Highway pavements in this study were constructed in the mid-1980s as non-reinforced, dual-lane, roads ranging in thickness between 200 mm and 300 mm, with skewed joints reinforced with dowels. Deterioration was initially recognized by Iowa DOT staff as a darkening of joint regions, which occurred for some pavements as soon as four years after construction. Pavement condition ranges from severe damage to none, and there appeared to be no unequivocal materials or processing variables correlated with failure.
Numerous studies of cores from these pavements have produced a variety, and some conflicting, opinions as to the causes of the deterioration, and the significance of microstructural features. Jones [1] examined cores drilled from US 20 in Webster County and found low air contents in regions where cracking was occurring. He speculated that the paver vibration and supplemental vibration around joint regions might be responsible for the low air void content leaving the pavements vulnerable to freeze-thaw damage. Ouyang [2] examined Wisconsin and Iowa pavements and concluded that internal and external sulfate attack and freeze-thaw attack were the most probable causes for premature deterioration of the US 20 pavements. He also felt that improper construction methods resulted in a low air content in the surface layer of the pavements. He noted that deposits of ettringite in the entrained air voids may have reduced resistance to sulfate attack and to freeze-thaw damage, and that freeze-thaw cracking could have facilitated the transport of external sulfate ions into the pavements.
Pitt et al. [3] examined the role of sulfate impurities originating from deicing salts on the durability of Portland cement mortar and pavements. They speculated that deicing salt impurities may have contributed to early pavement failures previously attributed to D-cracking. They note that concrete incorporating aggregate that passed freeze-thaw tests can exhibit failure in the field. These failures were observed to be especially severe in joint areas, and may actually have been the result of sulfate attack, even though the crack patterns appeared similar to that of D-cracking. Specimens exposed to gypsum-containing brines appeared to deteriorate more rapidly compared to those exposed to pure water. They observed an increase in ettringite and Friedel’s salt after exposure to deicing salts. It was thought that these phases filled voids making the mortars more vulnerable to freeze-thaw cycling damage.
Stark [4] concluded that the expansion caused by alkali-silica reaction (ASR) was the cause of the pavement deterioration. He felt that the ASR reaction was associated with chert and shales in the coarse sand sizes and that the problem pavements incorporated high alkali cements and Class C fly ash. Similarly, Thalow [5] found alkali-silica reactivity involving chert in the sand fraction and concluded that it was the primary cause of deterioration. He considered the filling of the entrained air voids to reflect water saturation. Evidence of an overall paste expansion, possibly indicating a delayed ettringite formation, was not found. Kofoed [6] examined a set of cores and found alkali-silica reaction involving shale in the sand and a poor entrained air void system and concluded that both ASR-related expansion and freeze thaw cycling led to cracking. In contrast to these findings, Marks [7] tested properties of selected Iowa sands used in construction and noted that while many failed the ASTM P-214 test (now ASTM C 1260 [8]), that in the past they had a good performance record in pavements.
Marks and Dubberke [9] observed the filling of entrained air voids with ettringite and felt that the voids would serve as a center-of-pressure as ettringite precipitation progressed, resulting in cracking. While no measurements were performed, they considered the volume of ettringite to be excessive and, that sodium chloride, likely from road salts, may have accelerated the deterioration by causing the ettringite to expand and then dissolve. This is in contrast to observations of Stark [10] finding ettringite to be stable in sodium chloride solutions, and that monosulfate appears to be converted to ettringite under such conditions. Schlorholtz and Amenson [11] determined that two or more deterioration processes were acting simultaneously. They concluded that freeze-thaw damage was most probable for the cracks in the cores extracted from I-35 and US 20. Alkali-silica reaction products were observed in all cores but were considered secondary in importance relative to the freeze-thaw processes. In regions showing less severe deterioration, damage was attributed to mixing or plastic concrete problems. D-cracking and marginal air void systems were also thought to be potential causes of deterioration.
Gress [12] noted a relationship between tine groove depth and distress. Regions with shallow tine depths and apparently poor finishability, generally denoted regions where distress was more likely. He considered these an indication of possible setting problems and that the loss of entrained air in these regions may be due to over vibration. Cracking was observed within the matrix with some cracks in the coarse aggregates and small entrained air voids were filled with what is thought to be ettringite. Empty cracks intersecting these filled voids led Gress to conclude that growth of ettringite in the voids created the pressure that cracked the concrete. Additionally, he felt these concretes contained "excessive ettringite", though no measurements were made.
Ouyang and Lane [13] examined freeze-thaw-cycled laboratory specimens prepared with and without Class C fly ash. They found a relationship between filling of entrained air voids and loss of freeze-thaw durability. They also felt that the degree of filling coincided with the length of moist curing prior to cycling, as well as the usage of Class C fly ash. This led to a subsequent study by Ouyang and Lane [14] where they devised an experimental program to examine the role of ettringite filling of entrained air voids and freeze-thaw durability in which a correlation between filling and freeze thaw performance was found. They felt that the filling was influenced by the period of moist curing and wet-dry test cycling, and by the materials used in the concretes. They proposed limits for materials in an effort to minimize the formation of ettringite.
A study by Northwestern University [15] of pavements from the Midwest, including those from Iowa, concluded that damage resulted from of multiple factors. They recommended investigation of the role of fly ash in pavements, avoidance of high alkali and high sulfate cements, the use of a permeable base, and also stressed the need to account for synergistic effects of combinations of variables in future testing.