Influence of the Coupling Action of Flexural Load and Freezing–Thawing on the Chloride Diffusion of Marine High-Performance Concrete

Abstract

The chloride diffusion of marine high-performance concrete under a couple of actions, flexural load and freezing–thawing, was investigated by a fast freezing and thawing test in NaCl solution. Concrete specimens of 100 × 100 × 515 mm were tested under bending load and 300 freeze–thaw cycles under the stress levels of 15%, 30%, and 50% of the ultimate fracture modulus. The change in the microstructure of the concrete was analyzed by SEM and MIP. The results indicated that the chloride diffusion coefficient of concrete under the coupling effect of flexural load and freezing–thawing or simple flexural load increased with the increasing in the flexural stress level, and the chloride diffusion coefficient was approximately exponential to the flexural stress level, as D = 0.8777e^1.668σ for a couple of actions of flexural load and freezing–thawing, and D = 0.8336e^1.3231σ for a simple flexural load. The resistance ability of concrete to chloride diffusion was reduced by the freezing–thawing procedures, the resisted ability dropped more severely under a couple of actions of flexural load and freezing–thawing than simple flexural load at the same stress level. Micro-cracks at the interfacial transition zone between the aggregate and the paste matrix in concrete was induced under a couple of actions of flexural load and freezing–thawing, which increased the average pore size and total pore volume, resulting in the modification of the pore size distribution in the concrete. The influence of a couple of actions of flexural load and freezing–thawing on the concrete was greater than that of simple flexural load.

1. Introduction

The concrete structure of a seaport is in the marine environment, and the actual stress of the structure is a static load or alternating load with low frequency. An actual engineering investigation showed that the corrosion damage caused by chloride is the main factor affecting the durability of concrete structures in harbor engineering, and the durability deterioration of concrete structures caused by chloride ion erosion is very common [1,2,3]. The time, environment, the concrete material itself, and other factors have been widely studied for the diffusion process of chloride ions in concrete, and many durability analysis models of reinforced concrete structures in the marine environment have been obtained [4,5].

However, the chloride ion diffusion coefficient is affected mostly by the load and actual environmental factors in engineering practice [6,7]. For example, concrete bridges in cold regions simultaneously bear the load, freeze–thaw cycles, and chlorine salt erosion together. Studies with consideration of single effect factors cannot truly reflect the actual environment of engineering and objectively predict the service life of concrete structures. Damage caused by chloride ion erosion under many factors is not the simple sum of single factors [8,9]. The failure process of actual service concrete is complicated and the conclusions and experience formula obtained with consideration of a single factor is limited.

The influence of the interaction of multiple factors on the chloride ion diffusion coefficient of concrete has drawn the attention of more and more scientists [10,11]. Research on the effect of load and chloride corrosion factors on concrete [11,12] shows that repeated loading significantly reduces chloride corrosion resistance. Jiang and Wang [13,14] quantified the influence of freeze–thaw damage on the chloride diffusion coefficient, providing a theoretical basis for the durability prediction of concrete structures subjected to chloride corrosion and freeze–thaw action. Wang [15] studied the coupling effect of freezing–thawing and uniaxial compression load but did not find out the mechanism of their interaction. It is generally believed that the pore structure and cracks are important factors affecting the freeze–thaw performance of concrete in severe cold regions, and the load action can make the structure produce new cracks and expand the original cracks, reducing the freeze–thaw resistance of concrete [16,17,18,19,20]. The research on the effect of load and freeze–thaw cycle on the chloride ion diffusion coefficient of concrete mainly individually focuses on the influence of different load proportions or the number of freeze–thaw cycles [21]. Although some scholars have discussed the influence of load on the chloride diffusion coefficient of concrete, most of them only consider the influence of load or the freeze–thaw cycle alone [22]. In some studies on the influence of load and freeze–thaw cycles on chloride diffusion coefficient, the object of study is the unloaded concrete sample, not the concrete sample subjected to both load and freezing–thawing, and the combined effect of load and environmental erosion cannot be fully considered [23,24].

In the northern region of China, the actual environment and load of concrete structures in seaports generally act as the interaction of load, chloride, and freezing–thawing. The chloride ion diffusion coefficient of concrete under the interaction of this load and environmental erosion factors will become a key research direction. In this study, based on the widespread adoption of marine high-performance concrete materials under chlorine salt environments [25,26] and the stress state of actual engineering structures in service, the combined effect of bending load and the freeze–thaw cycle on chloride ion diffusion in concrete is studied by the rapid freeze–thaw test of loaded specimens in NaCl solution.

2. Materials and Experiments

2.1. Materials

P. II 42.5 Portland cement produced by Zhujiang Cement Co., Ltd. (Guangzhou, China), class II fly ash produced by Zhujiang power plant, and class S95 GGBS (ground granulated blast furnace slag) from Guangdong Shaoguan Jia Yang as a binder materials were used. The particle sizes of granite rubble of 5~20 mm from Guangzhou Zhuhai Jian Bang quarries and the sand with a fineness modulus of 2.7 from Xijiang River in Guangdong used coarse and fine aggregates. A polycarboxylate HSP-V with a 32.2% water-reducing rate from Guangzhou Fourth Harbor Materials Technology Co., Ltd. (Guangzhou, China) was used as a superplasticizer. The chemical compositions of the binder materials are shown in

Table 1. The chemical compositions of the binder materials (wt%).

Materials	SiO2	Al2O3	Fe2O3	CaO	MgO	Na2O	K2O	SO3	LOI
Cement	20.24	4.43	5.30	63.83	0.97	0.08	0.64	2.85	2.68
Fly ash	60.81	24.12	5.76	3.03	0.55	0.32	0.71	1.31	3.26
GGBS	34.20.	13.39	1.20	38.20	9.02	0.20	0.51	0.28	2.73

.

2.2. Experiments

Marine high-performance concrete C50 for the splash zone of a typical port construction was designed. The size of the specimen was 100 mm × 100 mm × 300 mm for the specimen under uniaxial compressive loading and 100 mm × 100 mm × 515 mm for the specimen under bending loading. After 24 h, the specimen was removed from the mold and placed in the concrete specimen curing room for standard curing. The mix proportion and the mechanical properties of the specimens are shown in

Table 2. Mix proportion and 28 d mechanical properties of concrete C50.

NO.	Binder Materials (%)			W/B	Mix Proportion (kg/m3)				Compressive Strength (MPa)	Bending Strength (MPa)
	Cement	Fly Ash	GGBS		Binder	Sand	Rubble	Water
FS	40	20	40	0.35	420	723	1084	147	61.8	6.4

.

The concrete specimens with a size of 100 mm × 100 mm × 515 mm were prepared for mechanical properties testing at 28 d. The specimens were tested under constant three-point bending. According to the analysis of the stress state of the typical components in the actual high-pile wharf in China [27], the stress levels were respectively set as 15%, 30%, 40%, and 50% of the ultimate flexural strength of concrete, and a group of specimens without load was taken as a contrast.

Each group has three specimens and the concrete specimens under flexural load and freeze–thaw are shown in Figure 1.

The experimental process can be listed as follows.

(1) The concrete specimens respectively with the flexural stress level of 15%, 30%, 40%, and 50% were tested by a fast freezing and thawing test machine for concrete in NaCl solution of 1.5% concentration and the test was finished when the freeze–thaw cycles reached 300.

(2) The concrete specimens respectively with the flexural stress level of 15%, 30%, 40%, and 50% were put in the water level fluctuation area of a large indoor seawater simulation test chamber. The chloride concentration was 1.5% and the temperature was 20 °C. The flooding and ebbing tide occurred four times a day. The test time was the same as the fast freezing and thawing test.

(3) With the same experimental program, two groups of control specimens without loading were respectively put in the fast freezing and thawing test machine and the water level fluctuation area of the large indoor seawater simulation test chamber.

2.3. Experimental Methods

Powder samples were obtained from the different depths from the exposed surface of each concrete specimen at the tension zone of the pure flexural load by grinding using a diamond drill. The chloride content in the powder samples was measured by a 785DMP automatic potentiometric titrator made in Metrohm AG (Herisau, Switzerland).

The microstructure and composition of the interfacial transition zone were determined by JSM6490 scanning electron microscopy and an energy spectrometer which was made by a Japanese electronics firm. The acceleration voltage was 5–20 kV, and the scanning depth was about 5–10 μm.

The pore structure was determined by the series Pascal140 and Pascal240 mercury porosimetry made by Thermo Finnigan (Waltham, MA, USA).

3. Results and Analysis

3.1. Influence of Flexural Loading on the Chloride Diffusion Coefficient of Concrete

The chloride diffusion in concrete by concentration grade can be described by Fick’s second law. The chloride diffusion coefficient is an important index of the chloride penetration resistance of concrete. The chloride diffusion coefficient can be calculated by fitting the nonlinear regression analysis of the least squares method using the chloride content in the powder samples from the different depths and the solution of the error function of Fick’s second law. The influence of the flexural stress level on the chloride diffusion coefficient of concrete in a seawater simulation test chamber is shown in Figure 2.

As shown in Figure 2, the chloride diffusion coefficient of concrete in the seawater simulation test chamber increased with the increase in the stress level. When the stress increased from 0% to 50%, the chloride diffusion coefficient increased from 0.83 × 10⁻¹² m²/s to 1.60 × 10⁻¹² m²/s. The chloride diffusion coefficient of the specimens under a stress level of 50% is approximately double that of the specimens without load. The permeation rate of chloride in concrete was accelerated by the flexural load, and the ability to resist chloride diffusion was then reduced. By fitting the relationship between the chloride diffusion coefficient and the stress level, an expression can be obtained as y = Ae^Bx as shown in Equation (1).

D = 0.8336e^1.3231σ,     R² = 0.9973

(1)

where D and σ are the chloride diffusion coeffaicient and flexural stress loaded on the concrete specimen, respectively.

It can be seen in Equation (1) that the chloride diffusion coefficient has an exponential relationship with the flexural stress level.

Under the flexural load, the microstructure of concrete is altered by the tensile stress in the tension zone of the concrete specimens where the surface and interior produce micro-crack or even macro-crack damage. Under the flexural load, the critical energy of the matrix crack generation and expansion is reduced because the binding force between the hydration products and the aggregate particles in the concrete is relatively weaker. The direction of crack expansion and the normal stress of the load is vertical. The generation and expansion of each new crack will reduce the effective bearing area, further increase the stress at the critical crack tip, and promote the deformation and expansion of the crack, thus forming a crack bridge and connecting it with other cracks. The degradation phenomenon of the microstructure is more obvious under the greater flexural load. The formed cracks offer a convenient channel for chloride penetration, resulting in the reduction of the ability of concrete to resist chloride diffusion. Therefore, the influence of the flexural load should be considered based on the chloride diffusion of concrete.

3.2. Coupling Action of Flexural Load and Freeze–Thaw on the Chloride Diffusion Coefficient of Concrete

The coupling influence of flexural stress level and freeze–thaw on the chloride diffusion coefficient of concrete is shown in Figure 3.

As shown in Figure 3, the chloride diffusion coefficient of concrete under flexural load and freeze–thaw increased with the stress level. The chloride diffusion coefficient of concrete is 1.50 × 10⁻¹² m²/s under the stress level of 30%, while the chloride diffusion coefficient of concrete without load is 0.84 × 10⁻¹² m²/s. When the stress level increased to 50%, the chloride diffusion coefficient of concrete reached 1.97 × 10⁻¹² m²/s. The permeation rate of chloride in concrete is accelerated by the coupling action of flexural load and freeze–thaw. By fitting the chloride diffusion coefficient change with the stress level, an exponential equation similar to the case obtained for flexural load only y = Ae^Bx can be obtained as shown in Equation (2).

$$\mathrm{D}={0.8777\mathrm{e}}^{1.668\mathrm{\sigma }}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}{\mathrm{R}}^2=0.9864$$

(2)

It can be seen in Equation (2) that the chloride diffusion coefficient under the conditions of both the flexural load and freeze–thaw increases with the flexural stress level more significantly, compared with the flexural load only, as shown in Equation (1).

The results of the chloride diffusion coefficient from the seawater simulation test chamber and the freezing and thawing test chamber for the concrete specimens under flexural load were compared, as shown in Figure 4.

As shown in Figure 4, for the concrete specimens without any load, the chloride diffusion coefficient of concrete under freeze–thaw is slightly higher than the one without the freeze–thaw process. The chloride diffusion coefficient of concrete under flexural load and freeze–thaw or under the flexural load only increased with the increase in the flexural stress level. Under the same stress level, the chloride diffusion coefficient of concrete under freeze–thaw is higher than those under the flexural load only, and the difference between them increased with the increase in the flexural stress level. It has been illustrated that the ability of concrete to resist chloride diffusion is reduced by the freeze–thaw process. The greater the stress that is loaded, the more obvious the increase in the chloride diffusion coefficient with the freeze–thaw process.

During the freeze–thaw cycle process, the tensile stress of internal concrete is produced by hydrostatic pressure due to the frozen pore water. With the development of cracks in concrete, new cracks of concrete are produced simultaneously. The freeze–thaw cycles lead to freeze–thaw damage, and the damage degree of concrete increases with the increase in freeze–thaw cycles, resulting in more and greater permeability channels in concrete, thus increasing the permeability of concrete [14,28]. The temperature of the freezing and thawing test chamber is around 0 °C for the no-load specimens which is lower than the temperature in the seawater simulation test chamber for the specimens only under the flexural load. The lower the temperature of the specimens, the lower the chloride diffusion speed obtained [29]. Therefore, the chloride diffusion speed and chloride diffusion coefficient of concrete in the freezing and thawing test chamber should be lower than that in the seawater simulation test chamber. However, the chloride diffusion coefficient of concrete in the freezing and thawing test chamber is slightly higher than the one in the seawater simulation test chamber as shown in Figure 4. It is shown that the aggravation of the damage and degradation in concrete and the increase in the width and number of cracks lead to the increase in chloride diffusion velocity by enhancing the freeze–thaw cycles [30]. This influence is greater than the reduction of the chloride diffusion coefficient of concrete in the freezing and thawing test chamber caused by the low temperature.

In addition, under flexural load and freezing–thawing, when the cracks and micro-crack damage in the concrete are caused by flexural load, the expansion of damage is further increased by freeze–thaw cycles. It is equivalent to increasing the level of tensile stress in the concrete. According to previous research [31], it can be seen that the chloride diffusion coefficient and the tensile stress level follow an exponential relationship. In addition, the greater the stress level, the greater the influence on the chloride diffusion coefficient. The difference between the chloride diffusion coefficient of concrete under the coupling action of flexural load and freeze–thaw and the flexural load only increased with the increase in the flexural stress level. In the range of the test stress levels in this study, the influence of flexural load on the chloride diffusion coefficient is higher than the influence of freezing–thawing.

3.3. The Microstructure of Concrete under Flexural Load and Freezing–Thawing

3.3.1. Influence of Bending Load on Concrete Pore Structure

In this project, the mortar/concrete pore structure measurement (MIP) method was used to test the pore structure of mortar. The pore structure of concrete includes some parameters, such as porosity, average pore diameter, and pore diameter distribution. The types of pore structures in concrete include gel pores (<10 nm), transition pores (10~50 nm), pores (50~100 nm), and micropores (>100 nm). Generally, in the pore size distribution of concrete, the higher the proportion of gel holes and transition holes, the better the durability of the concrete.

After curing the specimen to the specified age, the same stress level as the concrete specimen is applied, and it is placed in the corresponding exposure environment. The influence of stress level on the porosity and pore size distribution of the paste specimen is shown in

Table 3. Influence of bending load on the pore structure of the hardened paste.

Stress Level	Total Pore Volume (cm3/g)	Average Pore Diameter (nm)	Pore Diameter Distribution (%)
			Gel Pore < 10 nm	Transition Pore 10~50 nm	Pore 50~100 nm	Micropore >100 nm
0%	0.085	19.92	18.66	59.64	16.92	4.78
15%	0.121	21.34	16.42	57.35	18.26	7.97
30%	0.141	23.72	14.34	57.46	18.43	7.77
50%	0.164	24.82	13.37	56.37	22.41	7.85

.

It can be seen from Table 3 that the total pore volume and average pore diameter of the slurry increase with the increase in bending stress level. The average pore diameter of the slurry under no load is 19.92 nm. When the stress level is 50%, the average pore diameter of the test piece reaches 24.82 nm, which is 25% higher than that of the 0% stress level. At the same time, it can be seen from the table that the proportion of gel pores and transition pores decreases with the increase in the stress level, and the proportion of gross pores and large pores increases with the increase in the stress level.

The proportion of gel pores and transition pores decreases with the increase in the load level, and the proportion of capillary pores and macropores increases with the increase in the load level, which indicates that the bending load will not only increase the pore volume and the average pore size of the slurry but also change the pore size distribution of hardened slurry, increasing the proportion of capillary pores and macropores. The research [32,33] shows that the higher the proportion of coarse pores and large pores, the lower the chloride ion penetration resistance of concrete. Therefore, increasing the bending load of concrete will change the pore structure of the concrete, thus reducing the chloride ion penetration resistance of the concrete.

3.3.2. Influence of Freeze–Thaw Cycles on the Pore Structure of Concrete

After curing to the age of 28 d, the specimens were placed in a freeze–thaw test chamber, and part of the samples was taken for pore structure measurement during 100, 200, and 300 freeze–thaw cycles. The influence of freeze–thaw cycles on the porosity and pore size distribution of composite admixture specimens is shown in

Table 4. Influence of freeze–thaw cycles on the pore structure of hardened paste.

Number of Cycles	Total Pore Volume (cm3/g)	Average Pore Diameter (nm)	Pore Diameter Distribution (%)
			Gel Pore < 10 nm	Transition Pore 10~50 nm	Pore 50~100 nm	Micropore >100 nm
0	0.074	13.01	23.21	52.40	15.70	8.69
100	0.070	17.39	16.07	56.39	17.89	9.63
200	0.090	19.07	10.61	61.48	18.79	9.11
300	0.088	21.64	9.34	61.27	21.88	7.49

.

It can be seen from Table 4 that the total pore volume of the slurry and the average pore diameter of the slurry generally increase gradually with the increase in freeze–thaw cycles. The average pore diameter of the slurry without freeze–thaw erosion is 13.01 nm. When the freeze–thaw cycles reach 300, the average pore diameter of the hardened slurry reaches 21.64 nm, with an increase of 67%. The proportion of macropores does not change significantly with the increase in freeze–thaw cycles, while the proportion of capillary pores increases with the increase in freeze–thaw cycles. The proportion of capillary pores increases when the total pore volume does not change significantly, indicating that the average pore diameter of hardened paste increases, while it also shows that the average pore diameter of paste increases with the increase in freeze–thaw cycles.

3.3.3. Influence of Bending Load and Freeze–Thaw Interaction on Concrete Pore Structure

The pore structures of concrete under the coupling action of the flexural load and freeze–thaw after 300 freeze–thaw cycles and under flexural load only at the same age are shown in

Table 5. Influence of test environment and stress level on the pore structure of the concrete.

Test Environment	Stress Level	Total Pore Volume (cm3/g)	Average Pore Diameter (nm)	Pore Diameter Distribution (%)
				Gel Pore < 10 nm	Transition Pore 10~50 nm	Pore 50~100 nm	Micropore >100 nm
freezing and thawing test chamber	0%	0.101	21.41	15.12	51.17	21.92	11.79
	15%	0.142	23.59	13.04	48.42	27.14	11.40
	50%	0.194	28.72	9.51	46.17	30.34	13.98

and Figure 5.

As shown in Table 5 and Figure 5e,f, for concrete under load only, the total pore volume and average pore diameter increased with the increase in the level of flexural stress, and similar trends occurred in the proportion of pores and macro-pores. When the stress level increases from 0% to 50%, the total pore volume of the concrete increases from 0.085 cm³/g to 0.164 cm³/g. The total pore volume and average pore diameter of concrete under flexural load and freeze–thaw increased with the increase in the stress level, and similar trends were observed in the proportion of pores and micropores.

It can be seen from Figure 5a,b that the gel pores and transition pores decrease with the increase in the pressure level. The harmful pores and micropores tend to increase with the increase in stress levels from Figure 5c,d. Compared with the pore structure under the same stress level in different test environments, the total pore volume and average pore diameter of concrete under flexural load and freeze–thaw are higher than that under flexural load only. On the contrary, the proportion of gel pores and transition pores is lower than that under flexural load only. It has been shown that the influence of flexural load and freeze–thaw on the pore structure of concrete is higher than that under flexural load only.

3.3.4. Micro-Cracks under the Interaction of Bending Load and Freeze–Thaw

There are many micro-cracks or defects in concrete. Under the action of external forces such as flexural load and freezing–thawing, the stress concentration occurs around these cracks or defects. According to the Griffith crack theory, the relationship between internal stress and the length of cracks in concrete can be represented as Equation (3).

$${\sigma }_e=\sqrt{\frac{2E\gamma }{\pi c}}$$

(3)

where σ_e is the critical stress of a stable crack, E is the Young modulus, c is the length of the crack, and γ is fracture surface energy.

Equation (3) was corrected by Rakesh and T. Luping [32,34] and the parameter average pore diameter $r_m$ was introduced as shown in Equation (4).

$${\sigma }_e=\sqrt{\frac{2E_0{T}_0}{\pi }}\frac{\left(1-p\right)}{r_m}$$

(4)

* p: porosity

Equation (4) shows that a certain length of crack c corresponds to critical stress. When the applied stress σ > σ_e, the crack expands and leads to an increase in the average pore diameter of the concrete.

Due to the lower tensile strength of concrete, the cracks lose stability and expand when the tensile stress in the concrete under flexural load is increased and greater than the corresponding critical tensile strength. The external form appears to be that the number and width of cracks increase, and this leads to an increase in the porosity, average pore size, total pore volume, and the proportion of pores and micropores. The tensile stress of concrete under flexural load and freezing–thawing is greater than that under flexural load only; the change in the pore structure of specimens under flexural load and freezing–thawing is then more obvious.

The interface transition zone between the paste matrix and the aggregate is considered the weakest region in concrete, and it is also the most important area that affects the chloride diffusion coefficient of concrete. The microstructures of the interface transition zone of the specimens under flexural load and freezing–thawing, or under flexural load only, were observed by SEM as shown in Figure 6.

As shown in Figure 6, the interface between the aggregate and the paste of concrete under no load is dense, and no obvious pores or cracks can be observed in Figure 6a. Under the flexural load, the interface between the aggregate and the paste produces cracks, and the width of the crack of concrete under a 30% stress level is bigger than that under a 15% stress level in Figure 6b,d, or in Figure 6c,e. By comparing Figure 6b,c, or Figure 6d,e, it can be seen that under the same stress level, the width of a crack in the interface transition zone of concrete under flexural load and freezing–thawing is greater than that under flexural load only. It can be seen that the coupling action of flexural load and freezing–thawing on the microstructure of concrete is greater than that under flexural load only at the same stress level.

4. Conclusions

The chloride diffusion coefficient of concrete both under flexural load and freezing–thawing and under flexural load only all increased with the increase in flexural stress. The flexural stress level has an approximately exponential relationship to the chloride diffusion coefficient. The ability to resist the chloride diffusion of the concrete is reduced by freezing–thawing. The greater the load is, the more obvious the decrease in the chloride penetration resistance of concrete under flexural load and freezing–thawing. Under the same stress level, the decrease in the ability to resist the chloride diffusion of concrete under flexural load and freezing–thawing is higher than that only under flexural load.

SEM and other microstructural analysis confirm that the change in the pore structure of concrete can be caused by flexural load and the freezing–thawing process or by flexural load only. The increase in the stress level increased both the average pore size and the total pore volume of the concrete and altered the pore size distribution as well. The influence of flexural load and freezing–thawing on the pore structure of concrete is greater than that under flexural load only. The change in the chloride diffusion coefficient is mainly caused by modification of the pore structure.