THE THEORY AND PRACTICE OF ASPHALT PAVING MIXTURES By A. W. Dow of Dow & Smith, Consulting Chemical and Paving Engineers, New York City Since the early days of asphalt paving many attempts have been made to devise methods by which the formula for a sheet asphalt mixture could be predetermined in the laboratory, but up to the present it has not been done and until certain fundamental principles relating to the interfacial surface tension between solids and liquids have been more thoroughly studied it will not be possible. I hope to give you in this paper a brief statement of what has so far been accomplished and discuss some of the fundamental principles that are still insufficiently understood. In the early nineties the laboratory of the Barber Asphalt Paving Company did considerable research work for the purpose of establishing a general formula that might be used as a guide for the manufacture of good sheet asphalt mixtures. At that time Mr. H. C. Bowen was chief chemist and the writer his principal assistant. From a study of the analyses of pavements that had done exceptionally well and of paving mixtures that worked satisfactorily Mr. Bowen suggested a formula for an ideal aggregate. The composition of this aggregate as published by me in my first reports as Inspector of Asphalt and Cements for the District of Columbia for the fiscal year ending June 30, 1894, is as follows: While the screens used in those days were not the same as are now in general use and the 200 mesh screen was unknown yet you can readily see that this is in accord with what is now considered a good paving aggregate, 10% passing the 80 mesh and a total of 23% passing the 100 and 200 mesh. At that time we did not feel justified in stating a definite percentage of bitumen or even giving a method by which it could be approximated other than by trial. We believed the surface area and voids in the aggregate were the governing factors and while we worked out a method for determining voids and a great deal of work was done on this, no conclusions were arrived at owing to the fact that actual practice did not appear to justify some of the theoretical deductions. For a long time we believed this was due to the inaccuracy of our method for determining the voids but it seemed impossible to devise a means by which the sand could be kept from segregating while being run into the void determining receptacle. For these determinations we first used a glass graduate but finally adopted the use of a heavy glass bottle with gradually sloping sides having a capacity of about 400cc. A bottle of this shape is about equivalent to the metal cone receptacles that have been in use for some years past. It has the advantage in being transparent so that you can readily see how badly the sand segregates. The method consisted of running the sand in a bottle and at the same time shaking and jarring it until the volume of the sand ceased to diminish. Possibly if recent experimenters on the void theory of sheet asphalt mixtures had used glass in place of iron vessels with which to make this determination they would have had less faith in their results and might have done some work on improving the method which is surely greatly needed. While this method for determining voids in the mineral aggregate has been in use up to the present the results have never been considered as anything more than an approximation and the writer has been suspicious for some time that the voids thus obtained were greater than those actually in the paving mixture and occupied by the asphalt cement. That this is most likely the case has been brought out by Messrs. Hubbard and Field in their paper read this year before the American Society for Testing Materials. They show that there is good reason to believe that in some cases at least the asphalt film between the sand grains is thinner than the air film surrounding these same sand grains in a dry condition. Void determinations on mineral aggregates and their relations to the design of sheet asphalt paving mixtures was most thoroughly investigated in the laboratory of the Barber Asphalt Paving Company under Mr. Richardson and the results published in his book "The Modern Asphalt Pavement." He showed that by adding filler in increasing amounts to a sand, that the voids diminished until the filler reached a certain percent (41.7%) and that after that the voids started to increase. Mr. W. S. Wilkinson of the Asphalt Block Industry also did considerable work on voids in mineral aggregates, but found that by mixing the limestone dust with the asphalt cement before it was added to the mixture that he could diminish the quantity of bitumen, but more will be said of this later. My first report as Inspector of Asphalt and Cements mentioned. above is also interesting because it is the first published discussion on the desirability of using graded sand and the necessity of having the voids between the larger grains filled with smaller grains thus producing a mixture with very small voids. In a paper read by me before the Washington meeting of the American Society for Municipal Improvements in 1898, I still further dis cuss sheet asphalt mixtures and point out that aggregates with small size voids made more stable mixtures, thus allowing the use of softer asphalt cement and in this paper it will also be noted that for the first time mention is made of the difference in the stability of mixtures due to the surface conditions of the sand grains, smooth grained sands making less stable mixtures than rough grained sands. In 1905, Mr. Richardson published the first edition of his book on "The Modern Asphalt Pavement," which is an account of the research work carried on by the laboratory of the Barber Asphalt Paving Company up to that date. The chapters on mineral aggregate and on surface mixtures are most instructive and should be carefully studied by every one who is at all ambitious to devise new types of mixture. Mr. Richardson for the first time points out one of the most important factors in the designing of sheet asphalt mixtures and strange to say the one that has received the least attention. This factor is the property of asphalt cements to coat or wet some mineral surface better than others. He cites the case of two sands of practically the same mesh composition and percentage of voids where one required 2% more asphalt cement than the other to make equally rich mixtures. In his chapter on surface mixtures he says: "It has become very evident from what has been said in the preceding pages that the amount of bitumen or asphalt cement in any mixture is very variable, depending on the grading of the mineral aggregate and upon the peculiar surface of the sand grains. . . . The proper amount of bitumen may vary from 9 to over 14%.' On page 349, he calls attention to some determinations made by the Bureau of Soils on the thickness of absorbed aqueous films on different minerals where they found the following values: Silica 167.00x10-6cm. He believes that this data makes it possible to conceive of a sand made up of particles which may vary as largely in thickness of the film. of asphalt cement required to coat them as do the minerals experimented with above. This may explain why one sand with the same grading and voids as another may require different percentages of asphalt cement. Anyone that has been at all observant must have noted on examining some surface mixtures under a microscope that some grains of sand are more heavily coated with asphalt cement than others. In fact same are coated so thinly that the coating is hardly visible. There are numerous examples that can be given to illustrate this selective action of solid surfaces with liquids. As for instance, copper prefers kerosene to water, quartz prefers water to kerosene, sulphide ore prefers oil to water while the gangue associated with it prefers water to oil. Upon these properties is based the well known. See Bancroft's Applied Colloidal Chemistry. flotation process for the concentration of sulphide ores. It is a known fact that the asphalt films on most sands are readily replaceable under certain conditions by water containing a small quantity of soap or soda, hence the disintegration of asphalt gutters in which soapy water is allowed to run. This is also the principle of Walczak and Rice's method for recovering the bitumen from mineral aggregates. In none of these cases is the asphalt dissolved. The separation depends upon the fact that the mineral grains have a greater absorptive property for the one liquid than for another and the asphalt film is replaced by the water containing either soap or soda. There has been a great deal of interesting work done by the Bureau of Soils on the adsorption of solids from solutions by different soils. In their bulletin No. 52 by Patten and Waggaman is a most interesting treatise on the work that has been accomplished along these lines. They mention the adsorption of different dyes by quartz powder and other mineral constituents of soil. I, myself, have done some work on the relative adsorptive value of different mineral powders for asphalt cement by shaking the powdered mineral in a 0.5% solution of asphalt cement in carbon disulphide and noting the degree of decolorization that takes place. Some clay soils adsorb as high as 98% of the asphalt cement from the solution; powdered quartz only about 15%; limestone with high percentages of lime about 40%; cement rock, composed of limestone and shale, about 60%. To show you the great difference in adsorptive powers of some minerals. I have here a dust that was sent us for examination recently and a clayey earth. I am not acquainted with the origin of the dust but on analysis it has a composition similar to Portland cement. This material adsorbs no asphalt cement from the solution while the clay soil adsorbs a large quantity as can be seen. On further examining this dust, it was found to be most remarkable in that after mixing with a hot asphalt cement it produced a dark brown powder that, judging from appearances, you would think had an insufficient quantity of asphalt cement and yet when this powder was put under pressure in a mold pure asphalt cement squeezed out of it. It also has other peculiar properties which I have not the time to mention here but which go to show that there is a tremendous difference between the surface energy of the mineral particles of this dust and the mineral particles of the clay soil. When the asphalt cement has been adsorbed by many clayey soils, it is so strongly held by them that it cannot be dissolved out with carbon disulphide. So as to illustrate to you the great difference in the adsorptive values of some sands for asphalt cement I have made up mixtures from two sands both with practically the same amount of voids. Figure 1 is a photograph of pat papers made up under the same conditions with these two sands, designated A and B. No. 1 is sand A with 12.5% bitumen; No. 2 is sand A with 10.5% bitumen; No. 3 is sand B with 12.5% bitumen. You can readily see the great difference in the character of these mixtures which must be due to the difference in surface adsorption of the mineral particles. One of my first experiments on surface conditions of sand grains carried on about 1904 was to show that the roughness of the sand grains was not necessarily a dominant factor in the thickness of the asphalt film it would hold. I took three sands, one from Sandusky Bay, Lake Erie, another from Nutley Hall on the Potomac River below Washington, and the other crushed pure quartz. These three sands were first washed thoroughly and all 200 mesh material screened from them. They were then separated into their several meshes by sifting and recombined so as to be of the same mesh composition. They were then separately mixed with 10% of asphalt cement and pat tests made of these mixtures in the usual manner. The Lake Erie sand gave a dark stain, crushed quartz a medium stain and the Nutley Hall sand a light stain. The crushed quartz had the highest per cent of voids and the roughest and most angular shaped grains and yet it took less asphalt to coat these grains than in the case of the Nutley Hall sand which was comparatively round. I could go on and mention numerous other experiments which show the selective adsorption of different sands grains for asphalt cement but I think I have said sufficient to convince you that it is one of the most important factors influencing the design of asphalt mixtures and yet we know little or nothing about it. I know of no investigations having been carried on on this subject. Mr. Richardson over twenty years ago showed by his writings that this factor had an important influence on the designing of asphalt mix |