Fundamentals of Non-Load Associated Cracking

There has been considerable research on the engineering basis of pavement cracking. Those interested in some of the basic studies on cracking might consult volume 41 (1972) of the Proc. Association of Asphalt Paving Technologists. Many of the concepts develop there were the basis of the PG grading system with regard to low temperature properties of asphalt. While those papers are 40 years old, they lay the basis of technical progress in understanding cracking. Later studies have been oriented toward understand how cracking can be predicted.

However, it is not the purpose of this blog to go into the engineering of pavement design but rather to speak of the basic physics involved.

Failure occurs either from tensile stress or crack propagation. The maximum tensile strength of asphalt and hot mix is about 1000 psi and that only happens if the asphalt is cold or stressed at a high rate. At higher temperatures or lower rates of strain the stress at failure would be less. When cracks appear, the stresses are concentrated at the apex of the crack accelerating the formation of a crack. Thus no matter what the crack might look like, it is caused by too much tensile stress.

Literature suggests that when the temperature drops down below about 100-110°F of the softening point of the asphalt in pavement one would expect damage to the pavement. That damage accumulates eventually resulting in transverse cracks showing up. The distance between cracks is related to hardness of the asphalt. If the temperature rapidly drops to, perhaps, 150° F below the softening point, the crack may occur that day. I actually observed that in the late ‘80s. There had been a very sharp drop in temperature in Spokane, Washington on one day. I was called in for several cases where even fairly new pavements showed block (traverse) cracking, including a new tennis court. The only answer was that the temperature drop had caused it.

If we recognize that the softening point of aged asphalt might approach 200° F it can be seen that the fast drop in temperature in deserts at night could even cause damage at surprisingly higher temperatures. Pavements can reach over 170° F in the deserts.

The effect of crack propagation can be seen in parking lots where asphalt pavement is adjacent to a portland cement area where there are 90° corners. A crack will be seen radiating out of the corner even if there is no other evidence of cracking in the pavement. If small cracks are formed inside a pavement and don’t heal themselves, they will grow and eventually show up.

When cracking occurs, the asphalt in a pavement is no longer performing as a liquid, but more as a solid. It responding to stress from cooling by pulling apart horizontally. When the pavement heats up again, the crack remains, although if they are small, traffic can knead them back together. If it can act as a liquid it flows vertically upward as the temperature increases and downward as the temperature decreases. The solution to cracking is to allow the asphalt to retain its liquid properties as long as possible.

As asphalts from different crude sources behave differently, there is no golden rule. Non-electrolytic solution chemistry can be involved but that is a discussion for another day.

One of the remedies for reducing the temperature related cracking in pavements is to seal them so that the rate of hardening of the asphalt is reduced. Also the HMA needs to be protected from water, both liquid and vapors. Even in the desert water accumulates under the pavement. If the bond between the asphalt and aggregate is susceptible to being compromised by the presence of water, the bond will be broken and failure will occur. Traffic accelerates the loss of strength as water propelled by changing pore pressure scours the asphalt off of the pavement. Even water vapor has been seen to do this. Weakening the bond between rock and asphalt will then be allowed to grow under less stress.

I also like to see primes used under the pavement to discourage water from entering the mix. Reducing the rate of hardening of the asphalt so it retains its liquid properties and protecting the pavement from water damage can reduce the rate of formation of non-load associated cracks.

Robert L. Dunning,,



Expecting Binder Research to Solve the Problem

In a previous article “Fundamental Causes of Cracking, Potholes, Raveling, and Rutting in Asphalt Pavements” I touched on some of the causes of rutting. I wish to expand on this subject. However I wish to exclude rutting from studded tires as that problem has not been solved at this time.
Prior to the establishment of the Strategic Highway Research Program (SHRP) I attended a meeting in which it was stated that the goal was to develop an asphalt that could solve all of the problems that occur in pavements. The philosophy that problems reside primarily with the asphalt is still deeply encountered, however, in my opinion it is all “vanity and blowing into the wind”. That does not mean that there isn’t a place for asphalt research because great strides have been made in resolving pavement problems with modified asphalts. In fact, with rutting, I am sure many will state that with such and such binder, the rut tester shows an improvement in rut resistance. And I am sure that their data is correct. However I would suggest that those modifications only affect the rate at which rutting occurs not the basic cause. The misconception is that it is the properties of the asphalt that allows rutting. That is false. A properly performing asphalt is a liquid and is purposely designed to not resist rutting or any other stress that might prevent it from flowing. In fact one of the solutions to low temperature cracking is to modify the asphalt so that it can flow to relax thermal stresses before they reach the point where the asphalt fractures.

If one wishes a life time research project on rutting, concentrate only on the binder and work only with oversanded aggregate gradations. Do I mean that the aggregate gradation is part of the problem? Yes. In fact the aggregate gradation is the problem; and the present gradation specifications specifically allow oversanded mixes, thus, allowing the construction of tender and rut prone pavements to be built. I learned this from Vaughn Marker and Went Lovering of the Asphalt Institute back when I had more hair, and it was not so grey. (Went Lovering also had worked for CALTRANS and was a great source of knowledge and wisdom. He was instrumental in the development of the Hveem Design.)

How can we get oversanded mixes. First draw a straight line on the 0.45 power gradation chart from the % passing of first sieve that retains aggregate to that of the – #200. If you want an oversanded mix, make sure that the -# 4 gradation is above that line. If it is below the line, you can still meet that goal of an oversanded rut prone mix if the gradation in the -#30 range goes above that line. It is true that messing with the ability of the binder to flow will help reducing the rate of rutting, but, of course, non-load associated cracking is associated with the lack of ability for the binder to relax thermal stresses. In this manner research can be continually funded so that one can be an expert on how to not to stop rutting and tenderness.

You do want an oversanded mix for hydraulic structures, however.

Robert L. Dunning,


Spreadsheet for Comparing McLeod and Kearny Designs

Chip seals are widely used for maintenance of pavements. Two of the popular design methods are those developed by Dr. Norman McLeod and Mr. J. P. Kearny. I am finishing the development of a spread sheet that can be used to compare the two procedures, and that can be taken out in the field during construction so that changes can be made with regard to pavement conditions, traffic during construction, % oil in the binder or other changes that might occur.

The spreadsheet may also be used by agency engineers to establish budgets and for contractors in bidding projects.

At the present time the Kearny method on the spreadsheet does not include the modifications that Jon Epps et al. have made, but those will be added soon.

For more information contact Robert Dunning at,


Variables Involved

The introduction of the Superpave Mix Design for asphalt pavements was herald as one of the greatest improvements of the 20th century by some. The effort in developing the new design parameters was scattered among some of the most skilled asphalt technologists in the world.

I was involved in the SHRP program, and have experience in actually using the Marshall, Hveem and Superpave (Gyratory) mix designs for use by various agencies in doing so have reach the conclusion that perhaps we have not gained anything with the Superpave design and perhaps the benefits do not outweigh the deficiencies. In this blog I wish to outline the variables involved and expand on them in later blogs.

It should be mentioned that many people concentrate on the properties of the asphalt as being the determining factors defining performance. While the asphalt properties, especially the low temperature ones, are indeed important, the properties of the aggregate is vital, especially its gradation and its ability to resist the stripping of the asphalt off of the aggregate by liquid and gaseous water. A case can be made that the asphalt properties are over-defined for a well made pavement, but that is for other blogs.


Mold Size. One of the benefits for some of the mix designs with large sized aggregate with mixes on the coarse side is the 6” mold. Smaller molds can result in over-estimating the asphalt requirement caused by bridging of the aggregate. This can especially be true with the 4” Marshall Design; however with the use of the 6” Marshall design, this problem has been overcome.

Record of Continuous Compaction History. In problem solving the record of the compaction by gyrations allows one to obtain much better information of how things are compacting, and to predict how other variables will affect the compaction.


Too “Academicized”. It seems to me that if the procedure would have been turned over to state technicians to proof test, the procedure would have been better.

VMA Requirement too High. The VMA requirement was taken out of the Marshall design procedure. For a nominal ½” mix that is 14%. The old requirement for VMA from the FHWA Hveem specifications was 13%, and many successful pavements have been made in that range. As a result with certain mixes it is needlessly necessary to blow out fines just to meet the VMA requirements. (In my opinion neither the VMA or voids filled are the best specification parameters. I prefer effective asphalt and film thickness but would report the VMA for information purposes.)

Gradations Allow for Tender and Rut Prone Mixes. It amazes me that the effort to control rutting is to concentrate on the properties of the asphalt and neglect the true controlling factor: the gradation. The properties of the asphalt can affect the rate of rutting, however the gradation controls the extent of rutting. No one seemed to have listened to Dick Davis (Retired from Koppers) at meetings where he explained the importance of letting the aggregate carry the load. Both my son (Dr. Michael R. Dunning) and I know how to set the gradation to greatly reduce rutting (except from studded tires) and tenderness.

Lack of Understanding of the Gradation in the – #30 + #100 Range. Some mixes need a fine sand to be added as a filler, otherwise asphalt will have to be used to fill the void.

Lack of a Measurement of Strength. I frankly don’t understand leaving out a strength measurement. Adding an indirect tensile strength would have been easy.

Insufficient Emphasis on Water Damage. Many aggregates prefer to be wetted by water than by asphalt. Even though the rocks may be coated with asphalt, if there is any break in the film, water will get inside and lift the asphalt off. Even water vapor can cause damage and at one time was measured by the Moisture Vapor Susceptibility Test.  For protection, amines and lime has been used, however there are data that suggest that the protection may be transitory, the mechanism for which will be described in a later blog. There is a new product that overcomes that problem.

Another problem can be that the test criteria for measuring stripping are not severe enough.

The research done by the SHRP projects was very valuable, however some have found that the Marshall and Hveem designs have features not found with the more costly Gyratory designs. The purpose of this blog is to suggest areas of concern from one who has practical experience.

Robert L. Dunning,,