UNIVERSITY OF CINCINNATI

8/16/2006 Date:___________________
Amarendranath Deshini I, _________________________________________________________,
hereby submit this work as part of the requirements for the degree of:

Master of Science
in:

Civil Engineering
It is entitled:

Finenes of Densified Microsilica and Dispersion in Concrete Mixes

This work and its defense approved by:

Anastasios M . Ioannides Chair: _______________________________ Richard A. Miller _______________________________ Issam A. Minkarah _______________________________ Ossama M. Salem _______________________________

_______________________________

Fineness of Densified Microsilica and Dispersion in Concrete Mixes

A thesis submitted to the Division of Research and Advanced Studies of the University of Cincinnati in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in the Department of Civil and Environmental Engineering of the College of Engineering 2007 by Amarendranath Deshini B.E. (Civil Engineering), Osmania University Hyderabad, India, 2001

Committee Chair: Dr. Anastasios M. Ioannides, P.E.

ABSTRACT

This study explores the effect of densification of microsilica on the mechanical and other engineering properties of concrete used on Ohio Department of Transportation (ODOT) projects. American Society for Testing and Materials (ASTM) C 1240 requires wet-sieved microsilica to pass a No. 325 sieve with no more than 10% retained. Densified microsilica samples submitted to ODOT sometimes do not meet this specification, since the sieving process may not be able to break the bonds formed due to densification. During this study, No. 325 sieve tests on three microsilica types (undensified, densified, and abused by prolonged exposure to moisture) were performed at the ODOT laboratory, but none of the materials tested were found to conform to the ASTM fineness specification. This calls into question the application of this procedure to assessing the suitability of densified microsilica for use in concrete. In contrast, the compressive and flexural strengths of concretes mixed with each of the three microsilica types exceeded those envisaged by ODOT Item 499.03 Concrete-General: Proportioning. As expected, undensified microsilica concrete yielded higher values than its densified and abused microsilica counterparts at all ages, but this advantage was rather limited. This was true for both natural and crushed coarse aggregate mixes. With very few exceptions attributable to material and testing variability, trends observed with regard to the effects of microsilica and coarse aggregate types, age and specimen size on the development of strength were also as anticipated. Therefore, it is concluded that

iii

densified microsilica can be used on ODOT projects for the construction of pavements and bridges.

iv

ACKNOWLEDGEMENTS

I would like to thank my graduate advisor Dr Anastasios M. Ioannides and coadvisor Dr Richard A. Miller, without whom this thesis would not have been possible. The constant support I received from them helped me to complete my Masters degree successfully and also to shape my professional career. I would also like to thank Dr Sam M. Salem, Dr Isaam A. Minkarah for readily accepting my request to be part of my thesis committee. The assistance and cooperation of Mr Bryan Struble, Ohio Department of Transportation (ODOT) Technical Liaison, Mr Lloyd Welker, the Research Manager, Ms Monique Evans, the Administrator for the Office of Research and Development at ODOT and Mr Herman Rodrigo, the Federal Highway Administration (FHWA) liaison in Columbus, OH, was a major contributor to the success of the study, and their support is gratefully acknowledged. Special thanks are also extended to Tim Jones, ODOT laboratory technician, who conducted the tests on microsilica. I am thankful to Mr Jim Martin with Martin Marietta Materials for providing sand and the aggregates, Mr Steve Reibold with CEMEX for donating cement, Mr Tony N. Kojundic with ELKEM Materials for microsilica and Mr Greg Wirthlin with Master Builders for supplying the air entrainer and the plasticizer. All these materials were provided free of cost for the research. The contributions of graduate students Kristy M. Walsh and Jeff C. Mills to the project are greatly appreciated.

vi

I am ever grateful to my parents, sister, wife and friends for the encouragement that I received from them from time to time. My special thanks to Mr Vamshidhar Thakkalapalli for his support and for suggesting University of Cincinnati for my graduate studies. Lastly I would like to express my gratitude to the Department of Civil and Environmental Engineering for giving me an opportunity to do my Master of Science in Construction Engineering and Management at University of Cincinnati. During my studies at the University, I received financial assistance in the form of Research and Teaching Assistantships (April 2002 till December 2003) and University Graduate Scholarship (September 2001 till April 2004).

vii

TABLE OF CONTENTS
Page

ABSTRACT ACKNOWLEDGEMENTS TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES LIST OF SYMBOLS AND ABBREVIATIONS LIST OF SPECIFICATIONS CITED
SI* (MODERN METRIC) CONVERSION FACTORS

iii vi vii viii xiv xvi xxi xxiii

INTRODUCTION
1.1 1.2 1.3 1.4 Problem Statement Objectives Technical Background Report Organization

1 1 2 3 5

LITERATURE REVIEW
viii

2.1 2.2 2.3 2.4

Introduction Production of Microsilica Properties of Microsilica Effect of Microsilica on Physical Properties of Concrete
2.4.1 2.4.2 2.4.3 2.4.4 2.4.5 2.4.6 2.4.7 Water Demand Cohesiveness Slump and Workability Air Content Setting Time Shrinkage and Cracking Heat of Hydration

7 8 9 10

11 12 12 13 13 14 17 17

2.5

Effect of Microsilica on Mechanical Properties of Concrete

2.5.1 2.5.2 2.5.3 2.5.4 2.5.5 Compressive Strength Flexural Strength Modulus of Elasticity Bond Strength Strength Development

17 18 19 19 20 22 24

2.6 2.7

Field Tests Do Undispersed Agglomerates Matter?

ix

MATERIALS AND PROCEDURES

3.1 3.2 3.3 Introduction Materials Used Tests on Microsilica
3.3.1 3.3.2 3.3.3 Preparation of Abused Microsilica Fineness Tests Gradation Tests

27 27 27 29 29 30 31 34 34 35 36 37 38 38 39 39 41 42 42 44

3.4

Tests on Aggregates
3.4.1 3.4.2 3.4.3 3.4.4 3.4.5 Specific Gravity and Absorption of Coarse Aggregate Specific Gravity and Absorption of Fine Aggregate Bulk Density of Coarse Aggregate Sieve Analysis of Fine and Coarse Aggregates Moisture Content of Fine and Coarse Aggregates

3.5

Mix Design
3.5.1 3.5.2 Constants and Variables Ingredients

3.6

Mixing, Casting, and Curing Methods

3.6.1 3.6.2 3.6.3 Mixing Concrete Casting Specimens Curing

3.7 3.8

Tests on Plastic Concrete Strength Test Procedures on Hardened Concrete

3.8.1 3.8.2 Compressive Strength Flexural Strength

44 46 46 47

TEST RESULTS
4.1 4.2 Introduction Microsilica
4.2.1 4.2.2 Microsilica Fineness Microsilica Gradation

49 49 49 49 50 50 50
51

4.3 4.4 4.5

Aggregates Concrete Mixes Mechanical Properties

DISCUSSION OF RESULTS
5.1 5.2 5.3 5.4 Introduction Microsilica Fineness Microsilica Gradation Variability of Mechanical Tests
5.4.1 Compressive Strength xi

72 72 72 74 74 75

5.4.2

Modulus of Rupture

76 76 77 79
79

5.5

Data Interpretation
5.5.1 5.5.2 Compressive Strength Modulus of Rupture

5.6

Effect of Microsilica Type on Mechanical Properties of Concrete

5.6.1 5.6.2 Compressive Strength Modulus of Rupture

79 81 82

5.7

Effect of Coarse Aggregate Type on Mechanical Properties of Concrete

5.7.1 5.7.2 Compressive Strength Modulus of Rupture

82 83

5.8

Effect of Specimen Size on Mechanical Properties of 84 Concrete

CONCLUSIONS AND RECOMMENDATIONS

6.1 6.2 6.3 6.4 Summary Conclusions Recommendations Implementation Plan
xii

110 110 112 114 115

REFERENCES

117

xiii

LIST OF TABLES
Page 4.1 Results of Tests Conducted on Microsilica Types at ODOT Laboratory 4.2 4.3 4.4 4.5 4.6 4.7 Hydrometer Test on Undensified Microsilica Hydrometer Test on Densified Microsilica Hydrometer Test on Abused Microsilica Dry Sieve Analysis of Undensified Microsilica Dry Sieve Analysis of Densified Microsilica Dry Sieve Analysis of Abused Microsilica 53 54 55 56 57 58 59 59 60 61 62 63 65 67 69 86 87 87 52

4.8 (a) Aggregate Sieve Analysis 4.8 (b) Physical Aggregate Properties 4.9 4.10 4.11 4.12 4.13 4.14 4.15 5.1 5.2 5.3 Ingredients by Mix (per yd3 of concrete) Specimens Cast by Batch Physical Properties by Batch Compressive Strength for Large Cylinders, fc Compressive Strength for Small Cylinders, fc Modulus of Rupture for Large Beams, MR Modulus of Rupture for Small Beams, MR Coefficients of Variation, COV (%), in Laboratory Test Results Average Compressive Strength for Large Cylinders (psi) Average Compressive Strength for Small Cylinders (psi)

xiv

5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11

Relative Compressive StrengthValues for Large Cylinders (%) Relative Compressive Strength Values for Small Cylinders (%) Best-Fit Compressive Strength Values for Large Cylinders (psi) Best-Fit Compressive Strength Values for Small Cylinders (psi) Average Modulus of Rupture Values for Beams (psi) Relative Modulus of Rupture Values for Beams (%) Best-Fit Modulus of Rupture for Small Beams (psi) Cylinder Size Factors (%)

88 88 89 89 90 90 90 91

xv

LIST OF FIGURES
Page 4.1 Grain Size Distribution of Undensified, Densified, and Abused Microsilica 5.1 5.2 5.3 5.4 Trend Line Curves for Large Cylinders Trend Line Curves for Small Cylinders Trend Line Curves for Small Beams Effect of Microsilica Type on Large Cylinders with Natural Aggregate 5.5 Effect of Microsilica Type on Small Cylinders with Natural Aggregate 5.6 Effect of Microsilica Type on Large Cylinders with Crushed Aggregate 5.7 Effect of Microsilica Type on Small Cylinders with Crushed Aggregate 5.8 Effect of Microsilica Type on Small Beams with Natural Aggregate 5.9 Effect of Microsilica Type on Small Beams with Crushed Aggregate 5.10 Effect of Aggregate Type on Large Cylinders with Undensified Microsilica 5.11 Effect of Aggregate Type on Small Cylinders with Undensified 102 101 100 99 98 97 96 71 92 93 94 95

xvi

Microsilica 5.12 Effect of Aggregate Type on Large Cylinders with Densified Microsilica 5.13 Effect of Aggregate Type on Small Cylinders with Densified Microsilica 5.14 Effect of Aggregate Type on Large Cylinders with Abused Microsilica 5.15 Effect of Aggregate Type on Small Cylinders with Abused Microsilica 5.16 Effect of Aggregate Type on Small Beams with Undensified Microsilica 5.17 Effect of Aggregate Type on Small Beams with Densified Microsilica 5.18 Effect of Aggregate Type on Small Beams with Abused Microsilica 109 108 107 106 105 104 103

xvii

LIST OF SYMBOLS AND ABBREVIATIONS

A: % absorption A: weight of oven-dry sample AASHTO: american association of state highway and transportation officials AC: abused microsilica concrete with crushed aggregate ACI: american concrete institute Al2O3: aluminum oxide AN: abused microsilica concrete with crushed aggregate ASG: apparent specific gravity ASR: alkali silica reactivity ASTM: american society for testing and materials B: weight of pycnometer with water B: weight of saturated surface dry sample in air BSGSSD: bulk specific gravity at saturated surface dry condition b: average width C: cement C: weight of pycnometer with water C: weight of saturated sample in water C0: zero correction CT: temperature correction factor CaO: calcium oxide C+P: cement + pozzolan

xviii

CA: coarse aggregate COV: coefficient of variation cwt: hundred weight D: weight of oven-dry sample D: diameter of particle DC: densified microsilica with crushed aggregate DN: densified microsilica with crushed aggregate d: average depth FA: fine aggregate Fe2O3: ferric oxide Fh: hydrometer percent finer, FM: fineness modulus f : compressive strength ft: feet G: grams G: mass of aggregate and measure G1: specific gravity of liquid in which the sample was suspended Gs: specific gravity of sample Gw = specific gravity of water Hrs: hours in.: inch K2O: potassium oxide

xix

kg: kilogram km: kilometer L: distance from the surface of the suspension to the level at which the density of the suspension is being measured L: liter Lb: pound LOI: loss on ignition l: effective span M: oven dry bulk density MgO: magnesium oxide Min.: minutes Mm: millimeter MnO: manganese oxide MR: modulus of rupture Ms = mass of oven dry material MSSD: bulk density of aggregate at ssd condition MW: mass of water m: meter mm: millimeter N: newton NA: not applicable NA: not available

xx

Na2O: sodium oxide NaPO3: Sodium Metaphosphate NCHRP: national cooperative highway research program No.: number NT: not tested n: coefficient of viscosity of water ODOT: ohio department of transportation oz: ounce P: load at failure P: pozzolan PCC: portland cement concrete P2O2: phosphorus oxide pcf: pounds per cubic feet psi: pound per square inch Ractual: actual hydrometer reading Rc: corrected hydrometer reading r: rate of application of load S = rate of increase in extreme fiber stress S: weight of ssd sample SEM: scanning electron microscopy SiO: silica oxide SiO2: silicon-di-oxide

xxi

SO3: sulfur trioxide Sq: square SSD: surface dry condition T: interval of time from beginning of sedimentation to the taking of the reading T: mass of measure TEM: transmission electron microscopy TiO2: titanium dioxide UC: undensified microsilica with crushed aggregate UN: undensified microsilica with natural aggregate US: united states of america V: volume Vair: volume of air VSSD: volume at saturated surface condition W: density of water W: weight WCEM: weight of cement WMS: weight of microsilica WMS (SSD): weight of microsilica at ssd WW (SSD): weight of water at ssd w: moisture content w/c: water cement ratio yd: yard

xxii

%: percentage C: degree celcius F: degree fahrenheit = absolute viscosity of the suspension water : micro m: micro meter w: unit weight of water

xxiii

LIST OF SPECIFICATIONS CITED

AASHTO M 154 05 Standard Specification for Air-Entraining Admixtures for Concrete AASHTO T 22-03 Compressive Strength of Cylindrical Concrete Specimens ASTM C 29/C 29M 97 Standard Test Method for Bulk Density (Unit Weight) and Voids in Aggregate ASTM C 39/C 39M 01 Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens ASTM C 78 02 Standard Test Method for Flexural Strength of Concrete (Using Simple Beam with Third-Point Loading) ASTM C 127 01 Standard Test Method for Density, Relative Density (Specific Gravity), and Absorption of Coarse Aggregate ASTM C 128 97 Standard Test Method for Specific Gravity and Absorption of Fine Aggregate ASTM C 136 96a Standard Test Method for Sieve Analysis of Fine and Coarse Aggregates ASTM C 143/C 143M 00 Standard Test Method for Slump of Hydraulic-Cement Concrete ASTM C 192/C 192M 00 Standard Practice for Making and Curing Concrete Test Specimens in the Laboratory ASTM C 231 97 Standard Test Method for Air Content of Freshly Mixed Concrete by the Pressure Method ASTM C 260 01 Standard Specification for Air-Entraining Admixtures for Concrete ASTM C 430 96 Standard Test Method for Fineness of Hydraulic Cement by the 45-m (No. 325) Sieve ASTM C 566 89 Standard Test Method for Total Moisture Content of Aggregate by Drying

xxiv

ASTM C 1240 01 Standard Specification for Use of Silica Fume as a Mineral Admixture in Hydraulic-Cement Concrete, Mortar and Grout ASTM D 421 85 Standard Practice for Dry Preparation of Soil Samples for ParticleSize Analysis and Determination of Soil Constants ASTM D 422 63 Standard Test Method for Particle-Size Analysis of Soils ASTM D 854 06 Standard Test Methods for Specific Gravity of Soil Solids by Water Pycnometer ODOT Item 499.03 Concrete-General: Proportioning ODOT Item 499.03 Concrete-General: Proportioning; Slump ODOT Item 703.01 Aggregate-General: Size ODOT Item 703.02 Aggregate for Portland Cement Concrete: Fine Aggregate ODOT Supplemental Specification 848 Bridge Deck Repair and Overlay with Concrete Using Hydro-Demolition ODOT 2002 Construction and Material Specifications U.S. Army Corps of Engineers CRD-C 13 Standard Specification for Air-Entraining Admixtures for Concrete

xxv

xxvi

INTRODUCTION

1.1

Problem Statement
Microsilica has proven to be an excellent admixture for Portland cement concrete.

Addition of microsilica to a concrete mix usually results in significant improvements in strength, durability and permeability. Some of the improvements to the concrete properties occur because microsilica is a pozzolan. Pozzolans are finely divided silica that combine with free lime (calcium hydroxide) to create more calcium silicate hydrate. When conventional Portland cement hydrates, it produces both calcium silicate hydrate, the chemical glue that makes the cement hard, and free lime, which is weak and highly soluble. This strengthens significantly the cement matrix and decreases its permeability. Microsilica is estimated to be about a hundred times finer than cement, giving it the ability to plug voids between cement particles, and helping it increase the density of the cement matrix (Malhotra, et al., 1987). In its natural state, microsilica is extremely fine, having a particle size about the same as that of cigarette smoke, and this makes it difficult to handle. Microsilica had originally been used as a slurry, but this form was also found to be inconvenient for ready-mix plants. The solution to this problem is densification. The microsilica is bonded into larger particles that are far easier to handle (Malhotra, et al., 1987). There is some concern, however, as to the effect of densification on the quality control process. The American Society for Testing and Materials (ASTM) C 1240 01 Standard Specification for Use of Silica Fume as a Mineral Admixture in Hydraulic-Cement 1

Concrete, Mortar and Grout requires in its Table 2 that wet-sieved microsilica pass the No. 325 sieve (45 m) with no more than 10% retained, and advises that care be exercised to avoid retaining agglomerations of extremely fine material. In its Appendix X2, however, ASTM 1240 01 also states: The 45-m (No. 325) sieve specification is to be used to determine the amount of foreign material present; good judgment must be exercised to differentiate between easily dispersible agglomerates and foreign materials. The Ohio Department of Transportation (ODOT) has found that densified microsilica samples submitted sometimes do not meet this specification, apparently because wet-sieving is not capable of breaking the bonds formed during the densification process. On the other hand, it may be argued that densification bonds are temporary and are easily overcome during concrete mixing itself. It is also possible that densified microsilica will still have adequate field performance even if a certain percentage of its particles do not pass the No. 325 sieve.

1.2

Objectives
In this study, two important questions that arise out of the processes involved in

packaging and subsequent mixing microsilica into fresh concrete are explored. The first pertains to increased particle size that results during densification. The second issue is related to the possible repercussions of densification into larger particle sizes on the mechanical and other engineering properties of microsilica concrete. Because this type of concrete is produced and utilized around the world in increasing amounts (Helland, et al., 1988), and because these two questions have not been answered conclusively yet in

the published literature, research reported herein assumes considerable significance, with direct financial and engineering consequences. Whether densification increases the amount retained on the No. 325 sieve to more than 10%, and whether such an increase is detrimental to the engineering properties of microsilica concrete are issues that need to be investigated using a well-designed factorial of experiments.

1.3

Technical Background
Microsilica is obtained as a by-product in the manufacture of silicon and

ferrosilicon, from a procedure that involves the reduction of high purity quartz with coal at a temperature of 3300F (1816C) in an electric arc furnace. Consisting in excess of 85% of amorphous non-crystalline silica (SiO2), microsilica is collected as the tiny particulate matter present in the emissions from this combustion process, a material that would otherwise have to be landfilled. Individual microsilica particles are spherical in shape and measure about 3.937 micro inch (0.1 m) in diameter, i.e., they are about 100 to 150 times smaller than Portland cement particles. The bulk density of microsilica is in the range of 10 to 15 pcf (150 to 250 kg/m3), and its specific surface area is on the order of 10,000 to 13,000 yd2/lb (20,000 to 23,000 m2/kg) (Malhotra, et al., 1987). Microsilica added to fresh concrete reacts with the calcium hydroxide produced during the hydration of Portland cement to produce increased amounts of calcium silicate hydrate. This results in a much stronger bond between the cement paste and the coarse aggregate, thereby leading to increased compressive strength. Moreover, the additional calcium silicate hydrates produced are much more resistant to chemical attack than the

weaker calcium hydroxide. Another beneficial mechanism operative when microsilica is used derives from the fineness of its particles and is referred to as the micro-filler effect. Filling of voids in the matrix leads to a much denser pore structure, and reduces the number and size of capillaries that would enable contaminants to infiltrate the concrete (www.norchem.com/appl-works.html; accessed: 07/21/05). The combined action of microsilica as a pozzolan and as a filler, results in concrete that can be of very high strength and durability. No more than thirty years ago, 6,000 psi (41.37 x 106 N/m2) concrete was considered to be high strength; using microsilica, compressive strengths of up to 20,000 psi (137.90 x 106 N/m2) are reported in the literature. Similarly, the modulus of elasticity and the flexural strength at 28 days are also higher than in ordinary Portland cement concrete (Helland, et al., 1988). Improvements in durability and in scaling resistance result from greatly reduced fluid permeability and ionic diffusivity and the concomitant increased resistance to penetration by chloride ions, most notably present in deicing or marine salts. Microsilica concrete can also exhibit very good freeze-thaw durability provided the air entrained is controlled. Reduction in the alkalinity of the pore solution and in the diffusion of alkali ions and water lead to a decrease in expansion and in alkali-aggregate reactivity. High early strengths and resistance to abrasion are additional benefits (Malhotra, et al., 1987). Microsilica used in concrete is available in three forms: water slurry, dry uncompacted powder, and dry densified (compacted) powder. The microsilica content of the slurry form is about 50% by weight, the remainder being water. While used commercially in this from, slurried microsilica can be difficult to handle in ready-mix

plants without special equipment. On the other hand, handling of the uncompacted powder poses a potential health risk, since it may be breathed in by construction personnel. Because of handling problems with both uncompacted and slurried microsilica, compacted or densified microsilica is preferred. Dry compacted microsilica is believed to have the same performance characteristics as the uncompacted material. At typical densities of 40 pcf (640.8 kg/m ), its handling qualities approach those of Portland cement, whose density is usually around 94 pcf (1505.88 kg/m ). The bulk density of uncompacted microsilica is typically 15 pcf (240.30 kg/m ). Compacted microsilica is virtually free of dust and lumps, flows readily in pneumatic lines or along bucket elevators, and can be stored in ordinary cement silos or transported in bulk cement tankers (Malhotra, et al., 1987).
3 3 3

1.4

Report Organization
This report is divided into six chapters. The first chapter provides a definition of

the problem explored, identifies the objectives of the study, and outlines the technical background of the research. Chapter 2 presents a literature review into the use of microsilica in concrete mixes, and the resulting effects on the physical, mechanical, and durability properties of concrete. The third chapter offers a detailed description of the mixing, casting, curing and testing procedures adopted during this project. Results from the tests conducted are presented in Chapter 4, which is subdivided into three sections, one for each of the microsilica types used, viz., densified, undensified and microsilica abused in the laboratory by wetting and drying. The discussion of these results in the 5

fifth chapter involves a further subdivision, this time depending on the coarse aggregate type employed, i.e., natural or crushed. In addition, the effects of specimen age and size on the compressive and flexural strengths of microsilica concrete are examined. On the basis of these tests, a number of recommendations to ODOT are formulated regarding the future application of microsilica in concrete pavements and bridges. These recommendations along with summary and conclusions are presented in Chapter 6.

LITERATURE REVIEW

2.1

Introduction
This chapter summarizes the literature on the use of microsilica as a concrete

admixture, and describes its effect on the physical, mechanical and durability properties of the resulting mix. In general, the properties of concrete depend on a number of variables, making it hard to identify the exact cause of a particular behavior change. The best available information from previous investigations and case histories is presented in this chapter. Microsilica was first introduced on an experimental basis as a concrete additive in the Scandinavian countries in the 1950s (Helland, et al., 1988). The development of high range water reducers in Europe and Japan in the early 1970s led to a reduction in the water/cement ratio of microsilica concrete, while ensuring acceptable workability, and contributed to the more widespread use of the product. Microsilica concrete highway applications in the United States did not begin until the mid-1980s, with trial placements of full-depth decks and overlays in the state of Ohio (Whiting and Detwiler, 1998). By the early 1990s, microsilica was being used by nearly 30 state agencies to varying degrees. Today, states like Ohio and New York place microsilica concrete overlays every year, while others explore the use of microsilica on an experimental basis. Microsilica contents have ranged from 5 to 12%, and typical water-to-cementitious material ratios have been between 0.3 and 0.4. There is no consensus yet on the optimum values of these percentages (Whiting and Detwiler, 1998). 7

2.2

Production of Microsilica
As already noted in Chapter 1, microsilica is a by-product of the ferro-silicon

alloy and silicon metal industries. Silicon, ferro-silicon, and other alloys of silicon are produced in electric arc furnaces, where quartz is reduced by carbon at very high temperatures. In the process, the silica oxide (SiO) vapors produced, oxidize and condense to form very tiny spheres of noncrystalline silica (silicon dioxide or SiO2). The latter is highly pozzolanic, and is recovered by passing the outgoing flue gas through a baghouse filter (Malhotra, et al., 1987). It seems that microsilica containing more than 78% SiO2 in amorphous form can be used in cement and concrete. Malhotra and Mehta (1996) maintain that the current world production of microsilica appears to be about one million tons per year with Norway and the United States numbering among the major producers. Nonetheless, they also note that in spite of several technical advantages, only a small percentage of the current supply of microsilica is being used as a mineral admixture in the cement and concrete industries, attributing this to the high cost of the material, as well as handling difficulties. Microsilica particles are extremely fine and have low bulk density, which makes handling and transportation of the material difficult. Therefore, microsilica is generally transported and used in the form of slurry or pellets, or as a densified powder (www.silicafume.org/general-concrete.html; accessed: 07/22/05). Undensified microsilica, harvested directly from the baghouse filters, is not commonly used. This form would be very useful as an admixture in concrete, but it is very difficult to handle because of its very low loose bulk density. Sometimes, undensified microsilica is

blended with cement to reduce handling and transportation costs. In order to make microsilica easier to handle and transport, it is usually densified. Densification doubles the loose bulk density of microsilica, and reduces the amount of dust created while handling it. Densified microsilica is less costly than the undensified form, and is, therefore, more economically attractive. The slurry form of microsilica contains 50% water. Slurried microsilica has a number of advantages over undensified microsilica. These advantages include lower transportation costs and the ability to be dispensed more accurately. Its main disadvantage is that slurried microsilica must be protected from freezing and from evaporation.

2.3

Properties of Microsilica
Microsilica particles are smooth, spherical, and have on average a particle

diameter between of 3.937 and 7.874 micro inch (0.1 and 0.2 m). This diameter is about a hundred times smaller than that of Portland cement particles. The specific surface area of microsilica is measured by nitrogen adsorption techniques. Most types of microsilica have an average specific surface area of 1.07 105 ft2/lb (22 m2/g), although this parameter can vary between 6.34 104 (13) and 1.37 105 ft2/lb (28 m2/g). Undensified microsilica has an average bulk density of 16 pcf (256 kg/m3), although this can vary between 5 (80) and 27 pcf (432 kg/m3). Densified forms of microsilica have bulk densities between 30 (480.6) and 45 pcf (720.9 kg/m3). The typical range for microsilica specific gravity is 2.2 to 2.3 (Luther and Smith, 1991).

When ordinary Portland cement hydrates, it produces calcium silicate hydrate, the chemical glue that makes the cement hard, as well as lime (calcium hydroxide), which is very weak and dissolves easily. When microsilica is added, the silica combines with the free lime to create more calcium silicate hydrate. The primary active ingredient in microsilica is silicon dioxide. Most types of microsilica have SiO2 contents above 85%. Loss on Ignition (LOI) is used to infer the carbon content; most types of microsilica have LOI values below 3% (Luther and Smith, 1991). Other chemical compounds like Al2O3, SO3, Fe2O3, MnO, TiO2, CaO, K2O, P2O2, MgO, Na2O make up the other percentages in microsilica. Wolsiefer, et al. (1995) tested mixtures of concrete containing sixteen different samples of microsilica. These samples differed with regard to the microsilica form employed, as well as its silicon dioxide content. Their tests involved undensified, densified, slurry, and a pelletized form of microsilica. The investigators found that using different forms of microsilica had no significant effect on mechanical properties of concrete, such as compressive and tensile strength. They also concluded that the shearing forces applied during mixing were sufficient to break up agglomerations of microsilica particles, even in pelletized form.

2.4

Effect of Microsilica on Physical Properties of Concrete

Due to its pozzolanic character, microsilica influences many of the physical

properties of concrete, including water demand, cohesiveness, bleeding, plastic shrinkage, etc., as discussed below.

10

2.4.1

Water Demand Microsilica particles tend to fill the void space between cement particles. This

void space is typically filled with so-called free water. Particle packing allows free water to become available for hydration, thereby decreasing water demand in a microsilica concrete mix. On the other hand, since microsilica particles have a relatively high surface area, more water is adsorbed by the particles, and this tends to increase water demand. Usually, such increases more than offset the water demand decreases stemming from improved particle size distribution, so that the net effect is an increase in water demand. Consequently, high range water reducers and superplasticizers are generally added in microsilica concrete mixes to compensate the water demanding characteristic of microsilica and maintain the required workability (Malhotra, et al., 1987). The net effect of microsilica on the physical properties of concrete depends on a number of factors, including the water-to-cementitious material ratio of the concrete, microsilica content, and the presence of water reducers or superplasticizers. Typically, when its concentrations are kept small, the overall effect of microsilica on water demand is negligible. When the microsilica content is increased, however, the water demand will also increase. Consequently, unless a superplasticizer is used, water must be added to maintain workability (Malhotra, et al., 1987). Jacobsen and Sellevold (1997) tested the frost resistance of microsilica concrete, and noted that one L/m3 of water should be used for every kg/m3 of microsilica added in order to maintain a consistent level of workability.

11

2.4.2

Cohesiveness When microsilica is added to concrete of high cement content or low water-to-

cementitious material ratio, the concrete appears to be more cohesive (i.e., sticky and gluey), making it more difficult to place and to consolidate. High cohesiveness often results in a reduction in surface bleeding, i.e., in the development of a layer of water at the top surface of freshly placed concrete. Bleeding occurs due to excess water in the concrete. Addition of microsilica reduces bleeding in concrete because of the high affinity of condensed microsilica to water, thereby resulting in less water remaining available for bleeding. Bleeding has mostly undesirable effects on concrete properties, e.g., low strength, yet it helps protect against plastic shrinkage cracking, i.e., cracking of the freshly placed concrete due to insufficient surface moisture. Consequently, it is important for contractors to cure the concrete properly (Malhotra, et al., 1987). Moreover, the addition of microsilica increases the concretes viscosity (i.e., its resistance to flow). Consequently, a concrete containing microsilica is less prone to segregation (i.e., the separation of its ingredients) than a mix without it. This effect ensures a more uniform mixture and, therefore, results in higher strength, especially in the presence of a superplasticizer (Malhotra, et al., 1987). 2.4.3 Slump and Workability The addition of microsilica as a concrete admixture has been shown to decrease both the slump and the workability of fresh concrete. In order to maintain a minimum level of workability, a contractor must either add more water or use a water reducing agent. In fact, it is generally considered that a somewhat higher slump must be achieved

12

with microsilica concrete in order to retain the same workability as ordinary Portland cement concrete. Yet, at a low water/cement ratio, and if proper amounts of superplasticizer are used, the workability of microsilica concrete may in fact be higher, on account of the small particles displacing some of the water present in the matrix (Malhotra, et al., 1987). 2.4.4 Air Content It is generally possible to ensure the same amount of air in a mix containing microsilica as in ordinary Portland cement concrete, even though the amount of an air entraining admixture may need to be increased slightly. If a constant dosage of air entrainer is used, the addition of microsilica can reduce the air volume in the mix. To retain constant air content, the demand for an air entraining agent is higher in concrete with microsilica than without it. This higher demand is caused by the high surface area of the microsilica, and possibly by its carbon content (Malhotra, et al., 1987). 2.4.5 Setting Time There is currently no consensus regarding how setting time of concrete is affected by microsilica. Pinto and Hover (1997) conducted an extensive experiment to answer this question. Their study consisted of two series of tests. In the first, the curing temperature was held constant and nine separate high-strength mortar mixtures were tested. In the second, the temperature was varied and six mixtures were examined. The water-to-cementitious material ratio varied from 0.27 to 0.33. Three levels of microsilica content were used: 0, 5, and 10%. Two levels of superplasticizer were also employed:

13

0.8 and 1.6%. The experiments showed that the presence of microsilica accelerated the setting behavior. On the other hand, during their experiments on the freeze-thaw durability of concrete, Soeda, et al. (1999) noticed that the inclusion of microsilica increases setting time by alleviating excessive viscosity stemming from a low water-binder ratio. Investigations by Malhotra and Mehta (1996) have shown that ordinary concrete mixtures incorporating small amounts of microsilica, i.e., up to 10% by weight of cement, exhibit no significant difference in setting time compared to conventional mixes. Since microsilica is typically used in combination with water reducers and superplasticizers, the effects of microsilica on setting time of concrete tend to be masked. For their part, Khayat and Atcin (1992) reported that concretes with microsilica take longer to set and achieve a given strength level than mixes without it. Moreover, the addition of microsilica to concrete without the use of a water reducer or superplasticizer delays setting time, especially when the microsilica content is high. Plasticizers are also known to increase the setting time of concrete, so the combination of microsilica and superplasticizers magnify the retarding effects of microsilica and make it difficult to determine how much of the increased setting time is due to each of the microsilica and the superplasticizer. 2.4.6 Shrinkage and Cracking As concrete containing microsilica shows little to no surface bleeding, the risk of plastic shrinkage is high. This can be a very serious problem under curing conditions of elevated temperature, low humidity, and high wind, all of which contribute to rapid

14

evaporation of water from freshly placed concrete. Crack formation can begin soon after casting, and can continue until the concrete starts to set (Malhotra, et al., 1987). To overcome plastic shrinkage, the surface of concrete should be protected from evaporation by covering it with plastic sheets or wet burlap, or by adding curing compounds and evaporation retarders (Mehta and Monteiro, 1993). Drying shrinkage depends heavily on the amount of microsilica used, the waterto-cementitious material ratio, and the length, as well as method of curing. Generally, the higher the amount of microsilica, the more prone a concrete becomes to drying shrinkage and cracking. Microsilica concretes with a low water-to-cementitious material ratio exhibit similar levels of shrinkage cracking as ordinary mixes. Microsilica concretes moist cured for fewer than seven days exhibit higher shrinkage cracking levels than mixes cured for at least seven days (Malhotra, et al., 1987). Whiting and Detwiler (1998) conducted experiments to assess the effects of microsilica on drying shrinkage and on cracking of bridge decks. Two classes of concrete mixtures were included in their study: full-depth mixtures, containing 620 pcf (9932.4 kg/m3) of cementitious material with a maximum aggregate size of in.; and overlay mixtures containing 700 pcf (11214 kg/m3) of cementitious material with a maximum aggregate size of 3/8 in. (10 mm). For each mix design, separate test mixtures were prepared over a range of water-to-cementitious material ratios and microsilica contents. The microsilica content ranged from 0 to 12% for both mixture classes, whereas the water-to-cementitous material ratio ranged from 0.35 to 0.45, and from 0.30 to 0.40 for full-depth and overlay mixtures, respectively. Specimens from full-depth

15

mixtures were moist cured for seven days before testing, while overlay mixtures were moist cured for only three days. Drying shrinkage was measured on beam specimens after 4, 7, 14, and 28 days, and after 8, 16, 44, and 64 weeks. Cracking tendency was measured according to a restrained-ring method developed under National Cooperative Highway Research Program (NCHRP) Project 12-37 (Krauss and Rogalla, 1996). In the first phase of the investigations, tests were carried out on both full-depth and overlay mixtures. In the second, more comprehensive series of tests were performed on the full-depth mixtures only, during which the more significant variables were examined using more replicates. Chloride ion penetration was measured by exposing the concrete to a chloride solution for 180 days, and measuring the chloride content close to the surface. This information was used to develop chloride diffusivity coefficients for each mixture. Compressive strength was measured on moist-cured cylinders after 7, 28, 56, and 90 days. The researchers stressed the importance of proper moist curing techniques on drying shrinkage. At early ages, inadequate curing, a high water-to-cementitious material ratio, and/or a high cement factor can increase shrinkage cracking by up to 40% in microsilica concretes used for bridge deck placement. Drying shrinkage increases with decreasing water-to-cementitious material ratio, but this effect is more pronounced as microsilica content increases. Specimens cured for only one day and containing between 6 and 9% microsilica showed a high tendency for cracking. In specimens water-cured for seven days, no significant correlation between microsilica content and cracking was observed.

16

2.4.7

Heat of Hydration The addition of microsilica can cause a rise in the mix temperature in the first

couple of days, but will also lead to a significant overall decrease in temperature at later ages. The initial temperature rise can be attributed to the acceleration of the cement hydration reaction (Helland, et al., 1988).

2.5
2.5.1

Effect of Microsilica on Mechanical Properties of Concrete

Compressive Strength Addition of microsilica increases the compressive strength of concrete. The main

reasons for the increase are the strong cement matrix formed due to the micro-filler effect and the pozzolanic characteristics of microsilica. Various factors contribute to this strength increase, including percentage of microsilica added to the mix, water-tocementitious material ratio, cementitious material content, dosage of superplasticizer, temperature, humidity, and length as well as method of curing (Malhotra and Mehta, 1996). Usually, the water demand of microsilica is very high. If slump is to be maintained constant without using a superplasticizer, the water demand is directly proportional to the percentage of microsilica replacement of cement. In this instance, the increase in strength development due to the microsilica is offset by the loss of strength that comes from a high water-to-cementitious material ratio (Malhotra, et al., 1987). Malhotra and Mehta (1996) note that in general, the use of superplasticizers is a prerequisite to achieve the proper contribution of microsilica to concrete compressive

17

strength. In fact, many important applications of microsilica in concrete depend strictly upon its utilization in conjunction with super plasticizing admixtures. The effect of microsilica on compressive strength is small when compared to that of a lower water-tocementitious material ratio. This effect on compressive strength is most pronounced when 0 to 6% of cement is replaced by microsilica, but beyond 6% the strength increases are minimal (Malhotra and Mehta, 1996). The extent of compressive strength gain due to the microsilica addition depends on the age of the concrete, the cement content, and the microsilica content. The age at which microsilica starts to contribute to strength gain depends on the cement content and the water-to-cementitious material ratio of the concrete. Generally, in concretes with high water-to-cementitious material ratios, the microsilica takes a longer time to contribute to compressive strength (Helland, et al., 1988). Prussack, et al. (2001) observed that most of the contribution of microsilica to compressive strength development occurs in the first 7 days under steam curing and in the first 28 days for moist-cured concrete. 2.5.2 Flexural Strength In general, the development of flexural strength in microsilica concrete is similar to that in ordinary Portland cement concrete, but the contribution of drying is more significant in the former. Moist cured concrete specimens show higher flexural strengths than air-dried specimens. With or without microsilica, concretes cured with water will have similar flexural-to-compressive strength ratios. When air cured, concrete containing

18

microsilica will generally have a lower flexural-to-compressive strength ratio (Malhotra, et al., 1987). 2.5.3 Modulus of Elasticity In general, the increase in compressive strength of concrete is accompanied by a small increase in the Youngs modulus of elasticity. The addition of microsilica reduces the porosity of the transition zone between the aggregate and the cement paste, allowing the stiffness of the aggregate to contribute more to the overall stiffness of the concrete (Helland, et al., 1988). 2.5.4 Bond Strength Internal bleeding in concrete can cause free water to accumulate around coarse aggregate and steel reinforcement. A high water content in a concrete mix can also reduce the adhesion between the cement and the aggregates, or between the cement and the reinforcing steel. As noted earlier, microsilica reduces the amount of bleeding, thereby strengthening the bond between concrete and reinforcing steel. This effect is attributable to a less porous transition zone between the cement and the aggregate. As a result, pullout strength increases as the amount of microsilica increases (Malhotra and Mehta, 1996). Fitch and Abdulshafi (1998) tested a total of 62 bridge deck overlay specimens for bond strength. After curing for 28 days, cylinders made with Portland cement concrete were cut in half and one circular face was sandblasted to simulate the surface of a bridge deck. The specimens were then placed at the bottom of a cylinder mold and material from various batches made with microsilica concrete was placed over them. Six

19

different batches of microsilica concrete were prepared: three had natural gravel, whereas the other three contained crushed limestone. Three different moisture conditions, viz., dry, saturated surface dry, and wet, were used for the Portland cement concrete base. The specimens were then subjected to tensile loading to see whether failure occurred at the overlay interface or within the Portland cement or the microsilica concrete matrix itself. All of the specimens with a dry surface condition showed higher overlay bond than concrete matrix strength. A majority of the specimens in the saturated surface dry condition exhibited higher bond than concrete matrix strength. Specimens with a wet surface condition generally had weaker bond than concrete matrix strength. These findings show that microsilica concrete overlays achieve their highest bond strengths when the Portland cement concrete base is dry. 2.5.5 Strength Development Malhotra and Mehta (1996) state that the main contribution of microsilica to concrete strength development at normal room [curing] temperatures takes place between the ages of about 3 and 28 days. The overall strength development patterns in microsilica concretes can vary according to concrete proportions and composition, and are also affected by the curing conditions. High curing temperatures have a greater strength accelerating effect on microsilica concretes than on comparable Portland cement concrete mixes (Malhotra and Mehta, 1996). As noted in previous paragraphs, curing conditions have a significant effect on a number of properties of both fresh and hardened concrete. The inclusion of microsilica as an admixture seems to magnify this effect, especially as far as strength development is

20

concerned. Microsilica concretes that are continuously moist cured show superior mechanical properties than air cured ones (Malhotra, et al., 1987). In fact, Carette and Malhotra (1992) found that air cured concrete containing microsilica may even show a slight decrease in compressive strength after 100 days or more. Typically, seven days of moist curing is sufficient to prevent this. Sasatani, et al. (1995) performed tests for a period of five years, in which they exposed concrete containing various admixtures to different weather conditions. Concrete mixtures containing fly ash, blast-furnace slag, and microsilica were prepared. Water-to-cementitious material ratios of 0.45, 0.55, and 0.65 were used for the control concretes containing only ordinary Portland cement; for concretes containing admixtures this ratio was fixed at 0.55. Replacement levels of ordinary Portland cement by fly ash, blast-furnace slag, and microsilica were 30, 50 and 10%, respectively. Specimens were cured in two different ways. In the first, specimens were cured in 68F (20C)-water for 28 days, whereas in the second they were cured in 68F (20C)-water for 7 days and then kept in a dry environment for another 21 days. After curing, the specimens were exposed to four separate environmental conditions: indoors, in 68F (20C)-water; indoors, at 68F (20C) and 60% relative humidity; on the roof of a building at Kanazawa University located 9.32 miles (15 km) from the sea; and at Matsuto Beach, facing the Sea of Japan. Chloride permeability and pore size distribution were measured after 28 days of environmental exposure. Following exposure times of 1, 3, and 5 years, compressive strength, pulse velocity, depth of carbonation, and chloride ion penetration were also measured.

21

Air-dried 10% microsilica concrete showed a decrease in compressive strength with exposure time over the first twelve months since casting. This may be ascribed to internal stress built up due to uneven drying between the surface and concrete matrix. Uneven drying probably results from the dense and discontinuous pore structure of microsilica concrete. The effects of curing conditions were not as pronounced for concrete exposed to outdoor conditions. This is because the concrete can absorb the water needed for hydration from the environment.

2.6

Field Tests
Fitch and Abdulshafi (1998) conducted brief visual inspections on 145 microsilica

overlaid concrete decks. Of these, 27 decks showed noticeable cracks in the overlay surface. The undersides of 84 of the decks were also visually inspected. Of these, 29 showed several transverse full-depth cracks, and evinced flow of water through the deck resulting in deep chloride penetration. In-depth condition surveys were also performed on 28 microsilica overlaid concrete decks. These surveys included manual sounding to detect horizontal cracking, and core samples to be tested in the laboratory. Approximately 39% of the concrete decks inspected exhibited cracking over more than 10% of their surface. Seven of the 28 decks showed maximum crack widths exceeding 0.008 in. (0.20 mm). Atcin (1990) conducted three field experiments testing various properties of concrete containing microsilica. In the first test, 1765.733 ft3 (50 m3) of microsilica concrete were used for part of the three-lane Highway 25 in Montreal, Quebec, Canada.

22

A reference mix of concrete not containing microsilica was used nearby. The second field experiment consisted of two sidewalks built in Sherbrooke, Quebec. Each sidewalk had sections of microsilica, as well as of ordinary concrete. These sidewalks were chosen because of the high amount of deicing salts they were routinely exposed to. In the third field experiment, two experimental columns were cast during the construction of La Laurentienne, a 26-story building in Montreal. Both columns contained microsilica, but only one column was part of the building. The other was used as a mock column to study the creep of the concrete. Compressive strength and chloride ion permeability tests were performed on both concrete cast during the experiment, and on cores drilled from the field for the first two experiments. In the third, cores were drilled from the mock column two and four years after casting. In the Highway 25 experiment, all the specimens had approximately the same compressive strength after one year, but the reference concrete experienced a greater strength increase between 28 days and one year. The chloride ion permeability was much lower in the microsilica concrete than in the reference concrete. In the second experiment, microsilica specimens cured in the laboratory had higher compressive strengths than those from ordinary concrete, but in the field the situation was reversed. After two years, no significant differences in compressive strength between the two sets of specimens were found. Again, the microsilica concrete had much lower chloride permeability. In the third experiment, the 28-day compressive strengths of the mock column and the lab specimens were nearly identical, while the long-term strength gain in the lab specimens was greater than that in the mock column. The chloride ion

23

permeability was extremely low. The combined results showed that microsilica concrete exposed to outdoor conditions performed just as well as ordinary concrete. Yet, it appeared that microsilica concrete suffered more from poor curing conditions. The chloride-ion permeability was extremely low in microsilica concrete even after four to six years.

2.7

Do Undispersed Agglomerates Matter?

The conventional wisdom on the repercussions of unbroken clusters in densified

microsilica is represented by the study conducted by Wolsiefer, et al. (1995), who tested mixtures of concrete containing sixteen different samples of microsilica. These samples differed with regard to the microsilica form employed as well as silicon dioxide content. Thus, tests involved undensified, densified, slurry, and pelletized form of microsilica. The investigators found that using different forms of microsilica had no significant effect on the mechanical properties of concrete, such as compressive and tensile strengths, etc. They also concluded that the shearing forces applied during mixing are sufficient to break up agglomerations of microsilica particles, even in pelletized form. On the other hand, in a deliberately provocative paper, Diamond and Sahu (2003) set out to redress the perceived failure of many respected sources of information in the industry to properly convey a clear picture of the nature and particulate characteristics of densified [microsilica]. They note with evident frustration, for example, that nowhere in the current [American Concrete Institute] ACI Guide for the Use of Silica Fume in Concrete is there mention of the fact that the actual size of the

24

densified silica fume, as supplied to the customer, is always in the range of hundreds of m...until and unless the fume is physically broken up by some process. Although the tone of this paper is unlike most in the professional literature, being rather aggressive and almost personal, the issues it raises are worth considering seriously. Diamond and Sahu (2003) insist that it is extremely unlikely that complete dispersion ever takes place in concrete mixing analogous to that achieved using an extremely powerful ultrasonic system. They cite a number of transmission and scanning electron microscopy (TEM and SEM) studies, that suggest that clusters were noticed not only in densified microsilica but also in undensified microsilica. This would be an indication that that these clusters are not formed during the process of densification of commercial densified microsilica, but would pose the same level of difficulty to break in the undensified material, as well. The study by Diamond and Sahu (2003) is motivated by two separate concerns stemming from the failure of agglomerates of densified silica fume to be completely dispersed in concrete mixing. The first is that to the degree that [undispersed] agglomerates remain in concrete the expected fine particle packing benefit is lost, since such agglomerates are often much coarser than cement. The second concern is potentially more important: agglomerates can clearly act as extremely aggressive alkalisilica reactive (ASR) aggregates, despite the fact that microsilica is ordinarily considered as mitigating the effects of possible ASR when reactive sand or coarse aggregate is components are present. The implications of the assertions by Diamond and Sahu (2003) are far reaching, but only a few of their facets can be explored within the scope of this study. For

25

example, it is not feasible to conduct TEM or SEM measurements, but more conventional means of establishing the gradation curves for the materials used herein can be employed. The repercussions of agglomerates on fine particle packing can be examined using unit weight determinations of the concrete prepared, as well as strength comparisons. Regrettably, the potential for ASR deterioration cannot be reliably assessed in a study as brief as the present one.

26

MATERIALS AND PROCEDURES

3.1

Introduction
The materials employed in this project are first enumerated in this chapter; they

include primarily the microsilica and the coarse aggregate types. Tests conducted on these materials are outlined next, to illustrate conformity with the prescriptions by the American Society for Testing and Materials (ASTM), by the Ohio Department of Transportation (ODOT), or by the American Association of State Highway and Transportation Officials (AASHTO). The derivation of the mix design for each of the specimen lots cast is then described, emphasizing again adherence to the pertinent governing specifications. The most arduous and time-consuming aspect of the project was mixing, casting and curing the test specimens, the procedures for which are discussed next. Finally, the tests conducted on the plastic and cured concrete samples are described.

3.2

Materials Used
Sand, coarse aggregate, Type I-II Portland cement, water (www.cincinnati-

oh.gov/water; accessed: 07/22/05) from greater Cincinnati water works, microsilica, superplasticizer, and air entrainer are the materials used in this project. The sand and coarse aggregate were supplied free of charge by Martin Marietta Materials, a leading supplier in the Cincinnati area. The sand was natural and came from their sand and

27

gravel facility in Ross, OH. Coarse aggregate was of two kinds, both in the No. 8 gradation with a nominal maximum aggregate size of 3/8 in.: natural river gravel, or crushed limestone. The gravel was obtained from their gravel facility in Fairfield, OH, whereas the stone was produced at their Phillipsburg quarry in Brookville, OH (Jim R. Martin: personal communication, 10/14/02; www.martinmarietta.com; accessed: 12/02/02). The natural aggregate had a rounded shape and a smooth surface, whereas the crushed aggregate had a more angular shape and a rougher surface. At given water/cement ratio, concrete made with crushed aggregate is usually expected to have higher compressive strength than concrete made with natural aggregate, as crushed aggregate creates a stronger bond with cement mortar due to its rough texture (Mehta and Monteiro, 1993). The Portland cement Type I-II was donated by CEMEX from their operation in Fairborn, OH (Steve Reibold: personal communication, 09/11/02; www.cemexusa.com; accessed: 08/14/02), and the microsilica by ELKEM Materials from their location in Alloy, WV (Tony N. Kojundic: personal communication, 08/07/02; www.materials. elkem.com; accessed: 07/24/02). The other admixtures were supplied at no cost by Master Builders, Inc. (Greg Wirthlin: personal communication, 08/07/02; www. masterbuilders.com; accessed: 07/24/02). These were: MB-AE 90 air entrainer, meeting the requirements of ASTM C 260, AASHTO M 154 and CRD-C 13, and recommended for obtaining adequate freeze-thaw durability in a properly proportioned concrete mixture, if standard industry practices are followed; and Rheobuild 1000 high range, water reducing admixture, formulated to produce rheoplastic concrete that flows easily,

28

maintaining high plasticity for time periods longer than conventional superplasticized concrete, and that meets ASTM C 494 requirements for Type A, water reducing, and Type F, high range water reducing, admixtures. The research laboratory facilities on the University of Cincinnati campus were used, except as noted.

3.3

Tests on Microsilica
Testing was performed on four types of microsilica, viz., undensified, densified,

and two types of abused microsilica. Undensified and densified forms are readily available on the market, and were procured commercially. Abused microsilica was prepared by the research team so as to represent a worst case scenario for the handling and storage of commercially available material, particularly with regard to its exposure to ambient moisture. 3.3.1 Preparation of Abused Microsilica Two different abused microsilica samples were prepared. For the first sample of abused microsilica (A), 19.8 lb (9 kg) of densified microsilica was weighed and soaked in a water tub. The microsilica was stirred rigorously at the beginning, and the water level was maintained to at least 3 in. (76 mm) above the surface of the settling solids. After a period of three days, the microsilica had settled and the water above it was clean enough to be removed using small plastic buckets. The wet sample was spread evenly onto plastic lined trays, placed on a flat horizontal surface, and was left to dry at room temperature for nearly three weeks. The second sample of abused microsilica (B)

29

resulted from exposing densified microsilica to high humidity in the curing room for two weeks before allowing it to air-dry. In both cases, any clumps forming during drying were broken from time to time using a small trowel. When completely dry, the microsilica was collected into plastic bags, which were sealed and stored in a cool and dry place, until the time of testing. 3.3.2 Fineness Tests Samples of all types of microsilica used in this project were transported to the ODOT laboratory in Columbus, OH on 05/14/03, for fineness testing by the agencys personnel, Mr. Tim Jones, who is responsible for checking conformance with ASTM C 430 96 Standard Test Method for Fineness of Hydraulic Cement by the 45-m (No. 325) Sieve. The following equipment was used: small dessicator bowl with pouring lip, 2 in. (50.8 mm) dia., 3/4 in. (19.05 mm) tall; twelve No. 325 sieves, 1 in. (38.1 mm) dia., 2 in. (57.15 mm) tall, each with 3 in. (88.9 mm) legs, each with its own calibration factor from the Cement and Concrete Reference Laboratory (CCRL); electronic scale (Mettler Model PC 180), calibrated to 2.2 X 10-6 lb (1/1000 g) precision; small paint brush; hot plate, 250-350F (121-177C) temperature range; city water faucet; pressure stabilizer; and temperature compensated pressure gage to 30 psi (0.21 x 106 N/m2), 0.1 psi (0.001 x 106 N/m2)) precision. The ambient temperature in the laboratory was 78F (26F), and the humidity was 53%. For each type of microsilica, three 0.0022 lb (1.000 g) samples were weighed in the dessicator bowl, and then pour into each sieve with the help of the brush and light tapping of the bowl. First, each sieve was quickly rinsed with de-ionized water from a

30

plastic spray bottle with a narrow tip. Next, each sieve was subjected to a 10 psi (0.069 x 106 N/m2) stream of tap water for a period of one minute (min). A constant stream of tap water was ensured by connecting both the pressure stabilizer and the pressure gage to the end of the faucet, which allowed easy control of the water pressure. The sieve was moved in a circular motion to ensure that the entire sieve was irrigated evenly. The sieve was rotated at approximately one rotation per second (s). For one sample from each type of microsilica, a small paint brush was used in addition to the stream of water. In such cases, the sieve remained stationary, but the brush was moved in a circular motion. Finally, each sieve was rinsed again with de-ionized water, and the bottom of the sieve was blotted dry using a towel. The sieves were then placed on a hot plate set to a temperature between 250 and 350F (121 and 177C). The samples were left here approximately for an hour and a half, or until all of the moisture had evaporated. Once dry, the samples were allowed to cool off for another hour, before the microsilica was removed from the sieves and was weighed in the dessicator bowl again. The weight was recorded as the weight retained, and the sieve correction factor was applied according to the specification. 3.3.3 Gradation Tests The gradation of each of the three microsilica types was also determined using ASTM D 422 63 Standard Test Method for Particle-Size Analysis of Soils (Hydrometer Test), conducted in conjunction with ASTM D 421 85 Standard Practice for Dry Preparation of Soil Samples for Particle-Size Analysis and Determination of Soil Constants, and ASTM D 854 06 Standard Test Methods for Specific Gravity of Soil

31

Solids by Water Pycnometer. In each case, 0.033 lb (15 g) of oven dry microsilica material previously washed through the No. 200 sieve, was mixed with 0.033 Gallons (125 mL) of 4% sodium metaphosphate (NaPO3) solution, in a 0.066 Gallons (250 mL) beaker. The solution, whose trade name is Calgon and whose function is that of a deflocculant or dispersant, had just been prepared by mixing 0.088 lb (40 g) of dry chemical with enough water to make 0.264 Gallons (1000 mL). The beaker was then covered with wet paper towels to minimize evaporation, and the contents were left to stand for an hour. The mixture was subsequently transferred to a dispersion cup, which was filled to about 2/3 of its volume with water, and was stirred for about 2 min. The suspension was then transferred into a sedimentation cylinder, more water was added up to the 0.264 Gallons (1000 mL) mark, and the cylinder was capped and agitated for 1 min. Two minutes after setting the cylinder down, the first reading was taken on the hydrometer, which had been inserted into the cylinder about 20 s before the measurement. Subsequent readings were taken at elapsed times of 4, 8, 16, 30, 60, 120, 240, 480, 960, 1920, 3840, and 5760 min. The temperature was also noted each time, and was used to determine the viscosity of the water in the cylinder. The hydrometer readings were taken to the top of the meniscus, and were later corrected as follows: Rc = Ractual C 0 + CT (3.1)

where: Rc = corrected hydrometer reading; Ractual = actual hydrometer reading; C0 = zero correction for impurities in tap water and for the use of a disperasal agent; CT = temperature correction factor (Bowles, 1992). The hydrometer percent finer, Fh, was then computed using:

32

Fh =

a Rc 100% Ms

(3.2)

in which:
a= 1.65 G s 2.65 (G s 1)

(3.3)

Ms = mass of oven dry material previously washed through the No. 200 sieve used in the test = 0.033 lb (15 g) (from above); and Gs = specific gravity of the solid particles in the sample. This percentage is adjusted in proportion to the material that had been retained on the No. 200 sieve, F200, using the following expression:

Fh adj = Fh F200

(3.4)

For each elapsed time, t, and depending on Rc, the effective sedimentation depth, L, is read from Table 2 in the specification, whereupon the particle diameter, D, can be obtained from Stokes formula, as follows:
D= 30 L 980 (G s G w ) t

(3.5)

with Gw = specific gravity of water in which the sample was suspended (=1.0); and = absolute viscosity of the suspension water, adjusted for temperature per the specification. Prior to the hydrometer tests, a 0.441 lb (200 g) of oven dried sample of each microsilica type was sieved mechanically through a stack of the following sieves: No. 10, 20, 40, 60, 140 and 200, following the procedures detailed by Bowles (1992). Moreover, the specific gravity of the microsilica solids was determined for each type using the procedure of ASTM D 854 06 Standard Test Methods for Specific Gravity of Soil

Solids by Water Pycnometer. ODOT Supplemental Specification 848 Bridge Deck

33

Repair and Overlay with Concrete Using Hydro-Demolition assumes that for microsilica
solids, Gs = 2.2.

3.4
3.4.1

Tests on Aggregates
Specific Gravity and Absorption of Coarse Aggregate ASTM C 127 01 Standard Test Method for Density, Relative Density (Specific

Gravity), and Absorption of Coarse Aggregate was followed in conducting this test.
Coarse aggregate was taken in sufficient quantity and was immersed in water at room temperature. After 24 4 hours of soaking, it was removed from the water and was rolled on a large absorbent towel, until all visible moisture was blotted from the surface of the aggregate. The aggregate was then considered to be in the saturated surface dry (SSD) condition. The test sample was weighed in air, before being placed in a container and immersed in water at 73.4 3F (23 -16 or -19C), where its submerged weight was also determined. The sample was then placed in an oven at 230 9F (110 13 or 23C) for a period of 24 4 hours, after which its oven-dried weight was measured. Three types of specific gravities were determined, viz., bulk specific gravity for the ovendried sample, bulk specific gravity for the saturated surface dry sample, and the apparent specific gravity of the sample. In addition, the absorption value for the aggregate was calculated. The following formulae were used:

Bulk Specific Gravity (Oven Dry), BSG =

A BC B BC

(3.6)

Bulk Specific Gravity ( Saturated Surface Dry), BSG SSD =

(3.7)

34

Apparent Specific Gravity, ASG =

A AC

(3.8)

( B A) Absorption, % Absorption = 100 A where: A = weight of oven dry sample in air, g; B = weight of saturated surface dry sample in air, g; and C = weight of saturated sample in water, g.
3.4.2 Specific Gravity and Absorption of Fine Aggregate

(3.9)

ASTM C 128 97 Standard Test Method for Specific Gravity and Absorption of Fine Aggregate was followed in conducting this test. A 2.2 lb (1000g) sample of sand was immersed in water at room temperature for 24 4 hours. The excess water was drained and the sand was then spread over a flat nonabsorbent surface exposed to a gently blowing current of air. The material was stirred frequently to ensure even drying. This process was continued until the specimen no longer stood on its own when subjected to the specified cone test, at which point the sand was considered to be in the saturated surface dry condition. A pycnometer was filled partially with water and 1.1 lb (500 g) of this saturated surface dry sample was added to it. The remaining volume of the pycnometer was filled with water, up to the calibration mark. The mixture was agitated slightly in order to remove the air bubbles. The total weight of the pycnometer with the sample and water was then recorded. The pycnometer was emptied and the sample was dried in an oven at 230 9F (110 13 or -23C) for 24 4 hours. The sample was taken out of the oven and its weight noted. The pycnometer was then filled only with water up to the calibration mark and its weight was recorded again. The following

35

calculations were performed to determine the three specific gravities and the absorption value for the fine aggregate: Bulk Specific Gravity (Oven Dry), BSG = A B+ S C (3.10)

Bulk Specific Gravity ( Saturated Surface Dry), BSG SSD = Apparent Specific Gravity, ASG = A B+ AC

B B+ S C

(3.11)

(3.12)

( S A) Absorption, % Absorption = 100 A

(3.13)

where: A = weight of oven dry sample in air, g; B = weight of pycnometer filled with water, g; S = weight of saturated surface dry sample, g; and C = weight of pycnometer with water, g.
3.4.3 Bulk Density of Coarse Aggregate

ASTM C 29/C 29M 97 Standard Test Method for Bulk Density (Unit Weight) and Voids in Aggregate was followed in performing this test. A sufficient amount of aggregate was oven-dried for 24 4 hours. The bucket measure was calibrated by filling it with water and determining the weight of water in it. The unit weight of the water was read off Table 3 (Density of Water) in the specification, as a function of its temperature. The volume was determined using the formula: V = Ww

(3.14)

where: Ww = weight of water (lb); and w = unit weight of water, pcf.

36

The effort to be expended in placing the aggregate in the bucket measure was decided based on the size of aggregate. If the nominal maximum aggregate size was 1 in. (38.1 mm) or less, a rodding procedure was used. This involved filling the measure in three layers and rodding each one with a tamping rod 25 times. If the nominal maximum aggregate size was greater than 1 in. (38.1 mm), a jigging procedure was used. This involved filling the measure in three layers, raising one side of the measure approximately 2 in. (50.8 mm), and allowing it to drop to the ground 25 times. After tamping the aggregate, the surface of the measure was leveled so that no aggregates protruded over the top of the measure. The mass of the measure plus contents, and mass of the measure alone were determined. The following formulae were used: Bulk Density (Oven Dry), M = (G T ) V (3.15)

A Bulk Density (Saturated Surface Dry), M SSD = M 1 + 100 (S W ) M Void Content, % Voids = 100 S W

(3.16)

(3.17)

where: MSSD = bulk density of aggregate at SSD condition, pcf; G = mass of measure plus aggregate, lb; T = mass of measure, lb; V = volume of measure, ft3; A = % absorption; S = bulk specific gravity; and W = density of water (=62.4 pcf or 1000 kg/m3).
3.4.4 Sieve Analysis of Fine and Coarse Aggregate

This test was conducted in accordance with ASTM C 136 96a Standard Test Method for Sieve Analysis of Fine and Coarse Aggregates. For the fine aggregates,

37

sieves No. 4, 8, 16, 30, 50, 100, and a pan were weighed and stacked in a column, and between 0.661 and 2.204 lb (300 and 1000 g) of each sample were placed on the No. 4 sieve. The sieves were locked on to a mechanical shaker and the shaker was run for 5 min. The sieves were then released and they were weighed along with the material, and the readings noted. The fineness modulus, FM, of the fine aggregate was calculated using the formula: FM =

Cumulative Percentages Retained

100

(3.18)

The procedure for sieve analysis of coarse aggregates was the same as that for fine, except that five additional sieves were used, viz., 3/8, 3/4, 1/2, 1, and 1 in. sieves. The fineness modulus was not calculated for coarse aggregates.
3.4.5 Moisture Content of Fine and Coarse Aggregates

ASTM C 566 89 Standard Test Method for Total Moisture Content of Aggregate by Drying was followed in this case. The weight of a moisture tin was recorded and an appropriate quantity of aggregate was placed in it. The weight of the moisture tin plus moist sample was also recorded. The specimen was then kept in an oven at a temperature of 230 9F (110 13 or -23C) for 24 4 hours. The material was weighed again after taking it out of the oven. The moisture content, w, was calculated using the formula: W D 100 w= D where: W = weight of moist sample, g; and D = weight of oven-dried sample, g. (3.19)

38

3.5

Mix Design
Concrete mix designs used in this project were prepared in accordance with

pertinent ODOT specifications, viz., ODOT Item 499.03 Concrete-General: Proportioning of the ODOT 2002 Construction and Material Specifications; and ODOT 848.
3.5.1 Constants and Variables

Cement content of 700 lb/yd3 (593.28 g/m3), water/cement ratio of 0.36 and microsilica content of 8% replacement of cement were maintained constant for all mixes. The amount of water was variable, however, since it depends on the physical properties of the aggregate used in any mix. The amount of MB-AE 90 air entrainer for each mix was set at 1.5 oz/100 lb (44.36 mL/100 kg) of cementitious materials, per manufacturer recommendations (1/4 to 4 fl. oz/cwt, or 16 to 260 mL/100 kg), in order to maintain an air content of 8 2%, as specified by the ODOT specification cited. The Rheobuild 1000 plasticizer was taken as 10 fl. oz/cwt (650 mL/100 kg) of cementitious materials for all mixes, conforming to the manufacturers recommendations of 10 to 25 fl. oz/cwt, or 650 to 1000 mL/100 kg . Proper workability did not require the use of another water reducer. Different properties of aggregate influence the water/cement ratio to a great extent. Therefore, in order to maintain a constant water/cement ratio for all mixes, the weights of the coarse aggregate (CA) and fine aggregate (FA) have to be adjusted for each mix, based on their respective properties so as to maintain the specified proportion of coarse to fine aggregate, CA:FA.
3.5.2 Ingredients

39

All determinations below are for the saturated surface dry condition, as indicated by the subscript SSD in the equations presented. ODOT 499 assumes SSD bulk specific gravity values of 2.62 for natural sand and gravel, and 2.65 for crushed limestone; the dry values of 2.20 for microsilica and 3.15 are also used, per ODOT 848 and 499.03. The mix design per cubic yard of concrete was adjusted after determining the actual BSGSSD values for the aggregates. The saturated surface dry (SSD) volumes for CA and FA were retained per the specification, and new SSD weights for CA and FA were backcalculated using the lab BSGSSD values. The mix design was further modified in order to achieve the desired percentage of cement-plus-pozzolan and air percentages % (c + p) and % air, respectively. The weight of microsilica for determinations at SSD condition was first calculated using the formula:

WMS ( SSD ) =

WCEM ( SSD )

% (c + p ) 100 % (c + p ) 1 100

(3.20)

where: WMS(SSD) = weight of microsilica for determinations at SSD condition; and WCEM(SSD) = weight of cement for determinations at SSD condition = 700 lb/yd3 (593.28 g/m3), in accordance with ODOT 848. As noted, both these weights are dry weights. The weight of water and the volume of air were then calculated using the formulae:

WW ( SSD ) = ( w / c) (WCEM + WMS ) V Air = 27 ft 3 %air 100

(3.21) (3.22)

40

where: Ww(SSD) = weight of water for determinations at SSD condition; (w/c) = water/cement ratio = 0.36, per the specification; and Vair = volume of air. The volumes of cement, microsilica, and water were then calculated using for each the formula:
VSSD = WSSD ( BSG SSD ) w

(3.23)

where: VSSD = volume at SSD condition; WSSD = weight at SSD condition; BSGSSD = bulk specific gravity at SSD condition; and w = unit weight of water = 62.4 pcf or 1000 g/m3 (assumed). The sum of the volumes of cement, water, microsilica, and air were then subtracted from a cubic yard (=27 ft3 or 0.729 m3) to obtain the available volume for CA plus FA. The proportions for CA and FA given by the specification were retained, and used to split the available volume between CA and FA. Finally, the SSD weights for CA and FA were backcalculated. The mix design charts are appended at the end of this report.

3.6

Mixing, Casting, and Curing Methods

ASTM C 192/C 192M 00 Standard Practice for Making and Curing Concrete

Test Specimens in the Laboratory, and AASHTO T 126-86 Standard Method of Test for Making and Curing Concrete Test Specimens in the Laboratory were adopted for mixing, casting, and curing the concrete specimens, as described below.

41

3.6.1

Mixing Concrete

The ingredients used for the mixing process were weighed according to the mix design. The concrete was mixed in a mechanical mixer of 5 ft3 (0.135 m3) capacity. The coarse aggregate and about 20% of the mixing water into which the air entrainer and superplasticizer had been dissolved were added in the mixer prior to turning it on. After the mixer was started, the fine aggregate, the cement with the microsilica, and the remaining water were added. After letting the mixer run for 3 min., it was rested for 3 min. It was then restarted, rotated for 2 more min., after which the concrete was poured into a clean pan placed on a smooth rigid surface. In order to cast the required number of specimens, nearly 4 ft3 (0.108 m3) of concrete needed to be mixed. For densified microsilica concrete, 7 to 8 ft3 (0.189 to 0.216 m3) were used, as several large specimens were cast for those mixes. The concrete mixer used by the research team had a yield of just 2.5 ft3 (0.068 m3), and consequently the material had to be mixed in 4 batches. To reduce the amount of concrete required, the investigators subsequently obtained ODOTs permission to cast smaller specimens for undensified and abused microsilica. This made the mixing process more efficient and consistent. Thus, two batches were mixed for concrete made from undensified and abused microsilica, while three to four batches were made for mixes including densified microsilica.
3.6.2 Casting Specimens

Four different kinds of specimens were cast from the concrete mixed, viz., small cylinders (4 8 in. or 101.6 203.2 mm) , large cylinders (6 12 in. or 152.4 304.8

42

mm), large beams (6 6 21 in. or 152.4 152.4 533.4 mm), and small beams (3 4 16 in. or 88.9 114.3 406.4 mm). The cylinders were tested to determine compressive strength, and the beams were used for the calculation of flexural strength. The concrete was placed in two layers in the molds for small cylinders, small beams and large beams, using a scoop, and consolidated using a standard rod. For large cylinders, the concrete was placed in three layers, and was rodded in the same way. Care was taken to avoid segregation and to ensure that the concrete placed in each mold was representative of the entire batch mixed. After the specimens were cast, they were transferred to rigid horizontal surface, and were cured in air overnight. A plastic cover was placed over the specimens to prevent undue moisture loss during this initial 24-hour period. A total of six mixes of concrete were made with three kinds of microsilica (densified, undensified, and abused) and two types of aggregate (natural and crushed). The mixes were named Densified Natural (DN), Densified Crushed (DC), Undensified Natural (UN), Undensified Crushed (UC), Abused Natural (AN), and Abused Crushed (AC). The number of specimens cast in each mix will be reported in Chapter 4. Large size specimens were more numerous than small ones in mixes DN and DC, whereas smaller specimens outnumbered the larger ones in mixes UN, UC, AN, and AC. The main motivation for switching to smaller specimens was the researchers desire to make the mixing and casting process more efficient. This change was implemented following ODOT staff consent. It is noted that ASTM C 192 / C 192 M specifies that the diameter of a cylindrical specimen or minimum cross-sectional dimension of a rectangular section

43

shall be at least three times the nominal maximum size of the coarse aggregate in the concrete. The maximum aggregate size used in this project is 3/8 in. (9.525 mm), so small cylinders (4 8 in. or 101.6 203.2 mm) and small beams (3 4 16 in. or 88.9 114.3 406.4 mm) are also permissible.
3.6.3 Curing

Upon casting, the specimens were immediately covered using a plastic sheet in order to avoid evaporation from the fresh concrete. After 24 hours, the specimens were extruded from the molds. Air pressure was used to demold the cylinders. The specimens were then cured submerged in water tubs in the University of Cincinnati moist room, until the day and time of testing, 7 to 90 days later. Calcium hydroxide (lime) was added to saturation in these tubs, so as to avoid leaching.

3.7

Tests on Plastic Concrete

Immediately after mixing each batch of concrete, slump, air content, and unit

weight determinations were conducted. Concrete slump is measured in inches, and is used as a measure of workability and of consistency from batch to batch. The procedure followed was as prescribed by ASTM C 143/C 143M 00 Standard Test Method for Slump of Hydraulic-Cement Concrete. The mold was first dampened and placed on a flat surface. The concrete was then placed in three layers, each occupying approximately a third of the total volume of the mold. The concrete placed in each layer was rodded 25 times with a standard rod. The excess concrete from the top layer was removed using a rolling action of the rod, and the mold was immediately raised off the specimen. The

44

slump was measured by determining the vertical distance between the top surface of the mold and the displaced original center of the top surface of the concrete specimen. The air content of concrete was measured as stipulated in ASTM C 231 97 Standard Test Method for Air Content of Freshly Mixed Concrete by the Pressure Method. The measuring bowl was placed on a level surface after the inner sides had been dampened. The concrete was then placed in three equal layers, consolidating each layer by 25 strokes of the standard rod. The extra concrete from the top layer was stricken off, the rim of the bowl was thoroughly cleaned, and the bowl was sealed using the covering lid. The air valve located between the air chamber and the measuring bowl was closed. The petcocks were opened and water was injected into one of them, until water came out from the other, which was then closed. Water was pumped into the other petcock, which was closed when the water overflowed out of it. Air was pumped into the air chamber, until the gage hand reached the initial pressure line. After a few seconds, the air valve was opened, and the percentage of air was read on the dial. The unit weight of concrete was calculated by dividing the weight into the corresponding known volume of concrete. The concrete was first weighed in a container of known volume after properly compacting it in three layers following the ASTM standard procedure. The weight of the container was also noted after filling it completely with water. The weight of the water alone was then calculated by subtracting the weight of container from the weight of the container plus water. Dividing the weight of water by its unit weight gives the volume of the container. The unit weight of concrete was then determined by dividing the weight of concrete by the volume of the container.

45

3.8
3.8.1

Strength Test Procedures on Hardened Concrete

Compressive Strength

The procedure followed was as specified by ASTM C 39/C 39M 01 Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens and by AASHTO T 22-03 Compressive Strength of Cylindrical Concrete Specimens. The diameter was measured twice at 90 to each other, and the average diameter was calculated, from which the average cross sectional area was also determined. With the hand finished surface at the top, the cylinder to be tested was placed on the loading table of a 400-kip Tinius Olsen universal testing machine (http://www.tiniusolsen.com/ products/hydraulic/hydraulic-2000kn.html; accessed: 07/21/05). The cylinder was placed such that its axis was aligned exactly with the center of thrust of the spherically seated upper block of the loading head. After adjusting the initial reading to zero, the upper block of the loading machine was brought down so that it just touched the top of the cylinder. Load was applied at a uniform rate of 60,000 25,000 lb/min. (27,215.242 11,339.809 kg/min.) for large cylinders, and 27,000 12,000 lb/min. (12,246.993 5443.108 kg/min.) for small cylinders, until the specimen failed completely. These rates conformed with ASTM C39/C 39 M, which specifies the rate of load application as between 20 and 50 psi/s (0.138 x 106 and 0.345 x 106 N/m2/s). Therefore, the lower limit is 2060d2/4 = 33,929 lb/min. 15,389.935 kg/min.) for large cylinders (diameter, d = 6 in. or 152.4 mm), and 15,080 lb/min. (6840.102 kg/min.) for small cylinders (d = 4 in. or 101.6 mm). Similarly, the upper limit is calculated as 84,823 and 37,699 lb/min.

46

(38475.065 and 17099.978 kg/min.), respectively. The compressive strength, fc, of the concrete specimens was calculated by dividing the maximum load at failure by the average cross sectional area of the cylinder.
3.8.2 Flexural Strength

ASTM C 78 02 Standard Test Method for Flexural Strength of Concrete (Using Simple Beam with Third-Point Loading) and AASHTO T 97 were adhered to. The beams were simply supported on each end, so that the effective span for large beams was 18 in. (457.2 mm) for large beams, and 13.5 in. (342.9 mm) for small beams. The specification requires that the effective span should be three times the depth, which was 6 and 4.5 in. (152.4 and 114.3 mm), respectively. After recording the dimensions of each beam on two opposite sides, the average dimensions were calculated. Two load-applying blocks were placed on the top side of the beam so that they were in contact with the surface of the beam at the third points between the bottom supports. Load was applied at a rate of 1800 300 lb/min. (816.466 136.077 kg/min.) for large beams and 775 125 lb/min. (351.534 56.699 kg/min.) for small beams, until the specimens failed completely, whereupon the failure pattern was recorded. These rates conform to ASTM C 78 02, which requires that the load be applied at a rate that constantly increases the extreme fiber stress between 125 and 175 psi/min. (0.861 x 106 and 1.207 x 106 N/m2/min.) until rupture occurs. To calculate such a rate, the specification provides the following equation: S b d2 r= L (3.24)

47

in which: r = loading rate, lb/min.; S = rate of increase in extreme fiber stress, e.g., 125 or 175 psi/min. (0.861 x 106 or 1.207 x 106 N/m2/min.); b = average beam width, in.; d = average beam depth, in.; and L = effective beam span, in. Thus, from the lower limit of S = 125 psi/min. (0.861 x 106 N/m2/min., r is found to be 125636/18 = 1500 lb/min. (680.388 kg/min.) for large beams, and 1253.520.25/13.5 = 656 lb/min. (297.556 kg/min.) for small beams. The values corresponding to the upper limit of S = 175 psi/min. (1.207 x 106 N/m2/min.) are 2100 and 919 lb/min. (952.543 and 416.851 kg/min.), respectively. The rates adopted are convenient rounded off values based on these calculation results. According to ASTM C 78 02, the modulus of rupture, MR, must be calculated using one of the formulae below, depending on the location of the failure plane: MR = or: MR = 3Pa b d2 (3.26) PL b d2 (3.25)

where: P = ultimate load, lb; and a = average distance between the line of fracture and the nearest support, in. Equation (3.25) is to be used when failure occurs inside the middle third of the span, whereas Equation (3.26) applies when failure occurs outside this region. None of the specimens tested necessitated the latter equation.

48

TEST RESULTS

4.1

Introduction
The main purpose of this chapter is to present the data collected during this

project, including all information recorded, as well as any observations made while performing the experiments. The raw data are tabulated along with derivative values calculated from them. The information provided in this chapter will form the database for the discussions to follow in Chapter 5, concerning the effect such parameters as age, specimen size, microsilica type, and aggregate type.

4.2
4.2.1

Microsilica
Microsilica Fineness

The results of the fineness tests conducted on the four microsilica types (undensified, densified, abused-A and abused-B) are presented in Table 4.1. These tests were performed at the laboratory of the Ohio Department of Transportation (ODOT), and some involved brushing the sample even though this is not prescribed by the American Society for Testing and Materials (ASTM) in ASTM C 430 96 Standard Test Method for Fineness of Hydraulic Cement by the 45-m (No. 325) Sieve.

49

4.2.2

Microsilica Gradation

The calculations for the grain size distributions of undensified, densified and abused microsilica are tabulated in Tables 4.2 through 4.7. The gradation curves plotted in Fig. 4.1 represent the results of both the sieve and hydrometer tests.

4.3

Aggregates
The results of sieve analysis for the fine and coarse aggregate types are tabulated

in Table 4.8(a), which also includes the required percentages passing according to Ohio Department of Transportation (ODOT) Items 703.01 Aggregate-General: Size and 703.02 Aggregate for Portland Cement Concrete: Fine Aggregate. Item 703.01 is identical with the American Association of State Highway and Transportation Officials (AASHTO) M 43. It can be seen that amounts passing are within the prescribed limits. Each aggregate test was conducted three times by independent researchers, from which average values and other statistics were determined. These results are tabulated in Table 4.8(b), in which the coefficients of variation (COV) are also given. It is observed that the COV did not exceed approximately 5% for most of the aggregate properties, except for the absorption values. It can be said, therefore, that the values obtained by the researchers are reliable from a statistical point of view.

4.4

Concrete Mixes
The amounts of each ingredient, used per cubic yard of concrete of each mix, are

given in Table 4.9. The mixes were named after the microsilica type and the nature of 50

the coarse aggregate used, as Densified Natural (DN), Densified Crushed (DC), Undensified Natural (UN), Undensified Crushed (UC), Abused Natural (AN), and Abused Crushed (AC). A number of batches were prepared in order to obtain the required volume of each mix. Table 4.10 enumerates the number of specimens cast from each batch, while Table 4.11 presents the physical properties of each batch, from which average values for each mix were calculated.

4.5

Mechanical Properties
The results of the compressive strength tests are tabulated in Table 4.12 for large

cylinders, and in Table 4.13 for small cylinders. At each age, a minimum of three cylinders were tested. All specimens gained a significant amount of strength by the age of 28 days, but explosive failures (typical of high strength concrete) occurred at almost all ages. Shearing of aggregate was observed in all specimens tested, and the ratio of the frequency of this failure mode to that by aggregate pullout was 4:1 in most cases. Modulus of rupture results can be found in Table 4.14 for large beams, and in Table 4.15 for small beams. All beams failed with a vertical failure plane near the center of the beam, between the two points of load application. In most of the beams, a crack was observed to originate at the bottom of the beam and to propagate up. Examining the specimens after failure, revealed that the failure mode was similar to that under compressive strength testing, i.e., there was much more shearing in the aggregate itself than aggregate pullout, typically in the ratio of 4:1.

51

Table 4.1 Results of Tests Conducted on Microsilica Types at ODOT Laboratory

Microsilica Type Undensified Microsilica Densified Microsilica Abused Microsilica A Abused Microsilica B

Sample No.

1 4 9* 6 7 8* 3 5 10* 2 11 12*

Weight Retained (g) 0.358 0.334 0.023 0.83 0.828 0.767 0.596 0.606 0.506 0.836 0.832 0.766

Correction Factor (%) 32.9 26.77 24.85 32.9 26.77 44.56 84.14 52.59 35.03 30.79 24.85 44.56

Corrected % Retained 47.58 42.34 2.87 110.31 104.97 110.88 109.75 92.47 68.33 109.34 103.88 110.73

Note: *Third replicate of each microsilica type pertains to a brushed specimen.

52

Table 4.2 Hydrometer Test on Undensified Microsilica

Act. Hyd. Rdg, Ra 20 18 17 16 14 13 12 12 11 11 11 11 10 Corr. Hyd. Rdg, Rc 18 16 15 14 12 10.7 10 9.7 8.7 8.7 8.7 8.7 7.7 Hyd. Corr. Only for Meniscus, R 21 19 18 17 15 14 13 13 12 12 12 12 11

Elapsed Time (Min.) 2 4 8 16 30 60 120 240 480 960 1920 3840 5760

Elapsed Time, t (s) 120 240 480 960 1800 3600 7200 14400 28800 57600 115200 230400 345600

Temperature (C) 24 24 24 24 24 24 24 23 23 23 23 23 23

Viscosity (Poise)

Actual % Finer 145.2 129.1 121 112.9 96.8 86.3 80.7 78.2 70.2 70.2 70.2 70.2 62.1

Adjusted % Finer

L (cm)

L/t

D (m)

0.00914 0.00914 0.00914 0.00914 0.00914 0.00936 0.00914 0.00936 0.00936 0.00936 0.00936 0.00936 0.00936

127.5 113.3 106.2 99.2 85 75.8 70.8 68.7 61.6 61.6 61.6 61.6 54.5

12.9 13.2 13.3 13.5 13.8 14 13 13 12 12 12 12 11

0.1075 0.055 0.027708 0.014063 0.007667 0.003889 0.001972 0.000986 0.000497 0.000248 0.000124 0.000062 0.000042

41.241 29.499 20.938 14.916 11.013 7.938 5.586 3.997 2.836 2.006 1.418 1.003 0.824

Calculated Gs = 2.060; a = 1.21; Percent finer than No. 200 sieve (wet sieving) = 87.80%

53

Table 4.3 Hydrometer Test on Densified Microsilica

Act. Hyd. Rdg, Ra 16 13 9 9 8 7 7 7 7 7 6 6 6 Corr. Hyd. Rdg, Rc 14 11 7 7 6 5 5 4.7 4.7 4.7 3.7 3.7 3.7 Hyd. Corr. Only for Meniscus, R 17 14 10 10 9 8 8 8 8 8 7 7 7

Elapsed Time (Min.) 2 4 8 16 30 60 120 240 480 960 1920 3840 5760

Elapsed Time, t (s) 120 240 480 960 1800 3600 7200 14400 28800 57600 115200 230400 345600

Temperature (C) 24 24 24 24 24 24 24 23 23 23 23 23 23

Viscosity (Poise)

Actual % Finer 118 92.7 59 59 50.6 42.2 42.2 39.6 39.6 31.2 31.2 31.2 31.2

Adjusted % Finer

L (cm)

L/t

D (m)

0.00914 0.00914 0.00914 0.00914 0.00914 0.00936 0.00914 0.00936 0.00936 0.00936 0.00936 0.00936 0.00936

29.4 23.1 14.7 14.7 12.6 10.5 10.5 9.9 9.9 9.9 7.8 7.8 7.8

13.5 14 14.7 14.7 14.8 15 15 15 15 15 15.2 15.2 15.2

0.1125 0.05833 0.03063 0.01531 0.00822 0.00417 0.00208 0.00104 0.00052 0.00026 0.00013 0.00007 0.00004

44.103 31.758 23.011 16.271 11.923 8.488 6.002 4.295 3.037 2.147 1.528 1.081 0.882

Calculated Gs = 1.970; a = 1.26; Percent finer than No. 200 sieve (wet sieving) = 24.93%

54

Table 4.4 Hydrometer Test on Abused Microsilica

Act. Hyd. Rdg, Ra 18 15 10 9 9 9 9 9 9 9 8 9 9 Corr. Hyd. Rdg, Rc 16 13 8 7 7 7 7 6.7 6.7 6.7 5.7 6.7 6.7 Hyd. Corr. Only for Meniscus, R 19 16 11 10 10 10 10 10 10 10 9 10 10

Elapsed Time (Min.) 2 4 8 16 30 60 120 240 480 960 1920 3840 5760

Elapsed Time, t (s) 120 240 480 960 1800 3600 7200 14400 28800 57600 115200 230400 345600

Temperature (C) 24 24 24 24 24 23 24 23 23 23 23 23 23

Viscosity (Poise)

Actual % Finer 140.5 114.1 70.2 61.4 61.4 61.4 61.4 58.8 58.8 58.8 50 58.8 58.8

Adjusted % Finer

L (cm)

L/t

D (m)

0.00914 0.00914 0.00914 0.00914 0.00914 0.00936 0.00914 0.00936 0.00936 0.00936 0.00936 0.00936 0.00936

34.2 27.8 17.1 15 15 15 15 14.3 14.3 14.3 12.2 14.3 14.3

13.2 13.7 14.5 14.7 14.7 14.7 14.7 14.7 14.7 14.7 14.8 14.7 14.7

0.11 0.057083 0.030208 0.015313 0.008167 0.004083 0.002042 0.001021 0.00051 0.000255 0.000128 0.000064 0.000043

45.35 32.669 23.765 16.92 12.357 8.737 6.178 4.421 3.126 2.211 1.568 1.105 0.902

Calculated Gs = 1.897; a = 1.32; Percent finer than No. 200 sieve (wet sieving) = 24.34%

55

Table 4.5 Dry Sieve Analysis of Undensified Microsilica Sieve Size (mm) Tare + Material (g) Weight of Material Retained (g) 35 35 15 10 70 20 5 0 0 190 % Cumulative Retained

Sieve Number

Tare (g)

% Retained

% Passing

10 20 40 60 140 200 270 325 Pan SUM

2 0.85 0.425 0.25 0.106 0.075 0.053 0.045

1250 1155 1120 1125 1110 1070 1130 1035 1140

1285 1190 1135 1135 1180 1090 1135 1035 1140

18.421 18.421 7.895 5.263 36.842 10.526 2.632 0 0 100

18.42 36.84 44.74 50.00 86.84 97.37 100.00 100.00 100.00

81.58 63.16 55.26 50.00 13.16 2.63 0.00 0.00 0.00

56

Table 4.6 Dry Sieve Analysis of Densified Microsilica Sieve Size (mm) Tare + Material (g) Weight of Material Retained (g) 0 35 105 45 85 35 10 0 0 315 % Cumulative Retained

Sieve Number

Tare (g)

% Retained

% Passing

10 20 40 60 140 200 270 325 Pan SUM

2 0.85 0.425 0.25 0.106 0.075 0.053 0.045

1250 1155 1120 1125 1110 1070 1130 1035 1140

1250 1190 1225 1170 1195 1105 1140 1035 1140

0 11.111 33.333 14.286 26.984 11.111 3.175 0 0 100

0.00 11.11 44.44 58.73 85.71 96.83 100.00 100.00 100.00

100.00 88.89 55.56 41.27 14.29 3.18 0.00 0.00 0.00

57

Table 4.7 Dry Sieve Analysis of Abused Microsilica Sieve Size (mm) Tare + Material (g) Weight of Material Retained (g) 0 35 90 45 60 25 10 0 0 265 % Cumulative Retained

Sieve Number

Tare (g)

% Retained

% Passing

10 20 40 60 140 200 270 325 Pan SUM

2 0.85 0.425 0.25 0.106 0.075 0.053 0.045

1250 1155 1120 1125 1110 1070 1130 1035 1140

1250 1190 1210 1170 1170 1095 1140 1035 1140

0 13.208 33.962 16.981 22.642 9.434 3.774 0 0 100.001

0.00 13.21 47.17 64.15 86.79 96.23 100.00 100.00 100.00

100.00 86.79 52.83 35.85 13.21 3.77 0.00 0.00 0.00

58

Table 4.8 (a) Aggregate Sieve Analysis Sieve Size or Number % Passing for No. 8 per Specification % Passing for Crushed No. 8 100 100 100 100 86 11 2 1 0 0 0 0 % Passing for Natural No. 8 % Passing % Passing for Natural Average for Sand Per Natural Specification Sand 100 100 100 100 100 100 100 100 100 100 95-100 100 70-100 94 38-80 72 18-60 43 5-30 14 1-10 2 0-5 0

1 in. 1 in. in. in. 3/8 in. No. 4 No. 8 No. 16 No. 30 No. 50 No. 100 Pan

100 100 100 100 85-100 10-30 0-10 0-5 0 0 0 0

100 100 100 100 94 20 1 0 0 0 0 0

Table 4.8 (b) Physical Aggregate Properties Natural Sand COV Average (%) No. 8 Natural COV Average (%) No. 8 Crushed COV Average (%)

Type of Test Apparent Specific Gravity Bulk Specific Gravity (Saturated Surface Dry) Bulk Specific Gravity (Oven Dry) % Absorption Fineness Modulus Dry Unit Weight (pcf)

2.72 2.62

3.7 6.1

2.77 2.61

1.21 3.04

2.8 2.59

1.52 1.45

2.57 3.21 3.11

7.79 50.22 17.94

2.52 3.7 108.21

4.69 52.51 4.8

2.47 4.78 94

1.29 0.67 5.87

59

Table 4.9 Ingredients by Mix (per yd3 of concrete) Mix Ingredient CEMEX Type 1/II cement (lb) Martin Marietta Materials air-dry coarse aggregate (lb) Martin Marietta Materials air-dry fine aggregate (lb) ELKEM microsilica (lb) City of Cincinnati water above air-dry (lb) MB-AE 90 air entrainer (mL) Rheobuild 1000 plasticizer (mL) UN UC DN DC AN AC

700 1339 1344 61 361

700 1319 1355 61 374

700 1339 1344 61 361

700 1319 1355 61 374

700 1339 1344 61 361

700 1319 1355 61 374

337.6 337.6 337.6 337.6 337.6 337.6 2070 2070 2070 2070 2070 2070

60

Table 4.10 Specimens Cast by Batch 2 3 4 Undensified Natural (UN) 2.0 2.0 Volume (ft3) 9 6 Small Cylinders 1 1 Large Cylinders 5 4 Small Beams 0 1 Large Beams Undensified Crushed (UC) 2.0 2.0 Volume (ft3) 10 7 Small Cylinders 1 1 Large Cylinders 5 4 Small Beams 0 1 Large Beams Densified Natural (DN) 3 2.0 2.0 2.0 2.0 Volume (ft ) 4 0 0 1 Small Cylinders 7 0 2 4 Large Cylinders 0 0 0 0 Small Beams 0 4 3 2 Large Beams Densified Crushed (DC) 3 2.5 2.5 2.5 Volume (ft ) 3 3 0 Small Cylinders 3 3 4 Large Cylinders 0 0 0 Small Beams 3 3 3 Large Beams Abused Natural (AN) 3 1.8 1.8 Volume (ft ) 8 9 Small Cylinders 1 2 Large Cylinders 5 4 Small Beams 0 0 Large Beams Abused Crushed (AC) 1.8 1.8 Volume (ft3) 8 9 Small Cylinders 1 2 Large Cylinders 5 4 Small Beams 0 0 Large Beams Note: -Specimens cast in 2004; remainder cast in 2002-2003. Batch No. 1 5 Total

1.0 7 0 0 0 1.0 8 0 0 0 1.0 7 0 0 0 1.0 7 1 0 0 1.0 8 1 0 0 1.0 7 1 0 0

5.0 22 2 9 1 5.0 25 2 9 1 9.0 12 13 0 9 8.5 13 11 0 9 4.6 25 4 9 0 4.6 24 4 9 0

61

Table 4.11 Physical Properties by Batch

Batch # 1 2 3 4 5 Average COV (%)

Undensified Natural (UN) Slump (in.) % Air Unit weight (pcf) Slump (in.) % Air Unit weight (pcf) Slump (in.) % Air Unit weight (pcf) Slump (in.) % Air Unit weight (pcf) Slump (in.) % Air Unit weight (pcf) Slump (in.) % Air Unit weight (pcf) 5.75 7.40 139.10 5.50 7.00 141.40 Undensified Crushed (UC) 5.75 8.70 139.40 5.25 7.10 141.60 Densified Natural (DN) 4.25 7.90 138.00 6.00 9.00 137.70 4.75 8.20 138.60 5.75 7.80 140.50 6.00 9.00 138.40 5.35 8.38 138.62 15.00 6.98 0.80 4.00 7.80 138.60 5.00 7.87 139.86 18.03 10.20 1.11 5.50 7.00 141.00 5.58 7.13 140.51 2.59 3.24 0.89

Densified Crushed (DC) 4.75 7.20 141.40 4.00 8.00 139.40 6.00 7.70 140.90 Abused Natural (AN) 7.50 9.00 137.30 6.50 9.50 135.50 Abused Crushed (AC) 6.50 10.50 134.80 7.00 10.00 134.60 6.50 10.00 135.80 6.67 10.17 135.08 4.33 2.84 0.46 5.50 8.00 138.30 6.5 8.83 137.04 15.38 8.65 1.02 6.00 9.50 138.10 5.19 8.10 139.95 19.02 12.22 1.06

Note: -Specimens cast in 2004; remainder cast in 2002-2003. 62

Table 4.12 Compressive Strength for Large Cylinders, fc

Mix/Age UN P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) UC P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) DN P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) DC P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) AN P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) AC P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV)
2 2 2 2 2 2

7 days C1 189958 27.69 6861

28 days

6861, D1 186277 27.98 6657 6657, A1 122993 28.57 4305 A3 120727 28.57 4226 4333, 2.87% B1 137483 28.87 4762 B2 130945 28.27 4631 4601, 3.88% G2 102490 27.98 3663 3663, I2 108070 28.28 3822 3822, I2 145031 28.87 5024 5024, G1 154065 27.98 5506 5506, B3 124667 28.27 4409 B1 157817 28.27 5582 A4 127698 28.57 4470 A1a 152968 28.27 5410 A1b 152723 27.98 5458 5488, 1.77% B2 162724 28.28 5755 5918, 7.45% B3 181443 28.28 6417 A4 156605 27.98 5597

63

Table 4.12 Compressive Strength for Large Cylinders, fc (Contd)

Mix/Age UN P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) UC P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) DN P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) DC P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) AN P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) AC P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV)
2 2 2 2 2 2

56 days C2 205818 28.87 7129 7129, D2 210985 28.57 7385 7385, A1 175569 28.27 6209 A3 172942 28.27 6117 6156, 0.78% B1 205544 27.98 7346 B2 171138 28.57 5990 6401, 12.81% G2 169969 27.98 6074 6074, I8 175482 28.57 6142 6142, I8 172221 28.28 6090 G9 181544 28.27 6421 B3 169421 28.87 5869 B3 207146 28.58 7249 A4 171845 27.98 6141 A1 182309 28.57 6381 A1

90 days

A4 180865 28.57 6331

A1 183160 28.57 6411

175511 28.57 6143

6316, 1.90% B6 190636 27.98 6813 7031, 4.38%

6421,

6090,

64

Table 4.13 Compressive Strength for Small Cylinders, fc

Mix/Age UN P (lb) Mean A (in. ) fc (psi) Mean fc (psi, COV) UC P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) DN P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) DC P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) AN P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) AC P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV)
2

7 days C1a 68791 12.57 5473 C1b 62941 C2 68063 C1a 85307 12.76 6683 D1a 104065 12.18 8546 A1 71430 12.37 5774 B2 81796 12.18 6717 G1 65132 12.37 5265 I1a 69133 12.57 5501 C1b 89279

28 days C2 90923 C6 92999 C4 86071 12.57 6849 D4 92015 12.37 7438

12.57 12.57 5009 5416 5299, 4.78% D1a D1b D2 70561 68308 71941 12.57 12.57 12.57 5615 5436 5725 5592, 2.61% A1 57494 12.76 4504 4504, B1 65536 12.76 5134 5134, G1A G1B G2 46014 51886 48793 12.37 12.37 12.57 3719 4194 3883 3932, 6.13% I1 I2a I2b 51251 56157 52718 12.57 12.37 12.37 4078 4539 4261 4293, 5.40%

12.96 12.76 12.57 6888 7123 7401 6989, 3.99% D1b D2 D6 95699 98017 89780 12.37 12.57 12.76 7735 7800 7034 7711, 7.22% A4 A5 61596 67457 12.76 12.76 4826 5285 5295, 8.95% B4 B5 74832 69245 12.76 12.57 5863 5510 6030, 10.29% G2a G2b G4 66524 63525 75059 12.18 12.37 12.97 5463 5135 5789 5570, 7.69% I1b I2 I6 65139 78617 65377 12.57 12.57 12.57 5184 6256 5203 5545, 7.84%

G6 75445 12.18 6196 I4 70121 12.57 5580

65

Table 4.13 Compressive Strength for Small Cylinders, fc (Contd)

Mix/Age UN P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) UC P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) DN P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) DC P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) AN P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV) AC P (lb) Mean A (in.2) fc (psi) Mean fc (psi, COV)
C1 99235 12.57 7897 C2a 99923 12.37 8077 56 days C2b 92891 12.97 7165 7489, 6.16% D1a 94369 12.57 7508 D1b 96110 12.57 7646 D2 109109 12.57 8680 7816, 6.26% A3 67126 12.57 5342 A6 72457 12.56 5766 5554, 5.40% B2 93327 12.37 7544 B1h 96488 12.57 7678 B1 78415 12.57 6240 6856, 12.84% G1 75416 12.57 6001 G2a 72367 12.57 5759 G2b 76508 12.57 6088 5929, 2.11% I1 72096 12.37 5828 I2a 79327 12.57 6313 I2b 71501 12.57 5690 5786, 5.97% I1 67273 12.57 5353 I5 72201 12.57 5746 I1a 77802 12.57 6191 I1b 91983 12.76 7206 I2 83566 12.76 6547 G1 76771 12.97 5921 G5 73841 12.57 5876 G1a 77445 12.76 6067 G1b 72096 12.76 5648 G2 80855 12.57 6434 B3 74933 12.57 5963 B1h 92455 12.77 7241 B8 80484 12.37 6506 6873, 7.55% G2b 69912 12.57 5563 6126, 7.14% I3 71992 12.37 5819 6441, 9.17% G7 80913 12.37 6540 G8 82982 12.76 6501 A1h 80545 12.37 6510 A1h 84950 12.76 6655 A7 75330 12.37 6089 6418, 4.58% D3 96845 12.57 7707 D5 94783 12.57 7541 D1 102908 12.37 8318 D2a 89939 12.76 7046 D2b 103200 12.37 8342 C3 89557 12.57 7127 C5 88843 12.37 7181 C1 97717 12.77 7652 C2a 96315 12.57 7665 C2b 95778 12.37 7742 90 days C7 91053 12.37 7360 7604, 2.21% D2c 107421 12.37 8683 7942, 7.77% D2d 102850 12.37 8313 D1 91528 12.37 7398 D7 95652 12.76 7494

66

Table 4.14 Modulus of Rupture for Large Beams, MR

Mix UN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi), COV UC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi), COV DN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi), COV DC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi) AN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi), COV AC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi), COV 7 days 28 days

A2 8068.50 6.00 6.00 672

A3 6446.30 6.00 6.00 537 625, 12.13%

A4 7967.20 6.00 6.00 664

A2a 11086.00 6.00 6.00 924

A2b 8304.00 6.00 6.00 692 747, 21.05%

A3 7567.30 6.06 6.00 624

B1 10874 6.00 6.13 870

B2 7755.5 6.13 6.00 633 700, 21.18%

B3 7158.4 6.00 6.00 597

B1 7952.7 6.06 6.00 656

B2 8623.1 6.06 6.00 704 702, 6.51%

B3 9061 6.06 6.00 747

67

Table 4.14 Modulus of Rupture for Large Beams, MR (contd)

Mix UN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) UC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) DN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) DC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) AN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) AC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) 56 days 90 days

A2 10246 6.00 6.00 854

A3 8746.2 6.00 6.09 707 755, 11.40%

A4 8436.8 5.94 6.03 703

B1 9221.2 6.00 6.06 753

B2 7906.6 6.00 6.16 626 686, 9.29%

B3 8226.9 6.03 6.00 678

68

Table 4.15 Modulus of Rupture for Small Beams, MR

Mix UN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) UC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) DN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) DC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) AN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) AC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) C1a 4077.70 4.50 3.53 981 7 days C1b 3724.80 4.47 3.53 902 942, 4.17% D1a 3733.90 4.59 3.88 731 D1b 3295.10 4.56 3.63 742 733, 1.13% D2 3113.20 4.56 3.56 726 D1 5127.20 4.59 3.75 1071 C2 3794.50 4.44 3.5 942 C1 4333.70 4.59 3.66 953 28 days C2a 4964.30 4.56 3.53 1178 1021, 13.38% D2a 4014.50 4.56 3.66 889 998, 9.66% D2b 4551.80 4.69 3.56 1033 C2b 4427.80 4.56 3.75 932

G1a 3017.30 4.50 3.53 726

G1b 2935.00 4.47 3.50 724 765, 9.03%

G2 3339.40 4.06 3.63 844

G1 3945.60 4.47 3.56 939

G2a 4359.00 4.50 3.56 1030 973, 5.11%

G2b 3988.10 4.63 3.50 950

I1a 3154.80 4.50 3.53 759

I1b 3718.50 4.50 3.5 911 824, 9.49%

I2 3296.10 4.53 3.5 802

I1 3653.30 4.56 3.53 867

I2a 4093.10 4.56 3.56 954 946, 8.01%

I2b 4442.30 4.56 3.69 1018

69

Table 4.15 Modulus of Rupture for Small Beams, MR (Contd)

Mix UN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) UC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) DN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) DC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) AN P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV) AC P (lb) bAVG (in.) dAVG (in.) MR (psi) Mean MR (psi, COV)

C1a 3995.30 4.56 3.66 884 D1a 3852.40 4.56 3.59 883

56 days C1b 4341.90 4.72 3.63 945 902, 4.17% D1b 3753.80 4.50 3.56 887 828, 11.85%

90 days C2 3734.80 4.53 3.56 877 D2 3219.10 4.63 3.63 715

G1a 4366.30 4.63 3.56 1004 I1a 4314.70 4.66 3.56 986

G1b 3742.00 4.53 3.59 863 919, 8.14% I1b 4160.00 4.59 3.63 930 938, 4.68%

G2 3844.20 4.59 3.56 890 I2 3951.00 4.59 3.59 899

70

100.0 90.0 80.0 70.0 Percent Finer 60.0 50.0 40.0 30.0 20.0 10.0 0.0 1.000000 Abused (Hydrometer) Densified (Hydrometer) Undensified (Hydrometer) Densified (Sieve) Abused (Sieve) Undensified (Sieve)

0.100000

0.010000 Particle Diameter, mm

0.001000

0.000100

Fig 4.1 Grain Size Distribution of Undensified, Densified, and Abused Microsilica

71

DISCUSSION OF RESULTS

5.1

Introduction
The interpretation and discussion of the data collected in this project are presented in

this chapter. The main problems encountered in this task are identified first: the questionable validity of results from the American Society for Testing and Materials (ASTM) C 430 96 Standard Test Method for Fineness of Hydraulic Cement by the 45-m (No. 325) Sieve) as performed by Ohio Department of Transportation (ODOT) personnel; and, the significant variability of the results of the mechanical tests conducted at the University of Cincinnati. In order to deal with the latter, an innovative approach to statistical data interpretation is proposed, which can lead to a number of meaningful conclusions by imposing a set of engineering boundary conditions, i.e., that strength increases with age at a decreasing rate.

5.2

Microsilica Fineness
It is anticipated that the smaller the particle size, the lower the percentage of material

retained, and the higher the fineness value. The undensified microsilica, as expected, had the lowest amount of material retained. The abused samples were expected to have higher amounts of material retained, but this was not the case. Abused microsilica A had the next lowest amount, while abused microsilica B and densified microsilica had almost identical amounts of material retained. It was also found that brushing the samples caused more material to pass through the sieves. 72

Looking at Table 4.1, it can be seen that many of the % Retained values exceed 100. Obviously, such values are not reasonable. Thus, for the purposes of this project, any % Retained amounts above 100 will be assumed to be 100%, indicating that none of the material passed through the sieves. The corrected results show that only the undensified microsilica and two of the abused microsilica A specimens were able to pass through the No. 325 sieve to any significant degree. The densified, as well as the abused microsilica B specimens failed to pass through the sieves. According to ASTM C 1240 01 Standard Specification for Use of Silica Fume as a Mineral Admixture in Hydraulic-Cement Concrete, Mortar and Grout, microsilica used in concrete should have a fineness value greater than 90%, i.e., the % Retained should be below 90%. Therefore, it can be concluded that even the undensified microsilica is not suitable for use in concrete, as 90% of the sample could only pass through the sieve when, deviating from the specification, the sample was brushed thoroughly. The results of testing by the water pressure method, as well as by using a brush, do not seem to produce any useful information other than that all of the microsilica samples failed this test. Looking solely at the weight retained, it can be seen in addition that the two abused samples of microsilica do not have the desired particle size. These samples were abused so that they would have a larger particle size than both the undensified and densified samples. The abused sample B had nearly identical results as the densified sample, while abused sample A was actually finer than the densified sample. It is concluded that this test provides no meaningful information for the purpose of assessing the engineering properties of densified (or even of abused) microsilica, and should be dropped as a means of quality assurance. decreasing rate. 73

5.3

Microsilica Gradation
The hydrometer method in conjunction with dry sieve analysis is not the ideal manner

to explore microsilica gradation, but it was the only approach feasible during this limited project. Nonetheless, the data obtained and presented in Fig. 4.1 corroborate some of the concerns expressed by Diamond and Sahu (2003), who point out that the actual size of the densified silica fume, as supplied to the customer, is always in the range of hundreds of m. In contrast, the data obtained in this study create no reasons for concern regarding clusters even in the undensified microsilica itself, as Diamond and Sahu (2003) suggest. Although dry sieving of undensified microsilica may lead to the impression that this type is hardly different from the densified and abused varieties, the superiority of the former is clearly visible in the hydrometer test results. For their part, the hydrometer gradations of the densified and abused microsilica in Fig. 4.1 are practically indistinguishable, indicating that abuse does not create agglomerations that survive the rigor of the hydrometer test methodology. This observation reinforces the conclusion reached above concerning the inadequacy of the microsilica test to distinguish among the microsilica types considered, yet by itself sheds no light on the suitability of using densified microsilica in construction. The latter question needs to be answered through additional testing, as discussed below.

5.4

Variability of Mechanical Tests

The coefficients of variation (COV) of the compressive strength and modulus of

rupture test data are discussed in this section. Even though the researchers had taken great care during mixing, casting and curing procedures to ensure consistency, considerable variation in the test results was observed ascribed mainly to the nature of concrete. From 74

prior experience, the following COV values may be expected: 0-5%: uncommonly low; 510%: excellent engineering work; 10-15%: good engineering work; > 15%: questionable reliability. The uniformity of the mixing processes employed is reflected in the low COV statistics obtained for the physical properties of each batch prepared. According to ODOT Item 499.03 Concrete-General: Proportioning; Slump, the slump may vary between 4 and 8 in. for concrete in which chemical admixtures are used. It can be seen from Table 4.11 that the average slump of all the mixes ranged between these values. Therefore, it can be said that the concrete used in this project meets the ODOT slump specification, even though slump was not a property that was controlled in itself. Similarly, the expected air content was 8 2%, according to ODOT Supplemental Specification 848 Bridge Deck Repair and Overlay with Concrete Using Hydro-Demolition. Consequently, the average air content values in Table 4.11 also met the ODOT specification, falling within the range specified, except for a small deviation in a single batch.
5.4.1 Compressive Strength

The coefficients of variation for various mixes at ages of 7, 28, 56 and 90 days are tabulated in Table 5.1. As noted in Chapter 4, a two letter designation is used to identify the mixes, in terms of the microsilica and coarse aggregate types used, viz., Densified Natural (DN), Densified Crushed (DC), Undensified Natural (UN), Undensified Crushed (UC), Abused Natural (AN), and Abused Crushed (AC). Moreover, compressive test results are designated as pertaining to large or to small cylinders, LC or SC, respectively. Low values were obtained in the compressive strength, fc, tests at almost all ages. From Table 5.1, it can be seen that the average coefficients of variation for the compressive strength tests varied mostly between 2 and 10% for all mixes, representing excellent engineering work. The 75

actual values of fc obtained can be found in Tables 4.12 and 4.13 for large and small cylinders, respectively.
5.4.2 Modulus of Rupture

Coefficients of variation obtained for the modulus of rupture, MR, tests are also tabulated in Table 5.1, in which they are designated as pertaining to large or to small beams, LB or SB, respectively. It can be seen that COV values were higher for modulus of rupture than for compressive strength testing. This is because beams are sensitive to even minor changes in mixing, casting or testing methods. Similar variability has been reported in the literature. The average coefficients of variation for MR varied between 7 and 15%, which represent excellent to good engineering work. The actual values of MR obtained can be found in Tables 4.14 and 4.15 for large and small beams, respectively.

5.5

Data Interpretation
Despite the researchers efforts to ensure uniformity and consistency in the results

obtained, which are reflected in COV values representing excellent to good engineering work, the data collected pose a difficult interpretation problem since the small number of specimens tested does not permit an exclusively statistical analysis. Consequently, it has been found useful to appeal in addition to a set of engineering boundary conditions to ensure that strength increases with age at a decreasing rate. This approach has been found to reinforce the statistical analyses performed, leading to a number of meaningful conclusions, as described below.

76

5.5.1

Compressive Strength

The average compressive strength of the various mixes at each age is tabulated in Tables 5.2 and 5.3, pertaining to large and small cylinder specimens, respectively. From these data, relative strength values at each age with respect to the corresponding 28-day average strength were calculated, as found in Tables 5.4 and 5.5. For example, in Table 5.5, the 7-day strength for the small cylinder mix with densified microsilica and natural aggregate (DN) is 85% of its 28-day strength, while the 56-day strength for the same mix is 105% of the corresponding 28-day strength. These values are calculated by dividing the strength at a particular age by the strength at 28 days; e.g., from Table 5.3, the compressive strength for small cylinders at 7 days for mix DN is 4504 psi (31.05 x 106 N/m2); similarly, the strength at 28 days for the same mix is 5295 psi (36.51 x 106 N/m2). Therefore, the relative strength at 7 days with respect to the 28-day strength is calculated as follows: (4504 / 5295) 100 = 85%. All relative strengths thus calculated were plotted against age, as shown in Figures 5.1 and 5.2, from the data for large and small cylinders, respectively. In each case, the researchers then fitted a trend curve through the points, by adjusting slightly the line predicted statistically, so as to conform to the boundary conditions that the slope should decrease and the strength should increase with increasing age. It can be noted from Figures 5.1 and 5.2 that the trend curve is only slightly different from the unadjusted statistical lines. The enforcement of boundary conditions results in a smooth curve that reproduces the expected trends. From the trend line thus obtained, adjusted relative strength values at each age were read. For example, in Fig. 5.1, the trend line indicates 75% strength at 7 days, and 98%, 112% and 117% at 28, 56 and 90 days of age, respectively. Thus, the strength gain

77

between two ages can be determined; e.g., the strength gain between ages 28 and 90 from Fig. 5.1 can be calculated as (117-98)/98 = 19%. In this manner, a set of best-fit compressive strength values were obtained for each mix, as presented in Tables 5.6 and 5.7, for large and small cylinder specimens, respectively, and will be used below in the interpretation of the test results. The entire procedure leading to the derivation of these best-fit strength values may be summarized as follows, for a particular specimen size: 1. 2. 3. Obtain the laboratory data pertaining to each mix at every age. Calculate the average strength value of each mix at every age. Normalize these average values with respect to the corresponding 28-day strength, to

obtain the relative strengths for each mix at every age. 4. Plot the relative strengths for all mixes against age, and obtain a statistically predicted

line through these points. 5. Impose the engineering boundary conditions of strength increase at a decreasing rate

with time, to adjust the statistical line and thus obtain a smooth trend curve. 6. From the adjusted trend curve, read off the percentage strength gain with each age

increment. 7. Return to the laboratory data, and inspect them to identify the optimum age for each

mix, to be used in the best-fit strength derivation. One way to do this, is to select the age, for which the coefficient of variation is the lowest. 8. Use the incremental strength gains from Step 6 in conjunction with the laboratory

strength at the optimum age for each mix from Step 7, in order to derive the best-fit strength values for each mix at every age. For example, if laboratory 90-day compressive strength of 78

mix DC is equal to 6873 psi (47.39 x 106 N/m2) (Table 5.3) and the lowest coefficient of variation occurs at this point as shown in Table 5.1, then the best-fit 7-day strength is 6873 multiplied by strength gain from 7 days to 90 days i.e., 6873 (72/110) = 4499. Note that 72% and 110% are obtained from the trend line from Fig. 5.2 at 7 days and 90 days, respectively.
5.5.2 Modulus of Rupture

The procedure followed in obtaining the best-fit modulus of rupture values is analogous to that for compressive strength. Because of the small number of large and small specimens tested, the analysis considered the two types of specimens collectively. The average values of actual test data obtained for each mix at each age tested are presented in Table 5.8, from which the corresponding values relative to the corresponding 28-day strength are obtained, as shown in Table 5.9. The trend line for the modulus of rupture can be seen in Fig. 5.3, from which the best-fit values in Table 5.10 are determined.

5.6
5.6.1

Effect of Microsilica Type on Mechanical Properties of Concrete

Compressive Strength

Natural Coarse Aggregate Mixes DN, UN, and AN were made using natural aggregate, each with one of the three types of microsilica, viz., densified, undensified and abused, respectively. The compressive strength comparisons based on the type of microsilica are plotted in Figures 5.4 and 5.5 for large and small cylinders, respectively. It can be observed that the compressive strength of undensified microsilica was greater than the other two microsilica concretes at all ages, for both large and small cylinders. At 28 days, large cylinders made with undensified 79

microsilica concrete had nearly 576 psi (3.971 x 106 N/m2) (10%) and 923 (6.364 x 106 N/m2) (17%) psi higher compressive strength than densified and abused microsilica concretes, respectively. In the compressive strength results of small cylinders, the strength of undensified microsilica concrete was 1025 psi (7.065 x 106 N/m2) (18.5%) and 1137 psi (7.839 x 106 N/m2) (15%) psi greater than densified and abused microsilica concretes, respectively. The compressive strength of densified microsilica exceeded 5000 psi (34.474 x 106 N/m2) at 28 days for both large and small specimens. In comparison, for mixes typically used in pavement construction, ODOT 499.03 requires an average compressive strength at 28 days of 4000 psi (27.579 x 106 N/m2) for Class C, 3000 psi (20.684 x 106 N/m2) for Class F and 4500 psi (31.026 x 106 N/m2) for Class S. Therefore, even though densified microsilica had lower compressive strengths than the undensified material, it can still be used on Ohio Department of Transportation (ODOT) projects, since it has adequate strength to meet the requirements of this agency. In fact, even the abused microsilica results appear adequate. Crushed Coarse Aggregate The effect of aggregate on concrete made with crushed aggregate and undensified microsilica can be found in Figures 5.6 and 5.7, respectively, for large and small cylinders. As was the case for concrete made with natural aggregate, the compressive strength of concrete made with undensified microsilica and crushed aggregate, showed higher strengths than densified and abused microsilica concretes. The abused microsilica exhibited in all cases lower strengths compared to the other two types, as expected.

80

At 28 days, the compressive strength of large cylinders made with undensified microsilica had 544 (3.751 x 106 N/m2) (9%) and 1088 (7.501 x 106 N/m2) (20%) psi higher strength than densified and abused microsilica concretes, respectively. Similarly, the small cylinders made with undensified microsilica had 1442 (9.942 x 106 N/m2) (24%) and 1714 (11.818 x 106 N/m2) (30%) psi greater strength than densified and abused microsilica concretes, respectively. Once again, the densified microsilica with crushed aggregate showed higher strengths than those prescribed by ODOT 499.03 at 28 days of curing in all cases. Therefore, densified microsilica meets the ODOT specifications in all cases and can be used in the construction of pavements and bridges.
5.6.2 Modulus of Rupture

Natural Coarse Aggregate The effect of microsilica type on the modulus of rupture of concrete made with natural aggregate can be found in Fig. 5.8. The beams made with undensified microsilica showed somewhat higher strengths compared to densified and abused microsilica concretes. Surprisingly, abused microsilica concrete had higher modulus of rupture than densified microsilica concrete and was comparable to undensified microsilica concrete. At 28 days, for example, the beams made with undensified microsilica had 22 psi (0.152 x 106 N/m2) (3%) and 73 psi (0.503 x 106 N/m2) (11%) higher strengths than densified and abused microsilica concretes, respectively. Crushed Coarse aggregate Figure 5.9 shows the effect of microsilica type on concrete with crushed aggregate. The results for modulus of rupture of concrete with crushed aggregate also showed some 81

surprising trends. Densified microsilica concrete exhibited lower values than undensified and abused microsilica concretes at all ages. Comparing Figures 5.8 and 5.9, it is observed that at 28 days concrete made with crushed aggregate showed slightly higher modulus of rupture values than concrete with natural aggregate, as expected.

5.7

Effect of Coarse Aggregate Type on Mechanical Properties of Concrete

5.7.1

Compressive Strength

Undensified Microsilica As expected, the compressive strength of large cylinders made with crushed aggregate showed higher strengths than the concrete made with natural aggregate for large and small cylinders at all ages and for all mixes. The effect of aggregate type on the compressive strength of concrete made with undensified microsilica can be found in Figures 5.10 and 5.11, for large and small cylinders, respectively. Concrete made with undensified microsilica and crushed aggregate showed higher strengths than the concrete made with same microsilica type and natural aggregate. As expected, at 28 days of age, the compressive strength of large cylinders from mix UC had 224 (1.544 x 106 N/m2) (4%) psi higher strength than large cylinders from mix UN. When the compressive strengths of small cylinders made with undensified microsilica are observed at 28 days, the concrete made with crushed aggregate shows an increase of 810 (5.585 x 106 N/m2) (12%) psi over the concrete made with natural aggregate. Densified Microsilica

82

Similarly, large and small cylinders made with densified microsilica and crushed aggregate showed greater strengths at all ages than the concrete made from the same microsilica and natural aggregate. The effect of aggregate type on compressive strength of large and small cylinders made with densified microsilica is plotted in Fig. 5.12 and Fig. 5.13. Moreover, the difference in strengths due to aggregate type is higher for small than for large cylinders. This trend was also observed for undensified microsilica concrete. At 28 days, large cylinders made with crushed aggregate had 256 (1.765 x 106 N/m2) (5%) psi higher strength than natural aggregate concrete, whereas for small cylinders, the difference in the strengths of crushed aggregate concrete and natural aggregate concrete was 393 (2.710 x 106 N/m2) (7%) psi. Abused Microsilica Abused microsilica concrete showed the same trends as undensified and densified microsilica concretes. Figures 5.14 and 5.15 show the effect of aggregate type on the compressive strength of abused microsilica concrete for large and small cylinders, respectively. Concrete made with crushed aggregate had higher strengths than the concrete made with natural aggregate at all ages, as expected. At 28 days, large cylinders made with crushed aggregate had 59 (0.407 x 106 N/m2) (1%) psi higher strength than the ones made with natural aggregate. In contrast, for small cylinders, the difference in strengths of crushed and natural aggregate was 300 (2.068 x 106 N/m2) (6%) psi.
5.7.2 Modulus of Rupture

The modulus of rupture values showed similar trends as compressive strength results, values being higher when crushed aggregate was used. The effects of aggregate type on the three types of microsilica are explained in the following sections. 83

Undensified Microsilica The effect of aggregate type on undensified microsilica concrete beams is plotted in Fig. 5.16. At 28 days, the beams made with crushed aggregate had 13 (0.090 x 106 N/m2) (2 %) psi higher flexural strength than beams made with natural aggregate. Densified Microsilica Rather surprisingly, beams made from densified microsilica concrete with natural aggregate showed greater strength than those made with crushed aggregate at all ages. Figure 5.17 shows the effect of aggregate type on densified microsilica concrete beams. It can be seen that, at 28 days, beams made with natural aggregate had nearly 10 (0.069 x 106 N/m2) (1%) psi higher flexural strength than beams made with crushed aggregate. Abused Microsilica The effect of aggregate type on abused microsilica concrete beams can be seen in Fig. 5.18. at 28 days, the modulus of rupture of beams with natural aggregate was surprisingly somewhat higher when compared to beams with crushed aggregate. At other ages, however, the beams with crushed aggregate exhibited greater values than the ones with natural aggregate, as expected. In all cases, the variability in the values calculated far exceeds any intrinsic differences due to aggregate type.

5.8

Effect of Specimen Size on Mechanical Properties of Concrete

Compressive Strength Size factors pertaining to cylindrical specimens for mixes at various ages are tabulated in Table 5.11. A size factor is calculated as the ratio of the compressive strength of 84

small cylinders to that of large cylinders, expressed as a percentage. Small cylinders, usually have 5 to 8% higher strength than large cylinders (Mehta and Monteiro, 1993); therefore, size factors are expected to be greater than 100%. From Table 5.11, it can be noticed that the size factors obtained in this project do not follow any particular pattern. It can be seen that in for mixes DN and AD, the compressive strength of the large cylinders was either higher than or equal to that of the small cylinders at all ages. Modulus of Rupture Size factors for modulus of rupture results could not be calculated as either large beams or small beams were tested at each age for all the mixes. Instead, a factor of 1.16 was assumed based on past experience to convert the large beam values for mixes AN and DC to the corresponding small beams values, so as to allow comparisons among aggregate and microsilica types. Such comparisons suggest that the choice of the factor of 1.16 was appropriate.

85

Table 5.1 Coefficients of Variation, COV (%), in Laboratory Test Results Age (days) Test fc (LC) fc (SC) MR (LB) MR (SB) fc (LC) fc (SC) MR (LB) MR (SB) fc (LC) fc (SC) MR (LB) MR (SB) fc (LC) fc (SC) MR (LB) MR (SB) fc (LC) fc (SC) MR (LB) MR (SB) fc (LC) fc (SC) MR (LB) MR (SB) 7 28 56 90 Average

4.78 4.17 2.61 1.13 2.87 12.13 3.88 21.18 6.13 9.03 5.40 9.49

Mix: Undensified Natural (UN) 3.99 6.13 13.38 4.17 Mix: Undensified Crushed (UC) 7.22 6.26 9.66 11.85 Mix: Densified Natural (DN) 1.77 0.78 8.95 5.40 21.05 11.40 Mix: Densified Crushed (DC) 7.45 12.81 10.29 12.84 6.51 9.29 Mix: Abused Natural (AN) 7.69 2.11 5.11 8.14 Mix: Abused Crushed (AC) 7.84 5.97 8.01 4.68

2.21 7.77 1.90 4.58 4.38 7.57 7.14 9.17

4.28 7.24 5.97 7.55 1.83 6.31 14.86 7.13 10.23 12.33 5.77 7.43 7.10 7.39

86

Table 5.2 Average Compressive Strength for Large Cylinders (psi) Age (days) 7 28 56 90 UN UC DN 4333 5488 6156 6316 DC AN AC

6861 7129

6657 7385

4601 5918 6401 7031

3663 5506 6074 6421

3822 5024 6142 6090

Note: The numbers in bold were used to calculate the best-fit values for each mix.

Table 5.3 Average Compressive Strength for Small Cylinders (psi) Age (days) 7 28 56 90 UN UC 5592 7711 7816 7942 DN DC AN AC 4293 5545 5786 6441

5299 6989 7489 7604

4504 5295 5554 6418

5134 6030 6856 6873

3932 5570 5929 6126

Note: The numbers in bold were used to calculate the best-fit values for each mix.

87

Table 5.4 Relative Compressive Strength Values for Large Cylinders (%) Age (days) 7 28 56 90 Mix UN 100 104 UC 100 111 DN 79 100 112 115 DC 78 100 108 119 AN 67 100 110 117 AC 76 100 122 121

Table 5.5 Relative Compressive Strength Values for Small Cylinders (%) Age (days) 7 28 56 90 Mix UN 76 100 107 109 UC 73 100 101 103 DN 85 100 105 121 DC 85 100 114 114 AN 71 100 106 110 AC 77 100 104 116

88

Table 5.6 Best-Fit Compressive Strength Values for Large Cylinders (psi) Age (days) 7 28 56 90 UN UC DN 4333 5662 6471 6760 DC AN AC

4774 6238 7129 7448

4945 6462 7385 7714

4529 5918 6763 7065

4068 5315 6074 6346

4113 5374 6142 6416

Note: The numbers in bold were used to calculate the best-fit values for each mix.

Table 5.7 Best-Fit Compressive Strength Values for Small Cylinders (psi) Age (days) 7 28 56 90 UN UC 5592 7378 8155 8543 DN DC AN AC 4293 5664 6261 6559

4977 6568 7259 7604

4201 5543 6127 6418

4499 5936 6561 6873

4066 5364 5929 6212

Note: The numbers in bold were used to calculate the best-fit values for each mix.

89

Table 5.8 Average Modulus of Rupture Values for Beams (psi) Age (days) 7 28 56 90 UN UC 733 988 828 DN DC AN AC

942 1021 902

625 747 755

700 702 686

765 973 919

824 946 938

Table 5.9 Relative Modulus of Rupture Values for Beams (%) Age (days) 7 28 56 90 UN UC DN DC AN AC

92 100 88

73 100 83

84 100 101

100 100 98

79 100 94

87 100 99

Table 5.10 Best-Fit Modulus of Rupture for Small Beams (psi) Age (days) 7 28 56 90 UN UC 733 864 916 DN DC AN AC

722 851 902

700 825 875

691 815 864

735 867 919

751 885 938

Note: Large beams (6 6 21 in.) have been tested for Mixes DN and DC, whereas small beams (3 4 16 in.) have been tested for all other mixes. For mixes DN and DC, large beam lab data were converted to the small beam values in this Table through multiplication by a factor of 1.16, selected on the basis of past experience. The numbers in bold were used to calculate the best-fit values.

90

Table 5.11 Cylinder Size Factors (%)

Age (days) 7 28 56 90 Average by mix Average by MS type

UN

UC

DN

DC

AN

AC

104 105 102 102 104 108

113 114 110 111 112

97 98 95 95 96 97

99 100 97 97 98

100 101 98 98 99 101

104 105 102 102 103

91

130 120 110 Relative Strength, % 100 90 80 70 60 50 40 0 10 20 30 40 50 Age, days 60 70 80 90 100 Test Data Trend Curve Statistical Line

Fig. 5.1 Trend Line Curves for Large Cylinders

92

130 120 110 Relative Strength, % 100 90 80 70 60 50 40 0 10 20 30 40 50 Age, days 60 70 80 90 100 Test Data Trend Curve Statistical Line

Fig. 5.2 Trend Line Curves for Small Cylinders

93

130 120 110 Relative Strength, % 100 90 80 70 60 50 40 0 10 20 30 40 50 Age, days 60 70 80 90 100 Test Data Trend Curve Statistical Line

Fig. 5.3 Trend Line Curves for Small Beams

94

9000 8000 7000 Compressive Strength, psi 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Undensified Densified Abused

Fig. 5.4 Effect of Microsilica Type on Large Cylinders with Natural Aggregate

95

9000 8000 7000 Compressive Strength, psi 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Undensified Densified Abused

Fig. 5.5 Effect of Microsilica Type on Small Cylinders with Natural Aggregate

96

9000 8000 7000 Compressive Strength, psi 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Undensified Densified Abused

Fig. 5.6 Effect of Microsilica Type on Large Cylinders with Crushed Aggregate

97

9000 8000 7000 Compressive Strength, psi 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Undensified Densified Abused

Fig. 5.7 Effect of Microsilica Type on Small Cylinders with Crushed Aggregate

98

1000 900 800 Modulus of Rupture, psi 700 600 500 400 300 200 100 0 7 28 Age, days 56 90 Undensified Densified Abused

Fig. 5.8 Effect of Microsilica Type on Small Beams with Natural Aggregate

99

1000 900 800 Modulus of Rupture, psi 700 600 500 400 300 200 100 0 7 28 Age, days 56 90 Undensified Densified Abused

Fig. 5.9 Effect of Microsilica Type on Small Beams with Crushed Aggregate

100

9000 8000 Compressive Strength, psi 7000 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.10 Effect of Aggregate Type on Large Cylinders with Undensified Microsilica

101

9000 8000 Compressive Strength, psi 7000 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.11 Effect of Aggregate Type on Small Cylinders with Undensified Microsilica

102

9000 8000 7000 Compressive Strength, psi 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.12 Effect of Aggregate Type on Large Cylinders with Densified Microsilica

103

9000 8000 7000 Compressive Strength, psi 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.13 Effect of Aggregate Type on Small Cylinders with Densified Microsilica

104

9000 8000 Compressive Strength, psi 7000 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.14 Effect of Aggregate Type on Large Cylinders with Abused Microsilica

105

9000 8000 Compressive Strength, psi 7000 6000 5000 4000 3000 2000 1000 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.15 Effect of Aggregate Type on Small Cylinders with Abused Microsilica

106

1000 900 800 Modulus of Rupture, psi 700 600 500 400 300 200 100 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.16 Effect of Aggregate Type on Small Beams with Undensified Microsilica

107

1000 900 800 Modulus of Rupture, psi 700 600 500 400 300 200 100 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.17 Effect of Aggregate Type on Small Beams with Densified Microsilica

108

1000 900 800 Modulus of Rupture, psi 700 600 500 400 300 200 100 0 7 28 Age, days 56 90 Natural Crushed

Fig. 5.18 Effect of Aggregate Type on Small Beams with Abused Microsilica

109

CONCLUSIONS AND RECOMMENDATIONS

6.1

Summary
This project explored the use of densified microsilica in concrete used in the

construction of pavements and structures by Ohio Department of Transportation (ODOT). Of the various kinds of microsilica that are commercially available in the market, only two were considered by the research team: undensified microsilica and densified microsilica. A third type of microsilica, viz., abused microsilica, was also explored. It is usually assumed that undensified microsilica will result in higher strengths than the densified material, which is the form most commonly used in practice. Wishing to compare densified microsilicas performance to a worst case scenario, the investigators prepared a quantity of abused microsilica by soaking the densified material in water and drying it, thereby encouraging the formation of clumps. Microsilica thus prepared was expected to in the weakest concrete, assuming that the consequence of densification is lower strength. Comparing the engineering performance of densified microsilica to undensified and abused materials brackets the range of situations that may be encountered in the field. Undensified microsilica is expected to pass the requirement that it should have a fineness value greater than 90% found in American Standard for Testing and Materials (ASTM) C 1240 01 Standard Specification for Use of Silica Fume as a Mineral Admixture in Hydraulic-Cement Concrete, Mortar and Grout when it is wet-sieved in accordance with ASTM C 430 96 Standard Test Method for Fineness of Hydraulic Cement by the 45-m (No. 325) Sieve, whereas densified and abused microsilica may not pass the sieve test. 110

Abused microsilica was prepared by the research team in University of Cincinnati concrete laboratory, while densified and undensified microsilica were obtained free of charge from ELKEM chemicals. Cement, concrete admixtures, and aggregate were provided free of charge by CEMEX, Master Builders, Inc., and Martin Marietta Materials, respectively. The research team utilized the research facilities at University of Cincinnati for performing the various tasks in this project. After procuring all the materials needed for the project, the research team conducted tests on the aggregates to determine their properties, and to formulate a mix design for the concrete. Two kinds of coarse aggregate, viz., natural and crushed, with a single gradation of No. 8 were used. ODOT Supplemental Specification 848 Bridge Deck Repair and Overlay with Concrete Using Hydro-Demolition was used in preparing the mix design, in conjunction with ODOT Item 499.03 Concrete-General: Proportioning. Six concrete mixes were made with the three types of microsilica and the two kinds of coarse aggregates. Concrete specimens were tested to determine their compressive strength and flexural strength. Cylinders and beams were cast for this purpose. Specimens were made in two different sizes in each case, viz., small cylinders (4 8 in. or 101.6 203.2 mm) , large cylinders (6 12 in. or 152.4 304.8 mm), small beams (3 4 16 in. or 88.9 114.3 406.4 mm), and large beams (6 6 21 in. or 152.4 152.4 533.4 mm). Cylinders were tested to calculate compressive strength, whereas beams were used to determine flexural strength. Three cylinders and three beams were tested at each age in most cases. Large-sized specimens were used for mixes made from densified microsilica, whereas small specimens were used for the other mixes. The decision to change the specimen size from large to small was taken by the investigators as it helped achieve higher efficiency and consistency, without 111

deviating from the ASTM specification that requires the minimum specimen dimension to exceed three times the nominal maximum size of the coarse aggregate (ASTM C 192/C 192M 00 Standard Practice for Making and Curing Concrete Test Specimens in the Laboratory). The same mixing, casting and testing procedures were followed for all six mixes in order to maintain consistency. Tests were also conducted on the microsilica material itself to check if it passes the sieve test. The facilities at ODOT laboratory were utilized in conducting these tests. The effect of densification of microsilica on its properties was evaluated from the results of the above tests.

6.2

Conclusions
During tests conducted at the ODOT laboratory on the three microsilica types used in

this project, it was observed that none of the materials could actually meet the ASTM C 1240 01 requirement when subjected to the sieve test of ASTM C 430 96. Test results were not meaningful, since negative fineness values were obtained for some of the samples. Even undensified microsilica could only meet the specifcation when a non-prescribed brush was used. Therefore, it can be concluded that the sieve test is not effective in predicting the performance of microsilica concrete that is to be used for ODOT projects, and that it is not appropriate for ODOTs assessment of the suitability of microsilica for use in concrete. The tests conducted on the concrete specimens indicate that those made with undensified microsilica show higher flexural and compressive strengths than concrete made with densified microsilica and abused microsilica, for both natural and crushed aggregate. The strengths of undensified microsilica concrete were followed by that of densified and 112

abused microsilica concretes in decreasing order. Trends observed in almost all mixes with respect to increase in strength with age, microsilica type, aggregate type and specimen size were as expected. The strength of concrete specimens from all mixes increased with age; moreover, the strength increase was more rapid in the initial ages than during later ages. Both large and small cylinders attained compressive strengths in excess of those envisaged ODOT 499.03 at 28 days of curing. Even though lower compressive strengths were obtained for densified and abused microsilica concretes compared to undensified, they still had adequate strength as required for ODOT projects. Therefore, densified microsilica can be used for construction of pavements and bridges by ODOT even though it may fail the sieve test. Just like compressive strength results, the modulus of rupture of undensified microsilica concrete was greater than corresponding values of densified and abused microsilica concretes, when natural aggregate was used. Yet, the modulus of rupture of abused microsilica was nearly equal to that of undensified microsilica, and was greater than that of densified microsilica concrete at all ages, for both types of aggregate. Coefficients of variation were higher for the modulus of rupture than the compressive strength results. Concrete made with crushed aggregate showed higher compressive and flexural strengths than concrete made with natural aggregate in most of the mixes. As noted earlier, a combination of either three large cylinders with a small cylinder, or three small cylinders with a single large cylinder, was used for any given mix to determine compressive strength. It was observed that the small sized concrete specimens yielded greater compressive strengths than large sized specimens.

113

As noted earlier, the compressive and flexural strengths of abused microsilica did not differ much from that of densified microsilica. The abused microsilica was intended to represent the worst possible situation that might arise in the field. The clumps formed during the abusing process were broken using a trowel; therefore, it can be said that the clusters of microsilica that are formed in the field due to moisture can easily be broken during the mixing process.

6.3

Recommendations

During the sieve tests conducted at ODOT, none of the three types of microsilica could pass the ASTM C 1240 01 requirement. On the other hand, when results from compressive and flexural strength tests are considered, all microsilica types achieved strengths in excess of those envisaged by ODOT Item 499.03, irrespective of whether they passed the No. 325 sieve test or not. Therefore, it is recommended that the No. 325 sieve test be abandoned as a consideration in assessing the suitability of microsilica for use in concrete. This is not to imply that densification is no longer a concern for microsilica users, nor that the No. 325 sieve test does not serve a useful purpose when used by the manufacturers for quality assurance, as it was originally conceived to do. The recommendation simply asserts that as the No. 325 sieve is conducted by an agency such as ODOT, it yields no meaningful information. It is also recommended that the microsilica should be stored for limited time only, in areas of low humidity at room temperatures, and that the mixing process should be careful and thorough, to limit the amount of densification at mixing, and to permit any bonds to be broken.

114

6.4

Implementation Plan
IMPLEMENTATION STEPS & TIME FRAME: The recommendations above can

be implemented immediately by any ODOT District including microsilica in its concrete mix design. EXPECTED BENEFITS: The main benefits from this research will derive from the use of densified microsilica from respected manufacturers in pavement and bridge construction, if such use is justified based on the results from other, more specific and expensive, studies. Another benefit will derive from the elimination of the No. 325 sieve test at the ODOT laboratory for the purpose of assessing the suitability of microsilica for use in concrete mixes. EXPECTED RISKS, OBSTACLES, & STRATEGIES TO OVERCOME THEM: It is anticipated that there may be a hesitation to abandon what may currently be the only test conducted at the ODOT laboratory in order to assure the quality of densified microsilica used in pavement and bridge construction. It is suggested that ODOT make more stringent its microsilica procurement process, in order to ensure that material is obtained from reliable manufacturers alone, whose declarations of suitability may be accepted with confidence. The possibility of bonding the manufacturer to the performance of the pavement or bridge concerned may also be considered. OTHER ODOT OFFICES AFFECTED BY THE CHANGE: Any ODOT District including microsilica in its concrete mix design. PROGRESS REPORTING & TIME FRAME: To be determined by ODOT. TECHNOLOGY TRANSFER METHODS TO BE USED: The Final Report from this study will be made available to interested parties, either in hard copy, or in electronic form, 115

the latter to include either Word .doc format or pdf. At least one refereed journal paper documenting this investigation will be prepared within a year from the completion of this contract. IMPLEMENTATION COST & SOURCE OF FUNDING: There are no costs associated with implementing the findings of this study.

116

REFERENCES

Atcin, P.C. (1990), Long-Term Compressive Strength of Silica-Fume Concrete, Journal of Materials in Civil Engineering, Vol. 2, No. 3, August, 1990, ASCE, New York, pp. 164-170.

Bowles, J.E. (1992), Engineering Properties of Soild and Their Measurement, Irwin McGraw-Hill, Boston, MA, 241 pp.

Carette, G.G. and Malhotra, V.M. (1992), Long-Term Strength Development of Silica Fume Concrete, in Fly Ash, Silica Fume, Slag, and Natural Pozzolans in Concrete, Proceedings, Fourth International Conference, Istanbul, Turkey, May 1992 (SP132:Vol. II), Malhotra, V.M. (editor), American Concrete Institute, Detroit, MI, pp. 1017-1044.

Diamond, S. and Sahu, S. (2003), "Densified Silica Fume - Is It What You Think It Is?", Proceedings of Advances in Cement and Concrete Conference, Engineering Conferences International, August 10-14, Copper Mountain, CO.

Fitch, M.G. and Abdulshafi, O.A. (1998), Field and Laboratory Evaluation of Silica Fume Modified Concrete Bridge Deck Overlays in Ohio, Transportation Research Record

117

No. 1610, Transportation Research Board, National Research Council, Washington, D.C., pp. 20-27.

Helland, S., Acker, P., Gram, H.E., and Sellevold, E.J. (1988), Condensed Silica Fume in Concrete, Fdration Internationale de la Prcontrainte (FIP) Commission on Concrete: State of Art Report, Thomas Telford Ltd, London, UK.

Jacobsen, S. and Sellevold, E.J. (1997), Frost/Salt Scaling and Ice Formation of Concrete: Effect of Curing Temperature and Silica Fume on Normal and High Strength Concrete, in Freeze-Thaw Durability of Concrete, Marchand, J., Pigeon, M., and Setzer, M. (editors), E & FN SPON, London, England, pp. 93-106.

Khayat, K.H. and Atcin, P.C. (1992), Silica Fume in Concrete - An Overview, in Fly Ash, Silica Fume, Slag, and Natural Pozzolans in Concrete, Proceedings, Fourth International Conference on, Istanbul, Turkey, 1992 (SP-132:Vol. II), Malhotra, V.M. (editor), American Concrete Institute, MI, pp. 835-872.

Krauss, P.D. and Rogalla, E.A. (1996), Transverse Cracking in Newly Constructed Bridge Decks, NCHRP Report 380, Transportation Research Board, National Research Council, Washington, D.C., 126 pp.

118

Luther, M.D. and Smith, P.A. (1991), Silica Fume Fundamentals for Use in Concrete, in Cement Manufacture and Use, Proceedings, Engineering Foundation Conference, Brown, P. W. (editor), ASCE, NY, pp. 75-106.

Malhotra, V.M. and Mehta, P.K. (1996), Pozzolanic and Cementitious Materials, Gordon and Breach Publishers, Amsterdam, The Netherlands.

Malhotra, V.M., Ramachandran, V.S., Feldman, R.F., and Atcin, P.C. (1987), Condensed Silica Fume in Concrete, CRC Press, Inc., FL, pp. 1-67.

Mehta, K. and Monteiro, P. (1993), Concrete, Structure, Properties and Materials, PrenticeHall, Englewood Cliffs, NJ, pp. 49-79.

Pinto, R.C.A. and Hover, K.C. (1997), Effect of Silica Fume and Superplasticizer Addition on Setting Behavior of High Strength Mixtures, Transportation Research Record No. 1574, Transportation Research Board, National Research Council, Washington, D.C., pp. 56-62.

Prussack, C., Jones, D., and Khaleghi, B. (2001), High Performance Concrete Bridges in Washington State, Practical paper No. 01-0348, presented at the 80th Annual Meeting of the Transportation Research Board (Preprint CD-ROM), Transportation Research Board, National Research Council, Washington, D.C.

119

APPENDIX

Table of Contents
1. Compressive Strength Test Results of Densified Microsilica with Natural Aggregate Old 2. Compressive Strength Test Results of Densified Microsilica with Natural Aggregate New 3. Compressive Strength Test Results of Densified Microsilica with Crushed Aggregate Old 4. Compressive Strength Test Results of Densified Microsilica with Crushed Aggregate New 5. Compressive Strength Test Results of Undensified Microsilica with Natural Aggregate Old 6. Compressive Strength Test Results of Undensified Microsilica with Natural Aggregate New 7. Compressive Strength Test Results of Undensified Microsilica with Crushed Aggregate Old 8. Compressive Strength Test Results of Undensified Microsilica with Crushed Aggregate New 9. Compressive Strength Test Results of Abused Microsilica with Natural Aggregate Old 10. Compressive Strength Test Results of Abused Microsilica with Natural Aggregate New 11. Compressive Strength Test Results of Abused Microsilica with Crushed Aggregate Old

12. Compressive Strength Test Results of Abused Microsilica with Crushed Aggregate New 13. Flexural Strength Test Results of Densified Microsilica with Natural Aggregate 14. Flexural Strength Test Results of Densified Microsilica with Crushed Aggregate 15. Flexural Strength Test Results of Undensified Microsilica with Natural Aggregate 16. Flexural Strength Test Results of Undensified Microsilica with Crushed Aggregate 17. Flexural Strength Test Results of Abused Microsilica with Natural Aggregate 18. Flexural Strength Test Results of Abused Microsilica with Crushed Aggregate

Compressive Strength Test Results for Densified Micro Silica with Natural Aggregate Old

Compressive Strength Test Results of Densified Microsilica with Natural Aggregate New

Compressive Strength Test Results for Densified Microsilica with Crushed Aggregate Old

Compressive Strength Test Results for Densified Microsilica with Crushed Aggregate New

Compressive Strength for Compressive Strength Test Results for Undensified Microsilica with Natural Aggregate old

Compressive Strength Test Results for Undensified Microsilica with Natural Aggregate new

Compressive Strength Test Results of Undensified Microsilica with Crushed Aggregate old

Compressive Strength Test Results of Undensified Microsilica with Crushed Aggregate new

Compressive Strength Test Results for Abused Microsilica with Natural Aggregate old

Compressive Strength Test Results for Abused Microsilica with Natural Aggregate new

Compressive Strength Test Results for Abused Microsilica with Crushed Aggregate old

Compressive Strength Test Results for Abused Microsilica with Crushed Aggregate new

Flexural Strength Test Results for Densified Microsilica with Natural Aggregate

Flexural Strength Test Results for Densified Microsilica with Crushed Aggregate

Flexural Strength Test Results for Undensified Microsilica with Natural Aggregate

Flexural Strength Test Results for Undensified Microsilica with Crushed Aggregate

Flexural Strength Test Results for Abused Microsilica with Natural Aggregate

Flexural Strength Test Results for Abused Microsilica with Crushed Aggregate

Sasatani, T., Torii, K., and Kawmura, M. (1995), Five-Year Exposure Test on Long-Term Properties of Concretes Containing, in Fly Ash, Blast-Furnace Slag, and Silica Fume in Concrete, Proceedings, Fifth International Conference, Milwaukee, WI (Vol. I), Malhotra V.M. (editor), American Concrete Institute, Detroit, MI, pp. 283296.

Soeda, M., Yamato, T., and Emoto, Y. (1999), Frost Durability of High-Performance Concrete Incorporating Slag or Silica Fume, Proceedings, Second CANMET/ACI International Conference, Gramado, RS, Brazil, 1999, Malhotra, V. M., Helene, P., Prudencio, L. R., Jr., and Doll Molin, D. C. C. (editor), American Concrete Institute, Farmington Hills, MI, pp. 409-426.

Whiting, D. and Detwiler, R. (1998), Silica Fume Concrete for Bridge Decks, National Cooperative Highway Research Program (NCHRP) Report 410, Transportation Research Board, National Research Council, Washington, D.C., pp. 1-18.

Wolsiefer Sr., J., Sr., Sivasundaram, V., Malhotra, V.M. and Carette, G.G. (1995), Performance of Concretes Incorporating Various Forms of Silica Fume, in Fly Ash, Silica Fume, Slag, and Natural Pozzolans in Concrete, Proceedings, Fifth International Conference, Milwaukee, WI, 1995 (SP-153:Vol. II), Malhotra, V.M. (editor), American Concrete Institute, Detroit, MI, pp. 591-656.

120

APPENDIX

Table of Contents
1. Compressive Strength Test Results of Densified Microsilica with Natural Aggregate Old 2. Compressive Strength Test Results of Densified Microsilica with Natural Aggregate New 3. Compressive Strength Test Results of Densified Microsilica with Crushed Aggregate Old 4. Compressive Strength Test Results of Densified Microsilica with Crushed Aggregate New 5. Compressive Strength Test Results of Undensified Microsilica with Natural Aggregate Old 6. Compressive Strength Test Results of Undensified Microsilica with Natural Aggregate New 7. Compressive Strength Test Results of Undensified Microsilica with Crushed Aggregate Old 8. Compressive Strength Test Results of Undensified Microsilica with Crushed Aggregate New 9. Compressive Strength Test Results of Abused Microsilica with Natural Aggregate Old 10. Compressive Strength Test Results of Abused Microsilica with Natural Aggregate New 11. Compressive Strength Test Results of Abused Microsilica with Crushed Aggregate Old

12. Compressive Strength Test Results of Abused Microsilica with Crushed Aggregate New 13. Flexural Strength Test Results of Densified Microsilica with Natural Aggregate 14. Flexural Strength Test Results of Densified Microsilica with Crushed Aggregate 15. Flexural Strength Test Results of Undensified Microsilica with Natural Aggregate 16. Flexural Strength Test Results of Undensified Microsilica with Crushed Aggregate 17. Flexural Strength Test Results of Abused Microsilica with Natural Aggregate 18. Flexural Strength Test Results of Abused Microsilica with Crushed Aggregate

Compressive Strength Test Results for Densified Micro Silica with Natural Aggregate Old

Compressive Strength Test Results of Densified Microsilica with Natural Aggregate New

Compressive Strength Test Results for Densified Microsilica with Crushed Aggregate Old

Compressive Strength Test Results for Densified Microsilica with Crushed Aggregate New

Compressive Strength for Compressive Strength Test Results for Undensified Microsilica with Natural Aggregate old

Compressive Strength Test Results for Undensified Microsilica with Natural Aggregate new

Compressive Strength Test Results of Undensified Microsilica with Crushed Aggregate old

Compressive Strength Test Results of Undensified Microsilica with Crushed Aggregate new

Compressive Strength Test Results for Abused Microsilica with Natural Aggregate old

Compressive Strength Test Results for Abused Microsilica with Natural Aggregate new

Compressive Strength Test Results for Abused Microsilica with Crushed Aggregate old

Compressive Strength Test Results for Abused Microsilica with Crushed Aggregate new

Flexural Strength Test Results for Densified Microsilica with Natural Aggregate

Flexural Strength Test Results for Densified Microsilica with Crushed Aggregate

Flexural Strength Test Results for Undensified Microsilica with Natural Aggregate

Flexural Strength Test Results for Undensified Microsilica with Crushed Aggregate

Flexural Strength Test Results for Abused Microsilica with Natural Aggregate

Flexural Strength Test Results for Abused Microsilica with Crushed Aggregate