THE ESSENTIALS OF REVIEWING ASPHALT TEST REPORTS
By Danny Gierhart, P.E.
People in many positions, from agency/ owners of highways, to those in business having their parking lot paved, look to asphalt test reports for information regarding the quality of the material they have purchased. Sometimes the owner either performs or hires their own testing to ensure they are receiving a quality product (called Quality Assurance, or QA).
Sometimes the asphalt contractor performs their own testing to ensure they are providing a quality product (called Quality Control, or QC). Some test reports say something like “pass” or “fail” for a few basic tests. Others contain all of the raw data, calculations and allowable tolerances for a multitude of tests.
Some people are experts that can tell with a glance at a report exactly what a problem is, how to fix it, and how it will affect the asphalt’s long-term performance. Others may simply hope for passing test results, and be at a loss regarding what action to take, if any, when a report shows that a particular test fails to meet requirements.
If a test report shows that one or more sample test results fell outside the allowable tolerance, some essential questions should be asked. Are the results reasonable? What could this mean to the pavement longevity? How should the failure be addressed?
An imperative component of a successful quality program is that samples must be obtained, tested, reported and communicated as quickly as possible, preferably the same day or the next. It is not in the best interests of either the owner or contractor to receive a stack of test results to file away after the project is already over. Timely reports allow the plant operator to make any necessary changes to bring the material back into specification, thereby providing a high quality asphalt pavement.
The type of report that includes not only the results, but the raw weights and measurements that were used to calculate the results, is preferable to reports that either give a pass/ fail or simply report a number. If an owner has not been receiving the raw data, it should be requested to enable a complete evaluation of the test results. The tests specified most often as a quality requirement on loose asphalt mix are:
» combined aggregate gradation
» binder content (Pb)
» lab-molded air voids (Pa)
GRADATION
The purpose of requiring a gradation on field samples is to help ensure that the asphalt plant used the same aggregates as reported on the asphalt mix design, in the same proportions. The gradation of all of the aggregates mixed together, plus the design binder content is called the job mix formula (JMF). The JMF must have allowable tolerances on the gradation, usually on the order of ± 7 percent for aggregates sized 4.75 mm and larger, proportionally lowered down to about ± 2 percent for aggregates sized 0.075 mm.
Are the results reasonable?
Simple checks on the feasibility of the numbers include:
» Determine that the results are for the same size mix as the design. Results are most often reported as “% Passing.” In this case, the JMF typically (but not always) shows 100 percent passing the mix’s maximum aggregate size. If the field sample shows less than 100 percent passing a sieve that the JMF shows 100 percent passing, the field sample is coarser than the mix design. If the field sample shows 100 percent passing a sieve that the JMF shows less than 100 percent passing, the field sample is finer than the mix design. It is possible, but very rare, for a mix to fail on the maximum aggregate size. If it does, the reviewer should verify that the correct mix was sampled.
» If the results are based on percent passing (most are), the magnitude of each sieve must decrease as the sieve sizes get smaller. If the results are based on cumulative percent retained, the magnitude of each sieve must increase as the sieve sizes get smaller. The cumulative percent retained will increase, but the individual percent retained may go up and/or down from one sieve to the next.
» Determine whether the results fall within acceptable multi-laboratory precision when comparing QC and QA results from split samples. If the samples were taken at different times or places, using AASHTO multi-laboratory precision statements is not applicable. Most test procedures contain a table that contains both single-operator precision statements (the maximum amount results should vary when the same technician retests the same material) and multi-laboratory precision statements (the maximum amount results should vary when two different labs test the same material). These statements represent the difference between two test results that has only a five percent chance of being exceeded if the two tests were performed on material from the same population (or in this case, the same sample). The normal expectation would be that the difference between two results on a split sample would be considerably less than this maximum acceptable range. For gradations obtained from asphalt mix samples by ignition or extraction, AASHTO T 30 gives the following multi-laboratory precision, based on the percent passing range:
If the difference between test results from two different labs on a split sample exceeds the acceptable range, a trusted third party lab could be used as a referee to determine whether the QA or QC results should be used for acceptance and in determining any pay differentials.
What could this mean to pavement longevity?
Generally speaking, if the primary mix volumetrics, lab-molded air voids and VMA, are still within specifications, an out-of-spec gradation should not be too detrimental in the long run. However, there are some potential problems that volumetrics alone may not catch.
For example, if the upper sieves are either out of specification or running close to the lower limit (a.k.a. coarse side), the mix may be prone to segregation. Segregated mixes tend to not drain water effectively, allowing pathways for rain to enter the pavement structure and potentially cause stripping or raveling.
If the intermediate sieves (No. 4 to No. 8) are out on the coarse side, the mix may be inherently permeable. This could also lead to stripping or raveling of the mix over a period of years, and should be addressed.
If the percent passing the No. 200 sieve is too low (specifications typically allow a tolerance from JMF of about ± 2 percent), it may result in a high air void content, permeability, or field compaction problems. If the percent passing the No. 200 sieve is too high, it may result in a low air void content and VMA, lower binder film thickness, field compaction problems, or mix plasticity, with an increased chance of rutting.
How should the failure be addressed?
If the out-of-tolerance gradation is accompanied by out-of-specification mix volumetrics, another sample should be immediately obtained and tested to verify that the problem is inherent to the mix and not simply a result of poor sampling, improper material handling, or inadequate testing. If the problem is verified to be inherent to the mix, it needs to be addressed immediately. The most common sources of gradation problems include:
» One or more aggregate stockpiles at the plant are different than the material used for the mix design
» One or more of the cold feed bins at the plant are out of calibration
» Aggregate contamination from either intermingled stockpiles or improper feeding of the cold feed bin
» Failure to properly perform an aggregate correction for samples obtained through the use of an ignition oven. Aggregates will typically break down to some degree simply because of the extreme heat generated by the ignition oven, artificially increasing the amount of fines in the sample. The aggregate correction gives an average value for each sieve to subtract from the sample gradation, resulting in a better estimation of what the gradation was before the ignition process degraded the aggregate.
Keep in mind that there are a host of possible reasons for gradation problems, too many to be completely addressed in this article. The main points are to test regularly, report results promptly, quickly determine the reasons for the failing gradation results, and correct the problem without delay.
If the out-of-tolerance gradation is accompanied by in-specification mix volumetrics, many agencies simply adjust the JMF gradation to fit the field results. However, there is often a cap on the total number of JMF adjustments allowed before a new mix design is required.
BINDER CONTENT
The purpose of determining the binder content (Pb) of a field sample is to help ensure that the asphalt plant added the binder in the same proportion as reported on the asphalt mix design. The design (target) binder content is shown in the JMF. The design binder content mixed with the design aggregate structure should result in mix volumetrics that meet specifications. The JMF must have an allowable tolerance on the binder content, usually on the order of ± 0.3 percent to ± 0.4 percent.
Are the results reasonable?
Simple checks on the feasibility of the numbers include:
» Check the magnitude of Pb vs. that of Gmm. Gmm is the Theoretical Maximum Specific Gravity of an asphalt mixture, sometimes referred to as “Rice’s gravity.” The Gmm of a field sample is typically determined by AASHTO T 209 to calculate the asphalt mix volumetrics. Gmm can also be mathematically calculated by the following equation: Gmm – 100 / (Ps / Gse + Pb / Gb )
» where: Ps = the percentage of aggregate by total mass of the mix, or 100 Pb
Gse = effective specific gravity of the aggregate
Pb = the percentage of binder by total mass of the mix
Gb = the specific gravity of the binder
» The Gmm mathematically calculated using the sample Pb and the mix design Gse should be reasonably close to the Gmm determined on the sample using AASHTO T 209. However, the reviewer should keep in mind that Gse can change during production, so the calculation using the design Gse could begin to drift.
» As with the gradation, determine whether the results fall within acceptable multi-laboratory precision when comparing QC and QA results from split samples. For binder contents obtained from asphalt mix samples by ignition, AASHTO T 308 gives a multi-laboratory precision of 0.33. Remember that this value is not a typical difference between QC and QA lab results, but one that has only a five percent chance of being exceeded if the two tests were performed on material from the same sample.
What could this mean to pavement longevity?
The binder content of an asphalt mixture typically affects all other volumetric properties. Binder and air are what fill the voids created by the compacted aggregate skeleton in an asphalt mixture (VMA). Therefore, a higher than desired binder content means that some of the inter-aggregate space that was intended for air is now taken up by binder instead.
A binder content out of tolerance on the high side can lead to flushing or bleeding in the pavement because too little air is available to absorb binder movement. This can also cause the mix to experience plastic movement, exhibited by rutting and shoving of the asphalt. These issues, if they happen at all, tend to show up in the first couple of years of the pavement life.
A binder content out of tolerance on the low side can lead to premature aging of the pavement because too much air is present in the mix, allowing quicker oxidative stiffening of the binder. The excess air and prematurely stiffened binder can make pavement deficiencies such as stripping and raveling more likely to occur. These issues, if they happen at all, tend to show up in the later years of the pavement life.
How should the failure be addressed?
First, it should be determined whether an out-of tolerance binder test result is inherent to the mix and not simply a result of poor sampling, improper material handling or inadequate testing. The reviewer should be aware that sampling and reducing coarse mixtures to testing size has the potential to result in a segregated sample. A coarse split may show a low binder content because of the lower overall surface area of coarse aggregate in a given volume, even if the binder content of the mix as a whole is on target.
Conversely, a fine split may show a high binder content because of the higher overall surface area of fine aggregate in a given volume. If the problem is verified to be inherent to the mix, it needs to be addressed immediately. The most common sources of binder content problems include:
» Improperly calibrated binder pumps at the plant
» Inaccurate estimations of moisture content of the aggregate. If the estimation is too low, the aggregate weigh belt on a drum mix plant would interpret the greater weight due to moisture as more aggregate, triggering a higher volume of binder to be added, and vice versa.
» Failure to properly perform a binder correction for samples obtained through the use of an ignition oven. Typically some loss of mass during ignition can be attributed to aggregate loss instead of binder ignition, and must be accounted for. Note that the binder correction factor is exclusive both to the ignition oven used and the specific mix design. The asphalt binder correction factor from one mix design to another can vary greatly, up to about 2.0 percent, due to the different aggregate types being used in each design. However, the difference in correction factors from one oven to another should be relatively minor, typically no more than 0.2 percent.
Again, if the out-of-tolerance binder content is accompanied by in-specification mix volumetrics, many agencies simply adjust the JMF binder content to fit the field results.
LAB-MOLDED AIR VOIDS
Asphalt mix specimens are compacted in the lab using some type of equipment, most often a Superpave Gyratory Compactor (SGC) or a Marshall Hammer. The bulk specific gravity of these specimens, Gmb, is determined, and also the theoretical maximum specific gravity, Gmm. A simple calculation using these two values gives the percent lab-molded air voids, Pa:
Pa=100 x (1 – Gmb / Gmm )
Are the results reasonable?
Determining whether or not the results seem reasonable involves examining both Gmb and Gmm. Do they follow the correct trend? Is their magnitude reasonable considering Pb and Gse? Do the QC and QA results compare within reason? The components included in a Gmb determination are:
1. aggregate (which has the highest specific gravity, typically somewhere around 2.6),
2. binder (which has a lower specific gravity, usually somewhere around 1.02), and
3. air (which has a specific gravity of zero).
As the percentage of air voids and/or binder changes in an asphalt mix, the relative contribution of each component changes the total mix specific gravity. Because of this, Gmb should always increase as Pb increases, because binder is filling up space in the specimen that was once occupied by zero specific gravity air.
The components included in a Gmm determination are only aggregate and binder, and do not include air. Therefore, Gmm should always decrease as Pb increases, because that necessarily means that the percentage of higher-specific gravity aggregate decreases.
Gmm is one of the most repeatable asphalt mix tests, with a multi-laboratory precision of 0.024, which equates to a difference in lab-molded density in the range of 1.0 percent. Gmb can be quite a bit more problematic to match from one lab to another, with a SGC maximum allowable difference of 1.7 percent. AASHTO does not give a precision statement for Marshall specimens, but the author estimates it would be a slightly higher value than the SGC.
What could this mean to pavement longevity?
The air void content of an asphalt mixture is an integral component of all other volumetric properties. Historically, properly designed air voids and VMA are excellent indicators of long-term performance.
An air void content out of tolerance on the high side indicates that the mix will likely be difficult to compact in the field as well. Potential problems include low durability, causing raveling of the coated particles from the pavement surface. There is also potential for the mix to be permeable, allowing water to penetrate into the pavement. The weight and motion of vehicles could force water in and out of the high air voids, scouring and stripping the binder from the aggregate within the pavement structure.
An air void content out of tolerance on the low side indicates that the mix may be tender in the field. The main concern is that the mix may exhibit plastic movement and have a susceptibility to rutting and shoving. A low air void content may also predict a tendency toward low air voids in the field also. With limited space for movement, the asphalt binder can be squeezed to the surface, leading to flushing or bleeding.
How should the failure be addressed?
An air void content that is 1 percent or more from the target is reasonable cause to stop production immediately until the reason for the out-of-tolerance air void content is discovered and resolved. The most common sources of air void problems include:
» Aggregate changes during production that cause particles to fit together more tightly than the more gently handled aggregate in the lab.
» A high binder content that fills too much VMA space, causing high VFA and low air voids.
It is not uncommon for field mixes to be slightly altered from the mix design to increase production air voids and VMA. It is important that a trial batch be produced, tested and verified that it meets specifications before the mix is placed where it will be used in service.
In summary, it’s important for any person reviewing asphalt mix test results to understand why the specific tests are being run, how to tell if the submitted results are reasonable, what out-of tolerance results could mean for the longevity of the asphalt, and how to address failing test results. In order to respond properly, tests must be performed in a timely manner and the results distributed while there is still time to address any problems.
Gierhart is an Asphalt Institute Regional Engineer based in Oklahoma.