Guide for Curing Portland Cement Concrete Pavements, II

APPENDIX B: LITERATURE REVIEW

The best known early work on the effects of cement hydration, and hence curing, on cement microstructure was by Powers and Brownyard (1946-47).⁽⁷⁶⁾ Senbetta (1994)⁽⁵²⁾, Taylor (1997),⁽¹¹³⁾ and Meeks and Carino (1999)⁽⁴⁾ have written excellent summaries of this and other early work on effects of curing. In practical terms, several major themes emerge, as described by Neville (1996a).⁽¹¹⁴⁾

• The relative humidity in the cement pores must remain above 80 percent for hydration to continue, although Powers (1947)⁽⁸³⁾ and Ho, Cui, and Ritchie (1989)⁽¹¹⁵⁾ showed that hydration slows down significantly as RH drops below 100 percent.
• At water-cement ratios less than approximately 0.40, the cement paste will self- desiccate so that additional water must be added through curing if hydration is to continue.
• At water-cement ratios greater than 0.40, mixing water is sufficient to cure the paste if it is not lost to evaporation.
• Concrete within approximately 30 to 50 mm of the surface is particularly sensitive to drying, while concrete deeper than 50 mm is usually relatively immune to drying.
• Cementitious materials that develop properties slowly, such as those containing pozzolan, must be cured longer than systems not containing pozzolan or slag.

For purposes of this work, these phenomena are taken as largely representing the status of research up to about 10 years ago. The remainder of this section of the report largely focuses on research published in about the last 10 years, although literature on curing compounds is traced back somewhat farther.

A common feature in the recent research literature on concrete curing is the development of comparative data, which are useful in evaluating option, but often the data are difficult to use for actually designing guidance. For example, research will commonly compare effects among types of cementitious materials, among different mixture proportions, among different curing materials and practices. Conclusions of this research are often that one material or a class of materials are more sensitive than another class of materials, or that a certain procedure is better than another for a particular type of material. The design and intent of the research was not to determine optimum curing times or application rates of curing materials; thus, it is sometimes difficult to extract from this research recommendations for specific guidance on how to specify curing practices. However, the literature was reviewed with a particular focus on trying to get this kind of information from it.

Early research and guidance focused on effects of curing on properties of cement paste and on strength of concretes and mortars as the basic performance property. Guidance on curing is still largely based around strength. Strength is properly identified as the performance property of interest when determining when concrete can be load bearing. Durability against one or more of the degradation processes that affect concrete is also sometimes strongly affected by curing. Some degradation processes are affected by the state of hydration of the entire section of concrete, such as strength. However, other curing related problems are more affected by the properties of the concrete at or near the surface than by the strength of the entire section of concrete. Examples include cracking resulting from near-surface drying strains, rate of penetration of water and waterborne salts from the concrete surface, and abrasion resistance. Obtaining adequate curing of the near-surface zone is more likely to be disrupted by inadequate choice of materials and practices, while concrete more than about 50 mm from the surface is often relatively immune to the effects of bad curing practices. Research on the relationship between curing and development of physical properties has tended to recognize both the strength aspect and the surface properties aspect of the problem. Some reports focus on one or the other of these features, while others attempt to deal with both.

General Concepts

Burrows (1998) comments that lack of adequate curing is too often cited as the cause of cracking in structural concrete.⁽¹¹⁶⁾ Contrary to this conventional wisdom, he argues that it is possible to cure concrete too much and cites literature demonstrating this concept. The strength, modulus of elasticity, and the creep capacity of paste are all increased by amount of curing. The amount of drying shrinkage that develops increases with increasing amounts of cement paste and with the amount of hydration of that paste. Hence, the combination of these effects results in higher probability of cracking when concrete is cured beyond the minimum amount needed to develop the needed physical properties. Altoubat and Lange (2001) also demonstrate this phenomenon in their study of creep, shrinkage, and cracking of concrete at early ages.⁽¹¹⁷⁾

Cather (1994) discusses the problem of curing in general and emphasizes the difficulty in translating knowledge developed from research into guidance that can actually benefit the performance of concrete.⁽⁷⁵⁾ He comments on the importance of developing specifications that can be verified. Sometimes in the spirit of setting all specification requirements in the format of performance specifications (as opposed to prescriptive specifications), unenforceable specifications develop. He further comments that it is commonplace to attribute many of the deficiencies in concrete to inadequate curing. In actuality, it is difficult to separate inadequate curing from other problems that occur in the early age of a concrete structure. Plastic shrinkage cracking is an exception to this. Another interesting point is the concept of using time to capillary discontinuity as a guide to determine length of curing requirements.

Near-Surface Effects

The concrete near the surface of a concrete structure has been recognized as potentially different from concrete interior to the structure because of the potential for exposure to different environmental conditions, either during curing or during service. Curing deficiencies will likely have their strongest effect on this part of the concrete. A poorly cured near-surface zone is likely to be less durable than a well cured one because of the possibility of a less dense microstructure and the possible presence of some level of cracking. This truth has been known for a long time, as indicated by Gonnerman's (1930)⁽¹¹⁸⁾ study, of the effect of curing on strength and abrasion resistance. Recent literature further investigates this (Dhir, Hewlett, and Chan, 1989⁽¹¹⁹⁾ and 1991;⁽¹²⁰⁾ Parrott, 1995;⁽¹²¹⁾ and McCarter and Watson, 1997.⁽¹²²⁾ Much of the research on curing in the last 10 years involves the near-surface zone in one way or another, mostly as it affects durability; development of test methods that measure near-surface properties have dominated the literature on curing test methods. This is discussed in a later section.

This thickness of the near-surface zone or curing-affected zone varies from a few millimeters to about 50 mm, depending on the composition of the concrete and on the climatic conditions (Cather, 1994).⁽⁷⁵⁾

Hooton et al. (2002) examined effect of duration of curing on resistance of the near-surface zone to Chloride (Cl) penetration.⁽¹²³⁾ They found that there was little added benefit to curing beyond 3 days at 20 °C. They also found that the detrimental effects of poor curing were limited to the top 40 mm of concrete.

Moisture Movement and Evaporation

There has been some research effort in the past 10 years to develop a better understanding of the physics of water movement and evaporative loss from concrete. This research is driven by two needs. One is to better understand properties required of curing compounds and evaporation retardants. The other is to estimate time required to dry concrete. The latter is driven primarily by the floor covering industry.

Coleman, as reported in Wang, Dhir, and Levitt (1994), described drying of a solid mass as a process that can be approximately classified into three phases.⁽¹²⁴⁾ Phase one represents evaporation from a saturated surface and is equivalent to evaporation from a free-liquid surface. Phase two occurs when the rate of evaporation from the surface exceeds the rate of liquid movement to the surface from the interior of the solid mass. Phase three represents the propagation of a drying front moving into the solid mass. The rate of liquid loss is highest in phase one, decreasing through phases two and three. They evaluated the applicability of this three-phase model to the loss of water from fresh concrete. They found that this model did not describe observed evaporation well. Evaporation rates were found to be higher than phase one rates (free surface) throughout the first 16 h after placing. Evaporation rates were found to be dominated by temperature rises in the concrete from the hydration of calcium aluminate and gypsum in the portland cement. This action generally occurs in the first few minutes of hydration. Evaporation rates were also dominated by hydration of C₃S and reactions involving calcium-sulfoaluminate phases, typically occurring within the first few hours of hydration.

Research on moisture movement, the objective of which is to understand the drying process, contributes indirectly to understanding of curing concrete. This research mostly focuses on moisture movements in concretes that contain only water vapor and the relationship between concrete properties and rate at which this vapor can be reduced to below about 80 percent and lower. Surface adhesion systems work best when the RH is below 70-80 percent, and mold growth is retarded at this RH. The following references represent recent work in this area and provide good descriptions of the physics of moisture movement and sorption ⁽¹²⁵⁾ Suprenant, 1997;⁽¹²⁶⁾ Suprenant and Malisch, 1998a⁽¹²⁷⁾ and 1998b;⁽¹²⁸⁾ Xi, Bazant, and Jennings, 1995a;⁽¹²⁹⁾ Xi, Bazant, Molina, and Jennings, 1995b;⁽¹³⁰⁾ and Hedenblad, 1997.⁽¹³¹⁾

The rate with which water moves into concrete can have important implications for some kinds of curing practices. Low water-cement ratio concretes (<0.40) are known to sometimes consume all of the mixing water during hydration so that, if additional hydration is needed, water must be added to the concrete during curing. But the capillary continuity among pores and with the surface of concrete tends to decrease rapidly in low water-cement ratio concretes, so that introducing appreciable water into the concrete may be difficult (Meeks and Carino 1999⁽⁴⁾, Cather 1994⁽⁷⁴⁾). The onset of capillary continuity varies with water-cement ratio. Table 23 gives times to capillary discontinuity from three different references. The numbers are similar but not in exact agreement.

For purposes of curing concrete pavements, the principal point is that capillary discontinuity of 0.40 water-cement ratio concrete develops in 2 to 3 days, and that any attempt to introduce water into the concrete through wet curing methods will decrease sharply in efficiency after this time.

Berthane (1984)⁽¹³²⁾ measured water losses from fresh concrete specimens during the first few hours (up to 24) of exposure to a variety of evaporation conditions and compared the rates with those derived from the evaporation rate nomograph in figure 1 of ACI 308⁽³¹⁾ (also published in a number of other places). This nomograph is commonly attributed to equations developed by Menzel. Berthane concluded that Menzel's equations did not predict evaporations rates all that well.⁽¹³²⁾ The observed evaporation generally exceeded the rates predicted by Menzel's equations. This was attributable to the fact that the heat of hydration of the cement caused the specimens to heat up, driving off additional water. This affect was most pronounced after the first several hours when the cement would have been expected to have set. Cement is known to have a significant heat of hydration at this time. From examination of the data in Berthane's figures, it appears that the evaporation rates observed during the first 2 h, or so, were reasonably approximated by Menzel's equation.⁽¹³²⁾

Plastic Shrinkage Cracking

Ravina and Shalon (1968) conducted laboratory tests to investigate the effects of water-cement ratio, cement content,⁽¹³³⁾ evaporation conditions, and tensile strength of fresh mortar on development of plastic shrinkage cracks. The range of mixture proportions covered overlapped with paving mixtures only a little, but two important things pertinent to concrete paving were evident. One was that mixtures that developed high tensile strengths, like paving mixtures, did not crack under the experimental conditions used (evaporation rates of 0.5-1.2 kg/m²/h). Second, cracking in higher water-cement ratio mixtures usually occurred after evaporation of 2-3 kg/m² of mixing water. These cracks developed 0.5 to 2 h after all bleed water had disappeared. The authors concluded that tensile strength was a major variable determining susceptibility to plastic shrinkage cracking. Although not extensive with respect to numbers of mixtures and conditions, these results suggest that exposure for 30 min to evaporation conditions around 1 kg/m²/h might be acceptable. This limit exists in some State DOT guidance.

Radocea (1994) models shrinkage due to drying during the plastic phase, but does not attempt to include the tensile properties of the concrete and so the model does not extend to predicting cracking.⁽¹³⁴⁾

Wang, Shah, and Phuaksuk (2001) investigated effects of fly ash and fibers on plastic shrinkage cracking.⁽¹³⁵⁾ Fibers generally reduced crack areas on laboratory specimens as did a Class F fly ash when used as 30 percent replacement for cement. A Class C fly ash increased the amount of plastic shrinkage cracking.

Mora et al. (2001) investigated the effect of fibers and shrinkage-reducing admixtures on plastic shrinkage cracking. Again, fibers reduced the amount of shrinkage cracking. Shrinkage-reducing admixtures are principally intended to reduce amounts of drying shrinkage, but were also found to reduce plastic shrinkage.⁽¹³⁶⁾

Holt (2000) quantitatively measures evaporation, shrinkage, and development of cracking in laboratory concretes and describes the relationships among them.⁽⁷⁰⁾ Concrete used in the author's laboratory program showed that water losses in the range of 0.5-1.0 kg/m² would be sufficient to cause cracking if it occurred late in the initial curing period.

Hammer (2001)⁽¹³⁷⁾ examined shrinkage forces, as measured by pore-water pressure, in the presence and absence of silica fume and in the presence of mild drying. The author found that negative pore pressures developed within about 30 min under the drying conditions.

Autogenous Shrinkage

Autogenous volume change is defined in ACI 116R⁽¹⁾ as "change in volume produced by continued hydration of cement, exclusive of effects of applied load and change in either thermal condition or moisture content."

There are two autogenous phenomena that contribute to shrinkage (Powers et al., 1946-47).⁽⁷⁶⁾ One is the reduction in volume associated with the formation of hydrated cement paste, relative to the volume of the materials before hydration. The other is the partial desiccation of capillary pores due to consumption of mixing water by the hydration reaction.

At water-cement ratios greater than approximately 0.40 (depending on cement chemistry), these effects have long been considered to cause insignificant volume changes in concrete. With the more commonly used low water-cement ratios in much high-performance concrete, the contribution of autogenous effects is considered to be much larger (Baroghel-Bouny and Aitcin, 2001).⁽¹³⁸⁾ Autogenous shrinkage may be an increasing problem in paving as lower water-cement ratio mixtures seem to have become popular in recent years. Autogenous shrinkage problems are likely to appear in patching and fast-track concreting where use of low water-cement ratios are very common.

Bjontegaard and Sellevold (2001) investigated the combined effects of strains due to early changes in temperature associated with hydration and autogenous shrinkage.⁽¹³⁹⁾

Beltzung and Wittmann (2001) develop the theoretical aspects of autogenous volume changes and measure effects at very early ages (from time of adding water until just past time of setting). There is significant shrinkage associated with the earliest reactions, that has practical significance in concretes with high cement contents.⁽¹⁴⁰⁾

Sule and van Breugel (2001) investigated the interaction between autogenous shrinkage and cracks formed around reinforcing steel in very low water-cement ratio concretes (0.33).⁽¹⁴¹⁾ Reinforcement tends to distribute strains, resulting in longer times to cracking and the formation of many small cracks rather than a few large ones.

van Breugel (2001) discusses some autogenous shrinkage modeling efforts and shows how these are able to separate and analyze the many components of this phenomenon.⁽¹⁴²⁾

Altoubat and Lange (2001) make several significant points about autogenous shrinkage.⁽¹¹⁷⁾ They found that autogenous shrinkage was a significant component of shrinkage even before time of setting in concretes with a water-cement ratio as high as 0.50, even with no drying. A high rate of stress during this early period, even if not producing cracking, does seem to condition the concrete to cracking at later ages. Holt (2000) draws this same conclusion.⁽⁷⁰⁾ The ability of the concrete to creep in tension, which is inversely related to the amount of hydration that occurs, is critical in preventing cracking at early ages. For structures prone to cracking, a curing regimen that includes periodic wetting is of great benefit in relaxing shrinkage strains.

Durability

The effect of curing on concrete durability was recognized in some of the earliest literature, particularly the effects of curing on surface features such as abrasion resistance. Lack of adequate curing continues to be considered one of the most important causes of poor durability.

Stewart (1997) used probabilistic methods to calculate the influence of poor curing and compaction practices on the frequency of serviceability problems in concrete.⁽¹⁴³⁾ A survey of engineers in Australia showed that curing practice was considered to be poor on 44 percent of projects. The author attributes this to the sensitivity of concrete to the timing of start of curing and the attention to detail required in applying curing and sometimes throughout the curing period. He concluded that poor curing and compaction increased the probability of serviceability problems by an order of magnitude, with poor curing being responsible for most of this.

Kettle and Sadegzadeh (1987) reported the effects of different curing methods on abrasion resistance of concrete.⁽¹⁴⁴⁾ The principal variable was method of finishing, but adequate curing was also found to be important. Both plastic sheeting and curing compounds were effective in giving good abrasion resistance.

Malchow and Senbetta (1987)⁽¹⁴⁵⁾ demonstrated quantitatively the effect of good curing on abrasion resistance, scaling resistance, corrosion of steel, chloride penetration, shrinkage, and water absorptivity. Curing was continuous, either by sealing the surface of specimens with wax or curing compound, or by storage in a fog room, so no information on length of early curing could be derived.

Gowripalan et al. (1990) looked at effects of curing on properties that tend to correlate with susceptibility to a number of aggressive conditions for concrete: porosity, gas permeability, and water absorption.⁽¹⁴⁶⁾ The difference in these properties when compared between 2‑day and 7‑day moist curing was very large, but even 7-day curing did not completely overcome the slow hydration effects of fly ash or slag unless curing temperature was elevated to 35 °C.

Rasheeduzzafar, Al-Gahtani, and Al-Saaldoun (1989) investigated the effect of curing on time to corrosion in chloride environments and resistance to sulfate solutions.⁽¹⁴⁷⁾ Time to corrosion was linearly related to length of moist curing through 28 days, the maximum length investigated. Time to corrosion after 3 days was 12 percent of the 28-day value and was 25 percent after 7 days of curing. Sulfate resistance increased in a nonlinear way with length of curing. Very little improvement was realized after 14 days of curing. More than 50 percent of 28-day values (mass loss, strength) was realized with 7 days of curing.

Mangat and El-Khtib (1992) found that length of curing was not a good predictor of sulfate resistance, but that rate of sulfate attack was inversely related to depth of carbonation.⁽¹⁴⁸⁾

Burrows (1998) challenges some of the conventional wisdom in concrete technology, including beliefs about curing.⁽¹¹⁶⁾ While acknowledging that impermeability of concrete improves with amount of curing (up to some limiting value) and that this often improves durability, he contends that excessive curing results in concrete with high potential for drying shrinkage strains. Such concretes also have a relatively high modulus of elasticity and low capacity for creep, resulting in a high potential for cracking when under restraint. He contends that many of the cracked bridge decks are a result of such a combination of modern concretes (rich mixtures) and excessive curing.

Pozzolans and Slag

In the United States, pozzolans and slags include coal fly ash, GGBFS, and silica fume. The current understanding of the hydration of cementitious materials containing these is that they require more attention to curing than does concrete containing only portland cement. However, if cured properly, the resulting microstructure of pozzolans or slag is often more dense than in pure portland cement concrete. Typically, concretes containing pozzolans, slag, or both, require relatively longer curing times, and they tend to benefit more from elevated temperatures than do simple portland cement concretes. Haque (1990, 1996, 1998) has been a proponent of curing concrete containing Class F fly ash for at least 7 days based on strength development and water absorption data. Additional curing was found beneficial, but the effect was reduced after 7 days.^{(149,150,151)}

Swamy and Bouikni (1990) measured compressive and flexural strength, ultrasonic pulse velocity, and modulus of elasticity on concretes containing GGBFS.⁽¹⁵²⁾ Specimens containing 50 percent GGBFS were not more sensitive to curing than specimens containing only portland cement, but at 65 percent GGBFS they were. Other test properties, however, showed sensitivity to this absence of curing. Even 7 days of moist curing was not enough to totally prevent some microstructural problems that occur on drying. Longer curing times were not explored. This same conclusion was reached by Gowripalan et al. (1990) although elevated temperatures (35 °C) for 7 days, or moist curing for 28 days, was required to reach approximate equivalency to control concrete containing no pozzolan cured for 7 days.⁽¹⁴⁶⁾

Austin et al. (1992) measured strength, ultrasonic pulse velocity (UPV), rebound number, water absorption, and air permeability of 50 percent slag concretes on specimens cured in air, under wet burlap with plastic on top, with curing compound, with plastic sheet, and in water.⁽¹⁵³⁾ They found the UPV to be the most sensitive measure of development of physical properties. Rebound number was the worst. Curing beyond 7 days was not examined (except in curing compound cured specimens). Temperatures above 20 °C strongly boosted the strength. The burlap curing was superior to the plastic sheeting and to the curing compound, indicating the added water was important for these concretes.

Ballim (1993) investigated curing in concretes containing 30 percent Class F fly ash and 50 percent GGBFS, as a mass fraction of cementitious materials, using oxygen permeability and water absorption as measures of physical property development.⁽¹⁵⁴⁾ The effects of curing (water immersion) times of 1, 3, 7, and 28 days were examined. Water absorption tests seemed to give the most easily interpretable set of test results. Near-surface concrete (0-15 mm) was most affected by length of curing. Curing up to 7 days resulted in relatively large changes in the measured properties. There were additional benefits with longer curing, but the increase was much smaller. This was true even for the mixtures containing no pozzolan or slag, but the effect of increasing curing time was strongest in the GGBFS mixtures.

Afrani and Rogers (1994) investigated the effect of method of curing on salt scaling of ordinary portland cement (OPC), GGBFS, fly ash (F), and silica fume concretes.⁽¹⁵⁵⁾ All specimens were cured for 14 days with either wet burlap, curing compound, and in a fog room. Best results were

obtained with the wet burlap method. GGBFS mixtures did not do well in this test, even with wet burlap curing.

Haque (1990) investigated the effect of length of fog-room curing of OPC and Class F fly-ash concretes (30 percent of cementitious materials) on strength and depth of water penetration.⁽¹⁴⁹⁾ Curing times of 0, 7, and 28 days were examined. He concluded that at least 7 days moist curing was needed.

Ozyildirim and Halstead (1994) investigated concrete containing Type II and III portland cements, and various levels of replacement with Class F fly ash and silica fume.⁽¹⁵⁶⁾ Strength and chloride permeability were measured on specimens moist cured (presumably fog room) for 1, 3, and 14 days at 23 and 38 °C. Higher temperatures, longer curing, and addition of silica fume all significantly reduced chloride permeability. Specimens containing silica fume, fly ash, and Type III cement reached low permeability values after 1 to 3 days of moist curing if curing temperatures were at 38 °C.

Thomas and Matthews (1994) found that concrete containing no more than 30 percent fly ash needed no more curing than simple portland cement concrete, 3 days being generally adequate, as determined by permeability measurements, but that 50 percent fly ash concretes needed at least 7 days curing.⁽¹⁵⁷⁾ When well cured, the permeability of fly ash concretes is lower, but depth of penetration of carbonation is higher. Carbonation depths ranged from less than 5 mm to a maximum of 35 mm, correlating strongly with strength and type of exposure (interior concrete was more deeply carbonated than exterior concrete).

Ho, Cui, and Ritchie (1989) used surface absorption of water to evaluate the length of curing required of concretes containing fly ash and slag. They concluded that 7 days of moist curing is a minimum amount for these systems.⁽¹¹⁵⁾

High-Performance Concrete

Meeks and Carino (1999) recently completed a comprehensive review of information on curing of high-performance concrete.⁽⁴⁾ Their description below of the objectives of a proposed work unit on curing high-performance concrete are pertinent to this project.

The long-term goal of the proposed research is to provide the basis for modifying current ACI standards related to curing so that structures in service will perform as required. Curing requirements should consider economy of construction and not place undue demands on the construction team. On the other hand, curing requirements should assure the owner that the potential properties of the concrete in the structure are realized. To achieve these goals, emphasis should be placed on developing a system to verify the adequacy of curing on the job, such as for example, by requiring the measurement of in place properties of the [sic] at the end of the curing period. (p. 177).

The report concluded with a discussion of research needs, summarized briefly as follows:

• Evaluation of curing methods should be done on full-scale models that accurately simulate field concrete in both dimensions, exposure conditions, and temperature history.
• The idea that low water-cement ratio concretes will benefit from added-water types of curing should be investigated. There is a concern that added water may not penetrate into the concrete to any appreciable degree, so the effort may be wasted.
• In-place test methods should be developed.
• Curing requirements should be revised to better reflect the particular performance requirements of high-performance concrete (HPC).

A common feature of HPC design is development of high strength. Mixture proportions commonly involve low water-cement ratios. As has been mentioned elsewhere, water-cement ratios less than 0.40 are generally thought to create a condition where the concrete will internally desiccate due to consumption of all of the mixing water by hydration, and that if additional hydration is necessary, then water must be added during curing. But since low water-cement ratio concretes tend to be relatively impermeable, there is some question about how effective externally added water will be in penetrating the concrete. Persson (1997) studied relative humidity depth profiles in concrete at various water-cement ratios, with and without silica fume in the concrete.⁽¹⁵⁸⁾ As expected, the internal RH of low water-cement ratio concretes (<0.4) that were sealed against any evaporation begins to drop below 100 percent within 28 days, eventually dropping to about 75 percent after 450 days. This pattern is accelerated when silica fume is present. Similar specimens stored in water without any surface sealant show the same pattern, suggesting that the external water is in fact not getting into the concrete to any appreciable degree. The minimum depth analyzed was 50 mm.

Two types of HPC are pertinent to portland cement concrete paving: bridge decks and fast-track paving.

Bridge Decks

Bridge decks are not within the scope of this project, but information on some of the practices is useful and interesting in developing ideas about curing pavements. Information on bridge deck curing practices comes mostly from State DOTs and from several papers on current practice that have appeared recently. The principal concern with the concrete in bridge decks is usually that it not be excessively permeable to chemical compounds that promote corrosion of reinforcing steel, such as carbon dioxide and particularly chloride, and that they do not crack appreciably.

Concretes used in bridge decks commonly contain silica fume and or other pozzolans or slag, and are proportioned at low water-cement ratios. If properly cured, these concretes will have very low permeability. However, cracking has been a persistent problem. Mohsen (1999) investigated causes of bridge deck cracking.⁽³⁸⁾ The investigation included variables such as damage due to physical causes, for example vibration from passing traffic (partial road closures), concrete mixture proportions, and curing. It was concluded that curing, including both moisture and temperature management, was the most important variable. Excess water in the concrete was considered the next most important variable.

Bridge decks typically get more attention to curing than conventional concrete paving. Practices vary among State DOTs, but usually considerable attention is paid to keeping the surface wet continuously after finishing to prevent shrinkage cracking. Some combination of curing compounds, water, wet burlap, and curing blankets are used for the duration of the curing, which is typically about 7 days.

One popular practice that exists is to use either an evaporation retardant or fogging to keep the concrete wet until conditions are right to apply curing compound. After the curing compound is applied, some kind of wet curing is applied. The additional water probably does not penetrate into the concrete to contribute to hydration to any appreciable extent because of the curing compound cover, but is probably more important for temperature control and possibly to compensate for any curing compound deficiencies.

Healy and Lawrie (1998) investigated cracking in highway bridge decks in Maryland. Two large bridges had almost 100 percent of surface area cracked.⁽³⁷⁾ They investigated mechanical causes, but determined that shrinkage cracking was the principal cause and that curing practices were deficient. It was reported that the Maryland DOT standard specifications require curing with curing compound along with burlap, plastic sheet, or cotton mats (apparently dry) for 7 days. A number of practices were investigated, but they found the most improvement with use of fiber-reinforced concrete or with use of curing compound together with wet burlap. The effect of the wetted burlap apparently was to help control temperature. The curing compound with fogged burlap was adopted into revised guidance along with guidance on time-of-day scheduling of construction to avoid major solar heating effects.

Ozyildirim (1999) of the Virginia Highway Research Council reported bridge deck curing was executed with wet burlap covered under plastic sheeting.⁽¹⁵⁹⁾ Concrete was designed to be low permeability, using the rapid chloride permeability test (ASTM C 1202)⁽¹⁶⁰⁾ and the chloride ponding test (AASHTO T 259).⁽¹⁶¹⁾ Curing requirements are for 7‑day moist curing or 70 percent f_c¢. Fogging is used to protect the surface from excessive evaporation until the burlap-plastic sheet curing system could be put into place. The bridge deck concrete contained slag or Class F fly ash with water-cement ratios of 0.40 to 0.45.

Waszczuk and Juliano (1999) of the New Hampshire DOT, reported on bridge deck concrete that contained silica fume with a water-cement ratio of 0.38.⁽¹⁶²⁾ Once concrete had been dragged, it was covered within 15 min to prevent plastic shrinkage cracking. Curing was with cotton mats (stitched burlap with cotton bat filler) which were installed dry, then wetted and kept wet for 4 days. Concrete was not placed unless evaporation rates were less than or equal to 0.5 kg/m²/h (it was not reported whether this was determined empirically or taken from the nomograph) and temperature was less than or equal to 29 °C.

Ralls (1999) reported on a bridge deck construction by the TxDOT. Bridge decks were cast in place.⁽¹⁶³⁾ Concrete was portland cement fly ash (approximately 30 percent), with a water-cement ratio of 0.43. Concrete was fogged until curing compound could be applied, followed by wet cotton mats as soon as the concrete could be walked on. TxDOT standard specifications require 10 days of wet mat curing for concrete containing fly ash.

Beacham (1999) reported on a bridge deck construction by the Nebraska Department of Roads.⁽³⁴⁾ Concrete was designed for a 56-day compressive strength of 55,158 kPa and a chloride permeability limit of 1800 coulombs at 56 days. Abrasion resistance was also a property of interest, but not a specification requirement. The concrete contained 445 kg/m³ of Type IP cement and 44.5 kg/m³ of Class C fly ash, at a water-cement ratio of 0.30. Concrete could only be placed if evaporation rates were below 0.75 kg/m²/h as determined by the nomograph. Maximum allowable ambient temperature for placing was 27 °C. Requirements were for 8 "curing days." A curing day is one in which the shade temperature is greater than or equal to 10 °C for at least 19 h, or surface temperature of concrete is at leas 4 °C for the entire 24 h. The project plan apparently required misting until time of setting, then curing with wet mats. Misting was found to be impractical because of wind, so curing compound application was substituted, then the wet mats, as required. This appeared to work well.

Fast-Track Paving

Fast-track paving is a technology by which road pavements are placed and opened to traffic quickly, usually within 24 to 48 h. Strength gain of most conventional concrete mixtures is too slow to allow this. One technique that has been used in fast-track paving is to take advantage of the heat of hydration of the cementitious materials to create an elevated curing temperature, thus accelerating strength gain. A more common approach is to proportion the concrete with a high cement content and low water-cement ratio (<0.40) and perhaps to use a high-early strength cement, like ASTM C 150⁽⁴⁴⁾ Type III cement. Even though no steps are taken to accelerate hydration, sufficient strength can be obtained in 24 h to bear traffic loads. Curing is of necessity abbreviated in time, which is not too important in the context of strength development, since that has been accommodated by mixture proportioning. However, there might be some concern about the near-surface qualities of the concrete.

The Iowa DOT was involved in much of the early work on fast-track techniques (Knutson and Riley, 1988;⁽¹⁶⁴⁾ Grove et al., 1990;⁽¹⁶⁵⁾ Public Roads 1988⁽¹⁶⁶⁾). This work used a Type III portland cement with a supplemental 12‑hour strength requirement of 8,963 kPa. Curing blankets were also used.

Grove (1989) explored the use of insulated blankets to elevate curing temperatures in the first 24 h after placing.⁽¹⁶⁷⁾ The objective was to try to accelerate strength enough to avoid the special purpose Type III cement. Temperature and strength differentials of rich concrete mixtures (355-415 kg/m³) at 5 to 8 °C were achieved 7 to 17 h after placing relative to uninsulated concrete. A leaner mixture (263 kg/m³), also containing fly ash (44 kg/m³), showed almost as much temperature rise, but the strength increase was much smaller. The results showed that insulation could not completely compensate for the higher strength gain of the special Type III cement, but that sufficient strength gain could be achieved to meet many project needs for rapid opening. The most advantageous aspect of using blankets is that it prevents the nighttime temperature drops and evens out the temperature of the concrete. One potential disadvantage is that solar radiation is blocked from the concrete, which may be a desirable source of heat in cool weather.

The American Concrete Pavement Association (ACPA) pamphlet "Concrete Paving Technology: Fast-Track Concrete Pavements" (ACPA 1994) has considerable advisory information on curing this type of concrete technology, as summarized below.⁽¹⁶⁸⁾ In addition to the information summarized above, the ACPA guidance recommends accelerators be considered. Also, monitoring heat development in the concrete is recommended so insulation can be adjusted. This is critical to anticipate the "sawing window," strength development (which determines when pavement can be opened for traffic), and to prevent cracking. Application of maturity testing concepts is recommended (ASTM C 1074).⁽³⁹⁾

White pigmented curing compounds (ASTM C 309, Type 2, Class A) are recommended with an application rate of 5.0 m²/L.⁽²³⁾ In mountainous and arid climates, heavier applications of 3.75 m²/L and use of white pigmented resin based materials (ASTM C 309, Type 2, Class B) is recommended.⁽²³⁾ Bonded overlays less than 152 mm require application rates of 2.5 m²/L because of the larger surface-volume ratio. Since fast track technologies cause rapid consumption of mixing water, the concrete is more susceptible to plastic shrinkage cracking. Early curing is very important; therefore, curing compound should be applied as soon as possible. Tire friction will wear some of the curing compound off, but the cement is believed to have sufficient hydration at this point that no problems result.

Blanket insulation is needed in cool weather, along with curing compound. The ACPA pamphlet contains a tabulation that allows the user to make use of ambient temperature projections and available curing times (allowable road closure times) to determine whether insulation is needed or not.⁽¹⁶⁸⁾ There is no consideration for the variable heat of hydration of various cementitious materials. "R" ratings of at least 0.035 m² Kelvins per watt (K/W) are recommended for insulating materials. Insulation constructed of a layer of closed-cell polystyrene foam with a protective layer of plastic is recommended. Additional blankets may be needed if temperatures fall below 4 °C. Time of day of concrete placing is important because of the interaction between early hydration reactions and temperature gain from thermal radiation.

When evaporation rate exceeds 1.0 kg/m²/h, using the evaporation-rate nomograph in ACI 308, plastic-shrinkage cracking is considered likely.⁽³¹⁾ To be conservative, it is recommended that particular attention be paid to moisture conditions in the concrete when evaporation exceeds 0.5 kg/m²/h.

It is recommended that flexural strength be used to determine the time when pavement can be opened to traffic, and hence deliberate curing ends. Recommended guidance is given on the flexural strength (ASTM C 78) needed for different situations.⁽¹⁶⁹⁾ It is also recommended that other field measurement methods might be preferable to C 78 and gives some guidance on correlating these with flexural strength.⁽¹⁶⁹⁾

Curing Compounds

Curing compound use dates to the early 1930s. Jackson and Kellerman (1939) describe the historical development of pavement curing methods.⁽²⁾ The traditional method was to cover the concrete with wet burlap as soon as possible after finishing. The concrete was kept wet until the following day, when the burlap was replaced with a covering of earth or straw kept continuously wet for 7 to 10 days. This method worked well, but it "required constant and efficient supervision to insure full compliance," (p. 482).⁽²⁾ This method was practical since daily placement rates were rather low. As construction techniques became more efficient, daily placement rates increased, so that the traditional methods became too expensive. "As was bound to happen, this condition has resulted within the last several years in the introduction of numerous substitute methods of curing, designed to accomplish the same purpose without the use of water," (p. 483).⁽²⁾ Early products included waterproof paper, sodium silicate, liquid bituminous products, rubber emulsions, and calcium chloride. Use of calcium chloride and liquid bituminous products has ceased over the years. Sodium silicate has been found to have little value as a curing compound. Currently available curing compounds typically fall either into an organic-solvent type or a water based type. The membrane-forming compounds may be either waxes or some type of polymerizing resin or oil.

The earliest reference to the use of curing compounds found in this literature search was in the construction of a canal lining in the Yakima Reclamation Project, Washington (Ruettigers and Whitmore (1930)).⁽¹⁷⁰⁾ Water curing was the predominant form of curing used on this project, but the last 6 miles were cured with an asphaltic compound as an experiment. There was no comment on the resulting performance.

Gonnerman (1930) conducted a systematic study of curing compounds being marketed at that time.⁽¹¹⁸⁾ This work was specifically in support of new curing methods for use in highway construction. Prior to this, highway concrete curing was accomplished by maintenance of surface moisture with earth, straw, ponding, etc. The curing compounds were asphaltic products (McEverlast paint, Tarvia K.P., and Curcrete), linseed oil, and paraffin.

Application rates were generally about 4.5 m²/L, applied immediately after finishing. The effects of curing on both strength development and durability were recognized in this work. Performance measures were strength gain (compressive and flexural), abrasion resistance, water absorption, surface hardness. Performance of these curing materials was found to be approximately equivalent to that of 14 days of moist curing.

Meissner and Smith (1938) reviewed use of currently available curing compounds for use by the U.S. Department of Interior's Bureau of Reclamation.⁽¹⁷¹⁾ These were mostly coal-tar or asphalt-based materials that had been dissolved in an organic solvent (cut backs). At that time, the Bureau of Reclamation required concrete be moist cured for 14 days, and so required that any curing compounds demonstrate equivalent performance in lab tests. Curing compounds at that time were found to perform well in moderate weather conditions, but did not do well in very hot weather. Another problem with the curing compounds of the day was the dark color, which caused solar heating and sometimes resulted in thermal cracking problems. Application rates were determined by what could be made to adhere to a vertical surface. These ranged from 3.5 to 4.7 m²/L. Moisture loss at 100 °C, 15 percent RH ranged from about 0.5 to 1.0 kg/m² at 9 days, depending on application rate and product. Abrasion resistance, along with moisture loss, was the performance measure. It was concluded that these products could give the equivalent of 14 days moist curing. It was strongly recommended that curing compounds be applied immediately after finishing.

Jackson and Kellerman (1939) published results of an evaluation of a number of curing methods for use on concrete pavements.⁽²⁾ They found the curing compounds of the day to be distinctly inferior to wet burlap curing, but that their performance improved substantially if the test specimens were wet cured for 24 h before application of curing compound. Their study, which was on mortar specimens, was criticized as employing unduly harsh conditions.

In another Bureau of Reclamation study, Blanks, Meissner, and Tuthill (1946) again reviewed performance of currently available curing compounds, comparing fog-room cured specimens with specimens coated with curing compound.⁽¹⁷²⁾ Elastic modulus and moisture loss were the measures of performance. They recognized then that other physical properties, such as surface hardness and abrasion resistance, could be used, but thought moisture loss was a particularly simple and direct measure. They classified curing compounds into three classes: (1) bituminous-coal tar and asphaltic cut backs, and asphaltic emulsions; (2) clear compounds-solvent type and emulsions; and (3) white pigmented. The clear, solvent types included a number of materials including resins, waxes, drying oils, and water repellent solids. The white pigmented compounds were basically clear solvent types to which a white pigment had been added.

Based on these comparisons with fog-cured concretes, they concluded that moisture loss in a standard test configuration could be used to evaluate the effectiveness of curing compounds. A test method and specification were presented. The test method is very similar to ASTM C 156⁽¹⁸⁾ (ASTM C 156-44T⁽¹⁷³⁾ existed at that time, but it is not clear from this report whether the Bureau of Reclamation test method is the same or some slight variant). Moisture loss of a standard mortar specimen was measured after 7 days at 38 °C, 21 percent RH (saturated CaCl₂). Application rate was 3.7 m²/L. A maximum moisture loss of 40 g (equivalent to 0.87 kg/m²) of water was determined to represent material that would result in water retention equivalent to 14 days moist curing (the standard Bureau of Reclamation moist-curing requirement). The method was reported to have "reasonably close duplication."

For simplicity of inspection and to avoid thermal problems with the use of the black bituminous compounds, the Bureau of Reclamation prescribed use of white pigmented curing compound that met the above moisture loss requirement, applied in a single coat at 3.7 m²/L.

Rhodes and Evans (1949) reported a major investigation of curing using curing compounds. They found that curing compounds work well if two conditions are met.⁽¹⁷⁴⁾ One is the early concrete temperature and the other is prevention of excessive drying before application of curing compound. Time of day is also heavily emphasized. Concrete placed in the early morning tends to develop maximum peak temperatures (through a combination of environmental and heat of hydration effects), and concretes placed early showed a temperature rise of 19 °C, while those placed late in the afternoon showed a 3 °C rise.

Burnett and Spindler (1952) measured moisture loss, abrasion resistance, and compressive strength of specimens in response to different times of application of curing compound, ranging from immediately after finishing to 18 h after.⁽¹⁷⁵⁾ Moisture loss increased with increasing time of application, but strength and abrasion resistance reached peak values when application was made 30 to 100 min after finishing. The explanation for this was that allowing bleeding and then evaporation of bleed water before curing compound application resulted in an effective reduction in the water-cement ratio. If more time passes before application of the curing compound, then too much water evaporates, affecting performance. The recommendation was to apply curing compound at time of setting, plus or minus 30 min. Application of the curing compound before bleed water evaporated did not substantially impair the physical integrity of the resulting membrane, although it did not appear to bond as well to the concrete surface as when it was applied later.

Mather (1953),⁽¹⁷⁶⁾ in a discussion of Burnett and Spindler (1952),⁽¹⁷⁵⁾ took exception to their conclusion that time of setting was a good indicator of the proper time to apply curing compound. He argued that the time when surface water has just disappeared is the best time to apply curing compound, and that this does not necessarily correspond to time of setting.

Carrier and Cady (1970) reported an investigation on some details of moisture distribution-from the surface of concrete to several inches of depth-that result from different curing compound practices.⁽⁵¹⁾ The curing compound was apparently a very good one, given moisture loss of 0.26 kg/m², as tested by C 156.⁽¹⁸⁾ They found that below about 5 mm, the application rate of the curing compound was inconsequential in causing changes is relative humidity in the concrete exposed to a dry environment. Application rates of 2.5-10 m²/L were used. They concluded that application rate was not a particularly important part of practice.

Mather (1987,⁽¹⁷⁷⁾ 1990⁽¹⁷⁸⁾) reviewed the state-of-the-art of curing practice, and specifically curing-compound practice in the latter reference. This information largely provided the basis of the current ACI 308.⁽³¹⁾ He chaired the ACI Committee 308 when the original version of that document was written. It was last revised in 1992, but has remained the standard practice to the present time. A new guide is currently being written by ACI Committee 308.

Not much literature was found reporting research on curing compounds in the last 10 years; however, two papers were found that seem to be particularly pertinent to developing improved practice. Two others were found that touch tangentially on the subject.

Dhir, Levitt, and Wang (1989) developed a concept for describing the performance of curing compounds that is based on measuring a water vapor permeability coefficient of a product.⁽²⁸⁾ This coefficient is based on a Darcy's Law application to evaporation through a thin layer. This coefficient includes membrane thickness and time in its calculation, so the property is independent of a specific application rate or time of measurement, as with the current moisture loss requirement. The water vapor permeability coefficient concept would allow a more sophisticated approach to designing curing to be developed because it more completely describes the full time-dependent and application-dependent properties. The paper also describes a test method and test apparatus for measuring this property. The within-laboratory coefficient of variation of this test method was found to be 4.2 percent.

Wang, Dhir, and Levitt (1994)⁽¹²⁴⁾ in follow up work to the work reported in Dhir, Levitt, and Wang (1989),⁽²⁹⁾ described water evaporation from an air-cured concrete as generally being higher than expected from simple evaporation from a free-water surface. The higher rates occur at two times in the early history of the concrete. The first occurs immediately after placing from the free bleed water that forms on the surface; however, its rate of evaporation exceeds that expected from a free-water surface because of the thermal warming from the heat of hydration occurring in the first few minutes involving the aluminate and calcium sulfate phases of the cement. The second peak in water loss occurs after several hours and appears to be driven by the hydration of cement phases at that time. Tricalcium silicate (C₃S) and calcium-sulfoaluminate phases typically evolve significant heat. Application of curing compound tends to eliminate water loss due to the second phenomenon.

Three types of curing compounds were also investigated in this work: resin-solvent based, chlorinated-rubber based, and wax-water based. The wax-water based compound appeared not to work as well as the others, but only one product was examined.

This paper also reported the effect of time of application of curing compound on retardation of water evaporation. It was concluded that the current general guidance that curing compound be applied after surface water has disappeared may not be optimal. Improved performance was observed when curing compound was applied while surface water was still present. This effect was particularly strong in rich concrete mixtures. The effect was hypothesized to be the result of capillary suction potential that develops when the surface of concrete dries. This potential in effect "sucks" holes in the curing compound film before it develops adequate strength. While recognizing the potential problems associated with application of curing compound when concrete is too dry, Mather (personal communication) believes that the bulk of experience has shown that application of curing compound when surface water is still apparent does not result in performance that is as good as when applied to a surface that has just lost this surface water.⁽¹⁷⁹⁾

This result and recommendation is counter to existing guidance in ACI 308, which directs that curing compound not be applied until the bleed water has disappeared by evaporation.⁽³¹⁾ There is the caution that when the evaporation rate exceeds the bleeding rate, the surface bleed water will disappear before the bleeding has stopped. If curing compound is applied then, it might soak into the first few millimeters of the surface, then as the bleed water continues to form, will lift off that few millimeters of surface. This can happen when the evaporation rate exceeds 1 kg/m²/h. ACI 308 lists precautions to take to avoid this situation.⁽³¹⁾

Kettle and Sadegzadeh (1987), in a study of effects of construction practices on abrasion resistance, found that concrete surfaces that were rough tended to show reduced abrasion resistance because of poor curing.⁽¹⁴⁴⁾ This was attributed to the curing compound flowing off of high spots after application, leaving insufficient coverage in spots. This paper also reported that the single water-based curing compound included in the study performed worse than the two resin-based curing compounds used.

Gowripalan et al. (1990) found that the two curing compounds included in their general study of durability, performed almost as well as curing with ponded water for 3 days, but that exactly equivalent properties were achieved only at a depth of 50 mm into the concrete.⁽¹⁴⁶⁾ Concrete nearer the surface did not perform quite as well. This result contrasts with older literature on curing compounds that indicated they gave curing approximately equivalent to 14 days moist curing.

Recent U.S. Environmental Protection Agency regulations require curing compounds meet low VOC limits of £ 350 g/L VOC. Many products are now available that meet this requirement. Whiting (2003) compared water retention performance of some VOC compliant compounds with non-VOC compliant materials and found no patterns in performance that distinguish these classes of materials.⁽¹⁸⁰⁾ Low VOC curing compounds are reputed to have drying time problems in very humid climatic conditions, but no literature was found on this.

White and Husbands (1990)⁽⁸⁵⁾ investigated the relationship between water-retention properties of curing compounds and curing performance, as measured by water absorption (ASTM C 1151).⁽²⁷⁾ Curing compounds varied in water loss, as measured in ASTM C 156, from 0.12 to 1.40 kg/m² at 72 h.⁽¹⁸⁾ ASTM C 309 limits water losses to 0.55 kg/m². In ASTM C 1151, a slab of mortar is cured with the subject materials and cores taken, prepared and water absorption measured on the exposed surface and on the interior surface.⁽²³⁾ Water-cured and air-cured (evaporation rates of 1.65 kg/m²/h) specimens were used as reference conditions. When water losses were up to approximately 1 kg/m², the membrane-cured specimens showed water absorption values essentially identical to water-cured specimens and to the surface of the specimen that represented the interior of the mortar. The curing compound that allowed loss of 1.4 kg/m² showed evidence of deficient performance in the water absorption test.

Tining concrete pavements is a common practice which effectively increases the surface area of the concrete. Shariat and Pant (1984) examined the effect of this on water retention and found that the amount of curing compound necessary for adequate water retention needed to be increased, relative to the control condition, in approximate proportion to the increase in surface area.⁽⁷⁴⁾ Increases in surface area of 25 to 50 percent can occur, depending on the tining pattern.

Test Methods

Test Methods for Curing Materials

Jackson and Kellerman (1939) published one of the earliest test methods for curing materials.⁽²⁾ The method was based on solid rectangular lean-mortar specimens exposed to a temperature of 38 °C and a relative humidity of 32 percent for 7 days. Performance was measured by flexural strength and by moisture loss. Test results were expressed as a curing efficiency, using results from wet cured and uncured specimens as a frame of reference. The method was criticized as being too harsh in that the temperature and RH conditions were extreme. Also the lean mortar was much more prone to drying than a well-proportioned concrete would have been, so that some curing methods or materials believed to be satisfactory, in practice were reported to have relatively low curing efficiencies. The method appears to be very similar to ASTM C 156, which was introduced several years later.⁽¹⁸⁾

ASTM C 156⁽¹⁸⁾ was first published as C 156-44T.⁽¹⁷³⁾ This method has been widely criticized for poor precision and has been modified significantly to improve on this (Leitch and Laycraft, 1971).⁽¹⁸¹⁾ Poor between-laboratory precision is particularly undesirable, especially for buyer-seller transactions. Blanks, Meissner, and Tuthill (1946) reported that a similar Bureau of Reclamation method has reasonable precision, but no quantitative data were offered.⁽¹⁷²⁾ It is entirely likely that their estimation of the precision is based on within-laboratory data. This statistic often tends to be substantially better than between-laboratory precision.

Mather (1990)⁽¹⁷⁸⁾ commented on the data of Leitch and Laycraft (1971),⁽¹⁸¹⁾ which showed that by modifying the requirements for C 156⁽¹⁸⁾ the standard deviation could be reduced from 0.13 kg/m² to 0.05 kg/m². Mather (1990)⁽¹⁷⁸⁾ suggested that a repeatable moisture loss method could be developed.

Although there are many variants of C 156 in use, there have not been many methods proposed that offer substantial improvement.⁽¹⁸⁾ As reported in Gowripalan et al. (1990)⁽¹⁴⁶⁾ there was an effort to develop a test method that works very much like ASTM C 156⁽¹⁸⁾, but is simplified by using a desiccant in a container with the specimen to control RH. The mass change of the desiccant was used to detect the amount of water that passed the curing membrane. It was found that the mass change of the desiccant was relatively insensitive to the difference between concretes treated with a curing compound and concretes not treated at all. The original report on this work was not located.

As mentioned above, Dhir, Levitt, and Wang (1989) developed a test method for measuring the permeability coefficient of curing membranes.⁽²⁹⁾ To our knowledge, this method has not been standardized.

Cabrera, Gowripalan, and Wainwright (1989)⁽¹⁰⁷⁾ describe a test method for determining the effectiveness of curing compounds based on the oxygen permeability of mortar specimens. Determinations are made on uncoated and coated specimens, and curing efficiency is calculated as a percent reduction in the permeability of the uncoated specimen. Within-laboratory precision (CV) is about 3 percent for three materials with a curing efficiency of more than 80 percent, and about 7 percent for one material with a curing efficiency of 50 percent.

Test Methods that Validate Curing

Test methods that verify the quality of curing have traditionally relied on strength or some correlate of strength. There is a more recent trend towards measurement of near-surface properties, such as surface hardness, permeability, and water absorption. The International Union of Laboratories and Experts in Construction Materials, Systems and Structions (RILEM) Technical Committee (TC) 116-PCD on "Permeability of concrete as a criterion of its durability," is developing methods for measuring near-surface properties as a general approach to evaluating potential durability of concrete (RILEM, 1999).⁽¹⁸²⁾ In addition to these, Cather (1994) cited capillary porosity as a potentially useful method. He mentions ultrasonic pulse velocity as a potential test method, but dismisses it as probably not sensitive to the near-surface zone.⁽⁷⁴⁾

Dinku and Reinhardt (1997) developed an in situ gas-permeability test for this purpose.⁽¹⁸³⁾ In this test, a hole is drilled into the concrete and hardware is arranged so that a fixed volume of pressurized gas (in this case nitrogen) is delivered to the hole. The rate at which the pressure drops as the gas permeates the concrete around the hole is related to the permeability of the of the near-surface concrete. The test is affected by the moisture condition of the concrete, which is determined separately and a correction is applied.

Ho and Lewis (1984,⁽¹⁸⁴⁾ 1992⁽¹⁸⁵⁾) describe test methods for water sorptivity, carbonation, and abrasion resistance (based on ASTM C 418⁽¹⁸⁶⁾) as measures of the quality of curing. The Roads and Traffic Authority of New South Wales ( Australia )⁽¹¹⁰⁾ uses water sorptivity as an acceptance test for curing of concrete. Specification limits on this property are presented in this report under the section on Australian curing guidance.

Carrier and Cady (1970) developed an RH-sensitive button (results in a color change) that could be used as a field test for the presence of surface water, but no other information on the subsequent development or use of this method was found.⁽⁵¹⁾

A state-of-the-art report prepared by RILEM TC 116-PCD⁽¹⁸²⁾ on performance of concrete as a criterion of its durability (Kropp and Hilsdorf 1995)⁽⁵⁰⁾ contains two chapters on test methods for physical properties of concrete that are believed to be critical for concrete durability. These properties are also considered quite sensitive to the degree of hydration of the portland cement, and hence their development is strongly related to curing. The report also summarizes methods for determining the moisture conditions in concrete.

Chapter 9 of Kropp and Hilsdorf (1995), by Geiker, covers laboratory test methods.⁽⁵⁰⁾ Methods include steady-state water permeation, nonsteady state water penetration, capillary suction, gas permeability, gas diffusion, ion transport (along concentration gradients), and ion transport (along electrical potential gradient).

Chapter 10 of Kropp and Hilsdorf (1995), by Paulmann and Molin, covers onsite methods.⁽⁵⁰⁾ These include capillary suction, water penetration, and gas permeability. The methods are all relatively simple in concept, but not particularly simple in execution. The required equipment could reasonably be assembled without excessive cost, but individual determinations appear to require enough time to setup and collect the data. Thus, they probably are not suitable for routine inspection and verification of curing. These methods do appear to be plausible for field investigations of relatively small areas of concrete.

Appendix A of Parrott (1995) covers methods to determine the moisture conditions in concrete.⁽¹²¹⁾ These include destructive methods, RH, resistivity, dielectric properties, thermal properties, infrared absorption, and neutron scattering. Most of the methods do not appear to be suitable for monitoring of large amounts of concrete pavement, but one method, infrared absorption, appears to have potential. Water has characteristic infrared absorption wavelengths that can be used as an indicator for the amount of water present. Equipment exists that can illuminate an area of a surface with an infrared beam of the required wavelength, the reflected light is sensed and the moisture content of that surface analyzed. Not discussed by Parrott but plausible, is the possibility that curing compounds have an infrared spectrum that can be used as the basis for measuring membrane thickness.⁽¹²¹⁾

Another, infrared method that could be used for concrete curing is infrared thermography. This method uses an infrared sensitive video camera image of a concrete surface. Moisture conditions affect heat flow through the surface, resulting in variations in surface temperature that could be sensed by the infrared camera. This method would appear to have application in analyzing the quality of cover of curing compounds.

Al-Manaseer and Aquino (1999)⁽¹⁸⁷⁾ reported an evaluation of the Windsor probe (ASTM C 803⁽⁹⁵⁾) for in situ measurement of concrete strength. They reported good results, but the method seems to rely on properties of the concrete considerably deeper that the near-surface zone, which is critical to evaluating curing. The method appears to have value for estimating in place strength, but probably has less value as a method for evaluating the near-surface zone.

White and Husbands (1990) investigated several methods for evaluating the effectiveness of curing compounds, including capillary porosity, water absorption, abrasion resistance, chemically combined water, and splitting tensile strength.⁽⁸⁵⁾ Abrasion resistance, capillary porosity, and water absorption appear to be the best from a sensitivity and precision viewpoint. Water absorption was found to be the most practical.

Maturity

Strength gain in concrete is a time-temperature dependent phenomenon commonly called maturity. The bulk of concrete testing and concrete research is conducted under standard temperature conditions, usually about 23 °C. Estimating strength nondestructively under nonstandard temperature conditions may be of considerable practical value in actual construction. If some fundamental properties of a concrete are known, the rate of strength gain can be represented by a mathematical model that allows incorporation of nonstandard temperature conditions if the time-temperature history is known. ASTM C 1074 is a standard test method based on this principle, called the maturity method.⁽³⁹⁾

Two basic approaches are described in C 1074.⁽³⁹⁾ One is based on a simple integration of the time-temperature curve to give a measure of maturity in units of degree-days. Maturity values are then calibrated to strength estimates using a concrete mixture similar to the job mixture. The maturity model is called the Nurse-Saul model. The Arrhenius model is also used in C 1074.⁽³⁹⁾ This model is based on the chemical kinetic model of that name, which represents the dependency of simple chemical reaction rate constants on temperature. The rate of strength gain is treated as though it were the rate constant of a simple chemical reaction.

There is considerable literature on determining the mathematical representations of the maturity phenomenon, and on verifying the accuracy of the method (Carino, 1984,⁽¹⁸⁸⁾ Parsons and Naik, 1985;⁽¹⁸⁹⁾ Chengju, 1989;⁽¹⁹⁰⁾ Malhotra and Carino, 1991;⁽¹⁹¹⁾ Carino, Knab, and Clifton, 1992;⁽¹⁹²⁾ Carino and Tank, 1992;⁽¹⁹³⁾ Kjellsen and Detwiler, 1993⁽¹⁹⁴⁾). Grove et al. (1990)⁽¹⁶⁵⁾ and Okamoto and Whiting (1994)⁽³⁶⁾ demonstrated the application of this technique for predicting strength development in fast-track paving. The method required some adjustments to the calibration procedure to account for the effects of the early temperature rise of the concrete. Pavements were placed at approximately 30 °C, but reached temperatures of 50-60 °C during the first 8 h. Hover (1992)⁽¹⁹⁵⁾ reported on use of this method in cold weather concreting to help predict strength development in structural concrete.

The maturity method might also be used to calculate the length of curing necessary under some anticipated time-temperature regimen.

Concrete Temperature

For purposes of protection of newly placed concrete from temperature or temperature-gradient extremes, it is of considerable practical value to be able to anticipate project concrete temperature from properties of the concrete materials and proportions and from anticipated ambient conditions. These properties and conditions include heat of hydration of cement, composition of the concrete, dimensions of the structure, and heat transfer rates to and from the surrounding environment. Kapila et al. (1997) report one mathematical development for this kind of analysis and their model predicted temperature-time measurements on a bridge deck placement very well.⁽⁶⁰⁾ These same concepts are being developed into a computer program, called HIPERPAV^TM, under a contract with FHWA.⁽¹³⁾ This program allows the user to input information on concrete materials, mixture proportions, environmental conditions, and design details of a pavement, and then to have output on strength gain and tensile stresses associated with thermal effects to predict cracking. In its final form, the program will also incorporate effects of curing procedures on temperature.

Temperature gradients can be the cause or contribute to cracking in concrete. As reviewed under standard guidance, there is guidance on the types of gradients that can usually be tolerated in concrete. These values are typically in the range of 10 °C to 13 °C between the concrete surface and the interior (e.g., 50 mm). Lykke et al. (2000) report that values in the range of 15 °C to 20 °C can be tolerated in pavements.⁽¹⁹⁶⁾

It is well known that the temperature history of concrete affects the early and the ultimate strength of concrete in opposite ways. High temperatures increase early strength and decrease ultimate strength. There are some data on this in ACI 305R on hot weather concreting.⁽¹²⁾ Kim et al. (1998) simulated some high early-temperature events and found that concrete cured at 40 °C for the first 24 h developed strengths 2.4 times that at 20 °C, and that the 28-day strengths were about 13 percent lower.⁽¹⁹⁷⁾

New Technology

Yu et al. (1998) investigated the effect of "electron water" on rate of strength development of concrete.⁽¹⁹⁸⁾ Electron water is produced by charging electricity through water in a tank for several hours or days. Such water has been found to have higher activity in biological systems, as measured by growth rates. Strength increases of 10 to 20 percent in concretes cured with electron water were observed, compared to curing in normal water. The effect is attributed to changes in the way water molecules cluster. This is reminiscent of the literature from the Soviet Union in the 1960s and 1970s about the higher concrete strengths produced by magnetic water, i.e., water that has passed through a strong magnetic field before being used as mixing water in concrete.

Weber and Reinhardt (1997),⁽¹⁹⁹⁾ Reinhardt and Weber (1998),⁽²⁰⁰⁾ and Bentur et al. (2001)⁽²⁰¹⁾ investigated use of saturated lightweight aggregate as a reservoir for curing water in low water-cement ratio concretes. The intention was to overcome the problem of getting added water into such concretes, which tend to internally desiccate due to consumption of mixing water by hydration reactions. There was evidence that this technique resulted in continued hydration through 1 year. This technology was used when more water was calculated to be needed to produce the required expansion when Type K expansive cement was used for tunnel plugs than could be tolerated as water-cement-ratio water. The water needed beyond that allowed by selected water-cement ratio was introduced as absorbed water in the low-density aggregate.

Several years ago, a series of papers was published describing performance of a curing admixture (Dhir et al. 1994;⁽²⁰²⁾ Dhir, Hewlett, and Dyer 1995,⁽²⁰³⁾ 1996;⁽²⁰⁴⁾ Klapperich, Potter, and Willocq, 1995⁽²⁰⁵⁾). This admixture acts to reduce the vapor pressure of water by forming hydrogen bonds with the water molecules in the liquid phase, thus reducing evaporation rates. Six products were included in the investigation. Performance with four of the products was not equivalent to water or curing compound curing, but two products produced almost equivalent results. Some products caused an acceleration of cement hydration. One resulted in reduced strength, apparently because it interfered with the formation of calcium hydroxide, which normally comprises the predominant material in the cement-aggregate interfacial zone. No later literature, including commercial literature, was found on this technology.

Knutson and Riley (1988) briefly mention that the ACPA was investigating use of a biodegradable foam for use as a curing material in cooler weather.⁽¹⁶⁴⁾ This material would have an advantage in that it could be washed away after concrete had reached adequate strength.