Scholars' Mine

Masters Theses Student Theses and Dissertations

1970

Clay mineralogy and compaction characteristics of residual clay soils used in earth dam construction in the Ozark Province of Missouri

Arthur David Alcott

Follow this and additional works at: https://scholarsmine.mst.edu/masters_theses

Part of the Civil Engineering Commons Department:

Recommended Citation Alcott, Arthur David, "Clay mineralogy and compaction characteristics of residual clay soils used in earth dam construction in the Ozark Province of Missouri" (1970). Masters Theses. 7177. https://scholarsmine.mst.edu/masters_theses/7177

This thesis is brought to you by Scholars' Mine, a service of the Missouri S&T Library and Learning Resources. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact [email protected]. CLAY MINERALOGY AND COMPACTION CHARACTERISTICS OF RESIDUAL CLAY SOILS USED IN EARTH DAM CONSTRUCTION IN THE OZARK PROVINCE OF MISSOURI

BY

ARTHUR DAVID ALCOTT, 1941-

A THESIS submitted to the faculty of UNIVERSITY OF MISSOURI - ROLLA

in partial fulfillment of the requirements for the

Degree of

MASTER OF SCIENCE IN CIVIL ENGINEERING

Rolla, Missouri 1970 19C853 ii

ABSTRACT

The clay mineralogy and Standard Proctor compaction characteristics of clayey residual soils developed on the Salem Plateau of the Ozark Physiographic Province and used in earth dam construction were studied. A review of the geology of the Salem Plateau is presented along with a discussion of the factors controlling the development of soils on carbonate bedrock in a high leaching environment. The predominant soil clay mineral was found to be a poorly crystallized kaolinite rather than halloysite as has been postulated by others. Minor amounts of triple-layer minerals and an undetermined amount of an alumino-silicate gel are indicated. The effects of particle size and degree of crystallinity on the tests used to determine clay mineralogy are discussed. The effects of several soil properties on the laboratory compaction of the clayey residual soils were

investigated. Since the clay mineralogy was similar for all of the soils investigated, this factor was discarded as not being responsible for compaction differences. The percent clay was found to have the highest correlation with the maximum dry density and the optimum moisture

content of the soils. The plastic limit and cation­ exchange capacity were also closely correlated with the compaction characteristics. lll

It was determined that the pedular structure character­ istic of the residual soils was not destroyed upon laboratory compaction, particularly compaction dry of optimum moisture content. A theory of the compaction of clay soils is postulated which deals primarily with the behavior and changes in this ped structure with increasing moisture content. Scanning electron microscopy was used to illustrate points brought out in the discussions on the morphology of the soil clays and on the compaction characteristics of the clays. iv

ACKNOWLEDGEMENT

I wish to thank my adviser, Dr. N. 0. Schmidt, for his assistance and guidance during the period that I have attended graduate school and particularly during the period that this paper has been written. I also wish to express my appreciation to Dr. William Hayes, Chief Geologist, and the Missouri Geological Survey for financing the expenses related to the gathering of the test soils and other field data and for allowing the author to initiate an electron microscope study at the Survey's expense. Without this assistance, the project would not have been feasible. The clay mineralogy investigation required equipment and technology found only from outside sources. I wish to thank Mr. W. M. Dressel for permitting the DTA and TGA studies to be performed at the U.S. Bureau of Mines, Rolla, Missouri. I also acknowledge with thanks the assistance of Mr. R. Rickey and Mr. M. Adams for performing the tests and for their suggestions with regard to interpretation. The transmission microscopy was performed by Mr. Kuldip Chopra at the Ceramics Department, University of Missouri - Rolla. The scanning electron microscopy investigation was performed at the Materials Research Center, also at the University of Missouri - Rolla. My sincere thanks for their assistance. Dr. W. D. Keller was most helpful in allowing the v

author to use the x-ray diffraction facilities at the University of Missouri - Columbia. His discussions regarding testing technique and data interpretation were invaluable to this study. Special thanks is also offered to Dr. J. L. Eades of the Geology faculty of the University of Illinois. His independent analysis of the nature of the clay mineralogy substantiated that of the author and is most appreciated. The encouragement and assistance over the past two years of Mr. J. H. Williams and Mr. E. E. Lutzen of the Missouri Geological Survey is gratefully acknowledged. Finally I wish to thank my wife, Molly, for her understanding and help, particularly during the period when this paper was being prepared. vi

TABLE OF CONTENTS

Page ABSTRACT . . . . ii ACKNOWLEDGEMENT iv

LIST OF FIGURES X LIST OF TABLES . xiv I. INTRODUCTION 1

II. LITERATURE REVIEW . . 5

A. GEOLOGY ...... 5

B. SOIL FORMATION 8 1. Soils derived from limestones . . . . 8 2. Formation of Kaolins 11 3. Kaolinite morphology . . . . . 12 C. COMPACTION CHARACTERISTICS . . . . 15 1. Soil gradation and index properties . 16 2. Cation-exchange capacity 18

3. Clay particle structure . 26

II I. FIELD INVESTIGATION ...... 32 IV. MATERIALS ...... 35 v. TEST PROCEDURES ...... 40 A. SOIL GRADATION ...... 40 1. Sample preparation . . . 40 2. Coarse sieve analysis . . . . 40 3. Hydrometer analysis ...... · 41 B. ATTERBERG LIMITS TESTS ...... · · · 41 1. Liquid limits .. . . . 41 vii

Table of Contents Continued Page 2. Wet liquid limits 41 3. Dispersed liquid limit . 42 4 . Plastic limits . 53 C. MOISTURE-DENSITY RELATIONSHIPS 53 1 . Natural samples . . . . . 53 2 . Dispersed samples . . . 53 D. X-RAY DIFFRACTION ...... 55 1. General ...... 55 2 . Preparation of samples . . . . 57 3. Powdered samples ...... 58 4. Potassium acetate treatment . . . 58 E. DIFFERENTIAL THERMAL ANALYSIS . 60 1. General . . . . . 60 2 . Sample preparation and testing . . . . 61 F. THERMOGRAVIMETRIC ANALYSIS ...... 63 G. TRANSMISSION ELECTRON MICROSCOPY (TEM) 64 H. SCANNING ELECTRON MICROSCOPY (SEM) . . . . 66 I . LOW-POWER MICROSCOPY (LPM) . . . 72 J. SOIL CLAY CHEMISTRY ...... 72 1 . General ...... 72 2 . Chemical analysis ...... 72 3 . Soil pH in water . . . . . 73 4. Cation-exchange capacity . . 75 5. Flame photometry . . . . . 75 6. Absorption spectrophotometry . . . . . 77 K. SPECIFIC GRAVITY ...... 77 viii

Table of Contents Continued Page

L. CORRELATION TESTS 79

VI. ANALYSIS OF RESULTS 82

A. CLAY MINERALOGY . 82 1 . X-ray analysis . . . . 82 2 . Differential thermal analysis 90 3 . Thermogravimetric analysis 100

4. Electron microscopy 102

5 . Chemical analysis 110 6. Summary . . . 112 B. SOIL CLASSIFICATION AND CATION-EXCHANGE CAPACITY . . . . . 113 C. COMPACTION CHARACTERISTICS 123 1. Atterberg Limits . . . 123 2. Gradation and cation-exchange capacity 135 3. Soil clay structure 135 VII. CONCLUSIONS 147 A. GENERAL . 147 B. CLAY MORPHOLOGY 147 C. SOIL CLASSIFICATION AND CATION-EXCHANGE CAPACITY ...... 148 D. CORRELATIONS WITH COMPACTION RESULTS 150 E. THEORY OF RESIDUAL SOIL COMPACTION 150 F. APPLICATION OF TEST RESULTS . 152 1. Correlations 152 2. Dam design - construction 152 G. RECOMMENDATIONS FOR FURTHER STUDY 157 VIII. APPENDIX A - Dam Reports 159 lX

Table of Contents Continued Page IX. BIBLIOGRAPHY . 176

X. VITA .... 184 X

LIST OF FIGURES

Figure Page 1. Clay Distribution in Clarksville Soil (from Scrivner, 1960). 10

2 • Location of Dam Sites. 33 3 . Particle-Size Analysis of Masters #1 and Masters # 2 . 44 4 . Particle-Size Analysis of Terre Du Lac and Timberline . 45

5 . Particle-Size Analysis of Sayers, Little Prarie, and Blackwell. 46 6 . Particle-Size Analysis of Floyd, Ft. Westside, and Rice . 47 7 . Particle-Size Analysis of Sunrise Central, Sunrise South, and Fabick. 48

8. Particle-Size Analysis of Hornsey, Elsey, and Sunnen . 49 9 . Plasticity Chart Plot of Ozark Soils . 52 10. X-Ray Diffractograms of the Sayers Soil. 85 11. X-Ray Diffractograms of the Masters #1, Timberline, and Terre Du Lac Soils . 86 1 2 . Shift in 001 Peak of Georgia Kaolinite When Treated With Potassium Acetate . 87 13. Differential Thermal Analysis of Georgia Kaolinite and Indiana Halloysite . 93 14. Differential Thermal Analysis of Ozark Soils . 94 15. Differential Thermal Analysis of Ozark Soils . 95 16. Differential Thermal Analysis of Ozark Soils . 96 17. Differential Thermal Analysis of Ozark Soils . 97 18. Thermogravimetric Analysis . 101 xi

List of Figures continued Figure Page 19. TEM Terre DuLac Suspended Sample (57000X) ...... 103 2 0 • TEM Terre Du Lac Suspended Sample (52000X) ...... 103 21. TEM Terre Du Lac Suspended Sample (28000X) ..... 104

2 2. TEM Terre Du Lac Suspended Sample (20000X). . .. 104

2 3. SEM Terre Du Lac Powdered Sample (1000X) . 105 24. SEM Hornsey Powdered Sample (3000X) 105 25. SEM Hornsey Ped Sample (3000X) ... 106 26. SEM Hornsey Ped Sample (10000X)...... 106 2 7. SEM Hornsey Ped Sample (10000X)...... 108 28. SEM Terre Du Lac Ped Sample (10000X) ...... 108 29. SEM Timberline Ped Sample (10000X) .... 109 30. Clay Content Versus Liquid Limit (Clay Residual Soils) ...... 115 31. Clay Content Versus Plastic Limit (Clay Residual Soils) 116 32. Cation-Exchange Capacity Versus Liquid Limit (Clay Residual Soils) ...... 118 33. Cation-Exchange Capacity Versus Plastic Limit (Clay Residual Soils) ...... 119 34. Cation-Exchange Capacity Versus Clay Content (All Soils) ...... 120 35. Clay Content Versus Maximum Dry Density (Clay Residual Soils) ...... 124 xii

List of Figures continued Figure Page 36. Clay Content Versus Optimum Moisture Content (Clay Residual Soils)...... 125 37. Cation-Exchange Capacity Versus Maximum Dry Density (Clay Residual Soils) . 126 38. Cation-Exchange Capacity Versus Optimum Moisture Content (Clay Residual Soils) .. 127 39 . Liquid Limit Versus Maximum Dry Density (Clay Residual Soils). 128 40. Liquid Limit Versus Optimum Moisture Content (Clay Residual Soils). . .. 129 41. Plastic Limit Versus Maximum Dry Density (Clay Residual Soils). . . .. 130 42. Plastic Limit Versus Optimum Moisture Content (Clay Residual Soils)...... 131 43. Clay Content Versus Maximum Dry Density (All Soils) .... 132 44. Clay Content Versus Optimum Moisture Content (All Soils) . 133 45. Optimum Moisture Content Versus Maximum Dry Density (Clay Residual Soils) 134 46. LPM Terre Du Lac (Black and White) 137 a.) Compacted 3% Dry of Optimum b . ) Compacted 5% Wet of Optimum 47. LPM Terre Du Lac (Color) 138 a.) Compacted 3% Dry of Optimum b.) Compacted 5% Wet of Optimum

48. SEM Terre Du Lac (lOOX) •• 140 a.) Compacted 3% Dry of Optimum b.) Compacted 5% Wet of Optimum 49. SEM Terre DuLac (300X). 141 a.) Compacted 3% Dry of Optimum b.) Compacted 5% Wet of Optimum 50. SEM Terre Du Lac (3000X) . . . . 142 a.) Compacted 3% Dry of Optimum b.) Compacted 5% Wet of Optimum xiii

List of Figures continued

Figure Page 51. The Effect of Dispersion of the Standard Proctor Compaction of Clayey Residual Soils. . 146 52. Graphical Relationship of the Maximum Dry Density to the Plastic Limit and the Clay Content of the Clayey Residual Soils . . 153 53. Graphical Relationship of the Optimum Moisture Content to the Plastic Limit and the Clay Content of the Clayey Residual Soils . . 154 54. Graphical Relationship of the Maximum Dry Density to the Plasticity Index and the Liquid Limit of all Soils. . 155 55. Graphical Relationship of the Optimum Moisture Content to the Plasticity Index and the Liquid Limit of all Soils . . 156 xiv

LIST OF TABLES

Table Page I. Affect of Crystal Structure on Cation­ Exchange Capacity (after Kelley and Jenny). 22

II. Material of Dams Investigated 36 I I I. Comparison of Natural Field Moisture With Atterberg Limits . 39

IV. Summary of Particle Size Analysis . 43 V. Comparison of Wet Liquid Limit And Natural Atterberg Limits. so VI. Comparison of Natural and Dispersed Atterberg Limits. so VII. Summary of Atterberg Limits Tests 51 VIII. Result of Moisture-Density Tests. 54

IX. X-Ray Machine Specifications. 56 X. Summary of Major DTA Temperature Peaks. 62 XI. Chemical Analysis of Ozark Soils. 67 XII. Summary of pH and Cation-Exchange Capacity of Ozark Soils. 74 XIII. Concentration of Mg, Ca, Na, and K Cations in meq/100 grams. 76 XIV. Specific Gravity of Ozark Soils 78 XV. Correlations for Clayey Residual Ozark Soils. 80 1

I. INTRODUCTION

Missouri has approximately 1600 man-made earth dams, as many or possibly more than any other state in the Union. About 700 of these dams are located in the Ozark Physiographic Province, a series of plateau surfaces developed principally on Paleozoic carbonate bedrock formations. A mantle of clayey residual soils up to 200 feet thick overlies these bedrock surfaces. Earth dams constructed in this area, particularly those constructed of the clayey residual soils, have a history of poor performance. Excess seepage, slope instability, piping, and total embankment failures have occured. Further problems will emerge as urban expansion and increasing numbers of recreational lake developments extend into the area. The purpose of this investigation is to study the nature of these residual soils, the factors which affect the compaction of the soils, and to recommend procedures to improve the performance of the dams. The stony nature of most of the embankments prohibited the collection of near surface undisturbed samples with hand tools. Owner resistence and the expenses involved in mobilizing and using mechanized equipment further prohibited undisturbed sampling. Only dams with a reservoir surface area greater than 15 acres were investigated. The expectation was that this restriction would limit the investigation to dams which were 2

constructed using accepted engineering practices. This was not the case. Almost anything passes for accepted construc­ tion practice in this state where there is no legislation to regulate dam design and construction. Therefore, attempts to relate dam performance with the nature of the embankment materials or with the embankment performance were not feasible. One dam with flat slopes utilizing good materials but poorly constructed might perform less satisfactorily than another constructed with steep slopes and less suitable materials, but well compacted. The investigation has been directed, therefore, to a study of the nature of the soils used in dam construction and to the factors which affect the compaction of these soils. With this knowledge, it should be possible to assess the potential for dam construction if good construction techniques are followed. The effect of significant soil properties on Standard Proctor compaction parameters were investigated. These properties were chosen considering the clayey nature of the soils. They are: (1) clay mineralogy, (2) Atterberg Limits and grain-size distribution, (3) cation-exchange capacity, and (4) soil clay structure. Evidence indicates that the Tertiary climate of the Ozark region was semi-tropical or tropical. Where carbonate bedrock is located in this type of climate, chemical leaching is prominant and conditions favoring the formation 3

of kaolinitic and halloysitic soils are developed. Terzaghi (1958) reports the characteristics of a halloysitic soil of tropical Kenya to be similar to that of the Ozark soils, Graham (1969) sites the probable presence of halloysite in the Ozark soils as being responsible for the low maximum dry density and the high optimum moisture content of the Springfield soil, an Ozark residual clayey soil. The mineralogical investigation was designed to determine more explicitly the nature of the clay and possible changes in its morphology from one portion of the Ozark Province to another. X-ray diffraction, differential thermal analysis, thermogravimetric analysis, chemical analysis, transmission electron microscopy, and scanning electron microscopy techniques were chosen for this investigation. The results of several of these tests are shown to be affected by mineral particle size, degree of crystallinity, and other factors. These factors can lead to errors in interpretation if they are not recognized and considered. The electron microscopy studies, particularly those of the transmission electron microscope, provide keys to clay mineral identification. Harris (1969) has made numerous correlations between independent soil properties as the percent fines, the percent clay, the liquid and plastic limits and the Standard Proctor compaction test variables, the maximum dry density and the optimum moisture content. It was decided to determine if 4

similar correlations exist for the Ozark residual soils. It was chosen to study the nature of the clay structure of the laboratory compacted samples with the scanning electron microscope at various magnifications 1 particularly samples compacted wet and dry of the optimum moisture content. 5

II. LITERATURE REVIEW

A. GEOLOGY The geology of the Missouri Ozarks is illustrated on the Geologic Map of Missouri published by the Missouri Geological Survey. The general geology of the Ozark Plateaus is that of a horizontal sequence of sedimentary formations deposited by shallow marine seas which moved onto and off of a Precambrian dome centered in the St. Francois Mountains. The Ozark Province has been divided into four subprovinces by Thornbury (1965): (1) the St. Francois Mountains - the Precambrian igneous core of the province; (2) the Salem Plateau - underlain by Cambrian and Ordovician dolomites with some sandstones; (3) the Springfield Plateau - underlain primarily by Mississippian and Pennsylvanian limestones; and (4) the Boston Mountains - rugged topography developed on shales, limestones and sandstones. The topography of the northern flank of the Salem Plateau where the darns investigated are located is char­ acterized by steep-sided ridges with closely spaced drainage. Broad uplands are developed in portions of Phelps, Dent, and Crawford Counties on the Roubidoux and Jefferson City formations. The youngest formation exposed in the area is the Ordovician Jefferson City dolomite, and oldest is the 6

Cambrian Potosi dolomite. The stratigraphy of these formations is described in detail in The Stratigraphic Succession in Missouri published by the Missouri Geological Survey. Bretz (1965) attributes the plateau development to weathering cycles which extended through three periods of peneplanation forming the three plateau levels. These stages of peneplanation correspond to, and are the result of, tectonic pulses triggered within the province dome. Evidence of later, minor pulses are noted in terraces and rejuvenated streams. He dates the plateaus as follows: (1) Boston Mount­ ains (pre-Tertiary?), (2) Springfield Plateau (early Ter­ tiary?), and (3) Salem Plateau (post-early Eocene). The minor pulses are dated as Pliocene (?) and pre-Pleisto­ cene(?). According to the theoretical cycles of pene­ planation, the Salem Plateau is now in the early mature stage. Bretz bases his postulation on the following evidences of peneplanation noted in the area: (1) accordant summits, (2) flat interfluvial divides, and (3) monadnocks. Quinn (1956) disputes the hypothesis of peneplanation and postulates that the plateaus, ". . were eroded in place by alternating periods of stream entrenchment providing the escarpments, and periods of escarpment retreat (pedimentation), forming the plateau surfaces." 7

Valley entrenchment occurs during periods of high moisture which results in rejuvenation of streams resulting in erosion. During alternating periods of arid conditions

streams aggrade and mechanical weathering predominates. Quinn relates these alternating cycles to correspond with the glacial and interglacial climatic conditions.

The two proposals vary considerable in dating the

development of the Salem Plateau. Bretz dates the plateau

as post-Eocene, pre-Pliocene; Quinn dates the plateau as late Pleistocene. The age of the residual soils must coin­

cide with or be somewhat younger than the plateau itself.

The soils must be older than the Pleistocene loess deposits which often covers the residium, particularly in the north­

eastern portion of the plateau. The exact age of the soils

has not been determined. During the Tertiary period, what is now the Gulf of Mexico extended far up into the Mississippi River Valley.

This near-shore (probably tropical) environment could

have provided the moisture required to produce a climate favorable to laterization forming the clayey residual soils on carbonate bedrock. Toward the end of the Tertiary with the advent of the Pleistocene glacial stages, these favorable conditions ceased and the existing soil forming

cycle was terminated. 8

B. SOIL FORMATION 1. Soils derived from limestones. The residual soils investigated in this study are classified as Ultisols (U.S.D.A., 1960). Earlier systems classified the soils as Red-Yellow Podzolics. Where Red~Yellow Podzolics are usually developed in tropical climates, Ramann (as reported in Miller, 1965) noted a tendency of tropical soils to pen­ etrate into temperate climates on limestone formation.

This is possible because the permeable limestones, or dolomites, permit rapid percolation with a resultant decrease in ground water influence on soil formation. The principle clay minerals in these Red-Yellow Podzols developed on limestones are kaolinites and halloysites. The soils also exhibit an increase in iron and clay content with depth (Scrivner, 1960). The laterization process is considered to be the dominant process in the genesis of these residual soils (Simonson, 1949). There is little evidence that eluviation and illuviation have acted as a major process to concen­ trate clay in the B horizon. Simonson further suggests that the dominant process in these soils is the destruction of clays in the upper horizons and the formation of new clay minerals in the deeper horizons. Jenny (1941) notes that in humid regions the limestone derived soils are closely related to the impurities in the 9

parent material. Carroll and Hathaway (1953) discuss the soils developed on the Lenoir limestone in Virginia. The limestone contains partings of hydrous mica and montmoril- lonite. Their investigation shows that the mica and mont- morillonite were changed during the leaching process to kaolinitic and chloritic clay soils. The clay content, particularly that of kaolinite, and the cation-exchange capacity increased with depth through the soil profile. Scrivner (1960) notes this variation of clay content with depth and a graph from his study illustrating this variation is presented as Figure 1. In a study of the Lebanon soil in Missouri, Scrivner (1960) reports that a limestone which contains illite weathered by chemical leaching to produce a soil profile which contains predominantly kaolinite. The soils also contain some montmorillonite, an intermediate product of illite weathering. Miller (1965) studied four Ozark soils developed by chemical weathering of four different carbonate bedrock formations. The formations ranged in age from Ordovician to Mississippian and the sources of the samples were separated by distances of up to 100 miles. There is a marked similarity in the soil profiles and mineralogy, 0 which consists primarily of kaolinite and some 14 A minerals. The only clay mineral present in any significant amount in the bedrock formations is illite. 10

10

Clay ( 0.002 mm) as 20 percent of less than 2 mm soil

30 Limestone 40

50 ,.-..... 0 60 ..__,·r-f ...c: .j.J 70 0. Q) 0 80

90

100

110

120

0 10 20 30 40 50 60 70 80 90 100 Clay Content (%) FIGURE 1. Clay Distribution in Clarksville Soil (from Scrivner, 1960) 11

2. Formation of Kaolins. The most common minerals of the kaolin group or family of minerals are allophane, halloysite (hydrated and non-hydrated), and kaolinite. Investigators present two primary modes of kaolin for­ mation; as precipitated gels and as particles from solution. Hauser (1952) postulates that after chemical leaching, hydration and adsorption of ions, the silica and alumina gels react to form halloysite. Then, by conden­ sation, kaolinite and montmorillonite, mica and other minerals may be formed depending on what ions are present. Bates (1952) favors the theory of formation of particles from solution. The kaolin mineral produced depends on many factors; time, temperature, pressure, humidity, pH, and the composition of the solution. Sherman (1952) relates the kaolinite formation to the intensity of the chemical weathering (amount, intensity, and distri­ bution of annual rainfall) and the length of time the material has been subjected to weathering. Kaolin formation is maximized by moderate, non-cyclic annual rainfall and good internal drainage. With increased age and/or greater annual rainfall, the kaolinitic clays decompose into their individual hydrated oxides which are concentrated to form laterites. Keller (1968) presents a detailed study of the cry­ stallization of clay minerals from solution which is beyond the scope of this investigation. In his summary he states, 12

" . it may be said that the factors which are important in the development of kaolin-type minerals from alumino-silicates are the complete removal of the metal cations (other than Al and Si) and the introduction of H ions (Ross, 1943), and a relatively low silica~alumina ratio." The conditions required to produce the above factors are a humid climate where precipitation exceeds evaporation, permeable bedrock to allow continued leaching, and an oxidizing environment to remove Fe from solution. Scrivner (1960) presents a mechanism whereby alumino­ silicates arrive in solution from which they are subse- quently precipitated as kaolins. His research describes the destruction o£ the illite, alumina feldspars, and quartz from the carbonate bedrock by an acid soil environment to produce a solution of alumino-silicates. The metallic Mg and Ca ions are removed by leaching from the solution and the Fe is oxidized to Fe2o3 . The requirements outlined above by Keller have been satisfied. A pH change immediately above bedrock from acidic to basic keys the precipitation of a kaolinitic mineral out of solution.

3. Kaolinite morphology. In the preceeding paragraphs the formation of kaolinitic minerals from solution has been discussed, but the morphology of these minerals has not been considered. A number o£ minerals from well-crystallized kaolinite to halloysite to 13

allophane may be the product of this precipitation.

Keller et al (1970) emphasize the question of kaolin group morphology when they state, " A long-standing problem of the kaolin group minerals has been the question: 'What geologic or geochemical environments determine whether well­ crystallized kaolinite; b-axis disordered, 'fire-clay' mineral; kaolinite of hexagonal, elongate plate, tubular or other external morphology; endellite, halloysite, allophane, dickite of differing IR absorptions; or possible other varieties, are formed?' A corallary question asks 'is there a sequence, either possible or necessary, in either direction, between allophane, endell1te, kaolinite, dickite. . ?'" In a discussion of the interrelationship of morphology and chemistry, Bates (1959) cites the following factors that determine the degree of curvature and therefore the shape of the minerals. The first is the degree of misfit between the tetrahedral and octahedral sheets which are affected by the number and size of cations. The second is the strength of the interlayer bond which is a product of the size and polarizing power of the cations and the distribution of H ions. The platy kaolinites are favored over the lath or tubular materials as the H20 content decreases and as the Si02 :Al 2o3 increases (as the number of cations decreases). Bates further postulates that there is a complete morphological cycle in the kaolin group from the well­ crystallized hexagonal plates of kaolinite to elongate 14

plates, to laths with crystallographic terminations to curved laths to tubes of halloysite (4H20). A halloysite-allophane transition has also been postulated (Sudo and Takahashi, 1956). It is therefore possible that a complete morphological series from allophane to well-crystallized kaolinite exists. In his discussion of soil clays Bramao (1952) notes three morphological classes: (1) hexagonal plates, (2) cylindrical or rod-like particles, and (3) irregular layered particles having curved surfaces. This third class exhibits x-ray and thermal properties of both halloysite and kaolinite. The clay particles are generally fine, have roughly hexagonal or elongate shape with rounded edges, and a roughness of texture. Bramao further summarizes the conditions under which this irregular soil clay kaolin would form. "The small particle size and evident poor crystallinity of much of the soil kaolin are results of the weathering conditions in soils. Soils are subject to wetting and drying cycles; a kaolinite crystal may begin to synthesize in one part of the cycle, and its growth may be interrupted or retarded during another part. The next increment of crystal growth might be more hydrous, or have a slightly different chemical composition and physical structure. Soil formation is characterized by the simultaneous presence or a large number of free ions and colloidal systems; the latter, particularly the hydrous oxides of aluminum and iron, may well 15

affect the development of clay mineral crystals. Where halloysite and kaolinite can grow undisturbed in an ideal environment, they can attain their characteristic euhedral forms. Where weathering conditions are dynamic and reflect the interplay of a host of variables, as in soils, perfect crystal formation would be only fortuitous." Bramao further postulates that halloysite may be a special form of kaolinite, one which has been deformed by the presence of excess water molecules. He believes that the poorly developed kaolin which forms in the above type of environment is not necessarily a transitional crystal- line form but one which, ". . may be an equilibrium form peculiarly developed in a soil environment."

C. COMPACTION CHARACTERISTICS The compaction of soils is affected by many factors including such natural variables as mineralogy, gradation, classification, moisture content, and mechanical variables as type and intensity of compactive effort. This inves- tigation proposes to study the effect of three of these variables; soil gradation and index properties, cation­ exchange capacity, and clay-particle structure, on the compaction characteristics of the Ozark soils. Corre- lations are made of the interrelationships between these factors and the compaction characteristics of the soils. These characteristics are the maximum dry density and 16

optimum moisture content as defined by ASTM D 1557-66. 1. Soil gradation and index properties. The effect of the gravel content of the soil (percentage of material retained on the No. 4 sieve) has been discussed by several authors. Holtz and Lewitz (1957) note that the compaction of the fines is not affected if the soil contains less than about 30% gravel. For soils containing from 30-50% gravel there is some interference with the compaction of the fines. As the plasticity of the fines increases, the soil is able to tolerate a greater percentage of gravel before the com­ paction of the fines is affected. Zeigler (1948) discusses the effect on the moisture­ density relationship which occurs when varying amounts of gravel are added to a loamy sand. The addition of SO% gravel (material finer than 3/4" and retained on the No. 4 sieve) increases the maximum dry density from 119.2 pcf to 133.5 pcf and the optimum moisture content from 13.5% to 7.5%. The trend of the increase approaches linearity. Harris (1969) has conducted a detailed investigation into the interrelationships between independent variables (soil gradation, index properties) and dependent variables (engineering properties as cohesion, angle of internal friction, and compaction characteristics). The major objection of his study was to develop predictive relationships for evaluating the dependent variables. 17

Two recent studies noted by Harris concentrate on the prediction of the moisture-density values of soils. Ring et al (1962) conducted studies for the Bureau of Public Roads. The first study produced curves relating the liquid and plastic limits to the maximum dry density and the optimum moisture content. A later study reported by the same authors developed equations and curves relating the plastic limit and fineness average (defined as one-sixth of the summation of the percentage of particles finer by weight than the following sizes in millimeters: 2.0, 0.42, 0.074, 0.020, 0.005, 0.001). It was found that the values of optimum moisture and maximum dry density are more closely predicted when related to plastic limit and fineness average than when related to the plastic limit and liquid limit. These studies utilize linear regression to analyze the relationships. Harris summarizes a report by the U.S. Army Engineer Waterways Experiment Station (1962) which presents the relationships between compaction, consolidation and strength characteristics and the index properties of the soils. The study develops correlations between index properties and soil gradation, index properties and specific gravity of the soils, and the index properties and the compaction characteristics. Harris (1969) concentrates his study to correlations 18

involving plastic fine-grained soils. He finds that the compaction values are significantly related to the plas­ ticity indices and the percent clay. The liquid limit provides the highest degree of simple correlation. A very high degree of correlation was obtained for the optimum moisture content versus maximum dry density. His study indicates that a high degree of correlation between the compaction characteristics and the liquid limit holds also for residual soils. A lesser correlation is noted with the plastic limit. The correlation of compaction values with soil gradation is less significant. For the glacial soils the gradation is more closely correlated with the optimum moisture and maximum dry density than with the index properties. The specific gravity correlates poorly with other properties. Few of the residual soils studied by Harris exhibit the high percentage of clays, the high degree of plasticity, or the low density characteristics typical of the soils tested in this study of Ozark soils. It may well be that the relationships determined by Harris will not hold for the Ozark residual soils. 2. Cation-exchange capacity. That cations exdst in the clay-water system and that they affect the properties of the system has been known for some time. During the 18th century, it was determined that the process of cation- 19

exchange occurred and that the phenomenon was restricted to the clay fraction of the soil (Grim, 1952). Other early studies were carried out by investigators in the field of soil chemistry. Hauth and Davidson (1950) and Grim (1953) discuss three methods by which clay minerals adsorb these cations. In the first, cations are attracted to the edges of the clay plates which hold a negative charge. The second occurs as a result of substitution within the clay lattice. Substitution of divalent ions for trivalent aluminum ions results in a deficiency of positive charges which is satisfied by cations which become strongly attached to the clay particles. The third method involves the exchange of the hydrogen in exposed hydroxyls by other cations. The first method, while active in all clays, is responsible for most of the exchange capacity of the kaolinite family minerals. The platy kaolinite minerals which have a layer of hydroxyls exposed along their basal cleavage surfaces may also derive some of their exchange potential from method three. The substitution of cations within the clay lattice is most prevalent in the triple-layer minerals where it is responsible for up to 80 percent of the total cation-exchange capacity (Grim, 1953). The exchange reaction in clay is very complex and as of 1958 no one had been able to develop a qualitative 20

hypothesis which completely describes the reaction or the sequence in which different cations would become involved in the exchange reaction (Grim, 1958). As early as 1939, Grim gave the following approximate order of cation exchangeability; Na>K>NH 4>Mg>Ca. It appears that divalent cations replace monovalent cations more readily than vice versa. Other factors involved in the exchange reaction include the total population of exchange positions, the concentration of the replacing ions, the nature and concentrations of the anions, the size and polarity of the ions, and the hydration tendency of the ions (Grim, 1952). Certain exchange relationships have been observed and reported in the literature. Grim (1953), and a number of other authors, report typical values of the cation-exchange capacity of clay minerals:

CEC meq/100 gr. Kaolinite 3 - 15

Halloysite zH 2o 5 - 10 Halloysite 4H 2o 40 - 50 Illite 10 - 40 Montmorillonite 80 - 150

Grim and Bray (1936) examine the properties of a number of ceramic clays and showed that the cation-exchange capa­ city increased as particle size decreases. Harmon and 21

Fraulini (1940) studied the properties of a kaolinite as a function of particle size. They note a pronounced increase in the cation-exchange capacity, about four-fold, as particle size decreased from 10-20 microns to less than 0.1 microns. The authors also devoted a portion of the study to the relationship between cation-exchange capacity, specific surface, and permeability showing an increase in permeability with increasing cation-exchange capacity and decreasing particle size. Kelley and Jenny (1936) determined that the cation­ exchange capacity of clay minerals increased markedly upon the breakdown in crystal structure caused by grinding. Table I modified from Kelly and Jenny shows that the cation-exchange capacity of a kaolinite increased from 8 to 100 meq/100 gr. as the mineral was ground for 7 days. Data for other minerals are also included in the Table. These authors postulate that upon grinding, many OH ions of the kaolinite lattice become exposed as breaks are produced across the octahedral layer or parallel thereto. The increased exchange capacity is due to these exposures of unsatisfied negative charges. These studies seem to indicate that the cation-exchange capacity is a more dominant factor in montmorillonites than in kaolinites and halloysites, and as the clay particle size and degree of crystallinity decrease the cation­ exchange capacity increases. 22

TABLE I Affect of Crystal Structure on Cation-Exchange Capacity (after Kelley and Jenny)

Minerals 100 Mesh Grinding Time 48 hr 72 hr 7 days

Muscovite 10. 5 . . . 76.0 . . . Biotite 3.0 62.0 72.5 . . . Kaolinite 8.0 57.5 70.4 100.5 Montmorillonite 126 . 238 . . . I . .

Baver (1930) shows that the divalent cation Mg and Ca flocculate at much lower concentrations than Na and K cations and that they produce a structure which is favorable to the movement of water and air; a structure which is relatively stable. Baver (1956) notes that, "The high hydration and dispersive action of the Na ion makes the plasticity of the Na-saturated soil greater than those soils saturated with divalent ions." Winterkorn and Moorman (1941) present data illustrating the effect of different exchange ions on Atterberg Limits, optimum moisture content and maximum dry density, total consolidation and rate of consolidation, permeability, and shear strength of compacted samples. Saturating the soils with different cations also causes changes in the results 23

with different cations also causes changes in the results of hydrometer particle size analysis showing the effect of each cation on the flocculated or dispersed nature of the clay particles. Davidson and Sheeler (1952) relate the cation-exchange capacity of loess in southwestern Iowa to certain engineer­ ing properties of the soil. At exchange capacities greater than about 20 meq/100 gr. the curves relating CEC to Atterberg Limits and percent clay appear to approach linearity. At least a very strong correlation is established. Along with these investigations on the nature of the cation-exchange reaction and its affect on the engineering properties of clay soils were investigations into the clay­ water system itself and how the cation affected the system. Winterkorn (1940) describes the oriented water film surrounding clay particles and how if affected the plasticity indices. Grim (1952 and 1958) has expanded this explanation. The properties of the clay-water system are a function of: (1) the bond between the particles, (2) the amount of water between the particles, and (3) the nature of the water adsorbed on the surfaces of the clay mineral particles. The bond between particles is dependent largely on the cations held on the basal and edge surfaces of the clay 24

particles. The valence, geometric size, and tendency of the ion to hydrate affect the bonding strength. The geometric nature of this bonding restricts to a degree the amount of water which can penetrate between particles. The exchangeable ions also affect the configuration of the water molecules that envelope the clay surfaces. The water for some distance out from the surface is not a true liquid since the molecules are probably arranged in a preferred orientation different from that of liquid water. There is an indication that the nature of the water molecule orientation may vary with the exchangeable ion and the thickness of the adsorbed water later. The Na ions tend to develop thick surrounding layers of oriented water whereas Ca and Mg ions develop only thin layers; tens of molecular layers for the Na ion versus only about four for the Ca and Mg ions. In addition there appears to be no distinct boundary between the oriented and liquid water systems when the Na ion is involved (Grim, 1958). The "swarm" of water molecules and cations surrounding the clay particles have been referred to as the double layer. A decrease in the thickness of the double layer causes a reduction in the electrical repulsion between clay particles. Lambe (1960) has graphically illustrated the effect of electrolytic concentration, ion valence, dielectric 25

constant, and temperature on the electrical potential of the double layer system. The higher the ion concentration of the pore fluid, the greater the particle-to-particle attraction and the less diffuse the double layer. A similar relationship exists as the valence of the ions in the pore fluid increases. Lambe and Martin (1960) note that the only exchangeable ions likely to be important in soils are Na, K, Ca, Mg, Fe, and Al and that of these Ca and Mg generally account for 50 to 90 percent of the exchangeable ions except on extremely acid soils of humid regions. Recently in the field of soils stabilization research the effects of exchangeable cations have been extensively studied and utilized. Vees and Winterkorn (1967) found that the liquid and plastic limits of a kaolinitic soil increases with the valence of the exchangeable ions and that this indicates an increase in the flocculation effect of the higher valence ions. The addition of lime to soils acts to decrease the dry density and to increase the optimum moisture content of compacted samples (Ladd et al, 1960). The addition of 10 percent lime caused a drop in dry density from 98 to 87 pcf for a heavy clay with a liquid limit of 60 and a plastic index of 20. Lambe (1962) extends the results of earlier studies to cover the general effects of the addition of aggregates and 26

dispersants to compaction samples. The aggregates produce a strongly random structure which resists compaction effort and results in a lowering of the maximum dry density and an increase in the optimum molding water content. Dispersants such as sodium tetraphosphate reduce the electrical attraction between particles causing a decrease in cohesion. This leads to a significant decrease in the liquid limit. The repelled clay particles can be easily moved relative to each other and may be molded into a dense mass by mechanical compaction. The result is an increase in compaction with the dispersed structure and a lowering of the optimum moisture content.

3 • Clay particle structure. Lambe (1953) shows that marine clays have a more flocculated structure than fresh­ water samples. His research also shows that remolded samples have a more oriented (parallel) fabric. Mitchell (1956) used a petrographic microscope to illustrate the improved orientation of clay samples with remolding. Lambe (1958) contributed a classic paper postulating the effect of structure on compacted clay samples. In this paper, Lambe reviews the factors which affect the structure of natural soils. He also lists several variables which can be altered by the engineer which will change the structure: (1) type of compaction, (2) amount of compaction, (3) amount of water, and (4) additives. Lambe concludes that: (1) 27

clays compacted dry of optimum owe their low density to a flocculated structure; (2) this flocculated structure results in part from a high electrolytic concentration of ions with strong bonding capacity which prohibits the development of a thick diffused double layer; and (3) wet of optimum a dispersed structure is developed which allows more efficient particle arrangement. Seed and Chan (1959) conducted a number of tests on clay samples compacted by different methods wet and dry of optimum. They investigated and reported the effect of particle arrangement or structure on sample shrinkage, swelling, and total and effective strength characteristics. The results of these tests when interpreted in light of Lambe's theory substantiate his conclusions with regard to particle structure. The authors also note the effect of different methods of compaction. They found that the soils which had undergone the greatest amount of shear strain had the greater degree of particle orientation and a lower shear strength. In his discussion of the physico-chemical properties of soils, Michaels (1959) reports the possibility of "packets" of parallel clay particles. He believes the packets act as "rigid solid entities" which resist attempts at being forced into a coherent mass because the area of contact between packets is so small. Relative to 28

individual clay particles, there are few points of attraction (cohesion) between packets of compacted, dry clay. Trollope and Chan (1960) discuss a mechanism whereby packets are formed. They attribute the packets to a strong electrolytic environment which forces particles to move together. The authors contend that only in zones of large shear strains are the particles reoriented and that compacted clay soils which are generally considered to be remolded are for the most part not remolded. Alymore and Quirk (1962) used transmission microscopy to study the natural structure of clay soils. They report a "turbostatic" microstructure of clay particles. "Turbostatic" structure refers to a twisted, swirl-like clay particle arrangement similar to that noted by Borst and Keller (1969) and discussed in a later section of this paper. This type of structure is noted in natural clays, particularly kaolinite and illite. Sloan and Kell (1966) used transmission microscopy to illustrate the structure of compacted kaolin samples. Their study illustrates the predominance of packets of clay particles in the compacted samples. Few individual clay particles were noted. The packets appear to exhibit the compacted arrangement (flocculated and dispersed) that Lambe had postulated for single particles. The kaolin compacted dry of optimum exhibited random packet orientation. 29

Zones affected by higher shear strains (as those close to the top of each layer in a compacted sample) show greater orientation. As water is added beyond that required for optimum, greater packet orientation is observed. Sloan and Kell further state:

"The addition of molding water in amounts less than that required to produce a slurry or viscous suspension would probably not disrupt the packets; hence persistence of the packets through the compaction process over the relatively limited moisture range in this study could be expected. The relative absence of individual particle edge-to-edge relationships would seem to support this view."

The investigation by Barden and Sides (1970) utilized the scanning electron microscope to illustrate the compacted structure of a clayey soil. The structure of the soil compacted at one-half Standard Proctor compaction effort at 2.8% dry of optimum and 5.2% wet of optimum was investigated. Their study indicates that at low to moderate magnification (20-SOOX) the structure of the compacted samples appears to vary appreciably depending on the molding moisture content. Dry of optimum, distinct "pellet-like" macropeds were predominant. Macropeds are packets or groupings of clay particles. Wet of optimum the structure was more homogeneous, offering little evidence of macropeds. At higher magnification (1900-9500X) there was no marked 30

difference between the micrographs of samples compacted wet and dry of optimum. This would indicate that the intrapacket or microstructure of samples is not dependent on the molding water content. Indeed, there appeared to be little difference between high magnification micrographs of the natural soils and those of the compacted samples. Barden and Sides postulate the following compaction sequence. At low compaction, moisture, the macropeds appear to resist deformation and macrospaces filled with air exist between the macropeds. As more water is added the macropeds become weaker and distort under compactive effort to reduce and finally eliminate the macrospaces. At this point the soil appears fairly homogeneous and the dry density is at a maximum. The addition of more water causes a reduction in the dry density as the water layers between soil particles increases. Obviously this explanation varies considerably from that postulated by Lambe (1958) and Seed and Chan (1959) where the change in particle orientation with change in molding water content determines the compacted density. To varying degrees Trollope and Chan (1960) and Sloan and Kell (1966) had hinted at the compacted nature of clay soils as outlined by Barden and Sides. The significance of this "new" explanation of compacted clay soils is evident. The theories of soil swelling, 31

shrinkage, consolidation, and shear strength of compacted clays based on the Lambe (1958) theory must undergo some re-examination and revision to allow for the presence and influence of the pedular nature of the compacted samples. 32

III. FIELD INVESTIGATION

Sixteen Missouri Ozark earth dams were investigated. Where possible soils were obtained from the borrow source of the dams. Where this was not possible, soils from the vicinity of the dams were sampled. Figure 2 shows the locations of the dams investigated. The physical parameters of the dams were determined with a 100 ft. chain, an abney, and a Brunton compass. The parameters determined were the width, height, length, location of water level, and the dimensions of upstream and downstream embankment slopes. The upstream slope values are based on the angle of the slope above the reservoir level. The exact location of the darns were spotted in the field on U.S.G.S. Quadrangle sheets. Early attempts to obtain moisture content and undisturbed samples from the darns with agricultural probes and small augers were abandoned when it became obvious that the hand equipment was not adequate to sample the stony embankment material. The location of the principle source of embankment material at each darn was generally apparent or was obtained from someone in the vicinity of the site. At these sites approximately 40 pounds of soil were obtained and placed in a plastic sample bag. Each darn was visually inspected to assess its per- forrnance. Particular attention was given to evidence of

34

slope instability and seepage below, around, and through the darns. The field investigation was completed during the dry summer months of 1969 making seepage relatively easy to locate and to differentiate from precipitation. An attempt was made to determine from the owners which darns were constructed with a cut-off trench. The general characteristics of the soils and bedrock were noted in the field and compared with geologic and soils maps to determine the soil series and bedrock geology at each site. Mr. James H. Williams, engineering geologist with the Missouri Geological Survey, visited a number of the sites with the investigator and was very helpful in differentiating between the dolomitic formations. The bedrock formations, soil series, the darn parameters, and a summary of the laboratory soils tests for each darn site are presented in Appendix A as Darn Reports. A typical cross-section or cross-sections for most of the darns are also illustrated. 35

IV. MATERIALS

The soils investigated basically fit into three cat­ egories: (1) alluvial-colluvial soils, (2) residuals developed on sandstones, and (3) residuals developed on carbonate rocks. Table II outlines the geologic formations and soil series present at each site. The Sunnen sample is an alluvial deposit or reworked Gasconade clay residium. The Masters #1 and Elsey soils appear to be colluvial deposits also derived from Gasconade clay residium. The Rice, Fabick, and Ft. Westside soils were developed all or in part from the sandstones within the Roubidoux formation. The remaining soils are residuals developed on the Jefferson City, Gasconade, Eminence, and Potosi formations. The clay residuals are classified as the Clarksville gravelly or stony loams. The Hanceville loam and the Lebanon silt loam are characteristic of the non-clay residuals. An attempt was made to collect each sample from the lower B and C horizon. Some, as the Sayers, Terre Du Lac, and Sunrise Central were collected immediately above bed- rock. Depth to bedrock at the other sites was not apparent. The Clarksville cherty residium is the typical soil developed on the carbonate bedrock of the Salem plateau. The percent of chert in the soil is highly variable from site to site and throughout the profile at any one site. The surface and upper horizons often have a heavy 36

TABLE II Materials of Dams Investigated

Soil (Dam) Location Bedrock Soil Series (county) Formation

Timberline St. Francois Potosi Clarksville gv.

Terre Du Lac St. Francois Potosi Clarksville gv.

Masters #1 Dent colluvial (?) Clarksville st. Gasconade Masters #2 Dent Gasconade Clarksville st.

Sunrise South Jefferson Potosi Clarksville st.

Sunrise Cent. Jefferson Potosi Clarksville st.

Elsey Washington Gasconade Clarksville st.

Blackwell Dent Gasconade Clarksville st.

Floyd Dent Gasconade Clarksville st. I Sayers Washington Potosi Clarksville st.

Hornsey Washington Eminence Clarksville st.

Sunnen Washington alluvial Clarksville st. Gasconade Rice Crawford Roubidoux Lebanon slt. lm.

Fabick Dent Roubidoux Hanceville lm.

Ft. Westside Crawford Roubidoux (?) Lebanon slt. lm.

Lt. Prarie Phelps Jefferson Ct. Lebanon slt. lm. 37

accumulation of chert. Sloping surfaces accumulate chert as fines are washed downslope. The Terre Du Lac and Masters #2 samples contain almost no chert, whereas the Blackwell and Floyd samples contain up to 50% chert. The quartz druze (secondary crystalline quartz deposits) typical o£ the Potosi formation was found in the Timberline and Sunrise samples. The clay residuals exhibit many similarities from one location to another. They are generally deep red in color~ an indication of high iron content. A high degree o£ blocky to sub-angular structure is developed throughout the profiles. The natural aggregate structural units of soil particles are called peds. The peds occur from coarse sand to gravel sizes. As a result of this structure which exists throughout the profile, the natural soil deposits are well-drained. There are several factors related to the origin of the soils which may have contributed to forming this pedular structure. The formation of the soil is associated with the solution and lowering of the bedrock level. This lowering is accompanied by an adjustment (drop) in the level of the residual mantle. This adjustment could fracture the soil mass resulting in the ped structure. The slicken-sided ped surfaces may reflect this movement. The ped structure may also be the result of dessica- tion. Later sections of this investigation will show that 38

the residual soils have rather high liquid limit values. It was noted that the field moisture of the soils was closer to the plastic limit than the liquid limit, Table III. This is generally regarded as an indication of dessication. The soils were formed ( and probably continue to be formed at a reduced rate) in a humid warm climate. Baver (1956) notes the specific effect on aggregate formation caused by drying of the clay-soil mass. He related the unequal strains developed in the soil caused by non-uniform drying as a cause of "shrinkage" cracks which form the peds. Wetting of the dried peds causes unequal swelling which results in ped fragmentation and development of smaller peds. Upon wetting, unequal air pressure in the peds caused by capillary adsorption of water also acts to break down the larger peds. The iron may play an important role in prohibiting the breakdown of the peds by both water and mechanical action. Iron acts to "cement" the particles together and/or to increase the cationic concentration and attraction between particles. The alluvial deposits do not exhibit well-developed profiles. The residual clay structure has been modified but not totally destroyed in most cases. As the residuals were water-worked, some of the clay was lost and sand was concentrated. The result is a deposit containing less clay and more sand than the parent residuals. 39

TABLE III Comparison of Natural Field Moisture With Atterberg Limits

Field Soil (Dam) I LL PL Moisture

Fabick 25 16 17 Rice 27 15 14 Lt. Prarie 57 23 26 Masters #1 84 25 35 Masters #2 103 34 44 Floyd 85 35 38 40

V. TEST PROCEDURES

Tests on the Ozark soils were performed to determine results in four general areas: (1) soil gradation and Atterberg Limits, (2) compaction characteristics, (3) clay mineralogy and (4) soil clay chemistry. The procedures used to determine these values are presented in sections essentially as outlined above.

A. SOIL GRADATION 1. Sample preparation. A portion of the field sample, approximately 1 lb., was prepared for soil classification tests essentially as directed in ASTM D421-58, Dry Preparation of Soil Samples for Particle-Size Analysis and Determination of Soil Constants. A Lancaster mixer was used to break down the air-dried clay soil aggregates after the stone portion, where size exceeded that of a No. 4 sieve, had been removed by hand. A mortar and pestle was used to further pulverize the soil for particle-size analysis (No. 10 sieve) and tests for soil index properties (No. 40 sieve). 2. Coarse sieve analysis. The pedular structure of the clayey residual soils investigated in this study does not break down completely during dry or wet sieving. Representative 500 gr. samples which passed the No. 4 sieve were presoaked for 24 hours in a 4% sodium hexametaphosphate (Calgon) solution, mixed in a dispersion 41

cup, and washed through a No. 10, a No. 40, and a No. 60 sieve. The amount of the sample retained on each sieve after oven drying was recorded and calculations made to determine the amount finer than each of the sieves. These values are recorded in Table IV. 3. Hydrometer analysis. Hydrometer analyses were performed as recommended in ASTM D422-63 except that the sedimentation cylinders were not placed in a constant temperature bath during tests. Temperature corrections were made later. After sedimentation, the samples were passed through a No. 60 and a No. 200 sieve for comparison with the hydrometer analysis. The results of the particle-size analyses were plotted on grain-size distribution charts which are presented in Figures 3 - 8. The results are summarized in Table IV.

B. ATTERBERG LIMITS TESTS

1 . Liquid limits. Liquid limit test samples were prepared and values determined in accordance to ASTM D423-66 except that a Casagrande grooving tool was used. Water was added to the pulverized samples until the soil appeared to be slightly wet of liquid limit and then allowed to age for 24 hours to allow water to be adsorbed by the clay particles. The liquid limit tests were performed wet to dry. 2. Wet liquid limits. Graham (1969) noted a sizeable difference between the wet and air-dried liquid limit 42

values of a residual soil similar to the soils investigated in this study. Several of the thesis soils were tested to determine if a similar pattern was noted. The wet liquid limit was determined by taking the soil at field moisture, soaking it for 24 hours in distilled water, mixing in a dispersion cup, and passing the mixture through a No. 40 sieve. The prepared mixture was then placed in a dish and allowed to air dry to a consistency just wet of the liquid limit. The liquid limit was then determined as previously. These values are reported along with normal air dry values in Table V, and compared with Graham's results for a similar soil. 3. Dispersed liquid limit. In another portion of this study, samples were mixed with sodium hexametaphosphate, a dispersant, prior to performing compaction tests. Small portions of these air-dried samples were prepared and tested to obtain Atterberg Limits as described in the previous sections. It was noted that the samples were particularly sen­ sitive to manipulation and that the liquid limit values vary depending on how long the samples were manipulated before testing. An attempt was made to be consistent in preparing and mixing samples in each test, but it is suggested that the values obtained are only approximate. These values are reported in Table VI along with normal air dry values for the same soils. 43

TABLE IV Summary of Particle Size Analysis

PERCENT FINER Soil (Darn) #10 #40 #60 #200 10]J 2]J l]J

Terre Du Lac 100.0 99.9 99.4 98 94 92 90

Hornsey 99.4 98.6 98.5 97 95 93 91

Timberline 99.5 94.8 91.2 85 80 79 78

Sunrise Cent. 98.9 94.9 93.4 92 87 83 82

Sunrise South 98.8 92.5 91.0 89 80 72 70

Sunnen 99.6 96.9 92.3 87 57 38 37

Elsey 99.2 97.6 96.0 94 86 71 69

Masters #1 100.0 99.8 99.6 96 83 76 71

Masters #2 100.0 99.6 99.4 99 99 95 94

Sayers 99.8 99.1 98.6 98 96 93 90

Blackwell 98.6 95.3 93.0 85 77 64 56

Ft. Westside 99.8 98.2 94.2 83 56 38 37

Rice 99.8 88.4 70.8 58 35 21 18

Fabick 99.6 94.0 78.0 61 38 19 16

Lt. Prarie 99.5 97.8 97.8 95 80 58 44

Floyd 99.3 98.0 98 98 97 87 83 Sieve Size No. 4 No. 10 No. 40 No. 200 100 - I' II I I I I r--- ~---- I I ---~ @ ~ I I I I I : ~ I 80 I I I r------~ I II I I I I I +J I ..c: I u I L 0 ~ bO I I I •rl I I , I I I .0 I I 1 ~

~ ~ FIGURE 3. Particle-Size Analysis of Masters #1 and Masters #2 Sieve Size No. 4 No. 10 No. 40 No. 200 100 fT I I r--...... I I -- ~ I 0 I I ...... I ~ ..__ 80 : l I I II I I I I I +J I I .£.: ! 0 bO I I I •rl I I Q) ,! I ~ 60 ! I I :>. I I .0 I I ! II I H Q) I I I I s:: I I •rl 40 I ! ~ II I I I I +Js:: I I I Q) u II I I I H 1 Terre Du l ac I I I 20 I I ~ 1l 2 Timbe line I I I II I I I I T I I I I I 0 II I I 0 0 rl rl r-l r-l 0 rl 0 0 rl 0 0 0 0 Particle Diameter in ~1illimeters .j:lo. U'1 FIGURE 4. Particle-Size Analysis of Terre Du Lac and Timberline Sieve Size No. 4 No. 10 No. 40 No. 200 100 I' I' --...... I I - I It ~ K r--..... II I """""' ~ --- I I I r---...... I ~ 80 II I I r----.::: -...... 0 . I I ..0 I I I ' ll I 1\" ~ Q) I I I I \ s:: I I I •r-l 40 I r..... lj I .f-1 1 Sayer I I I:: I I I Q) () I I I 11 ~ 2 Littl Pra jrie I I ~ 20 I I II I I 3 Black, ell I I I I I I I lj I I I l 0 _l _l ~ r-1 0 0 r-1 r-1 r-1 0 r-1 . 0 0 r-1 0 . 0 0 . 0 Particle Diameter in ~1illimeters .p. 0\ FIGURE 5. Particle-Size Analysis of Sayers, Little Prarie, and Blackwell Sieve Size 100 No. 4 No. 10 No. 40 No. 200 I I ~ I ~ I ,! ~ I ~ II I I "" l I !'...... 80 I 1\ ~ Q" II " II ' I~ I I ..c:+' I !\ ! ~ bO I •r-i II I \ I (I) I I 3 60 It II ! ~ ,.,. 1\ I :>. I \ ..0 I I ! ~ ~0 ~ I (I) I I I " J:! I I •r-i 40 I ! "" 1\ ~ 11 I " -... 1 Floyd I I +' I (3)' !'--.. J:! I I I ~ (I) '" {) I I I ~ 2 Ft. W stsi de 'I I I I l ~ 20 II " r--- 3 Rice I I I I I I I I I "' I I I 0 II ! l ,..; 0 0 ,..; ,..; ,..; 0 ,..; . 0 0 ,..; 0 0 0 . 0 Particle Diameter in ~1illirneters

.j::. -....] FIGURE 6. Particle-Size Analysis of Floyd, Ft. Westside, and Rice Sieve Size No. 4 No. 10 No. 40 No. 200 100 II r--.:: I I I r-::: : I II ~ ~ I T- ---...... ___ I I I I ~ r-- 0 80 I i \ I I II I I I I I ~'---._ f.J I II I ! ' ..c:: \ •0' bO I I I •o-1 I I (l) ~ :3 60 ! ! r---..!.. I I I :>, I I ..0 ! I I H (l) I I I I s:: I I I •o-1 40 I \ ~ [I .I "' I I f.J 1 Sunri: e Ce ~tra : I s:: !I I I 0' (l) {) II I I I H 2 Sunrit e So ~th I I I ~ 20 I ! " I I " I'-- I 3 Fabid I : I I I "" I I I I I I I 0 tl I I rl 0 0 rl rl rl 0 0 rl 0 rl 0 0 0 0 Particle Diameter in :1i llimeters

~ 00 FIGURE 7. Particle-Size Analysis of Sunrise Central, Sunrise South, and Fabick Sieve Size No. 4 No. 10 No. 40 No. 200 100 'I II I I I .....,._I G) I ---~ - r---. ~ II I ""'" ...... I I I H-- I 80 I I I ~ 0 !I I I I I I \ .f.J I .c: I I I \ ' ~ bO I I I I - •.-! I Q) I I :3 60 I I ! \ I I I I ~ I I ..0 I I I ' 1\. ~ Q) I I I I 1=: I I •.-! 40 I ! ~ JJ... I I ..___ I I I " (j)' .f.J 1 Horns ~y 1=: I I I I Q) () I I I ~ 2 Elsey I I I I ! ~ 20 I I I I 3 Sunne~ I II I I I I I I I I I I 0 tl II I I rl r-1 0 0 rl rl 0 0 0 rl 0 rl 0 0 0 Particle Diameter in 11illimeters ... \.0 FIGURE 8. Particle-Size Analysis of Hornsey, Elsey, and Sunnen 50

TABLE V Comparison of Wet Liquid Limit And Natural Atterberg Limits

% Change Soil (Dam) Wet LL LL PL in LL

Terre Du Lac 132 96 35 37 Hornsey 121 90 43 34 Sunnen 52 42 21 24 Springfield 144.8 105.6 34.0 37

* from Graham (1969)

TABLE VI Comparison of Natural and Dispersed Atterberg Limits

Natural Dispersed I Soil (Dam) LL PL PI LL PL PI

Terre Du Lac 96 35 61 73 34 39 Floyd 85 35 50 55 28 27 51

TABLE VI I Summary of Atterberg Limits Tests

% Soil (Dam) LL PL PI Clay

Timberline 69 30 39 79

Terre Du Lac 96 35 61 92

Masters #1 84 25 59 76

Masters #2 103 34 69 95

Sunrise South 68 30 38 72

Sunrise Central 70 33 37 83

Elsey 84 25 59 71

Blackwell 60 23 37 64

Floyd 85 35 so 87

Sayers 105 41 64 93

Hornsey 90 43 47 93

Sunnen 42 21 21 38

Rice 27 15 12 21

Fabick 25 16 9 19

Ft. Westside 48 21 27 38

Little Prarie 57 23 34 58 52

80

0 0 Clayey Residual Soils D 60 Sandy Residual Soils 0 0

"'"'~ '--"

~ QJ '"0 l=l H C) 0 ·r-l +J 0 (/) Cll r-1 A< D 20

D D

0 0 20 40 60 80 100 Liquid Limit (%)

FIGURE 9. Plasticity Chart Plot of Ozark Soils 53

4. Plastic limits. Water was added to the powdered samples until the approximate consistency of the plastic limit was reached; the samples were then aged for 24 hours. The plastic limit was determined in accordance with ASTM D424-59, and the plasticity index calculated. The Atterberg Limits for all soils investigated are reported in Table VII. Figure 9 illustrated how these values plot on the Plasticity Chart.

C. MOISTURE-DENSITY RELATIONSHIPS 1. Natural samples. Soils for compaction tests were prepared and tested in accordance with ASTM D698-70, Moisture-Density Relations of Soils Using 5.5-lb. Rammer and 12-in. Drop. Method A in which all material larger than the No. 4 sieve is discarded, was followed. After mixing with an amount of water predetermined to produce the desired moisture, the samples were aged 24 hours before testing. 2. Dispersed samples. Approximately 20 to 25 lb. samples of several soils were crushed to pass the No. 4 sieve. Each sample was soaked in a 4% solution of sodium hexametaphosphate (Calgon). The mixture was stirred to assure that all of the sample was wetted. After 24 hours of soaking, the sample was further worked by mixing in a dispersion cup. Additional distilled water was required to dilute the mixture to a consistency which could be handled by the dispersion apparatus. Each 54

TABLE VI I I Result of Moisture-Density Tests -- Soil (Dam) Maximum Optimum Dry Density Moisture (pcf) (%) -- Timberline 82 36 Terre Du Lac 79 37

Masters #1 84 25

Masters #2 81 36

Sunrise South 88 31 Sunrise Cent. 85 33 -- Elsey 89 26

Blackwell 92 24 Floyd 83 34

Sayers 74 42

Hornsey 77 38

Sunnen 108 18 Rice 121 13 Fabick 121 13

Ft. Westside 100 22

Lt. Prarie 94 27 Floyd - D 30 87 -- Sayers - D 75 35 T. D. Lac - D 80 33 Sunnen - D 107 17 55

portion of the sample was mixed until all lumps had disap­ peared, about 5 minutes. The dispersed mixture was then placed in two, 2' by 2' pans to a depth of less than 1 inch. The pans were placed before a 20 inch fan and a small electric heater until the soil dried; generally 3 days were required. The dried soil was again crushed to pass a No. 4 sieve, the moisture content determined, and the sample separated into portions. Predetermined amounts of water were mixed with each portion of soil to attain the desired moisture content and the soil placed in plastic bags to age. After 24 hours the samples were compacted by a procedure identical to that used for the standard moisture-density tests. The results of all compaction tests are given in Table VIII. Portions of the dispersed, dried, and crushed soil mixtures were saved for Atterberg Limits testing.

D. X-RAY DIFFRACTION 1. General. A number of authors including Bragg (1949), Brindley (1951), Grim (1953), and Bradley (1964) discuss the basic principles of x-ray diffraction theory and technique. The essence of these discussions are out­ lined by Schmidt (1965) where it is noted that clay mineral analysis using x-ray diffraction is basically a determina­ tion of the thickness of a unit sheet of the mineral. The angle of incidence at which reinforcement occurs is directly related to the thickness of a monomolecular layer of the TABLE IX X-Ray Machine Specifications

Location of Type of Time Delay Counts per Machine Machine (seconds) Second

Geology Dept. Norelco Wide-Angle 4 100 - 200 i UMC Diffractometer

Geology Dept. General Electric 2 - 4 1000 - 2000 UMR Model XRD-5

U1 0\ 57

mineral. The results of x-ray scanning are recorded on a chart called a diffractogram which graphs twice the angle of incidence of the x-ray against the magnitude of the count of reflected x-rays. For economic reasons, two x-ray machines were utilized in this study. This is not ideal since analysis and comparison of diffractograms from different machines is difficult. Table IX gives pertinent data for the machines used. 2. Preparation of samples. An attempt was made to prepare sample slides using a procedure similar to that described by Schmidt (1965) where 50 gram portions of soil were soaked in distilled water for one week, mixed in a dispersion cup, and sedimented in a 1000 ml. hydrometer cylinder. Some samples flocculated and settled rapidly. The supernatent water was poured off, fresh distilled water added, and the sample sedimented again. After 3 cycles of mixing with fresh distilled water had failed to disperse these samples, 5 grams of sodium hexametaphosphate (Calgon) had to be added to the mix for the samples to be success­ fully dispersed. According to Stokes' Law, after 24 hours all particles larger than 2 microns (all sand and silt) had settled out of suspension and a portion of the clay fraction was siphoned off and concentrated by evaporation. A few drops of the concentrate was pipetted onto a 1" X 3" X 1 mm. 58

microscopic slide and dried before a heater. The clay

plates tend to orient themselves upon drying.

It was this tendency of the clay plates to so com­

pletely orient themselves which rendered the sedimentation method unsatisfactory in this study. As shown in Brindley

(1951), Bramao (1952) and Grim (1953), the comparison and

identification of the kaolinite group minerals required

in part an analysis of the 020 peak which is not developed

in the oriented sedimented slides.

3. Powdered samples. Samples, pulverized to pass a No. 200 sieve, were mixed with denatured alcohol before being poured onto the 1n X 3" slides. As the alcohol evaporated, a film of sample adhered to the slide. Best results were obtained when a thin film of powder was pro­ duced. This was especially evident at low 8 angles where

a thick film placed the reflection surfaces inside the focal plane of the scanner resulting in a scattering of the x-ray. Powdered samples also develop a preferential orienta­ tion due to their platy structure, but there is enough random orientation to allow development of other peaks not associated with the c-axis, including the 020 peak required in analysis of the kaolin group (Figures 10 and 11).

4. Potassium acetate treatment. Halloysite (4H20) 0 with a basal 001 spacing of about 10.0 A is known to 0 collapse to around 7.2 A as the soil is dried. Early 59

researchers thought this process was irreversible. Wada (1961) showed the process was reversible and expanded the 0 kaolin group mineral lattice to 14.1 A using potassium acetate. Upon washing with distilled water the kaolinites 0 collapsed back to 7.2 A whereas the halloysites collapsed 0 to about 10.1 A retaining a water molecule in the expanded lattice. Miller and Keller (1963) refined the procedure by using ethylene glycol which helps to retain the expanded 0 10.1 A lattice since ethylene glycol evaporates more slowly than water. In accordance with this method, with some modifica- tions discussed with one of the authors C Keller, 1970), 1.5 grams of potassium acetate was mixed in a small mortar and pestle with 2.0 grams of air-dried soil passing the

No. 200 sieve. This mixture was allowed to rest for 24 hours in a high humidity environment, after which it was washed several times through filter paper with distilled water. A final wash using a solution of equal portions of distilled water and ethylene glycol was used. The washed sample was diluted with ethylene glycol and placed on a glass slide in a thin film. The slide was subjected to x-ray scanning while the film was still moist. It is noted that the specimens prepared in this manner remained moist for longer than 24 hours. The method of sample preparation used in this study was tested to check its workability. A sample of pure 60

Georgia Kaolinite was diffracted to determine its base pattern. Then a sample was x-rayed which had been treated with potassium acetate and ethylene glycol as described 0 above. There was no shift in the 7.2 A peak indicating no permanent lattice expansion of the reference kaolinite. A sample of pure Indiana Halloysite was x-rayed along with a treated sample, Figure 12 shows the results for this reference mineral with a definite shift in the basal 0 0 001 peak from 7.2 A to 10.0 A for the treated sample. This indicated the sample was halloysite and that the method of sample preparation used in this study did produce the expected results in the differentiation between control specimens. The results of similar tests on some of the Ozark soils are described and presented in a later section.

E. DIFFERENTIAL THERMAL ANALYSIS

1 . General. Differential thermal tests were performed at the U.S. Bureau of Mines, Rolla, in two furnaces designed to conform with the specifications outlined by Berkelhamer (1944). Nickel specimen holders with a capacity of approximately 0.5 grams were used with chromel-alumel thermocouples, recommended for their sensitivity. A Leeds and Northrup automatic programmer was used to control the rate of heating, 12°C/min. A Leeds and Northrup multiple thermal differential recorded of the automatic pen-and-ink type recorded the curves. 61

Samples of Georgia Kaolinite and Indiana Halloysite were tested to obtain reference curves. They are shown in

Figure 13. Grim and Rowland (1942) show a number of curves

for various clay minerals which can be used for comparison with test curves of the Ozark residual soils.

2 . Sample preparation and testing. Specimens were prepared for DTA by sieving a sample of air-dried soil

through a No. 200 sieve. Mackenzie (1957) notes that

sieve size has little influence on the curves but that

the sample should be fairly fine (100 - 200 mesh) to

insure good packing of the specimens.

The nature of the low temperature peak system is dependent to a degree upon humidity variations, therefore,

the use of air dried samples may result in erroneous low

temperature peaks (Mackenzie, 1957). Similarly, using

samples which have been oven-dried at about 100°C may lead

to the eradication of the low temperature peaks altogether.

Mackenzie suggests that samples be equilibrated over a sat­ urated solution of Mg(N0 3 )·6H20 in a dessicator for several days. This procedure standardizes the moisture

content of the test samples. Eades (1970) suggests that the procedure while important for the triple-layer minerals

is not required for the kaolinite group minerals. The

samples tested in this study were not pretreated as described above. Where small or very reactive samples are tested, they 62

TABLE X Summary of Major DTA Temperature Peaks

Peak Temperatures (OC) Low Major Soil (Dam) Endothermic Endothermic Exothermic -- Masters #1 100 560 minor ? Masters #2 120 550 925

Sunnen 90 ? 545 905 Terre Du Lac 115 560 925 Sunrise South 120 550 920 Sayers 120 545 915 Blackwell 110 535 minor ? Timberline (1) 118 555 916 Timberline (2) 120 555 915 Timberline - leached 120 540 945

Hornsey 105 ? 550 915 Hornsey - leached 128 562 930 Elsey (1) 125 540 920 Elsey (2) 110 545 920 Floyd 110 560 930 Floyd - oven dried 3 935 hr. at 700C no peak 530 Indiana Halloysite 120 580 973 same - oven dried 12 595 987 hr. at 70°C 135 Georgia Kaolinite (1) no peak 605 966 Georgia Kaolinite (2) no peak 610 970 63

are often mixed with an inert material such as calcined kaolinite or aluminum which has been ground to a particle size which approaches that of the sample. Base line drift and sample shrinkage can often be corrected with this procedure. The DTA base line is a horizontal line along which the curve should return after recording each endothermic and exothermic reaction. A disadvantage of mixtures with inert materials is that they effect the areas under the peaks which distorts quantitative determinations. Samples tested in this study were not mixed with inert material. Some base line drift was encountered in early tests, Figure 14, but this was corrected by packing the samples more firmly in the specimen holder. Factors such as heating rate, size and construction of specimen holders, and the degree of packing can radically alter the resulting DTA curves. Consistency of technique is the key to obtaining curves which can be compared and analyzed. Figures 14 thru 17 show DTA curves for a number of the Ozark soils. Table X summarizes the temperatures of the important reactions.

F. THERMOGRAVIMETRIC ANALYSIS Portions of the minus 200 mesh samples prepared for DTA were saved for thermogravimetric analysis, TGA. A Ugine-Eyraud Model B-60 Automatic Recording Balance and furnace were used. The heating rate, 10°C/min., was 64

programmed by a Bristol's Model 565 Round-Chart Dynamaster Pyrometer. Test results were recorded by a Bristol's Model 760 Strip-Chart Dynamaster. The equipment and method of testing is discussed at length in Duval (1963). One hundred mg. samples of the powder were placed in the sample holder in the furnace. The balance was zeroed and the sample heated to approximately 50°C which temper- ature was held until there was no further noticeable weight loss. It is believed that most of the hygroscopic moisture is expelled in this manner. A heating process was utilized in which the temperature was raised in steps of about 50 0 C and held until the weight was stabilized. The tests were terminated at 1000°C, The data has been transformed to dehydration curves. Curves for the standard minerals, Georgia Kaolinite and Indiana Halloysite, were determined, converted to dehydra­ tion curves, and plotted along with the curves for several of the Ozark soils in Figure 18.

G. TRANSMISSION ELECTRON MICROSCOPY (TEM) A Hitachi Model HU-llA transmission electron microscope was used to study the size and shape of clay particles of several soils in this investigation. Both powdered and colloidal suspension specimens were prepared for analysis. The major reference for transmission microscopy is an extensive study performed for the American Petroleum Institute (Davis et al, 1950). Later studies by Bates and 65

and Cromer (1955), Brindley and Cromer (1956), and Bates and Cromer (1959) concentrate on applying transmission microscopy in determining the morphology of the kaolin group minerals. A thin carbon film was evaporated (sputtered) onto glass slides in a vacuum chamber. The powdered sample (finer than 200 mesh) was dusted onto the film which was then floated off of the glass slide in a beaker of water. An electron microscope grid was used to lift the carbon film with sample from the beaker and mount the specimen in the microscope. The magnified image of the particles so mounted was projected onto a screen for viewing and glass plate photographic slides were obtained. The 200 mesh powder specimens produced images of aggregates of clay particles rather than of individual particles. These were difficult to analyze for particle parameters such as size and shape. Soils which had never been air-dried were suspended in distilled water for a week before testing. They were agitated often to separate the particles. Before testing, the suspension was again agitated and a drop of the mixture placed on the floated carbon film. The sample was then mounted as before. This method of sample preparation resulted in a much better segregation of the particles. Photographs taken of the soil particles when utilizing this method of sample preparation are presented in Figures 19-22. 66

H. SCANNING ELECTRON MICROSCOPY (SEM) The scanning electron microscopic investigation was conducted at the Materials Research Center, University of Missouri - Rolla. A JEOLCO (Japan Electron Optics Labora­ tory Company, Ltd.) Scanning Type Electron Microscope, Model JSM-2 was used. Samples were mounted on copper disks, placed in a vacuum and subjected to electron scanning. Micrographs were taken with a mounted Polaroid camera. Initially, specimens were prepared by mounting air­ dried pulverized minus 200 mesh samples on the copper disk with silver paint. Figures 23 and 24 present several micrographs taken from this portion of the study. Borst and Keller (1969) present a technique of mounting samples for scanning with a minimum of sample disturbance. Air-dried peds were broken to expose a fresh surface. These chips, about 1 em. by 1 em., were mounted with silver paint on the disks and coated with a thin film of gold in a vacuum chamher. The specimens were then exposed to scanning. Micrographs resulting from this portion of the study are presented in Figures 25 thru 29. Compacted clay samples were also analyzed. Chips with fresh surfaces, as described above, were prepared for testing. These chips were mounted to expose to scanning planes oriented perpendicular to and parallel to the dir­ ection of applied compactive effort. Micrographs of this TABLE XI Chemical Analysis of Ozark Soils U.S. Bureau of Standards M.G.S. Lab Analysis Name of Lake Sample #98 S amp 1 e # 9 8 Blackwell Plastic Clay Plastic Clay Analysis %

Si02 59.11 57.90 50.40 Al 2o3 25.54 27.28 24.48 Fe 2o3 2. OS 2.33 6.81 Ti02 1. 43 1. 03 0.46 MnO 0.005 --- -- CaO 0.21 0.16 0.50 MgO 0. 72 0.70 1. 60 Na 2o 0. 2 8 0.35 1. 97 K20 3.17 2.40 1.15 Ignition Loss 7.28 8.17 12.18 -H2o 105°C 0.84 3.93 + H20 1000°C 7.33 8. 2 5 HCl Insol. Residue 83.72 ----- HCl Soluble Si02 0.71 ---- - HCl Soluble Al 2o3 5.36 --- -- HCl Soluble Fe 2o3 1. 04 7.11 0\ Oxides of Al, Fe, 30.64 31.60 '-] Ti, Mn, P. Totals 100.32 99.40 TABLE XI - continued

Name of Lake Sayers Sunnen Sunrise South

0 0 Analysis % ~0 'o

Si02 40.86 53.06 41.04 Al 2o3 27.06 23.53 27.97 Fe 2o3 9.99 7.84 11. 49 Ti02 0.43 0. 71 0.46 MnO ------CaO 0.46 0.18 0.38 MgO 2.20 1.10 1. 31 Na 2o 0.23 0.20 0. 13 K20 0.94 1.14 0. 7 2. Ignition Loss 17.96 12. 51 16. 55 - H2 0 105°C 7.04 4.97 6.36 + H20 1000°C 10.92 7.54 10.19 HCl Insol. Residue 41. 52 ------HCl Soluble Si02 0.79 ------HCl Soluble A1 2o3 25.57 ------HCl Soluble Fe 2o3 9.98 7.59 11.18 Oxides of Al, Fe, 37.48 32.08 39.92 Ti, Mn, P. 0\ Totals 100.13 100.27 100.05 00 TABLE XI - continued

Name of Lake Timberline Terre du Lac Hornsey

Analysis % % !!:0

Si02 42.04 42.6 8 42.32 A1 2o3 28.08 28.60 29.21 Fe 2o3 9.99 10.16 8.33 Ti02 0.51 0.46 0.36 MnO ------GaO 0.4.8 0.40 0.40 MgO 1. 30 1. 07 1. 61 Na 2o 0.13 0.18 0.13 K2o 1. 03 0.65 0.87 Ignition Loss 16.13 16.05 16.73 - H 0 2 105°C 5.59 5.73 6.03 + H20 1000°C 10.54 10.32 10.70 HCl Insol. Residue ----- 45.20 43.11 HCl Soluble Si02 ----- 0.75 0.87 HCl Soluble Al 2o3 ----- 26.47 27.56 HCl Soluble Fe 2o3 10.06 10.47 8.91 Oxides of Al, Fe, 38.58 39.22 37.90 Ti, Mn, P. (]\ \.0 Totals 99.69 100.20 99.93 TABLE XI - continued

Name of Lake Floyd Masters #1 Elsey

~ Analysis % % 0

Si02 52.20 56.88 50.68 Al 2o3 24.24 21.17 22.76 Fe 2o3 7.16 6.16 7.42 Ti02 0.46 0.41 0.46 MnO ------CaO 0.10 0.37 0.42 MgO 1. 2 7 1. 58 2.15 Na 2o 0.13 0.13 0.23 K20 1. 30 1. 20 1. 33 Ignition Loss 11.96 12.44 15.08 - H 0 2 105°C 3.38 4.81 6.44 + H20 1000°C 8.58 7.63 8.64 HCl Insol. Residue ------HC1 Soluble Si02 ------HCl Soluble Al 2o3 ------HCl Soluble Fe 2o3 7.15 6.15 6.95 Oxides of Al, Fe, 31.86 27.74 30.64 Ti, Mn, P. "'-J 0 Totals 98.66 100.30 100.45 TABLE XI - continued

Names of Lake Sunrise Central Masters #2

Analysis % %

Si02 40.96 46.30 Al 2o3 29.44 24.68 Fe 2o3 10.41 9.66 Ti02 0.41 0.36 MnO ------CaO 0.52 0.64 MgO 1. 38 1. 77 Na 2o 0.13 0.16 K2o 0.73 1. 00 Ignition Loss 16.29 16.56 - H 0 2 105°C 5.08 6.79 + H20 1000°C 11.21 9.77 HCl Insol. Residue ------HCl Soluble Si02 ------HCl Soluble Al 2o3 ------HCl Soluble Fe 2o3 10.38 8.78 Oxides of Al, Fe, 40.26 34.70 Ti, Mn, P. '-I Totals 100.28 101.17 "'"' 72

portion of the study are presented in Figures 48, 49, and SO.

I. LOW-POWER MICROSCOPY (LPM) A low-power (10-SOX) microscope study of the compacted samples was conducted at the U.S. Bureau of Mines, Rolla, Missouri. Micrographs were taken at 17X both in black-and­ white and in color of samples of Terre Du Lac compacted 3% dry of optimum and 4.5% wet of optimum. Micrographs were taken with a Polaroid MP-3 Land Camera adapted to accept Bausch and Lomb lenses. A 72 mm lense was used. These micrographs are presented in Figures 46 and 47.

J. SOIL CLAY CHEMISTRY 1. General. All samples tested for chemical properties were first air-dried and passed through a No. 4 sieve. Where further preparation was required for specific tests, it will be noted in the discussion.

2 . Chemical analysis. In general, methods of chemical analysis as outlined by Scott (1948), Standard Methods of Chemical Analysis, were followed. All tests were performed on the clay fraction of the soils and major tests are out- lined below with results presented in Table XI.

a. Si02 - gravimetric. Clay is fused with sodium car- bonate (Na 2co 3), taken up in dilute hydrochloric acid, dehydrated, filtered, ignited and silica volatilized with hydroflouric acid. (Scott, p. 797-801) 73

b. Al 20 3 - gravimetric. filtrate from Si0 2 is used. Ammonium hydroxide is added to precipitate hydro­

xides of Al, Fe, Ti, etc. Al 2o 3 is calculated by difference after determination of Fe 2o 3 , Ti02 , P2o5 , and MnO. c. Fe 2o 3 - volumetric. Oxides are fused with potas­ sium acid sulfate (KHS0 4), taken up in dilute sulfuric acid (H 2so 4), reduced with test lead, and titrated with potassium permanganate.

(Scott, p. 470)

d. Tio 2 - colorimetric. Determination of titanium with hydrogen peroxide. (Scott, p. 987) e. CaO - volumetric. Oxalate is added to filtrate from precipitation of Al, Fe, hydroxides, etc. forming a precipitate of calcium oxalate - dissolved

in dilute H2so4 and titrated with potassium p erma n g an ate . ( S cot t , p . 21 0 ) f. MgO - gravimetric. Ammonium acid phosphate

(NH 4) 2HP0 4 is added to filtrate of calcium oxalate precipitation - forms precipitate which is ignited to form magnesium pyrophosphate from which MgO is calculated. (Scott, p. 532)

3. Soil pH in water. Various soil-water ratios have been used in pH determinations with a pH increase noted as the soil-water ratio decreases. As recommended by Peech (1965) in Methods of Soil Analysis a 1:1 ratio was used in 74

TABLE XI I Summary of pH and Cation-Exchange Capacity of Ozark Soils

Soil (Darn) pH CEC (rneq/100 gr) Si02/Al 20 3

Timberline 5.1 27.8 2.55 Terre Du Lac 7.1 27.9 2.53 Masters #1 5.4 21.9 4.57 Masters #2 7.5 35.8 3.20 Sunrise South 6.5 24.6 2.52 Sunrise Central 4.9 30.6 2.38 Elsey 7. 7 26.9 3.79 Blackwell 7.5 24.1 3.50 I Sayers 7. 5 44.9 2.56 Hornsey 4.8 42.1 2.46 Sunnen 4.9 11.5 3.82 Rice 5.2 8.7 ---- Fabick 5.6 7.6 ---- Ft. Westside 4.6 19.1 ---- Little Prarie 7.8 13.7 ---- Floyd ------3.66 75

this study. Twenty grams of soil were added to 20 grams of distilled water, stirred for several minutes, and allowed to sit for one hour. A Corning Model 12 Research pH Meter with a glass electrode was used to measure the pH. Two readings were taken and their average recorded in Table XII. 4. Cation-exchange capacity. The cation-exchange capacity, CEC, was determined essentially as outlined by Chapman (1965). The method utilizing ammonium acetate to saturate the sample was used. Two hundred ml. of 1 N ammonium acetate were filtered through 10 grams of soil placed in a Gooch crucible over a 500 ml. Erlenmeyer flask. The leachate was saved for fur- ther analysis. The soil was then filtered with 100 ml. of ethyl alcohol to wash out the excess ammonium acetate. The leached soil was then placed in a Kjeldahl flask with about 30 grams of magnesium oxide, a catalyst, and 200 ml. of distilled water and boiled for distillation. Fifty ml. of 4% boric acid was mixed with 150 ml. of the distillate with the mixture then titrated with .1 N hydrochloric acid. The amount of acid in ml. required to reach the end point is noted and recorded as the value of the cation-exchange capacity in meq/100 grams. Two samples of each soil were tested, the results averaged and reported in Table XII. 5. Flame photometry. Na+ and K+ were determined by use of a Coleman Flame Photometer with a Coleman Electric 76

TABLE XIII Concentration of Mg, Ca, Na and K Cations in meq/100 grams

Soil (Dam) Mg++ Ca++ K+ Na +

Timberline 9.3 10.1 0.50 0.40

Terre Du Lac 10.1 12.2 0.40 0.90

Masters #1 11.7 5.8 0.36 0.39

Masters #2 20.6 11.6 0.76 1. 50

Sunrise South 11.0 9.4 0.48 0.66

Sunrise Cent. 9.3 9.0 0.75 0.46

Elsey 21.3 6.5 0.79 0.41

Blackwell 14.6 13.6 0.60 0.30

Sayers 34.3 12.5 0.71 1.15

Hornsey 13.5 10.8 0.90 0.41

Sunnen 2. 8 1.4 0.45 0.20

Rice 1.4 5.1 0.31 0.19

Fabick 1.8 2.7 0.20 0.21

Ft. Westside 3.6 4.0 0.42 0.22

Lt. Prarie 7. 7 15.3 0.38 0.68 77

Colorimeter. The theory is explained in the article by Rich (1965) in Methods of Soil Analysis. The leachate from the CEC determination was placed in a 250 ml. volumetric flask and brought to volume with the addition of ammonium acetate. Standard solutions contain- lng. k nown amounts o fN+ a an d K + ..ln ammonlum acetate were tested and the results used to construct graphs of colori- meter readings versus ion concentration of meq/100 gr. of soil. Two samples of each soil were then tested for each ion, correlated with the graphs to determine concentration, the values averaged, and reported in Table XIII. 6. Absorption spectrophotometry. The leachate from the CEC determination was also used in the determination of ++ ++ Mg and Ca . Five ml. of the leachate was placed in a 50 ml. volumetric flask and brought to volume with 1000 ppm. lanthanum. A Jarrell-Ash Atomic Absorption Flame Emission Spectrophotometer was used to analyze the samples. Standard solutions of Mg++ and Ca++ were tested to construct graphs relating the cation concentrations and the spectra- photometer values. Two samples of each soil were tested, the values averaged, and reported in Table XI I I. This method of testing is described by Prince (1965) in Methods of Soil Analysis.

K. SPECIFIC GRAVITY Specific gravity tests were performed essentially as outlined by ASTM D854-58 (1965). The results of testing 78

TABLE XIV Specific Gravity of Ozark Soils

Soil (Dam) Specific Gravity

Fabick 2.68 Rice 2.70

Ft. Westside 2. 7 2 ' 2.72 Sunnen I 2. 7 2' 2.73 Blackwell 2.74 Lt. Prarie 2.77 Elsey 2.79

Floyd 2. 8 0' 2.80 Sunrise Cent. 2.82 Sunrise South 2. 8 2 Terre Du Lac 2.82 Timberline 2.83 Masters #1 2.83 Masters #2 2.84

Sayers 2.84

Hornsey 2.84, 2.84 79

are presented in Table XIV. The dessicated nature of the samples required some special procedures to be followed to assure sample dispersion. The samples tested were air­ dried. Two samples identical to the one tested were oven- dried for 24 hours at 105°C. The average moisture content for the two samples was calculated and used as the moisture content for the tested sample. The samples were soaked for longer than 24 hours in distilled water before testing. The method of air removal by boiling the sample for 10 minutes was followed. As Table XIV indicates, several soils were tested twice and reproductive results were obtained~

L. CORRELATION TESTS Data to be correlated was analyzed by a computer program written by Capt. W. Vogel, U.S. Army Corps of Engineers, a graduate student in Civil Engineering at the University of Missouri - Rolla. A P-PLOT option was implemented to plot the data and curves developed by the analysis. A least squares polynomial of the form:

Y = A(O) + A(l)X + A(2)X 2 + •.• + A(N)XN was used to analyze the data and to evaluate A(O), A(l), ... A(N).

Two values, sigma (a) and the correlation coefficient (r) were used to evaluate the curve fit where:

where: x = X - X

y = y - y 80

TABLE XV Correlations for Clayey Residual Ozark Soils

x- axis y-axis Sigma Correlation Coef. Coordinate Coordinate (a) (r)

Plastic Limit Optimum Moist. 7. 9 0.92 Liquid Limit Optimum Moist. 10.6 0.90 Cation-Exc. Cap. Optimum Moist. 14.3 0.86 % Clay Optimum Moist. I 9.7 0.90

9- D Clay* Optimum Moist. 8.8 0.95 Plastic Limit Max. Dry Density 8.1 -0.89 Liquid Limit Max. Dry Density 19.2 -0.88 Cation-Exc. Cap. Max. Dry Density 23.0 - 0. 8 7

9- 0 Clay Max. Dry Density 5.9 -0.96

% Clay* Max. Dry Density 10.9 -0.97 Optimum Moist. Max. Dry Density 15.1 0.90

Cation-Exc. Cap. * % Clay 155 0.88 -- Cation Exc. Plastic Limit 9. 2 0.91 Cap. -- Cation Exc. Cap. Limit 139 0.81 Liquid -- % Clay Plastic Limit 13.4 0.87

% Clay Liquid Limit 61.1 0.95

*Correlations include values for sandy residual soils 81

and,

2 E(y - Ycalc) where: N = no. of data points a = N - k - 1 k = degree of polynomial Ycalc = calculated y value for a given x value

The correlation curves resulting from this analysis were plotted and presented as Figures 30 thru 45 along with the values of sigma (cr) and the correlation coefficient (r). These values are summarized in Table XV. The correlation coefficient, r, has a range in values from -1.0 to +1.0. Perfect correlation is demonstrated where r = ±1.0. Sigma (cr) is a common statistic used in the earth sciences to express the summation of deviations of data from a curve fitted through that data. The value of sigma which indicates a significant relationship depends on the order of magnitude of values being correlated. 82

VI. ANALYSIS OF RESULTS

A. CLAY MINERALOGY The mineralogy of the residual soils will be discussed in sections corresponding to the method of testing. No method of testing in itself produced definitive data on which to identify the clay mineralogy of the soils, but when they are considered along with the probable environ- ment of soil formation, a clear picture of the mineralogy is developed. 1. X-ray analysis. Several x-ray diffractograms of the soils investigated in this study are shown in Figures 10 and 11. Powder slides were x-rayed to produce the diffracto- grams. The major peaks shown on these diffractograms are located at the following 28 angles: 12.0-12.4 0 , 20.0 0 , 0 21.0°, 25.0-25.4°, and 26.7°. The peaks at 21.0° (4.26 A) 0 and 26.7° (3.34 A) are quartz peaks. The 12.0-12.4° 0 0 (7.15± A) and 25.0-25.4° (2.54± A) peaks represent basal 0 001 and 002 kaolinite reflections and the peak at 20.0 0 (4.45± A) represents the 020 kaolinite mineral reflection. The reflections which occur for some soils at 20 angles of 0 0 24°± (3.71 A) and 33°± (2.72 A) appear to represent the 021 and 022 reflections of the same group. 0 The reflections which occur at 10± A (Masters #1 and Floyd) are probably basal illite peaks whereas those in the 83

0 area of 13.4-15± A (Timberline, Terre DuLac, etc.) probably represent chlorite or montmorillonite group minerals in the soil. The weak peaks indicate that only minor amounts of these minerals are present. The apparent major constituent of the soils studied, based on the spacing~ magnitude and size of the reflection; is a mineral of the kaolin family, probably halloysite or kaolinite. Brindley (1961) describes the x-ray character- istics of each in detail. In attempts to distinguish between the two~ he used the relative intensity of the 0 basal 001 peak and those near 4.4 A, the 020 reflection. For kaolinites, the usual situation is for the magnitude of the 001 reflection to be twice that of the 020 reflection. For poorly crystallized kaolinites, the reflections may be more nearly equal; for halloysite the 020 reflection is usually greater than that of the 001 reflection (Figure 10) . Bramao et al (1952) compares the 002 and 020 reflec­ tions. Where the intensity of the 002 peak is greater than that of the 020 peak, kaolinite is indicated; where the intensity of the 020 peak exceeds that of the 002 peak, halloysite is predominant. In the diffractograms of the residual soils studied in this investigation, the 020 peak intensity exceeds that of both the 001 and 002 basal peaks; therefore, halloysite is indicated. The broadness of the basal 001 and 002 peaks is also a characteristic of hal­

loysites (Grim, 1953). 84

Several features of the x~rays indicate that halley~ site may not be the major clay constituent of the Ozark soils. The 001 peak for most of the samples occurs between 0 7.0 and 7.2 A. This peak is generally located between 0 0 7.35 and 7.5 A for halloysites versus 7.15 A for kaolinites, even poorly crystallized kaolinites. Tests of several samples treated with potassium acetate and ethylene glycol, as recommended by Miller and Keller (1963), indicate no permanent expansion of the crystal lattice from the area of the 001 peak at 7.2 A. Figure 12 illustrates the results of these tests. This indicates that halloysite is not a major constituent of these samples. Grim (1953) discusses three factors which affect the diffractograms of the kaolin minerals. The first is the effect of displacement of the unit sheets with respect to each other, particularly in the a and b~axis directions. The displacement results in broad diffraction patterns rather than sharp peaks. Secondly, a particular kaolin mineral may be interlayered with another kaolin mineral or a number of other clay minerals. Weakened and displaced peaks result. Finally he notes that the shape and size of particles affect the diffractograms.

Brindley (1961) discusses the affects of b~axis disordered kaolinites (similar to the kaolinite pM of

Mackenzie, 1957) on x~ray patterns. He states that the 85

Treated with Potassium Acetate

Untreated

002

8 6 30 26 24 22 20 18 16 14 12 10 2 8 in Degrees

FIGURE 10. X-Ray Diffractograms of the Sayers Soil 86

Masters Ill

Timberline

Terre Du Lac

28 24 22 14 2 8 in Degrees

FIGURE 11. X-Ray Diffractograms of the Masters #1, Timberline, and Terre Du Lac 87

Shift in

Kaolinite Treated with Potassium Acetate

Untreated Georgia Kaolinite

26 24 22 20 18 l6 14 12 10 8 6 2 6 in Degrees FIGURE 12. Shift in 001 Peak of Georgia Kaolinite When Treated With Potassium Acetate 88

only visible effect of this disordering is the strengthening of the 020 diffraction band and a sharp low-angle termin­ ation of that peak. These features are evident in the diffractograms of most of the soils tested in this study, including the Sayers soil, Figure 10. Brindley and Robinson (1947) illustrate broadening and lowering of intensity of peaks with a decrease in cry­ stallinity. Murray and Lyons (1956) present data to sub­ stantiate this effect of degree of crystallinity. They also demonstrate that particle size varies directly with the degree of crystallinity, larger particles being better crystallized. Conflicting data for the special case of a b-axis disordered kaolinite are presented in Robertson et al (1954). This kaolinite exhibits very good crystallinity even with very fine particle size, on the order of 0.2 microns. The low intensity and broad peaks of the diffractograms are the result of fine particle size, not of poor crystallinity. It would appear that in most soils, both decreasing clay mineral particle size and degree of crystallinity act to broaden and weaken the reflections of x-ray peaks. The residual soil x-rayed in this study shows a marked degree of asymmetry of the 001 peak sloped to the low-angle side. Keller (1968), in a study of flint clays which shows similar patterns, attributes this feature to 89

mixed layering of the kaolinite, probably with water but possibly with illite. Mixed layering of kaolinite and chlorite (Brindley and Hillery, 1953) also produces 0 similar asymmetry. The reflection around 10 A and 0 13-15 A noted in several of the sample diffractograms may be an indicator of mixed layering. Grim (1953) states that chlorite may be confused with kaolinite, especially if the clay is iron rich. The iron 0 can result in weak first order (14 A) and third order 0 (4.7 A) basal peaks. This leaves the second order 0 chlorite peak at 7.2 A to be confused with the 001 basal peak of kaolinite. A sample of Sayers soil was leached with HCl and showed no reduction of the intensity of the 0 7.2 A peak. This indicates that chlorite which is more soluble in HCl than kaolinite is probably not a major soil constituent. The possibility that iron oxide might "coat" the clay particles in some manner scattering the x-rays and contri- buting to the broadness and lack of intensity of the 001 peak was considered. The Sayers sample which was leached with HCl to remove the acid soluble oxides, including iron, when x-rayed showed no sharpening of the 001 peak. This indicates that the iron oxide is not the factor responsible for the shape of the peaks (Figure 10). In summary, the diffractograms show some character­ istics of halloysite, kaolinite, some mixed layering, and 90

minor amounts of chlorite, illite, and montmorillonite. First appearances suggest halloysite to be the major clay mineral in the soils. Further study shows that these fea­ tures which suggest halloysite may also be the product of a fine-grained and/or poorly crystallized kaolinite which possibly exhibits some b-axis disordering. Some mixed layering, probably with water, is indicated and adds to the confusion. Results of tests on the samples treated and x-rayed as outlined by Miller and Keller (1963) are a solid indicator that halloysite is not the major clay mineral. Several authors note the risks of clay mineral identification based on x-ray data alone since so many factors may combine to affect the final diffrac­ tograms. 2. Differential thermal analysis. Differential ther­ mal curves for the clay minerals vary with several factors related to both test procedures and the nature of the minerals tested. Careful and consistent sample preparation and testing should eliminate variations between curves due to these factors. The curves obtained in the current study all show to varying degrees the following characteristics: (1) a low temperature endothermic peak which generally occurs between 100-135°C, (2) a major endothermic peak which occurs in the area of 535-560°C, and (3) an exothermic peak generally developed between 910-930°C. A number of 91

these curves are presented as Figures 13 thru 17. A summary of the temperatures at which the endothermic and exothermic peaks occur are reported in Table X. Two samples of several soils (Timberline, Elsey, Georgia kaolinite) were tested and the peak temperatures of the two samples agree very well. According to Mackenzie (1957) the low temperature (100-135°C) peak is due to the removal of: (1) water adsorbed on the surface of clay particles, (2) interlayer water, or (3) water associated with amorphous silica or alumina gels. The major endothermic peak at 535-560°C is related to the expulsion of water from the crystal lattice. There is disagreement as to the cause of the exothermic peak, but researchers believe it is due to the formation

of a-Al 2o3 and/or mullite. This reaction begins at a temperature of about 900°C and extends well past 1000°C. Bradley and Grim (1951) present a more detailed investi­ gation of the nature of the high temperature reaction which goes beyond the scope of this investigation. Kaolinite pM (Mackenzie, 1957), which has disordered shifts in the stacking of the unit sheets, and halloysite produce differential thermal curves which appear much like those obtained in this study. The curves also have many of the features of a mixture of illite and kaolinite as studied by Grim (1947). Mixtures of kaolinite with other triple-layer minerals might also produce these features. 92

Two features of the curves tend to indicate that the soils are not predominantly halloysite. A sample of the reference halloysite, Indiana Halloysite, was dried in an oven for 12 hours at 70°C and then subjected to DTA. This sample shows a pronounced low temperature water loss even after oven drying. A sample of Floyd soil was oven dried for only 3 hours at 70°C, but all trace of the low temperature water loss disappeared. If a significant amount of halloysite was contained in the Floyd sample, some low temperature water loss should have been indicated. These comparisons are illustrated in Figure 16. A sample of Masters #2 which was oven dried for 12 hours at 70°C still showed a significant low temperature peak. Bramao (1952) studied the shape of the major endo­ thermic curve, particularly the degree of asymmetry devel­ oped when halloysite soils were analyzed. He expresses the asymmetry as the Slope Ratio (S.R. = tana/tanS), and found values for kaolinites varied from 1.78 to 2.39 with a mean of 1.55 while values for halloysite varied from 2.50 to 3.80 with a mean of 3.11. Correlation with test results of other published DTA curves for the two minerals confirmed his observations. The maximum slope ratio for soils tested in this study is in the order of 2.10, indi­ cating that halloysite is not a major constituent of the soil, but other factors as particle size and degree of crystallinity probably affect values of slope ratio. ;0... 0.

GEORGIA KAOLINITE

INDIANA HALLOYSITE

FIGURE 13. Differential Thermal Analysis of Georgia Kaolinite and Indiana Halloysite

1.0 VI MA S T E RS no. 2

MA S T E R S no. 1

BLACKWELL

EL S E Y

\.0 FIGURE 14. Differential Thermal Analysis of Ozark Soils ~ ·.., 0.-

TIMBERLINE no.l

.Ill

TI MB E R Ll NE no. 2

He ated for 12 Ho u r s at 7o· c

Leached in HCI

FIGURE 15. Differential Thermal Analysis of Ozark Soils •o.. Ill 1.0 (/1 0 Ill o- HORNSfY

HORNSfY-Ieached

fL 0 YD

FLOYD- dried for 3 h ou r s a t 7o' c

FIGURE 16. Differential Thermal Analysis of Ozark Soils \D 0\ "o "0.

SUNRISE SOUTH

TERRE DULAC

SAYERS

~ FIGURE 17. Differential Thermal Analysis of Ozark Soils ~ "'-l ...... ,~ 98

Some features of these curves do not correlate with curves of kaolinite, even the disordered kaolinite pH.

They show the major endothermic peak reaction around 535~ 560°C versus 550-580°C for kaolinite pM and about 580°C for halloysite. The depressed exothermic peak for the soils tested in this study occur at 910-930°C versus

960-980°C for kaolinites and 970~980°C for halloysite. Several factors might act singularly or together to lower these peak temperatures. Speil (1944) reports sev­ eral effects of a decrease in particle size induced by dry grinding including: (1) a gradual increase in the size of the low temperature endothermic peak, (2) a decrease in the size and reduction of the temperature of the major endothermic peak, and (3) a reduction in the size of the exothermic peak. The curves of the Ozark soils, when compared with the standard kaolinite and halloysite, exhibit these effects. Carthew (1955) substantiates Speilts findings but found the effect of particle size only related to the minus 2 micron fraction. In reporting their study of refractory clays, Grimshaw et al (1945) produce data which conflicts with Speil's and Carthewts finding in that they noted a lowering of the peak temperatures due to a decrease in crystallinity, but not due to a decrease in particle size alone, This conclusion has been substantiated in a study by Volostnykl 99

(1966) in which he points to the regularity, or lack of regularity, of the stacking of kaolinite plates as the dominant factor controlling the temperature of the major endothermic peak. Obviously, there remains a question as to the effect of the decreasing particle size on the temperature at which the major endothermic reaction occurs. The effect of iron oxides on the thermal activity of kaolinites was studied by Saunders and Giedroyc (1950). They report that iron oxides lower the temperature of the exothermic peak (but not to a temperature as low as that of the Ozark soils). As shown in Figure 15, leaching the acid soluble oxides, including iron, from the Timberline soil resulted in an increase in temperature of the exother­ mic peak from 915°C to 945°C. Caillere et al (1946), as reported in Mackenzie (1957), found the presence of alkali ions decreases the size of the exothermic peak. Kerr et al (1949) discuss several factors which affect the temperature at which the exothermic reactions occurs. They note that fine-grained, poorly crystallized, or interlayered kaolinites present exothermic peaks at much lower temperatures, peaks which are more rounded. Reactions for interlayered or disordered kaolinites have been observed which barely show an exothermic peak, presumably because of the small particle size and/or the broad temperature range over which the reaction occurs. In summary of the results of differential thermal 100

analysis, the similarity of curves for most of the soils studied indicates similar mineralogy. Analysis of the DTA curves and the factors which might have affected them indicate the probability that the major constituent of the soils is a kaolinite with perhaps a minimum amount of hal­ loysite with the possibility of triple-layer clay minerals. The size and generally low temperature at which the various endothermic and exothermic reactions occur is probably the result of poor crystallinity and small particle size of the clay minerals. The iron content, the presence of alumino-silicate gels, possibly allophane, and the nature and concentration of adsorbed cations each probably have an effect to an undetermined degree on the curves. 3. Thermogravimetric analysis. Ross and Kerr (1931, 1934), Nutting (1943), and Mielinz et al (1953) present a number of dehydration curves of kaolin group minerals. Dehydration curves of several of the Ozark soils are pre­ sented in Figure 18 along with the curves of the reference halloysite and kaolinite. Except for the curve obtained from the test on Georgia Kaolinite, the curves display some low temperature water loss, major water loss between 500-600°C (related to the major DTA endothermic peak) and very little water loss above 600°C. Grim (1953) notes, "Variations in the specific char­ acteristics of the minerals being studied, such as grain size, crystallinity, nature of adsorbed ions, etc., affect ] 0 1

16~-4------~------~~~~======!

10

,...... _ ..._,~

UJ UJ 0 ,...l 8 1-< (!) 4-J ~

6 • Sayers • Terre Du Lac 4 • :f\fasters 1 • Georgia Kaolinite .. Indiana Ha1loysite 2

0 _____. ______~------~~------~------~ 200 400 600 800 1000 0 Temperature C

FIGURE 18. Thermogravimetric Analysis 102

dehydration results." The Ozark soils do exhibit rather high initial water loss which even exceeds that of the reference halloysite. The factors of fine particle size, poor crystallinity, and/or presence of alumino-silicate gels probably produce this unusual water loss. It is not possible to isolate the exact clay mineralogy from these curves. 4. Electron microscopy. Clay particle size and shape parameters are best determined from transmission electron micrographs. The moderately well-developed hexagonal clay plate and the elongated plate shown in Figures 19 and 20 illustrate the highest degree of crystal development of any particles observed during this investigation. This hexagonal plate was the only one observed in a study of two soils (Hornsey and Terre DuLac); the elongated plates were fairly common. The several particles in the aggregates shown in Figures 21 and 22 are somewhat less well developed and illustrate the most common degree of crystallinity observed in the micrographs. These micrographs also illustrate the predominant particle size range determined from the transmission microscopy portion of the study. This size is in the order of 0.4 to 1.0 microns. The scanning electron micrographs of the powdered samples also indicate that the clay minerals are poorly crystallized. Most of the particles appear to exist in FIGURE 19. TEM Terre Du Lac ¥IGURE 20. TEM Terre DuLac Suspended Sample Suspended Sample (57000X) • ( SZOOOX)

I-' 0 VI FIGURE 22 . TEM Terre Du Lac FIGURE 21. TEM Terre Du Lac Suspended Samp le Suspended Sample (20000X) (28000X)

4 t-' 0 -+:> FIGURE 23. SEM Terre Du Lac FIGURE 24. SEM Hornsey Powdered Sample Powdered Sample (lOOOX) ( 3000X) f-' 0 (Jl FIGURE 25. SEM Hornsey FIGURE 26. SEM Hornsey Ped Sample Ped Sample (3000X) (lOOOOX) t-' 0 0\ 107

aggregates of several particles which are difficult to analyze. A number of the particles or aggregates display curved surfaces. These aggregates average 1 to 10 microns in length (Figure 23 and 24). The scanning micrographs of the broken chips of the air-dried peds show several features of interest. They substantiate the general low degree of crystallinity of clay particles shown in the earlier micrographs except for the Terre Du Lac sample where a fair degree of platy hexagonal or elongate crystallinity has been developed, Figure 28. The crumpled, swirl-like laminated structure exhibited by the micrographs of the Hornsey sample, Figures 25 and 26, and to a degree by the Terre Du Lac micrographs is almost identical to that developed in the kaolinitic ball clays described and photographed by Borst and Keller (1960). They attribute the probable cause of the deforma­ tion in structure to post-depositional stresses. Forces, possibly due to dessication, which were responsible for the pedular nature of the Ozark soils may also be responsible for the distorted structure. It had been hypothesized by this investigator that as the dessication process removed moisture from the soil-water system and the interparticle bonding became stronger, the clay particles would be drawn into a low density flocculated position. This does not appear to be the FIGURE 27 . SEM Hornsey FIGURE 28. SEM Terre Du Lac Ped Sample Ped Sample !-" (lOOOOX) (lOOOOX) 0 00 109

(]) !=: •r-1 1'"-i (]) ~ 1'"-i ,--.,. (]) p.. :>< ..0 IS 0 IS C1l 0 •r-1 rJ) 0 E-< 0 '"d 1'"-i (]) '--' ~ p.. ~ rJ)

Gl N

~p:: :::::> (.9 H ~ 110

case as a strong degree of parallel particle orientation is observed in the scanning micrographs. Michaels (1959), Trollope and Chan (1960) and other authors have postulated this parallel orientation of particles in packets. Microscopy studies by Sloan and Kell (1966) and others have shown this structure to exist. The peds exhibit a noticeable volume of voids which is probably responsible in part for the low density of the clay soils in their natural state. The microscopy study shows that halloysite, if present at all, is a minor soil constituent. No transmis­ sion micrographs and only one scanning micrograph, Figure 29, shows a tabular or tubular form which may be halloysite, but could also be chrysotile or a number of other clay minerals. 5. Chemical analysis. The results of chemical anal­ yses performed at the Missouri Geological Survey on a number of the soils investigated in the study are presented as Table XI. A summary of pH, cation-exchange capacity, and silica-alumina ratio is presented in Table XII. The content of Na, K Mg, and Ca in meq/100 gr. as determined at the University of Missouri - Columbia is presented in Table XIII. Grim (1953) presents typical values of the chemical composition of a number of clay minerals; Ross and Kerr 1 1 1

(1934) give values for allophane. A common value used to compare clays is the silica-alumina ratio, Si02/Al 2o3 . Average values of the silica-alumina ratio as reported by the above authors for kaolinite (theoretically 2:1) are 2.00:1 to 2.12:1; for halloysite, 2.00:1 to 2.10:1; for allophane, 1.20:1 to 1.80:1; for illite, 3.20:1 to 4.00:1; and for montmorillonite, 5.10:1 to 5.95:1. As Table XII indicates, the typical values for the Ozark clay residual soils are 2.45:1 to 2.55:1. Higher values are noted in the alluvial-colluvial soils where silica has been concentrated. Higher values are also noted in the soils which have a high chert (silica) content as Floyd and Blackwell. The silica-alumina ratio for the residuals of 2.45:1 to 2.55:1 may be caused by a mixing of some illite and montmorillonite in with kaolinite. Another cause may be that some octahedral layers may have iron replacing alumina which would result in a higher ratio. Still another cause could be the presence of silica-rich gels in

the clay. The Ozark soils exhibit a relatively high (6-12%) iron content. A number of possible effects of this occurrence have been discussed in earlier sections. Table XI shows that for three soils tested (Sayers, Terre Du Lac, Hornsey) approximately 90 percent of the This indicates that Al 2o3 is soluble in 12.1 normal HCl. 112

a majority of the alumina in the soil is not crystallized or is poorly crystallized. When results of other tests are considered along with this evidence, the latter, poorly-developed crystalline state is indicated. Keller (1970) notes that soil acids can cause the breakdown of even well-crystallized kaolinites. This effect is thought to be much more severe for poorly-crystallized or disordered kaolinites. 6. Summary. The x-ray diffraction, differential thermal analysis and the thermogravimetric analysis com­ bined with a study of the electron micrographs indicates that the major clay minerals of all of the residual Ozark soils investigated in this study are poorly crystallized kaolin minerals. If placed in Bates' (1959) morphological series, these minerals would be classified as elongate kaolinite crystals with possibly some particles as a transitional phase between the elongate crystals and halloysite. Two or more phases of clay morphology may be repre­ sented in the Ozark soils. One may be the better crystallized, elongate platy crystals. Another phase represented by less well-developed crystallinity may be a precurser (Keller et al, 1970) of kaolinite or halloysite. This phase or other phases may be the result of post­ formational crystal destruction enhanced by soil acids. Bramao's discussion of the probable mode of formation 113

of these soil clay kaolins aptly describes the Ozark setting in which the clays were formed. It is to be expected that the result of changing soil conditions and interrupted formational cycles would be poorly developed and varying clay morphology. Minor amounts of chlorite, illite, and montmorillonite are indicated. The presence of an undetermined amount of allophane or an alumino-silicate gel of some nature is also indicated.

B. SOIL CLASSIFICATION AND CATION-EXCHANGE CAPACITY The results of Atterberg Limits tests are presented in Table VII with results plotted on the plasticity chart, Figure 29. Results of gradation tests are presented as Figures 3 thru 8. Table IV summarizes the results of the gradation tests. Before sieve analysis, all materials retained on the No. 4 sieve was discarded. This material was predominantly chert. Earlier sections of this discussion noted that the proportion of chert throughout the soil varies widely and that the percentage in any one sample is not a diagnostic value. The fines is defined as that material passing the No. 200 sieve. Clay is defined as that material (colloids) finer than 2 microns (0.002 mm). Except for the sandy residual soils, the soils invest­ igated have 85 percent or more fines and 60 percent or more clay. Several of the clay residuals have greater than 90 114

percent clay. ThB sandy residual soils have 60 to 80 percent fines but only 20 to 40 percent clay. These soils are composed primarily of fine sand and silt size particles. All but three of the soils plot as CH or CL on the plasticity chart. The Hornsey soil plots just below the A-line as a MH soil. The Rice and Fabick sandy residuals are CL-ML soils. Definite relationships are noted between the Atterberg Limits and the percent clay. Davidson and Sheeler (1952) note a general relationship between the clay fraction and the liquid limit and plastic limit. The clay content of the clayey residual soils was correlated with the liquid limit and the plastic limit, Figures 30 and 31. A linear curve fits both sets of data fairly well, but a concave upward second or third degree curve fits both sets of data more closely. It is suggested that at higher clay contents, notably above 90 percent, that the soil is more "sensitive" to other factors than to increases in percent clay. For high clay content soils, as water is added up to the liquid limit the electro-chemical nature of the soil-water system has a noticeable effect on

the Atterberg Limit values. The cation-exchange capacity is a measure of this electro-chemical potential of the clay soils. Values of the cation-exchange capacity determined for the Ozark soils are presented in Table XII along with

soil pH and silica-alumino ratio. 100 Linear y = -3.81 + 1.05 X rr = 61.1 r = 0.95 2 Best Fit y = 49.47 - 0.61 X+ 0.012 x tr = 50.2

I 0 """'IN! '-./ 80

.u ·r-1s •r-1 0 ....:l I ~~ 0 "0 •r-1 ;:1 0" 60 •r-1 ....:l

40

20_~------~------~------L------~------~------L------~------jL 55 -~ 6u 65 70 75 80 85 90 95 Clay Content (%) f---0 f---0 FIGURE 30. Clay Content Versus Liquid Limit (Clay Residual Soils) Ul Linear 0 y = 2.52 + 0.36 x cr = 13.4 r = 0.87 2 Best Fit y = 32.70- 0.56 X+ 0.007 X cr = 9.6 o I 40 I I I I "";;-e 35 "-'

+J 0 -Ma -M 0? ~ 0 '/ ·rl +J 30 / {/) 0 .....til /0 fl.! //

25 // / 0 0

0// , /, 0 0--:------20 ' I I I I I I I 30 40 50 60 70 80 90 100 Clay Content (%) f-' f-' FIGURE 31. Clay Content Versus Plastic Limit (Clay Residual Soils) 0\ 117

The Sayers and the Hornsey soils both contain 93 percent clay but the liquid limit of the Sayers is 105 compared to 90 for the Hornsey. The cation-exchange capacity of the Sayers is 44.9 whereas that of the Hornsey is 42.1. Masters #2 has 95 percent clay and a cation­ exchange capacity of only 35.8. The result is that Masters #2 has a lower liquid limit and plastic limit than does Sayers. This does not mean that the cation-exchange capacity is by itself the factor controlling the limit indices. The cation-exchange capacity was correlated with the liquid and plastic limit, Figure 32 and 33. The degree of correlation is similar to that of the correlation of the clay content versus the limit indices. This similarity suggests that a combination of influence of clay content and cation-exchange capacity on the limit indices exist. Data presented above indicates that at higher clay contents, the cation-exchange capacity of the soil has the dominating effect on Atterberg Limits. Cation-exchange capacity is a property of the clay fraction of the soil. The degree of correlation between these two factors, cation-exchange capacity and clay content, was determined and presented in Figure 34. A fairly linear relationship appears to exist between these factors up to a clay content of about 90 percent. Above that point the curve flattens and it is postulated that again the cation- exchange capacity becomes more important. The soils which 40 0

20L_--~L---~----~----~----~----~----~----~ 5 10 15 20 25 30 35 40 45 Cation-Exchange Capacity (meq/100 gr.) I-" I-" 00 FIGURE 32. Cation-Exchange Capacity Versus Liquid Limit (Clay Residual Soils) 0

40 Linear and y = 12.04 + 0.66 X cr = 9.3 r = 0.91 Best Fit

...... _ 35 ...... ,N 0 / 0 .j.J r •rl a •rl ...:1 t) •rl .j.J 30 r 0 (I) t\1 ...-t ~

25 0 0

0 0

20 5 20 25 30 35 40 45 Cation-Exchange Capacity (meq/100 gr.) f-1 f-1 \.0 FIGURE 33. Cation-Exchange Capacity Versus Plastic Limit (Clay Residual Soils) 100 0 0 _....~ A"',..... / 80 0 ,o/ // /./'0 0

"'"" // ~o/ IN!.._, 60 I 0 // .j.J ~ Q) / . .j.J ~ ~ ~ 0 u :>. ~ r-i 40 u 0

20 Linear y = 15.81 + 2.02 X a = 155.1 r = 0.88 2 Best Fit y = -17.74 + 5.21 X- 0.06 X a = 89.2

oL---~----~L------L------L------J------~------~----~-- 35 40 45 10 15 20 25 30 1-' N Cation-Exchange Capacity (meq/100 gr.) 0

FIGURE 34. Cation-Exchange Capacity Versus Clay Content (All Soils) 121 appear to control the deviation from linearity of the curve (Terre Du Lac, Sayers, Masters 2) are the same soils which have high cation-exchange capacities. The concentrations of the more common exchangeable cations which compose the cation-exchange capacity are presented in Table XIII. These observations are consistent with discussions by Grim and others covered in the literature review with regard to the clay-water system and the effect of cation concen- tration on the system. Increased cation concentration results in an increase in the attraction between the particles. More water is required to dilute the concentra- tion to reduce particle attraction, and to reduce flocculation, which allows the liquid limit shear failure to develop. With regard to the plastic limit, the higher cation concentration produces a flocculated microstructure which resists deformation at a fairly high moisture content. Conversely, soils with lower cation concentrations have a more dispersed structure which allows manipulation and water expulsion until a lower moisture content is reached. The effect on Atterberg Limits of dispersing the clay soil with sodium is illustrated in Table VI. It is thought that the sodium acts to lessen the attraction between clay particles and allows more water between particles. The result is a lowering of the limit values as the degree of flocculation and particle attraction decreases. 122

Table IV illustrates the difference in the liquid limit values of soils which have never been air dried and those which have. Though the cation-exchange capacity of wet samples was not determined, it is thought that the exchange capacity for soils which have never been dried would be higher than that of soils which have. This reduction in capacity would occur if upon drying the concentration of cations in the clay soil increases as water evaporated and micropackets of clay particles are formed. These packets would tend to trap cations, thereby reducing the cation-exchange potential. Upon rewetting with distilled water for the limit tests, the packets are not completely destroyed and all cations would not be available in the clay-water system, some still being bound up between plates in the packets. This decrease in cation-exchange capacity would result in lower particulate attraction and

lower Atterberg Limit values. Another aspect is that the micropackets which are formed upon air drying and which do not disperse when mixed with distilled water effectively reduce the percent clay-size fraction of the soil. Though the packets them­ selves probably have a water film as would the individual clay particles, the attraction between packets is less; therefore, less water is required to reach the liquid limit. The liquid limit of samples which have never been dried and have the higher cationic concentration and greater particle attraction would require more water to disperse the structure 123

to a point of failure in the liquid limit device.

C. COMPACTION CHARACTERISTICS Three factors; soil gradation and Atterberg Limits, cation-exchange capacity, and clay structure, were studied with relation to their effect on laboratory compaction of the Ozark soils. No gross difference was noted in the clay mineralogy of the soils investigated in this study so this was eliminated as a prime cause of compaction differences. The results of soil compaction tests are presented in TableVIII. Figures 35 thru 45 illustrate the relation­ ships between the factors listed above and the optimum moisture content and maximum dry density of the soils tested. Table XV illustrates the correlation coefficient and sigma value for each relationship tested. 1. Atterberg Limits. As discussed previously, Harris (1969) showed the best correlation tested was between the liquid limit values and the maximum dry density and optimum moisture content. The correlation coefficient for this relationship for residual soils was on the order of 0.90. This same range of values (0.88 for maximum dry density and 0.90 for optimum moisture content) was noted for the Ozark clayey residuals. The plastic limit was found to exhibit even a slightly better degree of correlation than the liquid limit. A correlation coefficient in the order of 0.90 signifies a high degree of relationship (Harris, 90

...... 4-4 tJ .....,~ 85 0 ....,:>...... til 0 ~ Q) 0 ~ :>. 0 ~ 80 ~ ~-- ~ ... ,...... ~ ~ ...... ~ Linear y = 125.09 - 0.51 X 75 a= 5.9 r = -0.96 2 Best Fit y = 138.85 - 0.93x + 0.003 x a = 5.4 0

70 I I I I I I I I I 55 60 65 70 75 80 85 90 95 Clay Content (%) ~ N .s;:. FIGURE 35. Clay Content Versus Maximum Dry Density (Clay Residual Soils) 0

40 Linear and y = 2.73 + 0.37 X (j = 9.7 r = 0.90 Best Fit

,-... ~ ~ '>.,/ 0 +.1 35 !::: Q) +.1 !::: 0 u Q) I-I ::l +.1 fJl 30 •rl jl

·rl~ +.1 p.. 0 25

20 ~~ 60 65 70 75 80 85 90 95 Clay Content (%) I-' N tn FIGURE 36. Clay Content Versus Optimum Moisture Content (Clay Residual Soils) 120

Linear y = 108.40 - 0.80 X a = 23.0 r = -0.87 110 2 3 Best Fit y = 158.16 - 6.66 X+ 0.21 X - 0.002 X 0 = 18.8 0 /""'> 4-1 () 0. '-../ \ \ :>. -IJ ·ri 100 11.) Q '\ (I) " A :>. 0 ~ 90 ·ri~ >< :2

80

70 10 15 20 25 30 35 40 45 f-' Cation-Exchange Capacity (meq/100 gr.) N 0\ FIGURE 37. Cation-Exchange Capacity Versus Haximum Dry Density (Clay Residual Soils) 40 I 0 /. 0 I 35

""'~ '-'

.j..J Q I 0 Q) .j..J 30 Q 0 u Q) ~ ::l .j..J 0 rn ..... 0 :S 25 ~ 0 ~ / 0 ...... j..J 0. 0

20- Linear and y = 13.83 + 0.62 X cr = 14.3 r == 0.86 Best Fit 0

15 10 15 20 25 30 35 40 45 f-' Cation-Exchange Capacity (meq/100 gr.) N "-J FIGURE 38. Cation-Exchange Capacity Versus Optimum Moisture Content (Clay Residual Soils) 110 t- Linear y = 117.80- 0.41 X cr = 19.2 r = -0.88 '\.o 2 Best Fit y = 154.50 - 1.44 X+ 0.007 X cr = 13.8

"'""11-1 " () llo '-' " 100

~ " ~ •rl ~ Ill " s:: Q) " Q

~ 90 Q'"' or!~ :< ~ 0 80 '

70 30~------~------~------~------L------~------~------~~------40 so 60 70 80 90 100 Liquid Limit (%) ..... N 00 FIGURE 39. Liquid Limit Versus Maximum Dry Density (Clay Residual Soils) 50 Linear y = 8.93 + 0.30 x a= 10.6 r = 0.90 Best Fit y = -17.90 + 1.06 x- 0.005 x2 a= 7.6

N"""" "-../

+J 40 !:I Ill +J !:I 0 u Ill ~ +J .~ 30 ~

I / j+J ~ . o 0.. 0 20

10 30 40 50 60 70 80 90 100 Liquid Limit (%) 1-' N \D FIGURE 40. Liquid Limit Versus Optimum Moisture Content (Clay Residual Soils) 110 Linear and y = 109.1 - 0.80 x a = 8.1 r = -0.89 Best Fit

,...... ~ u 0.. '-./ 100 :>. .j..l •r-i (I) I:: Q) I --...... _ 0 t:l :>. 1-1 ~0 Q 90 t- ...... 0 ~ 13 •r-i X :ill I 0 80

7010 ~------~------~------L------~------~------~~------~~----- 15 20 25 30 35 40 45 Plastic Limit (%) ...... tN 0 FIGURE 41. Plastic Limit Versus Maximum Dry Density (Clay Residual Soils) 50 Linear y = 3.79 + 0.90 X a= 7.9 r = 0.92 Best Fit y = -29.72 + 3.12 x -0.04 x2 a = 5.6

;-.. N '-" .j.Jg .j.J 40 1=1 0 c.>

<1.1 ~ .j.J Ul •r-1 30 ~ § 0 - ~ ~g .j.J p.. 0 ' 20

1010 ~------~------~------L------~------~------~~------~------15 20 25 30 35 40 45

Plastic Limit (%) f-' VI f-' FIGURE 42. Plastic Limit Versus Optimum Moisture Content (Clay Residual Soils) 120 Linear y = 128.30 - 0.55 X 0 = 10.9 r = -0.97 2 3 Best Fit y = 149.52- 1.79 X+ 0.019 X - 0.0004 X 0 = 7.6

110 "·', "'"'Il-l () ', ...... ,Q. '-..., t' 100 'rl 0 ', !ll ' 1::: Q) ~ ',o >. ' ""-, 0 ~'"' ...... 90 m ', -o 'rl -...... : X )! 0 0 ~...... ~._, 80 . ::--.o, 0~ 0

20 30 40 so 60 70 80 90 Clay Content (%) ...... Vl N FIGURE 43. Clay Content Versus Maximum Dry Density (All Soils) Linear and 50 y = 6.30 + 0.33 X a= 8.8 r = 0.95 Best Fit

,....._ N '-" I 0 +J l=l 40 Q) +J l=l u0 I ~ 0 Q) 1-4 ;:I +J til •rl 30 ~

•rl~ +J p. I 0 0 20 I 0

10 20 30 40 50 60 70 80 90 Clay Content (%) ...... (.N (.N FIGURE 44. Clay Content Versus Optimum Moisture Content (All Soils) 120 Linear y = 121.19 - 1.14 X (J = 15.1 r = -0.90 2 3 Best Fit y = 296.70- 19.01 X+ 0.58 X - 0.006 X (J = 7.3

110L \ 0

""'4-l \ t) p. \ '-" I ~ :>-. -1-1 1001 "T"4 Cll !:l (!) ~ Q :>-. ~ Q 90 m •.-1 X til ~ 0 80 '

70 10 15 20 25 30 ' 35 40 45 1-' Optimum Miosture Content (%) V-1 +:>. FIGURE 45 • Optimum Moisture Content Versus Maximum Dry Density (Clay Residual Soils) 135

1969). 2. Gradation and cation-exchange capacity. The percent fines was not found to have a significant relation­ ship with either the maximum dry density or optimum moisture content. The correlation between the percent clay of all the soils and the compaction characteristics was very high, -0.97 for the maximum dry density and 0.95 for the optimum moisture content. The cation-exchange capacity also has a high degree of correlation with the optimum moisture content (0.86) and the maximum dry density (-0.87). The high degree of correlation between the Atterberg Limits and the compaction characteristics indicate that the factors which influence the values of the limit indices also influence the compaction characteristics. The clay content and the cation-exchange capacity were determined to strongly affect the limit values; similarly, they have been shown to be strongly related to the optimum moisture content and the maximum dry density. Lambe (1958) has postulated that the compacted nature of clayey soils is a result of clay particle structure. Clay content, cation concentration, and other factors not studied in this investigation have a strong influence on the compaction process which is manifested in their affect on the clay structure. 3. Soil clay structure. The Atterberg Limits and the cation-exchange capacity are considered indirect measures 136

of clay structure. No direct quantitative measures of differences in clay structure in compacted samples were determined, but a qualitative low-power and scanning micro­ scopy investigation of the clay structure was conducted. Particular attention was made in comparing the apparent validity of the particle-to-particle concept of the structure of compacted samples and changes in this structure with different compacting moisture contents as postulated by Lambe (1958) to the packet-to-packet compacted structure as postulated by Sloan and Kell (1966) and Barden and

Sides (1970). Low-power micrographs of the Terre Du Lac soil are presented in Figures 46 and 47 for two samples; one com­ pacted 3% dry of optimum, the other compacted 5% wet of optimum. Samples compacted at all moisture contents contain a large number of particles which exhibit residual pedular structure which apparently has not been destroyed or altered

significantly by the compaction process. Some differences are observed between the samples compacted dry and wet of optimum. The dry sample has a much "rougher" texture, the wet sample an apparent uniform texture. A large volume of voids, or macrospaces, are also obvious in the dry sample and which do not appear in the wet sample. The size of the packets, or macropeds, in the dry sample is highly variable whereas less variation

is exhibited in the wet sample. a.) Compacted 3% Dry of Optimum b.) Compacted 5% Wet of Optimum

FIGURE 46. LPM Terre Du Lac (Black and White)

t-' lN -....} 'A·.. .f '•i . )' ~·ij,:r· . ~ ~k.~ , f •.•.' .., ~~ • io " .. • • ,":f ~ •fv '9. " 1~ . ! ' ·~ • '·' .. ~ ~ .; . ,~ .• . K . - . . ~· ~, .a;j ·\ : ·~. ·,,. ·.• IE1 aJ liiC · .. t. -~~ ,. P!l ~ r ~' 'I ~~. ·~·t·:li ".· ·~ I . •I'. ,. ... ' • • ji • _. ~ I r" , ·-~ ~ ~' ..· ~ . ~ . ;;r 'f. . ··., ~.4·--··...... :·,., ~ [J!l :.,)-, • • ~ I' • ~ ~- f~ . ~-...... ).' \ .. '-4···. . R I -.~ ......

a.) Compacted 3% Dry of Optimum b,) Compacted 5% Wet of Optimum

FIGURE ~7. LPM Terre DuLac (Color)

f-> (A co 139

It is suggested that the sample compacted wet of optimum has been more altered by wetting and compaction than that compacted dry of optimum. Part of this alteration is probably manifested in the breakdown upon wetting of the peds due to differential swelling pressure and differential air pressure as suggested by Baver (1956). A study of the microstructure was initiated to further investigate the macropeds and to investigate the particle­ to-particle structure of compacted samples with variations in the compaction moisture content. Figures 48, 49, and 50 compare the structure of Standard Proctor samples compacted dry and wet of optimum at magni­ fications of lOOX, 300X, and 3000X. All micrographs are of sample surfaces oriented parallel to the direction of applied compactive force. At a magnification of lOOX, the outline of the macropeds are still apparent with similar patterns of texture, macrospaces, and ped sizes as was evident in the low-power micrographs. The scanning micro­ graphs (lOOX) enhance the significant textural differences between the samples compacted wet and dry of optimum; the dry sample with a very random pattern of macropeds, the wet sample with a uniform pattern. These patterns are still evident in the 300X micrographs. Quite a different pattern is observed in Figure 50 for those micrographs taken at 3000X. The microstructure, or "intra-pedular" structure, is very similar for both wet

'-- · · ---..a._ -

t;,.,

T a.) Compacted 3% Dry of Optimum b.) Compacted 5% Wet of Optimum

FIGURE 49. SEM Terre Du Lac (300X)

I-' ~ I-' ""

a.) Compacted 3% Dry of Optimum b.) Compacted 5% Wet of Optimum

FIGURE SO. SEM Terre Du Lac (3000X) ,__. -+>o N 143

and dry samples, apparently independent of compaction moisture content.

The theory of the compactive process which emerges from this study is considerably different from that postulated by Lambe (1958). There is little indication in the micro­ graphs of single particle activity being prominant in the compaction process. This is particularly true for samples compacted dry of optimum. On the wet side of optimum where samples appeared uniform under relatively low magnification, clay particles seem to form a cement between macropeds. The following theory, similar to that postulated by Barden and Sides (1970), of laboratory compaction density is primarily a function of the degree of fit between macropeds of clay. Dry peds at low moisture content are resistant to deformation and a relatively loose compacted structure is obtained. With the addition of moisture, two factors are thought to influence the compacted density. First, upon adsorption of moisture, the macropeds tend to break down forming a better gradation of particles which can result in a more dense arrangement without major ped deformation. Also, as Barden and Sides postulate, with additional moisture the macropeds become "softer" and can be forced by outside forces into a more compact arrangement.

At the optimum moisture content, the effects of ped breakdown and deformation are maximized. Macrospaces are minimized. 144

Wet of optimum the influence of individual particles and the smaller peds are predominant. With increased moisture and further deterioration, more particle-to­ particle reaction may occur. As water layers are adsorbed by particles and small peds, they are pushed apart and density decreases. A decrease in cation concentration may play an important role as the water film expands. Even wet of optimum, a large portion of the peds remain relatively unaffected by the increased moisture and their internal structure is essentially unaltered. The micrographs of the natural air-dried soil has indicated a natural essentially parallel particle orienta­ tion. When this structure is affected by moisture, no major reorientation of particles may be required, as from flocculated to parallel structure; just a separation of parallel plates. Assuming the natural ped structure is the result of dessication, the forces responsible for the particle orientation were very large. Once the ped structure has been destroyed by moisture, the effort expended in a compaction test in an attempt to force the particles back into a dense state is considerably less than were the dessication forces. The water film surrounding the particles and peds resists the compactive effort. The natural low density of the peds (Scrivner, 1960) which is not destroyed by the compaction process is one 145

factor responsible for the low compacted densities of the Ozark clayey residual soils. As would be anticipated, the soils tested with higher non-clay contents (primarily quartz, S.G. - 2.66) have higher compacted densities. Soils with high cation-exchange capacities, particularly those high in the sodium ion, appear to adsorb moisture more readily with an accompanying breakdown in ped structure. The greater the degree of particle reorientation, the lower the compacted dry density and the higher the optimum moisture content. Figure 51 demonstrates the results of the Standard Proctor tests on the soils dispersed with sodium metahexa- phosphate. The dispersed soils demonstrate a higher maximum dry density and lower optimum moisture which agrees with Lambe's (1962) discussion of the effects of aggregants and dispersants. It is postulated that the sodium treatment acts to breakdown the peds more than the normal addition of water reducing the macrospaces and allowing a more dense particle arrangement at a lower optimum moisture content. As more water is added, particularly past the optimum, the particles adsorb water layers reducing the particle-to-particle attraction and the dry density. 90

Floyd - natural 0~ Floyd - dispersed Terre Du Lac - natural 85 / ~ Terre Du Lac - dispersed Sayers - natural ...... 4-1 0 / Sayers - dispersed s:l. '-J ""

~ .j,J ...... !I) !:1 Q) 80 Q

~ ~ Q

...... ~ i 75

70

20 25 30 35 40 45 50

Moisture Content (%) I-' .f::o FIGURE 51. The Effect of Dispersion of the Standard Proctor Compaction 0\ of Clayey Residual Soils 147

VII. CONCLUSIONS

A. GENERAL

This study of residual soils used in earth dam construction in the Ozark Province of Missouri centered investigation in four general areas. First was a determina- tion of the clay morphology of the soils and the extent of any changes in this morphology from one site to another. Second was a study of the soil gradation, Atterberg Limits and cation-exchange capacity. This was followed by a study of the correlation between these soil properties and the compaction maximum dry density and optimum moisture content. Finally, the effect of the pedular nature of the clay soils on the Standard Proctor compaction characteristics was investigated.

B. CLAY MORPHOLOGY Regardless of differences in the parent material or the geographical location of the dam sites within the portion of the Ozark Province investigated, the morphology of the clayey residual soils is similar. The relative location of the samples within the vertical soil profile appears to determine to a great extent the clay content in percent. It is shown that the climatic conditions which existed in the Ozarks during the Tertiary Period were probably ·zation process which conducive to the development o f a laterl has resulted in the formation of halloysite-rich soils in other areas under similar conditions. The numerous tests 148

which have been performed to determine the mineralogy of

the clays have shown that the residuals do not contain

halloysite in any significant amounts. It is suggested that the continuous, extensive leaching process associated with halloysite formation was never completely developed within

the Ozark area. Instead of halloysite, clays with a morphology more closely resembling kaolinite were formed.

The major clay morphology is expressed in poorly

developed elongate platy kaolinite crystallinity. Other kaolin morphologies even less well developed, including possibly some alumino-silicate gels, are indicated by some of the tests. A minor amount of illite, chlorite, and possibly some montmorillonite is also present. The soils are characterized by the presence of an aggregated structure which separates the soils into units called peds. This ped structure is probably the result of soil collapse as the bedrock is chemically leached and of dessication of the soil mass. These peds are thought to affect the results of all tests performed on the soils to determine their engineering properties.

C. SOIL CLASSIFICATION AND CATION-EXCHANGE CAPACITY When the Atterberg Limits test results are plotted on the plasticity chart, the clayey residuals and the alluvial­ colluvial soils are classified as CL and CH. The liquid limits of the clayey res1·d ua 1 so1·1 s were as high as 105 with a corresponding plasticity index of 69. The stone-free 149

residuals contain up to 95% clay.

The cation~exchange capacity of the residual soils was as high as 45 meq/100 gr. for those soils with greater than 90% clay. This is very high for kaolinitic soils. The fine clay particle size and poor crystallinity have been shown

to act to increase the cation~exchange capacity of kaolinitic soils. These high values may also reflect the presence of some illite and/or montmorillonite. The value of the liquid limit of samples never allowed

to air~dry in considerably (30~40%) higher than that of

samples which were tested after air~drying. It is thought that the process of air-drying enforces and strengthens the natural ped structure to a point that it fails to break down completely when wetted for the liquid limit test. The result is that only moderate particle-to-particle (ped-to­ ped) attraction develops and less water is required to facilitate shear failure in the liquid limit device. The liquid limit of the sodium dispersed samples was found to be less than the natural air-dried liquid limit for several of the soils. This agrees with results presented by White (1958) in his discussion of the effects of the exchangeable cation on the liquid and plastic limit of clay minerals. An explanation fitted to the pedular nature of the soils is that the samples treated with a sodium dispersant reacts in one of two ways or in a combination of the two: (1) the dispersant completely breaks down the ped structure allowing the sample to behave as described by White, and/or 150

(2) the ped structure is partially retained but the sodium attracts to the peds a film of water which acts to separate

the clay peds and reduce ped~to~ped attraction. In either case, the liquid limit value is lowered,

D. CORRELATIONS WITH COMPACTION RESULTS Several factors were studied to determine their correlation with the Standard Proctor laboratory compaction parameters of the clayey residual Ozark soils: (1) soil gradation, (2) Atterberg Limits, and (3) cation-exchange capacity. The percent clay had the best correlation tested with both optimum moisture content (r = 0.95) and the maximum dry density (r = -0.97). The plastic limit also correlated well with both factors, particularly the optimum moisture content; the liquid limit exhibited only fair correlation. The cation-exchange capacity also had only fair correlation with the compaction characteristics. This might be anticipated since the cation-exchange effect is more pronounced in a more fluid environment than that of the

normal compaction test. There appeared to be no significant relationship between the percent fines and the optimum moisture content and the maximum dry density. The relationship between the optimum moisture content and the maximum dry density is good

(r = 0.90).

E. THEORY OF RESIDUAL SOIL COMPACTION The Standard Proctor compaction tests yielded low values 151

of the maximum dry density with corresponding high values of the optimum moisture content for the clayey residual soils. The electron microscopy investigation has shown that the pedular structure of the clayey residual soils is not destroyed during laboratory compaction. This is thought to be the key factor controlling the low density of the compacted samples. The low-power micrographs illustrate that dry of optimum most of the clay remains bound in macropeds with a large volume of macrospaces between them. As more water 1s added and optimum is approached, the moisture breaks down some of the larger peds to form smaller peds and other peds are probably completely destroyed. A better ped-particle gradation and perhaps a softening of some peds with increased moisture allows the soil to be forced into a more dense arrangement with a minimum of macrospaces. Wet of optimum, even more peds are reduced in size and/or destroyed with clay particles and small peds develop water layers which force them apart, reducing the compacted density. These particles appear to form a cementing bond between the remaining peds, which still comprise the majority of the compacted sample. High magnification micrographs show that the micro­ structure of the clay peds is not destroyed during the compaction process. Indeed, it does not appear to be changed significantly from that of the natural soil peds. 152

F. APPLICATION OF TEST RESULTS

1. Correlations. Several correlations have been presented which may be useful in estimating the optimum moisture content and maximum dry density from more commonly available data. The clay content and plastic limit were noted to be closely correlated to the optimum moisture content and maximum dry density. Curves relating these factors have been plotted and presented as Figures 52 and

53. Similar curves are presented, Figures 54 and 55, relating the plasticity chart values, liquid limit and plasticity index, to the optimum moisture content and maximum dry density. At best these curves yield approximate values, show trends. Better estimates can be obtained using the equations presented along with the illustrations which show the correlations. 2 . ::.D..:::a=m~d.:;_:_e..:::s_i~g"n=---c.:...:..o_n_s_t_r_u_c_t_i_o_n_. The ma j or pro b 1 em observed in the earth dams investigated was seepage through and around the dams. The seepage is often noted high up on the downstream slope of dams, even on those dams with moderate slopes. The seepages are often associated with surface slips on the steeper dam slopes. The additional driving force exerted by the saturated areas overcomes the resisting forces.

It is probable that the resisting forces are reduced signifi­

cantly as the nature of the ped structure of the embankment

clays is altered upon saturation. A large portion of the peds

are believed to swell and deteriorate, destroying much of the

intergranular frictional resistance which is thought to provide 75 80 •77\

401- a~' •74 \ ~~~ \ ~ "-s~ \ \ v\, \ \ 75 ~ f5 / ~4,. e83 \. 79 ...... ~s'\- ~ 3l ~~ '--" \ • 81 ~ \ +J "'-~ • 85 'r'ls \ •n ~\,~ \ ~ 30 \0 \ t) • 88 .\ 80 ·r-1 +J 82 \ 1:1) til \ ...-! ll-o \ \ \95 \ 25L \ • 84 85 \ \. 89 \ • 94 e92 \ e108 \ \90 201 I I \ 30 40 50 60 70 80 90 100 Clay Content (%) ...... U1 FIGURE 52. Graphical Relationship of the Maximum Dry Density to the w Plastic Limit and the Clay Content of the Clayey Residual Soils 40 I 38 • I 35 \ ~~~ 42 • 40 ~~~ \ \ .s~~ '-.: \ 6\o 30 "40 / I \ 35 ~~~ 34 ·" • 37 ,-... va~<'\- ~ \ " • 36 '-" ~~ •33 s<'\;, .1-J ...... 35 •r-1 ~~-<..; 25 " 13 •r-1 \ ....:l 30 ~Ay# \ e31 (.) \)~ ~~ '1"'1 .1-J Cll \ t1l .-1 p.. \ ' ,30 25 ' • 26 •25

• 27 •24 ' ...... " 25 • 18

40 50 60 70 80 90 100 Clay Content (%)

FIGURE 53. Graphical Relationship of the Optimum Moisture Content to the 1-' V1 Plastic Limit and the Clay Content of the Clayey Residual Soils +;>. 80 •"'-o-v'~ --#~"" \....S~"V

~(j\. • 81 "" fO 60 -v-¢s'\-~ 84, "'"' 1 79• ~ '-" "V'Q;-4,.

~ QJ ~~~ 90 70 s:: "'H :>. 40 100 / .az .j.J -M 9Z. ba • • 0 -M I 1 /as .j.J 94. {/J !11 .-I l p.. 120 \-100 20 r 108• \ \ •113 \ 113. \ 100 0 I \ V I I I I 0 20 40 60 80 100 Liquid Limit {%) ._. FIGURE 54. Graphical Relationship of the Maximum Dry Density to the tn tn Plasticity Index and the Liquid Limit of all Soils C)))l ~<'\,~

80 ~ ls<'\-" "\o 40 / ~~<'\, d:;)~~ 36. I ~ s<'\-~ 60 1- ~()'\. 30 • 25 '\>~~ ,..-.,. ·:P~ . N # I '-' I ()~<'\,~ I >< I G) "' 20 .e 40 24 :>. .j.J I • ..... • t) 27 /1 40 ..... 10 .j.J I en ~ ...... I p., I 20 I ,.18 I I 13.

0 I I 0 20 40 60 80 100 Liquid Limit (%) 1-' V1 FIGURE 55. Graphical Relationship of the Optimum Moisture Content to 0\ the Plasticity Index and the Liquid Limit of all Soils 157

the mechanism of slope stability (Terzaghi, (1958).

Where these clayey soils are utilized in dam construc­ tion, compaction on the wet side of optimum would signifi­ cantly reduce the susceptibility of the embankment to

seepage. The macrospaces which exist in the samples compacted dry of optimum offer ideal channels for water flow. Samples compacted wet of optimum contain less macrospaces and should be more resistant to water flow. Post-construc­ tion loss of shear strength due to ped deterioration upon saturation would also be reduced. The possibility of piping would also be reduced.

It is also recommended that the dam embankments be designed based on the results of Modified Proctor compaction results rather than Standard Proctor. This would result in higher densities with higher shear strength and less macrospaces with a corresponding decrease in seepage.

G. RECOMMENDATIONS FOR FURTHER STUDY The discussion presented above relating dam design and construction to the nature of compacted soils assumes that the soils compacted in dam embankments behave as those compacted in the laboratory. It is believed that the theory of compaction as outlined previously does apply to embankment soils, but that the degree of soil structure evident in the laboratory soils is not developed in the field. The major recommendation based on this study is that a darn constructed of the residual clay soils be thoroughly 158

investigated. This investigation should be centered around the gathering and testing of undisturbed samples from the dam itself. This investigation would be facilitated by selection of a dam constructed of relatively stone-free clay. Where ped structure is exhibited by the in place soil, even slight sample disturbance could alter the apparent density. Both a low-power and a scanning electron microscopic investigation of the sample taken from the embankment would aid in determining the validity of the ped or packet theory of clay soils compacted in the field. 159

VIII. APPENDIX A

Dam Reports DAM: Fa RESERVOIR SIZE: 23--Acres !BEDROCK FORMATION: ___--t Roubidoux LOCATION: Dent County EMBANKMENT: HEIGHT 3 3 ft · LENGTH 530 ft. lsoiL SERIES: Hanceville Sec. 4 T.35 N. R.7 W. Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION:

LL 25 PL 16 PI 9 I

GRADATION: % FINER

#Ul 100 #200 61 #40 94 10t.L 3 8 #60 78 2J1. 19 #100 64 tp_ 16 + ' zo COMPACTION CHARACTERISTICS ASTM D 1557-,.66 MAX. DENJ21 lb/cu foPT. MOIST. --=.1~3.JL.% __

!PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC-7.6 pH-5.6 Gs-2.68 water level 20 ft. below crest water not passing through dam_

drop inlet W/ 3 ft. ¢ concrete pipe I-' 0\ j cutoff trench 0 DAM: Horns ey RESERVOIR SIZE: 40 + Acres BEDROCK FORMATION: ___-t Eminence LOCATION: Washington County EMBANKMENT: HEIGHT 50 ft · LENGTH 875 ft. SOIL SERIES: Clarksville Sec. 6 T.37 N. R.Z E. Stony Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 9 0 PL 4 3 PI 4 7 I

GRADATION: % FINER

#Ul 99 #200 9 7 #40 99 10/,L 95 #60 98 2J.L 9 3 #100 9 8 1~ 91

COMPACTION CHARACTERISTICS ASTM D 1557" 66 MAX. DEN. 77 lb/cu foPT. MOIST. 38 %

PREDOMINANT CLAY MINERALS: Kaolinite COMMENTS: CEC- 42.1 pH- 4. 8 Gs-2.84 7 ft. freeboard steep slopes w/ minor slides near toe--tilted fence line along toe of downstream slope I-' 0'\ cutoff trench I-' DAM: Blackwell RESERVOIR SIZE: l ~ Acres jBEDROCK FORMATION: ___--f Gasconade LOCATION: Dent County EMBANKMENT: HEIGHT 2 2 ft. . LENGTH "530 ft. !soiL SERIES: Clarksville Sec. 26 T.35 N. R.4 W. Stony Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION:

LL 60 PL 23 PI 3 7 I

GRADATION: % FINER

#l.Q. 99 #200 85 #40 95 10t,t. 77 #60 93 2JL 64 #100 89 1/L 56 l COMPACTION CHARACTERISTICS ASTM D 1557-66

MAX. DEN. 92 lb/cu ftoPT. MOIST. 24 %

'PREDOMINANT CLAY MINERALS: ,------~ I K 1· . ao 1n1te

COMMENTS: CEC-24.1 pH- 7 .5 G -2.80 s 6 ft. freeboard concrete spillway drop inlet w/ 24 in. CMP !---" previous embankment f ailure ( ? ) 0\ N DAM: Masters #2 RESERVOIR SIZE: 24 Acres BEDROCK FORMATION: Gasconade - Eminence LOCATION: Dent County EMBANKMENT: HEIGHT 36 ft. LE_NGTH 700 ft . SOIL SERIES: Clarksville Sec. 16 T. 35 N. R. 5 W. Stonx: Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 103 PL 34 PI 69 I GRADATION: % FINER I

#1.0 100 #200 99 #40 100 10tL 99 #60 99 2JL 95 #tOO 99 1J.L 94

SE.c /vftJ:srsli!S #I COMPACTION CHARACTERISTICS ASTM D 1557-66 MAX. DEN. 81 lbLcu ftoPT. MOIST. 36 %

PREDOMINANT CLAY MINERALS:

COMMENTS: CEC-35.8 pH-7.5 Gs-2.84

see Masters #1

1--' 0\ tN DAM: Sunnen RESERVOIR SIZE: 40+ Acres BEDROCK FORMATION: ___-' Colluvial on Gasconade LOCATION: Washington County EMBANKMENT: HEIGHT 50 ft . LENGTH 1300 ft. SOIL SERIES: Clarksville Sec. 4 T.37 N. R.l W~ S tony Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 4 2 PL 21 PI 21 I

GRADATION: % FINER

#Ul 100 #200 87 #40 WAS H&O 97 10J.L 57 #60 92 2JJ- 38 #100 88 1JL 3 7

COMPACTION CHARACTERISTICS ASTM D 1557-66 MAX. DEN.l08 lb/cu fbPT. MOIST. ~1..;;.8...::.%__

PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS:CEC-11.5 pH-4.9 G -2.72 s 13 ft. freeboard 150 ft. paved spillway drop inlet w/ two 4 ft. ¢ concrete 1--' pipes 0\ tilted slope indicators +>- DAM: Masters-~# 1 RESERVOIR SIZE: 24 Acres BEDROCK FORMATION: ___-! Gasconade - Eminence LOCATION: Dent County EMBANKMENT: HEIGHT 36 ft. LENGTH 700 ft • !soiL SERIES: Clarksville Sec. 16 T.35 N. R.S w: Stony Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 8 4 . PL 2 5 PI 5 9 I

GRADATION: % FINER

#Ul 100 #200 96 #40 100 1011. 83 #60 99 2JL 76 #100 98 1/.L 71

COMPACTION CHARACTERISTICS . ASTM D 1557-66 MAX. DEN. 84 lb/cu ftoPT. MOIST. --.::2:.;;;:;5....::%__

"PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC- 21. 9 pH-5.4 Gs -2.83 5 ft. freeboard 36 in. CMP w/ hooded inlet

cutoff trench I-' spring below toe of dam 0\ V1 DAM: Sunrise South RESERVOIR SIZE: .i:::> ACres JBEDROCK FORMATION: ____, Potosi LOCATION: Jefferson County EMBANKMENT: HEIGHT 35 ft · LENGTH 550 ft. lsoiL SERIES: Clarksville Sec. 36 T.39 N. R.4 E. Grvl. Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 6 8 PL 3 0 PI 3 8 I

GRADATION: Ofo FINER -413'~ #Ul 99 #200 89 ~ #40 93 10J,L 80 #60 91 2JL 7 2 #100 90 1J.L 70 t

COMPACTION CHARACTERISTICS ASTM D 1557 -66 MAX. DEN. 88 lb/cu ftoPT. MOIST. __;;.3..;;;1~% __

PREDOMINANT CLAY MINERALS: (rJ Kaolinite COMMENTS:CEC-24 . 6 pH-6 .5 G -2.82 s 6 ft . freeboard 60 ft. wide unpaved spillway ...... seepage--marshy at toe of dam 0\ 0\ DAM: Sunrise Central RESERVOIR SIZE: 2 S Acres BEDROCK FORMATION: ____, Potosi LOCATION: J efferson County EMBANKMENT: HEIGHT 35 ft · LENGTH 525 ft. lsoiL SERIES: Clarksv ille Sec. 36 T .39 N. R.4 E: Grvl. Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 7 0 PL 3 3 PI 3 7 I

GRADATION: % FINER

#l.Q 9 9 #200 92 #40 95 10J,L 8 7 #60 93 2P- 83 #100 93 1J..L. 82

COMPACTION CHARACTERISTICS ASTM D 1557-66 MAX. DEN. 85 l b / c u fbPT. MOIST. 33 %

PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC- 30.6 pH-4.9 G -2 . 82 s E~tooi!D 5 ft . freeboard 90 ft. wide paved spillway seepage through east abutment ...... eroded toe 0\ -.....) lake 20 ft. from toe of dam DAM: Ft. Westside RESERVOIR SIZE: !BEDROCK FORMATION:------4 Roubidoux (?) LOCATION: Crawford County EMBANKMENT: HEIGHT 23 ft · LENGTH 9 00 ft · lsoiL SERIES: Lebanon Sec. 2 T.39 N. R.4 W~ Silt Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 4 8 . PL 21 PI 2 7 I

GRADATION: % FINER

#1.0 100 #200 83 #40 99 10J,L 56 #60 94 2JL 38 #100 89 1J.L 37

COMPACTION CHARACTERISTICS ASTM D 15 57 ,.66 MAX. DEN}-00 lb/cu ftoPT. MOIST. 22 %

PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC-19.1 pH-4.6 Gs -2. 7 2 3 ft. freeboard earth spillways steep slopes--but no apparent 1--' seepage 0'\ cutoff trench 00 DAM: Rice RESERVOJR SIZE: 15 Acres jBEDROCK FORMATION: ___---t Roubidoux LOCATION: Crawford County EMBANKMENT: HEIGHT 4 2 ft. LENGTH 960 ft. !soiL SERIES: Lebanon Sec. 17 T.39 N. R.3 W. Silt Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION:

LL 2 7 PL 15 PI 1 2 1 _, _1s'\- GRADATION: % FINER #Ul 100 #200 58 #40 88 10J,L 35 #60 71 2JJ.. 21 #tOO 60 1J.L 18

COMPACTION CHARACTERISTICS ASTM D 1557-66 MAX. DEN.121 lb/cu ftoPT. MOIST. l 3 %

.PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC- 8. 7 pH-5.2 G -2.72 s 3 ft. freeboard seepage below toe into marsh varment dens in backslope at water ...... 0\ line 1.0 no cutoff trench ? DAM: Say ers RESERVOIR SIZE: 43 Acres BEDROCK FORMATION: ___-! Potosi LOCATION: Washing ton County EMBANKMENT: HEIGHT 60 ft · LENGTH 900 ft · !soiL SERIES: Clark s v ille Sec. 2 8 T .38 N. R.l E. Stony Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION:

LL 1 0 5 PL 41 PI 6 4 1

GRADATION: % FINER

#Ul 100 #200 9 8 #40 99 10t,t. 9 6 #60 99 2t,i. 9 3 #100 9 8 1J.L 90

COMPACTION CHARACTERISTICS ASTM D 155 7-6.6

MAX. DEN. 7 4 l b / cu f t OPT. MOIST. __;,4.;;;.2..;;.% __

PREDOMINANT CLAY MINERALS: Kao lini,te COMMENTS: CEC-44 . 9 pH-7 . 5 G -2 . 84 s 3 f t. f r eeboar d s l ides and seepage through steep crest section ...... 40ft . wide - spillway- eroded -....] 0 cutoff trench (?) DAM: Litt-le Prar1e RESERVOIR SIZE: l.UU Acres jBEDROCK FORMATION: ___--t Jefferson City LOCATION: Phelps County EMBANKMENT: HEIGHT 40 ft · LENGTH 1300 ft · lsoiL SERIES: Lebanon Sec. 21 T.38 N. R.7 W. Silt Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION:

LL 57 PL 23 PI 34 I

GRADATION: % FINER

#Ul 99 #200 9 5 #40 98 10/.L 80 #60 98 2JL 58 #100 9 7 1/L 44

COMPACTION CHARACTERISTICS ASTM D 1557-66 MAX. DEN. 94 lb/cu ftoPT. MOIST. 27%

PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC-13.7 pH-7.8 G -2.72 s 8 ft. freeboard granular toe drain drop inlet ...... --..J cutoff trench DAM: - -~ - RESERVOIR SIZE: 6 0 Acres fBEDROCK FORMATION:___ --t Gasconade LOCATION: Dent County EMBANKMENT: HEIGHT · E Clarksville St. LENGTH 9 . ~~ ~~ · ISOIL SERI S;_:.______Sec. 20 T.35 N. R.S W.

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 85 PL 35 PI 50 I

GRADATION: % FINER

#1.0 9 9 #200 98 #40 98 10tL 9 7 #60 98 2JL 87 #100 9 8 1JJ- 83

COMPACTION CHARACTERISTICS ASTM D 1557-66 MAX. DEN. 8 3 lb/cu ftoPT. MOIST. ~3....:.4....::;% __

PREDOMINANT CLAY MINERALS: . Kaolinite COMMENTS: Gs-2.80 6 ft. freeboard unlined spillway--eroded· very wet to~--marsh f-' several zones of seepage through -....] dam and abutments N DAM: Elsey RESERVOIR SIZE: j_ 1 ACres !BEDROCK FORMATION:___ -! Gasconade LOCATION: Washington County EMBANKMENT: HEIGHT 18 f t . LENGTH 770 ft. -- lsoiL SERIES: Clarksville Sec. 30 T.37 N. R.2 E. Stony Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION:

LL 8 4 PL 2 5 PI 5 9 1

GRADATION: % FINER

#Ul 9 9 #200 94 #40 98 10tL 86 #60 96 2JL 71 #100 9 5 1J..L. 69

COMPACTION CHARACTERISTICS ASTM D 1557-66 MAX. DEN.89 lb /cu ft .OPT. MOIST. 26 . 5%

PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC- 26. 9 pH-7.7 G -2.79 s 4 ft. freeboard cutoff trench

1-' two 50 ft. wide shallow spillways -.....! VI DAM: I~II~ D:u I.a!; RESERVOIR SIZE: BEDROCK FORMATION: Potosi LOCATION: St. Francois County EMBANKMENT: HEIGHT LENGTH SOIL SERIES:Clarksville Grvl. Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 96 PL 35 PI 61

GRADATION: ~ FINER

#Ul 100 #200 98 #40 100 10tL 94 92 #60 99 2P. #100 98 tJj. 90

COMPACTION CHARACTERISTICS

MAX. DEN. 79 lb/cu ft OPT. MOIST. 37%

PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC- 2 7 • 9 pH-7.1 Gs-2.82 10 ft.± freeboard rock spillway massive rock downstream bench ...... marshy valley floor "'-1 ~

-~ DAM: Timberl1ne RESERVOIR SIZE: ~~ Acr~~ BEDROCK FORMATION: Potosi LOCATION: St. Francois County EMBANKMENT: HEIGHT 75 ft. LENGTH · 67S±ft. SOIL SERIES: Clarksville Grvl. Loam

TYPICAL DAM SECTION SOIL CLASSIFICATION: LL 69 PL 30 PI 39

GRADATION: % FINER

#Ul 99 #200 85 #40 95 10tL 80 #60 91 2JL 79 #100 88 1J.L. 78

COMPACTION CHARACTERISTICS

MAX. DEN. 82 lbLcu fbPT. MOIST. 36%

PREDOMINANT CLAY MINERALS: Kaolinite

COMMENTS: CEC- 2 7. 8 pH-5.1 Gs-2.83 4 ft.± freeboard rock spillway eroded seepage through dam at several f-' locations -....J bench added to stabilize slope V1

------176

IX. BIBLIOGRAPHY

AYLMORE, L.A. G. and QUIRK, J. P. (1962). The Structural Status of Clay Systems, Clay and Clay Minerals Vol. 9, Pergamon Press, London, p. 104-130. ' BARDEN, L. and SIDES, G. R. (1970). Engineering Behavior and Structure of Compacted Clays, ASCE Vol. 96 No . SM4 , p. 1171-12 0 0 . ' ' BATES, T. F. (1952). Interrelationships of Structure and Genesis in the Kaolinite Group. Symposium on Problems of Clay and Laterite Genesis AIME New York, p. 144-153. ' ' BATES, T. F. (1959). Morphology and Crystal Chemistry of 1:1 Layer Lattice Silicates, Amer. Min., Vol. 44, p. 78-114. BATES, T. F. and COMER, J. J. (1955). Electron "Microscopy of Clay Surfaces, Clays and Clay Minerals, Natl. Acad. Sci. - Natl. Res. Council Publ. 395, p. 1-25. BATES, T. F. and COMER, J. J. (1959). Further Observations on the Morphology of Chrysotile and Halloysite, Clay and Clay Minerals, Pergamon Press, London, p. 237-248. BAVER, L. D. (1930). Relation of the Amount and Nature of Exchangeable Cations to the Structure of a Colloidal Clay, Soil Sci., Vol. 29, p. 291-309. BAVER, L. D. (1956). Soil Physics. John Wiley & Sons, Inc., New York, 4~. BERKELHAMER, L. H. (1944). An Apparatus for Differential Thermal Analysis, USBM R.I. 3762, 17 p.

BORST, R. L. and KELLER, W. D. (1969). Scanning Electron Micrographs of API Clay Minerals and Other Selected Samples, Proceedings, Int. Clay Con£. Vol. 1, Isreal Universities Press, Jerusalem, p. 871-901. w. F. (1964). X-Ray Diffraction Analysis ~f Soil BRADLEY, Clays and Structures of Clay Minerals, (1n) . Soil ClC* Miner~logy, University of North Carol1na Press, apel H1ll, p. 113-124. 177

BRADLEY, W. F. and GRIM, R. E. (1951). High Temperature Therm~l Effects of Clay and Related Materials, Am. M1n., Vol. 36, p. 182-201. BRAGG, W. H. (1949). The Crystalline State. G. Bell and Sons, Ltd., London, 352 p. BRAMAO, L., CADY, J. G., HENBRICKS, S. B., and SWERDLOW, M. (1952). Criteria For The Characterization of Kaolinite, Halloysite, and a Related Mineral in Clays and Soils, Soil Sci., Vol. 74, p. 273-287. BRETZ, J. H. (1965). Geomorphic History of the Ozarks of Missouri. Mo. Geol. Sur., Seconu series, Vol-.-61, 137 p. BRINDLEY, G. W. (1951). The Kaolin Minerals, (in) X-Ray Identification and Crystal Structure of Clay Minerals, Cpt. IT; p. 3 2-7 5, M1neralogTcar-B"o ciety of Great Britian Monograph.

BRINDLEY, G. W. and COMER, J. J. (1956). The Structure and Morphology of a Kaolin Clay From Les Eyzies, France, Clays and Clay Minerals, Natl. Acad. Sci. - Natl. Res. Council Publ. 456, p. 61-66.

BRINDLEY, G. W. and GILLERY, F. H. (1954). Mixed-Layer Kaolin-Chlorite Structure, Clays and Clay Minerals, Natl. Acad. Sci. - Natl. Res. Council Publ. 327, p. 349-353. BRINDLEY, G. W. and ROBINSON, K. (1947). Fireclay, Trans. Brit. Ceram. Soc., Vol. 46, p. 49-52.

CARROLL, D. and HATHAWAY, J. C. (1953). Clay Mi~erals in a Limestone Soil Profile, Clay & Clay M1nerals, Natl. Acad. Sci. - Natl. Res. Council Publ. 327, p. 171-182.

CARTHEW, A. R. (1955). The Quantitative Estimate o~ Kaolinite by Differential Thermal Analys1s, Am. Min., Vol. 40, p. 107-117.

CHAPMAN, H. D. (1965), Cation-Ex~hange C~pacity,. (in) Methods of Soil Ana1ys1s, ~mer1c~n Soc1ety of Agronomy-,-I~ Mad1son, W1scons1n, p. 891-904. D T and SHEELER, J. B. (1952). Cation-E~chan¥e DAVIDSON, · · · ty of Loess and Its Relation to Eng1neer1ng Capacl bl N 142 p 10-28 Properties, ASTM Sp. Pu · o. ' · · 178

DAVIS, D. W., ROCHOW, T. G., ROWE, F. G., FULLER, M. L., KERR, P. F. and HAMILTON, P. K. (1950). Electron Micrographs of Reference Clay Minerals API Project 49, Pre. Rpt. No. 6, Columbia Unive;sity, 17 p.

DUVAL, C. (1963). Inorfanic Thermogravimetric Analysis. Elsevier Pub ishing Company, New York, 684 p. EADES, J. L. (1970). Personal communication. GRAHAM, J. W. (1969). The Effect of Structure on the Consolidation of Compacted Unsaturated Clay Soils, Unpublished Thesis, University of Missouri - Rolla. GRIM, R. E. (1947). Differential Thermal Curves of Prepared Mixtures of Clay Minerals, Am. Min., Vol. 32, p. 493-501. GRIM, R. E. (1952). Ion Exchange in Relation to Some Properties of Soil-Water Systems, ASTM Sp. Tech. Publ. No. 142, p. 3-10. GRIM, R. E. (1953). Clay Mineralogy. McGraw-Hill Book Company, Inc., New YOrK, 384 p.

GRIM, R. E. (1958). Organization of Water on Clay Mineral Surface and Its Implications For The Properties of Clay-Water Systems, H.R.B. Sp. Rpt. No. 40, p. 17-23. GRIM, R. E. and BRAY, R. H. (1936). The Mineral Constitution of Various Ceramic Clays, Jour. Am. Cer. Soc., Vol. 19, p. 307-315. GRIM, R. E. and ROWLAND, R. A. (1942). Difhferentdial Thermal Analysis of Clay Minerals and Ot er Hy rous Materials, Am. Min., Vol. 27, p. 746-760, 801-818. R. W., HEATON, E., and ROBERTS, A. L. (1945). GRIMSHAW, Constitution of Refractory ~lays, Part II Thermal­ Analysis Methods, Trans. Br1t. Ceram. Soc., Vol. 44, p. 76-92. G d FRAULINI F. (1940). Properties of Kaolinite HARMON, C. as .aa~unction of' its Particle Size, Jour. Am. Cer. Soc., Vol. 23, p. 252-258. 179

HARRIS, M. T. (1969). A Study of the Correlation Potential of t~e Optimum Moisture Content, Maximum Dry Dens1ty,_and_Consol~dated Drained Shear Strength of Plast1c F1ne-Gra1ned Soils with Index Properties Unpublished Masters Thesis, University of Missouri ' - Rolla.

HAUSER, E. A. (1952). Genesis of Clay Minerals. Symposium on Problems of Clay and Laterite Genesis AIME New York, p. 100-106. ' ' HAUTH, W. E. and DAVIDSON, D. T. (1950). Studies of the Clay Fraction in Engineering Soils: Part II. Particle Size Distribution and Cation-Exchange Capacity, H.R.B. Proc., Vol. 30, p. 449-464. HOLTZ, W. G. and LOWITZ, C. A. (1957). Compaction Characteristics of Gravelly Soils, ASTM Sp. Tech. Rpt. No. 232, p. 67-86. JENNY, H. (1941). Factors of Soil Formation. McGraw-Hill Book Co., New YorK,~ p. KELLER, W. D. (1968). Principles of Chemical Weathering. Lucas Brothers Publishers, Columbia, M1ssour1, 111 p. KELLER, W. D. (1968). Flint Clay and a Flint-Clay Facies, Clays and Clay Minerals, Vol. 16, Pergamon Press, London, p. 113-128. KELLER, W. D. (1970). Personal communication.

KELLER, W. D., HANSON, R. F., HU~NG, l~. H. an~ CERVANTES, A. (1970). Active Alterat1on of Volcan1c Ro~k to Endellite and Its Precursor at a Hot Spr1ng in Michoacan, Mexico, Manuscript Submitted to Clays and Clay Minerals.

KELLEY, w. P. and JENNY, H. (1936). The Relationship of Crystal Structure to Base Exc~ange_and Its Bearing on Base Exchange in Soils, So1l Sc1., Vol. 41, p. 367-382.

KERR, P. p KULP J L and HAMILTON, P. K. (1949). niffere~ti~l Thermal Analy~is of Reference Cl~~ Mineral Specimens, API ProJeCt 49, Pre. Rpt. · 3, Columbia University, 48 p.

KOENIG, J. W. editor (l961). The Stratigraphic Succession 40 in Missouri. Mo. Geol. Sur., Second Ser1es, Vol. r16 p. 180

LADO, C. C.,_LAM~E, T. W., and MOH, Z. C. (1960). Recent So~l-L1me Research at M.I.T., Paper Presented at the 39th Annual Meeting of the Highway Research Board.

LAMBE, T. W. (1953). The Structure of Inorganic Soil Proceedings, ASCE, Vol. 79 Separate No. 3i5 49 p. ' '

LAMBE, T. W. ~1960). Compacted Clay (a symposium), Trans­ actlons, ASCE, Vol. 126, p. 681-756. LAMBE, T. W. (1962). Soil Stabilization. (in) Foundation Engineering, Edited by G. A. Leonards, McGraw­ Hlll Book Co., New York, p. 351-437. LAMBE, T. W. and MARTIN, R. T. (1960). Composition and Engineering Properties of Soils V, H.R.B. Proc., Vol. 36, p. 693-702. LAWS, W. D. and PAGE, J. B. (1946). Changes Produced in Kaolinite by Dry Grinding, Soil Sci., Vol. 62, p. 319-336. MACKENZIE, R. C. (1957). The Differential Thermal Investigation oFIClay. Mineralogical Society of Great Britian MOnographs, The Central Press, Aberdeen, 456 p. MICHAELS, A. S. (1959). Physico-Chemical Properties of Soils. Discussion, ASCE, Vol. 85, No. SM2, p. 91-102.

MIELENZ, R. C., SCHIELTZ, N.C., and KING, M. E. (1954): Thermogravimetric Analysis_of Clay and Clay-L1ke Minerals Clays and Clay M1nerals, Natl. Acad. Sci. - N~tl. Res. Council Publ. 327, p. 285-314.

MILLER, B. J. (1965). A Characteri~ation_of Four Limestone Derived Soils From The M1ssour1 Oza~ks, . Unpublished Thesis, University of M1ssour1 - Columbia. MILLER, W. D. and KELLER, W. D. (1963). Differe~t~tion Between Endellite-Halloysite and Kaol1n1te By Treatment with Potassium Acetate and Ethylene Glycol, Clays and Clay Minerals, Vol. 12, Pergamon Press, London, p. 244-253. J. K. (1956). The Fabrick of Natura~ Clays and MITCHELL, Its Relation to Engineering Propertles, H.R.B. Proc., Vol. 35, p. 693. 181

MURRAY, H. H. and_LYONS, S. C. (1956), Degree of Crystal Perfectlon of ~aolinite, Clays and Clay Minerals, Natl. Acad. Scl. ~ Natl. Res. Council Publ. 456 p. 31-40. ' NUTTING, P. G. (1943). Some Stand Thermal Dehydration Curves of Minerals, U.S.G.S. Prof. Paper 197E p. 197-216. PEECH, M. (1965). Hydrogen-Ion Activity (in) Methods of Soil Anal~sis, American Society of Agronomy-- Inc., Madlson, Wisconsin, pl 914-926. ' PRINCE, A. B. (1965). Adsorption Spectrophotometry, (in) Methods of Soil Analysis, American Society of Agronomy, Inc., Madison, Wisconsin, p. 866-878. QUINN, J. H. (1956). Origin and Age of Ozark Plateau Surfaces in Northwest Arkansas, GSA Bull., Vol. 67, p. 1726 (abstract). RICH, C. I. (1965). Elemental Analysis by Flame Photometry, (in) Methods of Soil Analysis, American Society of Agronomy, Inc~adison, Wisconsin p. 849- 865. RING, G. W., SALLBERG, J. R., and COLLINS, W. H. (1962). Correlation of Compaction and Classification Test Data of Soils. Public Roads, Vol. 32, No. 4, pp. 7 7- 8 7 . ROSS, C. S. and KERR, P. F. (1931). The Kaolin Minerals, U.S.G.S. Prof. Paper 165E, p. 151-175. ROSS, C. S. and KERR, P. F. (1934). Halloysite and Allophane, U.S.G.S. Prof. Paper 185G, p. 135-148. SAUNDERS, H. L. and GIEDROYC, V. (1950). Differential Thermal Analysis in Controlled Atmosphere, Trans, Brit. Ceram. Soc., Vol. 49, p. 365-374. SCHMIDT, N. 0. (1965). A Study of The_Isolat~on o~ Organic Matter as a Variable Affect1ng Eng1neer~ng Properties of Soi~s,_Doctor's Dissertat1on. University of Ill1no1s. SCOTT, w. w. (1948) Standard Methods of Chemkcaiz~4alysis. D. Van Nostrand Co., Inc., New Yor , P· SCRIVNER, C. L. (1960). Morphology, Mineralogy, and . Chemistry of the Lebanon Silt_Loam,_Unpubllsh~d Dissertation, University of M1ssour1 - Columbla. 182

SEED, H. B. and CHAN, C. K. (1959). Structure and Strength Characteristics of Compacted Clay ASCE Vol 85 No . SM 5 , p . 8 7 - 1 0 9 • ' ' . '

SIMONSON, R. W. (1949). Genesis and Classification of Red-Yellow Podzolic Soils, Proc. Soil Sci. Soc. Am., Vol. 14, p. 316-319. SLOAN, R. L. and KELL, T. F. (1966). The Fabric of M~chanically Compacted Kaolin, Clays and Clay M1nerals, Vol. 14, Pergamon Press, London, p. 289-296.

SPEIL, S. (1944). Application of Thermal Analysis To Clays and Aluminous Materials, USBM R.I. No. 3764, 36 p.

SUDO, T. and TAKAHASHI, H. (1956). Shapes of Halloysite Particles in Japanese Clays, Clay and Clay Minerals, Natl. Acad. Sci. - Natl. Res. Council Publ. 327, p. 158-171. TERZAGHI, K. (1958). Design and Performance of Sasumua Dam. Proc. Institution of Civil Engineers, London, Vol. 9, p. 369-394. THORNBURY, W. D. (1965). Regional Geomorphology of the United States. John Wiley & Sons, Inc., 6og-p. TROLLOPE, D. H. and CHAN, C. K. (1960). Soil Structure and the Step-Strain Phenomenon, Transactions, ASCE, Vol. 126, p. 1650-1698. U.S. DEPARTMENT OF AGRICULTURE (1960). Soil Classification­ A Comprehensive System (7th Approx1mat1on). Washington, U.S. Govt. Pr1nting Office. VEES, E. and WINTERKORN, H. F. (1967). Engineering Properties of Several Pure Clays as Functions_o~ Mineral Type, Exchange Ions, and Phase Compos1t1on, Highway Research Record No. 209, p. 55-65.

VOLOSTNYKH, G. T. (1965) Temperature of Mai~ Endothermic Effect of Kaolinite. Zap. Uses. M1n. Obshch., (Mem All Union Min. Soc.), Vol. 94, No. 4, p. 407-419- 95 reported in Min. Absts., Vol. 17, p. 555 (1966). WADA, K. (1961). Lattice Expansion of Kaolin Minerals by Treatment with Potassium Acetate, Am. Min., Vol. 46, p. 78-91. 183

WHITE, W. A. (1958). Water Sorption Properties of Homoionic Clay Minerals~ Illinois Geological Survey, R.I. 208, 44 p. WINTERKORN, H. F. (1940). Physico-Chemical Testing of Soils-Application of the Results in Practice, H.R.B. Proc., Vol. 20~ p. 798-806. WINTERKORN, H. F. and MOORMAN, R. B. B. (1941). A Study of Changes in Physical Properties of Putnam Soil Induced by Ionic Substitution, H.R.B. Proc., Vol. 21, p. 415-434. ZEIGLER, E. J. (194 8) . Effect of Mate rial Retained on the Number 4 Sieve on the Compaction Test of Soils, H.R.B. Proc., Vol. 128, p. 409-414. 184

X. VITA

Arthur David Alcott was born on 28 April, 1941 in Asheville, North Carolina. He received his primary and secondary education in Sumter, South Carolina. After high school graduation, he attended Colorado School of Mines, Golden, Colorado where he received a Geological Engineering Degree in June, 1964. He worked as a highway engineer for the Colorado Department of Highways until January, 1966, and as a soils and foundations engineer for Sverdrup & Parcel and

Assoc., Inc., St. Louis~ Missouri until August, 1968. He has attended the University of Missouri - Rolla since that time and was a graduate assistant during the 1969-1970 school year. The author is a registered "Engineering-in-Training" with the Colorado Board of Professional Registration. He is married to the former Molly Ann Orr of Lancaster, Pennsylvania and has two daughters, Laura Ann and

Jennifer Ashley.