The Laboratory Speaks

The Laboratory Speaks ARIZONA SUBGRADE STUDIES AND THEIR RELATION TO THE OILED ROADBy J. W. POWERS Engineer of Materials

PERUSAL of most of the literature on the subject of pavements will disclose that failures are common. This same literature sets forth various reasons for such failures. Even now the use of cotton fabric is being advocated to cut failures, so that recognition is still given to the fact that they are likely to occur. Our attitude is that we should study our projects so as to eliminate as far as possible any failures. For convenience in our study of the problem, we have separated failures into two general classifications: 1, Failure of the oil cake itself, and 2, failure of the subgrade.

Failure of the cake itself is not under consideration, but I might say in passing that they are rare indeed on an adequate base. In 1928, when the Arizona State Highway Department began the use of oil, very little heed was given to the base course or subgrade that underlay the oil mix. Subsequent developments showed that the supporting value of the soil on which the oil cake was to be placed was possibly of more importance than the grading and mixture of the cake itself.

At the time of our first oil construction, the most prominent method of determining the adequacy of a soil was the method outlined by A. C. Rose. His findings in the Pacific Northwest, and published in Public Roads in 1924 ("Public Roads," August, 1924), gave limits at which soils were presumably safe to use. Based upon his work, we began to examine and select our base materials using his methods, but decreasing the limits, especially the lineal shrinkage. Mr. Rose, in his recommendations, set 5 per cent as the maximum limit that would be allowable; we decreased this to 3 per cent, but still there were failures. Obviously it was necessary to delve further into the matter to determine the reason for these failures. When in 1931 the Bureau of Public Roads released their information on soil constants, we immediately took them up to see if they would throw any additional light on the soils we were using. While most of the materials found in Arizona do not have the grading of the soils on which the bureau constants are based, they do contain appreciable percentages of these gradings, mixed with sand and gravel.

Relying on these constants to give us further information, we made a survey of the oil roads built to that time. The resulting information was very enlightening. Among the things that were found was that even though the shrinkage values as made by the Rose method were within the limits we had set, the portion of the soil used in the determination of the soil constants might have double, or even three times the shrinkage value as determined by the B. P. R. method. The reason for this was that the grading of the material used in the Rose determination might be such that the type of binder did not show its true value. Particularly was this true of the high shrinkage plastic type clays.

We have been following the soil constant method of soil classification, as I stated, since 1931; and it is now our opinion, based on these tests and field observation, that the following limits will insure in most cases a material having a satisfactory value.

Centrifuge moisture equivalent, less than 18.

Plastic index, less than 10.

Lineal shrinkage, less than 4 per cent.

Even back in 1924, when Mr. Rose published his work, he recognized the necessity of maintaining a low moisture content in the soil, for he says "Closely related to the size and character of the particles composing the soil, is its ability to hold water against the force of gravity, i. e., its capillarity or water-holding capacity, and this relationship is so definite, and the determination of the moisture-holding properties is so much simpler than mechanical analysis, that it is perferred to the latter by the majority of investigators as a means of classifying soils."

The limits I have just mentioned are our findings at which the water-holding capacity will remain at the minimum, thus insuring the maximum supporting value.

Nothing thus far has been said about grading. It is preferable that the grading be uniform without too high a percentage of any one size predominating. I do not believe it advisable to be specific in connection with the grading requirements, as a comparatively large range of grading seems to be satisfactory. However, although I do not agree that the grading need be so rigid, I should like to give you a proposed grading to be used on future base course applications by the Bureau of Public Roads on Forest Projects. It is as follows: Passing 1 inch sieve, 100 per cent.

Passing 3-4 inch sieve, 85 to 100 per cent.

Passing 4 inch sieve, 55 to 85 per cent.

Passing 10 inch sieve, 40 to 65 per cent.

Passing 40 inch sieve, 24 to 50 per cent.

Passing 200 inch sieve, less than 25 per cent.

In addition they have stipulated that the plastic index should not exceed 6, and the liquid limit should not exceed 25.

You will note that the latter part of this specification does not follow our findings. We cannot agree that the liquid limit has a bearing on all soils. One of the best soil types which we have is caliche and it will, except in a very few instances, have a liquid limit well above 25. This same statement is true of cinders, yet they also will make a satisfactory base.

I have not made any attempt thus far to separate materials which may be imported, borrowed from the side or applied as a base course, because in our opinion, they can all serve equally well if they fulfill the above requirements.

So far, I have treated only of the satisfactory soil type. Of course we encountered the poorer types. When such poorer types, as evidenced by soil constants in excess of those previously mentioned, are encountered, we have applied varying layers of carefully selected material over them. In some cases such layers have served satisfactorily, and in other cases, not. From the success or failure, as the case may be, we have added further to our limits, and now believe that the following are the maximum soil constants on which intermediate layers will serve to make them stable enough to support an oil road. These values are as follows: Centrifuge moisture equivalent, 18-30.

Plastic index, 10-20.

Lineal shrinkage, 4-8.

Such soils usually can be made to carry an oil cake with an application of from 8 inches to 12 inches of good material. When soils are encountered which exceed these limits, it is advisable to import satisfactory material and waste the excavation, if there is any, or dispose of it possibly as a sodding material, along the roadway slopes. In other wordsit should not be used for embankment.

ARIZONA HIGHWAYS AUGUST, 1936

The procedure now standard practice with the Arizona Highway Department is to put test pits along the centerline at stated intervals and in all cuts to grade. After analysis, recommendations are made as to the disposition of cut material and whether imported borrow should be used. This outline, while rather sketchy, still is broad enough without going into details, to lay the ground work for any work you may wish to do along these lines. You may have been led to believe that this procedure is the answer to all the soil problems-but such is not the case. Even though it is generally true, still there is a difference in the supporting value of soils that come within the limits I have previously given. Considerable discussion in our laboratory has evolved around this recognized fact, and it has been only recently that we have been able to apply what seemed to indicate a satisfactory method to try and solve the problem.

We believe that, of the soil types encountered in this state, caliche has the highest supporting value of any. Basing design of our apparatus upon the apparatus described by D. S. Berry in the Proceedings of the A. S. T. M. (Vol. 35, 1935), we have partially confirmed the previous statement. This apparatus which we have adapted to our purpose is that described in the latter part of the article which deals with the shear value of a soil. It is interesting to note that the force normal to the shear plane has a very definite effect on the resulting shear value. As the normal pressure is increased, so are the shear values increased. Mr. Berry in his article makes this statement in connection with this phenomenon "Internal resistance or stability of granular materials depends primarily on mechanical arrangement or interlocking of the individual particles of the mass which is generally affected by the degree of compaction and particle shape of angularity and to some extent by gradation."

All of Mr. Berry's work has been done with materials in the dry state. As our problem is entirely due to the presence of moisture in the material, we have used the apparatus to study the effect of varying amounts of moisture upon several different materials. To date, we have noticed two apparent facts: First, the shear value of caliche a good material---is approximately constant for all variations in moisture content.

Second, with increase in moisture content, the shear value of good material does not drop greatly until the moisture content is greater than the capillarity (represented probably by the centrifuge moisture equivalent) of the material. While with poor material the shear value drops to quite a low point while the moisture content is still below the capillarity of the material. To date all our work with the shear apparatus has been only on the passing 3-mesh material.

We hope to augment the information we have gained from use of the soil constants with information from this test. Since, however, it is in the embryo state, I cannot give any concise recommendations concerning it.

It may be of interest to you to know the soil types that have been encountered in this state. So far we have found all except A-8. The preponderance of the soils fall into five classes-A-2, A-3, A-4, A-5, A-7. A-6 is rather an exception, and encountered very infrequently. Such is also the case of soil A-1.

What relation does all this have with reference to the oil road? Its whole purpose is to eliminate failures caused by the use of soils having too great an affinity for moisture. Our investigations have shown that the preponderance of our failures are directly traceable to moisture, either in the cake itself or in the soil beneath.

All of you in the less populous western states recognize that the oil road, whether using road oil or cutback, is the only type road that you can build over long stretches for some time to come; hence, it is necessary to build that type of road for a long life.

Failures on the oil road cause high maintenance costs, are inconvenient to the traveling public, are not easy to repair satisfactorily, hence any method of evaluating the one item that contributes most to its failure is a step forward to insure the long life that is needed, and becomes a credit to the engineer who has supervision over the design and construction of such roads.

In addition to decreasing to a minimum the oil road failures, study of and making adequate provision for a good subgrade will allow the use of the most economical section of oil mix. We believe that a 2-inch thickness is ample where subgrade conditions are known to be satisfactory.

In concluding this paper, the thought I wish to leave with you is that the oil road can be built to have a long life if it is built on an adequate base and an adequate base can be acquired only by careful selection of material.

COTTON FABRIC

The material is very open and coarse, which is necessary to allow the infiltration of the preservative asphalt, which is built before the roadway surface is placed over it.

And now the purpose for which it is intended to serve: All of you who have ridden on oiled roads have observed, from time to time, that there are depressions and broken places, which you have probably blamed on the failure of the mix itself. This, however, is not the case, as the oil cake, or oil mat as it is sometimes called, depends absolutely upon the subgrade for its stability and support.

It is thought the use of this cotton fabric which I have described above would act as a reinforcing agent, similar to the reinforcing action that one gets in placing chicken-wire or wire mesh in thin sections of concrete.

The project or projects on which this material is to be used are purely experimental, to test out whether this theory is valid or not.

The construction of this material would be somewhat as follows: The roadway would be built as usual, with the necessary base course and necessary gravel surfacing, as though an oil mix were to be placed without the introduction of the cotton fabric. However, on those projects which are chosen to be used as experimental projects, the cotton fabric will be placed longitudinally on the road oil, to tack it to the subgrade and to act as a preservative to the fabric itself. Upon this fabric itself will be placed the oil mix, as done in the past.

Time alone will tell whether or not this particular method of construction will prove satisfactorily. In order that a fair comparison may be had as to the efficacy of this material, similar sections on the same roadway and in the same vicinity will be built, except that cotton fabric will not be placed immediately between the oil cake and the subgrade.