Entrained Air-Void Systems for Durable Highway Concrete (2021)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

3 Background Laboratory testing and field experience have shown that highway concrete must be properly air entrained to resist the action of F-T in cold and wet climates (ACI Committee 201, 2016). Resistance to F-T depends on the characteristics of the concrete air-void system. Entrained air voids are stabilized by using chemicals in fresh concrete during mixing. An effective air-void system consists of small, closely spaced, stable voids that provide space for expanding water and ice formation during freeze conditions. These air-void system characteristics are influenced by material selection and mixture proportioning, mixing, placement, and finishing practices (FHWA 2007). The research team reviewed national and international research reports, standards, and practices dealing with the following topics: • Constituent materials and the air-void system, • Mechanisms of entrained air-void formation in concrete, • Mechanisms of F-T action in hardened concrete, • Air-void structural characteristics and the measurement of these characteristics, • F-T durability testing, and • Effects of air voids on mechanical properties. Freeze–Thaw Durability of Highway Concrete The ability of a concrete system to survive a cold and wet environment is primarily dependent on the air-void system in the hardened concrete, although other aspects of the mixture, such as mixture ingredients and paste porosity, may also influence performance, as discussed below. Mixture Ingredients The amount of freezable water in the paste capillaries affects the risk of damage. A reduction in the volume and size of the capillary voids enhances the F-T resistance of concrete because there is less freezable water in the system, and the freezing point is lowered in smaller voids due to surface tension effects (Pigeon and Pleau 1995). The pore system, in turn, is influenced by both the physical and chemical characteristics of the cement and the degree of hydration. Over time and with continued hydration, the continuity of the capillary pores in the paste decreases due to the formation of hydration products (Powers 1960, Garboczi and Bentz 1995). Another effect of cement chemistry is that air content may increase with increasing alkali levels because more of the air-entraining agent remains in solution during mixing. However, some researchers do not support this hypothesis (Greening 1967, Pistilli 1983, Pigeon et al. 1992). C H A P T E R 2

4 Entrained Air-Void Systems for Durable Highway Concrete With proper curing, concrete containing blended cement, slag, or fly ash is reported to have enhanced long-term strength and durability characteristics, partly because such concrete generally has a more refined capillary pore system (Pigeon and Pleau 1995, ACAA 2003). The presence of reactive carbon in fly ash may cause air-void stabilization problems and, therefore, may have a negative impact on F-T resistance. Due to their porosity, the carbon particles absorb the air-entraining agents and reduce their effectiveness (Klieger and Perenchio 1976, Ramachandran 1995, Karakurt and Bayazit 2015). The formation of the entrained air-void system during mixing is influenced by the concrete aggregate’s characteristics, such as aggregate size, gradation, shape, and texture, as well as the amount of contaminants (Verbeck and Landgren 1960, Waugh 1961, Gaynor 1967, Powers 1975, Kaneuji et al. 1980, Hudec 1989, Page and Page 2007). Because these factors also significantly influence the form of the interfacial transition zone, they also affect concrete F-T resistance (Rhoades and Mielenz 1946, Helmuth 1961, Dunn and Hudec 1965, Sawan 1987, Chatterji and Jensen 1992). Fine aggregates can influence the air-void system (Kosmatka and Wilson 2016). Increasing the amount of medium-sized fine aggregate tends to entrain more air (PCA 1962) because the aggregate provides a screen to hold bubbles in the paste (Deno 1966). Conversely, the use of very fine particles (<150 µm) results in a reduction of entrained air (Kosmatka et al. 2002). The benefits of entraining air were accidentally discovered in the late 1930s when researchers observed that concrete blended with portland and natural cement containing “crushed oil” was more resistant to surface scaling (Jackson 1944). Researchers also observed that concrete pavements with certain cements were more durable in a freezing environment than other pavements. Researchers found that cements used for the more durable concretes were manu- factured with grinding aids, including beef fat, calcium stearate, and fish oil, which acted as air-entraining agents. The properties and constituents of currently available air-entraining admixtures are summa- rized in Table 1. Other chemical admixtures are used in concrete to modify concrete properties such as workability and setting time. The interactions between these admixtures and air-entraining admixtures can be critical because they may affect the air-void system. Paste Porosity The form and source of pores in a hydrated cement paste system has a strong influence on F-T resistance. Calcium silicate hydrate (C-S-H) is a principal product of cement hydration. Classification Chemical Description Air Generation Prolonged Mixing Bubble Size Compatibility With Most Admixtures Vinsol resin, wood rosin Tricyclic acids, phenolics, and terpenes; typically anionic Rapid Air loss Medium Compatible Tall oil Fatty acids and tricyclic acids, typically anionic Slower Air may increase Small Compatible Vegetable oil acids Coconut fatty acids and alkanolamine salt, typically anionic Slowest Moderate air loss Coarse Compatible Synthetic detergents Alkyl aryl sulfonates and sulfates, typically anionic Rapid Minor air loss Coarse May be incompatible with some high-range water reducers (HRWRs) Source: Whiting and Nagi 1998. Table 1. Classification and performance of common air-entraining admixtures.

Background 5 Powers (1955) recognized that this product determines the properties of cement paste. C-S-H pores (known as gel pores) are a few nanometers in size; C-S-H porosity is about 25% (Cordon 1966). The small size of these pores means that water cannot freeze inside them at normal temperatures, although the water in them may be supercooled (Powers 1958). During hydration, the voids left behind as the mixing water is removed from the system are known as capillary voids (Jennings et al. 2008); the size of these capillary pores ranges from approximately 5 nm to 1 µm. The volume of capillary pores depends on both the degree of hydration and the original water/cementitious material (w/cm) ratio (Verbeck and Klieger 1956, Powers 1960, Cordon 1966, Aligizaki 2006, Jennings et al. 2008) (see Figure 1). Air voids in hardened cement paste are either entrapped or entrained. Entrapped air voids occur due to insufficient consolidation. These air voids are typically greater than 3⁄64 in. (1 mm) in size and are irregular in shape (ASTM C125). Entrapped air voids are typically isolated from other entrapped air voids and provide no benefit to the concrete (Powers 1954). In contrast, entrained air voids are spherical, or nearly so, and typically 10 µm to 1,000 µm in diameter. They are discrete and uniformly distributed, and therefore have little effect on concrete perme- ability. A comparison of the sizes of pores and concrete components is presented in Figure 2 (with capillary pores labeled as capillary voids). Table 2 presents a comparison of each type of pore and their functions in F-T resistance. Air-Void System Total air content is defined as the total volume of air voids expressed as a percentage of the bulk volume of concrete (ASTM C125). This parameter has been the basis of acceptance testing in the AASHTO T 152 test method. Despite the fact that the critical parameter is the spacing factor, total air content has been a satisfactory measure because there was good correlation between the two parameters for the materials available. With the use of different materials and changes to system chemistry, this correlation has diminished and the need to measure other parameters that indicate potential durability has increased. Powers (1954) suggested that air voids perform two functions in concrete: They limit the hydraulic pressure in the paste during the early stage of freezing and they limit the formation Figure 1. The components of cement paste in concrete.

6 Entrained Air-Void Systems for Durable Highway Concrete of ice bodies. Based on the former function, the space between air voids (sometimes expressed as the spacing factor) becomes critical because, as water expands with decreasing tempera- tures (below ∼ 39°F), it has to flow through saturated pores to find empty air voids before it freezes. Even with the same air content, the spacing factor can be quite different between mixtures. The standard approach to measurement is based on the ASTM C457 microscopical method, as discussed below. Clustering is the phenomenon in which entrained air bubbles preferentially collect around aggregate particles. This clustering reduces the bond between the paste and the aggregate (Hover 1989, Lamond and Pielert 2006, Sutter 2007) and thus has the potential to reduce compressive strength (Cross et al. 2000, Kozikowski et al. 2005). However, Riding et al. (2015) reported no correlation between clustering and strength. In general, a system of small, stable bubbles close together is desirable to provide F-T protection for concrete exposed to cold, wet weather. Compared with capillary pores, spherical bubbles are much more difficult to fill with water under capillary action. This has the effect Note: CH = calcium hydroxide. Figure 2. Size of elements in hardened concrete. Pore Type Diameter Location Effects on F-T Gel pores Intra-gel: < 0.6 nm Inside C-S-H Water in gel pores may travel into capillary pores to reduce solution concentration due to osmotic pressure Inter-hydrate: 0.1 nm–100 nm Space between C-S-H and CH Capillary pores Small: 2 nm–50 nm Between cement grains and products of hydration Ice crystals form in capillary pores (> 50 nm) and may generate stresses that damage the pasteLarge: 1 µm–10 µm Air voids Entrained: 10 µm–1,000 µm Between cement grains and products of hydration Provide boundaries for water to be forced out in capillary pores due to ice formation; limit hydraulic pressure Entrapped: 1 mm No effect Sources: Aligizaki 2006, Mehta 1986. Table 2. Pore types and functions in freeze–thaw.

Background 7 of reducing the degree of saturation of the system at any given time. Unsaturated bubbles provide a space into which expanding water and ice can move without incurring pressures within the hydrated cement paste. The bubbles need to be close together to reduce the distance that expanding water has to travel, and they need to be small to reduce the impact on the mechanical properties of the mixture. The principal objective of using air-entraining agents (AEAs) in concrete is to provide and stabilize an air-void structure that is able to protect concrete from the potential deterioration brought on by freezing and thawing. AEAs are surfactants that are readily adsorbed at air-water or solid-water interfaces and serve as bubble stabilizers (Figure 3). The literature includes two explanations for how the air-void system forms during mixing. One explanation is that, during concrete mixing, air layers are trapped between the folding surfaces of the concrete, and the fine aggregate acts as a so-called three-dimensional screen to hold air bubbles within the network of particles (Dolch 1996). Powers (1968) described another process in which air bubbles are formed in a vortex as the mixture is stirred. Mixing provides the energy to build the interface between air and liquid in large air voids and then splits the large voids into smaller bubbles. Mechanisms of Freeze–Thaw Action on Hardened Concrete Unlike most substances, which become denser during freezing, water expands and becomes less dense when the temperature falls below 39°F. The freezing point of water inside concrete pores depends on the pore sizes, the internal pressure, and the presence of solutes. The freezing temperature decreases at higher pressures (> 1 atm); the presence of solutes such as K+, Na+, Ca2+, and Cl– in water also depresses the initial freezing point (Akyurt et al. 2002). The damage due to frost action is mainly dependent on the degree of saturation (Fagerlund 1993). Damage is unavoidable if the critical degree of saturation of approximately 86% for all of the voids is reached (Li et al. 2012). The mechanisms believed to be primarily responsible for F-T deterioration are (Cordon 1966, Plum and Hammersley 1984): • Hydraulic pressure generated by freezing in capillaries, • The diffusion of supercooling gel water into capillaries after freezing, • Osmotic pressures resulting from the partial freezing in capillaries of solutions that have local concentration differences, and • Differential strains due to localized shrinkage and swelling. These mechanisms are influenced by the ability of fluids to travel through the system, as well as the saturation of the pores. It is difficult for water to penetrate spherical air voids through Figure 3. Stable air bubble with air-entraining admixture.

8 Entrained Air-Void Systems for Durable Highway Concrete capillary action due to the nature of surface tension forces; however, external pressures in the pore system may drive fluids into an air void. The extent of damage is significantly influenced by the fluid transport characteristics of the system, as well as the size, shape, and connectivity of the pores. Effects of Air Voids on Mechanical Properties The mechanical properties of concrete are influenced by each component in the matrix of the mix. While the major impact of air voids on concrete is on frost resistance, air voids also have an impact on mechanical properties such as compressive strength and drying shrinkage. It is well understood that increasing the air content of concrete may cause a commensurate loss in strength, which, as a result, may impact the durability of concrete (Cross et al. 2000, Sutter 2007, Kosmatka and Wilson 2016). A 1% increase in air content reportedly leads to a 2% to 6% reduction in compressive strength (Whiting and Nagi 1998). The dominant factor behind this reduction in strength is likely that air voids provide a shortcut for crack propagation because they have no strength themselves. Measurement Air-Void Systems For fresh concrete, the main challenges in measuring the air system are the timeliness of the results, measurable parameters, accuracy, and ease of use. For hardened concrete, the challenges are accuracy, measurable pore sizes, resolution, and sample preparation. Three con- ventional methods are used in the field to determine total air content for fresh concrete: the pressure meter method, the volumetric method, and the gravimetric method. The AASHTO T 152/ASTM C231 pressure meter method is the most widely used method for determining total air content. According to Boyle’s law, applying pressure to a known volume of concrete compresses the voids in it, which in turn reduces the volume of the concrete itself. The change in pressure of a known volume and the initial pressure of a vessel that is connected to the concrete allow the calculation of the volume of air in the concrete based on the assumption that only the air in the system is compressible. This method is not applicable for concrete that uses high-porosity aggregates. The AASHTO T 196 volumetric method is used to determine the total air content of concrete with lightweight aggregate. For the volumetric method, a container is filled with concrete, water, and isopropyl alcohol and inverted a number of times to release the air from the concrete and allow it to collect in a calibrated glass vessel at the top of the equipment. Air trapped in the aggregate does not affect the test results (Lamond and Pielert 2006). This method is more operator-sensitive and tedious than the pressure meter method (Pigeon and Pleau 1995, Lamond and Pielert 2006). This method presents a potential for underestimation of air content because extracting the smallest air voids from the paste can take more than an hour (Ozyildirim 1991). The AASHTO T 121 gravimetric method is an indirect method based on comparing the measured unit weight of concrete with the theoretical unit weight (Pigeon and Pleau 1995, Lamond and Pielert 2006). This method is not appropriate for concrete with lightweight aggre- gate because of the variability of the moisture content and specific gravity of the aggregate. In addition to the three conventional methods, the newer SAM method, using the device shown in Figure 4, has the convenience of the pressure method while reporting numbers that have been shown to correlate with F-T performance (Ley et al. 2017).

Background 9 The most commonly tested parameters for expressing the characteristics of an air-void system in hardened concrete are total volume of air, specific surface, and spacing factor. Conventionally, the average chord length of voids is used to calculate the specific surface and spacing factor. Also, analysis of air-void connectivity, clustering, and spatial distribution is possible with more recent technologies, such as flatbed scanners used with bubble counter analysis software (Anzalone 2007) and x-ray imaging (Sutter 2007). Most existing test methods only measure air voids that are at least a few microns long. Based on these measurements, it is widely accepted that the spacing factor of an air-void system should typically be less than 0.008 in. (0.20 mm) to provide the greatest protection against F-T damage (ACI Committee 221R 1996). However, recent studies have found that much smaller air voids and spacing factors also play an important role in F-T deterioration (Zhang et al. 2012, Liu et al. 2014, Ng et al. 2014). The ASTM C457 microscopical method has long been a reference method in determining the spacing factor and specific surface of air voids in concrete. The linear traverse and point count techniques use stereology to evaluate air-void structure. Freeze–Thaw Because environmental conditions in the field vary and F-T durability is not an inherent property, no single F-T test can determine the durability of a field installation. Also, because of the slow deterioration of concrete due to F-T damage, accelerated test methods have been developed to assess the ability of concrete to resist cold weather. The current AASHTO T 161 method includes two versions (procedures A and B). In proce- dure A, the tested samples are frozen and thawed in water; in procedure B, the samples are frozen in air and thawed in water. Both methods specify sample testing to begin at an age of 14 days immediately after curing and a cooling rate of 13.3°F per hour between 0°F and 40°F. The rela- tively early age, rapid cooling rate, and high degree of saturation are intended to obtain results Figure 4. Super air meter.

10 Entrained Air-Void Systems for Durable Highway Concrete quickly, but these conditions yield poor correlation with field conditions (Jacobsen et al. 1997, Ferreira 2004, Tang and Petersson 2004, Penttala 2006, Valenza II and Scherer 2007). The following factors affect results: • Cooling Rate and Time. Rapid cooling produces small ice crystals that do not represent those occurring in the field. AASHTO T 161 requires a maximum time between cooling and warming of 10 minutes. Researchers have noted that the time the sample is held in the freezing condition directly influences the observed performance. • Degree of Saturation. The high degree of saturation of AASHTO T 161 specimens results in a severe state that is not representative of most concrete structures, especially pavements. The capillary suction, internal distress, and F-T test (CIF/IDC) and capillary suction of deicing solution and freeze test (CDF) proposed by the International Union of Laboratories and Experts in Construction Materials, Systems, and Structures (RILEM) introduce water through capillary suction in a one-dimensional process similar to absorption (Setzer et al. 1996, 2004) that more closely mimics field conditions and are generally accepted as more representative tests. • Conditioning. AASHTO Method T161 calls for specimens to be tested immediately after curing or lime conditioning in the case of field-cut specimens. However, in practice, drying and resaturating the samples produces a better F-T response. The Kansas Department of Transportation (KDOT) requires drying and resaturating before testing begins (Riding et al. 2015), which corresponds to the drying period specified in RILEM TC 117 and TC 176 (Setzer et al. 1996, 2004). • Cycles. Current F-T tests are run for either 300 cycles or until the dynamic modulus drops below some percentage (typically 60%). KDOT requires F-T testing to 660 cycles because a number of sections that performed well to 300 cycles did not exhibit good durability in the field (Riding et al. 2015). Systems containing some aggregates tend to deteriorate between 500 and 600 cycles (Crovetti and Kevern 2018). Ideally, a test could be developed that indi- cates long-term performance without significantly increased testing time. RILEM TC 117 and TC 176 use 28 cycles for samples exposed to deicers and 56 cycles for those in water. • Deicer Solution. Early work by Klieger (1957) indicated that damage most rapidly occurred in systems that were tested in solutions with a deicer concentration of 3% to 4% (the value adopted in most test methods). The mechanism behind this damage is reported to be the “glue spalling” effect, where the frozen solution on the surface of the sample contracts more quickly under cooling than the concrete and sets up shear surface stresses (Valenza II and Scherer 2006). The coefficient of thermal expansion is reportedly greatest at this concentration. Scaling The ASTM C672 test method measures the resistivity of concrete against salt scaling. This test uses specimens that have been moisture cured for 14 days and air dried for an additional 14 days. The samples are then covered in a 4% calcium chloride solution and frozen for 16 to 18 hours, after which they are thawed at the laboratory air temperature for an additional 6 to 8 hours. Visual examination is conducted every 5 cycles on the flushed surface of the samples, and the calcium chloride solution is replaced every 15 cycles. Specimens with moderate scaling, i.e., a rating of 3 or lower on a scale of 5, pass the test. Some states monitor mass loss after 50 cycles and require a value of 0.3 lb/ft2 or less. Another less severe test has been suggested by the Bureau de normalisation du Québec that involves different curing conditions and finishing techniques for specimens (BNQ 2002 NQ 2621-900).