There has been much discussion recently about mixture tenderness with some Superpave mixtures during field compaction. These mixtures appeared tender during placement even though they conformed to requirements, meeting Ninitial and not passing through the restricted zone. The tenderness is normally noticed as the rollers move on the hot mat in the mid range of cooling temperatures, 240¡F to 190¡F.
While some projects using Superpave mixtures with Performance Graded (PG) binders like PG 64-28, PG 64-34 and PG 58-40 have exhibited this problem, other projects with the same binder grades used, for example by the Utah and Colorado DOTs, have been successfully placed and are performing well. Why the difference?
Field compaction is one of the most important processes in asphalt pavement construction. Achieving proper compaction density is critical to the ultimate performance of the pavement. According to the Asphalt Handbook, MS-4, published by the Asphalt Institute, dense-graded mixes with good in-place air voids (6-8%) reduce the detrimental effects of air and water, especially raveling and stripping. Experience has shown that dense graded mixes with in-place air voids greater than 8% are likely to disintegrate and fatigue more quickly.
There are five basic factors influencing HMA field compaction: 1) physical properties of the HMA component materials; 2) mat thickness; 3) environmental conditions at the time of compaction; 4) support of underlying materials including the subbase; and 5) type of compaction equipment and roller pattern used during construction. All of these factors are critical to the compactive effort. While the coarse- graded Superpave mixtures and the mat thicknesses may be even more important to compaction, this discussion will focus only on the physical components: asphalt binder, aggregates and moisture.
Most of our asphalt pavements have been constructed using a limited range of viscosity-graded asphalts, primarily AC-10, AC-20 and AC-30. Years of field experience with these asphalts have led us to believe that compaction needs to be achieved before the mix cools to 175¡F. At this temperature, the mixture's internal friction and cohesion increases sufficiently so further compaction produces minimal additional density. Why, then, do the PG asphalt binders exhibit the mixture tenderness when viscosity-graded asphalts did not?
PG asphalt binders
Performance-graded asphalt binders may or may not be equivalent to viscosity-graded asphalts. Where there is not a great difference between the high and low PG temperatures, straight-run asphalts meet the specifications, and will exhibit the same behavior as their equivalent viscosity and penetration- graded asphalts. For the more severe grades modification is necessary.
Modifiers can significantly change the physical properties of the binder. PG binders are formulated to meet the specified high, intermediate and low temperature properties for the project's environmental and traffic conditions. Viscosity-graded asphalts are specified primarily by their viscosity at high temperature. Two different AC-20s may have the same viscosity at 140¡F, but very different low-temperature properties. AC-20s from different sources currently on the market have been graded as PG 64-16, PG 64-22 and PG 64-28.
Typically, a grade with a wider temperature range, such as PG 64-34, is formulated by first choosing a base asphalt (such as an AC 2.5) conforming to the low-temperature stiffness and m-value requirements, as tested using the bending beam rheometer specified in AASHTO's MP-1.
In practice, the m-value generally needs to be higher than the minimum 0.300 to allow for variations in testing. Although there are wide variations in crude sources, normally a straight-run AC-2.5 or a premium AC-5 meets the -34 requirements. A conventional unmodified AC-2.5 should meet PG 46-34, and very high quality AC-5s may meet PG 52-34.
Once the base is selected, it is then modified to increase its temperature range and stiffen the high-temperature properties sufficiently to meet the specification as measured by the dynamic oscillatory shear rheometer. To do this, manufacturers may choose one of several methods, including polymer modification, air oxidation and chemical modification. In our example, the PG 46-34 can become a PG 64-34 with modification.
In conventional mixtures, the optimal compaction temperature is determined by the asphalt binder's high temperature viscosity. It has been generally assumed that the same holds true for Superpave mixtures; that is, that the compaction temperature should be determined by the binder's stiffness at high temperature.
Because of the differing rheologies imparted by different types of modification, this may or may not be a valid assumption. With some materials it may be the viscosity of the base asphalt, not the modified material, that correlates best with compactibility. In our example of the PG 46-34 modified to meet the PG 64-34, compaction temperatures typical of an AC-2.5 may be more applicable than temperatures for an AC-20. A straight-run PG 64 is normally an AC-20.
Obviously, for conventional materials, the compaction temperature for the AC 2.5 would be significantly lower than that for an AC-20. It is suggested that the purchaser follow the manufacturer's recommendations for mixing and compaction temperatures when using modified asphalts.
Some Indiana Superpave mixtures (a modified Superpave PG asphalt cement in a mixture designed using the Superpave Gyratory Compactor, SGC) have a 0.2% higher asphalt content than a conventional mixture (viscosity- graded asphalt in a Marshall design).
Total fluids is defined as moisture content plus total liquid binder minus the binder absorbed by the aggregate. The total fluids content is key to providing the lubrication necessary to achieve adequate field density.
The fluids content is further complicated with the presence of poor quality mineral fillers or plant returned fines (#200 and #325 sieve material) that may act as binder extenders. Superpave implementation is resulting in several unexpected consequences.
According to design criteria, Superpave mixtures can accommodate unwashed aggregates and mineral fillers, provided the Sand Equivalency, the Fine Aggregate Angularity and the Dust/Asphalt Ratios are within specified limits. Economic forces are driving an increased use of these less expensive material to fill the voids.
The result is a lower Voids in Mineral Aggregate (VMA) and ultimately a lower percent asphalt content at the 4% air voids required at Ndesign. Under the right conditions, poor quality mineral fillers or plant-returned fines filling the available VMA extend the binder, and therefore the apparent total fluids content.
Obviously the moisture also is key in total fluids content. Inadequate drying of the aggregate through the hot-mix plant can cause a fluid imbalance that may aggravate the tenderness problem. Today's drum mix plants often have rated capacities of 400 tons per hour and higher. Increasing production increases profitability. Unfortunately, heating wet stockpiled aggregates at high production rates can cause vapor pressure to rise inside the mixing drum. Until that free moisture is converted to steam and expelled through the exhaust stack, the aggregate materials will not heat above 212¡F, the boiling point of water. The air intake of the burner and dwell times developed during high production periods are often insufficient to overcome the pressure and adequately dry the aggregates. Free moisture may still be present in the pores of the aggregate, while the aggregate surface is flash dried. The PG asphalt binder is then introduced onto a saturated surface dry aggregate particle, and the process of releasing the trapped moisture begins.
A small percentage of free moisture may enable the asphalt binder to more quickly coat the aggregates. However, it is generally accepted that moisture contents be no more than 0.5% by weight of mixture upon discharge. If too much moisture remains in the mixture after construction, the resulting pavement may be very susceptible to stripping or other forms of moisture damage.
Thicker binder films, while beneficial for mixture durability, can hinder the release of moisture trapped in the pores of the aggregate. In this scenario, the mixture may not exhibit the classic signs of excessive moisture, such as slumping or foaming in the haul trucks, or streaks in the mat directly behind the paver. The goal of producing any HMA mixture should be to remove as much moisture as possible while providing good asphalt binder coatings on the aggregate.
Using conventional guidelines for mix and compaction temperatures based upon laboratory viscosity measurements generally results in recommendations that are well above optimum temperatures for field construction when using elastomer-modified asphalt. Additionally, the active total fluids content of the mixture may be higher. Is the tenderness, then a surprise? What can be done to correct the problem?
The PG asphalt binder viscosity optimal for achieving compaction may be influenced more by the base asphalt than the increased stiffness provided by the modification. The rheology imparted by the type of modification is the key.
If it is suspected that the compaction temperatures are contributing to the tenderness, lower temperatures should be considered. The PG asphalt binder supplier should be able to help with recommendations for mixing and compacting for the specific material being used.
The total fluids content plays an important role on the compactibility of Superpave mixtures, based on possible increased binder contents and the higher acceptable levels of fines. Too much available lubrication can make the mix appear unstable. Additional drying time through the mixing drum or adjustments to the percentages of mineral filler or plant returned fines may be necessary.
It is especially important to recognize the possibility of increased entrapment of moisture when using aggregates that are highly water absorptive. The quality of the mineral filler itself, including baghouse fines, might also be a factor. Rigden voids or other measures of filler quality can frequently identify potential problems, particularly for SMAs or other high filler mixtures.
The lift thickness, in addition to the asphalt binder and aggregate material interaction, plays an equally important role in achieving adequate field density. While beyond the scope of this discussion, it is accepted that the thicker the lift, the easier it becomes to achieve pavement density in the field. It is the author's opinion that reducing or maintaining the pavement layer thickness used in the past could be contributing to some of the problems encountered during construction of Superpave mixtures. Consideration should be given to increasing layer thickness to overcome this problem.
Superpave has made many improvements to the durability of asphalt cement concrete pavements. As with any new system, however, allowances must be made for the unexpected changes. The tenderness problem discussed here is only an occasional occurrence, and the problem is not insurmountable, as demonstrated by the many tons of Superpave mixtures already successfully placed throughout North America.
To capture the value of the SUPerior PERforming Asphalt PAVEments, it is important to recognize the differences from the conventional methods, and adapt laboratory and field practices as necessary to take full advantage of the improvements.