Freeze-thaw cracking of aggregate may be related to both the high degree of saturation and the loss of the adjacent air void system as no correlation between aggregate lithology and cracking was noted. Modeling simulations of deterioration processes indicate that cracking may be explained by both aggregate and paste expansions.
Wang et al. [43] found that ice did not completely fill the pores in cement paste and that supercooled water was always present due to alkalinity of the cement paste. Ice propagation in aggregates is faster than in cement paste and at -10 º C, ice was found in large pores but not small ones. Excess water in aggregates is expelled when concrete is frozen. Depending on size, permeability, and aggregate saturation, water flow from the aggregate may be too rapid to diffuse into the surrounding paste creating a pressure that may disrupt the paste at the paste / aggregate interface. Ice formation is from pure water so pore solution becomes increasingly concentrated. This creates an osmotic pressure that may be sufficient to result in failure.
The original air void system parameters generally appear to be uniform from top to base of the concrete. The excessive vibration, considered to be contributory [9], appears to not have adversely affected the entrained air void system for specimens taken away from the direct points of immersion probe vibrator insertion. The single vibrator trail specimen showed a substandard entrained air void system in the upper third of the core, and at the depth of the probe. Joint specimens exhibiting greater filling of the entrained air voids, which may reflect greater water availability to concrete adjacent to joints. The filling of the entrained air void system generally appears uniform with depth, although some mid-panel specimens appear to have increased filling with depth. These observations suggest that the degradation is not a "top down" process by infiltration of gypsum-contaminated road salt solutions. The increased filling at depth of some of the mid-panel specimens indicates a significant influence of surface drying and / or a source of water at the base of the slab. The low-permeability base may actually have facilitated deterioration by wicking water to the base of the pavement or by hindering drainage.
Entrained air affects plastic concrete by providing a potential for improved workability, improved resistance to segregation, settlement, and bleeding. The improvement of these properties is dependent upon the total air volume, the size distribution and dispersion of the voids, and the material properties of the concrete. However, the primary purpose of the entrained air void system is to facilitate frost resistance in hardened concrete. Air voids, being much larger than the capillary pores, remain open with water filling them only at sub-freezing temperatures, or when under pressure [44].
Early theories of frost damage in concrete were based upon the expansion of ice upon freezing and the subsequent stress increase [45]. Later, consideration that concretes contain enough air space to accommodate this expansion resulted in thoughts that stresses are produced by the flow of the water during freezing. The development of an entrained air void system for protection of concrete allowed room for excess water to migrate upon freezing. Finally, an osmotic pressure theory was developed based upon ice formation in voids leaving an alkali-concentrated supercooled water. Subsequent movement of this unfrozen water toward the air voids may create an osmotic pressure sufficient for concrete disintegration. Other factors in frost damage that must be considered are the availability of water, with the volume of freezable water (the degree of saturation) influenced by initial porosity, concrete age, curing conditions, and environmental exposure [44].
ASTM C 457 [46] provides a means to measure air-void size and size-distribution. Entrained air void bubble sizes range from a few micrometers to one millimeter. Larger-sized voids, while contributing to frost resistance, are considered to be entrapped air. The total paste volume protected by these large voids is much less than that for the smaller voids. The air volume is a measure of the total volume of the air bubbles in the hardened concrete. The spacing factor, considered to be the most significant measure of frost resistance, is defined as the theoretical maximum distance from any point in the paste to the edge of the nearest air void. ASTM C 457 recommends a spacing factor between 0.10 mm to 0.20 mm and notes that these values are for moderate exposures. Concretes in more severe environments, such as pavements that may be exposed to water upon freezing, may require smaller values than those stated here. Specific surface is comparable to "fineness" in cements, and is calculated by dividing the cumulative surface area of the voids by their cumulative volume. This variable is expressed as a surface area per unit volume with higher values indicating a finer size distribution. ASTM C 457 provides acceptable values for typical concrete being 24 mm2/mm 3 to 43 mm2 /mm3.
Freezing of water within aggregate pore systems may also pose problems in concrete durability [45, 47]. Similar to freezing of water in cement paste, saturated aggregate pore systems will expel water during freezing. This may result in a hydraulic pressure induced stress within the aggregate and in the aggregate / cement paste interfacial zone. The 'onion skin'-appearing failure of many of the aggregates may result from this pressure. This effect is increased with aggregate size and is influenced by freezing rate and degree of saturation. T.C. Powers reports that the effects of entrained air in accommodating pressures from freezing of aggregates appear to be minimal [45].
The influence of immersion-type vibrators on the entrained air void system has been shown to have a limited zone of influence. Simon et al. [48] demonstrated that concrete close to the immersion probe may experience a reduction of 50 % of the total air and a 100% increase in specific surface. They found a decrease in the large entrained air voids at the point of insertion, along with an increase in small air voids. They concluded that some of the larger bubbles are being broken up into smaller bubbles. However, beyond a radius of 125 mm, vibration has little effect on air void systems. For the Iowa concretes, the air void system quality, if judged primarily on spacing factor did not appear to be adversely affected by vibration outside of the vibrator trails, and in many cases was improved near the surface. The core taken directly from a vibrator did exhibit a sub-standard air void system for about half of its depth, roughly the probe immersion depth.
Mehta and Monteiro [47] indicate that aggregate gradation affects the volume of entrained air, whose volume may be decreased by excess sand and the presence of fine-sized mineral admixtures. Mixing, transportation, and placement procedures need to be optimized as insufficient mixing, over mixing, excessive transport and handling time, and over vibration all can reduce the air void system.
Neville [49] also acknowledges the influence of mixing on air content: "the cement should be well dispersed and the mix uniform before the air entraining agent is introduced. If the mixing time is too short, the air entraining agent does not become sufficiently dispersed, but overmixing may expel air - there is an optimum mixing procedure."
Kosmatka and Panarese [50] identify mixing action as one of the most important factors in production of entrained air in concrete. The lack of a suitable entrained air void system in most of these cores indicates a need to examine the effects of production procedures on the development of the entrained air void system.