Sample Preparation for Direct Tension Testing: Improving Determination of Asphalt Binder Failure Stress and Test Repeatability

Abstract

During Superpave direct tension testing (DTT) of asphalt binders, the failure stress and failure strain values were found to depend strongly on the molecules of asphalt in the DTT mold forming a stable network. An asphalt specimen in which molecules have formed a network with aligned dipole moments has a higher DTT failure stress value. A sample preparation method based on this theory was developed. In this method, ceramic tiles holding the DTT molds were heated in an oven with controlled temperatures specific to the performance grade (PG) of the binder. These heated ceramic tiles provided a means for the poured asphalt to stay fluid and cool slowly and uniformly. The theory, detailed method development, and sample preparation procedure are described. With this method, the DTT failure stress of asphalt binder is higher than in samples prepared without controlled cooling. The variability of failure stress and failure strain values of the six specimens within the same run is usually less than 15 percent for the softer asphalt, for example, PG52–34, and somewhat higher for the harder asphalt, for example, PG64–28. The average results of the best four out of six specimens usually agreed with another run within 10 percent, regardless of asphalt type. Important considerations during preparation of DTT samples are discussed. The incorporation of these details for the sample preparation method in the DTT procedure will lead to better agreement of direct tension results among different laboratories.