Effects of Steam Curing on the Resistance of Cementitious Materials to Chloride Penetration

CHEN Xuefei, HE Yongjia＊, LÜ Linnü, ZHANG Xulong, HU Shuguang

(1. State Key Laboratory of Silicate Materials for Architectures (Wuhan University of Technology), Wuhan 430070, China; 2. School of Materials Science and Engineering, Wuhan University of Technology, Wuhan 430070, China; 3. Hubei Key Laboratory of Theory and Application of Advanced Materials Mechanics, School of Science, Wuhan University of Technology, Wuhan 430070, China; 4. Hubei Provincial Academy of Building Research and Design Co., Ltd, Wuhan 430071, China)

Abstract: The resistance to chloride penetration of cement-based material with different curing regimes was investigated by solution titration, XRD, LF-NMR and rapid chloride migration test. The results show that the curing regime has influences on the types and structure of the hydration products, which in turn affects their ability to bind chloride ions. The binding capacity of cementitious materials to chloride ions, porosity and chloride ion migration coefficient increased with the increase of water-cement ratio, while steam curing increased the porosity and chloride ion migration coefficient at the same time as it increased the chloride ion binding capacity of the materials. At lower water-cement ratios, the effect of steam curing on the resistance of cementitious materials to chloride ingress is negligible.

Key words: chloride penetration; chloride binding capacity; pore structure; steam curing; cementitious materials

1 Introduction

Durability is an important performance of modern concrete, especially for concrete used in coastal areas or corrosive environment. One of the main deterioration mechanisms for reinforced concrete is the steel bar corrosion, and chloride ions plays important roles in the corrosion process. Chloride ions ingress through the concrete cover towards the steel when the structure is exposed to sufficient chlorides, and then the ions take part in the electrochemical reactions between steel and oxygen and accelerate this process. There are three main forms of chloride ions present in cementitious materials, some are chemically bound by tricalcium aluminate (C3A) and its hydration products to form stable compounds-Friedel’s salt (C3A·CaCl2·10H2O), some are physically adsorbed in the lamellar structure by C-S-H gels, and the remaining are existing in the pore solution as free chloride ions[1-5]. Only the free chloride ions play roles in the corrosion process of steel reinforcements[6], which means that increasing the binding of chloride ions in cement-based materials is conducive to reducing the content of free chloride ions in pore solution, so as to reduce the risk of reinforcement corrosion. Martín’s research showed that the service life of reinforced concrete was improved by 238% and 93% at 0.5 and 2.5 mol/L chloride environments, respectively,after considering the influence of Cl-binding in the life prediction model of concrete structures[6].

With the development of the construction technology, precast concrete components are more and more widely used. In order to improve efficiency and shorten production cycle, precast concrete products are usually produced by steam curing or autoclave curing. In these hydrothermal curing processes, the hydration process of cement is obviously accelerated, which leads to the changes of the types and morphologies of the hydration products, as well as the pore structure. Generally speaking, after hydrothermal curing, the grain size of the hydration products becomes bigger and the pore structure of the cement matrix becomes coarser. These changes have negative impacts on the erosion resistance of the material[7-9].

When the pore structure is coarsened and the porosity increases, the ion concentration difference inside and outside of the material is enhanced, which further intensifies the diffusion of chloride ions to the inside[10].In this paper, the effects of curing regime and water to cement ratio on the chloride corrosion resistance of cementitious materials were investigated by rapid chloride ion migration test and low field nuclear magnetic resonance (LF-NMR).

Table 1 Chemical composition of cement/wt%

2 Experimental

2.1 Raw materials

P.I 42.5 Portland cement was used for the test,and its chemical compositions are shown in Table 1.The water used is deionized water, and the sand is the standard quartz sand according to Chinese national standard.

2.2 Samples preparation

Mortar specimens with different water-cement ratios were prepared according to the mixtures shown in Table 2, respectively.

Table 2 Mixtures for the mortar specimens

Fig.1 Steam curing regime

The mortar specimens for binding capacity tests were well mixed, casted into 2 cm × 2 cm × 2 cm cubic molds and covered with cling film. The specimens for rapid chloride migration test were prepared by casting the mortars into cylindrical molds of 100 mm in diameter and 50 mm in height. The specimens for LF-NMR were cut into cylinders of 20 mm in diameter and 30 mm in height.

For standard curing specimens: after two days, the specimens were demolded and then placed in the curing room (temperature 20 ℃, relative humidity above 98%) for curing up to 28 days. For steam curing specimens, after casting and covering with cling film, the specimens are left for 4 h and then steam cured, and the curing regime is shown in Fig.1. After steam curing,they are placed in the curing room (same as standard curing) and cured until the same age as standard curing.

2.3 Test methods and procedures

2.3.1 Chloride binding capacity

The chloride binding capacity of specimens was determined by potentiometric titration, and the method is improved on the basis of Tang’s[11].

The specimens were removed from the core and crushed to particles of about 1-2 mm in diameter when they were cured until 28 d age.

20 g specimen particles with different curing conditions and water-cement ratios were immersed in the same concentration of NaCl solution of 0.8 mol/L (denoted asC0). After 14 days of soaking, the solution reached the overall adsorption equilibrium. The solution and the particles were taken out separately and washed with deionized water. The chloride concentration in the solution was measured at this time (denoted asC1). The total bound chloride ion content can be calculated using Eq.(1).

The cleaned particles were immersed in saturated calcium hydroxide solution for 14 days waiting for their physical adsorption behavior to reach equilibrium.Then the chloride concentration (denoted asC2) in the solution was measured by Eq.(2). This process is considered as ideal for the sake of experimental effectiveness.

Finally, the chemically bound chloride ion concentration was calculated by Eq.(3).

where,Cbtis the total bound chloride ion concentration of the sample, mg/g;c1is the chloride ion concentration of the equilibrium solution, mol/L;c0is concentration of NaCl solution configured with saturated Ca(OH)2,mol/L;Mis mass of the sample, g;Cbpis concentration of physisorbed chloride ions of the sample, mg/g;c2is concentration of chloride ions precipitated in the Ca(OH)2, mol/L;cbcis chemically bound chloride ion concentration of the sample, mg/g.

2.3.2 X-ray diffraction (XRD)

The specimens were soaked in isopropyl to terminate hydration, ground and sieved through 200 mesh and then subjected to quantitative XRD testing. An Empyrean X-ray diffractometer (Malvern Panalytical)was used with a scanning angle range of 5°-60° and a scanning rate of 2 (°)/min. 10 wt% of Al2O3powder was added as an internal standard before testing and the specimens were quantified using the Rietveld method.The total of all phases in the XRD was normalized to 100 wt% and the Friedel’s salt content was calculated.

2.3.3 LF-NMR

The pore structure of the sample was tested by LF-NMR from Suzhou Niumag Analytical Instruments Co. The magnetic field strength of the instrument was 0.42 T, the magnet frequency was 18 MHz, and the magnet temperature was maintained at 32 ℃±0.02 ℃.The number of echoes (NECH) was 2 000, the time interval for repeated sampling (TW) was 3 000, and the cumulative number of samples (NS) was 128. The samples were polished to a diameter of 25 mm and a length of 30 mm, and were saturated with water for 12 h before testing.

Fig.2 Schematic of RCM

2.3.4 Rapid chloride migration (RCM)

In this paper, the determination of RCM was carried out with reference to the Chinese national standard(GB/T 50082-2009). Three specimens were tested in each group， and the specimens were vacuum-preserved in saturated calcium hydroxide solution for 24 h before the test. The specimens were wrapped in a silicone sleeve and tightened with a stainless steel ring,and the bottom of the specimens were submerged in a 10% mass concentration NaCl solution and the top in a 0.3 mol/L NaOH solution. Initially, 30 V DC voltage was applied, and the test voltage and time were determined according to the initial current. After rinsing the specimen at the end of the test time, the specimen was split into two halves using a press, and then immediately sprayed with AgNO3solution with a concentration of 0.1 mol/L for color development, and the white part of the specimen was measured after 15 min,i e, the depth of the chloride ion penetration area.

3 Results and discussion

3.1 Chloride binding capacity

Fig.3 Chloride binding capacity of specimens under standard curing

Fig.4 Chloride binding capacity of specimens under steam curing

Fig.3 and Fig.4 show the diagrams of the chloride binding content of the three specimens with water-cement ratios of 0.42, 0.32 and 0.25 after standard and steam curing, respectively. The total bound chloride ion content after standard curing is about 12-15 mg/g. The sample C30 with the highest water-cement ratio of 0.42 has the highest total bound chloride ion content, chemically bound chloride ion content and physically adsorbed chloride ion content among the three. The sample C80 with the lowest water-cement ratio of 0.25 had the lowest combined chloride ion content of all three.The total bound chloride ion content of the three samples after steam curing was in the range of 15 to 20 mg/g. The binding of chloride ions in the samples followed a similar binding pattern to that in the standard condition, and the content decreased with the decrease of the water-cement ratio. However, in the longitudinal comparison, the steam-cured specimens showed an increase in the chloride binding content, with the total bound chloride ion content increasing in the range of 2 to 5 mg/g. This is because steam curing promotes the generation of HO-AFm in them, which leads to an increase in their binding capacity for chloride ions[12]. In addition, in the steam curing environment, the C-S-H gel changed from the granular shape in the standard curing to the honeycomb shape[13]morphology, and structure of hydration products in hardened pastes of three kinds of blended cement (cement-silica fume, cement-quartz powder and cement-silica fume-quartz powder), which improves the specific surface area, while improving the physical adsorption capacity of chloride.

3.2 XRD analysis

The specimens after immersion in 0.8 mol/L NaCl solution were recorded as C30-0.8, C50-0.8, and C80-0.8, respectively.

Fig.5 shows the comparison of XRD patterns before and after immersion in NaCl solution for cement pastes with water-cement ratio of 0.42, 0.32, and 0.25 under standard curing environment, respectively. As can be seen from Fig.5(a), the Ca(OH)2diffraction peak of the C30 specimen under standard curing is the strongest, followed by C50 and somewhat weaker for C80. Among them, the diffraction peaks representing the cement minerals dicalcium silicate (C2S) and tricalcium silicate (C3S) are enhanced with the decrease of water-cement ratio. Fig.5 shows that the degree of hydration of the three decreases with decreasing water-cement ratio.

Fig.5 XRD patterns of cement pastes before and after isothermal adsorption in NaCl solution under standard curing

The diffraction peak of Friedel’s salt is observed at 11.28° in Fig.5(b)[14]. It is known from the literature[15]that when the concentration of chloride ions in the solution to which the material is subjected decreases, the generated Friedel’s salt may convert to Kuzel’s salt (C3A·1/2CaSO4·1/2CaCl2·11H2O). When testing the physical adsorption capacity of the material, the particle samples were transferred from the NaCl solution to the Ca(OH)2solution, so Kuzel’s salt may be generated.The diffraction peaks of Kuzel’s salt are marked with red circles in Fig.5(b). The sum of the intensities of the diffraction peaks of Friedel’s salt and Kuzel’s salt indicates the chemical binding ability of the sample with chloride ions[14]. In this paper, Kuzel’s salt was negligible due to its low content.

Table 3 Friedel’s salt content generated from cementitious materials with different water-cement ratio

Fig.6 shows the comparison of XRD patterns before and after immersion in NaCl solution for cement pastes with water-cement ratio of 0.42, 0.32, and 0.25,respectively, under the steam curing environment. As can be seen from Fig.6, the diffraction peaks of the cement mineral phases C2S and C3S in the XRD of the cement specimens after steam curing are reduced compared to those of Fig.5 for the standard curing,indicating that steam curing promotes the hydration of the cement minerals of the specimens. In contrast, from the XRD patterns of the cement specimens soaked in NaCl solution and their enlargements, it is seen that a slight increase in the intensity of the Friedel’s salt diffraction peak compared to the standard curing, which is consistent with the experiment results described above that the chloride binding capacity of the specimens is improved after steam curing.

The quantification of Friedel’s salt generated from cementitious materials under standard and steam curing environments using the internal standard method is list in Table 3.

The quantitative results match the test results above and provide support for the test results.

3.3 LF-NMR

Fig.6 XRD patterns of cement pastes before and after isothermal adsorption in NaCl solution under steam curing

The pore structure is an important parameter in assessing the durability and erosion resistance of cementitious materials, and usually the higher the porosity,the worse the erosion resistance. Low-field NMR, as an effective pore structure test method, is highly efficient and non-invasive compared to other test methods (mercury-pressure and nitrogen adsorption), and can characterize the overall porosity and pore size distribution of samples. The number of echoes (NECH), repeated sampling interval (TW) and cumulative sampling times(NS) were 2 000, 3 000 and 128, respectively. The time it takes to return to equilibrium (relaxation time,T2time) is affected by different environments in which the hydrogen atoms are located. The relaxation time characterizes the corresponding pore size, and the pore size distribution can be determined from the distribution of the T2spectrum. The relationship between the relaxation time T2and the pore size is positively correlated,i e, the larger the pore, the longer the relaxation time of hydrogen atoms in water. Therefore, the pore size variation pattern of the hydrated sample can be explored directly by comparing the distribution of T2spectra without calculating the pore size.

The T2signal spectra showed that the pores of all samples consisted of a wide primary signal peak in the range of 0.01 to 1 ms and a secondary signal peak in the range of 1 to 1 000 ms. The positions of the primary signal peaks were in about the same interval for all samples, indicating that the pore size distribution was similar for these specimens. Since the signal intensity per unit volume is used as the vertical coordinate, the decrease in signal peak intensity represents a decrease in the number of pores. It is generally believed that there are two main types of pores in the hardened cement paste: capillary pores formed by mixing water that does not participate in the hydration reaction, and gel pores formed in the gel of calcium silicate hydration products[16]. The first wave peak in the low-field NMR spectrum represents the gel pore, the second wave peak represents the capillary pore, and the gel pore area is larger[17].

Fig.7 T2 spectra of cementitious materials

Fig.7 shows the T2signal spectra with porosity of cement mortars with different water-cement ratios after 28 d in standard and steam curing. The cement specimens with three different water-cement ratios show the same trend under different curing regimes — the number of pores decreases as the water-cement ratio decreases. The main peak signal of the steamed hardened cement paste shifts to the right, the peak intensity increases, the area of the main peak increases, and the number of pores rises. This is because the cement paste forms a large number of hydration products in the process of continuous hydration, especially the gel products that are generated from the surface of the particles to the surrounding water-filled space, forming a dense microstructure with a high gel pore ratio, and the steam curing improves the hydration of the specimens with relatively low water-cement ratio to a certain extent. The increase in the water-cement ratio represents an increase in the degree of hydration of the material.The unhydrated particles in the material with lower water-cement ratio are closely packed, and a small amount of hydration products hydrate on the surface of the unhydrated particles and fill the pore structure to make it denser, and the increase of its porosity after steaming is not significant.

Normally, chloride ions enter the interior of the material by diffusion. When steam curing increases the porosity and pore size of the material, the degree of chloride ion penetration will increase. The pore size structure of the material affects both the bound and free chloride ion content. There is a dynamic balance between pore size structure and chloride binding capacity.

3.4 Rapid chloride migration

Rapid chloride migration (RCM) coefficient is one of the important indicators to characterize the resistance of concrete to ionic erosion. The method is to apply a voltage at both ends of the specimen to promote the rapid migration of chloride ions in the concrete, by measuring the chloride ion diffusion depth and other calculations to obtain the migration coefficient of chloride ions to characterize the resistance of concrete to ionic erosion performance. The method was first proposed by Tang Luping and later adopted as the NordTest NTBuild492, and China has also included the method in the national standard. After rapid chloride migration tests and calculations for all samples, the non-stationary chloride ion migration coefficientDnssmcan be calculated according to the following equation:

whereDnssmis the non-stationary chloride migration coefficient (×10-12m2/s),Uis the applied voltage (V),Tis the average of the initial and ending temperatures of the NaOH solution (°C),Lis the thickness of the specimen,xdis the average depth of chloride ion penetration(mm), andtis the duration of the test (h).

Fig.8 represents the comparison of chloride ion migration coefficients of cement mortars at different water-cement ratios. The solid line represents the standard curing conditions and the dotted line represents the steam curing conditions.

It can be observed from Fig.8 that the chloride migration coefficient of mortar decreases as its water-cement ratio decreases, while steam curing increases the chloride migration coefficient. However, for the C80 specimens with lower water-cement ratio, steam curing has less effect on the chloride migration coefficient.From the previous conclusions, it is known that steam curing enhances the material’s ability to bind chloride ions, but also increases its porosity. According to the RCM results, it can be concluded that porosity occupies a major position in the material’s resistance to chloride ion attack, while the bound chloride ions are only a small amount. And the water-cement ratio reduction is the most fundamental solution.

4 Conclusions

a) For cementitious materials with different water-cement ratios, steam curing enhances the chemical binding and physical adsorption of chloride to the material.

b) Steam curing significantly increased the cement chloride migration coefficient, and make the pore structure of the specimen pore structure coarser. When the water-cement ratio is reduced to a certain degree, steam curing has little effect on the chloride ion migration coefficient. The reduction of water-cement ratio enhances the resistance of specimens to chloride erosion.

c) The effect of the pore structure on the material’s resistance to chloride penetration is greater than the effect of its own chloride ion binding capacity. If it is desired to improve the resistance to erosion by increasing the chloride binding capacity of the material, other aspects should be considered, such as changing the raw material.