xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200397Ka 4500
controlfield tag 006 m d s
007 cr mnu
008 041209s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000528
Hanhan, Amin A.
Influence of the SO content of cement on the durability and strength of concrete exposed to sodium sulfate environment
h [electronic resource] /
by Amin A. Hanhan.
[Tampa, Fla.] :
University of South Florida,
Thesis (M.S.C.E.)--University of South Florida, 2004.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 112 pages.
ABSTRACT: The objective of this investigation was to assess the influence of the SO content on the durability and strength of portland cement. Four portland cements were used in this study. The cements had a variable tricalcium silicate, tricalcium aluminate, and alkali contents, as well as differences in the amount and form of calcium sulfates. The SO content of the cements was increased by replacing part of the cement by gypsum according to ASTM C 452-95. Mortar bars and cubes were prepared for the as-received as well as for the cements with an SO content of 3.0% and 3.6%. The durability of the as-received and doped cements was determined by measuring the length change of the mortar bars that were exposed to sodium sulfate environment. The compressive strength of the mortar cubes prepared for the same mixes was measured at different ages for sets of cubes cured both in sodium sulfate solution and in saturated lime solution.It was concluded at the end of this study that there is an optimum SO content for the lowest expansion that is different from that determined for the highest compressive strength. Optimum values also differed from one cement to another and from one age to another for the same cement. The results also indicate the dependence of SO content on tricalcium aluminate and alkali content of cements. In addition, for all cements examined in this study with alkali content of less than 0.60%, increasing the SO content above 3.0% had negative effects on durability assessed by strength or expansion measurements. For the cement with highest alkali and tricalcium aluminate content, increasing the SO content from 3.0% to 3.6% delayed the onset of strength drop; however, at 360 days the strength drop experienced by both doping levels was the same.
Adviser: Zayed, Abla.
x Civil Engineering
t USF Electronic Theses and Dissertations.
Influence of the SO3 Content of Cement on the Durability and Strength of Concrete Exposed to S odium Sulfate Environment by Amin A. Hanhan A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Civil Engineering Department of Civil and Environmental Engineering College of Engineering University of South Florida Major Professor: Abla Zayed, Ph.D. Rajan Sen, Ph.D. Ram Pendyala, Ph.D. Date of Approval: November 5th, 2004 Keywords: Sulfur Trioxide, Compressive St rength, Expansion, Gyps um, Sulfate Attack Copyright 2004, Amin A. Hanhan
i TABLE OF CONTENTS LIST OF TABLES iii LIST OF FIGURES v LIST OF SYMBOLS AND ABBREVIATIONS ix ABSTRACT xi CHAPTER 1. INTRODUCTION 1 1.1 Objective 1 1.2 Phases in Clinker and Portland Cement 2 1.3 Composition of Portland Cement 4 1.3.1 Compound Composition 4 1.3.2 Impurity Oxides 6 1.3.3 Sources of Sulfates 6 1.4 Hydration of Portland Cement 8 1.5 Review of Previous Research 11 1.5.1 C3A and C4AF Hydration 28 1.5.2 Effect of Alkalis 32 1.5.3 Effects of Temperature 39 CHAPTER 2. EXPERIMENTAL PROCEDURE 40 2.1 Materials 40 2.1.1 Cements 40 18.104.22.168 Oxide Chemical Composition of the Cements 40 22.214.171.124 Bogue Calculations 40 2.1.2 Gypsum 41 2.1.3 Sand 41 2.1.4 Water 41 2.1.5 Sodium Sulfate 41 2.1.6 Fly Ash 41 2.2 Strength and Durability Tests 42 2.2.1 Mortar Cubes 42 2.2.2 Mortar Bars 46 2.3 X-Ray Powder Diffraction Analysis of Mortar Cubes and Bars 49
ii CHAPTER 3. RESULTS AND DISCUSSION 50 3.1 Characteristics of the As-Received Cements 50 3.1.1 Cement Fineness 50 3.1.2 Oxide Chemical Composition 51 3.2 Compound Composition 52 3.2.1 Mineralogical Composition According to Bogue Calculations 52 3.2.2 Mineralogical Composition Acco rding to Internal Standard Method and Rietveld Refinement Method 53 3.3 Strength and Durability 56 3.3.1 Compressive Strength of Mortar Cubes 56 3.3.2 Expansion of Mortar Bars in Sodium Sulfate Solution 76 3.4 X-Ray Diffraction Analysis of Mortar Cubes 86 3.5 X-Ray Diffraction Analysis of Mortar Bars 89 CHAPTER 4. CONCLUSIONS AND RECOMMENDATIONS 95 REFERENCES 97
iii LIST OF TABLES Table 1. Mix Proportions for the Mortar Cubes Prepared from the As-Received Cements 42 Table 2. Mix Proportions for the Mortar Cubes Prepared from the Cements with 3.0% SO3 Content 44 Table 3. Mix Proportions for the Mortar Cubes Prepared from the Cements with 3.6% SO3 Content 44 Table 4. Mix Proportions for the Mortar Bars Prepared from the Cements With 20% Fly Ash Replace ment and 3.6% SO3 Content of the Remaining Cement 47 Table 5. Proportions of Class F Fl y Ash, LOI 4.8%, for 20% Replacement of Cement 48 Table 6. Blaine Fineness 50 Table 7. Oxide Chemical Composition of the As-Received Cements 51 Table 8. Mineralogical Composition According to Bogue Calculations 52 Table 9. Mineralogical Composition Based on th e Internal Standard Method 53 Table 10. Rietveld Refinement Results for the As-Received Cements 55 Table 11. Relative Intensity Ratios of Ettringite and Gypsum for the 3.0% SO3 Content Mortar Cubes at the Age of 360 Days in Sulfate 86 Table 12. Relative Intensity Ratios of Ettringite and Gypsum for the 3.6% SO3 Content Mortar Cubes at the Age of 360 Days in Sulfate 87
iv Table 13. Relative Intensity Ratios of Ettringite and Gypsum for the 3.0% SO3 Content Mortar Bars at the Age of 480 Days in Sulfate 89 Table 14. Relative Intensity Ratios of Ettringite and Gypsum for the 3.6% SO3 Content Mortar Bars at the Age of 480 Days in Sulfate 90
v LIST OF FIGURES Figure 1. Compressive Strength verses SO3 Content for C Cement in Lime 57 Figure 2. Compressive Strength verses SO3 Content for C Cement in Sulfate 57 Figure 3. Compressive Strength verses Age for C, C-3.0, and C-3.6 in Lime 58 Figure 4. Compressive Strength verses Age for C, C-3.0, and C-3.6 in Sulfate 58 Figure 5. Compressive Strength verses SO3 Content for Cement D2 in Lime 60 Figure 6. Compressive Strength verses SO3 Content for Cement D2 in Sulfate 60 Figure 7. Compressive Strength verses Age for Cement D2, D2-3.0, and D2-3.6 in Lime 61 Figure 8. Compressive Strength verses Age for Cement D2, D2-3.0, and D2-3.6 in Sulfate 62 Figure 9. Compressive Strength verses SO3 Content for Cement E in Lime 64 Figure 10. Compressive Strength verses SO3 Content for Cement E in Sulfate 64 Figure 11. Compressive Strength verses Age for Cement E, E-3.0, and E-3.6 in Lime 65 Figure 12. Compressive Strength verses Age for Cement E, E-3.0, and E-3.6 in Sulfate 65
vi Figure 13. Compressive Strength verses SO3 Content for Cement P in Lime 67 Figure 14. Compressive Strength verses SO3 Content for Cement P in Sulfate 67 Figure 15. Compressive Strength verses Age for Cement P, P-3.0, and P-3.6 in Lime 69 Figure 16. Compressive Strength verses Age for Cement P, P-3.0, and P-3.6 in Sulfate 69 Figure 17. Compressive Strength verses Age for the As-Received Cements in Lime 70 Figure 18. Compressive Strength verses Age for the Cements with 3.0% SO3 Content in Lime 71 Figure 19. Compressive Strength verses Age for the Cements with 3.6% SO3 Content in Lime 71 Figure 20. Compressive Strength verses Age for the As-Received Cements in Sulfate 72 Figure 21. Compressive Strength verses Age for the Cements with 3.0% SO3 Content in Sulfate 73 Figure 22. Compressive Strength verses Age for the Cements with 3.6% SO3 Content in Sulfate 73 Figure 23. Compressive Strength of Doped Cements at 360 Days in Lime and Sulfates 74 Figure 24. Compressive Strength of Doped Cements at 360 Days in Sulfate 75 Figure 25. Expansion of Bars Prepared with the As-Received Cements 76 Figure 26. Expansion verses Age for the C, C-3.0, C-3.6, and C-FA Bars in Sulfate 78 Figure 27. Expansion verses Age for the D2, D2-3.0, D2-3.6, D2-4.0, and D2-FA Bars in Sulfate 79
vii Figure 28. Expansion verses Age for the E, E-3.0, E-3.6, E-4.2, and E-FA Bars in Sulfate 81 Figure 29. Expansion verses Age for P, P-3.0, P-3.6, and P-FA Bars in Sulfate 83 Figure 30. Expansion verses Age for the Cements with 3.0% SO3 Content in Sulfate 84 Figure 31. Expansion verses Age for the Cements with 3.6% SO3 Content in Sulfate 85 Figure 32. Expansion verses Age for the Cements with 20% Replacement by Fly Ash 85 Figure 33. Relative Intensity Ra tios of Ettringite and Gypsum for the 3.0% SO3 Content Mortar Cubes at the Age of 360 Days in Sulfate 87 Figure 34. Relative Intensity Ra tios of Ettringite and Gypsum for the 3.6% SO3 Content Mortar Cubes at the Age of 360 Days in Sulfate 88 Figure 35. Relative Intensity Ra tios of Ettringite and Gypsum for the C-3.0 and C-3.6 Mortar Bars at the Age of 480 Days in Sulfate 90 Figure 36. Relative Intensity Ra tios of Ettringite and Gypsum for the D2-3.0 and D2-3.6 Mortar Bars at the Age of 480 Days in Sulfate 91 Figure 37. Relative Intensity Ra tios of Ettringite and Gypsum for the E-3.0 and E-3.6 Mortar Bars at the Age of 480 Days in Sulfate 91 Figure 38. Relative Intensity Ra tios of Ettringite and Gypsum for the P-3.0 and P-3.6 Mortar Bars at the Age of 480 Days in Sulfate 92 Figure 39. Relative Intensity Ra tios of Ettringite and Gypsum for the 3.0% SO3 Content Mortar Bars at the Age of 480 Days in Sulfate 92
viii Figure 40. Relative Intensity Ra tios of Ettringite and Gypsum for the 3.6% SO3 Content Mortar Bars at the Age of 480 Days in Sulfate 93
ix LIST OF SYMBOLS AND ABBREVIATIONS ASTM American Society for Testing and Materials SEM Scanning Electron Microscope XRD X-Ray Diffraction QXRD Quantitative X-Ray Diffraction Cement Chemistry Abbreviations A Alumina, Al2O3 C Calcium Oxide, CaO F Ferric Oxide, Fe2O3 H Water, H2O S Silica, SiO2 Sulfur Trioxide, SO3 C3A Tricalcium Aluminate, 3CaO.Al2O3 C4AF Tetracalcium Aluminoferrite, 4CaO.Al2O3.Fe2O3 C2S Dicalcium Silicate, 2CaO.SiO2 C3S Tricalcium Silicate, 3CaO.SiO2 CH Calcium Hydroxide, Ca(OH)2 C H2 Gypsum, Ca2SO4.2H2O
x C H0.5 Bassanite, Ca2SO4.0.5H2O C Anhydrite, Ca2SO4 C-S-H Calcium Silicate Hydrate, nCaO.SiO2.mH2O C6A H32 Ettringite, 3CaO.Al2O3.3CaSO4.32H2O C4A H12 Monosulfoaluminate, 3CaO.Al2O3.CaSO4.12H2O
xi INFLUENCE OF THE SO3 CONTENT OF CEMENT ON THE DURABILITY AND STRENGTH OF CONCRETE EX POSED TO SODIUM SULFATE ENVIRONMENT Amin A. Hanhan ABSTRACT The objective of this investigation was to assess the influence of the SO3 content on the durability and streng th of portland cement. Four portland cements were used in this study. The cements had a variable tricalcium silicate, tricalcium aluminate, and alkali contents, as well as differences in the amount and form of calcium sulfates. The SO3 content of the cements was increased by replacing part of the cement by gypsum according to ASTM C 452-95. Mortar bars and cubes were prepared for the as-received as well as for the cements with an SO3 content of 3.0% and 3.6%. The dur ability of the as-received and doped cements was determined by measuring the length change of the mortar bars that were exposed to sodium sulfate environment. Th e compressive strength of the mortar cubes prepared for the same mixes was measured at different ages for sets of cubes cured both in sodium sulfate solution a nd in saturated lime solution. It was concluded at the end of this study that there is an optimum SO3 content for the lowest expansion that is different from that determined for the highest compressive strength. Optimum values also differed from one cement to another and from one age to
xii another for the same cement. The resu lts also indicate the dependence of SO3 content on tricalcium aluminate and alkali content of cements. In addition, for all cements examined in this study with al kali content of less than 0.60%, increasing the SO3 content above 3.0% had negative effects on durability assessed by strength or expansion measurements. For the cement with highest alkali and tri calcium aluminate content, increasing the SO3 content from 3.0% to 3.6% delayed the onset of strength drop; however, at 360 days the strength drop experienced by bot h doping levels was the same.
1 CHAPTER 1. INTRODUCTION 1.1 Objective When portland cement clinker is ground w ithout the addition of a retarder, the reactions with water are usually so rapid that quick set o ccurs. Consequently it is the common practice in the manufact uring process of portland cements, to add a retarder to control the rate of th e initial reactions. At present, gypsum (calcium sulfate dihydrate) and hemihydrate are the retard ers generally used. The addition of the correct amount of gypsum to the clinker is one of the most im portant steps in the ma nufacturing process. This added amount of gypsum, together with other sulfate sources determines the final sulfur trioxide (SO3) content of the cement. Other su lfate sources in portland cement clinker include raw materials used for manu facturing the clinker and from products of fuel combustion. Although gypsum was found to be beneficial in slowing down th e early hydration reactions of tricalcium aluminate (C3A), and thus contro lling the setting time; nevertheless, and in order to provide suffici ent protection against the abnormal expansion that could result from the use of excessi ve amounts of gypsum, placing a limit on the SO3 content of portland cement in standard speci fications has been the tradition. ASTM C 150 limits the SO3 content for Types I and II cements to 3.0%, for Type III cement 3.5%,
2 and for Types IV and V cements to 2.3% (these limits apply for the case when the tricalcium aluminate C3A is 8% or less). Alexander a nd co-workers  stated that Â“according to surveys published by CEMBUR EAU and Cement and Lime Manufacture, the limit in various national standards ranges from 2.5 to 5.0% SO3, depending on cement fineness or composition or bothÂ”. During the la st decades, the Blaine fineness of typical cements has increased dramatically, mainly in an attempt to increase concrete early strength. Since increasing cem ent fineness leads to higher early rate of hydration, and since gypsum is basically need ed to control the early hydra tion reactions of tricalcium aluminate (C3A), cement fineness would be an importa nt consideration in determining the quantity of gypsum required for proper retarda tion. Also during the last decades, a large increase in the sulfate content of the five ASTM types of cement clinkers occurred. This increase was to some extent the result of environmental restrictions on sulfur and other emissions, which in turn resulted in th e change of the burning and dust recycling practices. Since the total SO3 content is limited in specificatio ns, regardless of the origin of SO3, this change had the effect of limiting the amounts of gypsum needed for proper retardation without exceeding th e limits set in specifications. The objective of this research was to assess the influence of increasing the SO3 content of cement on the durability and strength. 1.2 Phases in Clinker and Portland Cement The manufacturing of portland cement is in principle very simple and relies on the use of abundant raw materials. A mixture, usua lly of limestone and clay, is heated in a kiln to 1400 Â– 1600 C (2550 to 2900 F), which is the temperature range in which the
3 two materials interact chemically to form the calcium silicates. The purity and uniformity of the raw materials determines the quality of the produced cement. A source of calcium oxide and a source of silica are the main sources needed. The most commonly used source of calcium oxide is limestone (calci um carbonate) although other sources such as chalk, shell deposits, and calca reous mud, are used. Iron-bear ing aluminosilicates are the primary source of silica. Clays or silts are usua lly preferred because they exist in a finely divided state; but shales, sc hists, and other argillaceous rocks are also used. Although quartz is the major form of pure silica in natu re, it is a relatively unreactive material and, moreover, pure lime-silica mixes have very high fusion temperatures. The aluminum and iron oxides coming from the siliceous raw mate rials, although they can lead to problems of durability and abnormal setting behavior, they act as fluxing agents, lowering the fusion temperature of the portion of the raw mix to a practical firing temperature . The heat treatment of the raw feed is te rmed clinkering, to distinguish it from sintering (where no melting occurs) and fusion (where complete melting occurs) . In the cement Kiln, partial melting takes place; only about one-fourth of the charge is in the liquid state at any time, but it is in this fr action that the necessary chemical reactions proceed. The kiln is a long st eel cylinder lined with refract ory brick and inclined a few degrees from the horizontal. It is rotated at about 60-200 rev/h about its axis. The raw feed enters the kiln at the high end, and is slowly moved down the le ngth of the kiln by the actions of rotation and inclination. As the raw feed moves down, it moves gradually into zones of increasing temperature, wh ere four distinct processes take place: evaporation, calcinations, clinke ring, and cooling. Free water is lost by evaporation in the dehydration zone and consequently the ch arge quickly heats up to calcinations
4 temperature. At 600 C water is lost from the arg illaceous material and at about 900 C limestone decomposes and carbon oxide is lost In the calcinations zone, the charge transforms into a reactive mixture of oxides that can enter into new chemical combinations. At the later part of the calcinations zone and around 1200 C, calcium aluminates and ferrites form through solid-state reactions. These compounds act as fluxes, melting around 1350 C to begin the clinkering zone. This is the heart of the kiln where final chemical combinations occur to form the calcium silicates. Finally, as the charge moves past the flame in the final few me ters, it rapidly drops off in temperature in the cooling zone. As the liquid phase solidif ies again, it produces hard, dark-gray porous nodules (6 to 50 mm in diameter) known as clin ker. The clinker is conveyed to ball mills, where it is ground to a fine powder. A small amount of gypsum is inter-ground with the clinker to control the early reactions of tric alcium aluminate that, without the addition of gypsum, can cause flash setting of the clinker . Portland cement is clinker inter-ground with gypsum; without the gypsum, it is only ground clinker. 1.3 Composition of Portland Cement 1.3.1 Compound Composition The typical chemical composition of portland cement consists of tricalcium silicate (3CaO.SiO2), dicalcium silicate (2CaO.SiO2), tricalcium aluminate (3CaO.Al2O3), tetracalcium aluminoferrite (4CaO.Al2O3.Fe2O3), and calcium sulfate dihydrate or gypsum (CaSO4.2H2O). The oxide notation by which these compounds are traditionally written and frequently used in ceramic ch emistry, gave rise to a unique shorthand notation that has universal use among cement scientists. Thus tri calcium silicate is
5 usually written as C3S, dicalcium silicate as C2S, tricalcium aluminate as C3A, tetracalcium aluminoferrite as C4AF, and calcium sulfate dihydrate as C H2 where represents SO3. The compound composition of portland cement is usually estimated by calculation using the ideal compound stoichiometries a nd oxide analysis determined by standard methods (usually X-ray fluorescence spectro scopy) and available from the cement supplier (mill certificate). The calculation of the phases from the composition is known as the Bogue calculations. The values obtained by Bogue calculations are only approximate and do not necessarily repres ent the true values. This is mainly because Bogue calculations are based on seve ral assumptions that are not usually found in the cement chemistry. Among these assumptions are chemical equilibrium and the consistency in the composition of all the phases . It is know n that chemical equilibrium rarely occurs inside the kiln during cooling, and it was found that a large number of substitutional ions can be incorporated in most of the phases. Thus, the phases do not have the compositions assumed for them. Nevertheless ASTM C 150 gives the Bogue calculations that are suitable for most purposes, alt hough more sophisticated proce dures have been developed. Knowing the compound composition of portland cement makes it possible to predict the properties of the cement, but more important is the fact that manipulation of the compound composition can be used to modify cert ain properties of the cement so that the cement will perform more satisfactorily in particular applications. C3A and C3S are the most reactive compounds, where as C2S reacts much more slowly. The calcium silicates provide most of the strength developed by portland cement; C3S provides most of the early strength (in the first three to four weeks); while C2S provides most of the ultimate
6 strength. C3A and C4AF control the setting time and the presence of gypsum slows the early rate of hydration of C3A. The reaction of C4AF + gypsum + water is believed to be somewhat slower than C3S, whereas the hydration of C4AF without gypsum is faster. Gypsum also increases the rate of hydration of the calcium silicates, which also compete for sulfate during hydration. 1.3.2 Impurity Oxides Since only approximate chemical equilibrium can be attained in the rotary kiln, it is to be expected that all ceme nt compounds will contain small amounts of the other oxides present in the clinker. The calcium silicates probably cont ain about 3% by weight of impurity oxides, principally Al2O3, Fe2O3, and MgO . Impure C3S, as it exists in portland cement, is known as alite and impure C2S as belite. Both alite and belite are more reactive than the pure silic ates and hydrate more rapidly. C3A contains considerable amounts (about 10% by weight) of SiO2 and Fe2O3, while C4AF contains considerable SiO2 and much MgO . 1.3.3 Sources of Sulfates The most important source of sulfates in portland cement is calcium sulfate. Various forms of calcium sulfate (anhydrite, hemihydrat e, dihydrate) are adde d to clinker during cement grinding to control the cement setting characteristics as was stated previously. These sulfates may be added in the form of natural or industrial-grade calcium sulfate dihydrate (gypsum) or anhydrite. Additional sulfat es originate from the clinker, in which they are formed during the manufacturing of the clinker from the raw materials, and from
7 the products of fuel combustion. Predominantl y, they are present in the clinker in the form of alkali-and calcium-alkali sulfates (doub le salts), and occasiona lly, in the form of calcium sulfate anhydrite or other phases [3 ]. The most common sulfate phases present in clinkers are arcanite, K2SO4, calcium langbeinite, KC23, and aphthitelite K3N 4. As stated earlier, and as a result of environmental restrictions on sulfur and other emissions, a large increase in the sulfate content of currently produced ceme nts occurred. Thus, the proportion of sulfate that is pr esent in typical cemen ts in the form of alkali-and alkalicalcium sulfates, originating from the clinker, could be in some cases higher than it was in the past . Occasionally, additional sources of sulfate in concrete may be the sulfates (e.g. gypsum) or sulfides present in aggregate. Al so, sulfates may be components of mineral and chemical admixtures. Thus, when using an unknown or new admixture in concrete applications potentially expos ed to sulfate conditions, it is advisable to check the chemical or mineralogical nature of all co ncrete materials . Finally, mixing water could be a possible source of sulfate, but this is considered to be an improbable source of serious damage. Although the chemical requirements of portland cement given in ASTM C 150 allow wide variations in chemical compositi on, since it has been found that cements with quite different chemical compositions may have suitable physical behavior, the only limits placed on all cement types are the MgO and SO3 levels and, optionally alkalis.
8 1.4 Hydration of Portland Cement It is very important to understand the r eactions that take place during the hydration of portland cement in order to assess the effect of the SO3 content of the cement. Although the hydration of portland cement invol ves a complex of chemical reactions, three main reactions are the ones involving the calcium silicates and calcium aluminates. The hydration reactions of the two calcium silicates ( Equations 1 and 2 ) are stoichiometrically very similar, differi ng only in the amount of calcium hydroxide formed:2C3S + 11 H C3S2H8 + 3CH (1) Tricalcium water C-S-H calcium silicate hydroxide 2C2S + 9H C3S2H8 + CH (2) Dicalcium water C-S-H calcium silicate hydroxide The principle hydration product is a cal cium silicate hydrate. The formula C3S2H8 is only approximate because the composition of this hydrate is actually variable over quite a wide range. C3S2H8 is a poorly crystalline material that forms extremely small particles in the size range of colloidal matter (less than 1m) in any dimension . Its name, C-S-H (or C-S-H gel) reflects these properties. In contrast, calcium hydroxi de is a crystalline material with a fixed composition. Although C2S hydrates in a similar manner to C3S, it is much slower because it is a less reactive compound than C3S.
9 The hydration of C3A in portland cement involves reac tions with sulfate ions that are mainly supplied by the di ssolution of gypsum. The prim ary initial reaction of C3A is C3A + 3C H2 + 26H C6A 3H32 (3) Tricalcium gypsum water ettringite aluminate The hydration product which is a calciu m sulfoaluminate hydrate is called Â“ettringiteÂ”. Ettringite is a stable hydration product only while there is an ample supply of sulfate available . If the sulf ate is all consumed before the C3A has completely hydrated, then ettringite tran sforms to another calcium su lfoaluminate hydrate containing less sulfate:2C3A + C6A 3H32 + 4H 3C4A H12 (4) This second hydration product is si mply called monosulfoaluminate. Monosulfoaluminate may sometimes form before ettringite if hydrating C3A consumes the sulfate ions faster than they can be supplied by dissolution of the gypsum in the mix water. It was found that the formation of ettringite slows down the hydration of C3A by creating a diffusion barrier around C3A grains. This barrier is broken down during the conversion to monosulfoaluminate and allows C3A to react rapidly again . The more gypsum there is in the system, the longer the ettringite will remain stable. Conversion to monosulfoaluminate will occur in most cemen ts within 12 to 36 h, after all the gypsum has been used to form ettringite. The forma tion of monosulfoaluminate occurs because in most cements there is not sufficient gypsum necessary to form ettringite from all the
10 available aluminate ions. When monosulfoalumi nate is brought into contact with a new source of sulfate ions, then ettrin gite can be formed once again:C4A H12 + 2C H2 + 16H C6A 3H32 (5) This potential for reforming ettringite is the basis for sulfate attack of portland cements when exposed to an external supply of sulfate ions. If gypsum is not present, C3A will react rapidly with water to form calcium aluminate hydrates: C3A + 21H C4AH13 + C2AH8 (6) These hydrates are not stable and later convert to C3AH6 (hydrogarnet): C4AH13 + C2AH8 2C3AH6 + 9H (7) If C3A is very reactive, even with the presence of gypsum, small amounts of hydrogarnet may be found in a hydrated cem ent. When small amounts of gypsum are present, there may still be unreacted C3A present when all of the ettringite has been converted to monosulfoaluminate. In this case, a solid solution between C4A H12 and C4AH13 is formed, the two hydrates having the same crystal structure. This solid solution is written as C3A(C ,CH)H12. The hydration of the ferrite phase C4AF forms similar hydration products to C3A in both cases with or without gypsum, however th e reactions are slower and involve less heat. Changes in the composition of the ferrite phase affect on ly the rate of hydration; as iron content is raised, hydrat ion becomes slower. Practical experience has shown that cements low in C3A, but high in C4AF are much more resistant to sulfate attack . This means that the formation of ettringite from monosulfoaluminate (Eq. 5), does not occur. It has not been establishe d why this is so; it may be that an iron-substituted
11 monosulfoaluminate cannot react to form ettr ingite. Alternatively, the presence of the amorphous product (F,A)H3 may in someway prevent the reaction described in Eq. 5 from occurring. 1.5 Review of Previous Research The phenomenon of optimum sulfate levels in portland cements has occupied researchers for decades. Beginning in 1870 Mich aelis discovered the most remarkable effect of adding only 2% of raw gypsum to ground cement on regulating setting, that is to lower the setting time to 5-7 hours . This discovery of Michaelis then became a very significant issue to th e whole world especially when it was noticed that adding gypsum, not only regulates the rapid setting, but even improves the strength of cement mortars . An early research work done on the setti ng time concluded that there would be no advantage in adding more than 2% CaSO4 and that specifications should call for such a limit. In this work, one type of clinker was grounded with gypsum at 0.5% intervals from 0.5 to 7.0% and time of set was measured with a Vicat-Type apparatus. Maximum retardation was obtained with 1.5% CaSO4 and additional gypsum did not further delay initial set for this clinker. ASTM adopted a limit of 2.0% SO3 in 1920, but the recommended limits for SO3 have increased consistently sinc e then due to several factors, including higher sulfur fuels used in burning clinker and increased clinker fineness . Several researchers observed an increase in strength with increas ing gypsum content and the first publication indicating an optimum amount of gypsum for strength development was published in 1924. The researchers also observed that gyps um inter-ground with clinker produced higher stre ngth mortars than cements made by blending ground clinker
12 plus gypsum. This suggests that gypsum finene ss plays an important role in strength development, a significant finding which has been confirmed more precisely in recent years . A study conducted by ASTM Committee C-1, Subcommittee B in 1931 concluded that more than 1.75% SO3 was detrimental in cements with 8% C3A but that higher C3A cements (11-16%) coul d contain up to 2.5% SO3 without detrimental effects. A close look at their da ta indicates that 2.5% SO3 was optimum for strength development measured at one day, but that 3.0% SO3 gave higher strengths at later ages. Other investigations also showed similar trends but indicated that for a certain optimum gypsum level, a set of conditions should be well defi ned including the fineness of the cement, the form and reactivity of the calcium sulfate, the curing conditions, the temperature during storage, and the age of the specimen at test Kanare and Gartner [4 ] stated that Â“the sulfate addition level which gives optimum st rength development may not give the best volume stability under all circumstancesÂ”. In 1941, Lerch , began a thorough study of the effects of gypsum and lime on cement hydration. Twelve commercial cli nkers covering the range of chemical composition generally found in portland cemen ts were selected for study. Each of these compositions was used with various SO3 contents with fineness held constant at about 1900 sq. cm. per gram. In addition five of the clinkers were ground to varying degrees of fineness with SO3 constant at about 1.8 per cent. The results of the study showed that with some cement compositions, the gypsum retards the initial hydration and set, while with others it acts as an accelerator. It also s howed that for many cements, the strengths can be increased and the drying shrinkage decreased by the use of larger additions of gypsum than were permitted by th e specifications at th at time. The results
13 also showed that the fine ness, the alkalis, and the C3A content all influence the gypsum requirements. Lerch defined a Â“properly retard edÂ” cement on the basis of the shape of the heat-liberation curve during the first 30 hr. of hydration. A Â“properly retardedÂ” cement, according to Lerch, Â“is the one that contains the minimum quantity of gypsum required to give a heat-liberation curve th at shows two cycles of asce nding and descending rates and that shows no appreciable change with larg er additions of gypsumÂ”. The results of the tests showed that when considering ce ments ground from a given clinker, those containing the proper amount of gypsum to gi ve this type of curve will develop the highest strength and th e lowest contraction. As for the influence of the alkalis and C3A content of the cement on the gypsum requirements for proper retardation, the results showed that with cem ents of low alkali content, those of high C3A content require larg er additions of gypsum than those of low C3A. For cements of the same C3A content, those high in alkalis react with gypsum more rapidly and require larger a dditions of gypsum than those low in alkalis. The author stated that Â“there is some ev idence that cements containing Na2O require larger amounts of gypsum than do similar cements cont aining an equivale nt quantity of K2OÂ”. The author adds that Â“it appears that at least part of the alkalis of the cement are present in the aluminate phases, and that aluminate phases containing alkalis react with water more rapidly than do similar phases which are alkali-free or of lo wer alkali content. Thus the cements of higher alkali content require larger additions of gypsum for proper retardation than do similar cements of lower alkali contentÂ”. The author also explained that increas ing the specific surface of the cement increases the quantity of aluminate phases ava ilable for reaction with the water at early
14 ages and thereby increases the quantity of gypsum required for proper retardation of cements of moderately high or high C3A content. It was stated that Â“cements of low C3A and low alkali content without added gypsum can be mixed with water without the occurren ce of a flash-setÂ” . The explanation of the author for this statement was that with cements of this type an amorphous hydrated calcium ferrite precipitates on the surface of th e cement particles and seals the surface in a manner such as to retard subsequent hydra tion, while with adde d gypsum a crystalline hydrated calcium sulfoferrite is formed wh ich does not seal the surface and the hydration is accelerated. The results of the physical tests showed th at for many cements the strengths can be increased and the contraction on drying or th e expansion in water storage decreased by the use of larger additions of gypsum than were permitted by specifications. In some instances, the strengths were increased by as mu ch as 20 to 50 percent and the contraction decreased by as much as 30 to 50 percent. For cements of low C3A and low alkali content, the strengths were not increased nor were the contractions decreased by larger additions of gypsum. The cements high in C3A regardless of their alkali content or cements high in alkalis regardless of C3A content require larger additions of gypsum. Lerch concluded that gypsum could be added in larger amounts than was permitted by the specifications at that time wit hout danger of delayed expansion. Although some researchers found that LerchÂ’s work could be faulted in some areas, other researchers reached essentially the same conclusions. Perhaps the most important conclusion which can be drawn from LerchÂ’s study is that most cements show different
15 optimum gypsum requirements for different prop erties, such as strength or shrinkage, at different ages. In 1956, the standard specification for portl and cement C 150 was revised to reflect the thinking that SO3 content should depend upon cement type and composition, and by 1961 up to 4.0% SO3 was permitted in Type III cements with more than 8% C3A . Several researchers have attempted to fi nd correlations between clinker composition, cement fineness, and the level of gyps um which produces optimum strength development, however none of the given equa tions proved to be suffi ciently accurate or reliable for use as predictive tools. Kanare a nd Gartner  stated that, Â“it seems that the technology of optimizing the sulfate content in portland cements has remained virtually unchanged since the turn of the century wh en technologists first made cements with several levels of gypsum to see which got str ongest. This is still the best and only method to use for the purpose of maximizing early strength in commercial cementsÂ”. Some research work was done on the effect of fineness of calcium sulfates in cements on the optimum sulfate content. Observ ations by several researchers suggest that gypsum interground with clinker might be Â“sm earedÂ” on the surfaces of clinker particles rather than simply ground into a fine powder which is intimately mixed with the clinker. Calcium sulfate particles several hundred mi crometers in size are often observed in commercial cements . These coarse gypsum particles could be effectively Â“inertÂ” during early cement hydration, possibly as a result of encapsulation by hydrated cement phases. In such a case these large particles ma y influence the attempts to optimize sulfate content.
16 Panigrahy et.al. , studied the differe ntial comminution of gypsum in cements ground in different mills. Identical mixes co ntaining fixed amounts of ordinary portland cement clinker and gypsum were ground in two types of industrial cement mills, ball mill (BM) and vertical roller mill (V RM), to identical Blaine fineness to examine the effect of any possible differential comminution of gypsum on cement setting times. The authors found that in every occasion, the VRM cement r ecorded much lower setting time than the BM cement. It was also found that although the SO3 contents in all the cement samples are almost identical, their gypsum XRD pulse counts were quite different and they were always significantly less in the BM products than in the VRM products. Since the XRD spectra did not record any other crystalline Ca lcium sulfate phase, it is evident that the BM product contained a signifi cant portion of Calcium sulfate in amorphous form. The effect of grinding on amorphism was also re ported by other researchers and the authors stated that Â“it is only the differential amorphism of gypsum caused during grinding which is solely responsible for vari ation in setting timesÂ”. It wa s concluded from the study that during comminution of cements, the degree of crystallinity of gypsum as determined by X-ray diffraction (XRD), changes with used grinding mills and this causes changes in setting times of similar cements even wh en ground to identical Blaine fineness. Goswami, Mohapatra, and Panda , studied gypsum dehydration during comminution and its effect on cement propertie s. The authors concluded that gypsum in cements, ground to the same fineness in differe nt mills, may be subjected to different degrees of dehydration. During cement grinding in an industrial mill, gypsum is often dehydrated to hemihydrate. It was found that the presence of hemihydrate in the cement increases ettringite formation during early hydration, retards the setting times, and
17 reduces the strength of the cement by about 10% The authors suggested that the fall in strength is particularly due to re-hydrati on of part of the hemihydrate present in the cement. Among the earliest findings concerning the optimum gypsum content were those of Lerch  who found that the optimum gypsum for minimum shrinkage of mortars was slightly greater at 28 days than at 3 da ys. Another important finding was that of Alexander and co-workers  who st udied the effects of variable SO3 in concretes. In one study, using three commercial cements, each produced at four different levels of SO3, creep under compressive loads was found to be very sensitive to the SO3 content of the cement, and that the optimum SO3 for minimum creep, generally 3-4% SO3, was at least 0.6% higher than the optimum gypsum for minimum drying shrinkage. The SO3 content of the cement was also shown to have no si gnificant effect on concrete modulus of elasticity. Some researchers investigated the effect of SO3 on the hydration of the silicate phases. They concluded that calcium sulfat e has an accelerating e ffect upon the hydration of C3S . One indication among th e findings of these researchers is that any level of gypsum which keeps the hydration liquid saturate d with sulfate ions at early ages will provide acceleration of the al ite hydration. In addition to th e fact that gypsum accelerates the hydration of the silicate phases, it is belie ved that it also influences the composition of the hydration products. It has been suggest ed that the silicate hydration products containing sulfate are intrinsically inferior to those without sulfat e, but the opposite has also been claimed . It app ears that there is a trade-off between the increased amount of gel produced by the addition of gypsum and its infe rior quality at later ages. It seems that
18 a certain combination of the amount and qua lity of gel was optimum for compressive strength. Sulfate ions are isostructural with silicate ions and might be expected to substitute for the silicate in CSH gels . Bentur , studied the effect of gyps um on the hydration and strength of C3S pastes. He studied pastes hydrated at 0.43 wate r-to-solid ratio of mixtures of C3S with 0,2,4, and 9% analytical-grade gypsum. He found that the effect of gypsum content on compressive strength changes with time. At early ages (1 and 2 days) the compre ssive strengths of the pastes containing 2,4, and 9% gypsum were similar, whereas the pure C3S was much weaker. At 3 and 7 days, the compressive st rength was maximum at 2% gypsum content. At 28 and 90 days, the strengths of pure C3S and the paste contai ning 2% gypsum were markedly greater than those of the pastes containing 4 and 9% gypsum. He also studied the C/S values of the pastes with different gypsum content. He found that in pure C3S past, this value decreases as hydration advances; beyond 70% hydration, it tends to stabilize. In the pastes co ntaining gypsum, the C/S ratio increases up to approximately 60% hydration and decreases thereafter. Bentur also found that in 70% hydrated pastes, the C/S ratio increases with the /S ratio. This trend indicate s that the reaction of the sulfate causes chemical changes in the struct ure of the CSH gel. He assumed that the variations in the strength of pastes having the same degree of hydration are related with changes in intrinsic strength. At 40 and 50% hydration, the C/S ratio is independent of gypsum content and so is the intrinsic stre ngth. At higher degrees of hydration, the C/S ratio increases with gypsum content and the intrinsic strength d ecreases with it. He suggested that the difference in the gypsum effect at lower and higher degrees of hydration might be the result of a slow reacti on between the sulfate and the CSH gel, so
19 that the influence of gypsum on the chemical constitution and mechanical quality of the CSH gel begins to be important only afte r a certain period of hydration. Thus the optimum gypsum content is the value at which the optimum combination of quantity and quality of the CSH gel occurs. MeÂ’neÂ’trier, et. al. , a nd in order to better understa nd both the effect of gypsum on cement hydration and the morphology of the hydration products, c onducted studies on C3S hydration in the presence of gypsum and then incorporated this into a mechanical study of cement hydration. The results of th is study confirmed an increase in C3S dissolution rate in the presen ce of gypsum. The study also re vealed that a substantial amount of sulfur is incorporat ed in the C-S-H, this was evidenced by the disappearance of gypsum from the system. It is the be lief of the authors that sulfate ions may replace the silicate ions in the amorphous structure of C-S-H. The authors also concluded that the morphology of the C-S-H fo rmed in the first minutes of hydration is similar to that of C-S-H formed in a satu rated lime solution rather than in pure H2O. However, the morphology developed in later st ages resembles that of usually-observed C-S-H. Observations of other researchers suggest that sulfate may affect the strength by modifying the morphology of the hydration prod ucts, or by modifying the pore structure which surrounds and permeates the solids. Other investigators found that although gypsum produced a higher hydra tion rate and, therefore, a mo re rapid strength increase, the resulting gel was less condensed (had more pores) and had higher lime-silica ratio, at the same degree of hydration, in the pastes containing gypsum. The lime-silica ratio is one of the most intensively studied aspect s of cement chemistry. At a given degree of
20 hydration, compressive strength wa s related to the inverse of the C/S ratio as determined by QXRD, with best strength at C/S = 1.9 (mol ar basis). C/S decreased with increasing degree of hydration and increased wi th increasing sulfate-to-silica ( /S) ratio . Some researchers estimated the maximum amount of sulfate which could be accommodated in the gel to be equivalent to a S/ molar ratio of 5.85, which is approximately equal to 4.5% SO3 by weight in ordinary portland ceme nt. Since most commercial cements contain far less than 4.5% SO3, all of the gypsum could, in principle, be accommodated in the CSH. The sulfate substitution for silicon in hydrated C3S paste appears to have a small contribution to the increase in strength since optimizing the SO3 content of a cement typically can produce a 25% increase in co mpressive strength of mortar cubes . The most important effect of gypsum is to accelerate C3S hydration in addition to the important role of retarding C3A and C4AF hydration. According to Kanare and Gartener , Â“it could be concluded that the uptake of SO3 by CSH is potentially an important fact or in determining the kinetics of SO3 depletion in cement hydration at early ages. Although by the end of the induc tion period (which usually takes few hours) only very little C3S is consumed and very little CSH is formed, the steady hydration of C3S that occurs after that resu lts in a rapid uptake of CaSO4 by CSH, well before the onset of the re newed rapid aluminate-phase hydrationÂ”. Ish-Shalom and Bentur , studied the e ffects of aluminate and sulfate contents on the hydration and strength of portland cement pa stes and mortars. They used three cement samples varying in C3A content and one varying in SO3 content. In the interpretation of their results, bound water was used as a measure of quantity of binding material, and free lime to bound water ratio (FLWR) as a measure of chemical constitution and quality. In
21 their work they stated that aluminate a nd gypsum can influence the hydration in two different ways: directly, by the presence of sulfoaluminate hydrate, or, indirectly, by affecting the rate of hydration of the silicate phases or the qu ality of the hydrosilicate gel that is formed or both. They assumed that th e extent of hydration re presents the quantity of binding material while FLWR represents some measure of its quality. Their results showed that when comparing different cem ents hydrated under the same conditions (period and temperature of hydr ation) high compressive streng th is associated with the higher FLWR but not with bound water. They suggested that in these cases high FLWR indicates the existence of a high quality gel. According to their explanation, high FLWR implies low Bound Lime to Bound Water Ra tio (BLWR) and since bound water is approximately proportional to the amount of hy drated silicates, then low (BLWR) means low CaO/SiO2 ratio in the hydrated gel. Thus, a hi gh FLWR indicates the existence of a low CaO/SiO2 hydrated gel and this gel indicates according to other researchers a large proportion of double tobermorite layered structur e in it, which is also connected with higher specific surface, which in turn leads to higher strength. They concluded from their work that the highest compressive st rength is achieved with medium aluminate conten t cement for all the temperatur es and ages investigated. A good correlation was observed between streng th and free lime water ratio (FLWR), which is a parameter of the chemical constitu tion of the cement gel. They also concluded that the increase in sulfate c ontent of the medium aluminate cement caused an increase in extent of hydration (bound water) and a reduction in the quality of the gel (FLWR). An optimum combination of both parameters (expressing quality and quantity of gel) was achieved at the medium sulfate cement which exhibited the highest compressive strength.
22 They also found that the medium aluminate cement exhibited an optimum response to the effect of high initial temperature: The benefi cial effect at early age (high strength and degree of hydration) was highest while the de leterious effect at later age was lowest. The effect of added gypsum on the compressi ve strength of por tland cement clinker was also studied by Soroka and Relis . Th e authors state that there is an optimum gypsum content which imparts the cement maximum strength and minimum shrinkage without excessive expansion. This optim um, generally speaking, depends on, and increases, with C3A and alkali oxide contents of th e cement and with its fineness. The observed optimum content in the strength curv e, and according to the authors, implies that the addition of gypsum involves two opposing effects. The first, which predominates the lower range of SO3 content, has a beneficial e ffect on strength, and brings about the ascending part of th e curve. The second, which pre-dominates the range of SO3 content greater than the optimum, has an adve rse effect and brings about retro-gradation in strength and the associated descending part of the curve. This adverse effect may be attributed to internal cracking which take s place when an excessive amount of gypsum (i.e., more than the optimum) is added to th e cement. It should be noted, however, that sulfate expansion is probably not the onl y mechanism involved because such retrogradation in strength was observed also in C3S and alite pastes, i.e., under conditions where no ettringite is formed. It was, acco rdingly, suggested by researchers that the hydration of C3S and alite in the presence of gypsum resulted in a C-S-H gel of inferior quality (a gel of a lower intrinsic strengt h), and it was shown that such a gel was characterized by lattice-substituted and a higher C/S ratio .
23 The test data of this study also indica ted that the addition of gypsum to portland cement clinker resulted in highe r strength, in particular at th e early ages of one and three days. This higher strength was associated w ith a lower degree of hydration and a greater porosity. Under test conditions, this observati on, implied that the beneficial effect of the gypsum on the strength of the cement was attrib utable to the improved quality of the gel which was produced in the presence of gypsum The study showed th at the addition of the gypsum resulted in hydrati on products of greater average density and it was suggested that this greater density brought about the higher strength. The authors also suggested that the greater density of the hydration products was due to the pressure generated on formation of ettringite due to the increase in the volume of the solids involved in the reaction. Alexander and Ivanusec , studied the long-term effects of cement SO3 content on the strength of concrete. Six brands of cements were each manu factured at four SO3 contents, in full scale plants. The strengths were determined, at up to one year, in concretes of high and low w/c ratios. The study showed that strength was usually independent of, or linearly related to, the SO3 content observed. Only rarely was a well defined optimum SO3 content observed. The study also showed that there is an appreciable increase in strength be tween 28 days and one year at all SO3 levels. The greatest strength increase with time occurred with low-C3A cement. Almost invariably, with each cement, the strength v. SO3 content relationships at 28 days and one year are similar, and the strength increase between these ages is therefore largely insensitive to the SO3 content of cement. The authors conclude d that the associati on between the 28-day strength of concrete and the C3S content of cement does not vary greatly with the SO3
24 content of cement. However, th e correlation coefficient for C3A and strengths shows considerable sensitivity to SO3 content. The one-year strengths of concrete made from under-sulfated cements showed a stro ng negative association with the C3A content of cement. Under these conditions, differences in the C3A contents could account for up to 10 MPa difference in strength. It was also conc luded that the strength of the association between one-year concre te strength and the C3A content of cement varies with the SO3 content of cement. At this age, the degree of sensitivity of the C3A coefficient to SO3 content depends, in turn, on w/c ratio. Since only few explanations have been sugge sted in order to e xplain the effect of gypsum on the mechanical properties of the cem ent and since these ex planations are not always complete and even sometimes c ontradictory, Soroka and Abayneh , conducted a study aimed to establish the effect of gypsum, if any, on the structure of the cement paste, and to try and relate, if possible, this effect to the mechanical properties of the paste. It was clearly evident from the study, that an optimum SO3 content existed in most cases, but not always, with respect to the strength and drying shrinkage of the pastes. Strength-wise, the optimum was 2 to 3% increasing to 4% in the finer cement. It was pointed out that the increase in the optim um content with the fineness of the cement was generally observed. The authors explai ned the higher value of 4% which was observed in the finer cement by the higher rate of hydration of such cements in comparison with their coarser counterparts. Su lfate expansion is gene rally attributed to the continued formation of ettr ingite in the set cement as a result of topochemical reaction between the gypsum and the C3A of the cement. As more gypsum is consumed at an earlier stage with a higher rate of hydration, and because only the gypsum which is left to
25 react at the later stages causes expansion, a greater gypsum content can be tolerated in the finer cements. The study also showed that for SO3 contents exceeding the optimum of 3%, and particularly at the later ages of 28 and 90 days, the presence of the gypsum significantly retarded the hydrat ion of the cements. It was concluded from the study that the strength of a cement at a given SO3 content in the range studied (i.e., to 5%), is significantly related to the degree of hydration, and that at a given degree of hydration or porosity, strength is increased with the increase in the SO3 content of the cement. The authors found that the improved strength with the increased SO3 content could not be related to the quality of the ge l, as it is reflected in the de nsity of the hydr ation products, nor to differences in pore-size distribu tion. The authors concluded that there are apparently some other factors invol ved and this warrents further study. The increase in strength with increasing SO3 content, can be explained by the fact that the presence of calcium sulfate accelerates the hydrati on of alite, but this does not explain the existence of a maximum. Kana re and Gartner , suggests one possible explanation as follows:Â“When the optimum am ount of calcium sulfate is present, the aluminate phases are retarded and the liquid phase remains saturated with respect to gypsum until after the alite has experienced its maximum rate of hydration and the cement has achieved final set. This permits the maximum rate of alite hydration and therefore maximizes early strength development. However, if too much sulfate is present, then the remaining aluminate hydration will be retarded and so w ill continue to produce ettringite after the paste has hardened, causi ng localized expansion of the paste structure and consequent reduction in strength deve lopment due to micro-crackingÂ”. This explanation is supported by the fact that the 1-day strength-versus-SO3 curves for low
26 C3A cements are virtually flat in the region beyond the SO3 level that gives maximum strength, whereas the curves for cem ents containing more than 5% C3A show significant decreases in strength at higher SO3 levels. When gypsum conti nues to be present beyond the time of final set, this need not be nece ssarily detrimental to the early paste structure because the growth of expansive ettringite can be accommodated and micro-cracks which may develop can be autogenously filled by ne wly formed CSH . Another important fact is that it is reasonable to assume that the strength increases as the total volume of hydrates increases and this includes the ca lcium aluminate hydrates, thus early gypsum depletion could be advantageous in th at it will result in a high degree of C3A hydration within the first day. Accordingly, a relatively small excess level of SO3 could retard this process and therefore lowers 1-day streng ths, although it need not necessarily produce any obvious expansion. SO3 levels well beyond the optimum are usually necessary to produce deleterious expansions . Tang and Gartner , studied the influe nce of sulfate source on portland cement hydration. Cements were synthesized by ble nding a Type I low alkali portland cement clinker with sulfate salts. The authors deve loped a quantitative X-ray diffraction method to measure the rate at which C3A and C4AF phases in these cements were consumed in pastes hydrated at 23 C. It was found that the initial high rate of aluminate phase consumption was influenced significantly by th e rate of solubility of the sulfate source and especially the rate at which it released CaSO4 into solution. In terground gypsum was far more effective than inte rblended gypsum in controlli ng aluminate hydration. It was also found that a better control of the initial aluminate hydration generally led to higher 28-day paste strengths. For all sulfated cemen ts, ettringite was the only aluminate hydrate
27 detected by XRDA over the first 30 min. Ho wever, by 24 hr. it was in most cases accompanied by detectable amounts of an AF m phase which usually appeared to be monosulfoaluminate. The authors stated that although it has been suggested that high alkali cements require more SO3 for optimization because the alkali accelerates aluminate hydration, their data did not suppor t this as a general rule. They added that other factors, such as the solubility or Â“reac tivityÂ” of the sulfate source, se em to be just as important. Thus, much less aluminate reacts initially in the low alkali cements containing hemihydrate than in the comparable cement made only with gypsum, because hemihydrate is more soluble than gypsum. Af ter 30 min. however, the difference is lost, but this initial control of the aluminate phase hydration could lead to higher 28 day strength. More remarkably, the cement ma de by inter-grinding clinker and gypsum showed significantly lower aluminate consum ption than the interblend at all ages, and even after 24 hr. its aluminate was severely retarded. This indicates that distribution of sulfates is just as important as their chem ical form, and that results obtained with interblends can never completely represent what would occur in commercially produced cements made from clinkers containing alkali sulfates. The study showed that there is an approximate inverse correlation between initia l aluminate consumption and initial minislump of the pastes. The data for the paste compressive strength show the positive effect of soluble alkali on 1-day stre ngths, but at 7 and 28 days, there was little significant difference between the high and low alkali cement groups. The practic al consequence of this work lies in the observation that a reduced initial rate of aluminate phase consumption tends to improve both the workability of the fresh cement paste and the ultimate strength of the hardened product .
28 1.5.1 C3A and C4AF Hydration Considerable controversy still exists regard ing whether the formation of ettringite is the primary mechanism for early retardation of C3A. According to Kanare and Gartner , some researchers sugge st that retardation of C3A in mixtures with CaSO4 is due to ettringite formation, but in cements it is due to amorphous hydration products. They also suggest that monosulfoaluminate can form ev en when gypsum particles are present, and thus local dissolution and transport rates ar e an essential part of the retardation mechanism. Other researchers dem onstrated that the hydration of C3A and C4AF is retarded in solutions saturated with both lime and gypsum and that the ferroaluminate phase is the one being more st rongly affected. They suggest th at ettringite forms first on the C4AF surface, and when the sulfate concentr ation becomes low, monosulfoaluminate is produced, the ettringite layer is broken up, and C4AF hydration accelerates. Other researchers stated that evidence was found th at a retarding coating of ettringite and hydrous alumina exists which disappears at later stages as the sulfate is used up, permitting renewed acceleration. The findings re ported by several researchers indicate that gypsum retards C3A only when calcium hyd roxide is also present. In this case, ettringite forms protective films on the surface of the C3A as long as the solute concentration exceeds 14.5 mg CaSO4/liter. When all the gypsum is used to form ettringite, further reaction of C3A converts ettringite to a solid solution of monos ulfoalumiante and hydrated tetracalcium sulfoaluminate . Other researchers hydrated cements in the presence of isotopically labeled gypsum and then treated the products with lime water to remove residual gypsum. They concluded that a calcium sulfoaluminate coating formed initially on the grains as a
29 slightly permeable film, and that the mini mum amount of sulfate required to give a complete film coating to the grains was the amount which would properly retard the cement. Skalny and Tadros , i nvestigated the mechanism of retardation of the C3A dissolution rate by CaSO4 and concluded that th e retardation of the C3A-H interaction in the presence of CaSO4 is not primarily the result of the formation of an ettringite film on its surface. Upon cont act with water, C3A dissolves incongruently, leaving an aluminumrich layer on the surface. Calcium ions adsorb on this surface, producing positively charged particles even though the medium is highly alkaline. The formation of such a structure appears to minimize the active disso lution sites, and the dissolution rate of C3A decreases. In the presen ce of small amounts of CaSO4, sulfate ions adsorb on the positively charged particles, resulting in further reduction of the dissolution sites which would otherwise be available for hydroxyl ions to catalyze the dissolution. Collepardi et.al. , studied the hydration of tricalcium aluminate in the presence of lime, gypsum, or sodium sulfate. The results of his work confirmed that the mechanism of C3A hydration by gypsum is based on ettringite coating C3A grains. The authors concluded that ettrin gite crystals are formed by a through-solution mechanism but that they are assumed preferenti ally to form on the surface of C3A because of the catalytic action of the C3A surface on the nucleation of ettr ingite. The retardation due to gypsum is more effective in the presence of CH as ettringite crystals are smaller and can fit to the irregular shape of C3A grains better than larger ettr ingite crystals obtained in the absence of CH. The same hypothesis was proposed to explain why CH retards C3A hydration. The authors also found that afte r some hours, the ettringite coating C3A grains
30 is converted to monosulfate because of the consumption of gypsum, and C3A hydration is renewed. The results of this work did not c onfirm two other proposed mechanisms for the retardation of C3A hydration in the presence of CH and gypsum. The first mechanism was the one based on the formation of C4AHx impervious layer coating the C3A grains, this was due to the fact that et tringite was observed instead of C4AHx during the initial period of C3A hydration. Also expansion caused by the formation of monosulfate from C4AHx was not observed. The second mechanism was the one based on the adsorption of sulfate ions on C3A grains, this is because the resu lts of this work showed that Na2SO4 does not retard C3A hydration as gypsum does. Also the renewal of C3A hydration which was observed after some hours when C3A hydrates in the presence of gypsum or gypsum and CH, can not be explai ned by this mechanism. The Rilem Committee 68-MMH, in their report about the hydration of tricalcium aluminate and tetracalcium aluminoferrite in the presence of calcium sulfate , summarized the state of knowledge regarding the hydration of C3A and C4AF in the presence of calcium sulfate at ordinary temper atures. The report stat es that there does not appear to be general agreement as to the mechanism by which C3A hydration is retarded in the presence of calcium sulfate. Most of th e experimental evidence favors the view that retardation is associated with AFt formati on. However, recent el ectron optical studies have shown the formation of a hydration produc t layer of uncertain composition that may control the rate of early C3A hydration. A number of inves tigations have analyzed the kinetics of ettringite formation and have suggested a diffusionally controlled mechanism. However, reported activation en ergies are not consistent w ith a diffusionally controlled process. Some researchers pr oposed a mechanism in which a thin layer of ettringite
31 rapidly forms around the surface of a C3A particle by a topochemichal mechanism. As this layer thickens, pressure resu lting from the volume increase as C3A is converted to ettringite, develops and causes fissures in th e layer. As the fissures form, calcium and sulfate ions and wate r gain access to the C3A surface and the fissures are filled by the formation of additional ettringite. Finally, when sulfate ion is exhaus ted, ettringite begins to convert to AFm in the presence of unreacted C3A. Other researchers disagree with the fissure mechanism. They stated that they observed the development of hollow, tubular amorphous calcium sulfoaluminate hydrate filaments early in the hydrati on process, which they reported to form as a result of osmotic swelling of amorphous calcium sulfoa luminate layers in itially su rrounding the C3A particles. These observations support ot her proposed mechanisms in which semipermeable membranes form around the C3A grains. Water, diffusing through these membranes, eventually results in their rupture. This results in the mixing of an aluminaterich solution with calcium a nd sulfate ions and the precipit ation of ettringite. Another proposed mechanism suggests th at the retardation of C3A hydration results from the formation of alumina gel in the interfacial region between the C3A surface and an outer hydrate layer. Observations have shown that CH and gypsum, when mutually present, are more effective in retarding C3A hydration than is gypsum alone. Other studies indicated that the size of the ettringite crystals was reported to increase with increasing temperature and, as a consequence, become less effectiv e as a diffusion barrier. Studies have also indicated that while ettringite forms duri ng initial hydration of portland cement, it is slowly converted to an AFt so lid solution in which hydroxyl i on partially repl aces sulfate ion. Fe+3 may substitute for Al+3 in the ettringite structure and a solid solution between
32 the aluminum and iron AFt phases exists. Othe r researchers conclude d that at a pH of about 12.8, the AFt phase is unstable with respect to the AFm phase and gypsum. As for the ferrite phase, in general, the hydrat ion of the ferrite phase in the presence of calcium sulfate follows the same mechanistic path as that of C3A: early formation of AFt phase, followed by its conversion to AFm on exhaustion of gypsum. Kanare and Gartner , stated that Â“No researcher has presented unequivocal evidence for the formation of a crystalline ettr ingite coating which can seal the surfaces of C3A grains and thereby retard hydration. While th e formation of crystalline ettringite is coincident with retardation, the experimental evidence suggests that a slightly permeable, gel-like layer is formed on the surface of C3A grains surmounted by several layers of varying composition ranging from amorphous to cr ystalline. Sulfate a nd other ions are to be found in this coating and its compos ition will determine its permeability and propensity toward further re-crystallizationÂ”. The effect of temperature is another im portant factor reported by researchers who showed that the hydration of C3A is very sensitive to temp erature, and can be greatly accelerated by heating. They also showed th at at much higher temperatures than 20 C, gypsum is far less effective in controlling th e rate of C3A hydration. 1.5.2 Effect of Alkalis It is known that sulfate solubility increas es and calcium solubility decreases with increasing alkali ion concentra tion . Some researchers have pointed out that the fineness, distribution, and minera l form of the interground calcium sulfate affects the rate at which calcium and sulfate can go into so lution and interact with the other phases already present. The presence of alkalis wi ll cause a change in the rate of gypsum
33 consumption and thus affects the levels of cal cium and sulfate in solution. The alkalis in portland cement clinker are found partly as solu ble sulfates and partly as constituents of the clinker minerals. The effects of alkali s on the strength deve lopment properties of hydrating cement can be attributed to cha nges in the composition of the liquid phase mainly caused by the alkali sulfates, or to changes in the hydraul ic properties of the clinker minerals caused by the presence of alkalis in their latti ce structure . Osbaeck and Jons , and in order to eval uate the importance of the distribution of the alkalis, prepared a series of laboratoryburned clinker, differi ng only in content and distribution of alkalis, and have been ground to cement fineness at various gypsum addition levels. Their investiga tion indicated that it is the soluble alkalis that affect strength. They concluded that the influence of the content of alkalis in clinker on cement strength is dependent on the content of SO3 in clinker as well as the content of gypsum in the cement. Increased SO3 levels in clinker of the same al kali content imply that a greater fraction of the alkalis will be in an easily soluble form. The effect of this transfer of alkalis will generally be an increase of early strength and a decrease in late strengths. However, the effects are modified by the c ontent of gypsum in the cement. Thus the effect on early strength seems to be absent when gypsum content is higher than the optimum content. By analogy high gypsum conten ts tends to diminish the negative effect of alkalis on late strengths The authors explained the above by assuming that the presence of soluble alkalis causes entrai nment of air in mortar during mixing by promoting premature structure formation in the paste (precipitati on of gypsum, syngenite or ettringite). Such an indirect effect of alkalis will introduce a general reduction of strength at all ages, but mo st pronounced at late ages. Th e authors also explained the
34 favorable influence of high gypsum contents on late strength of cements rich in soluble alkalis to the fact that more gypsum is n ecessary to counteract th e faster removal of gypsum from the system caused by the alkalis. Thus the presence of gypsum to control the composition of the liquid phase is believ ed to be favorable for the hydration of C3S. However, too much gypsum will be detrimenta l to strength development even after 28 days. This could be due to a prolonged retardation of the C3A and C4AF phases and thus a reduced contribution fro m these phases to the total hydrate formation. The research work done by Lerch  showed similar trend and this led Lerch to state that: Â“ For cemen ts of the same C3A content, those high in alkalis react with gypsum more rapidly and require larger additions of gypsum than those low in alkalisÂ”. Various studies have shown that the presence of alkali metal sulfates, and of anions other than sulfate, also have a marked effect on the reaction kinetics of C3A hydration. It is of particular intere st that, as the ratio of alkali sulfate to calcium sulfate is increased, the effect is first to retard C3A hydration, and then, at higher al kali levels, to accelerate it very severely . The results of these i nvestigations, although they were done on pure C3A, they indicate that the optimum level of SO3 in a cement is likely to be strongly influenced by the presence of soluble al kalis. The mechanisms and kinetics of SO3 reactions in cements are not fully understood an d the data that exists leads to conflicting conclusions. As stated previous ly, the availability of calcium hydroxide and sulfate ions in solution at very early ages is an impor tant factor in determining the amount of C3A which reacts in the first few mi nutes. This amount of reacted C3A and of sulfate uptake during the first few minutes is very importa nt to the overall react ion kinetics and may affect the whole course of th e reactions at later ages, b ecause it determines how much
35 gypsum is remaining to keep the liquid phase saturated during the initial hydration of alite. Jelenic, et. al. , studied the influe nce of gypsum content on the hydration and compressive strengths of two commercial por tland cements having the same amount of alite, a similar amount of C3A, but different amounts of readily soluble alkalis and sulfates. This work was intended to find out how much the effects observed on pure systems may be significant for the determin ation of the optimum gypsum content in the case of portland cements, especially in the pr esence of alkali sulfates which cause a higher concentration of SO4 2ions in the liquid phase of portland cement pastes. The effects observed on the strength of their samp les were in accordance with those observed by Lerch , mainly for a clinker rich in alite and C3A, and low in alkalis; the maximum strength shifts to the lower values of SO3 as the hydration proceeds. On the other hand for a clinker high in alite and moderately high in C3A but containing a considerable amount of readily soluble alkalis and sulfates, the sh ift of the maximum to the higher values of SO3 was noticeable. They found that gypsum had a remarkable effect in accelerating alite hydration at all ages for the clinker rich in so luble alkalis and sulfates For the clinker that is low in alkalis, the accelerating effect was noticed up to the age of 28 days but not at 90 days. They concluded that the difference in the optimum addition of SO3 is influenced not only by the difference in C3A content, but also by the difference in alkalis, as well as in the amount of the primary present water so luble sulfates. They al so found that only a part of the SO3 content (1.14%) is used in forming ettr ingite at the age of 1 day, and this is much lower than 2.58% which is the total SO3 value determined by chemical analysis, and since the calcium aluminate monosulfate hy drate has not been detected at 2% of SO3
36 added, they concluded that the rest of the total SO3 amount (i.e. ~ 1.4%) has been incorporated into the C-(S, )-H. They also conclude d that the amount of SO3 incorporated in the C-(S, )-H increases with the increase of total SO3, thus giving a badquality gel, and consequently lower compre ssive strengths. The two examined clinker samples behaved differently with respect to ettringite formation and the authors suggested that this was caused by the differe nce in the composition of the liquid phase, i.e. by the presence of a higher amount of r eadily soluble alkalis and higher amount of glassy phase in the clinker sample rich in al kalis. The authors also noticed differences in the morphological characteristics of the ettr ingite formed by the two different clinker samples. There were also indications that the nature of ettr ingite coatings on C3A grains has been influenced by the amount of the gypsum added; at higher SO3/C3A ratios the retardation of C3A hydration is more effective, most probably because a faster reaction forms a much denser coating. Jawed and Skalny , in their review about alkalis in cement, states that there is a worldwide trend towards higher alkali content in cements. This is due to various factors including changes in clinke rs pyprocessing technology, call for energy conservation, the limited availability of low-alkali raw material s, and tightened envir onmental restrictions, and the use of coal as the primary fuel sour ce. The authors stated that the presence of sulfur leads to reduction of alkali volatility during clinker formation and that alkali compounds in clinker can be divided into three main groups: a)Alkali sulfates; b)Alkali aluminates and al uminoferrites; c)Alkali silicates. Also in some cases, alkalis may occur in the form of carbonates. The authors added that the clinkerÂ’s SO3 makes prior demand on the alkalis, the resulting quantity of alkali sulfate is
37 determined by the ratio of total clinker sulfat e to total alkali. The remaining sulfate forms calcium sulfate. After allocating alkalis to sulf ate, the remainder appears to be distributed between the silicates, aluminates, and al uminoferrites. The rules governing the quantitative division of alkalis between silicates, aluminates and ferrites are not yet clear but, it is known that aluminates and ferrite s accommodate about half or more of the available alkalis. The introduction of alkalis into clinker mine rals modifies their crystal structure which in turn, can change their hydraulic reactivity. Some researchers have shown that introduction of alkalis into C3A modifies its normal cubic form to orthorhombic. This orthorhombic form of C3A is formed with a minimum alkali content of 2.8% equivalent Na2O or 1.8% equivalent K2O . The authors also pointed out the effect of SO3 on clinker in the presence of alkalis. A positive effect of gypsum addition to alkali containing raw materials on the formati on of clinker minerals has been reported by some researchers. The clinkering te mperature decrease d, the amount of C3S increased and a positive effect on the binding of CaO was noticed. The presence of alkali sulfates resulted in well developed alite and belite crystals. Different alkali containing raw materials required different amounts of gypsum It is claimed that the use of proper amount of gypsum resulted in higher kiln out put and reduction of fu el consumption . Some other research work showed that almost 17% of the total C3A in cement reacts during the first 20 minutes of hydration while ot her results from another research work in which gypsum was substituted by hemihydrate (whi ch has a high initial solubility) at the same total SO3 level in cements showed that at the first 20 minutes of hydration a decrease in the amount of formed ettringite and of consumed gypsum happened. Also a small but significant increase in the time to the second C3A heat peak happened. These
38 results contradict the previous ones b ecause they imply that the amount of C3A consumed in the initial period is reduced by the increase d sulfate availability resulting from using the hemihydrat . Also previous research work has shown that C3A hydration rates in cement pastes differ considerably from those observed in the C3A-CaSO4-Ca(OH)2 system, mainly with respect to the time of occurren ce of the second heat peak whic h is usually associated with depletion of gypsum . In pure C3A-CaSO4-Ca(OH)2 system, and at a /A = 0.7, the retardation of the second peak is expe cted to be about 50-70 hours at 20-25 C. Since most optimized Type I portland cements ( /C3A = 0.7-1.5) does not show any heat peak after 18 hrs, regardless of alkali content, this implies that sulfate uptake in cements is more rapid than what is expected in pure C3A/C4AF systems. One explanation is that it is most likely a result of its uptake by the CSH . The importance of CSH in gypsum depletion is a subject worthy of further ex amination. It was found that CSH can contain as much as 20% SO3 relative to SiO2 in cement pastes after 1day hydration at typical w/c ratios. Although these same past es show a decreasing level of SO3 substitution in CSH reducing to about 10% relative to SiO2 at 28 days, the fact that there is typically twice CSH present at 28 days compared to 1 day, this shows that the same total amount of SO3 is contained at both ages and the c oncentration is reduced by dilution . Another interesting observation wa s that the concentration of Al2O3 and Fe2O3 in the CSH in the same pastes were increasing from 1 to 28 days. This implies that initially and while sulfate is still available to form ettrin gite, there is little initial dissolution of Al3+ or Fe3+ in CSH. Once all the sulfate has been c onsumed, and AFm phases begin to form in large amounts, Al3+ and Fe3+ become more soluble and can apparently diffuse into the
39 CSH more readily. The results of a resear ch work showed that the reaction of C3A with sulfate-rich CSH, or the reaction of gypsum with Al3+-rich CSH, tends to result in the formation of ettringite, in preference to monosulfoaluminate. Some researchers concluded from these results that CSH has a stronger tenden cy to dissolve Al3+ than it does to dissolve SO4 2-. This implies that a mixture of monosulfoaluminate plus excess CSH will tend to disproportionate towards ettringite plus substituted CSH . 1.5.3 Effects of Temperature The temperature during mixing and curing is ex pected to have a si gnificant effect on the optimum level of gypsum. This is due to the rapid increase in C3A hydration rate with increasing temperature. It is wo rthwhile to note that the time to the second heat peak at a molar /C3A ratio of 0.5 decreases from 60 hours at 15 C to 15 hours at 30 C, to 5 hours at 40 C, and to less than 1 hour at 70 C . It was found that cements can accommodate more SO3 at higher curing temperatures with no detrimental effects, possibly because ettringite becomes increasingly less stable with respect to monosulfoaluminate, and the formation of the latter phase is apparently not destructive to the matrix. Also, more Al2O3 and SO3 may enter the CSH during high temperature curing .
40 CHAPTER 2. EXPERIMENTAL PROCEDURE 2.1 Materials The materials used in this study consisted of cements, gypsum, sand, distilled water, sodium sulfate, and fly ash. 2.1.1 Cements Four types of cements were used in this study. They were randomly labeled as C, D2, E, and P cement. 126.96.36.199 Oxide Chemical Composition of the Cements The cements were analyzed for their oxide chemical composition in an external laboratory. The laboratory report of chemical an alysis stated that the samples were fused at 1000 C with Li2B4O7 and the oxide analysis was done by X-ray fluorescence spectrometry. This analysis meets the prec ision and accuracy requirements for rapid methods per ASTM C 114-00 Â“Test Methods for Chemical Analysis of Hydraulic CementÂ”. 188.8.131.52 Bogue Calculations Bogue formulas as stated in ASTM C 150-00 Â“Standard Specification for Portland CementÂ” were used to calculate the theoreti cal mineralogical composition of the cements.
41 2.1.2 Gypsum The gypsum used in this study was Terra Alba No.1 gypsum as it meets the requirements specified in ASTM C 452-95 Â“Standard Test Method for Potential Expansion of Portland-Cement Mort ars Exposed to SulfateÂ”. The SO3 content of the gypsum was determined by an ex ternal laborato ry and was found to be 46.5%. This value was used in calculating the percentage of cement and gypsum required to provide a mixture containing 3.0 and 3.6 mass % SO3 according to the formula given in ASTM C 452-95. 2.1.3 Sand The sand used in this study was ASTM Graded Sand furnished by U.S. Silica Company. The sand conforms to ASTM De signation C 778-00 Â“Standard Specification for Standard SandÂ”. The sand was delivered in bags of 50 lbs each and was oven dried before use. 2.1.4 Water All the water used in mixing of mortar a nd in preparing of sodium sulfate solution for this study was pure distilled water conf orming to Type IV of Specification D 1193. 2.1.5 Sodium Sulfate All the sodium sulfate used in this study was a certified ACS sodium sulfate anhydrous obtained from Fisher Scientific. 2.1.6 Fly Ash The fly ash used in this study was a Cla ss F fly ash with a LOI of 4.8%. It was sieved and the amount used was reconstituted from the fly ash that was retaining on different sieves.
42 2.2 Strength and Durability Tests In order to assess th e effect of the SO3 content of the cement on the durability and strength, mortar cubes and bars were prepared for the as-received cements as well as for the cements with an SO3 content of 3.0% and 3.6%. The mortar cubes were tested for their compressive strength both in saturated lime solution and in sodium sulfate solution. The mortar bars were stored in sodium sulf ate solution and their length was measured at certain ages in order to assess the ex pansion that was happening to the bars. 2.2.1 Mortar Cubes Mortar cubes were prepared for each type of the four as-received cements. The cubes were mixed in accordance to ASTM C 305-99 Â“Standard Practice for Mechanical Mixing of Hydraulic Cement Pastes and Mo rtars of Plastic ConsistencyÂ”. The mix proportions are listed in Tabl e 1. All the mixes had 1 part cement to 2.75 parts of sand by mass. The water/cement ratio by mass was 0.485 Table 1. Mix Proportions for the Mortar Cube s Prepared from the As-Received Cements 9 cubes mix Cement g 740 Sand g 2035 Water ml 359 The cubes were molded according to the ASTM C 109-99 Â“Standard Test Method for Compressive Strength of Hydraulic Cement Mortars (Using 2-in. Cube Specimens)Â”.
43 For each type of cement, 39 cubes were prepared for the first mix. Three cubes were tested immediately after de-molding at the age of 1 day and 18 of the remaining 36 cubes were stored in saturated lime solution while the other 18 were stored in sodium sulfate solution. Three cubes from each set were tested at the ages of 3, 7, 28, 90, 180, and 360 days respectively. The cubes were tested accordi ng to the ASTM C 109-99 also and the testing machine was an MTS 809 Axial/Torsional Test System. The sodium sulfate solution was a 5 weight percent solution and was prep ared according to the ASTM C 1012. The sodium sulfate solution was changed for the re maining cubes of each set after the testing was done at any age. Before placing the cube s in the new solution, the pH of the solution was measured with a pH meter to ensure that the pH was within th e range of 6.0 Â– 8.0 as per ASTM C 1012. The same procedure was repeated for a second mix where another 39 cubes were prepared for each type of cement. The cubes we re stored and tested exactly as the first mix and the average strength of the first a nd second mix at any age (i.e. the average strength of 6 cubes, 3 from each mix) was ta ken to represent the compressive strength at that age. Mortar cubes were also prepared for the cements with an SO3 content of 3.0% and 3.6%. As was the case for the as-received cemen ts, the first mix for each type of cement with an SO3 content of 3.0% consisted of 39 cubes. The SO3 content of the cement was increased by replacing part of the cement w ith Terra Alba gypsum. The percentage of cement and gypsum required to provide the mixture for each type of cement was calculated according to the formula give n in ASTM C 452-95. The mix proportions for
44 the cements with SO3 content of 3.0% are listed in Table 2 while those for the SO3 content of 3.6% are listed in Table 3. Table 2. Mix Proportions for the Mortar Cubes Prepar ed from the Cements with 3.0% SO3 Content Cement C Cement D2 Cement E Cement P Cement g 736.78 732.26 732.92 737.96 Gypsum g 3.22 7.74 7.08 2.04 Sand g 2035 2035 2035 2035 Water ml 359 359 359 359 Table 3. Mix Proportions for the Mortar Cube s Prepared from the Cements with 3.6% SO3 Content Cement C Cement D2 Cement E Cement P Cement g 726.62 722.16 722.81 727.79 Gypsum g 13.38 17.8435 17.19 12.21 Sand g 2035 2035 2035 2035 Water ml 359 359 359 359 In both cases, the mortar was mixed accordi ng to the procedure for mixing mortars of Practice C 305, with the excepti on that after placing the mi xing water in the bowl, the
45 gypsum was added and mixed at the slow speed for 15 seconds, then the mixer was stopped and the cement was added and the mixi ng continued as prescribed in Practice C 305. This was done according to the procedure as per ASTM C 452-95. Three cubes from each set of cement were tested at the age of 1 day and the remaining 36 cubes were stored half in sa turated lime solution and the other half in sodium sulfate solution. As was the case for the as-received cements, the cubes were tested at the ages of 3, 7, 28, 90, 180, and 360 da ys. The sodium sulfate solution was also replaced every time testing was done. On a different date, a second mix was prep ared for each type of cement (i.e. 39 cubes per each type of cement with an SO3 content of 3.0%). The second mix was also cured and tested under the same conditions as the first mix and the average value of the compressive strength for the two mixes was ta ken to represent the compressive strength at any age. The same procedure that was done for the as-received cements and for the cements with an SO3 content of 3.0% was repeated for the case of the 3.6% and again the average value of the compressive strength for the fi rst and second mixes for each type of cement was taken to represen t the compressive stre ngth of the 3.6% SO3 content mortar at any age. In all the cases, the 5 weight percent sodi um sulfate solution was changed at the ages of 3, 7, 28, 90, and 180 days for the remaining cubes of each set on the same day when testing was done. Again the pH of the soluti on was measured every time before placing the cubes to ensure that th e pH is between 6.0 and 8.0
46 2.2.2 Mortar Bars For each type of cement, mortar bars were prepared for the three cases of the asreceived cement, 3.0% SO3 content cement, and 3.6% SO3 content cement. The bars were prepared according to ASTM C 1012-95a Â“Sta ndard Test Method for Length Change of Hydraulic-Cement Mortars Exposed to a Sulf ate SolutionÂ”. The mix proportions for the mortar used to prepare the bars were the sa me as those used for preparing the cubes and are listed in Tables 1, 2, and 3. In all the cas es the mortar bars were cured in saturated lime solution until they attained a compre ssive strength of 20.0 1.0 MPa ( 3000 150 psi), as measured using cubes made of the same mortar, before the bars were immersed in the sodium sulfate solution. The apparatus us ed for the determination of the length change was according to the ASTM C 490-00 Â“Standard Practice for Use of Apparatus for the Determination of Length Change of Hardened Cement Paste, Mortar, and ConcreteÂ”. The initial reading for the length of the bars was taken immediately before they were immersed in the sodium sulfate solution. A digital comparator manufactured by Humboldt Co. (model H 3250) was used for measuring the length ch ange of the bars. After the initial r eading, the bars were measured at 7, 14, 21, 28, 56, 91, 105, 120, 150, 180 days, and then every 15 days after that The sodium sulfate solution was changed every time the bars were measured and the pH of the new solution was measured before immersing the bars to ensure that it is in the range of 6.0-8.0 All the sets of bars were repeated anothe r time on different dates to verify the trend of the expansion found in the first sets. The second sets were prepared, cured, and measured in the same way and on the same ages as the first sets.
47 In addition to all the above sets of bars, an additional set of bars was prepared for each type of cement by replacing 20% of the cement by fly ash. The fly ash used was a Class F fly ash with LOI of 4.8%. The SO3 content of the remaining 80% of the cement was increased to 3.6% by replacing part of the cement by Terra Alba gypsum. The mix proportions for the fly ash ba rs are listed in Table 4. Table 4. Mix Proportions for the Mortar Bars Prepared from the Cements With 20% Fly Ash Repla cement and 3.6 % SO3 Content of the Remaining Cement Cement C Cement D2 Cement E Cement P Cement, g 581.30 577.73 578.25 582.23 Gypsum, g 10.70 14.27 13.75 9.77 Fly Ash, g 148 148 148 148 Sand, g 2035 2035 2035 2035 Water, ml 359 359 359 359 The fly ash was sieved before use and the 148 grams required for each mix were reconstituted from the fly ash retained on different sieves according to Table 5.
48 Table 5. Proportions of Class F Fly Ash, LOI 4.8%, for 20% Replacement of Cement Sieve No. Weight (grams) 40 0.91 50 1.12 70 1.60 100 3.31 200 26.61 325 107.84 Pan 6.61 Total 148.00 After weighing the fly ash, it was added to the dry cement and mixed together in a glass beaker with a spatula for several minutes until a homogeneous mixture was obtained. After placing the water in the mi xing bowl, the gypsum was added and mixed with the water at the slow speed for 15 s econds. Then the mixer was stopped and the cement-fly ash mixture was added and the mixi ng continued as prescribed in Practice C 305. As was the case with the other sets of ba rs, the fly ash bars were also prepared according to the ASTM C 1012-95a, and after de-molding, the bars were cured in a saturated lime solution until they attained a strength of 20 1.0 MPa (3000 150 psi), as measured using cubes made of the same mort ar, before the bars were immersed in the
49 sodium sulfate solution. The Fly Ash bars were measured at the same ages as the other bars, and the solution was changed every time a set of bars was measured. 2.3 X-Ray Powder Diffraction Analys is of Mortar Cubes and Bars X-Ray powder diffraction analysis was done for the mortar c ubes that were tested at the age of 360 days. Pieces of the crushed cube after testing, were ground in a mortar for several minutes and then sieved using si eve number 325. The amount passing the sieve was collected in a vial and 0.45 grams of it we re weighed and placed in another vial. An amount of 0.05 grams of TiO2 were weighed and added to the vial containing the 0.45 g sample. The sample was mixed inside the vial with a spatula for approximately 5 minutes until a homogeneous mixture was achieved. The sample was then placed in a diffractometer. The diffractometer used for an alyzing the sample was a Phillips XÂ’Pert PW3040 Pro diffractometer. In a similar manner, XRD analysis was also done for the mortar bars at the age of 480 days. For each type of bars, a piece of a pproximately two inches was taken out from the middle of the bar. The bars were broke n using a hammer and a screwdriver. The outside perimeter of the 2 inch piece was sa wed off at a thickness of 2-3 mm, and the sawed pieces were soaked in acetone for an hour, then ground in a mortar and the same procedure as described above was done for sieving, mixing with TiO2 and placing in the diffractometer.
50 CHAPTER 3. RESULTS AND DISCUSSION 3.1 Characteristics of the As-Received Cements 3.1.1. Cements Fineness As was stated earlier in Chapter 1, the limit on the SO3 content of cement in various national standards ranges fr om 2.5 to 5.0% depending on cement fineness or composition or both. Since increasing the cement fineness will lead to a higher rate of hydration due to an increase in the surface area of the cement that would be in contact with water, the cement fineness plays an important role in determining the quantity of gypsum required to obtain proper retardation and the proper optimum SO3 content. Based on this consideration, it was decided to eliminate cemen t fineness as a variable in this work. All of the four types of cements used in this study, have very similar fineness, but vary in their C3S, C3A, and alkali contents. Table 6 lists the Blaine Fineness values of the as-received cements. Table 6. Blaine Fineness Cement C D2 E P Blaine Fineness (cm2/g) 3840 3880 3800 3820
51 The values range between 3800 Â– 3880 cm2/g with cement E having the lowest value and cement D2 having the highest although the diffe rence is not significant. Based on this, the four types of cements are expect ed to have similar reactivity. 3.1.2 Oxide Chemical Composition Table 7 lists the oxide chemical com position of the as-received cements as determined by X-ray fluorescence spectroscopy. Table 7. Oxide Chemical Composit ion of the As-Received Cements Cement Type C D2 E P Analyte Weight % Weight % Weight % Weight % SiO2 20.52 20.55 21.15 20.78 Al2O3 4.92 4.4 4.78 5.47 Fe2O3 3.7 3.61 3.76 4.15 CaO 64.31 64.6 64.41 63.14 MgO 1.71 2.47 0.95 0.85 SO3 2.81 2.54 2.58 2.88 Na2O 0.01 0.03 0.18 0.26 K2O 0.41 0.54 0.34 0.6 TiO2 0.27 0.22 0.33 0.32 P2O5 0.03 0.05 0.07 0.18 Mn2O3 0.04 0.05 0.03 0.03 SrO 0.04 0.02 0.12 0.05 Cr2O3 < 0.01 0.02 < 0.01 0.02 ZnO < 0.01 0.03 0.02 0.02 L.O.I. (950 C) 1.08 0.99 1.15 1.3 Total 99.83 100.12 99.84 100.04 Alkalies as Na2O 0.27 0.39 0.4 0.65 Free CaO 0.92 2.31 1.05 0.44
52 P cement has the highest SO3 content followed by C cement, while cements E and D2 have lower SO3 content. P cement also has the highest Na2O and K2O content and consequently the highest Na2O equivalent. Cement C has the lowest Na2O content and Na2O equivalent. Cements E and D2 have similar Na2O equivalent values although E has much higher Na2O content while D2 has a higher K2O content. Among the four cements, only the equivalent alkali content of cement P marginally exceeds the limit established by the ASTM C 150. Cement D2 has the highe st MgO content followed by cement C although the values are well below the lim it established in the ASTM C 150 (6%). Cement D2 also has the hi ghest level of free CaO. 3.2. Compound Composition 3.2.1. Mineralogical Composition Acco rding to Bogue Calculations Table 8 lists the mineralogical compositi on of all the cements according to Bogue calculations. Table 8. Mineralogical Composition According to Bogue Calculations Cement Type Compound C D2 E P C3S 60 65 57 48 C2S 14 10 18 23 C3A 7 6 6 7 C4AF 11 11 11 13 C3S/C2S 4.3 6.5 3.2 2.1 The C3S content of the cements varies between 48% for cement P and 65% for cement D2. Cements P and C have a C3A content of 7% while cements D2 and E has a slightly
53 lower value (6%). Cements C, D2, and E have a C4AF content of 11% while cement P has a slightly higher value of 13%. It is worth noting that as the C3S content of the cements increases, so does the C3S/C2S ratio. Cement D2 has the highest C3S/C2S ratio (6.5) while cement P has the lowest ratio of 2.1 According to the Bogue calculations for the compound composition, the four types of cements can be considered as ASTM Type I portland cements. 3.2.2. Mineralogical Composition According to Internal Standard Method and Rietveld Refinement Method In order to determine a more accurate mine ralogical composition of the four cements used in this study, both the Internal Sta ndard Method and the Rietveld Refinement Method were used. Table 9 shows the amounts of C3S, C3A, C4AF, and MgO in the as-received cements based on the Internal Standard Method . Table 9. Mineralogical Composition Base d on the Internal Standard Method Cement Type Compound C D2 E P C3S 70 63 58 55 Cubic C3A 3 3 4 6 C4AF 14 11 10 15 MgO 1 2 0 0
54 The results for the C3S content of the cements accordi ng to the internal standard method differed from those according to Bogue calculations. Although cement P still has the lowest C3S content, the new value is 55% compared to 48% according to Bogue calculations. The new value for cement E is ve ry close to the previous one, but for cements D2 and C, the internal standard me thod showed that cement C has the highest C3S content of 70% (60% according to Bogue) followed by cement D2 63% (65% according to Bogue). For the C3A content, only cement P had a valu e close to that determined through Bogue calculations, while all the othe r three cements had a much lower C3A content than that determined by Bogue calculations. The in ternal standard method showed that cement P has the highest C3A content, followed by cement E, while cements C and D2 have the lowest C3A content. The C4AF content of the cements according to the internal standard method is similar to that according to Bogue calcula tions. Cements P and C had a higher value and both were higher than those for D2 and E cements. Table 10 shows the results obtained by the Rietveld Re finement Method . The results of the Rietveld refine ment method are similar to thos e of the intern al standard method. Cement P has the lowest C3S content and the highest C3A content. Cement C has the highest C3S content and the lowest C3A content. It is worth noting that the Rietveld refine ment method showed that cement E has the highest total calcium sulfates (3.6%) follo wed by cement C (2.8%). Gypsum was not found in cements C and D2, while insol uble anhydrite was only found in cement C (1.3%). Since the solubility of this form of anhydrite is extremely low, it is expected that
55 it will not participate in the hydration process and thus the actual total CaSO4 content for cement C can be considered to be 1.5%. This is similar to the values obtained for cements D2 and P. Although cements C and D2 have no gypsum, calcium sulfate is present in them in the form of Bassanite which is more soluble than gypsum. Table 10. Rietveld Refinement Resu lts for the As-Received Cements Cement Type Compound C D2 E P C3S 67 61 54 53 -C2S 15 19 25 23 Cubic C3A 2 3 4 8 C4AF 14 12 13 11 Gypsum (Ca2SO4.2H2O) --2.0 1.1 Bassanite (Ca2SO4. 0.5 H2O) 1.5 1.6 1.6 0.7 Insoluble Anhydrite (Ca2SO4) 1.3 ---Magnesite (MgCO3) ---1.8 Periclase (MgO) 0.6 1.8 --Dolomite (CaMg(CO3)2) ---0.8 Portlandite (Ca(OH)2) -1.2 --C3S/C2S 4.5 3.2 2.2 2.3 Total Ca2SO4 2.8 1.6 3.6 1.8
56 3.3 Strength and Durability The effect of the SO3 content of the cements on thei r strength and durability was determined in this study by measuring the comp ressive strength of mortar cubes prepared for the As-Received cements as well as for the cements with an SO3 content of 3.0% and 3.6%, and also by measuring the length change of mortar bars prepared for the same mixes and stored in sodium sulfate solution. 3.3.1 Compressive Strength of Mortar Cubes Sets of cubes for all the mixes and for the four types of cements were prepared and stored in both saturated lime solution and s odium sulfate solution. The cubes were tested for their compressive strength at several ag es. Figure 1 shows the compressive strength verses the SO3 content for C cement cubes stored in saturated lime solution while Figure 2 shows the same relationship but for the cubes stored in sodium sulfate solution. In saturated lime solution C-3.0 cubes had the highest compressive strength for all the ages up to 28 days, although the C-3.6 had a ve ry close strength at the age of 7 days. In sodium sulfate solution, the C-3.0 had the hi ghest strength only at the age of 3 days, while the C-3.6 had the highest strength at 7 days and the as-received at 28 days, although the differences between the strength s for the as-received and 3.0 and 3.6 at the same age were not that significant. Figures 3 and 4 shows the compressive stre ngth verses age for the as-received C cement as well as the cement with SO3 content of 3.0% and 3.6% in both lime and sodium sulfate solutions. The strength of C ce ment in lime was not affected by increasing the SO3 content with the 3.0 having a slightly higher strength.
57 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 2345 SO3 Content %Compressive Strength (psi) C 1 day in Lime C 3 days in Lime C 7 days in Lime C 28 days in Lime Figure 1. Compressive Strength verses SO3 Content for C Cement in Lime 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 2345 SO3 Content %Compressive Strength (psi) C 3 days in Sulfate C 7 days in Sulfate C 28 days in Sulfate Figure 2. Compressive Strength verses SO3 Content for C Cement in Sulfate
58 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0100200300400 TIME (DAYS)COMPRESSIVE STRENGTH (PSI) C-AR C-3.0 C-3.6 Figure 3. Compressive Strength verses Age for C, C-3.0, and C-3.6 in Lime 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0100200300400 TIME (DAYS)COMPRESSIVE STRENGTH (PSI) C-AR C-3.0 C-3.6 Figure 4. Compressive Strength verses Age for C, C-3.0, and C-3.6 in Sulfate
59 In all the three cases for C cement, the st rength keeps on increasing up to the age of 180 days with the value of the increase in th e strength between the ages of 180 days and 28 days being 390 psi for the as-receive d, 490 psi for the 3.0 and 540 psi for the 3.6. Only a slight increase in strength happens after the age of 28 da ys and this can be explained by the fact that C cement has a very high C3S content, which is responsible for the high early strength, and a low C2S content, which governs later-age strength development. In sodium sulfate soluti on, and for all the thr ee cases, the strength increases only up to the age of 90 days and th en starts dropping. The value of this drop between the ages of 180 days and 28 days wa s 440 psi for the as-rece ived and 720 psi for the C-3.0. For the C-3.6 and although a drop happ ened after 90 days, th e strength at 180 days was slightly more than that at 28 da ys by 110 psi. The strength of the C as-received in lime at the age of 180 days was higher than that in sulfate at the same age by 1150 psi while the same difference for the case of the C-3.0 was 1640 psi and for the C-3.6 was 730 psi. A big drop in strength of 2320 psi happened to the C-3.6 cubes in sodium sulfate exposure between the ages of 180 days and 360 days while the drop that happened for the C-3.0 between the same two ages was 360 psi onl y. Thus a big difference in the strength was observed between the CS-3.0 and CS-3.6 at 180 and 360 days in sulfate. Also the difference in the strength between the C-3.0 in lime and sulfate at the age of 360 days was 2010 psi while the same difference for the C-3.6 was 2970 psi. Thus it is clear that increasing the SO3 content of C cement has a very de trimental effect on the strength in sodium sulfate exposure but th is effect only becomes obvious at late ages. It is worth noting that C cement has the highest C3S content according to both the Internal Standard Method and the Rietveld Refine ment Method, and also has th e lowest alkali content.
60 The compressive strength verses the SO3 content for cement D2 in both lime and sodium sulfate solutions are shown in figures 5 and 6. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 2345 SO3 Content %Compressive Strength (psi) D2 1 day in Lime D2 3 days in Lime D2 7 days in Lime D2 28 days in Lime Figure 5. Compressive Strength verses SO3 Content for Cement D2 in Lime 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 2345 SO3 Content %Compressive Strength (psi) D2 3 days in Sulfate D2 7 days in Sulfate D2 28 days in Sulfate Figure 6. Compressive Strength verses SO3 Content for Cement D2 in Sulfate
61 In the saturated lime solution, the As-Receive d D2 had the highest strength at 1 day and at 28 days while the D2-3.0 had the highe st strength at 3 days. The D2-3.6 had the highest strength at 7 days. In the sodium sulfate solution, the D2-3.0 had the highest strength at the ages of 3, 7, and 28 days. Fi gures 7 and 8 shows the compressive strength verses age for the D2 As-Received, D2-3.0, and D2-3.6 in both lime and sulfate solutions. In lime solution and for the three cases, the strength keeps on increasing up to the age of 180 days. The value of the incr ease in strength for the as-received case between the ages of 180 days and 28 days wa s 650 psi while the increase for the D2-3.0 between the same ages was 1350 psi and for the D2-3.6 880 psi. In the sodium sulfate solution, and as it was the case of the C cement, for all the three cases of the D2 cement the strength keeps on increasing up to the age of 90 days and then starts dropping. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0100200300400 TIME (DAYS)COMPRESSIVE STRENGTH (PSI) D2-AR D2-3.0 D2-3.6 Figure 7. Compressive Strength verses Age for Cement D2, D2-3.0, and D2-3.6 in Lime
62 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0100200300400 TIME (DAYS)COMPRESIVE STRENGTH (PSI) D2-AR D2-3.0 D2-3.6 Figure 8. Compressive Strength verses Age for Cement D2, D2-3.0, and D2-3.6 in Sulfate Despite this drop, the value of the strength of the as-received was higher at 180 days than that at 28 days by 110 psi and for the case of the D2-3.0 was higher by 280 psi while for the D2-3.6, it was higher by 350 psi. At the ag e of 180 days, the difference in the strength between the lime and sulfate conditions was 510 psi while that for the D2-3.0 was 540 psi and for the D2-3.6 was 150 psi only. On the ot her hand, the D2-3.6 in sulfate that had the lowest difference in the strength between the lime and sulfate exposures at the age of 180 days, had a large drop of 970 psi at the age of 360 days with the D2-3.0 having a higher strength. D2 cement has the second highest C3S content (61% according to the Rietveld Refinement Method) and the second highest K2O content (0.54%), but has the lowest SO3 content (2.54%). The D2 cement handled the increase in its SO3 content in a better way than the C Cement, but again the detrimen tal effect was only obvious at late ages.
63 The relationship between the compressive strength and the SO3 content for cement E in lime and sodium sulfate solutions is show n in Figures 9 and 10. In the lime solution, the E As-Received had the highest strength at al l the ages of 1, 3, 7, and 28 days. In the sodium sulfate solution, the E As-Received only had the highest strength at the ages of 3 and 7 days, while the E-3.0 had the highest st rength at 28 days. In the saturated lime solution the strength of the th ree cases of E cement also ke eps on increasing up to the age of 180 days as was the case with the C a nd D2 cements. The increase in the strength between the ages of 180 days and 28 days wa s higher in both the cases of E-3.0 and E-3.6 than it was for the as-recei ved case. The value of the in crease was 1950 psi for E-3.0 and 1220 psi for the E-3.6 while it was only 900 psi fo r the as-received cemen t. In the sodium sulfate exposure and unlike both the C and D2 cements the E cement, and for its three cases, kept on increasing in its strength up to 180 days without any drop at 90 days. This is shown in figures 11 and 12 which show the relationship between the compressive strength and age for the E cement As-Received, E-3.0, and E-3.6 cubes cured in both saturated lime and sodium sulfate solutions. The increase in the strength between the ages of 180 days and 28 days for the E as-received in sulfate was 1610 psi while that for the E3.0 was 1680 psi and for the E-3.6 was 1250 psi only. At the age of 180 days, the difference in the strength between the lime a nd sodium sulfate exposures was only 70 psi for the E-3.0 and 660 psi for the E-3.6. On th e other hand, for the as-received case, the strength in sulfate exposure wa s slightly higher by 80 psi th an that in lime exposure. At the age of 180 days, the streng th of the E-3.6 in sulfate was much lower than the E-3.0.
64 After that age, both the E-3.0 and E-3.6 in su lfate had a drop in st rength, and although the drop for the E-3.0 was greater than that that of the E-3.6, it still had the higher strength. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 2345 SO3 Content %Compressive Strength (psi) E 1 day in Lime E 3 days in Lime E 7 days in Lime E 28 days in Lime Figure 9. Compressive Strength verses SO3 Content for Cement E in Lime 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 2345 SO3 ContentCompressive Strength (psi) E 3 days in Sulfate E 7 days in Sulfate E 28 days in Sulfate Figure 10. Compressive Strength verses SO3 Content for Cement E in Sulfate
65 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0100200300400 TIME (DAYS)COMPRESSIVE STRENGTH (PSI) E-AR E-3.0 E-3.6 Figure 11. Compressive Strength verses Age for Cement E, E-3.0, and E-3.6 in Lime 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0100200300400 TIME (DAYS)COMPRESSIVE STRENGTH (PSI) E-AR E-3.0 E-3.6 Figure 12. Compressive Strength verses Age for Cement E, E-3.0, and E-3.6 in Sulfate
66 It is worth noting that cement E has the second lowest C3S content and the highest C2S content (according to the Ri etveld Refinement Method). This explains its low early strength and high late strength. Ceme nt E also has the second highest Na2O content and accordingly the second highest Na2O equivalent. The relationship between the compressive strength and the SO3 content for cement P is illustrated in Figures 13 and 14. In the saturated lime solution, th e P-3.0 had the highest strength from the age of 1 day up to the age of 28 days. The P As-Received and P-3.6 had very similar strengths during th at period with the 3.6 having a s lightly higher strength. In the sodium sulfate solution, the P-3.0 also had the highest strength up to the age of 28 days. Figures 15 and 16 illustrate the relati onship between the comp ressive strength and age for cement P, P-3.0, and P-3.6 cured in bo th lime and sodium sulfate solution. Again in lime solution, the strengths of both the P cement and P-3.6 keep on increasing up to the age of 180 days, while the P-3.0 had a slight dr op in strength after 90 days. The increase in the strength between the ages of 180 days and 28 days in lime was 1060 psi for the asreceived, and was 810 psi for the P-3.6 case. Fo r the P-3.0 the value of the strength at 180 days was equivalent to that at 28 days. The increase in the strength of the P-3.6 in lime between the ages of 180 days and 360 days was also greater than that of the P-3.0 but the final strength of the P-3.0 was slightly higher th an that of the P-3.6. In the sodium sulfate solution the strengths of the P as-received a nd P-3.0 keep on increasi ng up to the age of 90 days and then begins to drop, while that of the P-3.6 keeps on increasing up to the age
67 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 2345 SO3 Content %Compressive Strength (psi) P 1 day in Lime P 3 days in Lime P 7 days in Lime P 28 days in Lime Figure 13. Compressive Strength verses SO3Content for Cement P in Lime 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 2345 SO3 Content %Compressive Strength (psi) P 3 days in Sulfate P 7 days in Sulfate P 28 days in Sulfate Figure 14. Compressive Strength verses SO3 Content for Cement P in Sulfate
68 of 180 days without any drop. The strength of th e P as-received in sulfate at 180 days was less than that at 28 days by 280 psi and fo r the P-3.0 was less by 570 psi. On the other hand an increase in strength of 670 psi happe ned for the P-3.6 in sulfate between the same ages. Also the difference in the strength at 180 days between the lime and sodium sulfate exposures were much less for both the P-3.0 and P-3.6 than it was for the P as-received. The difference for the P as-received was 550 psi while the difference for the P-3.0 was 420 psi and for the P-3.6 was only 40 psi. Also the difference in the strength between lime and sulfate at 360 days was approximately similar at 1000 psi for both the P-3.0 (1240 psi) and P-3.6 (1030 psi). P cement has the lowest C3S content and the second highest C2S content. This explains the low early strength of 1 and 3 da ys and the large increase in strength that happened after 28 days. On the other hand, P cement has the highest Na2O, the highest K2O, and consequently the highest Na2O equivalent. This could explain the continuous increase in the strength of the P-3.6 without any drop up to the age of 180 days where it got a much higher strength than both the as-received and the P-3.0 in sulfate.
69 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0100200300400 TIME (DAYS)COMPRESSIVE STRENGTH (PSI) P-AR P-3.0 P-3.6 Figure 15. Compressive Strength verses Age for Cement P, P-3.0, and P-3.6 in Lime 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0100200300400 TIME (DAYS)COMPRESSIVE STRENGTH (PSI) P-AR P-3.0 P-3.6 Figure 16. Compressive Strength verses Ag e for Cement P, P-3.0, and P-3.6 in Sulfate
70 Figure 17 illustrates the relationship between the compressive strength of the asreceived cements cured in saturated lime so lution verses age, while figures 18 and 19 shows the same relationship for the cements with an SO3 content of 3.0 and 3.6% respectively. In all the three cases, E cement had th e highest strength with the E-3.0 having a slightly higher value. C cement got the second highest w ith the C-3.0 having a slightly higher strength, followed by cement D2 with the D2-3.0 also having a slightly higher value. P cement had the lowest strengt h with the P-3.6 having a slightly higher value. It was found that a big difference in th e compressive strength of the cubes in lime exists between cements E and P although they have very similar C3S and C2S contents. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 050100150200 AGE (DAYS)COMPRESSIVE STRENGTH (PSI) E-AR IN LIME P-AR IN LIME C-AR IN LIME D2-AR IN LIME Figure 17. Compressive Strength verses Ag e for the As-Received Cements in Lime
71 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 050100150200 AGE (DAYS)COMPRESSIVE STRENGTH (PSI) E-3.0 IN LIME P-3.0 IN LIME C-3.0 IN LIME D2-3.0 IN LIME Figure 18. Compressive Strength verses Age for the Cements with 3.0% SO3 Content in Lime 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 050100150200 AGE (DAYS)COMPRESSIVE STRENGTH (PSI) E-3.6 IN LIME P-3.6 IN LIME C-3.6 IN LIME D2-3.6 IN LIME Figure 19. Compressive Strength verses Age for the Cements with 3.6% SO3 Content in Lime
72 Most of the cements exposed to the sodium sulfate environment had lower strengths than their equivalents that were cured in sa turated lime solution especially at the late ages. Figures 20, 21, and 22 illustrates the rela tionship between the compressive strength and age for the as-received cements, 3.0% and 3.6% SO3 content cements respectively, all cured in sodium sulfate solution. As was the case in lime solution, E cement had the highest strength in all the cases with the E-3.0 and the as-received having slightly higher late strengths than the E-3.6 cement. D2 ceme nt had the second highest strength with the D2-3.0 and D2-3.6 having a better late stre ngth also. Although C and D2 cements had relatively close strengths in the as-received condition, this was not the case in the doped cements as the C-3.0 and C-3.6 had a large drop in their late strengths. It was found that the largest drop in strength between the lim e and sulfate exposures occurred for the C cement and especially for the C-3.6 case were the drop was ~ 3000 psi at 360 days. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 050100150200 AGE (DAYS)COMPRESSIVE STRENGTH (PSI) E-AR IN SULFATE P-AR IN SULFATE C-AR IN SULFATE D2-AR IN SULFATE Figure 20. Compressive Strength verses Age for the As-Received Cements in Sulfate
73 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 050100150200 AGE (DAYS)COMPRESSIVE STRENGTH (PSI) E-3.0 IN SULFATE P-3.0 IN SULFATE C-3.0 IN SULFATE D2-3.0 IN SULFATE Figure 21. Compressive Strength verses Age for the Cements with 3.0% SO3 Content in Sulfate 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 050100150200 AGE (DAYS)COMPRESSIVE STRENGTH (PSI) E-3.6 IN SULFATE P-3.6 IN SULFATE C-3.6 IN SULFATE D2-3.6 IN SULFATE Figure 22. Compressive Strength verses Age for the Cements with 3.6% SO3 Content in Sulfate
74 Although P cement had the lowest strength in sulfate for the as-received case, increasing the SO3 content to 3.6% seems to increase the durability of the cement against sulfate attack. This can be s een by the fact that a drop in strength between the ages of 90 and 180 days did not happen for the P-3.6 as was the case for both the as-received and P3.0. Also, the drop in strength at 360 days fo r the P-3.6 cement was less than that for D23.6 cement, thus bringing both P-3.6 and D23.6 to a very close strength at 360 days of exposure to sodium sulfate solution. Figures 23 and 24 illustrates the difference in the strength between the 3.0 % cements and 3.6 % cements at the age of 360 days in both saturated lime solution and sodium sulfate solution respectively. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 C DOPED CEMENT D2 DOPED CEMENT E DOPED CEMENT P DOPED CEMENTCOMPRESSIVE STRENGTH (PSI) 3.0% SO3 AT 360 DAYS LIME 3.0% SO3 AT 360 DAYS SULFATE 3.6% SO3 AT 360 DAYS LIME 3.6% SO3 AT 360 DAYS SULFATE Figure 23. Compressive Strength of Dope d Cements at 360 Days in Lime and Sulfate
75 0 1000 2000 3000 4000 5000 6000 7000 8000 C DOPED CEMENT D2 DOPED CEMENT E DOPED CEMENT P DOPED CEMENTCOMPRESSIVE STRENGTH (PSI) 3.0 % SO3 AT 360 DAYS 3.6% SO3 AT 360 DAYS Figure 24. Compressive Strength of Dope d Cements at 360 Days in Sulfate It can be seen from the figures that increasing the SO3 content of the cements did not greatly affect the compressive strength in lime exposure, but this was not the case in sodium sulfate exposure. C cem ent which had the highest C3S content and the lowest alkali content was greatly affected by increasing its SO3 content. On the other hand, the cements high in its C3A content and alkali content like cement P got a beneficial effect from this increase. Cements D2 and E and due to their higher alkali content relative to cement C were able to handle the increase in a better way. Cement E also had a lower C3S content and a higher C2S content.
76 3.3.2 Expansion of Mortar Bars in Sodium Sulfate Solution The length change of mortar bars prepared for the as-received cements as well as for the cements with an SO3 content of 3.0% and 3.6% was m easured in order to assess the expansion of the bars that were cured in sodium sulfate solution. Figure 25 illustrates the expansion of the bars prepared from the as-received cements. 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 050100150200 AGE (DAYS)EXPANSION % C CEMENT ASRECEIVED P CEMENT ASRECEIVED E CEMENT ASRECEIVED D2-CEMENT ASRECEIVED Figure 25. Expansion of Bars Prepa red with the As-Received Cements It can be seen from the figure that the e xpansion of the mortar bars can be divided into two stages. The first stage is characte rized by a relatively minor expansion and can be referred to as the induc tion period, while the second stag e is characterized by a rapid increase in expansion which continues at th e same rate until failure. P cement had the shortest induction period of 105 days with an expansion of 0.06% at that age. After that the expansion started to incr ease dramatically and at 180 days of exposure to sodium sulfate solution, the P bars expanded by 0.293%. P bars had the largest expansion at 180 days. It should be noted that P cement had the highest C3A content and the highest alkali
77 content. C and D2 bars had the same inducti on period of 120 days w ith an expansion of 0.063% for C and 0.068% for D2. C bars had the second highest expansion followed by D2 bars. At 180 days, the expansion of C bars was 0.169%, while that for D2 bars was 0.145%. Although C and D2 cements had approximately the same low C3A content, both had a high C3S content with the C having the highest C3S content. The induction period of the E cement lasted much more longer than the other three cements and can be estimated to be 210 days with an expansion of 0.06%. The final expansion of the E bars at 180 days (0.044%) was way less than the othe r three cements and this made it difficult to exactly locate the end of the induction period. The expansion was much lower although E cement had a slightly higher C3A content than both C and D2 cements according to Rietveld Refinement Method. It is worth noting that E cement had the second lowest C3S content, the highest C2S content, and the highest total CaSO4 content of 3.6% with 2.0% of it in the form of gypsum. The expansion trend of the four cements suggests dividing them into three categories. The first one is that of high expansion and includes P cement, followed by medium expansion (C and D2 cements), a nd finally low expansion which includes E cement.
78 Figure 26 illustrates the expans ion of the C, C-3.0, C-3.6, and C-FA bars cured in sodium sulfate solution. 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 050100150200 AGE (DAYS)EXPANSION % C AS-RECEIVED C-3.0 C-3.6 C-FA Figure 26. Expansion verses Age for the C, C-3.0, C-3.6, and C-FA Bars in Sulfate Increasing the SO3 content of cement C led to an in crease in the expansion of the bars exposed to sodium sulfate environment. While the expansion of C bars at 180 days was 0.169%, the expansion of the C-3.0 bars was 0.179% and that of the C-3.6 bars was 0.342%. Although the increase in expans ion was not significant at early ages up to 105 days, a 102% increase in expansion happened for th e bars at the age of 180 days due to increasing the SO3 content of the cement from 2.81% to 3.6%. This increase in the expansion explains the results obtained for the strength of the cubes cured in sodium sulfate solution where the streng th was not significantly affected in the early ages, but a significant drop happened at late r ages. Also, the big differe nce in the strength observed
79 between the CS-3.0 and CS-3.6 at 180 and 360 days, can be explained by the big difference in the expansion of the C-3.0 and C-3.6 bars at these ages. It can also be seen from the figure that by replacing 20% of the C cement by fly ash, the expansion of the bars was lowered dramatically to 0.064% at 180 days. This huge drop in the expansion happened although the SO3 content of the remaining 80% of the cement was increased to 3.6%. Thus, incorporat ing Fly Ash in the mix greatly increased the durability of the cement against sulfat e attack despite the increase in the SO3 content. The expansion of D2, D2-3.0, D2-3.6, D2-4.0, and D2-FA bars in sodium sulfate solution is illustrated in Figure 27. 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 050100150200 AGE (DAYS)EXPANSION % D2 AS-RECEIVED D2-3.0 D2-3.6 D2-4.0 D2-FA Figure 27. Expansion verses Age for the D2, D2-3.0, D2-3.6, D2-4.0, and D2-FA Bars in Sulfate In the case of the D2 cement which has a low C3A content (3%) but the second highest C3S content and the second highest K2O content (0.54%), increasing the SO3
80 content to 3.0% led to a d ecrease in the expansion of the bars. On the other hand, increasing it further to 3.6% and 4.0%, led to an increase in the expansion with a final expansion greater than the as-received cem ent. Thus it seems that the optimum SO3 content for minimum expansion in sodium su lfate solution for the D2 cement is around 3.0%. The results of the cubes compressive st rength showed that the D2-3.6 cubes had a higher strength than the as-received at the ages of 3, and 7 days. At the ages of 28 and 90 days the strength of the as-received was highe r, but at the age of 180 days the D2-3.0 had a higher strength which is in compliance with the expansion results. This indicates that the optimum SO3 content for maximum compressive stre ngth is not the same for all ages of the D2 cement and also i ndicates that this optimum can be different than that for minimum expansion of bars. Similar resu lts were found by other researchers, among them was Lerch  who concluded that mo st cements show different optimum gypsum requirements for different properties, such as st rength or shrinkage, at different ages. As was the case with the C cement, repl acing 20% of the D2 cement with Fly Ash led to a significant decrease in the expansion of the bars even to a much lower value than that of the D2-3.6 bars. While the D2 as-recei ved bars had a final expansion of 0.145% at the age of 180 days, the D2-FA bars had an expansion of 0.052% at the same age. The SO3 content of the remaining 80% of the cemen t was increased to 3.6%, and still we had this big difference in the expansion of the ba rs. This indicates the effectiveness of the pozzolanic materials in improving the durabilit y of mortar and conc rete against sulfate attack.
81 Figure 28 illustrates the expansion of E, E-3.0, E-3.6, E-4.2, and E-FA bars in sodium sulfate solution. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 050100150200 AGE (DAYS)EXPANSION % E AS-RECEIVED E-3.0 E-3.6 E-FA E-4.2 Figure 28. Expansion verses Age for the E, E-3.0, E-3.6, E-4.2, and E-FA Bars in Sulfate The expansion of the E cement bars was generally very low compared to the other cements and increasing the SO3 content led to an increase in the expansion, although this increase was not much significant. Although th e E as-received had the lowest expansion, it did not have the highest compressive streng th at all the ages. The E-3.0 had the highest strength at 28, and 180 days wh ile the strengths at 90 days we re very similar with the asreceived having a slightly higher value. As was the case with the D2 cement, it was found that also for the E cement, the optimum SO3 content for minimum expansion of bars is different than the optimum for maximum compressive strength, and that the optimum for maximum strength differs with age. In general, for both cements, increasing the SO3 content to 3.0% led to an increase in
82 the strength especially at the late ages. It is worth noting that E cement had the second lowest C3S content and the highest C2S content. It also had the second highest Na2O content and accordingly the second highest Na2O equivalent. The increase in strength with increasing the SO3 content of other cements was also reported by other researchers  These researchers stated that the beneficial effect of gypsum on the strength of the cement was attrib utable to the improved quality of the gel which was produced in the presence of gypsum Their study showed th at the addition of gypsum resulted in hydration pr oducts of greater average de nsity and it was suggested that this greater density br ought about the higher strength. The expansion of the E-FA bars was very si milar to that of the as-received cement, thus incorporating 20% Fly Ash in the mix did not cause a signifi cant decrease in the expansion, most probably because the expansio n of the E cement is generally very low. The expansion of the P, P-3.0, P-3.6, and P-FA bars in sodium sulfate solution is illustrated in Figure 29. Increasing the SO3 content of cement P to 3.0% caused a significant decrease in the expansion of the bars, and in creasing it further to 3.6% caused an additional significant decrease in the expansion. Thus P cement requires a higher amount of SO3 content to reach the optimum for lowest expansion. The P-3.6 with an SO3 content of 3.6%, which is higher than the specified limit of 3.0% had the lowest expansion under sodium sulfate exposure. P cement had the highest C3A content and the lowest C3S content. It also had the highest Na2O, the highest K2O, and accordingly the highest Na2O equivalent.
83 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 050100150200 AGE (DAYS)EXPANSION % P AS-RECEIVED P-3.0 P-3.6 P-FA Figure 29. Expansion verses Age for P, P-3.0, P-3.6, and P-FA Bars in Sulfate The result that cements with high C3A content and high alkali content requires larger additions of gypsum to reach the optimum was also reported by other researchers. The results for the compressive strength of th e P cubes in sulfate showed that the P-3.0 had the highest strength up to the age of 90 days, and it can be seen from Figure 29 that the expansion of all the sets of P bars does not vary signi ficantly up to that age. The strength results also showed that a large drop in strength happened to both the as-received and P-3.0 at the age of 180 days while an increase happened in the strength of the P-3.6. The expansion results matches the strength re sults as it is clear from Figure 29 that a significant difference in expansion exists betw een the 3 sets at the age of 180 days with the P-3.6 having the lowest expansion. The reason why the P-3.6 had a higher drop in strength between the ages of 180 and 360 days than the P-3.0 can not be explained from
84 the expansion results as the P-3.0 continued to have a much higher expansion than the P3.6 at that age. Figures 30, 31, and 32, compares the expansi on of the different types of cements used in this study in the three cases of 3.0% SO3 content, 3.6% SO3 content, and replacement of 20% of the cemen t with Fly Ash respectively. 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 050100150200 AGE (DAYS)EXPANSION % C-3.0 E-3.0 P-3.0 D2-3.0 Figure 30. Expansion verses Age for the Cements with 3.0% SO3 Content in Sulfate
85 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 050100150200 AGE (DAYS)EXPANSION % C-3.6 D2-3.6 E-3.6 P-3.6 Figure 31. Expansion verses Age for the Cements with 3.6% SO3 Content in Sulfate 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 050100150200 AGE (DAYS)EXPANSION % C-FA D2-FA E-FA P-FA Figure 32. Expansion verses Age for the Ce ments with 20% Replacement by Fly Ash
86 The figures show the beneficial effect of increasing the SO3 content of cement D2 to 3.0% and cement P to 3.6% on decreasing their expansion in sodium sulfate environment. The D2-3.0 had the lowest expansion of the 3.0% sets, while a significant decrease in the expansion of the P bars happened in the cas e of P-3.6 as compared to the as-received case. Figure 32 shows that incorporating 20% Fly Ash in the mix, had a significant effect in reducing the expansion of all the cements, bringing them to very close low values regardless of the differences that exist in their chemical and mineralogical composition. This result also happened despite the fact that the SO3 content of the remaining 80% of the cement was increased to 3.6%. 3.4 X-Ray Diffraction Anal ysis of Mortar Cubes Tables 11 and 12 illustrates the Relative In tensity Ratios for ettringite and gypsum as measured by the X-Ray diffraction analysis done on pieces of the crushed cubes at the age of 360 days for both the 3.0% sets and 3.6% sets respectively. Table 11. Relative Intensity Ratios of Ettringite and Gypsum for the 3.0% SO3 Content Mortar Cubes at the Age of 360 Days in Sulfate Mortar cubes at 360 days in Sulfate Solution C-3.0 D2-3.0 E-3.0 P-3.0 Ettringite Intensity Ratio 0.166 0.245 0.206 0.288 Gypsum Intensity Ratio 0.196 0.220 0.215 0.271
87 Table 12. Relative Intensity Ratios of Ettringite and Gypsum for the 3.6% SO3 Content Mortar Cubes at the Age of 360 Days in Sulfate Mortar Cubes at 360 days in Sulfate Solution C-3.6 D2-3.6 E-3.6 P-3.6 Ettringite Intensity Ratio 0.169 0.262 0.199 0.313 Gypsum Intensity Ratio 0.206 0.279 0.300 0.276 The relative intensity ratios are also illustrated in Figures 33 and 34. 0 5 10 15 20 25 30 35 40 EttringiteGypsumRelative Intensity Ratio x 100% C-3.0 D2-3.0 E-3.0 P-3.0 Figure 33. Relative Intensity Ratios of Ettringite and Gyps um for the 3.0% SO3 Content Mortar Cubes at the Age of 360 Days in Sulfate
88 0 5 10 15 20 25 30 35 40 EttringiteGypsumRelative Intensity Ratio x 100% C-3.6 D2-3.6 E-3.6 P-3.6 Figure 34. Relative Intensity Ratios of Ettringite and Gyps um for the 3.6% SO3 Content Mortar Cubes at the Age of 360 Days in Sulfate From the above figures, it can be seen that for both the 3.0% se ts and the 3.6% sets, P cubes contained the highest amounts of ettr ingite as expected, since P cement has a much higher C3A content than the other three cements. P cubes also contained the highest gypsum content for the 3.0% set. For the C, D2, and E cements, which have a relatively similar C3A content, D2 cubes contained the highes t ettringite conten t in both 3.0% and 3.6% sets. It is worth noting that cement D2 has the highest K2O content of the three cements (0.54%) and the lowest SO3 content (2.54%).
89 As for the gypsum content, P-3.0 cubes had the highest gypsum content for the 3.0% set, while E-3.6 cubes had the highest gypsum co ntent for the 3.6% set. It is worth noting that increasing the SO3 content of E cement from 3.0% to 3.6% caused a significant increase in the gypsum content while the ettr ingite content only increased slightly. 3.5 X-Ray Diffraction Anal ysis of Mortar Bars Tables 13 and 14 illustrates the Relative In tensity Ratios for ettringite and gypsum as measured by the X-Ray diffraction analysis done on pieces from the outside perimeter of bars at the age of 480 days for th e 3.0% sets and 3.6% sets respectively. Table 13. Relative Intensity Ratios of Ettringite and Gypsum for the 3.0% SO3 Content Mortar Bars at the Age of 480 Days in Sulfate Mortar bars at 480 days in Sulfate Solution C-3.0 D2-3.0 E-3.0 P-3.0 Ettringite Intensity Ratio 0.268 0.296 0.252 0.306 Gypsum Intensity Ratio 0.368 0.455 0.450 0.292
90 Table 14. Relative Intensity Ratios of Ettringite and Gypsum for the 3.6% SO3 Content Mortar Bars at the Age of 480 Days in Sulfate Mortar bars at 480 days in Sulfate Solution C-3.6 D2-3.6 E-3.6 P-3.6 Ettringite Intensity Ratio 0.330 0.281 0.268 0.308 Gypsum Intensity Ratio 0.407 0.495 0.534 0.469 Figures 35, 36, 37, and 38 illustrates also the rela tive intensity ratios of the sets of bars for the C, D2, E, and P cements respectively. 0 10 20 30 40 50 60 EttringiteGypsumRELATIVE INTENSITY RATIO x 100% C-3.0 C-3.6 Figure 35. Relative Intensity Ratios of Ettr ingite and Gypsum for the C-3.0 and C-3.6 Mortar Bars at the Age of 480 Days in Sulfate
91 0 10 20 30 40 50 60 EttringiteGypsumRELATIVE INTENSITY RATIO x 100% D2-3.0 D2-3.6 Figure 36. Relative Intensity Ratios of E ttringite and Gypsum for the D2-3.0 and D2-3.6 Mortar Bars at the Age of 480 Days in Sulfate 0 10 20 30 40 50 60 EttringiteGypsumRELATIVE INTENSITY RATIO x 100% E-3.0 E-3.6 Figure 37. Relative Intensity Ratios of Ettr ingite and Gypsum for the E-3.0 and E-3.6 Mortar Bars at the Age of 480 Days in Sulfate
92 0 10 20 30 40 50 60 EttringiteGypsumRELATIVE INTENSITY RATIO x 100% P-3.0 P-3.6 Figure 38. Relative Intensity Ratios of Ettrin gite and Gypsum for the P-3.0 and P-3.6 Mortar Bars at the Age of 480 Days in Sulfate 0 10 20 30 40 50 60 EttringiteGypsumRELATIVE INTENSITY RATIO x 100% C-3.0 BARS D2-3.0 BARS E-3.0 BARS P-3.0 BARS Figure 39. Relative Intensity Ratios of Ettringite and Gyps um for the 3.0% SO3 Content Mortar Bars at the Age of 480 Days in Sulfate
93 0 10 20 30 40 50 60 EttringiteGypsumRELATIVE INTENSITY RATIO x 100% C-3.6 BARS D2-3.6 BARS E-3.6 BARS P-3.6 BARS Figure 40. Relative Intensity Ratios of Ettringite and Gyps um for the 3.6% SO3 Content Mortar Bars at the Age of 480 Days in Sulfate Figures 39 and 40 compares the relative intens ity ratios of ettri ngite and gypsum for the four types of cements in the two cases of 3.0% and 3.6% SO3 content respectively. From the above figures, it can be seen that increasing the SO3 content of cements C and E led to an increase in the amount of ettringite found in the bars. These resu lts explain the expansion trend that was found for the C and E cements with the 3.6 bars having a higher expansion than the 3.0 bars.
94 Similarly, the results of the 3.6% set s howed that the C-3.6 bars contained the highest amounts of ettringite followed by P-3.6 bars, then D2 -3.6 bars, and finally the E3.6 bars with the lowest amount. The expans ion trend for the 3.6% set was exactly the same at that age, with the C-3.6 bars having the highest expa nsion, followed by P-3.6 bars, then D2-3.6 bars, and finally E-3.6 ba rs with the lowest expansion. The gypsum content of the D2 and E bars was higher than that of the C bars in both the cases of the 3.0% and 3.6%.
95 CHAPTER 4. CONCLUSIONS AND RECOMMENDATIONS It can be concluded from this st udy that there is an optimum SO3 content for the lowest expansion and for the highest compressi ve strength of the portland cement mortars used, and exposed to sodium sulfate environment. This optimum was not the same for both expansion and strength. The optimum also differed from one cement to another and from one age to another for each cement. For cements C, D2, and E with similar and low C3A content, the compressive strength showed an optimum at 3.0% for ages up to 28 days except for E cement. In the case of the latter, the optimum was at the as-received SO3 content. For cement P in lime, with a moderately high C3A and alkali content, increasing the SO3 content to 3.0% increased the strength; howeve r, increasing it further to 3.6% decreased the strength. The effect became more pronounced as the SO3 content was increased beyond 3.0%. As for the expansion behavior, for cements with similar tricalcium aluminate content, increasing the SO3 content beyond 3.0% results in an increase in expansion on exposure to sodium sulfate solution. For cem ent P with higher tricalcium aluminate and alkali content, increasing the SO3 content to 3.6% did not increase the expansion in sodium sulfate solution.
96 The findings of this study did indicate that increasing the SO3 content above 3.0% for all the cements studied here results in d ecreasing the durability of mortar exposed to sulfate environment.
97 REFERENCES 1. Goswami, G., Mohapatra, B., and Panda, J.D. (1990). Gypsum Dehydration During Comminution and its Effect on Cement Properties Journal of the American Ceramic Society, Vol. 73, No. 3, pp. 721-723. 2. Mindess, S., Young, J.F., Darwin, D. (2003). Concrete, 2nd ed. Prentice Hall, Pearson Education, Inc. Upper Saddle River, NJ. 3. Skalny, J., Marchand, J., Odler, I., (2002). Sulfate Attack on Concrete Spon Press, New York, NY. 4. Kanare, H.M., and Gartner, E.M. (1984). Optimum Sulfate in Portland Cement Cement Research Progress, Chapter 13, pp. 213-251. 5. Lerch, W. (1946). The Influence of Gypsum on the Hydration and Properties of Portland Cement Pastes American Society for Testing Materials, Proceedings, Vol. 46, pp. 1252-1297. 6. Alexander, K.M., Wardlaw, J., and Ivanusec I. (1979). The Influence Of SO3 Content Of Portland Cement On The Creep And Ot her Physical Properties Of Concrete. Cement and Concrete Research, Vol. 9, pp. 451-459. 7. Osbaeck, B. and Jons, E.S. (1980). The Infl uence of the Content and Distribution of Alkalies on the Hydration Prope rties of Portland Cement. 7th International Congress on the Chemistry of Cement, Paris. Vol. 2, pp. 135-140. 8. Ish-Shalom, M., and Bentur, A. (1972). Effects of Aluminate and Sulfate Contents on the Hydration and Strength of Portland Cement Pastes and Mortars. Cement and Concrete Research, Vol. 2, pp. 653-662. 9. Bentur, A. (1976). Effect of Gypsum on the Hydration and Strength of C3S Pastes. Journal of The American Ceramic Society, Vol. 59, No. 5-6, pp. 210-213. 10. Jelenic, I., Panovic, A., Halle, R., and Gacesa, T. (1977). Effect of Gypsum on the Hydration and Strength Development of Co mmercial Portland Cements Containing Alkali Sulfates. Cement and Concrete Research, Vol. 7, pp. 239-246.
98 11. Skalny, J., and Tadros, M.E. (1977). Retardation of Tricalcium Aluminate Hydration by Sulfates. Journal of The American Ceramic Society, Vol. 60, No. 3-4, pp. 174175. 12. Jawed, I., and Skalny, J. (1977). Alkalies in Cement: A Review, 1. Forms of Alkalies and Their Effect on Clinker Formation. Cement and Concrete Research, Vol. 7, pp. 719-730. 13. Collepardi, M., Baldini, G., and Pauri, M. (1978). Tricalcium Aluminate Hydration in the Presence of Lime, Gypsum or Sodium Sulfate. Cement and Concrete Research, Vol. 8, pp. 571-580. 14. Menetrier, D., Jawed, I., and Skalny, J. (1980). Effect of Gypsum on C3S Hydration. Cement and Concrete Research, Vol. 10, pp. 697-701. 15. Soroka, I., and Relis, M. (1983). Effect of Added Gypsum on Compressive Strength of Portland Cement Clinker. Journal of The American Ce ramic Society, Vol. 62, No. 6, pp. 695-697, 703. 16. Alexander, K.M., and Ivanusec, I. (1981). Long-Term Effects of Cement SO3 Content on the Properties of Normal and High-Streng th Concrete. Part I. The Effect on Strength. Cement and Concrete Research, Vol. 12, pp. 51-60. 17. Soroka, I., and Abayneh, M. (1986). Effect of Gypsum on Properties and Internal Structure of PC Paste. Cement and Concrete Re search, Vol. 16, pp. 495-504. 18. Rilem Committee 68-MMH. (1986). The Hydration of Tricalcium Aluminate and Tetracalcium Aluminoferrite in the Presence of Calcium Sulfate Materials and Structures, Vol. 19, No. 110, pp. 137-147. 19. Tang, F.J., and Gartner, E.M. (1988). Influence of Sulfate Source on Portland Cement Hydration. Advances in Cement Research, Vol. 1., No. 2, pp. 67-74. 20. Panigrahy, P.K., Goswami, G., Pa nda, J.D., and Panda, R.K. (2003). Differential Comminution of Gypsum in Ceme nts Ground in Different Mills. Cement and Concrete Research, Vol. 33, Issue 7, pp. 945-947. 21. Shanahan, N.G. (2004). Influence of C3S Content of Cement on Concrete Sulfate Durability. Thesis, M.Sc. in Civil Engineering. Un iversity of South Florida, Tampa, FL. 22. Taylor, H.F.W. (1997). Cement Chemistry, 2nd Edition. Thomas Telford Publishing, London, UK.
99 23. Irassar, E.F., Bonavetti, V.L., and Gonzalez, M. (2002). Microstructural Study of Sulfate Attack on Ordinary and Limest one Portland Cements at Ambient Temperature. Cement and Concrete Research, Vol. 33, pp. 31-41. 24. Rendell, F., and Jauberthie, R. (1999). The Deterioration of Mortar in Sulfate Environments. Construction and Building Materials, Vol. 13, pp. 321-327. 25. Santhanam, M., Cohen, M.D., and Olek, J. (2001). Review, Sulfate Attack ResearchWhither Now? Cement and Concrete Research, Vol. 31, pp. 845-851. 26. Tian, B., and Cohen, M.D. (2000). Does Gypsum Formation During Sulfate Attack on Concrete Lead to Expansion. Cement and Concrete Research, Vol. 30, pp. 117123. 27. Tikalsky, P.J., Roy, D., Scheet z, B., and Krize, T. (2002). Redefining Cement Characteristics for Sulfate-R esistant Portland Cement. Cement and Concrete Research, Vol. 32, pp. 1239-1246. 28. Al-Dulaijan, S.U., Masle huddin, M., Al-Zahrani, M.M., Sharif, A.M., Shameem, M., and Ibrahim, M. (2003). Sulfate Resistance of Plain and Blended Cements Exposed to Varying Concentrations of Sodium Sulfate. Cement and Concrete Composites, Vol. 25, pp. 429-437. 29. Monteiro, P., and Kurtis, K. (2003). Time to Failure for Concrete Exposed to Severe Sulfate Attack. Cement and Concrete Research, Vol. 33, pp. 987-993. 30. Khatri, R.P., Sirivivatnanon, V., and Yang, J.L. (1997). Role of Permeability in Sulfate Attack. Cement and Concrete Research, Vol. 27, No. 8, pp. 1179-1189. 31. Rendell, F., and Jauberthie, R. (1999). The Deterioration of Mortar in Sulfate Environments. Construction and Building Materials, Vol. 13, pp. 321-327. 32. Al-Amoudi, O.S. (2002). Attack on Plain and Blended Cements exposed to aggressive Sulfate Environments. Cement and Concrete Composites, Vol. 24, pp. 305-316.