By Wendell Dubberke
Nearly 3000 ASTM C666-B durability tests have been run at the Iowa DOT since 1961. For the most part, the test was used to evaluate the coarse aggregate fraction of the concrete mix. The specimen size was 4X4X16 inches. Method B was used which means the PCC beams were frozen in dynamic air (to zero degrees F) and thawed in water (to 40 degrees F) during a 3 hour cycle. Normally 300 cycles constitute a complete test. The specimens were moist-room cured for 90 days (prior to testing) because very few specimens would fail with a 7 or 14 day cure even when faulty aggregates were used. Possibly the 90 day cure allows for some adverse chemical reactions to occur. Others theorize that the brittleness or pore sizes are affected by the 90 day cure. Research, to determine a cause, was not performed.
Before 1980, most of the early portland cement concrete pavement (PCCP) deterioration could be related to faulty coarse aggregates used in the PCC mixes. There were a few exceptions, like highway #25 (built in the mid 1960s) in southern Iowa where the whole pavement came unglued in 3 years. There never was an adequate explanation until the scanning electron microscope (SEM) revealed air voids, in the concrete, filled with ettringite, which was rare for Iowa PCCP that old. More information concerning ettringite can be found on other pages (Duggan test method and DEF).
The DOT has known since the 1940s that argillaceous (contains evenly distributed clay) carbonate coarse aggregates (calcites and dolomites) perform poorly when used in PCCP. Aggregates with clay contents over approximately 6% are not allowed to be used in PCCP because of poor performance when subjected to freeze/thaw cycles along with the use of sodium chloride deicing salts. Use of the ASTM C666 test method, with the 90 day cure, will properly classify these aggregates. However, there are many other test methods that can do it faster, cheaper and more accurately. Argillaceous aggregates can perform adequately in PCCP in southern states where deicing salts are not used and few freeze/thaw cycles occur. There is no magic-bullet aggregate test that will work for everyone. However, with the use of direct testing techniques (XRD, XRF, TGA and pore system analysis) correlations of aggregate (and other concrete components) properties with PCCP performance within a given location can be developed. Direct analyses of aggregates is the easy part. Developng a computerized, extensive PCCP (or AC, asphalt) service record history is the hard part. Fortunately, personnel in the Iowa DOT geology section began this task in the 1940's.
The Iowa DOT experienced other problems with the ASTM C666 test method. This test method usually will fail durable coarse aggregates that contain a minor amount (less than 3%) of expansive material such as tripolic chert or iron spal. Extensive PCCP service record information shows that aggregates containg less than 3 % chert and/or iron spal can be used in PCCP and perform well for over 30 years if surface aesthetics (popouts caused by tripolic chert or stains caused by iron spal) is discounted. To make the the ASTM C666 test method match service record information, personnel at the DOT remove chert and/or iron spal material from the aggregate sample before using it in the mix. By specification, aggregate samples containing more than 3% chert are not allowed to be used in PCCP mixes. The 3% limit on chert is arbitrary based upon the number of popouts that could be tolerated for aesthetic purposes. As far as early PCCP deterioration is concerned, the maximum amount of tripolic chert allowable is unknown. County PCCP using a gravel coarse aggregate, containing over 15% dense chert, has performed well for over 15 years.
Another problem with this test method is that it has a tendency to pass all dolomite coarse aggregates, regardless of their PCCP service records. In Iowa, where calcium choride is used as a starter deicer, pyritic dolomites and medium to fine grained dolomites are associated with a PCCP service records of less than 20 years. Pretreatment of the dolomite coarse aggregate, with NaCl, before incorporation into the specimen mix seemed to give results that matched field service records. Pretreatment of limestone (calcite) aggregates, with NaCl, did not work well.
The type of early PCCP deterioration encountered, when these reactive dolomite aggregates are used in the mix, will not be obvious when observing the surface of the pavement. Rotting, starting at the base near contraction joints, can be observed in cores taken from newer PCCP. As the rotting nears the surface, the concrete will break through, leaving a depression. There will be no surface cracks until it breaks through. This type of PCCP deterioration calls for a complete joint rehabilitation before overlaying with asphalt. Good joint sealing can delay but not stop this type of deterioration. With the current fashion of using 15 to 20 foot joint spacings, much money is being spent on rehabilitation. The use of longer joint spacings along with the use of better aggregates may be a viable alternative. Continuously reinforced concrete pavement (CRCP), made with this type of reactive dolomite aggregate, needs to be rubblized prior to being resurfaced with asphalt.
In southern states, where deicers are not used, the normal ASTM C666 test method would most likely classify these dolomites correctly. However, not many southern states use the ASTM C666 test method for obvious reasons. The effect of ocean mist brine on PCCP may be similar to the effects of deicers. The migration of magnesium ions (magnesium chloride and/or magnesium sulfate from the dedolomitization of dolomite aggregate aided by calcium chloride) into the concrete paste is thought to be the mechanism related to early PCCP deterioration. The generation of magnesium sulfate is restricted to pyritic dolomites. Wagon trains, traveling through the state of Iowa, stopped at certain dolomite outcrops to gather magnesium sulfate from the surface of the bedrock. A dry spell after a wet spell would produce enough magnesium sulfate on the surface of the bedrock (pyritic dolomite) to make it look like it had snowed. Magnesium sulfate is extremely soluble in water. The water in the pore system of the bedrock carries the magnesium sulfate to the surface. The magnesium sulfate is left behind when the water evaporates. The fine to medium grained dolomites usually contain an extensive capillary sized pore system that allows the brine to have access to a very large surface area within aggregate particles. It appears that extensive freeze/thaw damage can occur after the initial chemical reaction damage. Some in-house research, using fine grained calcites, indicate that freeze/thaw related damage can be reduced or eliminated if the prior chemical reactions do not occur.
One theory holds that the magnesium ions, in the paste, migrate to magnesium hydroxide sites, making them larger and consequently causing cracking of the paste. If this is true, then amorphous (glassy) blast furnace slag, containing magnesium, may help control this situation when used in a concrete mix, by furnishing an enormous mumber of extremely small nucleation sites to mop up stray magnesium ions before they can get to larger magnesium hydroxide sites that relate to MgO compounds in the cement used in mix. Blast furnace slag may or may not contain magnesium, depending on the type of carbonate used in their production. If the slag initially produces mobile magnesium ions rather than magnesium hydroxide, then the previous theory would not work.
Silica gel works in a similar fashion. Large, porous, reactive quartz (SiO2/chert) particles in concrete can generate enough silica gel (if the gel is viscous enough or if the pore system in the paste is insufficient) to cause cracking in the concrete. If these same large quartz (porous, reactive SiO2) particles are first pulverized, then they become a beneficial additive to a concrete mix. These extremely small reactive silica particles also generate silica gel, but on a scale too small to cause cracking of the pcc matrix. These relatively small silica gel nucleation sites are very efficient at mopping up alkalis (sodium and potassium ions). This is also the reason why proponents of the proposition that available alkalis in fly ash, when added to a concrete mix, can compound alkali related problems, may be wrong. Most likely, the silica gel generated by the highly reactive, amorphous, exremely small silica spheres in fly ash can not only mop the alkalis from the fly ash, but some of the alkalis from the cement as well. As far as chemical reactivity is concerned, most of the DOT, in-house research showed the addition of fly ash (up to 20%) to have an effect (on durability) ranging from beneficial to none. Economic considerations are another matter.
Initially, some PCCP investigators blamed the use of 15% fly ash (class C) replacement, in the PCCP mix, for some of the early Iowa PCCP deterioration. A short time later, other PCCPs (not containing fly ash) showed the same type of early deterioration. The scanning electron microscope (SEM) was used to verify the presence and amount of fly ash in the hardened concrete.
In Iowa, the correlation of ASTM C666-B test resuts to PCCP service records is not very good (0.42 r). Apparently the test model strays too far from what actually occurs with PCCP in the field. Rapid freeze/thaw cycles of saturated concrete prisms, from all six sides, traps the water in the concrete prism. This situation would rarely occur in field concrete. In field concrete, one or more sides are unfrozen as the freeze-front moves through the concrete. Water should be able to exit through the pore system, ahead of the freeze-front. It is also known that bulk ice will attract water through a capillary sized pore system. It is not clear if there is enough energy to crack the concrete or if the additional ice merely accumulates in the existing pore spaces. The same questions come to mind when discussing crystal (such as magnesium hydroxide) growth taking place in a porous medium (concrete).
The ASTM C666-B test method is very sensitive to the amount of air entrained in the concrete prism or beam. The number and distribution of air voids in the concrete matrix is most critical when fine-grained coarse aggregates containing an extensive, capillary sized pore system, are used in the fabrication of the concrete prisms. Conversely, some coarse-grained coarse-aggregates, used to fabricate concrete prisms, were able to pass the test even though no air entrainment was added to the mix. The grain and pore diameters of these large-grained carbonates was over 200 microns. The grain and pore diameters of the fine-grained carbonates (that require air entrainment to pass the ASTM C-666 test method) is under 20 microns.
The strange thing about all of the above, is that, in Iowa, there are many miles of PCCP, constructed prior to 1950, without air entrainment, using both coarse and fine-grained coarse-aggregate that have accumulated excellent PCCP service records over 40 years and in a few cases over 60 years. It is interesting to listen to the rationalizations put forth by old-line concrete investigators and petrogaphers in trying to justify thier absolute belief in the need for air entrainment in concrete to obtain freeze/thaw durability. What they have done, unfortunately, is place ASTM C-666 test results ahead of actual PCCP field results.
Years back, a professor from Canada, questoned the need for air entrainment after inspecting old concrete dams in Canada that were constructed without the use of air entrainment but were still in good condition after many years of use, even though they were located in a freeze/thaw environment and were in contact with water. He proposed a freeze/thaw test method where the prisms would freeze from the top and be insulated at the middle. He moved to another university before the research program could be implemented.
The use of air entrainment is optional or restricted in some European countries.
The use of air entrainment was initially envisioned as a cure for PCCP surface spalling where sodium chloride deicers were being used. Later, with ASTM C-666 test results being used as a club, the use of air entrainment, in PCCP, was deemed critical in areas where freeze/thaw cycles occur.
Air voids in concrete may be of great value, not necessarily as reservoirs for water/ice, but as reservoirs for silica gel and/or ettringite (gel?). When it was observed that air voids in concrete reduced surface spalling, many concrete investigators automatically assumed a freeze/thaw connection. An alternative explanation is that an alkali (sodium from deicing salt) overload caused excessive silica gel to form and the air voids became reservoirs for the excessive silica gel.