Conventional Portland Cement and Carbonated Calcium Silicate ...

Conventional Portland Cement and Carbonated Calcium Silicate–Based Cement Systems Performance During Freezing and Thawing in Presence of Calcium Chloride Deicing Salts Chiara Villani, Yaghoob Farnam, Taylor Washington, Jitendra Jain, and W. Jason Weiss concrete occur during the temperature reduction associated with the freeze–thaw cycle causing ice to form, which expands, resulting in hydraulic pressure (4, 5), osmotic pressure (6–8), or crystallization pressure (6, 9) and can result in the development of stress, cracking, and spalling. The chemical effects include the detrimental reactions between the deicers and the cement paste (10–12), enhancement of the aggregate–cement reactions (13), or corrosion when reinforcing steel is present in the concrete (14). Calcium chloride (CaCl2) is one deicing salt increasingly used owing to its ability to provide ice melting capabilities at relatively low temperatures (10, 14). However, the use of CaCl2 can result in damage to the concrete since it can leach calcium hydroxide and ettringite from the cement matrix (15). Additionally, the CaCl2 can result in a series of chemical reactions (10–12) causing strength losses of up to 80% (10). The chemical reactions between cementitious materials and CaCl2 can result in the formation of aluminate phases (Friedel’s salt), chloro-sulfo-aluminate phases (Kuzel’s salt) (16), or calcium oxychloride phases (17, 18). Each of these phases is described in detail in the following paragraphs. Friedel’s salt can form when the CaCl2 either reacts with the calcium aluminate phases and water as shown in Equation 1 (19) or when it reacts with monosulfate (AFm) as shown in Equation 2 (20).

The behavior of two cementitious materials during thermal changes associated with freezing and thawing in presence of calcium chloride deicing salts was examined. The two systems consisted of a conventional portland cement-based material and an alternative economically friendly cement that formed a solid by carbonating a calcium silicate–based cement. Lowtemperature differential scanning calorimetry was used to quantify the phase changes associated with ice formation, eutectic solution transformation, and calcium oxychloride formation. Longitudinal guarded comparative calorimetry was used to detect the damage that developed as a result of the expansive pressures created by these phases when they form. In both systems exposed to low salt concentration, the damage was primarily caused by hydraulic and osmotic pressure. This type of damage was moderate at low degrees of saturation (e.g., 12 (26)]. 0.5 CaCl 2 + 3CaO i Al 2 O3 i CaSO 4 i 12H 2 O → 3CaO i Al 2 O3 i

0.5CaSO 4 i 0.5CaCl 2 i 11H 2 O + 0.5 ( CaSO 4 i 2H 2 O) (3)

where 3CaO • Al2O3 • 0.5CaSO4 • 0.5CaCl2 • 11H2O is Kuzel’s salt. In systems containing calcium hydroxide, the formation of calcium oxychloride can also be expected; it derives from the reaction between CaCl2 and calcium hydroxide in presence of water as shown in Equation 4. Calcium oxychloride formation generally results in a large wellcrystallized needle structure, which is believed expansive and consequently deleterious for integrity of the cementitious materials (10–12, 17, 27, 28). The stability of calcium oxychloride appears easily altered by changes in temperature and moisture content (17, 29–31). 3Ca ( OH )2 + CaCl 2 + 12H 2 O → CaCl 2 i 3Ca ( OH )2 i 12H 2 O

(4)

where Ca(OH)2 is calcium hydroxide and CaCl2 • 3Ca(OH)2 • 12H2O is calcium oxychloride. Equations 1 through 4 show that tricalcium aluminate AFm, and calcium hydroxide are important phases that can react in the presence of CaCl2. As a result, it would appear that the chemistry of cementitious materials can be altered to develop systems that are less reactive in the presence of CaCl2. For example, if a system contained a reduced or negligible amount of calcium hydroxide (i.e., carbonated systems, systems containing supplementary cementitious materials), the addition of CaCl2 result in reduced or negligible formation of calcium oxychloride formation (10). Recent studies by Farnam et al. have suggested that the formation of calcium oxychloride is a chemical phase transition that can be characterized using low-temperature differential scanning calorimetry (LT-DSC) and whose damage can be quantified through longitudinal guarded comparative calorimetry (LGCC) (11, 12). Results obtained with these test methods showed that the temperature at which calcium oxychloride forms in a conventional mortar resulted above 0°C and is primarily dependent on the CaCl2 concentration. At high salt concentrations (>15%), the damage caused by the formation of calcium oxychloride is considerable and occurs even if concrete does not experience freeze–thaw cycles (11, 12). This paper will examine the performance of two cementitious systems that are exposed to CaCl2 solutions of different concentrations while undergoing temperature changes associated with freezing and thawing. The first system will be made using a conventional ordinary portland cement (OPC), and the second system will be made using a new environmentally friendly calcium silicate–based cement (CSC) that forms a solid through the carbonation process (32, 33). The conventional cement can be expected to contain tricalcium aluminate, monosulfate, and calcium hydroxide, while the CSC matrix consists primarily of a low-lime cement that is reacted with carbon dioxide in the presence of water to form a silica gel structure and a calcium carbonate (32). On the basis of information provided in Equations 1 through 4, it is hypothesized that the CSC will have a lower potential for damage development in the presence of CaCl2 deicing salts.

Experimental Program This paper will examine two cementitious material systems. The first system was made with an OPC, and the second system was made with a CSC, a nonhydraulic binder that forms a solid through the carbonation process (32, 33). For both systems, paste and mortar samples were prepared. Table 1 provides the mixture proportions for the materials used in this study. The OPC mortar was prepared with a water-to-cement ratio (w/c) of 0.42 and 55% of aggregate by volume (3). The conventional system was made with a Type 1 OPC with 60% tricalcium silicate, 10% dicalcium silicate, 9% tricalcium aluminate, and 10% tetracalciumaluminoferrite by mass, and Blaine fineness of 375 m2/kg. The total equivalent alkali was 0.86%, including 0.35% of sodium oxide and 0.77% of potassium oxide, by mass. Aggregates used to prepare mortar specimens consisted of natural sand with a maximum size of 4.75 mm, specific gravity of 2.61, fineness modulus of 2.89, and absorption value of 2.2% by mass. The CSC mortar (CSCM) was made with a w/c ratio of 0.35, using CSC cement and 55% of aggregates (ASTM C778-13). The OPC mortar (OPCM) was prepared by following ASTM C192-12a for mixing and casting the samples in 25.4 mm × 25.4 mm × 300-mm (1 in. × 1 in. × 11.81 in.) molds. The samples were demolded after 24 h, and all mortar bars were then sealed in double plastic bags and cured for 28 days in these sealed conditions at 23°C ± 0.5°C. After 28 days of sealed curing, the mortar bars exhibited a degree of hydration of 78%. The OPC paste (OPCP) samples were prepared by using a vacuum mixer (speed 400 rpm), and cylindrical samples were cast (1.5 in. × 2.0 in.). Paste samples, before the test, were cured in a sealed condition for 1 year. The CSCM was prepared following a modified ASTM C109. The CSC paste (CSCP) samples were prepared with a water-to-cement ratio (w/c) of 0.24. The CSC systems were then exposed to a moist CO2 rich environment at 60°C for 40 h, allowing the material to develop its strength by forming an hardened solid constituted of mainly calcium carbonate and silica gel (32). The w/c indicated for the CSC system has a different implication than it would for a system with a hydraulic cement. The water in CSC systems serves to aid in rheology initially; however, it is removed during the reaction and only a fraction remains as a carrier for CO2, allowing the reaction between CO2 and the binder (CSC) to form the final hardened solid. In hydraulic cements, the w/c is often used as a surrogate measure for porosity. The porosity was directly measured here as 20% and 22% for OPCP and CSCP systems, respectively. Further, these systems were designed to have a similar compressive strength. The experimental program consisted of performing three primary tests: 1. Dynamic vapor sorption (DVS) was used to provide an estimated pore structure/volume/size in an effort to better understand the pore structure of the two systems. TABLE 1 Mixture Proportions OPC, by Sample Type (lb/yd3)

CSC, by Sample Type (lb/yd3)

Item

Paste

Mortar

Paste

Mortar

Cement Fine aggregates w/c

2,284

1,030 2,419 0.42

3,022

1,135 2,391 0.35

Note: na = not applicable.

na 0.42

na 0.24

50

Dynamic Vapor Sorption: Pore Structure Desorption isotherms were measured with a DVS analyzer. The OPCM and CSCM samples (described as follows) were placed into a tared quartz pan in a relative humidity (RH) and temperature controlled chamber. The humidity was reduced in controlled FH steps, while the mass was recorded as a function of time (the balance has an accuracy of ± 0.1% over a dynamic range of 100 mg), and the temperature was kept constant (3, 32, 34). The test started equilibrating the sample at 97.5% RH and subsequently decreasing the RH to 90% RH (for the CSC sample) and to 87.5% RH (for the OPC mortar sample), and then in 10% RH steps until a RH of 0% was achieved. The sample was allowed to equilibrate for 12 h or 0.001% (of initial mass of the sample) change in mass over 15 min at each humidity (3, 32, 34). The total porosity of the samples was also evaluated following ASTM C642-01; however, the boiling method was replaced with vacuum saturation in lime water (32). The vacuum saturation was conducted at 7 ± 2 torr (ASTM C1202-12). For DVS and total porosity measurements, the OPCM samples were prepared after 28 days of sealed curing, while the CSCM samples were tested 90 days after the carbonation curing process. The curing time provided for both systems, owing to the different reaction occurring, (i.e., hydration for OPC systems and carbonation for CSC systems), is not directly comparable. The curing time for the CSC system enabled the reactions to take place during the first 40 h, and the structure does not change substantially over time. The samples were prepared by cutting thin slices (0.8 ± 0.05 mm) from mortars samples using a water-cooled diamond-tipped wafer cut saw. A 50- to 70-mg piece of sample was selected for the test. To avoid the influence of the interfacial transition zone properties, a piece mainly constituted of paste was selected for testing. Before the test, the sample was submerged in lime water for 24 h. The desorption curve of the OPCM and CSCM samples are shown in Figure 1, which reports the mass of water present in equilibrium condition in both materials at various levels of RH. The CSCM has a lower total porosity (mass of water/oven-dry mass) at 4.5% than the traditional OPCM at 9% (31). The lower porosity in the CSCM is seen for the entire RH range. The lower total porosity of CSCM was confirmed by using measurement performed under vacuum saturated condition (ASTM C642-01) with a total porosity of CSCM of 11% by volume, while for OPCM this value was measured at 18% by volume. Figure 2 provides an estimate of the pore size distribution for the CSCM and OPCM. The curves have been obtained neglecting the contribution of the absorbed water layer and assuming only capillary flow with reference to Kelvin-Laplace equation (32) (This will result in errors from smaller pores. However, this is simply done to provide an estimate of the sizes of pores that are emptying and the relative volume of each pore size.). The curves presented in

Mass of Water–Oven-Dry Mass (%)

The following sections of the paper provide a description of each technique, present results from each test method, and discuss the results as they relate to potential deleterious reactions.

10

CSCM OPCM

8 6 4 2 0 0

20

40

60

80

100

RH (%) FIGURE 1 Desorption curves of CSCM compared with OPC mortar obtained from DVS test.

Figure 2 show that CSCM is characterized by a higher volume of pores larger than 40 nm compared with the traditional OPCM. This result is in accordance with previously published results by the authors (32). The distribution of pores in the range between 1 and 40 nm can be considered similar to the pores of this size in OPCM in regard to the shape of the curve, even though the percentage of pores at each size is consistently lower in the case of CSCM. LT-DSC: Phase Formation LT-DSC tests were performed to assess the potential volume of deleterious phases that form when salts and binders react at different temperatures (11, 12, 19). Cementitious paste samples were selected for both the OPC and CSC systems since this test is intended to only evaluate the potential reactivity of the binder. The sample preparation procedure is reported below. The LT-DSC tests were performed with a low-temperature differential scanning calorimeter with temperature accuracy of ± 0.1°C and calorimetric precision of 0.05%. During the test, the temperature cycles ranged from 25°C to −90°C

Cumulative Porosity [g of water/g of oven-dry sample (%)]

2. LT-DSC was used to provide an information on temperature of phase changes (e.g., ice formation, calcium oxychloride formation, eutectic) as well as the volume of materials changing phase. 3. LGCC was used to provide information on the extent of damage that forms during these phase changes.

Transportation Research Record 2508

100

CSCM OPCM

80 60 40 20 0 0.1

1

10

Kelvin Radius (nm) FIGURE 2 Pore size distribution comparison between CSCM and traditional OPC mortar.

100

Villani, Farnam, Washington, Jain, and Weiss

51

0 Cumulative Heat (J/g of paste powder)

in the cooling path and from −90°C to 70°C in the heating ramp at a rate of a 4°C/min. Paste samples belonging to both systems (OPCP and CSCP) were then ground using a milling machine obtaining a powder subsequently sieved in a 75 µm sieve (No. 200) (11, 12). The powder obtained was oven dried at 65° ± 1°C for 3 days and the cooled in a desiccator for 2 h before testing to minimize carbonation and moisture absorption. The paste powder (9 to 11 mg) was mixed with 9 to 11 mg of solution (deionized water or CaCl2 solution) in a high-volume aluminum pan and tested in LT-DSC after 1 day. A reagent grade CaCl2 and deionized water were used to prepare several salt solutions having mass concentrations between 0% and 35%. A typical curve of the heating ramp obtained from the LT-DSC test is presented in Figure 3, where the heat flow (normalized with respect to the mass of powder) is shown as function of temperature. The curves shown in Figure 3 belong to samples obtained mixing CSCP powder and OPCP powder, respectively, with 25% CaCl2. Endothermic peaks characterizing the eutectic transition, which marks the beginning of ice melting to the completion of ice melting, have been marked for both materials tested. An additional peak is seen at temperature above 0°C in the case of the OPCP that corresponds to the formation of calcium oxychloride (11, 12). The formation of calcium oxychloride is deleterious for cementitious matrices because of its expansive nature (10–12). The amount of calcium oxychloride formed can be determined from the LT-DSC curve since it is related to the amount of heat released during calcium hydroxide formation. The energy absorbed during calcium oxychloride formation can be evaluated by calculating the area below the corresponding peak (Figure 3) with respect to the constant heat capacity baseline. Alternatively, the energy associated with calcium oxychloride formation can be estimated by measuring the magnitude of the drop recorded in the cumulative heat curve (Figure 4). The drop can be quantified measuring the vertical distance (along the y-axis) between the two inflection points indicated as solid circles (labeled as a and b) in Figure 4. The inflection points correspond to the transition

OPCP –20

–40

a

–60

Heat absorbed as a result of calcium oxychloride formation

–80

b –100 –80

–40 0 40 Temperature (C)

80

FIGURE 4 Cumulative heat absorbed for OPCP sample that reveals the drop associated with calcium oxychloride formation.

between the two linear regions where no phase changes occur. The second method (evaluation of the magnitude of the drop) has been used in the current study since it can be more easily automated. The magnitude of the drop showed in Figure 4 increases with increasing CaCl2 solution concentration, and so does the heat absorbed during this phase transition. Consequently, more calcium oxychloride is expected to form when increasing the salt solution concentration. This is generally true under the assumption of constant amount of calcium hydroxide available in all systems. The heat absorbed owing to calcium oxychloride formation with respect to CaCl2 solution concentration is pictured in Figure 5. The heat also provides qualitative information on the extent of damage expected in the system. Interestingly, the calcium oxychloride peak seen for the traditional paste did not appear for any CSCP samples tested in the entire range of CaCl2 solution concentration tested (0%, 10%, 15%, 20%, 25%, and 33%). This can be seen in Figure 6a, where the normalized 50

Ice formation

OPCP

0.5

Heat (J/g of paste powder)

Heat Flow (W/g of powder)

CSCP

Eutectic OPCP

Eutectic

–80

Ice formation –40 0 40 Temperature (C)

Calcium oxychloride 80

FIGURE 3 Heat flow for CSCP and OPCP with phase transitions marked (eutectic, ice formation, and calcium oxychloride formation peak).

CSCP

40 30 20 10 0 0

20 30 10 Salt Concentration (%)

FIGURE 5 Heat absorbed owing to calcium oxychloride formation for traditional OPC paste and CSC paste.

40

52


33% CaCl2 Eutectic

Calcium oxychloride formation

Eutectic 29.8% CaCl2 0.5 Eutectic 25% CaCl2



Eutectic

Ice formation

Ice formation

Eutectic

20% CaCl2 Eutectic

Ice formation

Calcium oxychloride Ice formation formation

Eutectic

Eutectic

Ice formation

Eutectic

15% CaCl2 Eutectic

Ice formation

Ice formation

Eutectic

–40

0

40

25% CaCl2

20% CaCl2

15% CaCl2 Calcium oxychloride formation

Ice formation

Ice formation –80


29.8% CaCl2

Calcium 10% CaCl2 oxychloride formation

0.5

10% CaCl2


35% CaCl2

80

Temperature (C) (a)

–80

–40

0 40 Temperature (C) (b)

80

FIGURE 6 Heat flow heating curve (normalized by the mass of powder) for (a) CSC system and (b) OPC system exposed to several calcium chloride concentrations; only the OPC system reveals the peak corresponding to calcium oxychloride formation.

Longitudinal Guarded Comparative Calorimeter: Damage Detection

and used to quantify and detect damage from cracking. Therefore, damage development can be monitored as any phase changes (i.e., freezing and thawing) occur in mortar specimens. The detailed test procedure has been previously described in Farnam et al. [(19), 2014a] and Farnam et al. [(35), 2014b]. Mortar specimens for each mixture were prepared cutting the original OPCM and CSCM prisms into shorter prismatic samples (25.4 mm × 25.4 mm × 50.8 mm). Before applying the temperature change cycle, mortar specimens were vacuum saturated in de-ionized water and water–CaCl2 solutions (5%, 10%, 15%, and 29.8% CaCl2 by mass). The vacuum saturation procedure was used to obtain three degrees of saturation (DOS): 100%, 95%, and 85%. Before saturation, all samples were dried in a vacuum oven at 65°C and at 20 mm Hg pressure for 7 days. Samples having the three DOS were prepared as follows:

For evaluation of the performance of the OPCM and CSCM during the temperature changes associated with freezing and thawing, several prismatic mortar specimens were prepared using the mixture proportions provided in Table 1. An LGCC was used to perform the freeze–thaw experiments. The advantage of using the LGCC was that acoustic emission sensors could be attached to the specimen

1. Samples having 100% DOS were vacuum saturated at 5 mm Hg of pressure. 2. Samples having 95% DOS were vacuum saturated at 25 mm Hg of pressure. 3. Samples having 85% DOS were vacuum saturated at 5 mm Hg of pressure and then stored in a temperature and RH-controlled

heat flow curves of CSCP samples mixed with different salt solutions are shown: only the eutectic and the ice formation peak are recorded. The OPCP samples exposed to CaCl2 solution (Figure 6b) revealed the calcium oxychloride peak at temperature above 0°C. Calcium oxychloride results from the chemical reaction between calcium hydroxide, CaCl2, and water (Equation 4). Consequently, the absence of calcium hydroxide in the CSCP system explains its nonreactivity in the presence of CaCl2 solution with reference to the formation of calcium oxychloride. The absence of this deleterious chemical reaction leads one to foresee the advantage of implementing CSC composites for enhancing the durability of concrete structure.

Villani, Farnam, Washington, Jain, and Weiss

53

chamber (50% RH and 23°C), where their mass was monitored until it reached 85% DOS. After saturation, the specimens were placed in the LGCC, and the temperature of mortar specimens was varied from 24°C to −40°C using a cold plate. The initial temperature of the test was set to remain at 24°C for 1 h to allow the specimen to equilibrate. After the initial temperature became stable, the bottom surface was cooled at a rate of 2°C/h. At −40°C, the temperature was kept constant for 4 h to allow the specimen to equilibrate. Then, the temperature was increased to 24°C at a rate of 4°C/h. The damage was evaluated measuring the ultrasonic wave speed (pulse velocity) before and after the freeze–thaw cycle using a pulsed wave generated by two coupled acoustic emission sensors through the length of the specimen (19, 35). Knowing the pulse velocity before and after the test enables a damage index to be determined for mortar samples. This can be used to calculate changes (i.e., reduction in the dynamic elastic modulus) that occur in the samples owing to damage caused by the exposure to deicing salt or by the thermal cycle (35). Figure 7 shows the damage index for CSCM and OPCM specimens after one freeze–thaw cycle. In both systems exposed to low salt concentration (lower than 15% CaCl2), damage is primarily caused by ice formation, in the form of hydraulic pressure and osmotic pressures. The damage at low concentration decreases as the degrees of saturation decreases. Mortar samples exposed to high salt concentrations (greater than 15% CaCl2) show damage mainly because of the formation of calcium oxychloride. This theory is supported by two aspects. First, the highly concentrated solution may not freeze at the temperature applied in this study. Second if it freezes, it will produce less ice, which will likely be not sufficient to explain the damage recorded. For the OPCM samples, a considerable amount of damage can be observed, mainly owing to the formation of calcium oxychloride. For CSCM samples, however, no damage was observed; this is related to the absence of calcium hydroxide in the CSCM samples that could interact with CaCl2 to form calcium oxychloride.

60 Damage caused by ice formation (hydraulic and osmotic pressure)

Damage Index (%)

50

OPC CSC (DOS 100%) CSC (DOS 95%) CSC (DOS 85%)

Summary This paper has investigated damage that forms in cementitious materials in the presence of CaCl2 deicing salts through a wide range of temperatures analyzing the damage expected to occur during freezing and thawing conditions in real concrete structures. The paper examined two systems. The first system uses a conventional OPC that solidifies primarily through the hydration reaction, while the second system is a CSC that solidifies primarily through carbonation. The CSC system was generally more resistant to deterioration caused by CaCl2 at higher CaCl2 concentrations. The improved performance that was observed in the CSC system can be attributed mainly to calcium hydroxide not being present in the CSC system. Use of CSC systems can greatly reduce and eliminate the potential for calcium oxychloride formation in cementitious systems, thereby reducing associated damage owing to salt exposure. Additionally, it is believed that the reduction in the calcium aluminate and monosulfate phases also aids the CSC in being more resistant to deterioration. LT-DSC provides clear evidence of the temperature of phase transformations when CaCl2 solutions are mixed with ground paste powders. The temperature ranges at which these transformations occur can be easily identified from the thawing cycle. In addition, the integral of the heat flow curve provides quantitative measures of the volume of phases (oxychloride, ice, and eutectic) forming in the system. This paper also provides results confirming that the damage that develops when CaCl2 is present in a cementitious system is complex. The damage developing at low concentrations of CaCl2 is similar in saturated OPC and CSC systems. This damage is primarily caused by ice formation owing to the hydraulic pressure created by the ice and the osmotic pressures that develop as the solution becomes more concentrated. Damage that occurs at low CaCl2 concentrations is heavily influenced by the degree of saturation. This damage appears to be most severe for concentrations below 10% to 15% CaCl2. For higher CaCl2 concentrations (>10% to 15%), the damage is primarily attributed to oxychloride formation and occurs at temperatures above freezing. This damage appears in the OPC system; however, it is eliminated in the CSC systems since it does not contain calcium hydroxide (a key component in the oxychloride reaction as shown in Equation 3). Acknowledgments

20

This work was supported in part by Solidia Technologies, which licensed core technology from Rutgers University. The authors gratefully acknowledge assistance in sample preparation for the CSC series, as well as financial support. The experiments reported in this paper were conducted in the Pankow Materials Laboratories at Purdue University. The authors acknowledge the support that has made this laboratory and its operation possible. The authors also acknowledge TA Instruments as the manufacturer of the LT-DSC device used in this study.

10

References

40

Damage caused by formation of calcium oxychloride

30

0 0

5

10 15 20 25 30 CaCl2 Concentration (% by mass)

35

FIGURE 7 Change in dynamic elastic modulus (damage index) for CSCM and OPCM specimens owing to damage caused by freeze–thaw cycle and calcium oxychloride formation.

1. Special Report 235: Highway Deicing: Comparing Salt and Calcium Magnesium Acetate. TRB, National Research Council, Washington, D.C., 1991. 2. Shi, X. The Use of Road Salts for Highway Winter Maintenance: An Asset Management Perspective. Presented at the 2005 ITE District 6 Annual Meeting, Kalispell, Mont., 2005. 3. Villani, C., R. Spragg, M. Pour-Ghaz, and J. Weiss. The Influence of Pore Solutions Properties on Drying in Cementitious Materials. Journal of the American Ceramic Society, Vol. 97, No. 2, 2014, pp. 386–393.

54

4. Powers, T. C. A Working Hypothesis for Further Studies of Frost Resistance of Concrete. Journal of the American Concrete Institute, Vol. 4, No. 4, 1945, pp. 245–272. 5. Beaudoin, J. J., and C. MacInnis. The Mechanism of Frost Damage in Hardened Cement Paste. Cement and Concrete Research, Vol. 4, No. 2, 1974, pp. 139–147. 6. Scherer, G. W. Crystallization in Pores. Cement and Concrete Research, Vol. 29, No. 8, 1999, pp. 1347–1358. 7. Mehta, P. K. Durability of Concrete in Marine Environment—A Review. SP-65: Performance of Concrete in Marine Environment, American Concrete Institute, Farmington Hills, Mich., 1980, pp. 1–20. 8. Valenza II, J. J., and G. W. Scherer. A Review of Salt Scaling: II. Mechanisms. Cement and Concrete Research, Vol. 37, No. 7, 2007, pp. 1022–1034. 9. Scherer, G. W. Stress from Crystallization of Salt. Cement and Concrete Research, Vol. 34, No. 9, 2003, pp. 1613–1624. 10. Collepardi, M., L. Coppola, and C. Pistolesi. Durability of Concrete Structures Exposed to CaCl2 Based Deicing Salts. SP145: Durability of Concrete—Proceedings Third CANMET, Nice, France, American Concrete Institute, Farmington Hills, Mich., 1994, pp. 107–120. 11. Farnam, Y., T. Washington, and J. Weiss. The Influence of Calcium Chloride Salt Solution on the Transport Properties of Cementitious Materials. Advances in Civil Engineering, Volume 2015, 2015. doi: 10.1155/2015/929864. 12. Farnham, Y., S. Dick, A. Wiese, J. Davis, D. Bentz, and J. Weiss. The Influence of Calcium Chloride Deicing Salt on Phase Changes and Damage Development in Cementitious Materials. Submitted to the Journal of Cement and Concrete Composite, Elsevier (under review). 13. Dubberke, W., and V. J. Marks. The Effect of Deicing Salt on Aggregate Durability. In Transportation Research Record 1031, TRB, National Research Council, Washington, D.C., 1985, pp. 27–34. 14. Wallbank, E. J., Great Britain Department of Transport, and G. M. Partners. The Performance of Concrete in Bridges. A Survey of 200 Highway Bridges. Her Majesty’s Surveying Office, London, 1989. 15. Wang, K., D. E. Nelsen, and W. A. Nixon. Damaging Effects of Deicing Chemicals on Concrete Materials. Cement and Concrete Composites, Vol. 28, No. 2, 2006, pp. 173–188. 16. Glasser, F. P., J. Marchand, and E. Samson. Durability of Concrete— Degradation Phenomena Involving Detrimental Chemical Reactions. Cement and Concrete Research, Vol. 38, No. 2, 2008, pp. 226–246. 17. Demediuk, T., W. Cole, and H. Hueber. Studies on Magnesium and Calcium Oxychlorides. Australian Journal of Chemistry, Vol. 8, No. 2, 1955, pp. 215–233. 18. Brown, P., and J. Bothe, Jr. The System CaO-Al2O3-CaCl2-H2O at 23±2°C and the Mechanisms of Chloride Binding in Concrete. Cement and Concrete Research, Vol. 34, No. 9, 2004, pp. 1549–1553. 19. Farnam, Y., D. Bentz, A. Hampton, and W. J. Weiss. Acoustic Emission and Low Temperature Calorimetry Study of Freeze and Thaw Behavior in Cementitious Materials Exposed to Sodium Chloride Salt. In Transportation Research Record: Journal of the Transportation Research Board, No. 2441, Transportation Research Board of the National Academies, Washington, D.C., 2014, pp. 81–90. 20. Suryavanshi, A., J. Scantlebury, and S. Lyon. Mechanism of Friedel’s Salt Formation in Cements Rich in Tri-Calcium Aluminate. Cement and Concrete Research, Vol. 26, No. 5, 1996, pp. 717–727. 21. Glasser, F., A. Kindness, and S. Stronach. Stability and Solubility Relationships in AFm Phases: Part I. Chloride, Sulfate and Hydroxide. Cement and Concrete Research, Vol. 29, No. 6, 1999, pp. 861–866.


22. Hirao, H., K. Yamada, H. Takahashi, and H. Zibara. Chloride Binding of Cement Estimated by Binding Isotherms of Hydrates. Journal of Advanced Concrete Technology, Vol. 3, No. 1, 2005, pp. 77–84. 23. Jensen, T. R., A. N. Christensen, and J. C. Hanson. Hydrothermal Transformation of the Calcium Aluminum Oxide Hydrates CaAl2O4 10H2O and Ca2Al2O5 8H2O to Ca3Al2(OH)12 Investigated by In Situ Synchrotron X-Ray Powder Diffraction. Cement and Concrete Research, Vol. 35, No. 12, 2005, pp. 2300–2309. 24. Florea, M., and H. Brouwers. Chloride Binding Related to Hydration Products: Part I: Ordinary Portland Cement. Cement and Concrete Research, Vol. 42, No. 2, 2012, pp. 282–290. 25. Jones, M. R., D. E. Macphee, J. A. Chudek, G. Hunter, R. Lannegrand, R. Talero, and S. N. Scrimgeour. Studies Using 27Al MAS NMR of AFm and AFt Phases and the Formation of Friedel’s Salt. Cement and Concrete Research, Vol. 33, No. 2, 2003, p. 177–182. 26. Csizmadia, J., G. Balázs, and F. D. Tamás. Chloride Ion Binding Capacity of Aluminoferrites. Cement and Concrete Research, Vol. 31, No. 4, 2001, pp. 577–588. 27. Beaudoin, J., S. Catinaud, and J. Marchand. Volume Stability of Calcium Hydroxide in Aggressive Solutions. Cement and Concrete Research, Vol. 31, No. 1, 2001, pp. 149–151. 28. Mori, H. Chemical Deterioration of Hardened Cement Pastes Immersed in Calcium Chloride Solution. 3rd International Conference on Sustainable Construction Materials and Technologies—SCMT3, Kyoto, Japan, 2013. 29. Sutter, L., K. Peterson, S. Touton, T. Van Dam, and D. Johnston. Petrographic Evidence of Calcium Oxychloride Formation in Mortars Exposed to Magnesium Chloride Solution. Cement and Concrete Research, Vol. 36, No. 8, 2006, pp. 1533–1541. 30. Peterson, K., G. Julio-Betancourt, L. Sutter, R. D. Hooten, and D. Johnston. Observations of Chloride Ingress and Calcium Oxychloride Formation in Laboratory Concrete and Mortar at 5°C. Cement and Concrete Research, Vol. 45, 2013, pp. 79–90. 31. Shi, C. Formation and Stability of 3CaO CaCl2 12H2O. Cement and Concrete Research, Vol. 31, No. 9, 2001, pp. 1373–1375. 32. Villani, C., R. Spragg, R. Tokpatayeva, J. Olek, and W. J. Weiss. Characterizing the Pore Structure of Carbonated Natural Wollastonite. Proc., 4th International Conference on the Durability of Concrete Structures. Purdue University, West Lafayette, Ind., 2014, pp. 1–9. 33. Sahu, S., and N. DeCristofaro. Solidia Cement. White paper. Solidia Technologies, Piscataway, N.J., pp. 1–12. http://www.solidiatech.com. 34. Pour-Ghaz, M., R. Spragg, and J. Weiss. Moisture Profiles and Diffusion Coefficients in Mortars Containing Shrinkage Reducing Admixtures. Proc., International RILEM Conference on Use of Superabsorbent Polymers and Other New Additives in Concrete Technical (O. M. Jensen, M. T. Hasholt, and S. Laustsen, eds.), Lyngby, Denmark, 2010, pp. 197–206. 35. Farnam, Y., D. Bentz, A. Sakulich, D. Flynn, and J. Weiss. Measuring Freeze and Thaw Damage in Mortars Containing Deicing Salt Using a Low-Temperature Longitudinal Guarded Comparative Calorimeter and Acoustic Emission. Journal of Advances in Civil Engineering Materials, Vol. 3, No. 1, 2014b, pp. 316–337. doi:10.1520/ACEM20130095. •

•

•

•

The contents of this paper reflect the views of the authors, who are responsible for the facts and the accuracy of the data; the results do not constitute a standard, specification, or regulation. The Standing Committee on the Durability of Concrete peer-reviewed this paper.