Abstract
This paper analyzed the air-void system in air-entrained concrete in the presence of calcium salts and developed a new approach method, random “parking” method, for studying it. The foam index test studied indicated that the introduction of calcium salts did result in a drop in foaming power at 0.5 min, but the foam ability after 60 min was enhanced in surfactant-water solution. The Powers spacing factor in the air-entrained hardened concrete decreased gradually in the presence of CaCl2 or Ca(NO3)2. The probability density function of air-void distribution showed that the air-void size in air-entraining concrete was reduced with the addition of CaCl2 and Ca(NO3)2. The random parking method developed herein was compared with the experimental Powers spacing factor, and the concept of the alternative spacing factor, defined as the average distance between the surfaces of the two nearest air voids in concrete, was used. It was indicated that the alternative spacing factor is in close proximity to the experimental Powers spacing factor.
1 Introduction
Air-entrained concrete has been becoming one of the most consumed composite materials used in cold regions all over the world. Since 1937, entrained air-void systems have consisted of a large number of uniform and stable small air voids, with pore sizes ranging from tens to hundreds of microns being introduced into the concrete mix to protect it against frost damage. A typical concrete without air-entraining admixtures (AEAs) contains about 1–2% volume of harmful and inevitable entrapped air voids during mixing. These air voids are different from the entraining air voids in both size and shape. Although entrapped air voids are detrimental to the durability of concrete, entraining air voids play a significant role in improving the resistance of concrete to damage by repeated exposure to freeze-thaw cycles and salt scaling.
CaCl2 is the most well-known accelerator for Portland cement, while Ca(NO3)2 is the most commonly used chloride-free set accelerating admixture. The influence of CaCl2 and Ca(NO3)2 on the hydration of cement has been the subject of a number of investigations [1–5]. It has been reported that these inorganic salts can increase the capillary pore size in the cementitious matrix, which acts to limit the severity of the freeze-thaw cycle and the deicing salt resistance of concrete [6]. In practice, the acceleration of the hydration reaction effect and the presence of air-entrained voids are usually considered to improve the workability and the durability of concrete in cold-weather construction. Therefore, it is essential to study the effect of inorganic salts on the air-entraining process in concrete.
Traditionally, it is assumed that the spacing factor proposed by Powers [7] is an appropriate way to describe the air-void structure. The Powers spacing factor (L̅) is an estimate of the longest distance to an air-void surface. The air-void spacing factor (L̅) is one of the most important factors in evaluating the freeze-thaw cycle and the deicing salt resistance of concrete. Much research has been conducted on the freeze-thaw mechanism and deicing salt resistance of concrete [8–11]. Recently, digital image techniques have been increasingly applied to investigate the parameters of the air-void system in hardened concrete [12–14], including the specific surface, air content, and spacing factor. The use of automated digital imaging techniques has several advantages: (1) they are a highly effective measurement technique as they do not require the operator to measure the air voids individually, thereby reducing operator time; (2) they can be easily repeated; and (3) operator bias is also an issue that is mitigated by automated techniques.
Given the significant interest in studying the effect of accelerators CaCl2 and Ca(NO3)2 on the air-void system analysis in air-entraining concrete, this paper presents a new statistical numerical simulation for the analysis of the air-void system in air-entrained concrete in the presence of calcium salts through the use of the two-dimensional random “parking” method [15, 16]. The results from the experimental study performed in accordance with ASTM C457 are incorporated into a statistically numerical simulation that extends Powers’ original model. The formulation provides a theoretical foundation for understanding the effect of accelerators, such as CaCl2 and Ca(NO3)2, on the air-void system in concrete. Moreover, the relationship between experiment and numerical simulation is established to evaluate the air-void system in concrete.
2 Materials and experimental methods
2.1 Materials
This study used Chinese-grade 42.5 Portland cement (Harbin Swan, Harbin, Heilongjiang Province, China) with a density of 3150 kg/m3 (initial setting of 195 min and final setting of 320 min) and siliceous shale with a continuous grading of 5–19 mm as the coarse aggregate (apparent density of 2700 kg/m3). The fine aggregate used was medium sand (fineness modulus of 2.3), with an apparent density of 2620 kg/m3. Sieve analysis of fine and coarse aggregates was performed, and the fitting curves were calculated (see Figure 1 for the numerical simulation). The mix proportions are shown in Table 1 and also the dosage of salts and AEAs. Analytical CaCl2 and Ca(NO3)2 were purchased from (Tianjin Kaitong, Tianjin, China). The different amounts of CaCl2 and Ca(NO3)2 incorporated into the mixes were labeled. The amount of Vinsol resin anionic AEA (Vinsol resin, Heibei, China) was 0.3‰ by mass of cement. In addition, 0.5% UNF-5 (naphthalene-based superplasticizer; Tianjin, China) was added to the concrete mix to control the workability. Deionized water was used. The concretes studied were prepared in accordance with Table 2.
A) Sieve analysis of coarse aggregates; B) sieve analysis of fine aggregates.
Usage of salts and AEAs.
| Samples | Inorganic salts (%, by mass of cement) | AEAs (‰, by mass of cement) | Molecular formula of Vinsol resin | |
|---|---|---|---|---|
| CaCl2 (A) | Ca(NO3)2 (B) | Vinsol resin (H) | ||
| H | – | – | 0.3 |  |
| HA0.5 | 0.5 | – | ||
| HA1.0 | 1 | – | ||
| HA2.0 | 2 | – | ||
| HB1.0 | – | 1 | ||
| HB2.0 | – | 2 | ||
| HB3.0 | – | 4 | ||
Mixture proportion of concretes.
| W/C | Water (kg/m3) | Cement (kg/m3) | Sand (kg/m3) | Crushed stone (kg/m3) | Superplasticizer UNF-5 (%, by mass of cement) |
|---|---|---|---|---|---|
| 0.4 | 170 | 425 | 630 | 1120 | 0.5 |
W/C represents water-to-cement ratio.
2.2 Experimental methods
2.2.1 Foam index test
The foam index test is used to predict the AEA-calcium salt compatibility [17]. The test procedure was modified for this work. We fixed the amount of surfactant, but varied the amount of either CaCl2 or Ca(NO3)2. The initial volume of the foam after agitation ceased indicated the foaming power, while the changes in the remaining volume after 3, 5, 30, and 60 min represented the stability of the foam. This modified foam index test was considered more applicable because it allowed a better comparison between the AEA with and without the effect of different concentrations of calcium salts.
The procedure for the modified foam index test was as follows:
One thousand milliliters of surfactant-water solution at 0.05% by the mass of water was prepared.
Fifty milliliters of the surfactant-water solution was poured into a 250-ml graduated cylinder (Harbin, China). An amount of 0.5%, 1.0%,1.5%, 2.0%, 2.5%, or 3.0% of either CaCl2 or Ca(NO3)2 was added when necessary.
The graduated cylinder was capped and shaken upside down continuously for 20 times. The initial volumes were recorded.
The cylinder was laid on its side for 3, 5, 30, and 60 min. The corresponding volumes were recorded.
2.2.2 Microscopic determination of the Powers spacing factor
The concretes studied were tested in accordance with the following ASTM standards: ASTM C231 for the fresh air content and ASTM C457 for the parameters of the air-void system. The samples were cut into 10-mm-wide slices and polished with successively finer abrasives until suitable for testing in accordance with ASTM C457 [18]. A microscope (HXS-1000A, Shanghai Shangguang, China) was used to determine the parameters of the air-void system in hardened concrete. Also, the air-void spacing factor was calculated by the linear traverse method. However, the diameter of air voids crossed by the linear traverse was recorded to study the distribution of air voids and to use it as the basis of the numerical model. The effect of CaCl2 or Ca(NO3)2 on the air-entraining process of Vinsol resin in hardened concrete was investigated by comparing with L̅.
2.2.3 Random parking method
The mixture proportion of concrete, diameter distribution of fine and coarse aggregates, air content, and diameter distribution of air voids were selected as the parameters of the simulation process. During recording, the chord length of the air voids according to ASTM C457 and the diameter of these air voids (Dn) were also recorded one by one. Note that the air voids and aggregates were assumed to have circular particles except for the filling cement phase, which is composed of unhydrated cement particles and hydration products. The random parking method is described in detail in Figure 2. Random locations were generated for the centers of the circles by a computer, and they were placed inside a square 100 mm long on a side. All the particles including the air voids and aggregates were sorted in order of decreasing radii and the other was filled by the cement paste, describing it as “parking.” This parking approach was used until all the circles were placed into the cube, allowing the computer to determine the size and location of every circle in the system. The computer could now calculate any desired measure of spacing. The algorithm mentioned above was encoded by the C++ programing language.
Flow chart for numerical simulation about air-entraining process. (Note that D represents the center distance between two particles and rn is radius of particle n.)
In this case, the information of the nth air void was compared to the others to find the nearest distance of a neighboring air void and to quantify the average distance (S̅N) between the surfaces of the two nearest air voids rather than that between the cores (S̅C). Note that, for example, the nearest to the 10th air void is the 502nd air void, but the nearest to the 502nd air void is not necessarily the 10th air void. Based on the simulated two-dimensional (2D) image of the distribution of air voids in concrete, the air-void spacing factor L̅S was calculated by the linear traverse method of ASTM C457. To further assess the validity of the approach, real air-entrained concretes were analyzed. The results are described subsequently. Significantly, the random parking method may provide a likelihood for characterizing the air-void system in practice.
3 Results and discussion
3.1 Foam index test
The foaming of air bubbles is a process of increasing systemic surface free energy; therefore, air bubbles are unstable thermodynamically and have a certain lifetime. However, when AEAs are used in the solution, the surface tension is reduced by a substantial amount [19]; in other words, systemic surface free energy decreases, so foaming forms more easily. In solution, the stability of foams is affected primarily by the surface tension, surface viscosity, Gibbs-Marangoni effect of films, and air diffusion between bubbles. In addition, there are plenty of environmental and experimental factors in concrete, such us the mixing action and mixture proportions, the role of paste, the temperature, as well as the quality of mixing water [20]. The foam index test indicated the effect of calcium salt on foam stability in surfactant-water solution.
As shown in Figure 3, the initial volume (V0) of foam at 0.5 min represents the foaming power, while the difference between the volume of 0.5 and 60 min (ΔV) is defined as the foam stability. Obviously, Vinsol resin improves the foaming power and the foam stability substantially. This performance, however, can be affected by interfusing the calcium salts. The introduction of calcium salts did result in a drop in foaming power at 0.5 min, but ΔV decreased, suggesting that the foam ability was enhanced. Specifically, V0 dropped from 46 to 15 ml with increasing CaCl2 concentration, while that of Ca(NO3)2 decreased to around 18 ml. Based on ΔV, the foam stability with CaCl2 or Ca(NO3)2 was enhanced, ranging from approx. 15 to 7 ml. Initially, depending on the increasing concentration of CaCl2 or Ca(NO3)2, the surface tension of salt solutions increased linearly [21], decreasing the foaming power. In the presence of either CaCl2 or Ca(NO3)2, very little foam was produced after 20 times of shaking, suggesting that there was weak foaming power, but the foam stability of Vinsol resin generally remained almost unchanged over time or slightly increased.
Modified foam index test results: A) CaCl2; B) Ca(NO3)2.
Although Vinsol resin lowers systemic surface free energy, resulting in the easier formation of air bubbles, air bubbles are still unstable thermodynamically and have a certain lifetime. Therefore, the air bubbles could rupture gradually with time in solution. But air bubbles can be kept inside the concrete after setting. In contrast, because the main composition of Vinsol resin is an anionic surfactant, the Vinsol resin sol solution had a negative charge. The Vinsol resin was precipitated by Ca2+ [20, 22–24]; the effective air-entraining component was reduced, which decreased the foaming power of the Vinsol resin. However, these flocculent precipitates absorbed in films enhanced the strength of the bubbles, improving the foam stability slightly [24], explaining why the higher the concentration of CaCl2 or Ca(NO3)2, the greater the decrease in foam power. Interestingly, the surfactant-water solution immediately produced plenty of flocculently white precipitated substance upon the addition of calcium salt.
3.2 Microscopic determination of the parameters of the air-void system in hardened concrete
Compared to studies previously described, the probability density function was used to analyze the air-void distribution in concrete, while the parameters of the air-void system in hardened concrete were also determined microscopically. Based on the records of the diameters of the air voids in the microscopic field of all the lines, the air-void distribution was calculated statistically by the number percent corresponding to one diameter of air void using the probability density function. As shown in Figure 4, the distribution with CaCl2 shifted to the left, ranging from 176 to 85 μm, while the distribution with Ca(NO3)2 decreased to around 150 μm compared to the value of 180 μm in absence of CaCl2 or Ca(NO3)2. In other words, either CaCl2 or Ca(NO3)2 is capable of decreasing the pore size of the air voids in the system during the air-entraining process. These results are in agreement with subsequently measured parameters of the air-void system in hardened concrete. As the concentration of CaCl2 or Ca(NO3)2 increased, the surface tension of salt solutions increased linearly [21]. The Vinsol resin was precipitated by Ca2+ [20, 22–24]; the effective air-entraining component was reduced, which decreased the foaming power of the Vinsol resin. However, these flocculent precipitates absorbed in films enhanced the strength of the bubbles, improving the foam stability slightly [24]. In contrast, these precipitates can delay the growth and the coalescence of air bubbles before setting in concrete [20]. Therefore, the introduction of CaCl2 and Ca(NO3)2 can make the air-void distribution narrower.
![Figure 4: The air-void distribution with CaCl2 (left) and Ca(NO3)2 (right) in air-entraining concrete. [*The capital letters are representing the type of surfactant and calcium salts in concrete: H-Vinsol resin, A-CaCl2, B-Ca(NO3)2, while the numbers are representing the dosage of corresponding calcium salt by mass of cement in concrete.]](/document/doi/10.1515/secm-2014-0380/asset/graphic/j_secm-2014-0380_fig_004.jpg)
The air-void distribution with CaCl2 (left) and Ca(NO3)2 (right) in air-entraining concrete. [*The capital letters are representing the type of surfactant and calcium salts in concrete: H-Vinsol resin, A-CaCl2, B-Ca(NO3)2, while the numbers are representing the dosage of corresponding calcium salt by mass of cement in concrete.]
ASTM C457 [18] stipulates that the maximum value of the spacing factor for moderate exposure of the concrete be taken to be 0.20 mm. Somewhat smaller values may be required for severe exposure, especially if the concrete is in contact with deicing chemicals; larger values may be adequate for mild exposure [18]. Note that, for a given air content in concrete, the larger the air-void diameter, the larger the spacing factor and specific surface area, and the lower the freeze-thaw resistance [20].
As shown in Table 3, the average chord length, average radius, and air content of the air-entrained hardened concrete decreased by different rates in the presence of CaCl2 or Ca(NO3)2, resulting in a decrease in the spacing factor. In this study, Vinsol resin and the calcium salt were blended separately during mixing in order to prevent them from coagulating. The decreasing effect of CaCl2 on the spacing factor was higher than that of Ca(NO3)2. Given these results, it was determined that neither CaCl2 nor Ca(NO3)2 interfered with the air-entraining treatment in concrete. That said, the Vinsol resin AEA could be precipitated by calcium cation in solution [22–24], forming a kind of fluffy insoluble substance that decreases the effectiveness of air-entraining composites. Ca2+ is the top diffuser of cations, while NO3- is located at the penultimate position, ranked in order of the effectiveness as accelerators [25–27]. Therefore, calcium nitrate can affect the air-void structure to some extent, but it is less effective than CaCl2. In addition, the effect of calcium salt on the air-void system in hardened concrete primarily depends on the usage of calcium salt.
Parameters of the air-void system in hardened concrete.
| No.a | Fresh air content, Af (%) | Average chord length, l̅ (mm) | Average radius, r (mm) | Specific surface, α (mm2/mm3) | Hardened air content, Ah (%) | Spacing factor, L̅ (mm) |
|---|---|---|---|---|---|---|
| H | 4.5 | 0.175 | 0.131 | 22.9 | 5.73 | 0.209 |
| HA0.5 | 5.5 | 0.154 | 0.115 | 26.0 | 4.32 | 0.201 |
| HA1.0 | 4.8 | 0.141 | 0.106 | 28.3 | 3.61 | 0.198 |
| HA2.0 | 3.4 | 0.117 | 0.088 | 34.2 | 4.36 | 0.159 |
| HB1.0 | 3.7 | 0.137 | 0.103 | 29.3 | 3.95 | 0.187 |
| HB2.0 | 4.3 | 0.150 | 0.113 | 26.6 | 3.75 | 0.197 |
| HB4.0 | 4.7 | 0.137 | 0.103 | 29.1 | 3.67 | 0.192 |
aThe capital letters represent the type of surfactant and calcium salts in the concrete: H, Vinsol resin; A, CaCl2; B, Ca(NO3)2; the numbers represent the dosage of the corresponding calcium salt by mass of cement in concrete.
It is widely accepted that the molecule of anionic AEA on the surface of an air void is precipitated by calcium cations in solution [22–24]. These chemically precipitated products possibly form a shell with some mechanical properties on the surface of the air void, stabilizing the air void [17, 19, 24, 28]. The rate of precipitated reaction is much faster than the hydration of cement particles, thereby forming a distinct shell around the air void where the unhydrated cement particles may be wrapped (probably by the precipitated substance). Subsequently, some hydration products begin to grow into air void. Most likely, the formation of the protective shell in fresh paste does not grow, coalesce, and/or break up in the air bubbles until the concrete sets; therefore, the air-void distribution size in air-entrained hardened concrete has a tendency to decrease.
The Powers spacing factor attempts to calculate the fraction of paste within some distance of an air void (paste-void proximity); many researchers [29–32] have made some modifications to improve this calculation. Air-entraining concrete usually contains hydration products, aggregates, unhydrated cement, and air voids. A modified numerical model that provides a new approach for calculating the air-void system including the effect of aggregates and that characterizes the real air-void system in concrete is desirable. It is significant to evaluate the relationship between experimental studies and simulated studies. The next section presents the results of a numerical model developed to calculate the air-void system compared to the experimental results of the Powers spacing factor for the same system.
3.3 Random parking method of the air-entraining process
The previous microscopic determination is a one-dimensional statistical method; however, the random parking method can provide a 2D image at a scale of 100×100 mm, including the information of all phases in concrete. The images include the air-void distribution based on the experimental information about the distribution and content of air voids, aggregates, and cement paste. The 2D simulation is a more thorough technique for characterizing the air-void system of concrete. As shown in Figure 5, the image represents the surface of a 100×100-mm section of concrete; one pixel was assigned to 10 μm. Based on these pieces of information about the 2D images, the alternative spacing factor (S̅N), which is the average value for distances between the surfaces of the nth void and its closest neighboring air voids, while the simulated Powers spacing factor (L̅S) was calculated from these simulated 2D images according to the linear traverse method of ASTM C457.
Numerical simulation of hardened concrete with Vinsol resin air-entraining admixtures (white-air voids, blue-fine aggregates, orange-coarse aggregate, and black-cement paste).
The alternative and the simulated values are considerably higher than that predicted by the Powers spacing factor in Figure 6. It is possible that the wall effect of aggregates during the simulation increases the spacing factor. In contrast, the Powers spacing factor assumes that the pore size of air voids is homogeneous in space, when, in fact, the air-void distribution of air voids in concrete is heterogeneous. Although the Powers spacing factors are <0.2 mm, the freeze/thaw cycles of partial specimens are not capable of improving the freeze/thaw resistance [30, 33]. Therefore, the magnitude of air-void spacing alone does not sufficiently characterize the air-void system for the purposes of freeze-thaw durability analysis [34]. Based on the results of the numerical simulation, the alternative spacing factor is a reliable way to evaluate the air-void distribution, while the simulated Powers spacing factors are completely higher by approx. 0.075 mm than the alternative spacing factor probably because of the wall effect of the aggregates. Although the wall effect of the aggregates resulted in an increase in the spacing factor, the random parking method can also evaluate the air-void distribution in concrete and characterize the changes with the addition of different concentrations of calcium salts.
Effect of two calcium salts, respectively, on various spacing factors: (A) CaCl2; (B) Ca(NO3)2.
4 Conclusions
This paper presents the results of an experimental study to determine the effect of calcium salts on the air-void distribution of air-entraining concrete. Based on these results, the random parking method was used to study the air-void system.
In surfactant-water solution, the introduction of calcium salts did result in a drop in foaming power at 0.5 min, but ΔV decreased, suggesting that the foam ability was enhanced.
Either CaCl2 or Ca(NO3)2 is capable of decreasing the pore size of air voids during the air-entraining process in concrete in the presence of Vinsol resin.
In air-entrained hardened concrete, adding CaCl2 or Ca(NO3)2 in conjunction with Vinsol resin decreased gradually the Powers spacing factor. However, further studies are needed to determine the frozen-thaw durability of air-entraining concrete in the presence of calcium salts.
The random parking method developed herein is a new approach for evaluating the spacing factor in concrete. The alternative spacing factor (S̅N), which is the average distance between the surfaces of the two nearest air voids, could also characterize the parameters of the air-void system compared with the Powers spacing factor. It was shown that the alternative spacing factor was in close proximity to the experimental Powers spacing factor.
In order to ensure an accuracy of 10 μm per pixel, future work should consider improving the algorithm in three-dimensional random parking.
Acknowledgments
Funding for this project was provided by the China National Science Foundation (no. 51278157). The authors also gratefully acknowledge the help of G. Geng and R. Taylor.
References
[1] Abdelrazig BEI, Bonner DG, Nowell DV, Egan PJ, Dransfield JM. Thermochim. Acta 1989, 145, 203–217.10.1016/0040-6031(89)85140-8Search in Google Scholar
[2] Juenger MCG, Monteiro PJM, Gartner EM, Denbeaux GP. Cem. Concr. Res. 2005, 35, 19–25.10.1016/j.cemconres.2004.05.016Search in Google Scholar
[3] Rosenberg AM. J. Am. Conc. Inst. Proc. 1964, 61, 1261–1268.Search in Google Scholar
[4] Skalny J, Odler I, Hagymassy Jr. J. J. Colloid Interface Sci. 1971, 35, 434–440.10.1016/0021-9797(71)90143-3Search in Google Scholar
[5] Young JF. Cem. Concr. Res. 1972, 2, 415–433.10.1016/0008-8846(72)90057-9Search in Google Scholar
[6] Yang WC, Ge Y, Yuan J, Zhang BS. J. Chin. Ceram. Soc. 2009, 4, 622–626.Search in Google Scholar
[7] Powers TC. J. Am. Conc. Inst. Proc. 1945, 41, 245–272.Search in Google Scholar
[8] Browne FP, Cady PD. ACI Spec. Publ. 1975, 47, 101–120.Search in Google Scholar
[9] Çopuroğlu O, Schlangen E. Cem. Concr. Res. 2008, 38, 27–39.10.1016/j.cemconres.2007.09.003Search in Google Scholar
[10] Foy C, Pigeon M, Banthia N. Cem. Concr. Res. 1988, 18, 604–614.10.1016/0008-8846(88)90053-1Search in Google Scholar
[11] Pigeon M, Lachance M. J. Am. Conc. Inst.Proc. 1981, 78, 282–201.Search in Google Scholar
[12] Jakobsen UH, Pade C, Thaulow N, Brown D, Sahu S, Magnusson O, De Buck S, De Schutter G. Cem. Concr. Res. 2006, 36, 1444–1452.10.1016/j.cemconres.2006.03.005Search in Google Scholar
[13] Peterson K, Carlson J, Sutter L, Van Dam T. Mater. Charact. 2009, 60, 710–715.10.1016/j.matchar.2008.10.005Search in Google Scholar
[14] Zalocha D, Kasperkiewicz J. Cem. Concr. Res. 2005, 35, 2041–2046.10.1016/j.cemconres.2005.05.004Search in Google Scholar
[15] Bentz DP. J. Am. Ceram. Soc. 1997, 80, 3–21.10.1111/j.1151-2916.1997.tb02785.xSearch in Google Scholar
[16] Bentz DP. Mater. Struct. 1999, 32, 187–195.10.1007/BF02481514Search in Google Scholar
[17] Corr DJ, Lebourgeois J, Monteiro PJM, Bastacky SJ, Gartner EM. Cem. Concr. Res. 2002, 32, 1025–1031.10.1016/S0008-8846(01)00751-7Search in Google Scholar
[18] C457/C457M-12, ASTM International: West Conshohocken, PA, 2012.Search in Google Scholar
[19] Atahan HN, Carlos C, Chae S, Monteiro PJM, Bastacky J. Cem. Concr. Compos. 2008, 30, 566–575.10.1016/j.cemconcomp.2008.02.003Search in Google Scholar
[20] Du L, Folliard KJ. Cem. Concr. Res. 2005, 35, 1463–1471.10.1016/j.cemconres.2004.07.026Search in Google Scholar
[21] Ge Y, Chang CL, Yang WC, Zhang BS, Yuan J, Yu JS. Concrete 2007, 6, 7–9.Search in Google Scholar
[22] Chatterji S. Cem. Concr. Compos. 2003, 25, 759–765.10.1016/S0958-9465(02)00099-9Search in Google Scholar
[23] Mielenz RC, Wolkodoff VE, Backstrom JE, Flack HL. ACI J. Proc. 1958, 55, 95–121.Search in Google Scholar
[24] Ley MT, Chancey R, Juenger MCG, Folliard KJ. Cem. Concr. Res. 2009, 39, 417–425.10.1016/j.cemconres.2009.01.018Search in Google Scholar
[25] Kantro DL. ASTM J. Test. Eval. 1975, 3, 312–321.10.1520/JTE10661JSearch in Google Scholar
[26] Double DD, Hewlett PC, Sing KSW, Raffle JF. Phil. Trans. R. Soc. Lond. A. 1983, 310, 53–66.10.1098/rsta.1983.0065Search in Google Scholar
[27] Wilding CR, Walter A, Double DD. Cem. Concr. Res. 1984, 14, 185–194.10.1016/0008-8846(84)90103-0Search in Google Scholar
[28] Monteiro PJ, Coussy O, Silva DA. ACI Mater. J. 2006, 103, 136–140.Search in Google Scholar
[29] Attiogbe EK. ACI Mater. J. 1997, 94, 588–591.Search in Google Scholar
[30] Mayercsik NP, Felice R, Ley MT, Kurtis KE. Cem. Concr. Res. 2014, 59, 16–23.10.1016/j.cemconres.2014.01.023Search in Google Scholar
[31] Philleo RE. Cem. Concr. Aggr. 1983, 5, 128–130.10.1520/CCA10263JSearch in Google Scholar
[32] Pleau R, Pigeon M. Cem. Concr. Aggr. 1997, 5, 128–130.Search in Google Scholar
[33] Jin S, Zhang J, Huang B. Constr. Build. Mater. 2013, 47, 126–130.10.1016/j.conbuildmat.2013.04.040Search in Google Scholar
[34] Attiogbe EK. ACI Mater. J. 1996, 93.Search in Google Scholar
©2017 Walter de Gruyter GmbH, Berlin/Boston
This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
