9) . A total of five cubes were tested for each of the four mix designs. The

PAGE 83

83 testing was done through applying uniaxial compression using a Tinius Olsen compression machine with a maximum load capacity of 400 kN. The rate of compression wa s set at 623 N/s until failure. Block Compressive Strength Blocks for the compressive strength test were capped with a 3 mm thick plywood on the top and bottom faces (Walker, 2004). This created an even surface thus facilitating uniform transfer of stre sses between platens and specimens. It also reduced friction between platens and specimen surfaces. The blocks were tested under uniaxial compression using a Gilson concrete compression machine with a maximum load capacity of 224 kN (Figure 3 10 ). The rate of compression was set at 623 N/s until failure. The test set up and procedure complied with ASTM C140. Figure 3 9. Compressive strength testing . A ) Beams broken in flexure . B ) Sawing off portions (cubes) of broken beams. C ) Capping cubes with gypsum pla ster . D ) Test setup (Photo courtesy of author).

PAGE 84

84 Figure 3 10. Compressive strength test setup (Photo courtesy of author). Block 3 Point Bending Test An Instron Universal testing machine with a maximum load capacity of 150 kN was used in running the 3 point bending tests. Blocks were simply supported and subjected to a single point loading during testing (Morel et al. 2002; Morel et al. 2007; Villamizar et al. 2012). The rate of compression was set at 267 N/min until an 80% load drop was reached. The values for the 3 point bending strength were computed using E quation 3 1 . A schematic drawing of the 3 point bending test set up is presented in Figure 3 11 . Figure 3 11. 3 point bending strength test .setup A) Photograph. B) Schematic drawing (Photo courtesy o f author).

PAGE 85

85 Beam Flexural Strength Test Beams were tested according to ASTM C1609. Testing was run using a Tinius Olsen compression machine with a maximum load capacity of 400 kN. The test procedure consisted of a simply supported beam with third point loa ding. The test setup included a mounting jig (yoke) to record mid span deflection. Linear variable displacement transducers (LVDTs) with a stroke of ±17.8 mm were mounted on each side of the centerline of beams to record mid span deflection ( Figure 3 1 2 ). Beams were rotated through 90° from their casting position before testing to minimize the influence of casting direction on results. Leather shims were placed on the specimen contact surface to provide an even surface and eliminate gaps during load app lication. Loading was deflection controlled at a rate of 0.25 mm/min. The test setup is shown in Figure 3 12 . The load deflection curves obtained during testing were used to calculate first peak strength (f 1 ) and peak strength (f p ) (Eq uation 3 2 ), equivale nt flexural strength ( f e ) (Eq uation 3 3 ), residual strength at deflections of L /600 ( ) and L /150 ( ) (Fig ure 3 13 and 3 14 ), and flexural toughness ( ); area under load the load deflection from 0 to L /150 (2.03 mm) . Figure 3 12 . Flexural strength test set up . A ) Photograph . B ) Schematic drawing (Photo courtesy of author).

PAGE 86

86 Figure 3 13. Schematic diagram of load deflection curve for calculating first peak strength, residual strength, and toughness where first peak load is equ al to peak load (ASTM C1609). Figure 3 1 4 . Schematic diagram of load deflection curve for calculating first peak strength, residual strength, and toughness where peak load is greater than first peal load (ASTM C1609).

PAGE 87

87 ( 3 2 ) ( 3 3 ) Where; f 1 = first peak strength; f P = peak strength ; P 1 = first peak load; P p = peak load ; = residual load at net deflection of L/60 0; = residual l oad at net deflection of L/150; b = average wid th of specimen at the fracture; d = average de pth of specimen at the fracture; = flexural toughness (Nmm); = ne t deflection at first peak load; = net deflection at peak load; = net deflection of L/150. Prism Compressive Strength Test Prisms for compressive strength testing were cast and tested using ASTM C 1314. Test specimens for compressive strength testing were capped with a steel p late before tes ting (Figure 3 1 5 ). The prisms were tested under uniaxial compression using a Forney FX 250/300 compression test machine with a maximum load capacity of 224 KN. The rate of compression was set at 620 N/s until failure.

PAGE 88

88 Figure 3 1 5 . Two block prism compres sive strength test setup (Photo courtesy of author). Prism Flexural Bond Strength Test The third point loading method from ASTM E518 12 was used in determining the flexural bond strength of the seven block prisms. A schematic drawing of the test setup is presented in F igure 3 1 6 . Flexural bond strength was calculated using E quation 3 4 . Figure 3 1 6. Flexural bond strength testing . A) Photograph of test setup. B ) Schematic drawing of test setup (Photo courtesy of author).

PAGE 89

89 (3 4 ) Where; R = gross area modulus of rupture, MPa, P = maximum applied load indicated by the testing machine, N, Ps = weight of specimen, N, l = span, mm, b = average width of specimen, mm, and d = average depth of specimen, mm. Micro structure Analysis Scanning Electron Microscopy (SEM) Scanning Electron Microscopy (SEM) is a microstructural analysis technique widely used in concrete. SEM use allows for the observation and characterization of heterogeneous organic and inorganic mater like images of the surfaces of a very wide range of materials (Acquaye, 2006). In the concrete industry, SEM use is primarily to study the effects of deterioration of concrete or its performance characteristics, qualitative phase identification, grain morphology, and distribution pattern ( Ramachandra n and Beaudoin , 2001). Cementitious matrices are quasi brittle and have low tensile strength and strain capacities. A practical means of enhancing the performance of such matrices is the inclusion of various types of fibers to enhance ductility, strength, toughness, and resistance to impact loads (Atahan et al. 2013). An understanding of fiber, matrix, and f iber matrix interaction is therefore essential to using fibers to enhance the performance of cementitious matrices. This is because failure of fiber reinforced cementitious

PAGE 90

90 matrices largely depends on fiber matrix interactions and the pullout (slip) charac teristics between fibers and matrices (Li and Kanda, 1998; Bindiganavile and Banthia, 2005). Microstructure analysis of CEBs is a subject area that has the potential of considerably enhancing the understanding of CEB performance. Understanding fiber matrix interactions of fiber reinforced CEBs, response of CEBs to different environments, and crack propagation and failure mechanisms of CEBs are some of the potential uses of microstructural analysis techniques such as SEM in enhancing the understanding of CEB s. The use of microstructural analysis techniques especially in fiber reinforced CEBs could help understand how fibers bridge micro cracks thereby preventing crack propagation and catastrophic failure. A micro level understanding of matrix interactions wi ll result in the formulation of appropriate macro level systems such as appropriate exposure conditions necessary for durability. Experimental w ork . All SEM analyses were conducted at the Major Analytical and Instrumentation Center at the University of Fl orida. An SEM JSM 6400 microscope ( Figure 3 1 7 ) was used in this research to analyze the fracture surfaces of specimens that were broken in flexure. Micrographs of the fractured surfaces were captured to enable evaluation of the failure mode of the fib ers, an important parameter for the evaluation of energy absorption (Atahan et al. 2013). Samples used for the SEM analysis were obtained from fracture surfaces of fiber reinforced beams broken during flexural strength testing. Samples of size 38 mm X 38 m m were obtained from 102 mm on specimen stubs for the specimen chamber of the SEM JSM 6400 microscope

PAGE 91

91 (Figure 3 1 8 A , B ). Unused fibers were also coated with gold palladium f or imaging (Figure 3 1 8 C , D ). The coating was to minimize or eliminate the accumulation of electric charges that build up rapidly on non conducting specimens when scanned using a beam of high energy electrons. Figure 3 1 7 . SEM JSM 6400 microscope for cap turing micrographs of fractured surfaces (Photo courtesy of author) . Quality Control/ Assurance Procedures As stated earlier, there are no standard protocols for producing and testing CEBs. Additionally, there are no standard protocols for fiber addition in to CEB matrices. The inherent variability in soil properties makes it important to adopt a systematic approach to producing quality CEBs. This section presents suggested best practice procedures in preparing soil, producing, and testing CEBs. These procedu res were identified to yield best results based on observations made during specimen production and test results.

PAGE 92

92 Figure 3 1 8 . SEM sample preparation . A ection of a fractured surface. B ) Gold palladium coated 38 mm x 38 mm sample mounted on a specimen stub r eady for the specimen chamber. C ) Unused fibers. D ) F ibers cut into lengths of 15 mm, coated with gold palladium, and mounted for im aging (Photo courtesy author). Soil Preparation Soil preparation is key to CEB production. The soil to be used for CEB production underwent an initial visual inspection (Figure 3 19 ). This was to assess the amount of visible organic material present in th e soil. Soils with high organic matter (containing roots, moss, sticks, leaves, etc.) are weaker in strength, have a high compressibility and therefore not suitable for CEB production (Rigassi, 1995; Coduto, 1999). When soils with a high organic content ar e used for CEB production, there is potential for the creation of voids within the CEBs when the organic matter breaks down. This has the tendency of weakening blocks. The soil used in this research was s i ft e d to rid it of

PAGE 93

93 excessively large elements such a s lumps, stones, and organic material with a 3.40 mm² mesh size (Figure 3 20 ). This was identified as one of the quality control procedures that reduced variation in test results of the final product. Figure 3 19 . Visual inspection of soil to be used for CEB production (Photo courtesy of author). Figure 3 20 . Soil sifting for CEB production (Photo courtesy of author).

PAGE 94

94 Fiber Addition The protocol adopted for fiber addition for the field produced CEBs was different from that used for the lab oratory prod uced short flexural beams. With the field produced blocks, mixing was done by hand. The fibers were gradually introduced into the mix after the initial hand dry mix of sand and OPC had been observed to be thorough. Fibers were added in quantities of 0.04 5 kg and thoroughly mixed. Mixing in the fibers took about 20 minutes. Wet mixing took between 10 to 20 minutes depending on the quantity of fibers added. The more the fibers, the longer the mixing process took. When the fiber content exceeded 0.6% (by mas s ), it took a longer time to attain good fiber dispersion and homogeneity. Workability was negatively affected as fiber quantity increased. The surface of the CEBs produced with fibers above 0.6% appeared rough in the fresh state and poro us after curing ( Figure 3 21 ). Strength also began to decline after fiber content exceeded 0.6%. Based on these findings, it is recommended not to exceed 0.6% of polypropylene fibers by mass for hand mixed CEB matrices. An ideal range of polypropylene fibers to add woul d be between 0.4 and 0.6%. With the lab oratory produced short flexural beams, mixing was done with a concrete mixer. For the fiber reinforced beams, the polypropylene fibers were gradually introduced into the dry mix of OPC and sand in batches of 0.045 kg every one minute. Mixing continued for an additional 3 minutes after the last batch of fibers were add ed. After the dry mix had been observed to be uniform and consistent with well dispersed fibers, it was watered gradually in a uniform manner while mixin g continued for an additional 10 minutes. Fiber balling was observed as fiber quantity exceeded 0.6% by

PAGE 95

95 mass . Fibers stuck out on the surface of beams produced with 0.8 and 1.0% fibers (Figure 3 22). Figure 3 21 . Cured CEBs showing porous surfaces with fi bers sticking out (Photo courtesy of author). Figure 3 22 . Short flexural beams showing with porous surfaces and fibers sticking out (Photo courtesy of author).

PAGE 96

96 Moisture Content After the wet mix of the unreinforced matrix had been visually assessed and d dropping the ball from a height of one meter onto a hard surface. If the ball c ompletely disintegrates, the mix is too dry; if it breaks up into four or five pieces, the moisture content is right; if it flattens out without breaking, or breaks into 2 pieces, the mix is too wet ( Figure 3 2 3 ) (Rigassi, 1995). Figure 3 2 3 . The d rop test (UN Habitat, 2001). Compression and Curing The retention time lag between mixing and compression did not exceed ten minutes (Rigassi, 1995 ; Walker, 2004 ). This protocol was observed to avoid flash set (ACI E 701.E3, 2001) . Curing was done by spray ing the blocks with water from the second day up to the seventh day. This was done while keeping the field produced blocks under a tarp and the lab oratory produced beams under plastic sheets. With the lab oratory produced beams the plastic sheets were taken off seven days before testing.

PAGE 97

97 The removal of the plastic sheets was to prevent condensed water trapped inside the plastic sheets from weakening the beams. The initial moist curing done was to help achieve maximum strength. This is typically common with c ementitious materials. Leaving blocks exposed to hot dry weather conditio ns has the tendency of causing surface cracks on CEBs (Adam and Agib, 2001). Production and testing sequence is presented in Figure 3 23. Figure 3 2 4 . CEB production and testing seq uence

PAGE 98

98 CHAPTER 4 TEST RESULTS AND DISCUSSION This chapter presents and discusses the test results of the experimental work undertaken to determine the influence of polypropylene fibers on soil cement matrices used for CEB production. Preliminary Result s for Fiber Length Selection Flexural Strength Failure of the 20 tested samples occurred at the middle third of the span length. Each tested sample exhibited linear elastic characteristics prior to initial crack, which occurred at peak load. The load defle ction responses of the fiber reinforced samples were different from the unreinforced ones. Typical load deflection curves of the tested samples are presented in Figure 4 1. Mix type #2 yielded the highest average MOR followed by mix types #5, #1, #3, and # 4 in that order. The average MOR of mix type #2 was 26% more than mix #1 (plain matrix). Mix #4 yielded the lowest MOR, which was 24% lower than mix #1 (plain matrix). The MOR for Mix #3 was 11.8% lower than mix #1 (plain matrix), while mix #5 was 2.5% hi gher than mix #1 (plain matrix). In general, the fiber reinforced matrices performed better in post crack behavior compared to the plain matrix. The findings also suggest that fibers affect the brittle behavior of the matrices. The unreinforced samples ex hibited catastrophic failure in all instances. None of the fiber reinforced matrices underwent complete failure even at 10 mm deflection. There was an observation of the fibers bridging the cracks ( Figure 4 1) explaining why there was not catastrophic failure.

PAGE 99

99 Figure 4 1. Typical load deflection curves for blocks . The blocks reinforced with 54 mm long fibers had the best performance. The increase in peak load recorded for the specimens reinforced with the 54 mm fibers compared to the unreinforced sp ecimens can be attributed to the fibers contributing to the bonding of particles surrounding individual fibers. This opposes particle movement and delays crack formation (Namango, 2006; Tang et al. 2006; Elenga et al. 2011). Similar to the ob servations mad e in this study, Bagherzadeh et al. (2012) reported a flexural strength increase in polypropylene fiber reinforced lightweight cement composites compared to the unreinforced composites. They also observed longer polypropylene fibers (12 mm) to perform bett er in flexural strength compared to shorter fracture section (Zhang et al. 2010). Kaufmann et al. (2004) and Bagherzadeh et al.

PAGE 100

100 (2011) established a relationship betwee n increased aspect ratio of polypropylene fibers and the ability of fibers to bridge micro cracks. In this study, the 54 mm (aspect ratio 50 ) fibers had a higher aspect ratio compared to the 27 mm fibers (aspect ratio 2 5 ). The maximum post crack load for t he matrices reinforced with 54 mm fibers was about 40% of the peak load recorded for the matrices. Extensible polypropylene fibers do not totally pull out of matrices when composites reach peak strength. Gradual fiber slipping and stretching results in a h igh post peak strength even at high deformation levels (Consoli et al. 2009). Comparing the results from the different matrices in this study, matrices with different proportions of the shorter fibers (27 mm) did not sustain as much fiber slippage as matri ces with only 54 mm fibers. In sandy soils as wa s the case in this study (87.3% sand), the primary reinforcing mechanism is load transfer from soil to fiber through interface friction. Increasing fiber length or aspect ratio therefore result ed in a highe r surface area, thus providing greater interface frictional resistance between fiber and soil (Maher and Ho, 1994). In this study, the 54 mm (aspect ratio 50) fibers had a higher aspect ratio and performed better in MOR compared to the 27 mm fibers (aspect ratio 25). The results of mixes # 4 and 5 (different proportions of both 54 and 27 mm fibers) varied. Composites with fiber lengths shorter than the critical length fail without fiber fracture, since the fibers are not long enough to generate fracture. Fa ilure in such cases is governed by fiber pullout followed by matrix fracture. On the other hand, failure of composites with fiber length longer than the critical length is governed by fiber fracture (Akkaya et al. 2000). This explains the varying results o btained for mixes with different proportions of both long and short fibers compared to mixes produced with fibers of the same length. The subsequent more

PAGE 101

101 gradual failure after initial crack of all the fiber reinforced matrices suggested an improved perform ance in ductility that can be attributed to the fibers. The MOR for each of the mix types is presented in Table 4 1 . Some polypropylene fiber reinforced specimens were tested to complete failure to investigate fiber matrix interaction at failure surfaces. Both fiber pullout and breakage w ere observed. Table 4 1 . Modulus of rupture of tested beams (MPa) Mix Type 1 2 3 4 Average Std. Dev. COV* (%) # 1 0.75 0.85 0.84 0.81 0.82 0.03 3.86 # 2 1.16 0.90 1.06 1.00 1.03 0.11 10.68 # 3 0.95 0.65 0.63 0.65 0.72 0. 15 21.34 # 4 0.76 0.50 0.58 0.66 0.63 0.11 17.79 # 5 0.89 0.82 0.86 0.79 0.84 0.04 5.23 *Coefficient of variation Compressive Strength of Cubes Obtained from Failed Beams The compressive strength results are presented in Table 4 2 . At 0.2% of polypropy lene fiber content by weight, compressive strength was 84%, 35%, 31%, and 35% higher for mix types #2, #3, #4, and #5 respectively, compared to the un rein forced specimens. Patel et al. ( 2012) attributed similar observations made in compressive strength of polypropylene fiber reinforced concrete specimens to the confinement provided by the fiber bonding. Binici et al. (2005) attributed observed increases in compressive strength of mud bricks to the interface layers and geometric shape of the fibrous reinfor cing materials used. Resistance to fiber sliding has also been cited as a reason for such increases ( Prasad et al. 2012). In this study, the matrix reinforced with only 54 mm fibers (mix # 2) recorded the highest compressive strength. The findings were sim ilar to the findings by Bagherzadeh et al. (2012) who observed longer polypropylene fibers (12 mm) to perform better in compressive strength compared to

PAGE 102

102 shorter fibers (6 mm) at 28 days. They also observed the fiber reinforced matrices in their study to pe rform better in compressive strength compared to the unreinforced ones (Bagherzadeh et al. 2012). Table 4 2 . Cube compressive strength (MPa) Mix Type 1 2 3 4 5 Average Std. Dev. COV* (%) # 1 2.34 2.40 2.10 2.26 2.31 2.28 0.11 4.98 # 2 4.36 4.62 3.94 3.7 4 4.36 4.20 0.35 8.47 # 3 3.37 3.85 3.08 2.29 2.81 3.08 0.59 19.03 # 4 2.57 3.76 2.85 3.07 2.69 2.99 0.47 15.74 # 5 2.48 3.14 3.66 3.08 3.06 3.09 0.42 13.57 *Coefficient of variation There was no significant difference between the values recorded for m ix # 3 and mix #5 even though both had different proportions of 54 mm and 27 mm fibers respectively. Cracks typically formed before peak load was reached during testing. This our ( Figure 4 2). Figure 4 2. Failure o f blocks under compression . A) C r ack pattern during testing. B ) T (Photo courtesy of author) .

PAGE 103

103 Influence of F iber Length on Flexural and Compressive Strength Two sample t tests were conducted to compare the difference in means of compressive strength and the difference in means of flexural strength for the two fiber lengths that were used in the experimental prog ram to determine if the length of the fiber reinforcement had an effect on strength values. The means used for the test were obtained from recorded compressive and flexural strength data for block specimens reinforced with only 54 mm and only 27 mm fibers. The alternative hypothesis that longer fibers increased the mean strength of the blocks was compared against the null hypothesis that length had no effect on the mean strength of matrices. The results of the t tests are summarized in Table 4 3 . The null h ypothesis of there being no difference between the means of compressive strength and no difference between the means of flexural strengths were rejected at the 95% confidence level, indicating that length of fiber had a statistically significant effect on strengths for the 54 mm and 27 mm fiber reinforced C Table 4 3 . Average flexural and compressive strength t test r esults for 54 mm and 27 mm fibers Calculated t value P value Compressive Strength 3.662 0.0032 Flexural Strength 3.327 0 .0079 Based on these initial set of results, subsequent experimental work on optimizing polypropylene fiber reinforced CEBs used the commercially available fiber length of 54 mm. Compressive Strength of CEBs The control sample (unreinforced samples) had an average compressive strength of 4.19 MPa (609 psi) . Compared to the unreinforced blocks, the average compressive strength of the fiber reinforced blocks with 0.2, 0.4, and 0.6%

PAGE 104

104 polypropylene fibers were 10%, 22.5%, and 3.0% higher respectively. The com pressive strength of the 0.8, and 1.0% fiber reinforced blocks were 1.6 and 11.5% lower respectively compared to the unreinforced blocks. The results are presented in Table 4 4 . Table 4 4 . Compressive strength of CEBs * Mass percentage **Coefficient of variation Adequacy of Sample Size In order to determine whether the number of CEBs tested (sample size) was adequate, statistical tests were perform ed using compressive strength of tested CEBs. In determining the adequacy of the sample size, t he initial step was to determine if the recorded results of the best performing fiber reinforced CEB could be attributed to the fi bers used. The difference in mean compressive strengths between the unreinforced CEBs (PP 0.00) and best performing fiber reinforced CEB (PP 0.40) was assessed to determine if it was significant using a two sample t test. The next step was to determine t he required sample size. The aim in determining the sample size required to obtain a significant difference in the compared properties was to reduce type I error , which is the probability of rejecting the null hypothesis of there being no difference betw een measured mechanical properties between fiber Mix Fiber Content (%)* 1 2 3 4 5 Avera ge Std. Dev. COV (%)** PP 0.0 0 0 .00 4.15 4.18 4.28 4.33 4.03 4.19 0.12 2.84 PP 0.2 0 0.2 0 4.90 4.49 4.44 4.38 4.89 4.62 0.25 5.48 PP 0.4 0 0.4 0 5.03 5.40 4.93 4.81 5.55 5.15 0.31 6.13 PP 0.6 0 0.6 0 4.30 3.39 4.08 5.00 4.31 4.32 0.41 9.50 PP 0.8 0 0.8 0 4.0 4 4.13 3.90 4.55 4.01 4.13 0.25 6.00 PP 1.0 0 1.0 0 3.31 3.14 3.76 4.67 3.68 3.71 0.59 16.02

PAGE 105

105 reinforced and unreinforced blocks when such differences actually existed. Determining an adequate sample size would therefore prevent getting a false positive while at the same time increasing the statistic al power of the test s . Using the recorded results , a power analysis was perform ed to determine a statistically adequate sample size to detect a meaningful difference between measured compressive strengths between the two selected CEB types . This was also t o ensure that the sample size used would promote a high degree of replicability of tested CEB properties and the production process used. The parameters of the test are summarized in Table 4 5 . Table 4 5 . Mean CEB c ompressive strength Parameter Mean Compres sive Strength (psi) Number of CEBs Unreinforced (PP 0.00) Fiber reinforced (PP 0.40) Mean Compressive Strength 609 746 5 Standard deviation ( ) 17.31 45.75 5 (PP 0.00 and PP 0.40) determining the sample size from the power analysis. Using Rejection Region:

PAGE 106

106 Reject if or The results of the test indicate a failure to reject the null hypothesis at the 5% significan ce level. Based on the results of the F tes t for equality of variances, the pooled standard deviation of the two samples was calculated using the assumption that Us ing a power of 0.99, an alpha of 0.05: Where is the inverse of the standard normal cumulative dis tribution Using as an estimator of the population variance The power analysis for s ample size is given in the following: Therefore, to detect a significant statistical differ ence between mean compressive/flexural strengths of unreinforced and fiber reinforced CEBs, a minimum of

PAGE 107

107 3 blocks were needed in each of the two dependent sample sizes. The initial set of results obtained used 5 samples which was adequate for the purposes of this research . Influence of fibers on compressive s trength . The relationship between fiber content and compressive strength is shown in Figure 4 3 . This relationship depicts an enhancement in average compressive strength with increasing fiber content until the fiber content reaches an optimal weight fraction of 0.4%. At 0.6% fiber content, average compressive strength was still higher than the average for the unreinforced blocks. The development of strength properties of soil cement fiber mixes mostly depends on the matrix, matrix on the dimension, surface conditions, and quantity of fiber present (Khedari et al. 2005). The number of contact points between fiber and matrix is respon sible for transmitting stress large quantity of fibers reduce strength (UN Habitat, 1992). Above 0.4% fiber content, a declining trend in compressive strength was observed. This occurred matrix and matrix matrix bonds and an increase in fiber fiber interactions. Large amounts of the fiber (more than 0.6%) also resulted in micro fractures at fiber soil interfaces contributing to the observed decreasing trend in compressive strength (Namango, 2006). During block production, th e surface of the blocks with fibers above 0.6% appeared rough in the fresh state and porous after curing. As a result of the stiffness of the fibers used (sticklike nature), a rebound of fibers was observed after de molding. Fibers near the surface of beam s stuck out after the rebound of fibers. This rebound of fibers caused soil particles to move, resulting in a lowering of the bond between fiber

PAGE 108

108 and matrices (Prasad et al.2012). Increasing the percentage of fibers beyond 0.6% therefore increased the occur rence of this rebound effect of the fibers. Figure 4 3. Average c ompressive strength versus fiber content of block s (n=30) According to Morel et al. (2007), typical compressive strength of manually pressed CEBs range from 2 .0 MPa to 3 .0 MPa. Higher values can be expected for blocks made with hydraulic presses as was the case in this research, which exceeded 3 MPa in compression. 2.07 MPa is the minimum required strength of earthen masonry materials in most US State s and local building codes like the State of New Mexico Earthen Building Materials Code (NMCB, 2009). It is worth noting that as the percentage of fiber content in the blocks increased, the coefficient of variation also increased. However, the highest coefficient of variation (16.02%) was still lo wer than the 20% deemed acceptable for CMU (Walker, 1995). 0.0 1.0 2.0 3.0 4.0 5.0 6.0 0.0 0.2 0.4 0.6 0.8 1.0 Average compressive strength (MPa) Fiber mass content of blocks (%)

PAGE 109

109 Using statistical regression analysis, a model is proposed to predict compressive strength of blocks ( c ) as a function of fiber weight fraction ( W f ) . All coefficients of variables included in th e model were significant at the 10% significance level as demonstrated by the p values in Table 4 6 . Within the limits of this experimental work, the relationship between compressive strength and fibe r content is expressed in E quation 4 1: c = 4.15 + 5. 25 W f 10.72 W f 2 + 5.04 W f 3 (4 1) Where; c is the compressive strength (MPa), W f is the fiber mass content (%), W f 2 is the fiber mass content squared (%), and W f 3 is the fiber mass content cubed (%). Table 4 6 . Model summary Mode l Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta (Constant) 4.150 .169 24.484 .000 Fiber Content ( W f ) 5.249 1.646 3.277 3.189 .004 Fiber Content Squared ( W f 2 ) 10.719 4.090 6.972 2.621 .014 Fiber Content Cubed ( W f 3 ) 5.039 2.686 3.306 1.876 .072 R 0.753 R Square 0.567 Adjusted R Square 0.517 Std. Error of the Estimate 0.38671

PAGE 110

110 Building codes specify the minimum compressive strength requirements for earthen masonry units as the average of five samples. For exa mple in the 2010 state of Florida building code, the section on the compressive strength of adobe is stated as tested in accordance with ASTM C 67. Five samples shall be tes ted and no individual 2010). For this reason, Eq uation 4 1 was used to predict the average compressive strength of five CEBs. The predicted compressive strength values showed a close fit to measured values of compressive strength with prediction errors below 6.0% as shown in Figure 4 4 and Table 4 7 . Figure 4 4. Influence of fiber mass content on measured and predicted average compressive strength . 0.0 1.0 2.0 3.0 4.0 5.0 6.0 0.0 0.2 0.4 0.6 0.8 1.0 Average compressive strength of 5 CEBs (MPa) Fiber mass fraction W f (%) Measured Predicted

PAGE 111

111 Table 4 7 . Comparis on of predicted and measured average compressive strength values (MPa) Mix Fiber content by weight (%) Predicted Measured Prediction error* (%) PP 0.0 0 4.150 4.197 1.112 PP 0.2 0.2 4.811 4.623 4.068 PP 0.4 0.4 4.857 5.147 5.625 PP 0.6 0.6 4.529 4.325 4.719 PP 0.8 0.8 4.069 4.129 1.444 PP 1.0 1.0 3.719 3.714 0.128 * 3 Point Bending Strength The 3 point bending streng th is the most widely used indirect way of measuring compressive strength (Morel et al. 2007). The results are presented in Table 4 8 . The highest average 3 point bending strength was for blocks reinforced with 0.4% fibers (PP 0.4). The ranking for second to last average 3 point bending strengths was mix types PP 0.2 (0.2% fiber), PP 0.6 (0.6% fiber), PP 0.0 (no fiber), PP 0.8 (0.8% fiber), and PP 1.0 (1.0% fiber) respectively. The average 3 point bending strength of mix type PP 0.2, PP 0.4, and PP 0.6 were 6%, 22 %, and 1.5% higher than mix type PP 0.0 (no fiber) respectively . The strength of mix types PP 0.8 and PP 1.0 were 26 and 31% lower than mix type PP 0.0 (no fiber) respectively. Influence of Fibers on 3 Point Bending Strength The blocks were subjec ted to a single (center point) loading point with simple supports. 3 point bending strength was calculated assuming pure bending and ignoring other factors such as shear and the arching action typically observed during failure (Morel et al. 2002; Morel et al. 2007). Center point loading may therefore result in an overestimation of strength since the values are evaluated using peak load (Walker,

PAGE 112

112 1995). When considered as a function of fiber content, the recorded 3 point bending strength values exhibited simi lar trends to that observed for compressive strength (Figure 4 5 ). Table 4 8. 3 point bending strength results of field produced blocks * Mass percentage *Coefficient of variation Most specimens failed with cracks initially emerging from the supports and propagating to the upper third of the span. The observed failure mode was consistent with findings reported by Morel et al. 2002 , Morel et al. 2007, and Villamizar et al. 2012. This failure mode is attributed to an arching action that occurs within blocks due to the small span to depth ratios of CEBs (Morel et al. 2002). In this research, the span to depth ratio was approximately 2 .0 . In fiber reinforced composites, fibers oppose crack formation as stress increases and prevent micro cracks from expanding (UN Habitat, 1992; Houben and Guillaud, the crack zone bore the tensile stresses transferred from the fractured sections resulting in a subsequent more gradual failure after initial crack of all the fiber reinforced blocks. This suggested an improved performance in ductility that can be attributed to the fibers Mix Fiber Content (%)* 1 2 3 4 5 Average Std. Dev. COV (%)** PP 0.0 0 .0 0.76 0.72 0.79 0.62 0.52 0.68 0.11 16.52 PP 0.2 0.2 0.69 0.87 0.65 0.56 0.80 0.72 0.12 17.24 PP 0.4 0.4 0.68 0.78 0.69 0.94 1.02 0.83 0.15 18.37 PP 0.6 0.6 0.88 0.91 0.48 0.49 0.68 0.69 0.20 29.26 PP 0.8 0.8 0.47 0.36 0.65 0.41 0.61 0.50 0.12 24.79 PP 1.0 1.0 0.52 0.39 0.58 0.37 0.47 0.47 0.08 18.4 7

PAGE 113

113 (UN Habitat, 1992; Zhang et al. 2010). Improvements in ductility can provide sufficient escape time during failure, which can be the factor de termining if people get out alive or remain trapped inside a collapsing structure (Segetin et al. 2007). Figure 4 5. Average 3 point bending results of CE s versus fiber content of blocks (n=30) Extensible polypropylene fibers do not totally pull out of ma trices when composites reach peak strength. Gradual fiber slipping and stretching results in a high post peak strength even at high deformation levels as was the case in this research (Consoli et al. 2009). The 3 point bending strength of the blocks was be tween 0.37 and 1.02 MPa , which falls within the range of 0.13 to 1.67 MPa reported by Walker, 1995, Galán Marín et al. 2010, and Villamizar, et al. 2012. The average 3 point bending strength for each mix types exceeded the minimum (0.35 MPa) stipulated in the New Mexico Earthen Building Materials Code (NMCB, 2009). The coefficients of variation was between 16.52 and 29.26%. Other studies have reported coefficients of variation between 9.2 and 48.5% (Walker, 1995) and between 5.2% and 36.5% (Walker, 2004). 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.2 0.4 0.6 0.8 1 1.2 Average 3 Point Bending (Mpa) Fiber mass content of blocks (%)

PAGE 114

114 B ased on the reported findings, the optimal fiber mass fraction based on the strength values and the coefficient of variation for 3 point bending strength was 0.4%. Influence of Fibers on CEB Deformation The vertical displacement recorded during compressive strength testing was used to generate strain values (Villamizar et al. 2012). Blocks reinforced with 1.0% fiber content had the highest deformations at the peak load of between 18.32 and 21.48% with an average of 20.5%. Blocks reinforced with 0.4% fiber c ontent had the lowest deformation at a peak load of between 4.58 and 6.58% with an average of 8.61% (Table 4 9 ). At 0.2, 0.4, and 0.6% of fiber content, there was an observed reduction in deformation at peak load compared to that in the unreinforced block s. As fiber content increased from 0.8 to 1.0%, an increasing trend in deformation was observed. F ailure of the unreinforced blocks was sudden and was generally initiated by one or two large cracks, while that of the fiber reinforced blocks was more ductil e and preceded by multiple cracks during compressive strength testing ( Figure 4 6 ). Whereas the unreinforced blocks underwent catastrophic failure resulting in a total separation of blocks into two halves during 3 point bending strength testing, none of the fiber reinforced blocks experienced catastrophic failure. The fibers through bridging cracks prevented catastrophic failure (Figure 4 7 ). The fibers prevented crack face separation mainly through a stretching process that provided an extra energy ab sorbing capacity while also reducing the stress around the micro cracked region surrounding the crack tip (Song, et al. 2005). The enhancement in ductility and flexibility as observed in this research improves the capacity of blocks to store elastic energy thereby providing some resistance to out of plane loads at the system level (UN Habitat, 1992; Binici et al. 2005; Islam and Iwashita, 2010 ).

PAGE 115

115 Table 4 9 . Results of compressive strength and strain Mix Type Compressive Strength (MPa) Strain % at Peak Load PP 0.0 CEB 1 4.15 12.17 CEB 2 4.18 11.11 CEB 3 4.28 10.53 CEB 4 4.33 9.79 CEB 5 4.03 12.53 PP 0.2 CEB 1 4.90 7.39 CEB 2 4.49 8.73 CEB 3 4.44 10.23 CEB 4 4.38 9.58 CEB 5 4.89 7.12 PP 0.4 CEB 1 5.03 5.78 CEB 2 5.40 4.99 CEB 3 4.93 5.32 CEB 4 4.81 4.58 CEB 5 5.55 6.38 PP 0.6 CEB 1 4.30 8.42 CEB 2 3.39 9.22 CEB 3 4.08 9.23 CEB 4 5.00 7.01 CEB 5 4.31 8.56 PP 0.8 CEB 1 4.04 19.90 CEB 2 4.13 18.46 CEB 3 3.90 17.97 CE B 4 4.55 19.55 CEB 5 4.01 18.76 PP 1.0 CEB 1 3.31 21.48 CEB 2 3.14 18.32 CEB 3 3.76 22.60 CEB 4 4.67 19.68 CEB 5 3.68 20.42

PAGE 116

116 Figure 4 6 . T ypical failure mode of blocks. A ) Unreinfor ced blocks showing two cracks. B ) Blocks rei nforced with 0.8% fibers showing multiple cracks (Photo courtesy of author) . Figure 4 7. Fibers bridging a crack during 3 point bending test (Photo courtesy of author).

PAGE 117

117 The level of cementation, compaction, fiber type, and surface conditions have been identified as key factors influencing the matrix strength of fiber reinforced soil composites (Consoli et al. 1998; Tang et al. 2007; Consoli et al. 2009; Kumar et al. 2010). The interfacial and frictional bonds activated when fibers are pulled out of matr ices improves the bonding between fibers and matrices (Sukontasukkul and Jamsawang, 2012). It was discovered in this research that, a major contributor to fiber matrix bond was the surface conditions of the fibers f ibrillation or in this case cross linking and embossment of the fiber surface ( embossed depths from peak to valley of about 0.005 to 0.006 mm). The embossment provided mechanical anchorage thereby resulting in the development of frictional bonds between the fibers and matrix. The frictional bond was further enhanced by matrix compaction. Relationship between Compressive Strength and 3 Point Bending Strength The average 3 point bending strength of the unreinforced blocks (PP 0.0) was 16.23% of the average compressive strength of the unreinforced b locks. The average 3 point bending strength of mix types PP 0.2, PP 0.4, PP 0.6, PP 0.8, and PP 1.0 was 15.58%, 16.12%, 15.97%, 12.12%, and 12.67% of their corresponding compressive strengths respectively. Generally, 3 point bending strength of the CEBs w ere in the range of 12.67% to 16.23% of compressive strength. The range reported here, falls within the 10 20% reported by other researchers (Walker, 1995; Walker, 2004; Namango 2006). Figure 4 8 shows a bar graph of both compressive strength and 3 point bending strength of the CEBs.

PAGE 118

118 Figure 4 8. Compressive strength and 3 point bending strength of blocks at different fiber mass fractions Load Deflection Response of Short Flexural Beams The load deflection curves obtained during testing of the beams is presented in Figure 4 9 . With the unreinforced matrices, load increased linearly with increasing deflection until first peak load was reached after which a sharp decline in load was observed. First peak load which in this case was the same as peak load w as reached at the onset of crack. Beams fai led catastrophically (Figure 4 10 ) when peak load was reached indicating the brittleness of the u nreinforced matrices. In the case of the fiber reinforced matrices, the load deflection response was similar to the plain matrices until first peak load was reached. 4.19 4.62 5.15 4.32 4.13 3.71 0.68 0.72 0.83 0.69 0.5 0.47 0 1 2 3 4 5 6 PP-0.0 PP-0.2 PP-0.4 PP-0.6 PP-0.8 PP-1.0 Average Compressive/3 Point Bending Strength (MPa) Fiber mass content (%) Compressive Strength (Mpa) Modulus of Rupture (MPa)

PAGE 119

119 0 500 1000 1500 2000 2500 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Load (N) Deflection (mm) Specimen 1 Specimen 2 Specimen 3 Specimen 4 Specimen 5 A 0 500 1000 1500 2000 2500 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Load (N) Deflection (mm) Specimen 1 Specimen 2 Specimen 3 Specimen 4 Specimen 5 B 0 500 1000 1500 2000 2500 3000 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Load (N) Deflection (mm) Specimen 1 Specimen 2 Specimen 3 Specimen 4 Specimen 5 C 0 500 1000 1500 2000 2500 3000 3500 4000 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Load (N) Deflection (mm) Specimen 1 Specimen 2 Specimen 3 Specimen 4 Specimen 5 D

PAGE 120

120 Figure 4 9. ) Unreinforced matrices. B) 0.2% fiber content. C) 0.4% fiber content. D ) 0.6% fiber content . E) 0.8% fiber content. F ) 1.0% fiber content . Figure 4 10. Failed beams. A ) Unreinforce d beam (catastrophic failure). B ) Fiber reinforced beam held together by fibers (Photo courtesy of author) . 0 500 1000 1500 2000 2500 3000 3500 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Load (N) Deflection (mm) Specimen 1 Specimen 2 Specimen 3 Specimen 4 Specimen 5 E 0 500 1000 1500 2000 2500 3000 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Load (N) Deflection (mm) Specimen 1 Specimen 2 Specimen 3 Specimen 4 Specimen 5 F

PAGE 121

121 At 0.2% and 0.4% fiber content by weight, a sharp drop in load carrying capacity was observed after first peak load, which in this case occurred at first crack. This was followed by a gain in load bearing capacity, which gradually dropped and flattened out. The observation made for the matrices with 0.4% fiber content was similar to that for 0.2% except that the drop in load carrying capacity after first peak load was not as sharp. The regain in load carrying capacity of the beams after first peak load did not go higher than what was recorded at first crack point. As the fiber dosage increased, the drop in loa d bearing capacity after the first crack point decreased (Figure 4 9 D , E , F ). Matrices with 0.6, 0.8, and 1.0% fibers began exhibiting a hardening behavior after initial load drop. Load carrying capacity typically increased above what was recorded at firs t crack point. At 0.8 and 1.0% fiber content, the matrices exhibited more of a ductile behavior (Figure 4 9 E , F ). Flexural performance of Short Flexural Beams Flexural performance was determined by the first peak strength (f 1 ), peak strength (f P ), equivalent flexural strength ( f e ), residual strength at deflections of L /600 ( ) and L /150 ( ), and flexural toughness ( ) at a deflection L /150 (Table 4 10 ). The first peak (f 1 ) strength depicts the flexural behavior of the beams (both reinforced and unreinforced) up to the onset of cracks. Peak strength (f p ) was calculated using the maximum load obtained from the load deflection curves (Fig ure 4 9 ). For the unreinforced matrices and matr ices with 0.2%, 0.4% fibers, first peak strength was the same as peak strength. On the other hand, matrices that exhibited deflection hardening (0.6%, 0.8%, and 1.0% fibers) had their peak strengths greater than their corresponding first peak strengths by about 8%, 28%, and 11% respectively. The results show that the average first peak strength of the fiber reinforced matrices were higher than the

PAGE 122

122 unreinforced matrix. .P eak strength increased up to a fiber dosage of 0.6%. After this, average first peak stre ngth and peak strength steadily decreased as fiber dosage was increased. This shows that the increases in strength observed is more of a function of fiber dosage optimization and not directly related to fiber dosage increase ( Wilson and Abolmaali, 2014) . Table 4 10 . Flexural properties (average of five specimens for each mix design) Mix Strength (MPa) Peak Strength (MPa) Deflection L/600 L/150 (kN) (MPa) (kN) (MPa) Tb (Nmm) fe (MPa) FB 0.0 0.47 0.47 FB 0.2 0.54 0.54 1.08 0.32 0.82 0.24 2089 0.30 FB 0.4 0.58 0.58 1.54 0.45 1.09 0.32 2901 0.42 FB 0.6 0.78 0.84 2.86 0.83 2.11 0.61 5210 0.75 FB 0.8 0.58 0.74 2.37 0.69 1.67 0.50 4312 0.62 FB 1.0 0.58 0.63 2.23 0.62 1.65 0.48 399 2 0.57 Equivalent flexural strength ( ) shows the effectiveness of fibers to bridge cracks and enhance the energy absorption ability of the reinforced matrices under loading. The accounted for the flexural strength up to a deflection of L/150 (2.03 mm). The interfacial and fricti onal bonds activated when fibers are pulled out of matrices (slipping and stretching) improves the bonding between fibers and matrices resulting in atrices (Consoli et al. 2009; Sukontasukkul and Jamsawang, 2012) . The best performing mix in terms of was mix FB 0.6 which had an average 21% higher tha n the next best performing mix FB 0.8. Out of the fiber reinforced

PAGE 123

123 matrices, mix FB 0.2 had the worst performance in terms of average , which was 60% lower than mix FB 0.6. Flexural toughness ( ) is a measure of the energy absorption capacity of a material cture reinforced composites since it shows the ability to absorb energy after cracking (ACI 544.1R, 1996). As fiber dosage was increased from 0.2% to 0.4% and 0.6%, flexural to ughness increased by 39% and 80% respectively. On the other hand when fiber dosage was increased from 0.6% to 0.8% and 1.0%, flexural toughness dropped by flexural toughnes s because they could absorb more energy (Lin et al. 2014). Residual strengths at L/600 (0.50 mm) and at L /150 (2.03 mm) in Table 4 9 the spec beams were able to maintain some load carrying capacity after first crack. At fiber co ntent of 0.2% and 0.4% (mixes FB 0.2 and FB 0.4), average residual strengths at a deflection o f L/600 ( the other hand at fiber content o f 0.6%, 0.8%, and 1.0% (mixes FB 0.6 and FB 0.8, and FB 1.0), the average residual strengths at a deflection of L/600 ( ) were hig her than their corresponding first peak strengths. Mix FB 0.6 had the highest average residual strength ( The average residual strength of the fiber reinforced beams at a de flection of L/150 ( ) were lower than their corresponding first peak strengths. Mix FB 0.6 had the

PAGE 124

124 highest average residual strength , which was about 22% lower than its corresponding first peak strength. Relationship between Fiber Content and Flexural Performance The development of strength properties of soil cement fiber mixes mostly matrix, matrix identified bonds are influenced by the dimension, surface conditio ns, and quantity of fibers present. Additionally, the number of contact points between fibers and matrices are responsible for transmitting stress a large quantity of fibers therefore reduces strength (UN Habitat, 1992; Khedari et al. 2005). Increasing f iber content in CEB CEBs leading to a reduction in strength (Namango, 2006). This explains the declining f 1 ), peak strength, equivalent flexural strength ( f e ), residual strength ( and ), and flexural toughness ( ) after exceeding 0.6% fiber content. Due to the stiffness of the fibers used (sticklike nature), a rebound of fibers was observed after de molding. Fibers near the surface of beams stuck out after the rebound of fibers. This rebound of fibers causes soil particles to move, resulting in a lowering of the bond between fiber and matrices (Prasad et al. 2012). Bond strength between fiber and matrix is res ponsible for composite strength. Increasing the percentage of fibers without adjusting the other matrix constituents upwards meant the quantity of fibers per block increased. This increase in fiber quantity meant there were more fibers rebounding per block . Other researchers have reported similar decreasing trends for strength after exceeding certain fiber thresholds. For example, Bouchicha et al. 2005 reported that at

PAGE 125

125 1.5% barley straw reinforcement , there was between an increase of between 10 20% in the compressive strength of CEBs depending on soil type used. However, increasing fiber content from 1.5% to 3.5% resulted in a 45% drop in compressive strength. Other reported thresholds for fiber content by volume and weight beyond which strength decreases i nclude 2% of salvaged steel fibers (Eko et al. 2012) and 0.75% of sisal fiber (Namanago, 2006). The lack of adequate test data to facilitate the accurate prediction of the structural performance of earthen construction materials has not helped with the structural advancement of CEB masonry (Adam and Agib 2001; Rigassi, 1995; Morton, 2008). Using statistical regression analysis on the test results, a model to predict equivalent flexural strength ( f e ) based on th fraction is proposed in this study (Eq uation 4 2 ). Table 4 1 1 presents a summary of the model details. All coefficients of the model were significant at the 10% significance level as demonstrated by the p values in Tab le 4 1 1 . Figure 4 11 shows the relationship between the equivalent flexural strength predicted using Eq uation 4 2 and those calculated from the experimental results. = 0.27 + 0.59 f 1 + 1.07 W f 0.573 W f 3 . (4 2 ) Where; f 1 = firs t peak strength (MPa), W f = fiber mass content (%), and W f 3 = fiber mass content cubed (%).

PAGE 126

126 Table 4 1 1 . Model summary Figure 4 11 . Predicted vs. experimentally recorded equivalent flexural strength Model Unstandardized Coefficients Standardized Coefficients t Sig.(p value) B Std. Error Beta (Constant) 0.268 0.085 3.169 0. 004 First Peak Strength ( f 1 ) 0.593 0.169 0.275 3.503 0.002 Fiber Content ( W f ) 1.066 0.139 1.422 7.692 0.000 Fiber Content Cubed ( W f 3 ) 0.573 0.125 0.803 4.598 0.000 R 0.949 R Square 0.900 Adjusted R Square 0.889 Std. Error of the Estimate 0.086 95

PAGE 127

12 7 Crack Patterns Crack patterns were observed during beam testing. The plain matrix beams underwent sudden failure at the onset of crack. With the beams reinforced with 0 .2 and 0.4 fiber content, cracks were typically straight and run from top to bottom in the middle third of beams. Failure was mostly localized at a single crack (Figure 4 12 A ). The fibers resisted the formation of large crack widths. As fiber content was i ncreased to 0.6%, 0.8%, and 1.0%, failure was typically characterized by multiple cracks (Figure 4 12 B ). This multiple cracking phenomenon observed is the result of the stress redistribution action of the fibers. After the formation of the first macro crack, the load carried by the matrix is transferred to the bridging fibers. The fibers acting as a bridge, transfers the load back to the matrix through the fiber matrix interfac e. This process leads to load build up, resulting in the development of another crack. As this process continues, multiple cracks develop. During the process of multiple cracking, the load carrying capacity of composite s can rise to exceed the first cracki ng strength of the composite, as was the case in this study (Arisory and Wu, 2008; Lin et al. 2014). increase in toughness, which is important for the serviceability and durability of structures under different loading conditions (Lin et al. 2014). The toughness recorded for mixes that exhibited multiple cracking ( FB 0.6, FB 0.8, and FB 1.0) was higher than mixes ( FB 0.0, FB 0.2, and FB 0.4) that failed with a single crack. According to Arisory and Wu (2008), a critical minimum fiber quantity is needed for multiple cracking t o occur. In this study, multiple cracking began occurring at 0.6% fiber content. The failure

PAGE 128

128 Figure 4 12. Typical crack patterns. A ) Beams reinforced with 0.2 and 0. 4% fibers. B) Beams reinforced with 0.6, 0.8, and 1.0% fibers (Photo court esy of author). Scanning Electron Microscopy (SEM) Analysi s T he micrographs of the fractured surfaces were to help understand the interfacial interactions between the fiber surface and soil matri ces. Some f iber reinforced beams were forced to complete fail ure to allow for the separation of crack surfaces. Figure 4 13 shows a close up photograph of a failed beam held together by the polypropylene fibers. Figure 4 13. Close up of fibers bridging crack (Photo courtesy of author) .

PAGE 129

129 After failed beams were for ced to complete failure, both fiber pullout and fra cture were observed (Figure 4 14 ). This observation was an indication that the fibers used had a high pullout capacity and resistance; a quality that was enhanced by matrix compression. During the transfer of stresses from matrix to fiber, the de bonding that takes place at the fiber matrix interface when fibers are pulled out from matrices generates frictional energy losses, which in turn contribute to composite toughness (Tonoli et al, 2011). Micrographs of unused fibers were captured for comparison with used fibers after composite failure. The abrasion observed on the surface of the fiber gives an indication of fiber matri x bond strength ( Figure 4 15 ). Figure 4 14. SEM microgra ph of fractured beam surfaces. A) Embedded fiber in matrix. B) F iber fracture and pullout . As stated earlier, it was observed in this research that, a major contributor to fiber matrix bond was the cross linking and surface deformations of the fibers (depths from peak to vall ey of about 0.005 to 0.006 mm) which resulted in frictional bond developing between the fib ers and matrix. From Figure 4 15 , it can be observed that the

PAGE 130

130 abrasions on the fibers were more pronounced at the location of fiber surface deformations. Figure 4 1 5 . SEM Micro graph of polypropylene fibers. A ) Unused fiber showing surface deformations meant to provide mechanical anchorage. B ) De bonded fiber showing surface abrasion after matrix failure. Mortar Compressive Strength Mortar compressive strength is pre sented in Table 4 1 2 . The mortar mix that contained the same percentage of cement as the CEBs (EM # 1) had an average compressive strength of 7.21 MPa. When the cement content was increased by a factor of 1.5 and 2 .0 , compressive strength increased to 12.2 4 MPa and 15.10 MPa respectively representing an increase of 56% and 109% respectively ested specimens ( Figure 4 16 ). Guillaud and Joffroy, ( 1995 ) suggested that increasing the proportion of c ement in earthen mortar by a factor of 1.5 or 2 relative to the cement content in CEBs results in mortar of similar strength to the CEB s . The results presented here shows that increasing the cement content in mortar by factors of 1.5 and 2.0 yielded mortar much stronger than the CEBs.

PAGE 131

131 While mortar batching was done in mass fractions in this research, it can be inferred that the increase in cement content proposed by Guillaud and Joffroy (1995) was in volume fractions. Table 4 1 2 . Mortar compressive strengt h (MPa) Mix Type 1 2 3 4 5 Average Std. Dev. Coefficient of Variation (%) EM # 1 6.94 7.23 6.70 7.75 7.14 7.21 0.32 4.44 EM # 2 12.28 11.04 11.62 11.10 10.07 11.24 0.84 6.86 EM # 3 14.76 15.81 15.73 16.03 13.18 15.10 1.18 8.00 Figure 4 16 . Typical hour glass type failure for tested specimens

PAGE 132

132 Prism Compressive Strength Masonry compressive strength depends upon characteristics of the masonry units used, mortar, and the nature of bond formed between the masonry units and mortar (Uday Vyas and Venkatarama Reddy, 2010). During testing, the prisms were compressed perpendicular to the bed joints. Failure planes were observed to run through both blocks and mortar from top to bottom. Cracks initially formed in the top block and travelled down throug h the mortar bed to the lower block. Prism compressive st rength is presented in Table 4 1 3 . Some blocks showed signs of crashing during testing. The polypropylene fibers held all failure planes together thus preventing catastrophic failure. Block mortar jo ints did not fail; there was no separation between mortar bed and face shell of blocks, giving an indication of good bonding ( Fig ure 4 17 ). The average prism compressive strength was 4.89 MPa with a coefficient of variation of 5.52%. Table 4 1 3 . Prism compressive strength Prism No. Age at Test (days) Avg. Width (mm) Avg. Height (mm.) Avg. Length (mm) Net Area (mm 2 ) Max Load (N) Compressive Strength (MPa) 1 28 205 255 200 41000 213070 5.20 2 28 205 255 190 38950 181265 4.65 3 28 205 255 205 42025 2135 15 5.08 4 28 205 255 205 42025 211290 5.03 5 28 205 255 195 39975 181043 4.53 6 28 205 255 200 41000 182000 4.44 7 28 205 255 190 38950 195500 5.02 8 28 205 255 205 42025 201650 4.80 9 28 205 255 195 39975 200010 5.00 10 28 205 255 200 41000 212000 5.17

PAGE 133

133 Figure 4 17. Typic al failure mode of blocks. A ) Tested specimen still held tog ether by fibers after failure. B ) F ailure plane pulled apart to allow for observation; mortar block bond remains intact after failure (Photo courtesy of author). Prism Fle xural Bond Strength Flexural bond strength was determined using ASTM E518. This test method is used to determine the flexural bond strength of unreinforced masonry assemblages. It is intended to provide an inexpensive and simplified means for gathering com parative research data on the flexural bond strength developed with different masonry unit/mortar types. It is also used as a quality control mechanism for checking the quality of materials and workmanship ( ASTM E518 12). The flexural bond strength values are presented in Table 4 1 4 . The flexural bond strength recorded was between 0.28 MPa and 0.50 MPa. According to Drysdale et al. 1999, flexural bond strength can range between 0.2 0 to 1.75 MPa. Other reported flexural bond strength values are 0.08 0.38 MPa (Sarangapani et al. 2005), 0.05 0.29 MPa (Venu Madhava Rao et al. 1996) and 0.13 0.34 (McGinley, 2004). It is worth noting

PAGE 134

134 that these studies used different block/mortar combinations and used the bond wrench test prescribed in ASTM C1072 or variations of it thereof. Table 4 1 4 . Flexural bond strength McGinley (2004) reported flexural bond strength using the third point loading method in ASTM E 518 to be in the range of 0.30 and 0.70 MPa, which was between 2.07 to 2.25 times higher than the recorded values using the wrench test. Reasons for the differences in results were attributed to either variations in specimen production, size, and curing conditions or the fact that the ASTM E 518 procedure only tests the central block/mortar joints that may be inherently stronger as a result of the consolidation action by the blocks on top of the a ssemblage during fabrication. Average Prism No. of courses Span Length (mm) Width (mm) Depth (mm) Weight (N) Max. Load (N) Modulus of Rupture (MPa) 1 7 669 203 121 307 1646 0.42 2 7 669 203 121 310 1779 0.45 3 7 669 203 121 305 1557 0.40 4 7 669 203 121 311 2000 0.50 5 7 669 203 121 307 1350 0.36 6 7 669 203 121 300 1250 0.33 7 7 669 203 121 310 1190 0.32 8 7 669 203 121 305 1380 0.36 9 7 669 203 121 308 1000 0.28 10 7 669 203 121 300 1500 0.39

PAGE 135

135 prism weight in this study was 306 N. Prism weight can result in the failure of bonds during specimen moving and set up when ASTM E 518 is used. Failure Mode Joint failure during testing was sudden in almost all cases . Two distinct modes of failure were observed during testing f ailure within the block un it and mortar joint (Figure 4 18 ) and failure within the block unit. Failure within the block unit and mortar joint was observed for about 40% of the tested prisms wi th failure within the block unit being the most common. Failure within the block unit occurred in two main forms (Figure 4 19 ); shallow fracture depths and deeper fractur e depths. This type of failure of plane flexural failure of the block f weaker to the strength of blocks (Walker, 1999). However, in this case, mortar strength was higher than block strength. The failure modes observed; failure of block and a combination of block and bond fa ilure was an indication of good bonding (Sarangapani et al. 2005). Figure 4 18. Failure between block unit and mortar joint.

PAGE 136

136 Figure 4 19. Opposite sides of the first form of t he second failure mode observed (shallow i ndentations/fractures ) . For the t ype of failure where deeper fractures occurred within the block, failure was not sudden. When such fractures occurred, the fibers within the blocks held both faces together thus preventing sudden fai lure in the process (Figure 4 20 ). Total separation occur red only after the fibers had either pulled out or fractured. Figure 4 20. Opposite sides of the failure mode showing deep block fractures and block fibers sticking out. .

PAGE 137

137 CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS Overview This dissertation investigat ed the effects of polypropylene fibers on the strength, deformability, and general flexural performance of soil cement matrices used for CEB production. The inclusion of the fibers was to improve ductility, damage resilience, and deformability as well as provide additional insight s into the behavior of fiber reinforced CEB matrices . Plain (unreinforced) and fiber reinforced CEBs and soil cement beams with fiber dosages of 0.2, 0.4, 0.6, 0.8, and 1.0% were produced and tested. The CEBs were tested for their compressive strength and 3 point bending strength while the short flexural beams were tested for their flexural performance. Fractured beam surfaces obtained during flexural strength testing were observed using scanning electron microscopy to help determi ne composite failure mechanism of the fiber reinforced matrices. Two and seven block prisms fabricated with polypropylene fiber reinforced CEBs and earthen mortar were tested for their compressive strength and flexural bond strength to determine the compat ibility of the mortar with the fiber reinforced CEBs. The CEBs used in for prism fabrication were produced with 8% OPC and 0.4% polypropylene fibers. The average compressive strength of the blocks was 5.15 MPa. The mortar for prism fabrication was produced with the same soil as the blocks and with twice the cement content. The mix design of the mortar was 1:6 (cement:soil) with an average compressive strength of 15.10 MPa. Conclusions Within the limits of the experimental program used, the following conclus ions were reached:

PAGE 138

138 Flexural Performance, Compressive/3 Point Bending and Prism Bond Strength The observed enhancement in ductility and flexibility of the fiber reinforced blocks as observed in this research improves the capacity of blocks to store elastic energy . Fiber inclusion resulted in a high degree of load retention after the first crack compared to the unreinforced matrices that were brittle and exp erienced catastrophic failure. The fibers prevented catastrophic failure of the reinforced beams even a t large without collapse. The inclusion of the polypropylene flexural performance as was demonstrated by the load deflect ion response of beams after initial crack, residual strength, flexural toughness, a nd equivalent flexural strength. an improvement in ductility as a result of polyprop ylene fiber reinforcement. There was no separation between mortar bed and face shell of blocks during prism compressive strength testing, giving an indication of good bonding. In addition ,the observed failure modes during prism flexural bond testing fail ure of block and a combination of block and bond failure was also an indication of good bonding . The earthen (cement:soil) mortar used was deemed compatible with the polypropylene fibers based on the prism compressive and flexural bond strength. SEM Analys is A major advantage of adding fibers to matrices is that after matrix cracking, fibers bridge and restrain cracks. The additional forces required to pull out or fracture the fibers either protects the integrity of matrices or improves their post crack l oad carrying capacity. The process of fiber pullout during the transfer of stress from matrix to fiber

PAGE 139

139 generates frictional energy losses contributing to composite toughness. Micrographs of fractured surfaces obtained during flexural strength testing showe d bo th fiber pullout and fracture. This was an indication of the pull out capacity of the polypropylene fibers used. The fibers used had deformations to provide mechanical anchorage between the fiber and matrices. De bonded fibers showed surface abrasion a fter matrix failure. The abrasions were more pronounced at fiber surface sections that had the deformations. Recommendations Suggested Mix Proportions/Procedure It is recommended not to exceed 0.6% polypropylene fibers by weight for CEB production. An ide al range of polypropylene fibers to add to CEB matrices is between 0.4 and 0.6%. For hand mixed CEBs , jt is difficult to achieve a uniform mix when fiber quantity exceeds 0.4%. For mechanically mixed CEB matrices. fiber balling begins to occur after exceed ing 0.6% fiber content. A declining trend in strength was observed when these fiber thresholds were exceeded. The addition of polypropylene fibers into CEB matrices should be done in measured quantities at a time while mixing continues. Adding about 0.045 kg of fibers every one minute to a dry mix of OPC and soil was determined to yield homogenous matrices with well distributed fibers. Mixing should continue for about 3 more minutes after the last batch of fibers have been added. For hand mixing, adding ea ch batch of 0.045 kg of fibers should be done after the previous batch has been well mixed into the matrix. Wet curing (sprinkling with water) during the first seven days after production should be done under tarps or plastic sheets. This procedure is imp ortant to ensure there is enough moisture within blocks to facilitate OPC hydration. It is essential to

PAGE 140

140 ensure that blocks are well dried before testing. Plastic sheets/tarps can be taken off about seven days before testing. Specimen span to depth ratio o f 3 .0 is critical to accurately evaluating flexural performance of fiber reinforced CEB matrices. Where actual fiber reinforced CEBs do not meet this span to depth ratios, it is recommended that specimens meeting a span to depth ratio of a least 3.0 be pro duced for flexural performance evaluation. Recommendations for Fut ure Research The findings of this research help to understand the behavior of fiber reinforced soil cement matrices used in CEB production. The following research areas will contribute to a better understanding of the material behavior of CEBs and CEB masonry and provide design professionals with valuable information on CEB performance : Research involving further testing of mechanical properties (compressive and 3 point bending strength) of polypropylene fiber reinforced CEBs at different specimen aspect ratios will be beneficial . Previous researchers have established that CEB strength is influenced by specimen geometry/aspect ratio therefore , making it important to determine how polypropyle ne fiber reinforced CEBs are influenced by specimen geometry/aspect ratio. Additional tests at the system level are required to quantify how the benefits of polypropylene fibers at the block level translate into performance enhancement at the system level . The durability of fiber reinforced CEBs should be studied further . Presently, durability of CEBs is mostly associated with compressive strength. Durability measurements on some of the samples used in the present research can be conducted and correlatio ns between those measurements and compressive strength and 3 point bending strength established.

PAGE 141

141 Similar studies to the one done in this dissertation using different soil types will broaden the understanding of fiber soil interactions and improve the pred ictability of results. Further research on the use of alternative stabilizing agents and cementitious binders and the optimal moisture content for different soil types will also help provide more insights into CEB performance.

PAGE 142

142 APPENDIX A FLEXURAL PERFORMA NCE RESULTS Table A 1 . Flexural performance of beams Fiber Content Specimens Average s.d.* COV** 1 2 3 4 5 First Peak Strength No Fiber 0.43 0.48 0.49 0.46 0.51 0.47 0.03 6.80 0.20% 0.49 0.41 0.66 0.51 0.60 0.54 0.10 18.19 0.40% 0.55 0.46 0.61 0.5 7 0.73 0.58 0.10 17.03 0.60% 0.80 0.76 0.71 0.89 0.76 0.78 0.07 8.54 0.80% 0.52 0.52 0.55 0.75 0.53 0.58 0.10 16.80 1.00% 0.60 0.54 0.54 0.68 0.52 0.58 0.07 11.52 Equivalent Flexural Strength (MPa) 0.20% 0.29 0.31 0.34 0.26 0.30 0.30 0.03 9.78 0.40% 0.40 0.40 0.49 0.48 0.31 0.42 0.07 17.16 0.60% 0.78 0.66 0.57 0.97 0.76 0.75 0.15 20.42 0.80% 0.56 0.78 0.55 0.69 0.51 0.62 0.11 17.91 1.00% 0.55 0.51 0.58 0.67 0.54 0.57 0.06 10.78 Peak Strength (MPa) No Fiber 0.43 0.48 0.49 0.46 0.51 0.47 0.03 6.80 0.20% 0.49 0.41 0.66 0.51 0.60 0.54 0.10 18.19 0.40% 0.55 0.46 0.61 0.57 0.73 0.58 0.10 17.03 0.60% 0.87 0.76 0.75 1.04 0.81 0.84 0.12 14.12 0.80% 0.67 0.83 0.62 0.79 0.79 0.74 0.09 12.14 1.00% 0.63 0.60 0.62 0.73 0.55 0.63 0.07 10.57 * Standa rd Deviation ** Coefficient of Variation

PAGE 143

143 Table A 1. Continued Fiber Content Specimens Average s.d.* COV** 1 2 3 4 5 Flexural Toughness (Nmm) 0.20% 2001 2156 2369 1813 2105 2088.80 204.37 9.78 0.40% 2765 2794 3425 3330 2192 2901.20 497.81 1 7.16 0.60% 5415 4588 3964 6804 5281 5210.40 1063.87 20.42 0.80% 3899 5423 3842 4804 3593 4312.20 772.14 17.91 1.00% 3843 3593 4082 4694 3750 3992.40 430.37 10.78 Residual Strength at L/150 0.20% 0.24 0.26 0.22 0.25 0.22 0.24 0.02 7.02 0.40% 0.38 0.29 0.32 0.34 0.27 0.32 0.04 13.06 0.60% 0.62 0.54 0.41 0.83 0.65 0.61 0.16 25.58 0.80% 0.47 0.62 0.45 0.56 0.40 0.50 0.09 17.76 1.00% 0.45 0.40 0.50 0.56 0.50 0.48 0.06 12.52 Residual Strength at L/600 0.20% 0.30 0.34 0.38 0.26 0.31 0.32 0.05 14.83 0.4 0% 0.38 0.43 0.54 0.54 0.37 0.45 0.08 18.76 0.60% 0.86 0.73 0.73 1.01 0.81 0.83 0.12 14.01 0.80% 0.62 0.83 0.62 0.78 0.60 0.69 0.11 15.38 1.00% 0.62 0.60 0.61 0.73 0.54 0.62 0.07 11.11 * Standard Deviation ** Coefficient of Variation

PAGE 144

144 APP ENDIX B REGRESSION MODEL DIAGNOSTICS MODEL FOR PREDICTING COMPRESSIVE STRENGTH c = 4.15 + 5.25 W f 10.72 W f 2 + 5.04 W f 3 [MODEL 1] Where; c = Compressive strength of fiber reinforced blocks W f = Fiber mass content W f 2 = Fiber mass content squared W f 3 = Fiber mass content cubed The model details are below: (1) Table B 1 . Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .753 a .567 .51 7 .38671 a. Predictors: (Constant), Fiber_Cubed, Fiber_Content, Fiber_Sq Table B 2 . ANOVA a Model Sum of Squares df Mean Square F Sig. 1 Regression 5.091 3 1.697 11.346 .000 b Residual 3.888 26 .150 Total 8.979 29 a. Dependent Variable: Comp_St rength b. Predictors: (Constant), Fiber_Cubed, Fiber_Content, Fiber_Sq

PAGE 145

145 Table B 3 . Coefficients a All c oefficients were significant at the 90% confidence level . (2) MODEL DIAGNOSTICS Unusual and Influential Data In identifying highly influential data points in this study, SPSS was set up to flag outliers above +/ 3 standard deviations. None of the observations were flagged. Therefore, no value exerted a significant amount of influence. Generally, there were no problems as far as extreme and highly influential/leveraged observations were concerned. The relevant st atistics are highlighted in yellow in Table B 4. Plots to show the distribution are presented in Figure B 1. Figure B 1 . Residual plots (a) Histogram of standardized residuals fitted with normal curve (b) Normal probability plot of residuals Model Unstandardized Coefficients Standardized Coefficients t Sig. 90.0% Confidence Interval for B B Std. Error Beta Lower Bound Upper Bound 1 (Cons tant) 4.150 .169 24.484 .000 3.860 4.439 Fiber_Content 5.249 1.646 3.277 3.189 .004 2.442 8.057 Fiber_Sq 10.719 4.090 6.972 2.621 .014 17.695 3.743 Fiber_Cubed 5.039 2.686 3.306 1.876 .072 .459 9.620 a. Dependent Variable: Comp_Strength

PAGE 146

146 Table B 4 . Residuals Statistics a Minimum Maximum Mean Std. Deviation N Predicted Value 3.7192 4.8567 4.3557 .41897 30 Std. Predicted Value 1.519 1.196 .000 1.000 30 Standard Error of Predicted Value .117 .169 .139 .022 30 Adjusted Predicted Value 3.4907 4.8661 4.3553 .42527 30 Residual .59873 .96084 .00000 .36617 30 Std. Residual 1.548 2.485 .000 .947 30 Stud. Residual 1.637 2.764 .000 1.019 30 Deleted Residual .70446 1.18925 .00034 .42496 30 Stud. Deleted Residual 1.695 3.226 .017 1.081 30 Mahal. Di stance 1.703 4.603 2.900 1.258 30 Cook's Distance .000 .454 .041 .086 30 Centered Leverage Value .059 .159 .100 .043 30 a. Dependent Variable: Comp_Strength Normality of Error Terms In order to test the validity of normality of error terms assumption in the MODEL 1 , both KS and Shapiro Wilk tests were performed on the residuals to test the null hypothesis that the residuals from the model are normally distributed against the alternative hypothesis that they are not normally distributed. The test resul ts are summarized in Table B 5. Ho: Residuals from Regression are normally Distributed Ha: Residuals from Regression are not normally Distributed Table B 5 . Tests of Normality Kolmogorov Smirnov a Shapiro Wilk (SW) Statistic df Sig. Statistic df Sig. U nstandardized Residual .137 30 .155 .956 30 .239 a. Lilliefors Significance Correction

PAGE 147

147 The results of the SW test indicate a failure to reject the null hypothesis that the residuals are normally distributed at the 95% confidence level as is evidenced b y the high p value of 0.239 compared to the significance level of 0.05 . The overall conclusion is that the normality of error terms assumption is validated or met with respect to MODEL 1 . Figure B 2 further supports the normality of error terms assumption. Homoscedasticity Variance of the residuals (error terms) appear homoscedastic and the residuals appear independent based on Figure B 3. This conclusion based on visual inspection e. The residuals were separated out into two groups about the median value of Compressive Strength ( 4.305 ). There appears to be equal scatter in the points all along the horizontal axis from the scatter diagram. From the box plot, average residuals is clo se enough to zero for both groups. The plots look OK. Figure B 2 . Residual diagnostics (a) Histogram of residuals fitted with normal curve (b) normal probability plot of residuals

PAGE 148

148 Figure B 3 . Variance of residuals (a)Scatter plot of residuals vs pr edicted values (b) Residual box plot below & above median compressive strength (4.305). Residual terms by Group MODEL FOR PREDICTING EQUIVALENT FLEXURAL STRENGTH The best model for predicting equivalent fl exural strength was determined to be; f e = 4.15 + 5.25 W f 10.72 W f 2 + 5.04 W f 3 [MODEL 2] Where f e = Equivalent flexural strength of fiber reinforced blocks f 1 = First peak strength W f = Fiber mass content W f 3 = Fiber mass content cubed

PAGE 149

149 (3) Table B 6 . Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate Durbin Watson 1 .949 a .900 .889 .08695 2.291 a. Predictors: (Constant), Fiber Content Cubed , First Peak Strength, Fiber Content b. Dependent Variable: Equivalent Flexural Strength Table B 7 . ANOVA a Table B 8 . Coefficients All coefficients were significant at the 90% confidence level . Model Sum of Squares df Mean Square F Sig. 1 Regression 1.770 3 .590 78.049 .000b Residual .197 26 .008 Total 1.967 29 a. Dependent Va riable: Equivalent Flexural Strength b. Predictors: (Constant), Fiber Content Cubed, First Peak Strength, Fiber Content Model Unstandardize d Coefficients Standardized Coefficients t Sig. 90.0% Confidence Interval for B B Std. Error Beta Lower Bound Upper Bound (Constant) .268 .085 3.169 .004 .412 .124 First Peak Strength .593 .169 .275 3.503 .002 .304 .882 Fiber Content 1.066 .139 1.422 7.692 .000 .830 1.303 Fiber Content Cubed .573 .125 .803 4.598 .000 .785 .360 a. Dependent Variable: EquivFlexStren

PAGE 150

150 (4) MODEL DIAGNOSTICS Unusual and Influential Data In identifying high ly influential data points in this study, SPSS was set up to flag outliers above +/ 3 standard deviations. None of the observations were flagged. Therefore, no value exerted a significant amount of influence. Generally, there were no problems as far as ex treme and highly influential/leveraged observations were concerned. The relevant statistics are highlighted in yellow in Table B 9. Plots to show the distribution are presented in Figure B 5. Table B 9 . Residual Statistics a Minimum Maximum Mean Std. Dev iation N Predicted Value .0148 .7747 .4420 .24706 30 Std. Predicted Value 1.849 1.347 .000 1.000 30 Standard Error of Predicted Value .022 .044 .031 .006 30 Adjusted Predicted Value .0175 .7397 .4401 .24490 30 Residual .24638 .19530 .00000 .08233 30 Std. Residual 2.834 2.246 .000 .947 30 Stud. Residual 2.991 2.611 .010 1.026 30 Deleted Residual .27457 .26392 .00190 .09699 30 Stud. Deleted Residual 3.622 2.981 .009 1.141 30 Mahal. Distance .930 6.573 2.900 1.409 30 Cook's Distance .000 .59 9 .046 .119 30 Centered Leverage Value .032 .227 .100 .049 30 a. Dependent Variable: Equivalent Flexural Strength

PAGE 151

151 Figure B 4 . Residual plots (a) Histogram of standardized residuals fitted with normal curve (b) Normal probability plot of residuals Normality of Error Terms In order to test the validity of normality of error terms assumption in the MODEL 2 , both KS and Shapiro Wilk tests were performed on the residuals to test the null hypothesis that the residuals from the model are normally distrib uted against the alternative hypothesis that they are not normally distributed. The test results are summarized in Table B 10. Ho: Residuals from Regression are normally Distributed Ha: Residuals from Regression are not normally Distributed Table B 10 . Te sts of Normality Kolmogorov Smirnov a Shapiro Wilk (SW) Statistic df Sig. Statistic df Sig. Unstandardized Residual .143 30 .120 .922 30 .030 a. Lilliefors Significance Correction

PAGE 152

152 The results of the SW test indicate a failure to reject the null hypo thesis that the residuals are normally distributed at the 95% confidence level as is evidenced by the high p value of 0.030 compared to the significance level of 0.05 . The overall conclusion is that the normality of error terms assumption is validated or m et with respect to MODEL 2 . Figures B 6 further supports the normality of error terms assumption. Figure B 5 . Residual diagnostics (a) Histogram of residuals fitted with normal curve (b) normal probability plot of residuals Homoscedasticity Variance o f the residuals (error terms) appear homoscedastic and the residuals appear independent based on Figure A 7.

PAGE 153

153 Figure B 6 . Variance of residuals: Scatter plot of residuals vs predicted values

PAGE 154

154 APPENDIX C ADDITION AL SEM IMAGES Figure C 1 . Fiber showi ng cross linking and surface deformation (embossment) Figure C 2 . Fractured CEB surface showing fiber pullout

PAGE 155

155 Figure C 3 . Fractured fiber after specimen failure Figure C 4 . Fiber embedded in matrix after specimen failure

PAGE 156

156 Figure C 5 . Magnified im age of fractured CEB surface Figure C 6 . Zoomed in image of fractured CEB surface.

PAGE 157

157 LIST OF REFERENCES Acquaye L., (2006). high curing temperatures on the strength, durability and potential of delayed ettringite formation in mass concrete st ructures . PhD dissertation, University of Florida, Gainesville, Florida. Ad am, E. A., and Agib, A. (2001). stabilised earth block manufacture in United Nations Educational, Scientific, and Cultural Organization (UNESCO), Paris, France. Agevi, E. (1999). "Technology dissemination in Kenya ." Development Alternatives Newsletter, 9 (11): 7 9. Aguwu, J. I. (2013). coir reinforced laterite blocks for buildings J . of Civ . Eng and Construction Tech., 4(4):110 115 . Akkaya, Y., Peled, related to fiber length and processing in cementitious composites Mater . and Struct . , 33(8): 515 524. Appropriate technology for socioeconomic development in third world countries ." J . of Tech . Stud . , 26 (1): 33 43. American Concrete Institute (ACI), (1996), "State of the art report on fiber reinforced concrete." ACI Committee , 544.1R , Detroit Michigan, USA. American Co ncrete Institute (ACI), (2001). materials for concrete Educatio n Bulletin E3 01. American Concrete Institute, (2008), "Guide for specifying, proportioning, and production of fiber ACI Committee, 544.3R , Detroit Michigan, USA. American Concrete Institute (ACI) , (2010). "Report on the physical pr operties and durability of fiber ACI Committee , 544.5R , Detroit Michigan, USA. Aruma la, J.O. and Gondal, T. (2007). earth building blocks for affordable housing Proc., T he Construction and Building Research Conf ., T he R oyal Institution of Chartered Surveyors Georgia Tech, Atlanta USA . ASTM (2000). test method for compressive strength of soil cement using portions of beams broken in flexure (Mo dified cube method D1634 , West Conshohocken, PA. ASTM (2005). andard test methods for liquid limit, plastic limit, and plasticity index of soils D4318 , West Conshohocken, PA.

PAGE 158

158 ASTM (2007). test methods for laboratory compaction characteristics of soil using standard effort (12 400 ft lbf/ft3 (600 kN m/m3) D698 , West Conshohocken, PA. ASTM (2009). practice for classification of soils and soil aggregate mixtures for highway construction purposes D3282 , West Conshohocken, PA. ASTM (2010). specification for fiber reinforced concrete . C1116/C1116M , West Conshohocken, PA. ASTM (2010). test methods for flexural bond strength of masonry E518/E518M , West Conshohocken, PA. ASTM (2012) . specification for mortar for unit masonry C270 , West Conshohocken, PA. ASTM (2012). test method for compressive strength of masonry prisms C1314 , West Conshohocken, PA. ASTM (2012). test method for flexural strength of soil cement using simple beam with third point loading D1635/D1635M , West Conshohocken, PA. ASTM (2012). method for flexural performance of fiber reinforced concrete ( using beam with third point loading C1609 , Philadelphia. ASTM (2012 ). test methods for sampling and testing concrete masonry units and related units C140 , West Conshohocken, PA. ASTM (2013). test method for compressive strength of hydraulic cement mortars (using 2 in. or [50 mm] cube specimens C10 9/C109M , West Conshohocken, PA. ASTM (2014). "Standard test method for preconstruction and construction evaluation of mortars for plain and reinforced unit masonry C780 , West Conshohocken, PA. Atahan, H. Behavior of pva fiber reinforced cementitious composites under static and impact flexural effects J. Mater. Civ. Eng. , 25(10), 1438 1445. Auroville Earth Institute (AEI ) (n.d.d). http://www.earth auroville.com/raw_material_introduction_en.php. A ccessed on 8/6/13 Auroville Earth Institute (AEI) ( 2010 ). http://www.earth auroville.com/auram_earth_equip ment_introduction_en.php. Accessed on 8/6/13 . Azeredo, G . , Morel J . C . , and Perazzo , B . N . (2007) . Compressive strength of earth mortars. J Urban Environ Eng., 1(1):26 35 .

PAGE 159

159 Bagherzadeh, R., Sadeghi, A., and Latifi, M. (2012a). polypropylene f ibers to improve physical and mechanical properties of concrete Textile Research Journal , 82:88 96. Bagherzadeh, R., Pakravan, H. R., Sadeghi, A ., Latifi, M., and Merati, A.A. (2012b). investigation on adding polypropylene fibers to reinforce light weight cement composites (LWC) J . of Engineered Fibers and Fabrics , 7(4):13 21. Banthia,N ., and Gupta, R. (2006). shrinkage cracking in concrete Cem . Conc . Res ., 36 (7): 1263 1267. Bin diganavile, V, and Banthia, N. fiber matrix bond in concrete Can. J. Civ. Eng., 32(5), 924 933 Bin ic i,H., Aksogan, O., and Shah, T. (2005). fiber reinforce d mud brick as building material . Constr . Build . Mater . , 19(4):313 318. Bland, D. W. plane cyclic shear performance of interlocking compressed earth block walls . . Bouhi cha, M., Aouissi, F., Kenai, S. (2005). composite soil reinforced with barley straw Cem . Conc . Compos . , 27: 617 621 Boynton, R, Gutschick, K. (1964). mortar and maso nry Masonry Mortar Technical Notes # 1. http://www.lime.org/documents/publications/free_downloads/durability mortar masonry.pdf Accessed 6/17/2013. Burroughs, V.S. (2001). Quantitative criteria for the selection of soils for rammed earth wall construc tion . PhD Thesis, University of New South Wales, Australia. Chan, C. M. (2011). natural fibres inclusion in clay bricks: physico mechanical properties . http://www.waset.org/journals/waset/v73/v73 9.pdf . Accessed 6/17/2013. Geotechnical engineering: principles and practices. upper saddle river NJ, Prentice Hall, Inc. Consoli, N. C., Prietto, P. D. M., Ulbrich L. A. (1998). fiber and cement addition on behavior of sandy soil J . Geotech . Geo environ . Eng ., 124(12):1211 1214. Consoli, N.C., Vendruscolo, M. A., Fonin i, A., and Dalla Rosa, A. (2 009). reinforcement effects on sand considering a wide cementation range Geotex . G eomembranes, , 27:196 203.

PAGE 160

160 Deboucha, S., and Hashim , R. R. (2011). A review on bricks and stabilized compressed earth blocks Sci. Res. Essays , 6 (3) : 499 506 Dhakal , S. K. (2012). very weak subgrade soil with cementitious stabilizers Mechanical College. Drysdale, R. G., Hamid, A Masonry structures: beh avior and design Second Edition , Boulder, CO: The Masonry Society . Eko. R.M., Mpele, M., Doumtsop, M. D., Min sili, L.S., and Wouatong, A.S. (2006) . "Some hydraulic, mechanical, and physical characteristics of three types of compressed earth blocks ." Agr ic . Eng . Int ., CIGR J , Manuscript BC 06 007, Vol. VIII. Eko, R.M., Offa, E. D., Ngatcha, T.Y., and Minsili, L. S. (2012). salvaged steel fibers for reinforcement of unfired earth blocks Constr. Build. Mater. , 35 : 340 346 Elenga R.G., Mabi ala B., Ahouet L., Goma Maniongui J., and Diris G. F. (2011). clayey soils from congo and physical properties of their compressed earth blocks reinforced with post consumer plastic wastes Geomaterials , 1:88 94. Elizabeth , L., and Adam s, C. eds. (2005). construction contemporary natural building methods John Wiley & Sons, New Jersey. Exelbirt, J. (2011). compressed earth bricks based on hygrothermal aging and wind driven rain erosion University of Florida, Gainesville, Florida. Galán Marín , Rivera Gómez, C., Petric, J. (2010). based composite stabilized with natural polymer and fibre Constr. Build. Mater. , 24 (8):1462 1468. processes for th e production of soil cement building blocks PhD Thesis, Warwick University. Guillaud, H., and Joffroy, T. (1995). Compressed earth blocks volume II. Manual of production. Vieweg, Eschborn, Germany. Harshit a Bairagi, R. K.Y. and Jain, R. (2014). ct of jute fibres on engineering properties of lime treated black cotton soi l Int. J. Adv. Res. Technol., 3(2) 1550 1552 Hazeltine, B., and Bull, C. (1999). technology: tools, choices and implications .

PAGE 161

161 Heathcot e, K. A. (2002). erodability of earth wall units PhD Thesis, University of Technology, Sydney, Australia. Herskedal, N. A. (2012). of plane properties of interlocking compressed earth block walls Maste University . Hejazi, S.M., Sheikhzadeha, M., Abtahi, S.M., and Zadhousha, A. (2012). simple review of soil reinforcement by using natural and synthetic fibers Constr. Build. Mater., 30:100 116 Houben, H., and G uillaud, H. (1994). Earth construction A comprehensive guide Intermediate Technology Publications, London. I nternational Code Council (ICC). (2012). Whittier, California. Islam M, and Iwashita K. (2010). tance of adobe reinforced by low cost traditional materials J . Nat . Dis . Sc i. , 2:1 21. Jad hao, P. D., and Nagarnaik, P.B. (2008). soil fly ash mixtures . Proc. 12th Int . Con f., Int . Assoc . of Computer Method s and Advances in Geomechanics (IACMAG) Goa, India. Jayasingh e, C. and Mallawaarachchi, R.S. (2009). strength of compressed Mater . Design , 30(9):3859 3868. Jerome, P., Chiari, G., Borelli, C. (1999). "The arch itecture of mud: constructio n and repair technology in the H adhramaut region of Y emen ." APT Bulletin 30 (2 3): 39 48 Jaquin, P., and Augarde, A. building: history, science and conservation BRE Press: Watford, UK. Kaufmann, J., Winnefeld, F., and Hesselbarth , D. (2004). Effect of the addition of ultrafine cement and short fiber reinforcement on shrinkage, rheological and mechanical properties of portland cement pastes Cem . Concr . Compos . , 26 : 541 549 . Kerali, A. (2001). co mpressed and cement stabilized building bricks Thesis, University of Warwick, School of Engineering, Warwick, UK. Kestner, D., Goupil, J., and Lorenz, E. eds. (2010). guidelines for the structural engineer American Society of Civil Engineers. Khedari J., Watsana sathaporn P., and Hirunlabh, J. (2005). fibre based soil cement block with low thermal conductivity Cem . Concr . Compos ., 27:111 116.

PAGE 162

162 Kibert C.J. (2003). start of a sustainable materials strategy for the built environment UNEP Indus . Environ ., 26(2 3):84 88. Kitingu, S.H. (2009). flexibility, alignment accuracy and load bearing PhD Thesis, University of Warwick, Sch ool of En gineering, Warwick, UK. Klinge r, R. E. (2010) . structural design . Hill Kumar, A., Walia, B.S. and Mohan, J. (2010). strength of fiber reinforced highly compressible clay Constr. Build. Mater., 20(10): 1063 1068. Lawrence. R . Proc. LEHM 5th Int.Conf. on Build. with Earth , Koblenz, Germany; 66 73. Li, C. (2005). response of fiber reinforced soil . Ph.D. Dissertation, University of Texas, Austin, USA. Li, V., and Kanda, T. forum: engineered cementitious composites for structural applications J. Mater. Civ. Eng. , 10(2):66 69. Lin, C., Kayali, O. , Morozov, E.V. and Sharp, D.J. (2014). fibre type on flexura l behaviour of self compacting fibre reinforced cementitious composites Cem. Conc. Compos ., 51:27 37 . Maher, M. H., and Ho, Y. C. (1994). properties of kaolinite fiber soil composite J . Geotech . Geo environ . Eng . , 120, (8):1381 1393. Manjun ath K. R., Venugop al G, and Rudresh A. N. (2013). of random inclusion of sisal fibre on strength behavior of black cotton soil Int. J. Adv. Res. Technol., 2 (7): 2227 2232 . Masonry Standards Joint Commi ttee (MSJC) (2008). ements for mason ry structure . 08/ACI 530 08/ASCE5 08 , USA . Mbumbia, L., Mertens de Wilmars, A., and Tirlocq, J. (2000). "Performance characteristics of lateritic soil bricks fired at low temperatures: a case study of cameroon ." Constr. Build. Mat er., 14: 121 131. McGinley, W.M. (2004). unit bond wrench test trials Final Report http://www.ncma.org/foundation/programs/Lists/Funded%20Grants%20and%20S cholarships/Attachments/19/F2004.002%20 %20NCA%20T%20 %20Bond%20Wrench%20 %20Final%20Report.pdf . Accessed 3/12/2013. Mesbah A., Morel, J. C., Walker, P. and Ghavami, K. (2004). direct tensile test for compacted earth blocks reinforced with natural fibers J. Mater. Civ. Eng., 16:95 98.

PAGE 163

163 Minke G., (2009). B uilding with earth: design and technology of a sustainable architecture Birkhäuser Publish ers for Architecture . Mohr, B.J. , El Ashkar, N.H., Kurtis, K.E. (2004). cement composites for housing construction: state of the art review Proc . Housing Research Agenda Workshop , National Science Foundation, Orlando, FL . Morel, J.C., Pkla, A. (20 02). model to measure compressive strength of compressed Constr. Build. Mater., 16:303 310 . Morel, J.C., Pkla, A., and Walker, P. (2007). strength testing of compressed earth blocks Constr. Build. Mater., 21(2): 303 309. Morony, J. J. (2004). data for heat wave a preliminary report Del Rio, Texa s: South West Texas Junior Coll ge. http://www.toolbase.org/PDF/ Techinv/LoggedDataHeatWave.pdf . Accessed on 6/17/13. Morton, T. (2008). design and construction guidelines BRE Press, Bracknell, Berkshire, UK. Montgomery, D.E. (2002). compacted cement stabilized soil blocks for low cost w alling . Warwick, UK. Murphy, M., and Taub, M. (2010). "The problem and potential of sustainable design in resource poor settings: cases from rawanda ", African Tech . Dev . Forum , 7(1/2):24 30. Namang o, S. (2006). cost effective earthen building material for housing wall construction: investigations into the properties of compressed earth blocks stabilized with sisal vegetable fibers, cassava powder and cement compositions PhD Thesis, Brandenburg University of Technology, Germany. National Concrete Masonry Association (NCMA). (1994). evaluation of the flexural tensile strength of concrete masonry. http://www.ncma.org/resources/design/Research%20Reports/MR10.pdf . Accessed 3/24/2012. National Institute of Building Sciences (NIBS). (2010). Whole building design guide. h ttp://www.wbdg.org/design/sustainable.php . Accessed 09/15/ 2014). New Mexico Construction Bureau (2009) . "New mexico adobe and rammed earth building code ." Construction Industries Division, General Construction Bureau, New Mexico. Norton, J. (1997) ding with e arth Intermediate Technology Publication London.

PAGE 164

164 Obonyo, E., Exelbi rt. J., and Baskaran, M. (2010). compressed earth bricks: assessing erosion resistance using the modified spray testing J . Sust . , 2(12):3639 3649 Obonyo, E. ( 2011). physical, mechanical and hygrothermal performance of compressed earth bricks J . Sust . , 3(4):596 604 Oti, J. E., and Kinuthia, J. M. (2012). Stabilized and sustainable use Applied Clay Sci . , 58:52 59. Patel, P. A. , Desai, A. K., and Desai, K. A. (2012). engineering properties for polypropylene fiber reinforced concrete Int. J. Adv. Res. Technol., 3(1): 42 45 Prasad, C.K.S, Nambi ar, E.K.K., and Abraham., B. M. (2012). fiber reinforced soil blocks as a sustainable building material Int. J. Adv. Res. Technol., 1(5):42 45 Prochazka (2010) . polypropylene fibres at higher temperature conditions Proc . 35th Conf . on Our World in Concrete & Structures, Singapore. Rigassi, V. (1995). earth blocks volume 1. manual of production Vieweg, Eschborn, Germany. Sarangapani, G., Venkatarama R eddy, B.V., and Jagadish, K.S. (2005). "Brick mortar bond and masonry compressive strength ." J. Mater. Civ. Eng., 17: 229 237 Rai C. D. (2007). ildings Final Report: A Earthquake Codes IITK GSDMA Project on Building Codes . http://www.iitk.ac.in/nicee/IITK GSDMA/EQ10.pdf . A cc essed on 6/17/13 . Ramachandra n, V.S., and Beaudoin, J.J. eds . (2001). Scanning electron microscopy, x ray microanalysis of concrete . . Sci . Tech ., Noyes Publication, US, 231 273. Accelerated masonry construction : Review and fu Proc. of Struct . Eng .M at er. ; 6:1 9 . Segetin M, Jayaraman K, Xu X. (2007). reinforcement of soil cement building materials: manufacturability and properties . Build . Env . , 42:3066 3079. Smith, E. W. (1981). brick productio n in New Mexico New Mexico Geology , 3(2):17 24 Song, P.S., Hwang, S., and Sheu, B.C. (2005). properties of nylon and polypropylene fiber reinforced concretes Cem. Concr. Re s., 35(8):1546 1550.

PAGE 165

165 Stirling, B. J . (2011). Flexural b ehavior o f i nterlocking compressed earth b lock shear walls s ubjected t o in plane l oading . University. Sukontasukku l, P., and Jamsawang, P. (2012). steel and polypropylene fibers to improve flexural performance of deep soil cement column Constr. Build. Mater., 29:201 205 Tennant, A., Foster, C., and Reddy, B. (2013). Flexural Behavior of Cement J. Mater. Civ. Eng. , 25(3), 303 307. Tang, C., Ga o, W. , Fengjun, C., and Cai, Y. (2007). mechanical behavior of short polypropylene fiber reinforced and cement stabilized clayey soil Geotextiles and Geomembranes , 25(3):194 202. Tonoli, G., Savastano, H., Jr., Santos, S., D ias, C., John, V., an d Lahr, F. (2011). reinforcement of sisal and polypropylene fibers in cement based composites. J. Mater. Civ. Eng., 23(2):177 187. Uday Vyas, C, H., and Venkatarama Reddy, B.V. (2010). solid block masonry prism compressive strengt h using FE model Mater . Struct ., 43(5): 719 735 . United Nations Centre for Human Set tlements (UN Habitat) (1992). "Earth construction technology ." Nairobi. United Nations Centre for Human Settlements (UN Habitat) (2001). "Poster 1 of 5 on low cost bui lding materials : Compressed earth blocks ." Nairobi. http://ww2.unhabitat.org/programmes/housingpolicy/documents/Compressed.pdf . Accessed on May 29, 2013. United Nati ons Centre for Human Settlements (UN Habitat) urban shelter: global report on human settlements http://www.unhabitat.org/documents/media_centre/ghs/GRHS05 PR1.pdf . Accessed on May 29, 2013. United Nations International Strategy for Disaster Reduction ( UNISDR ) (2007). Terminology on DRR www.unisdr.org/we/inform/terminology . Accessed 03/11/2014. Venkatar ama Reddy, B. V., and Gupta, A. (2006). bond strength of soil cement block masonry couplets using cement soil mortars J. Mater. Civ. Eng., 18(1):36 45 Venu Madhava Rao, K., Venkatarama Reddy, B. V., and Jagadish, K. S. (1996). bond strength of masonry using various blocks and mortars Mater. Struct ., 29(2), 119 124.

PAGE 166

166 Villamizar, M. C. N., Araque, V.S. , Reyes, C. A., and Silva, R.S. (2012). addition of coal ash and cassava peels on the engineering properties of com Constr. Build. Mater., 36: 276 286 Walker, P. (1995). durability and shrinkage characteristics of cement Cem. Conc. Compos., 17 (4):301 310 . Walker, P. , ement stabilized compressed Mater. Struct ., 30:545 551. Walker, P. (2004). erosion characteristics of earth blocks and earth block J. Mater. Civ. Eng., 16: 497 506 . Wei smann, A., and Bryce, K. (2006). uilding with Cob: A Green Books, Totnes. Wilson, A., and Abolmaali, A. (2013) material behavior of steel and synthetic fibers in dry cast application. Trans . Res . Rec ., 2332 , Transportation Research Board, Washington, DC, 23 28 . Wilson , A. and Abolmaali, A. (2014). Performance of synthetic fiber reinforced concrete pipes J. Pipeline Syst. Eng. Pract. , 5(3), 04014002. Zhang, P., Li, Q. and Hua, W. (2010). flexural properties of cement J. Mater. Civ. Eng., 22:1282 1287 .

PAGE 167

167 BIOGRAPHICAL SKETCH Peter Donkor was born and raised in Accra, Ghana. He received his Bachelor of Science degree in b uilding t echnol ogy at the Kwame Nkrumah University of Science and Technology in 2005. Upon receiving his undergraduate degree, Peter worked in the construction industry in Ghana for a year and then move d to the USA to pursue a Master of Science degree in building c onstru ction at the University of Florida, Gaines ville. After graduating with a m Manager for a construction firm and returned to the M.E. Rinker Sr. School of Construction Management in 2012 to pursue a PhD. Upon the completion of this PhD, Peter intends to work in a project management position in the construction industry. Peter is a certified Project Management professional (PMP).