THE EFFECTS OF FREEZlNG‘AND-THAWING ON STRUCTURAL STRENGTH OF SELECTED MTUMINOUS STAEEELIZES 3.5153 COURSE Thesis for the Degree of M. S. MECHKSAN STAYE UNWERSiTY Abbas Ghana-Bassiri i966 THESIS L [B k A R Y j Highs» Stun yahfl'tiry ..—._ —-—vv w- THE EFFECTS OF FREEZING-AND-THAWING ON STRUCTURAL STRENGTH OF SELECTED BITUMINOUS STABILIZED BASE COURSE By Abbas Ghane-Bassiri A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Civil Engineering 1966 ACKNOWLEDGMENTS I would like to express my deep gratitude to my advisor, Dr. Gail C. Blomquist for his able guidance and encouragement in the writing of this thesis. I am also grateful to the Asphalt Institute and the Michigan Asphalt Paving Association for their fi- nancial assistance. ii TABLE OF CONTENTS Page ACKNOWLEDGMENTS . . . . . . . . . . . . ii LIST OF TABLES . . . . . . . . . . . . iv LIST OF GRAPHS . . . . . . . . . . . . v LIST OF FIGURES . . . . . . . . . . . . vii INTRODUCTION . . . . . . . . . . . . 1 MATERIALS 0 o o o o o o o o o o o o o 3 LABORATORY TEST PROCEDURE. . . . . . . . . 5 FREEZING—AND-THAWING TESTS 7 Capillary Absorption Test. . . . . . . 7 Thermocouple . . . . . . . . . 8 Freezing- -and- -Thawing . . . . . . . . 8 Water Exposure . . . . . 10 REVIEW OF PHYSICAL PROCESSES OF FROST ACTION ll Freeze-Thaw of Concrete . . . . . . . 15 Frost Action in Stabilized Soil Mixtures. 15 RESULTS. . . . . . . . . . . . . . . 18 Laboratory Test Results . . . . . . . 18 Tables . .‘ . . . . . . . . . . 19 Graphs . . . . . . . . 21 Grain Size Distribution . . . . . . 32 CONCLUSIONS . . . . 3H APPENDIX . . . . . . . . . . . 36 BIBLIOGRAPHY . . . . . . . . . . . . 88 iii LIST OF TABLES Marshall sample parameters Marshall test results after "0", 6, 12 cycles of freeze-thaw . Marshall test results after 18 and 2“ cycles of freeze—thaw . Marshall test results at 100°F. Volume and height changes Summary of results for samples with cent A.C. 85/100. . . Sieve analysis on aggregate samples Sieve size sequences for grain size sis of aggregate. Temperature penetration . iv 0 A per analy- Page 38 40 A2 uu H6 117 76 77 82 Graph l. 10. ll. 12. LIST OF GRAPHS Page Marshall Stability (at 77°F) as a function of the freeze-thaw cycles (sand—gravel) . 50 Marshall Stability (at 77°F) as a function of the freeze-thaw cycles (sand-gravel, (submerged) . . . . . . . . . . 51 Marshall Stability (77°F) as a function of the freeze-thaw cycles (sand) . . . 52 Marshall Stability (77°F) as a function of the freeze-thaw cycles (sand, submerged). 53 Marshall Stability (77°F) as a function of the freeze-thaw cycles (gravel). . . . 5“ Marshall Stability (77°F) as a function of the freeze-thaw cycles (gravel, sub- merged). . . . . . . . . . . . 55 Marshall Stability (at 100°F) variation with asphalt content and freeze-thaw cycles (gravel) . . . . . . . . . 56 Marshall Stability variation with asphalt content and freeze-thaw cycles (S-G at lOOOF) o o o o o o o o o o o o 57 Marshall flow values as a function of freeze-thaw cycles (sand-gravel) . . . 58 Marshall flow values as a function of the freeze-thaw cycles (sand-gravel, sub- merged). O I O O O O O O O O O 59 Marshall flow values as a function of the freeze-thaw cycles (sand). . . . . . 60 Marshall flow values as a function of the freeze-thaw cycles (sand, submerged) . . 61 Graph Page 13. Marshall flow values as a function of the freeze-thaw cycles (gravel) . . . . . 62 in. Marshall flow values as a function of the freeze-thaw cycles (gravel, submerged) . 63 15. Marshall flow variation with asphalt cement content (at lOO°F) sand-gravel . . . . 6A 16. Marshall flow variation with asphalt cement content (at lOO°F) gravel. . . . . . 65 17. Marshall Stability variation (77°F) with asphalt cement content (A.C. 85/100) . . 66 18. Density variation with asphalt cement content. . . . . . . . . . . . 67 19. Effect of varying asphalt content on voids filled with bitumen. . . . . . . . 68 20. Volume changes as a function of the number of freeze-thaw cycles . . . . . . . 69 21. Water absorption as a function of the number of freeze-thaw cycles. . . . . 70 22. Degree of saturation as a function of the freeze-thaw cycles . . . . . . . . 71 23. Void ratio as a function of the freeze- thaw cycles (sand—gravel). . . . . . 72 2M. Void ratio as a function of the freeze— thaw cycles (sand) . . . . . . . . 73 25. Void ratio as a function of the freeze- thaw cycles (gravel). . . . . . . . 7H 26. Grain size distribution curves. . . . . 78 27. Grain size distribution curves. . . . . 79 Vi LIST OF FIGURES Figure Page 1. Electrical circuit for thermocouple con- nections . . . . . . . . . . . 81 2. Temperature vs. Time. . . . . . . . 83 vii INTRODUCTION In many parts of the country, natural sources of good, clean, base course material are becoming rapidly depleted. In metropolitan areas, as well as rural areas, it has become necessary to import from considerable distance or use a plant-processed product. In any event, costs run high and are continously climbing. One method of remedying this situation is to render the lower quality aggregate for use by stabilization with asphalt cement. The use of asphalt cement for the stabiliz- ing and for waterproofing of a coarse aggregate (gravel and sand) has been widespread and extensive for many years. The economic advantages of local natural aggregates for bituminous stabilized base courses can be realized only if certain limiting factors inherent to gravels, sand, and fillers, as they occur in nature, are properly taken into account in the design and control of the mix. It is believed that, in stabilizing a coarse aggregate (sand, gravel) with asphalt cement, the thin films of asphalt surrounding the particles serve to produce cohesion and superior load distributing properties, and increase grain-to-grain frictional (intergranular) resistance. [\3 There have been developed some criteria in evalu- ating the stability for some of these substandard materials during the past fifty years in the U. S. A. and Germany. Results of previous investigations have indicated, aSphalt stabilized base course materials, increase in stability, unconfined compression strength, and shear strength. The durability of asphalt stabilized soil mix- ture, when exposed to freezing-and-thawing, however, has not been investigated as a literature search indicates. If a pavement or base course (bituminous stabilized aggregate) design is to be based on an improved strength value, the expected loss in strength due to freezing—and- thawing must be known. It is the purpose of this investigation to attempt development of some criteria for a certain type of sub- standard base course materials, when exposed to the various cycles of freezing-and-thawing, through a program of physi- cal research and testing. MATERIALS The materials employed in preparation of the labo- ratory mixes were obtained from stockpiles of aggregate which were employed in the mixes for the experimental asphalt stabilized base course in the Alger Road Project located in Gratiot County, Michigan. The project was constructed for Gratiot County during the summer of 1963. These aggregates were substandard local aggregates which could not have been used satisfactorily without a sta— bilizing agent. The aggregate consisted of two types as follows: 1. The gravel which was uniformly-graded Effective size 0.0095-inch Uniformity coefficient 10.“ AASHO classification A-l-b(0) 2. A gap—graded sand Effective size 0.0078-inch Uniformity coefficient 2.3 AASHO classification A-3(0) The sieve analysis and grain size distribution curves for gravel, sand, and a mixture of 50 per cent sand and 50 per cent gravel are presented in Appendix III. Sieve analysis was performed in accordance with ASTM:C 136-u6 (A). (The materials retained on a 3/A-inch sieve were removed before mixing.) The asphalt cement used for preparation of all speci- mens, and in the base course construction of Alger Road, was 85/100 penetration. This material exhibited better stability and better compaction, according to previous in- vestigation (30). The specific gravity of asphalt at 77 (77°F) was equal to 1.026 (2). LABORATORY TEST PROCEDURE The mixtures for laboratory testing were prepared in three major groups, as following: Group A - A mixture consisting of 50 per cent gravel and 50 per cent sand, with 85/100 penetration asphalt cement. Group B - A mixture of fine—sand and A.C. 85/100. Group C - A mixture of gravel and A.C. 85/100. Asphalt cement content of the mixtures were varied over a range of increments of one-half per cent, from 3 per cent to 5 per cent. At each asphalt cement content in each group, at least 20 samples were prepared. All of the samples were prepared for Marshall Stability test at varying cycles of freezing- and—thawing. The samples were molded for Marshall Stability test- ing apparatus in accordance with ASTM Designation: D 1559- 58 T (l) with the following variations: 1. Both aggregates and asphalt cement were heated to 250—260°F. 2. The mixtures were molded and compacted at 210°F, since this gave the best compaction result on the field project. One batch was prepared for each sample. Fifty blows were applied to the each face of the specimen. Specimens then were allowed to cool to room temperature, and were tested after a fourteen day period. FREEZING-AND—THAWING TESTS Capillary Absorption Test Because of waterproofing characteristics of the asphalt cement, it was thought that it might also be effective in reducing moisture absorption by capillarity. To prove this theory, a simple laboratory procedure was employed. A sample of sand—gravel mixture with A per cent asphalt cement was placed in a rubber membrane to prevent moisture loss from ex- posed air surfaces, and placed with exposed surface on porous stones, then positioned in water so the water level was approximately l/A-inch below the base of the specimen. The specimen was covered to prevent moisture loss to the atmos— phere, weighed at different time intervals to determine the amount of moisture absorbed by the specimen. Tests revealed that the specimen absorbed practically no water during a seven day period of testing. It can be concluded that there is very little, if any, capillary moisture absorption. Therefore, there exists little possi- bility of formation of ice lense in this type of bituminous base courses. From the above it was concluded, there was no need for availability of free water supply to the samples during the freezing periods. Thermocouple To help correlate asphalt temperatures with depth, and to determine the required time for freezing and thaw- ing, two samples of sand-gravel mixture with four per cent asphalt cement content were frozen at a temperature of -20°C. The samples were then allowed to thaw slowly at room temperature (22°F). The temperatures of the samples were measured periodically throughout the freezing and the thawing periods by means of thermocouples in the samples. Thermocouples were placed at l l/A—inch depth from the top and the bottom of the specimens to check the tempera- ture within the specimens and observe the progress of freez- ing and thawing temperature in the specimens. The thermo- couples were inserted through the top of the specimens in holes punched with l/8-inch diameter drill. Entrance points were sealed with wax to waterproof and insulate the speci- mens. The electrical circuit for thermocouple connections and the table of temperature readings, and a plot of tempera- ture versus time are presented in Appendix IV. Freezing-and—Thawing Two types of freezing—and-thawing tests were carried out. The first was one in which the specimens were frozen at -20°C with all faces of the specimens exposed. The specimens were then allowed to thaw at room temperature of approximately 22°C. In the second type test, the speci— mens were immersed in fresh water at room temperature for 6A hours after which the specimens were frozen at -20°C. The specimens were allowed to thaw at room temperature for 2A hours. After thawing was completed specimens were im- mersed in water again for 2A hours to absorb water and the process of freezing-and—thawing was then repeated. Twenty-four hours of freezing temperature and 2A hours of thawing constituted one cycle of freeze-and—thaw. A 2A hour submergance was employed for the second group of speci- mens (submerged samples). Following the freeze-and-thaw cycle, the height and volume of each specimen was measured and samples of the second group were weighed after 2A hours of immersion. The height of the specimens were measured at three point, with 0.0005-inch accuracy. Data of these measurements have not been presented in this thesis, but the results including volumetric change and void ratio and water absorption following immersion of the second group are presented in Tables 3, A, 5, and 6 with corresponding graphs in Appendixes I and II. To determine the effects of freezing-and-thawing on stability of specimens, Marshall stability tests were carried out after "0", 6, 12, 18, 2A cycles of freeze-thaw, in accordance to ASTM Designation: D 1559—60T (l) with the following variations: lO 1. Samples in second group were brought to test- ing temperature in a water bath at 77°F or 100°F. 2. Samples were tested at 77°F, and 100°F rather than lA0°F, (l, 7) since many of the samples would have shown little if any, strength at 1A0°F. The 77°F and 100°F temperatures are more indicative of temperatures actually attained in highway base courses. The Marshall Stability test was selected to measure stability because the previous investigation on the same material at Michigan State University were based on this method (30). Generally, at least two samples were tested identically to obtain an average result. Water Exposure Samples in second type were submerged in fresh water (20°C), three days prior to the freeze—thaw test, and 2A hours after each thawing period, to accomplish saturation. Samples were weighed before and after immersion. Repre- sentative results, including water absorption and degree of saturation, are presented in Table 6 and Graph 27, in Appendixes I and II. REVIEW OF PHYSICAL PROCESSES OF FROST ACTION Temperature, together with soil moisture, is one of the main factors influencing the performance of a highway and behavior of a soil. Sudden temperature change, ac- companied by moisture, induce changes in mechanical stress in materials. These may deform highway pavements by crack- ing or spalling or by causing slabs to blow out (23). The expansion in volume of pure water or freezing of atmospheric pressure is approximately 9 per cent (36). Measurements of expansions as great as 60 per cent of the original volume of frozen soil masses have been reported. It is obvious that factors other than the expansion of the water originally contained in the voids of frozen soil must enter into physical process of frost action. These factors and the entire physical process of frost action have been studied by Taber (33, 35), A. Casagrande (10), and the others. Taber has summarized his concept (35) as follows: . . .Frost heaving is due to the growth of ice crystals and not to change in volume of water on freezing. Pressure is developed in the direction of crystal growth which is usually determined chiefly by the direction of cooling. Excessive heaving results when water is pulled up through the soil to build up layer or lenticular masses of segregated ice, which grow in thickness because water molecues are pulled into the thin film that 11 12 separate the growing columnar ice crystals from underlying soil particles. The first freeze breaks up a consolidated soil, increasing permeability and reducing its tensile strength so that less resistance to heaving is offered when refreezing occurs. Repeating cycles of freezing and thawing introduce pp new factors and do not alter the mechanics of frost heaving. In Taber's experiments natural conditions of soil freezing were duplicated by exposing the tOp of the speci- mens to freezing temperatures, insulating the sides with dry sand tem) ture soil (32, and maintaining the bottoms, whether sealed (closed sys- or open in a pan of water (open system) at a tempera- above the freezing point corresponding roughly to normal temperatures below the zone of frost penetration. The conclusions of previous investigations by Taber, 33, 3A, 35) Casagrande (10), Bouyoucos can be summarized (37) as follows: 1. Destructive frost heaving is almost invariably associated with the formation of segregated ice. 2. The total amount of frost heaving is very closely equal to the sum of the thickness of all layers of segregated ice in the frozen soil. 3. The total amount of frost heaving is in direct proportion to the increase in total water concent of the frozen soil. A. The soil must have a water content at least equal to a state of capillary saturation for ice segregation to take place. I3 5. A supply of water must be available for the growth of ice crystals, either from some portion of soil itself or from some external source, e.g., ground water table. 6. For normal field conditions of temperature a cer- tain minimum percentage of grains smaller than 0.02 mm is necessary for ice segregation. 7. One slow, gradual decrease in temperature well into the freezing range is necessary and sufficient to cause ice segregation and frost heaving. Subsequent thaw- ing and freezing may increase the severity of the frost heaving, but will not change the basic action. 8. A cumulative curve of degree hours of freezing plotted against time is a quantitative measure of the in- crease of frost heaving with time. (a) The following factors are all necessary for ice segregation and frost heaving. If any one of these factors is not present heaving will not occur: (1) Capillary saturation of the soil at the beginning of,or during the freezing process. (2) A free supply of water from within or without the soil. (3) A minimum percentage (3 to 10 per cent) of grains smaller than 0.02 mm. (A) A gradual decrease in temperature of air above the soil below freezing temperatures. iA It can be summarized that when a soil is subjected to freezing temperatures, several phenomena take place (23), as follows: 1. Moisture is supplied by upward flow from ground water to the growing ice lenses. 2. Under normal atmospheric pressure and at 32°F water is converted to ice, causing a decrease in density per unit volume as well as the 9 per cent expansion. 3. The initial moisture in the voids of the soil and the moisture supplied upward to the cold front start to freeze, forming ice lenses. A. Frost penetrates the soil. 5. Upon penetrating the soil, frost causes frost heaves of the soil or ground surface and pavements. Because of the upward supplied moisture, the frost heave is more than the increase in volume of 9 per cent of the initial moisture content originally contained in the voids of the soil before freezing. Hence frost acts fore- most through the conversion of water to ice. All these phenomena constitute what is understood as "Frost Action" (23, 2A). More recent research shows that the loss of support- ing strength of soil, particularly in the fine—grained soil, occurs during thawing period, when unsuitable soil below a pavement under unfavorable conditions of drainage become soft from saturation by melting ice. Thawing takes place from the surface downward, as well as upward from 15 underneath the frozen soil layer. The result is the loss of bearing capacity and stability of the soil causing damage to the pavement (21, 22, 23, 2A). Freeze-Thaw of Concrete Powers (28, 29) developed a theory to explain the freezing-and—thawing action in concrete. Concrete is believed to act somewhat as a closed system. Power's hypothesis rests mainly on the premise that, the destruction of concrete by freezing is caused by hydraulic pressure gener— ated by the expansion accompanying freezing of water, rather than by direct crystal pressure developed through growth of the bodies of ice. If the destructive action of freezing is due to hydraulic pressure, the resistance to movement of water must be the primary source of pressure, since concrete contains enough air-filled voids to accomodate water-to-ice expansion. The intensity of the hydraulic pressures developed during the freezing, depend on degree of saturation, pore size, and permeability characteristics. Frost Action in Stabilized Soil Mixtures Frost action in stabilized soil mixture has been studied by H. E. Winn (37) and Corps of Engineer (l9A3— l9A6). Results of these studies indicate that graded mixture resulting from a combination of 16 1/2 per cent sandy clay with 83 1/2 per cent of the pit-run gravel 16 was typical of the mixture in the investigation by Winn (37). The stabilizing admixtures used,among others, were bituminous materials including, tars, emulsions, cutbacks, and road oils. The following conclusions were drawn: 1. Many of the types of stabilized soil mixtures were liable to serious damage by ice segregation, retarded to various degree by the stabilization process, but occurring just as it would in natural soils. 2. Frost damage may be expected to occur in a sta- bilized soil mixture only when limiting conditions of in- itial and attainable moisture concent exist. 3. In general, natural fine-grained soils start to heave sooner, heave at a greater rate, and reach a greater total heave, than stabilized soil mixture exposed to the same conditions. A. Once capillary saturation is reached and ice segregation begins in a stabilized soil mixture, the rate of heaving is usually only slightly less than for natural soils. 5. The frost line penetrates a well-graded mixture at a greater rate than it does a natural fine-grained soil, resulting in less total damage of the well-graded soil. 6. All the admixtures were much more effective in reducing frost action when used with well-graded soil mix— tures than when used with natural fine—grained soils. 17 7. In general, the resistance to frost action of bituminous mixtures was directly proportional to the per- centage of admixture and the degree of saturation at the beginning of freezing period. When 100 per cent saturated, serious frost damage occurred. Limited studies were made by the Frost Effect Labo- ratory, of Corps of Engineers, from l9AA to 19A6 to deter- mine the effectiveness of various bituminous materials in preventing ice segregation in soils (21). It was found that ice segregation could be prevented by the addition of sufficient bitumen to render the soil impervious. However, the quantity required to reduce ice segregation to negli- gible amounts approached the asphalt content commonly em— ployed for construction of bituminous pavements. RESULTS Laboratory Test Results A summary of laboratory test results are as follows: 1. Sieve analysis and grain size distribution of sand, sand-gravel, and gravel. 2. Specific gravity of the sand, gravel, and asphalt cement 85/100. 3. Specific gravity of various type of mixtures. A. Bulk density of samples. 5. Per cent air voids in total mixtures. 6. Per cent voids filled by asphalt cement. 7. Marshall Stability and flow values of 77°F, as a function of the freeze-thaw cycles for submerged and un- submerged samples. 8. Marshall Stability and flow values at 100°F, in non-frozen samples of sand, sand-gravel, and gravel; and after six cycles of freeze-thaw in sand—gravel, and gravel mixtures. 9. Per cent air void changes as a function of the freeze-thaw cycles in submerged and unsubmerged samples. 10. Per cent volume changes, water absorption changes, and degree of saturation of the submerged samples as a function of the freeze—thaw cycles. 18 19 ll. Freeze penetration into the sample of sand-gravel versus time. These above are tabulated in Appendixes I through IV. Average quantity determined for each set of samples are presented. Tables Table l of Appendix I lists the following quantities: l. The sample number. 2. The sample series--series A, B, C are mixed samples of sand-gravel, sand, and gravel, respectively, and asphalt cement 85/100 penetration. ("Submg." in front of sample series, indicates the submergence of the samples before ex- posure to freeze—thaw cycles.) 3. The aggregate type—-S-G is sand-gravel, S is sand, and G is gravel. A. The asphalt cement content in per cent by weight. 5. The bulk densities of specimens tested at 77°F. \ The per cent air voids in total mixes. “\ION . The per cent voids filled with bitumen. Table 2 in Appendix I contains: 1. The sample number. 2. The Marshall Stability and flow value of non- frozen samples tested at 77°F. 3. The Marshall Stability and flow value and per cent air voids of the samples after exposure of six cycles of freezing-and-thawing, tested at 77°F. 20 A. The same values mentioned in item (3), but after 12 cycles of freeze-thaw. Table 3 contains: 1. The Marshall Stability and flow value and void ratio of the samples after 18 cycles of freezing-and- thawing, tested at 77°F. 2. The same values mentioned in item (1), but after 2A cycles of freezing—and-thawing. Table A contains: 1. The sample number. 2. Aggregate type. 3. The asphalt cement content. A. The Marshall Stabilities and flow values of the non-frozen samples, tested at 100°F. 5. The Marshall Stabilities and flow values of the samples after six cycles of freezing-and-thawing, tested at lOO°F. Table 5 contains: 1. The number of freeze—thaw cycles (submerged samples). 2. The per cent of increase in volume (é; x 100), measured at the end of thawing period. 3. The per cent of vertical expansion of the samples due to freeze—thaw cycles. 21 Table 6 contains: The water absorption in per cent, the degree of saturation, and the volume changes of the samples of sand- gravel, sand, and gravel with A per cent asphalt cement content, at "0", 6, l2, 18, 2A cycles of freezing-and—thaw- ing. Graphs Graph 1 through 6, indicate the variation of Marshall Stability as a function of the freeze—thaw cycles. Submerged Samples Generally, one can see from Graphs 2, A, and 6 that the Marshall Stability decreases as the numbers of freeze- thaw cycles increase. The samples with low asphalt cement content, lost more strength, and at higher rate, than the samples with higher asphalt cement content. But, non- submerged samples with higher asphalt cement content lost more stability than the samples with low asphalt cement content. It can be observed, that the submerged samples lost more stability than the identical samples not submerged before freezing. The rate of decrease of stability is high during the initial six cycles. Almost no change in sta— bility was shown after cycle number 6. Following cycle number 18, loss in stability is indicated by these graphs. The representative percentage of strength loss was as follows: 22 Type Per Cent Strength Lost Sand—gravel samples (submerged) 9-15% Sand samples (submerged) l2-20% Gravel samples (submerged) 11-18% Complete items are shown in Table 3, Appendix I. Another fact that can be realized from these graphs is that the samples with highest asphalt cement content lost the least amount of strength after 2A cycles of freezing- and-thawing. The sand samples lost a greater per cent of strength than did sand-gravel or gravel samples with the same asphalt cement content with other conditions being equal. Unsubmerged Samples It can be observed, from Graphs l, 3, and 5, that variations of the strength as a function of the freeze-thaw cycles is fairly straight line with a very small slope. The percentage of stability lost in various mixtures were as follows: Type Percentage Strength Lost Sand-gravel mixes 5 — 7.8% Sand mixes 7.A - 9.5% Gravel mixes A.2 - 7% 23 Complete items are shown in Table 3, Appendix I. The loss of stability in various mixtures may be contributed to two factors: 1. Effect of shrinkage and swell of a§phalt cemepp and for aggregates. Even the cleanest sands and gravels shrink considerably when subjected to low temperature (23, 2A, 38, A0). Asphalt cement has substantially greater contraction and expansion tendencies, when subjected to a temperature change (28). Excessive shrinkage may even cause cracks in some cases. It is unfavorable that the asphalt cement is most brittle and least ductile during the period of freezing, when the greatest shrinkage occur and greatest subgrade heave may occur (38, A0). The stability of hot-mixed base courses, however, is capable of over- coming the destructive forces of frost action in underlying subgrade materials. 2. The effect of water content, expressed as degree of saturation, resulting in increases of volume and re- sultant increases of void volume in the samples. Employing enough asphalt cement content, ice segre- gation was prevented by making the samples relatively im- pervious (8, 9, 21). The asphalt cement filled the aggregate voids to the extent necessary to reduce the moisture mi- gration. None of the previously developed frost action theories could be applied in the case of this test. The only action resultant was from expansion in volume of ice by 9 per cent and producing greater voids in the samples. 2A Another factor that could have resulted in reduction of stability were physical and chemical changes of the asphalt cement due to freezing-and-thawing (31). There is apparently no Marshall test criteria for designing mixtures at the test temperature equal to 77°F. The results obtained in this study cannot be compared with any existing criteria. However, the results of the Marshall tests for samples made of gravel and sand-gravel mixtures indicate good stability values at the test temperature of 77°F during 2A cycles of freezing-and-thawing. The Marshall results for sand samples were also good, but not as high as for the sand-gravel and gravel samples. The sand samples indicated less stability, greater flow, and greater percent- age of stability lost due to freeze-thaw action. Graphs 7 and 8 show stability variation with asphalt cement for samples tested at 100°F rather than 77°F, at "0" and 6 cycles of freezing—and-thawing. It can be observed that the gravel samples indicated more stability and less reduction in stability after six cycles of freeze—thaw test at 100°F test temperature, than did the sand-gravel samples. Comparison of stabilities of samples tested at 77°F and 100°F revealed that there is proportionally greater change in strength over the testing temperature interval at high asphalt cement content than at low asphalt cement contents. At the standard test temperature (1), 1A0°F, samples indicated a little, if any, stability especially those heated 25 in the water bath rather than air bath. The dependence of the stability and flow properties of bituminous mix- tures on temperature is due to changes in the rheological properties of the asphalt, the dominating factor being the great dependence of viscosity on temperature (1A, 21, A0). The resistance to deformation of bituminous material de- creasing rapidly with increase in temperature. Furthermore, it can be said on immersion in water at 1A0°F before Marshall Stability test, water penetrates rapidly into those regions of the mixture which have not been adequately protected by asphalt. This results in a rapid loss of cohesion and there- fore rapid loss in strength. At 1A0°F should gradually find its way into the asphalt-sand interfaces and destroy the cohesive bond. The temperature of 1A0°F which has been experienced in road surfacing is too high to apply in base courses. From the test result at 77°F and 100°F it can be con- cluded that at lower asphalt cement contents the mixture's stability is resultant from intergranular frictional, while at higher asphalt cement content cohesive forces produce most of the strength. Thus in stabilizing sand with asphalt cement, the thin film of asphalt serves to produce cohesive resistance as well as intergranular resistance (15, 17, 26, 30, 39). The percentage of the total strength supplied by each of the two actions depends on temperature and asphalt cement content of the mixture. The proportion of the total strength supplied by the cohesive resistance varies directly with asphalt cement content and inversely with temperature (30). Graphs 9 - l5 These graphs show the Marshall flow variation versus various cycles of freezing-and—thawing for varied types of mixtures. One can state that these graphs are rather erratic. The variations could have been in the reading of flow dial, since it is quite difficult to read the dial at precisely right instant, when conducting the test, because of speed of the test. The errors are a little more for test at 77°F than for those at 100°F. As it has been stated in previous work (30): A possible cause for even larger variations than attributable to reading difficulties stems from the nature of the sample itself. At lower temperatures, such as 77°F, as the sample is compressed, a combi- nation of frictional resistance and cohesion tend to hold sample together. As further strain occurs, numerous frictional resistance and cohesion points break eventually resulting in failure. It can be concluded that some samples may reach this failure point more rapidly than others but at no less sta- bility. At the higher temperatures the asphalt cement is softer and acts slightly as a lubricant. The higher temperature would result in decreased cohesion, the particles being slightly better lubricated then the lower temperature, and with co- hesive bonds weakened, will slide past each other more rapidly. Thus, at higher temperature the mechanism of failure would occur more regularly and more consistantly. Some results can be ascertained from these curves as follows: 27 l. The Marshall flow increases as the number of freeze-thaw cycles increase. This may be because of loss in cohesion, due to expansion and contraction of materials and the effect of water. 2. The rate of increase in the flow values is much higher between 6 and 18 cycle than the other cycles. 3. The sand, sand-gravel, and gravel samples yielded maximum flow at 5 per cent asphalt content at the end of any cycle. A. The sand mixtures indicated greater flow than sand-gravel or gravel mixes, and increased at a higher rate. In the same way sand-gravel samples had greater flow value and flow increased at the higher rate than gravel samples. 5. The flow values changes more regularly at 100°F, than at 77°F (Graphs 15, 16). Graphs 15 and 16 Graphs 15 and 16 indicate the Marshall flow variations with asphalt content for the sand-gravel and gravel samples tested at 100°F before and after six cycles of freezing-and- thawing. It can be said that the flow value increases as the asphalt cement content increases. After six cycles of freeze-thaw the flow values were greater than corresponding non-frozen samples, the rate of increase being approximately equal in both cases. As mentioned above, the dependence of flow properties of bituminous mixtures on temperature of testing is due to 28 changes in the rheological properties of the asphalt cement. The resistance to deformation of bituminous materials de- crease rapidly with increase in temperature (2A, 31, 38). Graphs 17, 18,19 In the graphs 17, 18, and 19, the Marshall Stability, density, per cent voids filled with asphalt of three groups sand-gravel, sand, gravel have been plotted as a function of the asphalt cement content. The purpose was to attempt to make a comparison between the three types of aggregate. It can be observed that the gravel mixture had superior stability, higher density and a greater percentage of voids filled with asphalt cement, than the two other types. The Marshall Stability value (Graph 17) passes through a maximum as the asphalt cement content increased (1) at 3 1/2 - A per cent asphalt cement content for sand-gravel, (2) at A per cent for gravel, and (3) at A 1/2 per cent for sand samples. Thus the test gave an "optimum asphalt cement content" for which the compressive force was maximum. The density (Graph 18) also goes through a maximum as the asphalt cement content increases and the Optimum asphalt cement content for density is at or slightly below to the optimum for stability. The gravel samples had higher densities than the sand and sand-gravel samples, and in the same manner, sand—gravel samples had greater densities than the sand samples (A0). This is because the well-graded gravel and sand-gravel have been compacted better and to 29 the higher densities than of the sand alone which is fine- grained. No optimum density has been reached for the sand samples in the range of 3 - 5 per cent asphalt cement con- tent. The voids filled with bitumen (Graph 19) increased as the asphalt cement content increases. The variation is linear in each type of aggregate but with different slope (30). At any particular asphalt cement content, gravel mixtures had a higher percentage of void filled than the sand-gravel and sand mixtures, since the gravel had a lower void ratio, and better gradation initially, than the sand- gravel and sand minerals. (See Grain Size Distribution, Appendix III). Based on this discussion, the gravel mixture is the best suited for hot-mixed bituminous stabilized base courses. Graph 20 This graph shows the per cent increase in volume as a function of the freeze-thaw cyCIes in samples of gravel, sand-gravel and sand with A per cent asphalt cement content. This is the typical curve for all samples, therefore, the curve for all samples were not presented. Generally, the volume of samples increased as the number of freeze-thaw cycles increases. It can be observed that the volume change was slightly greater in sand mixture than the sand-gravel and gravel mixtures and volume increased at a higher rate during the first 6 cycles. Almost no volume 30 changes occurred after cycle number 6. The maximum volume during the 2A cycles of freeze-thaw test did not exceed by more than 1.26 per cent the volume at time of molding of the sand-gravel.* From the above volume change determination it is evident that the total volume of pore space is limited and capillary absorption cannot exceed that at the time of mold- ing by more than that by 1.26 per cent volume increase. Another fact is obvious at this point. With the in- crease in volume, the densities should have decreased and this could be another reason for loss in stability due to the freeze-thaw exposure. The point at which samples remain uniform in volume was not attained after 2A cycles. Graph 21, 22 Graphs 21 and 22 show the per cent water absorption and degree of saturation, respectively, as a function of the freeze-thaw cycles, for samples of sand, sand-gravel, and gravel with A per cent asphalt cement content. It can be seen that both water absorption and degree of saturation increased as the number of freeze-thaw cycles increased. The increase in per cent water absorption and degree of saturation were very low during the first 18 a”.Criteria for soil-cement samples require that maxi- mum volume at any time during the 12 cycles of the freeze— thaw shall not exceed more than 2 per cent of initial volume-(38). 31 cycles for all samples, but it was increasing rapidly after cycle number 18. An examination of air void values (Graphs 23 - 25) and the degree of saturation, it is evidenced that the voids have not been completely filled with water and only 20 per cent of voids have been saturated after 2A cycles. The above results lead us to the conclusion that the mixture was rather impervious and only some of the voids have been available to be filled with water (11). However, the high hydraulic pressures could not be formed from freezing of water in the pores because a portion of the water was forced out of the sample as soon as freezing pressures were created. . It was indicated previously that the following factors are all necessary for ice segregation and detrimental frost action will not appear unless the following conditions are all present (21, 23, 2A, 35, 37): l. The material must be frost susceptible. 2. The capillary saturation of the material at the beginning and during the freezing process must be attained. 3. A source of water must be available. A. Gradual freezing temperature must penetrate the material. Referring to the capillary absorption tests and the Graphs 20, 21, and 22, it is evident that the asphalt cement content in hot-mixed bituminous stabilized base courses 32 would prevent the migration of water resultant of capillary action, necessary for ice lens formation. It is concluded that the ice segregation will be pre— vented by additions of sufficient asphalt cement to hot- mixed bituminous stabilized base courses, thus rendering the aggregate mass impervious (8, 9). Grain Size Distribution Graphs 26 and 27 in Appendix III represent the grain size distribution of the sand, gravel, sand—gravel mixture, respectively. As a preliminary criterion of the gradation of pit- run gravel, Germans utilize the Aggregate Grading Chart after Ruthfuchs (13, A2). Instead of drawing sieve sizes in a log-scale, these were shown on a square root-scale (Graph 26). Any Straight line drawn from the origin through the diagram, is in effect a "maximum density curve." Natu- ral well-graded aggregates very seldom approach this theo- retical voidless status, but generally show a "hump" above the minimum void line with aggregate voidage between 15 and 25 per cent. Nevertheless, the Ruthfuch's (A2) chart is considered a valuable tool for preliminary suitability investigations for local material for bituminous stabilized base course gradation. The flatter the gradation curve and the more it approaches the "minimum void line," the higher is the Marshall Stability as long as sufficient air voids remain 33 in the mixture. It can be noted that the gravel has flatter distribution curves than sand, and sand-gravel mixture, therefore the mixture of the gravel and asphalt cement would have more stability than the others. Three curves in Graph 27 represent the grain size distribution of the sand, gravel and sand-gravel minerals. The dashed curve represents ideal grading based on Weymuth's Theory (27) of particle interference. It can be noted that the grain size distribution and texture of the natural aggregate, influence the stability of the samples as well as the flow values. The finer the aggregate, the less stability would be expected. A fine- grained sand generally has a lower density than a coarse- grained gravel. The graphs show that the mixture of gravel and asphalt cement should be more stable and more dense mix than sand or sand-gravel mixes. CONCLUSIONS From the results obtained during the course of the re- search described in this thesis, the following conclusions can be drawn: 1. The hot-mixed bituminous stabilized base course exhibit the following benefits as regard freeze-thaw test: (a) Elimination of capillary water movement through sample and prevention of ice lens formation by addition of sufficient asphalt cement content. (b) The loss of strength during 2A cycles of freeze-thaw was not significant and ranged from A.2 per cent to 18 per cent of the initial non-frozen strength of the sample. (c) In submerged samples the greatest per cent of increase in volume of samples did not exceed 0.9 to 1.26 per cent of initial non- frozen volume, which is not significant and will not produce any cracking in the base courses. 2. The samples with greater air void and less density (fine—grained sand) lost more strength due to freeze-thaw action. 3A 35 3. Maximum strengths occur at or slightly below optimum densities for various cycles of freeze—thaw. A. Sand-gravel mixed with 3 1/2 per cent to A per cent asphalt cement content, gravel with A per cent asphalt cement content and sand with A 1/2 — 5 per cent asphalt cement content exhibited the maximum stability. 5. As a lower course of the pavement, base courses do not reach as high temperature as the surface, therefore, the test temperature of 77°F and 100°F was chosen for the course of this study. There is no information available in the literature concerning this type of construction and especially this low test temperature. There also is no Marshall test criteria for designing mixtures for this ser— vice. However, the result of the Marshall tests indicate good stability and flow values at the test temperatures used, before and after exposure of freeze-thaw. APPENDIXES 36 APPENDIX I TABULATED RESULTS 37 38 0.0m ow.sa o.mma m m : m om :.wm om.mH m.mma m\H : m = m ma :.mm o:.mm m.omH : m = m ma m.om om.:m H.mHH m\H m m = m NH :.wH w.sm m.wHH m m A.w5nsmv m SH 2.:m m:.ma w.:mH m m m ma m.om 0.0m m.mma m\H : m m 2H m.mm m.mm H.HNH : m m ma s.am m.mm o.omH m\H m m m ma m.wa H.sm H.mHH m m m Ha m:.wm mm.mH m.mmH m mum = < OH wH.mm wm.ma m.oma m\H : cum = < m m:.Hm mw.sH mmfi : cum = < m om.wm Ho.mH m.sma m\H m mum = < s ms.:m :m.ma A.mma m mum A.m5nzmv a m om.sm mH.mH H.0MH m mum < m Hm.:m mm.wa m.oma m\H : mum < : mn.om ms.na m.mma : cum < m om.mm sm.ma m.mma m\H m mum < m m.mm om.ma m.sma m mum < H A A A .ph\mnav A m mama moHLom honesz omwwwm mmwm> zmwMMmm wmmmmmw mpmwopmw< maqewm mHQEmm mmmemz¢m Apfimemo pampeoo UHo> LH< xazm uamnome mpwwmpww¢ mHQEmm oHQEwm UmSCHpGOOIIH mqm<9 A0 O.OH OH O:OO O.OH :H OOOm O.OH OOOO Om O.Hm :H OHOm m.om O.HH OOOm O.mH OOHO OH O.:m mH OOOm 0.0m O.OH O:sm HH OHOm OH 0.0m O.OH OOOm H.Om A OOOm OH omOm AH 0.0m O O:OH 0.0m O OOOH OH O:Hm OH I O.OH OOOm I O.mH OOOm OH OsOm OH I :H OOOm I O.HH OOOm HH OOHO :H I mH Ommm I O OO:m O.OH OOsm OH I OH o::m I O :m:m OH O:Om mH I O HOom I O OOom O OOmm HH O.OH OH mOsO :.sH HH OOOO .OH OmOO OH OO.OH mH OHOO O.OH 0.0 OOOO .O OHOO O O.OH O OOH: O0.0H O Omm: O OO:: O .O0.0H O O:OO Os.OH O OOOO O Omm: A OO.OH O OOH: O.OH s Ooo: .O OOO: O I OH OOOO I OH OOOO OH OmOO O I OH OHH: I :.O omH: OH oom: : I m.O om:: I 0.0 OOH: O OHO: m I O OOO: I H.O OOO: O OOO: m I 0.0 Oom: I m.> omm: m ONO: H HOV HOV OOHo> H.OH HO.OO H.OOHO OOH0> H.OH HO.OO A.mOHO H.OH H0.00 H.OOHO OHO zOHO OOHHHOOOO EHO OOHO OOHHHOOOO BOHO OOHHHOOOO hmOeOz maoemm macho NH macho m macho :o: NH .0 «:0: mmem< mBQDmmm Emma qq A.OH HO.OV A.OOHV OOHo> H.OH HO.OV A.mOHV A.OH HO.OO A.mOHO AHO OOHO OOHHHOOOO OHO OOHO OOHHHOOOO OOHO OOHHHOOOO Amnesz oHQEmm macho ma macho m macho =0: Umscfiucoollm mqm A.2H Ho.ov A.mnHv mUHo> A.QH Ho.ov A.mnHv CH .AOHQ HHH ZOHm OpHHthpm HH< BOHm OpHHHnmpm Honesz mHQEmm mHvo :m mHvo OH 3¢mBImNmmmm mo mmqowo am Qz¢ mH mmem< mBHDmmm Emma HH H.OH HO.OO H.OOHO OOHo> H.OH H0.0V H.OOHO OH .OOHO AHO OOHO OOHHHOOOO AHO OOHO OOHHHOOOO honesz OHQEwm mHomo :m mHoOo OH UmSCH QCOOIIM magi. AA 0.0H OOO m m QON N.OH OOOH N\H O m an 0.0 OOOH O m nOH 0.0 O:O m\H m m nAH Ompmme poz H.A mOA m m OOH I I HH OOO m m an I I OH OmA N\H O m QOH I I 0.0 OOO O m OOH I I H.O OOOH N\H m m QNH I I O.A OOA m m OHH OH OAOH 0.0 OOOH m OIO OOH a com O.A OANH N\H O OIm mm m.O OOO H.A OOOH O OIm OO O ONO A ONO N\H m OIm HOA A ONA O OOO m OIm OO O mHNH m.O mmNH m OIm mm m mAHH O.A mmOH N\H O Olm OO :.O OOO A OAOH O OIm mm O.A OOOH N.A OOO N\H m OIm Om A OOO O OAO m OIm mH A.QH H0.00 A.mnHv A.:H H0.0V H.OQHV moOOH moOOH OOOOH moOOH OOHO AOHHHOOOO OOHO OOHHHOOOO OOOOOOO OOOH Omnesz pHmnam< mpmwmhmw< mHQEmm mHoOO O mHOOO =O= mOOOH B< mBHDmmm Emma HH¢mmm .:H H0.0 > .cH H0.0 > .:H H0.0 OOH x II OOH x ll OOH x II .2 HO 3 H2 O.O OOOOO Amhesz mHoOO Avmmhmansmv AOOOHOEQSOV AOOOHOEQOOV .O.< AO csz HO>OHO .O.< OO Osz O:mm .O.< AO OOH: .OIO mmwz m mqm<8 A7 OH.H :O.H OH.H HOOH x mwv OOOOOO mesHo> 0.0m o:.HO O:.Om Om OO.m OH.: OO.m .OO OOH\OO OOOOHO OOOm: OOOO.O OHOO.O OOOO.O O>OOO :m :O.O OO.O HHO.O AOOH x mwv OO:OOO mesHo> H.OH :.OH m0.0H . Om AH.H OA.H Om.H .EO OOH\OO OOOOHO OmOO: OOOO.O OHOO.O OOOO.O HONOOOO AOOOOO OOOOO OH OO.O OO.O OA.O HOOH x mwv OOOOOO OesHo> OA.OH O.OH OO.HH Om mOO.O OOO.O OO0.0 .AO OOH\OO OOxOHO OmOO: mHOO.O OmO0.0 HmO0.0 AnOeHO m>mmm mH > OO.O mO.O m:A.O HOOH x HOV OOOOOO mesHo> m.O OA.A OO.OH Om OHO.O OHA.O OOO.O .HO OOH\OO OOOOHO OmOO; OOOO.O O:O0.0 ::OO.O AcocHO O>Omm O m.O m.O H0.0 ACOHmHmEEH Hmpmmv coHpmpspmm mo mmuwmm OOm.O o:m.O O:O.O HOOHOOOEEH AOOOOO .Hw OOH\oo O:OOHQ Hmpmz o o O AONOOLM hwummv m>mmm O Awmwpwehsmv Omwmensmv I Ommhmehsm .O.< O: .HOOOOO .w.< O: .OOOO .m.< O: .OIW OOHOOHOOOOO . OHoOO OOH\mO .o.¢ Bzmo mmm O mBHB mmqmz.....=—---- ~‘-. 2000- ' : \ 1500 T I I | F O 6 12 18 2A Freeze—Thaw Cycle Graph 3.-—Marsha11 Stability (77°F) as a function of the freeze-thaw cycles (sand). A500 . A000 3500 3000 Marshall Stability — lbs 2500 2000 1500 53 ~ ’/.\‘ 5--_- 0",‘u. ._ . ‘s, . : \‘\ ~~~~ . \ \ _— § \ ““‘ ‘ \ \ \ \ . \ \ \ \ ‘ \O - ’- ~ ——---—_ _. \.-—-¢—-""'-—- \‘ \.~ \‘ \. 1 l T I 6 12 18 2A Freeze—Thaw Cycle Graph A.—-Marshall Stability (77°F) as a function of the freeze-thaw cycles (sand, submerged). 3500 3000 5A l l 1 12 18 2A Freeze—Thaw Cycles OH ox- Graph 5.--Marshall Stability (77°F) as a function of the freeze-thaw cycles (gravel). gum—womqugmmmsa .‘E' I h a , p 55 6000 ‘ 5500 - 5900 w A500 - Marshall Stability - lbs. A000 ‘ 3500 w 3000 ‘ l l 1 12 18 2A Freeze-Thaw Cycles 0. ON- Graph 6.--Marshall Stability (77°F) as a function of the freeze-thaw cycles (gravel, submerged). Marshall Stability - lbs. 56 3000‘ "0" Cycle 9 L 6 Cycle 2500‘? @—.____§ pf4j\ x”/ "“‘43 2000- 13/ \\\ // E] 1500- 1000- '5001 1* I l i ‘r 3 3 1/2 A A 1/2 5 Asphalt Content - per cent Graph 7.--Marshall Stability (at 100°F) variation with asphalt content and freeze-thaw cycles (gravel). Marshall Stability - lbs. 57 I 3000- 2500~ 2000‘ "0" cycle 6 cycle 1500- /o a ‘ o l,”’ 100.. 49/3 500‘ I i r l I 3 3 1/2 A A 1/2 5 Asphalt Content - per cent Graph 8.--Marsha11 Stability variation with asphalt content and freeze-thaw cycles (S-G at 100°F). Marshall Flow — 0.01 in. 58 15; 5 1A- [A.S 13- /- \ \ //,// 12H 11-. /. ’/'l x” 1’ ______ ‘A 10" e —-——". rim—“7.2 at {Hm-'OO‘. I P ' 'I' 1 l l I 1 l 0 6 12 18 2A Freeze-Thaw Cycles Graph 9.--Marshall flow values as a function of freeze- thaw cycles (sand-gravel). Marshall Flow - 0.01 inch 59 16- 15— 1a- 13- 13- IL- 10- 8- o—--——--—o—-—--—-O 1 h I l l I I 0 6 12 18 2A Freeze—Thaw Cycles Graph 10.--Marshall flow values as a function of the freeze-thaw cycles (sand-gravel, submerged). flex-hag I 9 ¢ 4 firm—:O.... “O:mm I I Marshall Flow - 0.01 inch 60 18~» . ' 5% 17- \ / 16.. ’ . 15- A” A 1/2% lA- xr- ’’’’ '—"’/ 13- . \ / . /°\ 12: ’ f/ \\ \‘Ag / // I I I 0 6 l2 18 2A d q Freeze-Thaw Cycles Graph ll.--Marshall flow values as a function of the freeze-thaw cycles (sand). Iran-ax Marshall Flow — 0.01 inch 16- 15‘ 1A.. 131 12~ 11. 61 /‘ \o—————05% ..\ ~.\o A 1/2% //’A% 3% 1 rl I l ' 0 6 12 18 2A Freeze-Thaw Cycles Graph 12.-~Marshall flow values as a function of the freeze-thaw cycles (sand, submerged). Marshall Flow - 0.01 inch 16- 15‘ NH 13H 61 o/ " .‘~.\.. N“-0A 1/2% /”“4% 3% 1 i l j l 0 6 l2 18 2A Freeze-Thaw Cycles Graph 12.-~Marshall flow values as a function of the freeze-thaw cycles (sand, submerged). I”? _ I'O- zn-r-w—hi; gran- l Marshall Flow - 0.01 inch 19 1A- 13‘ 12. F‘ H P T \P 62 5% /J//——A 1/2% I I I I I 0 6 12 18 2A Freeze—Thaw Cycles Graph l3.--Marshall flow values as a function of the freeze-thaw cycles (gravel). Cab—Tr 2. _ ". .33?! I. ‘. . I Marshall Flow — 0.01 inch 63 15- lAd o 13. / \. 1.. / 11H \ / -ml“ \ /’,’ \_-\~-. 10" ----- ‘-""" Pi—r‘ 9- o--—' "”‘~-‘-:~:-——————Q ————— -—O p 0" ’~-— o\\ 81 ° ' \~ I \ \O, t. 7- / . I , A LA 3... 2. 1 l ' I l l 0 6 12 18 2A Freeze—Thaw Cycles Graph 1A.--Marshall flow values as a function of the freeze-thaw cycles (gravel, submerged). Marshall Flow — 0.01 inch 6A 12- Legend: O-——-O "0" Cycle ,O----O 6 Cycle 11- [j—ufl "0" Cycle (submg.) [J—-D 6 Cycle (submg.) 6 Cycle (submg.) 10- ’l,’ 6 Cycle 9“ 9- eii;—-46/55 , I3 8* Mach ’,E{/’§<\\\\\\\"O" cycle (submg.) 7- 6- 5- I Tr* I I I 3 3 1/2 A A 1/2 5 % Asphalt Content - per cent Graph 15.-~Marshall flow variation with asphalt cement content (at 100°F) sand-gravel. Marshall Flow - 0.01 inch 12-- 11‘ 10- 65 ’ 60 1 b . /,//\__Lc_e (su mg) I/’ . \ 6 Cycle H 0" C cle "O" C cle (submg.) . Legend: o—o "0" Cycle O---O 6 Cycle E}---E]"0" Cycle (submg.) III—436 Cycle (submg.) I I I I I 3 3 1/2 A A 1/2 5 % Asphalt Content - per cent Graph l6.--Marsha11 flow variation with asphalt cement content (at 100°F) gravel. Marshall Stability 66 6000- \ e 5500 A , \\{—Gravel @\~\/ 0 5000 " A500- . AM A A000 u \\\\\éf Sand-Gravel A 3500 ‘ E]\\ [———Sand 3000 'i / ‘ e /,/ ’,,,43 . ,El’ 2500 -' // // ID 2000 I I f l I 3 3 1/2 A A 1/2 5 Asphalt Content - per cent Graph 17.--Marshall stability variation (77°F) with asphalt cement content (A.C. 85/100). Density - lbs./ft.3 138: 137- 136; 135‘ 13A- 133* 132- 131— 1301 129- 128- 127- 126- 125- 12Af 123: 122. 121- 120O 119 67 I S-G lfll ./ / R Sand 1 I *I I *r 3 3 1/2 -A A 1/2 5 5 1/2 Asphalt Content - per cent Graph l8.-—Density variation with asphalt cement content. 68 AOO 38- 3A.. 32- 30‘ 28'- 26 - 2AI- Voids Filled With Bitumen - % 22-— 2O - 18-‘ 16- 1 I ' I 3 3 1/2 A A 1/2 5 Asphalt Content - per cent Graph l9.--Effect of varying asphalt content on voids filled with bitumen. 69 8 Legend: 2 Sand-gravel X lo 6- > ------ Sand Sand - A% A.C. l.lifi O -—»u~—--Grave1 (I) w . - , p l 2 Gravel Gravel I O a l o O "‘ sand 1 .a", 3(/ H /\ __,’/\‘_ ,0—4 ./// H l / Sand-gravel, A% A.C. :3 o 6— I I' O // CH /’I O 0. Ll ‘ III” 4.3 If 8 O 2 J‘ I” O . h H (i) 0 I i I I I I #5 I r I ' l 2 A 6 8 10 12 1A 16 18 20 22 2A Freeze-Thaw Cycles Graph 20.--Volume changes as a function of the number of freeze-thaw cycles. Water Absorption - cc/lOO gr. u 2 ------- Gravel 70 Legend: Sand—Gravel ———— Sand Sand-Gravel, 4% A.C. ,/’ ' /’//.’ 1 1 .———1:==-——:f::~""fiEGravel, 4% A.C. A T 1 , T r l 0 6 12 18 2M Freeze-Thaw Cycle Graph 21.-~Water absorption as a function of the number of freeze-thaw cycles. Degree of Saturation - % 71 Legend: Sand—Gravel -—----- Sand “9‘ ------ " Gravel l Sand uz A.C. ' 3o : /,. 20~ ’// S-G U% A.C. r % ’ 7% 15’ 10- - A .,a'1i4;" "" w”/ ’é/ 0‘7” O... 1 l 1 I I O 6 12 18 2M Freeze-Thaw Cycle Graph 22.--Degree of saturation as a function of the freeze-thaw cycles. Void Ratio - % 25- 2M- 23- 22.. 21-1 20+ 19- 18— 16- 15- O m- I...) [\J l—l CD M 4‘: Freeze-Thaw Cycle Graph 23.--Void ratio as a function of the freeze-thaw cycles (sand-gravel). f L4" ~ h MUm'm‘“.l‘fl~’l'tTM ; I I l ; 1 Void Ratio - % 314‘ 33- 32‘ 31- 30‘ 29- 28- 27. 26s 25. 2h— 23- 22~ 214 2o- 19- 18- 73 17 thaw cycles (sand). 0 6 12 18 24 Freeze-Thaw Cycle Graph 24.—-Void ratio as a function of the freeze— Void Ratio - % 29 28 27 26 25 24 23 22 21 2O 19 18 17 16 15 74 r* I 1 l I 0 6 12 18 24 Freeze-Thaw Cycle Graph 25.-—Void ratio as a function of the freeze— thaw cycles (gravel). APPENDIX III SIEVE ANALYSIS AND GRAIN SIZE DISTRIBUTION CURVES 75 TABLE 7 SIEVE ANALYSIS ON AGGREGATE SAMPLES Per Cent Passing Cumulative Sieve Gravel Sand S-G 3/4-inch 100 - 100 5/8-inch 98.5 100 99.25 1/2—inCh , 94.7 99.6 97.2 3/8—inch 91.3 98.5 94.8 No. 4 82.1 95.3 88.2 No. 10 72.5 90.5 81.1 NO. 40 44.8 58.3 51.2 No. 80 6.1 5.6 5.8 NO. 200 2.6 1.7 2.1 Specific Gravity 2.70 2.64 2.67 Void Ratio % 26.5 28.7 23.7 77 TABLE 8 SIEVE SIZE SEQUENCES FOR GRAIN SIZE ANALYSIS OF AGGREGATE Sieve OpenigghSize OpeninimSize Squififi Root 3/4-inch - 19.1 4.37 5/8-inch 0.625 15.9 3.99 1/2-inch 0.500 12.7 3.57 3/8-inch 0.375 9.52 3.09 No. 4 0.187 4.76 2.192 NO. 10 0.0787 2.00 1.42 No. 40 0.0165 0.42 0.65 No. 80 0.0070 0.177 0.42 No. 200 0.0029 0.074 0.273 Total per cent passing 78 100q Sand-Gravel 90* 80— 70- 60- 20- 10- l 1 2 3 4 Sieve Opening - /fifi Graph 26.--Grain size distribution curves. 79 .mo>h:o COHpanppmflU mmfim cfiwpoll.wm gamma as I mcflcmao m>mHm Hm>mmolocmm Ucwm dH dm .3 .3 .3 T8 .3. as do abOH Butssed queo Jed redo; APPENDIX IV TEMPERATURE PENETRATION CURVES 80 l 8 mHQEMm OB mCOpEmmeoot .mQOHpooccoo maddoooELmsu n0% pfisohwo Hmofippomamil.a onswfim JaqemOUOAIBQ pmmeQHpCmpom fnmdaoo Q Q-+._i IO- :54“.— C’) o o o no +0 OHmII. +ao humupwn pao> m m ::.: um +o -m om om 0 Im +.<. 1 N + as. om om w HHmo 4 .USm 0 U.. pmm mOH ow c. Looms 8 2 TABLE 9 TEMPERATURE PENETRATION Freeze Thaw Time Temp. Time Temp. Time Temp. Hrs. Min. °C Hrs. Min. °F Hrs. Min. °C 0 0 22 1 25 - 8.9 0 0 -13.2 10 18 30 — 9.6 5 -12.4 15 13.6 35 -10 10 8.3 20 12.5 40 10.5 15 7.4 25 10.2 45 -11 20 2.9 30 9.3 50 -11.7 25 1.3 35 7.4 55 -12 27 000 40 4.5 2 00 -13 30 5.4 45 3.1 35 +13.6 50 1.4 40 15 55 0 45 16.6 1 O -2.8 50 18 10 -703 55 2O 15 -8.0 60 22 20 -8.3 Sample: Sand-gravel with 4% Asphalt Content (A.C. 85/100) Height: 2 l/2—inch Diameter: 4-inch Thermocouple at 1 1/4-inch from the top Test Temperature: Room Temperature: -20°F 22°C .mpm OO.m 83 opzcfle I mEfiB om.H 0:.H Om.H om.H OH.H om . V . .mEHB .m> magpmhmqsmell.m mmsmwm om o: . I- MH\\|. mmomnm \V BMQB ;K om om OH o .$H .H Hm eannnJedwem APPENDIX V MATHEMATICAL RELATIONSHIPS USED IN CALCULATIONS 84 85 Mathematical Relationships Used in Calculations The following is a list of the mathematical relation- ships used for the calculation of parameters for each sample. 1. Bulk Specific Gravity - D b D = wa b Wa - WW wa D =— b Vb 2. Air voids in the compacted mix - Vv’ in % 3. Specific gravity of the aggregate - Gag Gag = 100 3213. G2 G3 4. Volume of aggregate as per cent of the total volume of the sample - % Vag, in % Pag x Wa 86 5. Voids in the mineral aggregate - in % VMA 100 - % Vag 6. Voids filled with bitumen - in % VMA — V ._______Y VFB = VMA 7. Uniformity coefficient of the aggregate - VC < (a u UIU F’(h c: o 8. Effective size of aggregate - ES, in millimeter E8 = Dlo Note: Items 1 through 6 were taken from reference (25). Items 7 and 8 from reference (38). 87 Symbols P1’ P2, P3 - Per cent of the asphalt cement, sand, gravel respectively, in the mix. G1, G2, G3 - Specific gravity of the asphalt, sand, and and gravel, respectively, in the mix. Wa — Sample weight in air, in grams. WW - Sample weight in water, in grams. Vb - Total volume of the sample, in cubic centimeters. Vag - Volume of the total aggregate in cubic centimeters. Pag - Per cent aggregate in the mix. Gag - Specific of the aggregate in the mix. Dlo - Maximum diameter of the smallest 10 per cent of the aggregate. D60 - Maximum diameter of the smallest 60 per cent of the aggregate. 10. BIBLIOGRAPHY American Society for Testing Materials. "Tentative Method of Test for Resistance to Plastic Flow of Bituminous Mixtures by Means of the Marshall Appa- ratus," ASTM Standards. Part IV. 1958. . "Standard Method of Test for Specific Gravity of Road Oils, Road Tars, Asphalt Cements, and Soft Tar Pitches," ASTM Standards. Part IV. 1958. . "Standard Method of Test for Specific Gravity of Compressed Bituminous Mixtures," ASTM Standards. Part IV. 1958. .________ "Standard Method of Test for Sieve Analysis of Fine and Coarse Aggregates," ASTM Standards. Part IV. 1958. American Road Builder's Association. Stabilization of Soil with Asphalt. Technical Bulletin No. 200. American Road Builder's Association, 1953. Asphalt Institute, The. Asphalt Base--Key to Better Pavement Performance. Information Series No. 118, June, 1962. ' . Mix Design Methods for Hot-Mix Asphalt Paving. Manual Series No. 2, April, 1965. Baskin, Charles M., and McLeod, Norman W. "Water- proofed Mechanical Stabilization," Proceedings, Association of Asphalt Paving Technologist, Vol. 12, 1940. Benson, J. R. "Bituminous Soil Stabilization," Roads and Streets, May, 1956. Casagrande, Arthor. "Discussion on Frost Heaving," Proceedings, Highway Research Board, Vol. 11, pp. 168-172, 1932. 88 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 89 Dillard, J. H., and Wittle, J. P. "An Examination of Mixing Times as Determined by the Ross Count Method," Bituminous Pavement.Permeability and Field .Compaction Studies on Asphaltic Concrete. Bulletin No. 358. Highway Research Board, 1962. Endersby, V. A. "Fundamental Research in Bituminous Soil Stabilization," Proceedings, Highway Research Board, Vol. 22, 1942. Fluss, P. "German Practice on Flexible Hot-Mixed ASphalt Bases," Roads and Streets, Vol. 101, No. "9, September, 1958. Herrin, Moreland. "Bituminous Soil and Aggregate Stabilization," Highway Engineering Handbook. New .York: McGraw Hill Book Co., 1960. Hudson, S. B. "Practical Graduation Limits for Natural Bituminous Concrete," Roads.and Streets, Vol. 101, No. 5, 6, May and June, 1958. . "Natural Aggregate Bituminous Concrete, Sand Asphalt, and Intermediate Mixing," Miller—_ Warden Technical Paper,.No. 61, 1961. Hudson, S. B. and Warder, W. B. "Hot-Mixed Black Base Construction Using Mineral Aggregate," £22? ceedings, Association oszsphalt Paving Technologists, Vol. 30, 1961. Herrin, Moreland. "Drying Phase of Soil-Asphalt Construction," Asphalt Soil Spabilization.. Bulletin No. 204. Highway Research Board, 1946. Highway Research Board. "Soil Bituminous Roads," Current Road Problems, No. 12, 1946. . Soil Stabilization with Asphalt, Portland Cement, Lime and Chemicals. Bulletin 241. Highway Research Board, 1960. . Soil Temperature and Ground Freezing. Bulletin 71. Highway Research Board, 1953. . Asphalt Soil Stabilization. Bulletin 204. Highway Research Board, 1958. Jumikis, Alfreds R. The Frost_Penetration§rob1em in Highway Engineering. New Brunswick: Rutgers University Press, 1955. 2“. 25. 26. 27. 28. 29. 30. 31. 320 33. 34. 35. 36. 90 Johnson, A. W. "Frost Action in Roads and Airfields, A Review of the Literature," Special Report. No. 1. Highway Research.Board, 1952. Lamb, T. W. Soil.Testing.for Engineers. New York: John Wiley and Sons, 1960. McKesson, G. L. "Research in Soil Stabilization With Emulsified Asphalt," Proceedings, American Sociepy ..of.Testinngateria1, Vol. 39. 1939. Powers, T. C. "A Discussion of C. A. G. Weymoth's Theory of Particle Interference." Papers of the. Portland Cement Association, Chicago, Ill., March, 1936. "Basic.Consideration Pertaining to Freezing and Thawinngests," Proceedings, American Society for Testing Materials, Vol. 55. 1955. . "A Working Hypothesis for Further Studies of Frost Resistance of Concrete," Proceedings, American Concrete Institute, Vol. 41, pp. 245- 272, 1945. Riley, J. Charles. "Asphalt Stabilization of Selected ,Sand and Gravel Base Courses." Unpublished Master'3~ thesis, Michigan State University, 1964. Rader, L. F. "Investigation.of the Physical Properties of Asphaltic Mixtures at Low Temperatures," Proceed- ..ings, American:Soeiety for Testinngaterials, Vol. -35: Part II: p. 559: 19350 Tabor, S. "Freezing and Thawing of Soils as Factors in Destruction of Road.Pavements," Public Roads, Vol. 2, pp. 113_132, 1930. . "Surface Heaving Caused by Segregation of Water Forming Ice Crystals," Letters of Engineering .News Record, Vol. 81, pp. 683-684, 1918. . . "Discussion on Frost Heaving," Proceedin s: ..Highway Research Board, Vol. 11, pp. 173-177, 1932. . . "Mechanics of Frost Heaving," Journal of ‘FEEEIogy, Vol. 38, pp. 303-317, 1930. Watson, W. "Change of Volume During Fusion and Latent Heat of Fusion," A Textbook in Physics, pp. 236- .239, 1911. 37. 38. 39. 40. Lil. 42. 91 Winn, H. E. "Frost Action in Stabilized Soil Mixtures," Proceedings, Highway Research Board, Vol. 18, Part I, 1938. Woods, K. B. (Editor) Highway Engineerinngandbook. New York: McGraw-Hill Book Co., 1960. Wooltorton, F. L. D. The Scientific Basis of Road Design. London: Edward Arnold and Co., 1954. Yoder, E. J. Principles of Pavement Desigp. New York: John Wiley and Sons, Inc., 1959. . "Principles of Soil Stabilization," Pro- ceedings of the 43rd Annual Purdue Road School, Series No. 92, 1959. Zoepf, W. T. "Notes on German Practice in Natural Aggregate Bituminous Concrete," Roads and Streets, Vol. 102, No. 2, February, 1959. "'TITt'I‘fii’HLflfi [17” [1 1771111? {1111 1111'“