This article summarizes key findings from a longer-term aging experiment in Mississippi that is a partnership of APAC-Mississippi, Ergon Asphalt & Emulsions, the Mississippi DOT and the Mississippi State University Construction Materials Research Center (CMRC).
Experiments began in November 2011 and are expected to continue until at least November 2021. Figure 1 provides several photographs of the outdoor aging site, which is located in Columbus, Miss. The overall goal of this partnership is to better understand how modern asphalt mixtures age in the southeastern U.S. climate and to develop laboratory conditioning methods and tools to reasonably replicate a few years of aging in this climate. To date, over 5,000 mixture specimens have been tested alongside hundreds of measurements on recovered asphalt binder; multiple mixtures, binder sources and binder grades have been included in the study.
Plans for this test section began well before 2011, and the site was originally used as part of an emergency paving demonstration for the Department of Homeland Security through Oak Ridge National Laboratory’s SERRI program. The purpose of originally building the full-scale test sections was to prove the concept that asphalt paving could be more effectively used in response to disasters such as hurricanes where power and infrastructure are often lost for a period of days to weeks over a widespread area. The original study utilized warm-mix technology (WMT) to show that haul distances easily exceeding six hours could facilitate producing asphalt far from the disaster where power and infrastructure were not damaged, and effectively using that material for paving the way into the disaster zone so that all other essential functions that make use of a functional path into the area can be less affected than in previous disasters. Once the emergency paving research and demonstration had successfully completed, the Figure 1 parking lot had been paved with 12 strips of asphalt with different WMTs, different haul times, and compacted to different air void levels.
Within days of completing the emergency paving demonstration, some participants began discussing the value of this as a longer-term aging site, especially since raw materials such as binder had been individually sampled during paving. These plans materialized quickly, and specimens began to be extracted from the parking lot (cores and slabs) for unaged assessments (November 2011). These unaged specimens have been evaluated relative to specimens aged over time at the test section, with most assessments being on specimens extracted from the section on a yearly recurring basis so they have received full yearly weather cycles (e.g., three years, four years; not 3.5 years or 4.6 years). Beginning in November 2012, gyratory-compacted specimens of a variety of additional mixtures began to be placed onto the test section in plastic sleeves.
Putting on the years
One of the major findings to date is the need to use combined effects damage mechanisms during laboratory conditioning to simulate longer-term aging in Mississippi. Table 1 shows several attempts to replicate the aging experienced in the field with laboratory protocols, and that the only manner that could consistently replicate longer lives was the one where oxidation (oven), moisture (64°C water) and expansion (freeze-thaw) mechanisms were present. Note that the first row of Table 1 is AASHTO R30, which is a current national standard method that did not simulate the duration of time claimed in the method (seven to 10 years).
Another major finding to date is that pressure-aging vessel (PAV) conditioning of binders via AASHTO R28 has used 20 hours of PAV time for several years under the premise that this would simulate five to 10 years of binder aging. Five combinations were evaluated in the field where corresponding raw binders were PAV-aged in the lab for up to 80 hours. The main finding was that only one of these five cases showed PAV conditioning for 20 hours actually simulating five years of field aging (less than five years was simulated in the remaining cases). When this finding is combined with Table 1, the overall assessment of the Columbus, Miss., experiment to date is that national laboratory conditioning protocols are not simulating as much field aging as one might expect at this location.
Future efforts are planned to look more into combined effects laboratory conditioning and actual field temperature/moisture conditions. Examples could include other combinations to complement Table 1 (perhaps more expedient combinations) that include oxidative, moisture and freezing mechanisms to replicate a few years of service in the Mississippi climate. Also, work is ongoing to improve understanding of the combined effects mechanisms actually occurring at the test section. Figure 1 shows a core drilled absent water (a moisture gradient is easily visible) and probes measuring temperature as a function of pavement depth in the laboratory compacted specimens and test strips. Plans are ongoing to measure temperature and moisture properties over an approximately one-year period and interpret this data to shed more light on improved manners to simulate combined effects in the laboratory. Data collected to date already has shown moisture contents in the parking lot test strips varying from 1.1% to 4.8%, and when combined with daily or seasonal temperature changes the impact of these values on time-dependent property changes is not fully understood.
Acknowledgements: This study would not be possible without the support of the aforementioned partners and direct assistance of the following individuals: Gaylon Baumgardner, Michael Bogue, Dwayne Boyd, Ben Cox, Codrin Daranga, Scott Glusenkamp, Bradley Hansen, Michael Hemsley, Trey Jordan, Alex Middleton, Drew Moore, Brent Payne, Carl Pittman and Braden Smith.