Study on Concrete Deterioration and Chloride Ion Diffusion Mechanism by Different Aqueous NaCl-MgSO₄ Concentrations

Abstract

Chemical erosion of reinforced concrete by Cl⁻, SO₄²⁻ and Mg²⁺ in saline soil is the main factor of steel corrosion and concrete damage. In this study, the effects of different molar ratios of aqueous NaCl-MgSO₄ on concrete macroscopic properties (appearance, weight change, compressive strength, and dynamic elastic modulus), ion content, microstructure, and porosity of concrete were investigated. The effects of different molar ratios on the macroscopic characteristics and erosion depth of concrete were revealed through concrete appearance, weight, mechanical properties, and SO₄²⁻ and Cl⁻ content. Analysis of the microstructural evolution process and complex mineral composition of concrete using various microscopic testing methods. The results showed that with increased salt concentration and erosion time, the weight change rate, compressive strength change rate, and relative dynamic elastic modulus of concrete samples had a trend of first increasing and then decreasing. The evolutionary process of transition from large pores to medium and small pores and then to large pores. In the early erosion stage, with increased MgSO₄, corrosion products were deposited in pores and cracks, which refined the concrete pore structure and reduced ion diffusion speeds of Cl⁻, SO₄²⁻, and Mg²⁺. In the later erosion stage, corrosion products caused matrix damage and produced intersecting cracks, which promoted ion diffusion rates and induced deterioration of concrete macroscopic properties. During experiments, the binding ability of SO₄²⁻ and Mg²⁺ ions to hydration products was found to be higher than that of Cl⁻.

1. Introduction

The durability of concrete is very important for the normal use of infrastructures, such as roads and railways. Among these, erosion by sulfate and chloride salts (SO₄²⁻ and Cl⁻) in saline soil and salt lake environments is the key factor that affects the durability of concrete structures [1,2,3]. This erosion of concrete is a complex problem that has been increasingly studied over the past few decades [4]. SO₄²⁻ and Cl⁻ diffuse into the interior of reinforced concrete through pores in the matrix, interface transition zones (ITZ), and cracks, with Cl⁻ mainly destroying the passive film on the surface of steel rebar, thus causing steel corrosion [5,6,7]. SO₄²⁻ mainly has complex physical and chemical reactions with cement hydration products, causing the destruction of concrete structures. In a salt lake and saline soil environment, there are a lot of Cl⁻, SO₄²⁻, and Mg²⁺ ions, which will diffuse into the concrete and have complex physical–chemical reactions with cement hydration products, resulting in a complex deterioration mechanism of reinforced concrete structures [8,9].

At this point, scholars at home and abroad have performed a large number of NaCl erosion tests and studied the diffusion and microstructure of Cl⁻ in concrete. Li Lin [10] found that the Cl⁻ diffusion coefficient gradually increased with the increase in stress level and ambient temperature through Cl⁻ erosion concrete tests under different environmental conditions and the mechanical properties of concrete decreased after prolonged erosion. Audenaert et al. [11] established the diffusion equation of Cl⁻ in concrete using Fick’s second law and found that the diffusion rate of Cl⁻ in concrete varied after different erosion times. Wang [12] has found that the essence of Cl⁻ diffusion in concrete is the migration of charged particles in the pore solution, and the driving force of diffusion is mainly concentration differences between the inside and outside of the concrete as well as the pressure difference, humidity difference, and convection of pore liquid. Q. Yuan [13] has found that after entering concrete, a portion of Cl⁻ will react with cement hydration products to form Friedel’s salt, which is called “bound” chloride ions, and another portion continues to diffuse into the concrete, called free chloride [14,15]. Among them, free chloride ions will continue to diffuse inward through pores, cracks, and ITZs to accumulate around the steel rebar [16,17]. When the concentration of free Cl⁻ on the steel surface is higher than a certain value, free Cl⁻ can destroy the bar’s passivation film, and form a corrosion micro-battery with Fe [18], which accelerates steel corrosion by Cl⁻ and disturbs the relationship between steel bar and concrete and reduces their bonding performance [19].

Relevant scholars have studied the corrosion products, microstructure, mechanical properties, and mechanism of aqueous SO₄²⁻ and Mg²⁺ erosion of concrete [20,21,22]. SO₄²⁻ diffuses into concrete and reacts with cement hydration products (calcium hydroxide and silicate), forming sulfate and aluminate hydration products and further reacting with SO₄²⁻ to produce ettringite (AFt). Another product of the reaction between calcium hydroxide and sulfate is gypsum (Gyp). Among these, AFt has a high expansion ability, which can lead to matrix cracking or spalling [23,24,25]. Calcium silicate hydrate (C–S–H), a hydration product of cement, reacts with sulfate, resulting in decalcification and softening of calcium silicate hydrate, thereby reducing the strength of hydration products [26,27]. In addition, it has been found in previous research that the MgSO₄ solution deteriorates concrete faster than the Na₂SO₄ solution [27]. This is mainly due to the reaction of Mg²⁺ with cement hydration products during its diffusion into concrete to produce brucite (Mg(OH)₂), as an M–S–H gel. However, the binding properties of M–S–H gels are poor. Therefore, in the process of aqueous SO₄²⁻ and Mg²⁺ erosion of concrete, deterioration is more serious [28,29].

However, there are many types and different concentrations of ions in saline soil and salt lakes. The process of concrete erosion is a complex physical and chemical reaction and the impact on concrete durability is extremely complex. Until now, research on the diffusion process of Mg²⁺ and SO₄²⁻ to Cl⁻ during external erosion of concrete with different molar ratios of NaCl/MgSO₄ composite solution is less, and no relevant theoretical model has been established. To make up for this lack of research on the mechanical properties, in this study Cl⁻ diffusion, pores, and microstructure of concrete in aqueous NaCl/MgSO₄ environments were investigated in terms of appearance changes, weight loss, and mechanical properties, including compressive strength and relative dynamic elastic modulus (DEM) of concrete. The evolution process of fine microstructure was also examined using different molar ratios of NaCl/Na₂SO₄ (0, 0.2, 0.5, 0.7, 1), revealing the effects of different concentrations of MgSO₄ on the migration of Cl⁻ ions. The pore structures from different solutions after 30, 180, and 360 d were studied using a capacitive mercury porosimeter. Scanning electron microscope (SEM) and X-ray diffraction (XRD) were used to analyze the changes in concrete microstructure, corrosion products, and the evolution of ITZ with MgSO₄ concentration.

2. Materials and Methods

2.1. Materials

In this study, Portland cement of P.O. 42.5 was selected to prepare cubic concrete specimens. The chemical composition of cement is shown in

Table 1. Chemical composition of cementitious materials.

Chemical Composition	Al2O3	SiO2	SO3	Cl	TiO2	Fe2O3	Na2O	K2O	MgO	CaO
Content (%)	5.08	20.1	2.02	0.028	0.341	2.94	0.700	0.350	1.50	60.7

(cement manufacturer: Jidong Heidelberg Jingyang Cement Co., Ltd. (Xianyang, China)). The diameters of coarse aggregate particles were between 2.5 and 10 mm. The fine aggregate is made of river sand (production place: Xingping, Shaanxi), with fineness modulus of 3.24 and water content of about 8%. Distilled water was used to mix concrete and salt solutions. The Mg₄SO₄ reagent (analytically pure, 99.5%) was obtained from Tianjin Beichen Fangzheng Reagent Factory (Tianjin, China). The mix of C35 concrete is shown in

Table 2. Mixture proportion of concrete in the present study.

Concrete Grade	w/c	Cement (kg/m3)	Sand (kg/m3)	Gravel (kg/m3)	Water (kg/m3)
C35	0.4848	402.16	631	1199	195

.

2.2. Sample and Solution Preparation

The corrosion conditions of different concentrations of NaCl-MgSO₄ solutions, five test chambers were placed with different molar ratios of composite solutions, named C; CS0.2-M; CS0.5-M; CS0.7-M; CS1-M; details of these solutions are shown in

Table 3. Corrosion conditions of specimens.

Specimen	Molar Ratio	Corrosive Environment (Mol/L)		Specimen	Molar Ratio	Corrosive Environment (Mol/L)
		NaCl	Na2SO4			NaCl	Na2SO4
C	0	1	0	CS0.7-M	0.7	1	1
CS0.2-M	0.2	1	0.2	CS1-M	1	1	1
CS0.5-M	0.5	1	0.5

. The solutions and samples were checked every month to ensure full solution immersion during experiments and that the concentrations of Mg²⁺, SO₄²⁻, and Cl⁻ were constant.

Before the concrete was poured, coarse aggregate, river sand, cement, and water, were poured into the mixer and mixed for 3 min. The prepared concrete was then poured onto a wet iron plate to produce the samples of 100 × 100 × 100 mm and 100 × 100 × 300 mm. The specimens were left at room temperature for 24 h and then transferred to a standard maintenance room (temperature 20 + 2 °C, humidity 95%). After standard curing for 28 d, four surfaces were covered with epoxy resin, with two opposite sides uncovered. After the surface coating dried, specimens were placed in different molar ratios of the compound solution.

2.3. Procedures of Experiment

After soaking for 30, 90, 180, 270, and 360 d, the macroscopic parameters of the samples were measured. Before testing, they were placed in a 50 °C oven for 48 h to ensure a consistent test environment.

2.3.1. Weight, Compressive Strength, and Relative DEM

The weight, compressive strength, and relative dynamic modulus of elasticity of the specimens after different times of erosion were tested using an electronic scale with an accuracy of 0.01 g, an HYE-2000BS electro-hydraulic servo pressure tester (Hebei Sanyu Testing Machine Co., Cangzhou, China), and an NELD-DTV type dynamic modulus of elasticity tester (Beijing Naerde Instrument Equipment Co., Ltd., Beijing, China). Wherein, the average value of 3 specimens was taken as the test value for the compressive strength and dynamic elastic modulus tests.

2.3.2. Cl⁻ and SO₄²⁻ Concentrations

Concrete powder from different depths under a sample’s eroded surface was obtained by impact drilling, sieved through a 0.63 mm diameter sieve, and baked in an oven at 105 °C for 2 h until the weight of the test powder was constant. Then, 2 g samples of powder after different sample soaking times and depths were mixed with 50 mL of distilled water, placed in a constant temperature vibrator for 5 min, and then allowed to stand for 48 h to ensure that free ions were fully dissolved into the water. The dissolved solution was then filtered to remove residues and the Cl⁻ and SO₄²⁻ concentrations in each sample were measured by a CLU-W type Cl⁻ content rapid analyzer (Beijing Zhongke Road Jian Instrument Co., Ltd., Beijing, China) and an ultraviolet–visible spectrophotometer (Shimadzu Corporate Management (China) Co., Shanghai, China). The average value of the three eroded specimens was measured as the relative concentration of Cl⁻ and SO₄²⁻.

2.3.3. SEM, EDS, and XRD Tests

A scanning electron microscope (Zeiss Sigma 300, Jena, Germany, place of manufacture: Shenzhen Hualing Precision Electromechanical Co Ltd, Shenzhen, China) was used to observe the microscopic morphology and elemental analysis of the specimens at the surface layer (0–10 mm) after different erosions. Prior to testing, the specimens were soaked in anhydrous ethanol to prevent hydration of the cement and baked in an oven at 50 °C to dry. X-ray diffraction (Shanghai Erdi Instrument Technology Co., Shanghai, China) was used to analyze the composition of the corrosion products on the surface layer (0–10 mm) after different erosions to test the corrosion products at the eroded surface layer (0–10 mm).

2.3.4. Pore Analysis

A YG-97A capacitive mercury piezometer (Wuxi Huio Instruments and Equipment Manufacture Co., Ltd., WuXi, Jiangsu, China) was used to measure the pore throat distribution and cumulative mercury saturation at 0–25 mm after 30 d, 180 d, and 360 d of erosion and 25–50 mm after 360 d of erosion, respectively. A TZ-1 benchtop rock corer drills a 25 mm–75 mm cylinder from the erosion surface and splits the specimen with a QP-100 benchtop high power slicer to obtain 0–25 mm and 25 mm–50 mm specimens. Before testing, the specimen needs to be baked in a 50 °C oven for 8 h to reduce the effect of moisture on the test results.

3. Results

3.1. Weight Changes and Appearance of Performance

After samples were soaked in different concentration solutions for 30, 90, 180, 270, and 360 d, the weight change rate of test samples was calculated using Equation (1). Photographs and weight change rates are shown in Figure 1 and Figure 2.

$$\Delta W(n)=\frac{W_n-W_0}{W_0}\times 100\%$$

(1)

where △W is the weight change after soaking for t days, %; W₀ is the weight before soaking (after 28 days of standard curing), g; W_n is the weight after soaking for n d, g.

The appearance changes in samples after immersion for different times under different molar ratios of NaCl/MgSO₄ (0, 0.2, 0.5, 0.7, 1) showed that, compared with C solution samples, a layer of white powdery material adheres to the surface and pores of the specimen after erosion of the composite solution containing MgSO₄ (Figure 1). With increased MgSO₄ in solution and increased soaking time, the corrosion and peeling phenomenon of the edges and corners of samples gradually advanced and increased, with the aggregation of white powder crystals in surface pores gradually increasing. After 360 d of erosion, the peeling phenomenon on the concrete surface was clearer. During this process, C sample surfaces basically did not change significantly. Therefore, the deterioration rate of the concrete surface in the other solutions gradually increased with increased MgSO₄ and erosion time.

The weight of concrete samples was generally affected by ionic corrosion, cement hydration products, and the degree of sample surface peeling. The weight change rate of samples under different molar ratios and erosion times showed that the weight under NaCl-MgSO₄ erosion exhibited a trend of first increasing and then decreasing (Figure 2). In the early stage of erosion (before 30 d), sample weights in solutions containing only NaCl (C) increased slowly and then increased slightly in the later stage. The weight change rate under the erosion with MgSO₄ increased with increased erosion time. After 30 d of erosion, CS0.7-M and CS1-M samples showed a downward weight change rate trend, while that of samples in other solutions continued to increase. The rates of CS0.2-M and CS0.5-M began to decrease after 180 and 90 d of erosion, respectively, but the quality of CS0.2-M samples was always higher than before erosion. After CS0.5-M, CS0.7-M and CS1-M were eroded for 120, 240, and 300 d, respectively, the weight change rate was below the x-axis, and the concrete weight after erosion was smaller than before erosion. The weight changes in concrete samples were found to show the same results as the observations of appearance.

The changes in concrete weight and appearance were mainly related to erosion time, ion species, and concentration gradients. The results showed that, with increased molar in solution, the ion diffusion rate into the concrete was accelerated. At this time, Cl⁻, SO₄²⁻, and Mg²⁺ ions diffused into the concrete and reacted with the hydration products, resulting in a large number of corrosion products deposited in the pores and cracks, which might have affected the concrete pores, ITZs, and micro-cracks. With extended erosion age, the weight and appearance of concrete samples changed more clearly. The appearance and weight changes in samples in the solutions containing only NaCl were lower, indicating that the influence of Cl⁻ on the appearance and quality of concrete was lower.

3.2. Compressive Strength and Relative DEM

3.2.1. Compressive Strength Changes and Failure Mode

Uniaxial compressive strength reflects the relationship between different erosion times, concentrations, and compressive strength on a macro scale, visually reflecting the damage characteristics of concrete in different environments. According to Equation (2), the rates of change in concrete compressive strength in different immersion solutions were calculated and expressed as:

$$\Delta f_c(n)=\frac{f_{c.n}-f_{c.0}}{f_{c.0}}\times 100\%$$

(2)

where $\Delta f_c(n)$ is the rate of change in compressive strength of concrete after n days of immersion, %; $f_{c.0}$ is the compressive strength of concrete before soaking (after standard curing for 28 d), MPa; $f_{c.n}$ is the compressive strength after soaking for n d, MPa.

The change in compressive strength of concrete after erosion for certain times under different molar ratios of NaCl/Na₂SO₄ (0, 0.2, 0.5, 0.7, 1) showed that change trends of concrete compressive strength under NaCl-MgSO₄ solution erosion were basically similar (Figure 3). There appeared to be three stages: rapid increase, slow increase, and decline. CS0.2-M, CS0.5-M, CS0.7-M, and CS1-M samples gradually entered the descending stage after being eroded for 180, 180, 90, and 30 d, respectively. With increased MgSO₄ concentration, the change rate of compressive strength gradually increased. After 360 days of erosion, C, CS0.2-M, CS0.5-M, CS0.7-M, and CS1-M concrete compressive strength change rates were 38.38, 37.22, 18.96, −6.10, and −18.61% of before erosion, respectively. It can be seen that as the concentration of MgSO₄ in the composite solution increases, the damage to the concrete gradually increases.

The change in damage appearance after erosion for 180 and 360 d from compressive strength tests of samples showed that the damage degree in solutions containing MgSO₄ was significantly higher than for samples in only NaCl solution (Figure 4). After 180 d of erosion, with increased MgSO₄ concentration, cracks on sample surfaces gradually widened and extended, shifting to the middle of samples. CS0.7-M had cracks that ran through each other and concurrent surface peeling. The mortar layer on the surface of CS1-M was severely peeled off and after 360 d of erosion, CS0.2-M showed severe peeling with intersecting cracks that penetrated each other. There are many vertical cracks on the surface of CS0.5-M and CS0.7-M, which exposed a large amount of coarse aggregate. The mortar layer on the surface of CS1-M was also severely peeled off, which showed that in NaCl-MgSO₄ solution, the damage degree in the appearance and morphology of concrete after compressive strength testing increased with increased MgSO₄ concentration and prolonged erosion time.

From the above analytical results, the change in compressive strength was clearly seen to be related to cement hydration, SO₄²⁻ and Mg²⁺ concentrations in the erosion solution, and erosion time. As erosion time increases, the concrete structure at the eroded surface changes. Due to the complex physical–chemical reactions of Cl⁻, SO₄²⁻, and Mg²⁺ with cement hydration products result in the deterioration of the concrete microstructure, which in turn affects the compressive strength and damage characteristics of the concrete.

3.2.2. Relative Dynamic Modulus of Elasticity

Ultrasonic measurement of the DEM belongs to nondestructive testing, using ultrasonic waves transmitted into the tested concrete and judging the internal damage degree according to the attenuation characteristics of ultrasonic waves in the process of propagation through the object. According to Equation (3), the change trend of the relative DEM after different erosion regimes was calculated.

$$E_{rd}=\frac{E_n}{E_0}$$

(3)

where E_rd is the relative DEM of a sample after immersion for n days; E₀ is the DEM before immersion, MPa; E_n is the DEM after immersion for n days, MPa.

The relative DEM of concrete after erosion under different molar ratios of NaCl-MgSO₄ solution mainly included three stages: initial acceleration, slow increase, and falling stage, as shown in Figure 5. Before 90 d of erosion, the relative DEM of C, CS0.2-M, CS0.5-M, and CS0.7-M were in the initial acceleration stage, but the initial acceleration stage of CS1-M was before 30 d erosion, which showed that the DEM in the initial acceleration stage was mainly due to diffusion of Mg²⁺, SO₄²⁻, and Cl⁻ into the sample and complex physical–chemical reactions with cement hydration products. With prolonged erosion time, the relative DEM of C, CS0.2-M, and CS0.5-M gradually entered a slow-increasing stage. C, CS0.2-M, and CS0.5-M samples were eroded for 270, 180, and 180 d, respectively, and their relative DEM gradually entered a declining stage. While CS0.7-M and CS1-M directly enter the slow descent stage and then the accelerated descent stage, in the later stage of erosion, the surface or internal structure of these samples was damaged by corrosion products, resulting in new pores or cracks and a rapid decline in sample relative DEM. These results showed that the change trends of relative DEM after erosion in different concentrations were basically consistent with compressive strength. The effects of high-concentration Mg²⁺ and SO₄²⁻ solutions on the compressive strength of concrete after erosion were further found to be greater than that of the relative DEM.

3.3. Chlorine Salt and Sulfate Concentration

The relative contents of Cl⁻ and SO₄²⁻ in samples at different depths during the erosion process in different solutions. The relative contents of Cl⁻ and SO₄²⁻ at different depths after erosion of different molar ratios of the composite solution were placed on the same axis for comparative analysis (Figure 6).

For different molar ratios of NaCl/MgSO₄ solutions during erosion experiments, the distribution of relative content of SO₄²⁻ and Cl⁻ content was consistent. The relative contents of Cl⁻ and SO₄²⁻ in different solutions were clearly seen to be higher near sample surfaces and decreased rapidly with depth. In the early stage of erosion, SO₄²⁻ and Cl⁻ contents in the surface layer gradually increased with increased erosion time and MgSO₄ concentration. This was due to the large concentration gradient between the sample inside and outside, which was conducive to SO₄²⁻ and Cl⁻ diffusion into the samples through pores and micro-cracks. With longer erosion time, the concentration gradient difference between samples inside and outside gradually decreased, and the SO₄²⁻ and Cl⁻ diffusion rates gradually slowed. After 180 d of erosion, SO₄²⁻ and Cl⁻ contents at the surface of CS0.7-M and CS1-M significantly decreased, while the ion content increased at 5–10 mm depth, which was mainly due to serious damage to the surface layer and internal structure deterioration, resulting in faster inward ion diffusion. To further examine the effects of different MgSO₄ concentrations on the migration law of Cl⁻ ions, the change trend of Cl⁻ at 2.5 mm was examined from samples in different solutions and erosion times (Figure 6f). After 30 d of erosion, the Cl⁻ content resulting from the solution containing only NaCl was the lowest and the highest after 90 d. As the erosion time progressed, the Cl⁻ content in solution also containing MgSO₄ was higher than that in samples containing only NaCl. After 360 d of erosion, the order of Cl⁻ content at 2.5 mm from high to low was CS0.2-M, CS0.5-M, C, CS0.7-M, and CS1-M. This showed that the presence of SO₄²⁻ and Mg²⁺ ions in solution had a certain influence on Cl⁻ diffusion into concrete. At low concentrations and in the early stage of erosion, corrosion products inhibited Cl⁻ diffusion. As the molar ratio and erosion time in the composite solution increases, the high concentration of MgSO₄ concentration promoted inward Cl⁻ diffusion.

3.4. Microstructural and Mineral Analytical Results

The effects of erosion time and concentration on samples in different solutions after 30, 180, and 360 d of erosion time were examined in terms of their microstructures and mineralogy testing.

3.4.1. XRD Tests

In XRD diffractograms of NaCl-MgSO₄ solutions under different concentrations and erosion environments, the peak of XRD diffractograms was basically the same, with differences in the intensities of diffraction peaks (Figure 7). The intensity of the diffraction peaks of Friedel’s salt and calcium hydroxide (CH) in solution C gradually increased, mainly due to the formation of Friedel’s salt by the hydration product of Cl⁻ ions and cement hydrated calcium sulfoaluminate (AFm), and the release of SO₄²⁻ [30]. The peaks of Friedel’s salt and Ca(OH)₂ gradually decreased with increased MgSO₄ concentration in solution and extended erosion time. This was mainly due to the reaction of Mg²⁺ and SO₄²⁻ with cement hydration products, AFm, and Friedel’s salt, to generate M–S–H, Mg(OH)₂, AFt, and Gyp crystals, which were produced after prolonged erosion. Therefore, the diffraction peaks of Mg(OH)₂, M–S–H, Aft, and Gyp in the sample gradually increased, while the intensity of Friedel’s salt and CH diffraction peaks gradually weakened. This showed that in the process of eroding concrete by a salt solution containing MgSO₄, Cl⁻ reacted with AFm first to form Friedel’s salt and then SO₄²⁻ and Mg²⁺ reacted with Friedel’s salt, Ca(OH)₂, and other products to produce refractory substances, including AFt, Mg(OH)₂ and M–S–H. However, in high concentration or long-term erosion, the AFt diffraction peak decreased, and the Gyp diffraction peak gradually increased.

The corrosion characteristics of cement hydration products in the process of NaCl-MgSO₄ erosion of concrete were further verified using infrared spectroscopic analysis of samples eroded for 360 d (Figure 8). The vibrations of the SiO₂ band near 455 cm⁻¹ were related to the MgSO₄ concentration. With increased solution MgSO₄, the band near 996 cm⁻¹ and the Si–O–T bond in the C–S–H gel material moved in the direction of smaller wavenumbers and the S–O band near the 579 and 1065 cm⁻¹ also moved to smaller wavenumbers. The O–H bond signal near 579 and 1065 cm⁻¹ moved toward higher wave numbers, indicating that, with increased MgSO₄ and 360 d of erosion, C–S–H gel and Ca(OH)₂ crystals gradually decreased and AFt and Gyp generation gradually increased, which was basically consistent with XRD analytical results.

3.4.2. SEM and EDS Analysis

SEM and EDS analyses were performed on the samples after different erosion times. The mineral composition and microstructure of the concrete after attack by the composite solution shows that the corrosion products are mainly Friedel’s salt, calcium alumina (AFt), gypsum (Gyp), magnesium hydroxide, and magnesium silicate (Figure 9, Figure 10, Figure 11 and Figure 12) [31].

The microstructures resulting from erosion with NaCl-MgSO₄ solutions with different molar ratios for different times showed that C samples had a loose granular structure inside the matrix after 30 d of erosion, with a small number of rod-like structures and a network structure that gradually appeared longer erosion time (Figure 9). EDS analysis of the network structure was performed and for C–S–H cementitious material, more irregular sheet-like structures appeared in samples after 360 d, which were judged to be Friedel’s salt (Figure 10). With increased MgSO₄ and erosion time, the rod-like, short-column-like, and sheet-like structures in matrix pores and in the original fractures gradually increased and became denser. EDS analysis indicated that this material was AFt crystals (Figure 10b), Mg(OH)₂, and M–S–H crystals (Figure 10c), while the Ca(OH)₂ crystal structures at this time had a large number of sawtooth structures around it, with relatively thin thickness (Figure 10d). There were M–S–H crystals on the surface of rod-like AFt and Mg(OH)₂ sheet-like structures found in other positions of the matrix. After 90 d of erosion in MgSO₄ solution, AFt and columnar Gyp crystals in the matrix gradually increased and filled into pores and cracks. At this time, the flaky and flocculated gels gradually decreased. There were then interwoven clusters of AFt and short-column Gyp crystals in the fractures, and the hexagonal thin plate layers of Ca(OH)₂ and flocculated C–S–H gradually decreased. After 270 d of erosion, interconnected pores in the matrix gradually increased and the surrounding pores and bulk lamellar crystals appeared in the interior. After 360 d, the matrix compactness decreased, interconnected pores gradually increased, and lamellar structures gradually increased. However, in the process of CS0.7-M and CS1-M sample erosions, the corrosion products AFt and Gyp in the matrix appeared earlier than in other solutions, which increased concrete compactness in the early erosion stage in higher or high-concentration solutions. As the erosion time extended CS0.7-M and CS1-M matrices gradually increased and intertwined with each other. With the appearance of a large number of sheet-like and rod-like structures, the surface of the aluminum phase was covered by a large number of erosion products and there were interwoven cracks. After 360 d, the corrosion products did not completely fill the pores, and the internal structure of the concrete relatively loose, showing an isolated block structure.

Cracks in the matrix are one of the main causes of concrete failure. Figure 11 shows the crack evolution in the specimen matrix before and after 360 d of erosion of the NaCl-MgSO₄ composite solution. Before erosion, the overall internal structure of the concrete was loose, with a large number of disconnected and connected pores, indicating that the concrete had certain initial defects (Figure 11a). After 360 d, there were relatively fine cracks in C samples and a large number of C–S–H network cementitious structures in the cracks (Figure 11b). With increased solution MgSO₄ concentration, the internal independent cracks gradually intersected and penetrated each other, and crack widths gradually widened. There were fewer corrosion products inside CS0.2-M cracks (Figure 11c). A large number of needle-like AFt and C–S–H gels were found in the internal cracks of CS0.5-M (Figure 11d). In CS0.7-M (Figure 11e) and CS1-M (Figure 11f) samples, intersecting cracks appeared in the matrix, and crack lengths gradually extended and widened. There were a large number of rod-like structures of AFt, short columnar Gyp, and sheet C–S–M or Mg(OH)₂ in the cracks.

There was an ITZ with a thickness of 10–50 μm between aggregate particles and cement hydration products [32], which was the weakest part of the concrete and the channel through which ions can easily diffuse. ITZ also easily accumulates corrosion products, with the erosion time and solution concentration increasing, corrosion products gradually increase and produce radial expansion stress to the ITZ, resulting in ITZ damage and crack extension widening. This was the area where initial cracks were most prone to crack propagation. The evolution of NaCl-MgSO₄ composite solutions with different molar ratios eroded for 360 d showed that, during the process of erosion in C samples, no new cracks were generated due to only being eroded by Cl⁻ (Figure 12). The ITZ in CS0.2-M samples was filled with corrosion products, such as AFt, Gyp, Mg(OH)₂, and C–S–M, and new cracks were also generated. The corrosion products in CS0.5-M samples gradually increased, which made the aggregate and matrix tightly connected and converted into a relatively complete structure. As the concentration of MgSO₄ in the composite solution continues to increase, the corrosion products C–S–M and Mg(OH)₂ in the ITZ gradually increase, and the cracks in the ITZ gradually widen and extend at this time. Therefore, increased MgSO₄ concentration in the composite solution was found to cause the crack width in the ITZ to gradually transition from a tightly connected state to a crack, which would then be conducive to the diffusion of Mg²⁺, SO₄²⁻, and Cl⁻ and shorten the service life of the structure.

3.5. Pore Analysis

Pores are important components of concrete mesostructure. The changes in pore structure were directly related to ion diffusion and mechanical properties (compressive strength, relative dynamic elastic modulus). Pore analysis of the specimens after 30 d, 180 d, and 360 d erosion of NaCl-MgSO₄ composite solution was carried out using the YG-97A capacitive mercury compression instrument. The pores in the specimens after different times of erosion were divided into: small pores (<0.05 μm), medium pores (0.05 μm–10 μm), and large pores (10 μm–63 μm).

The distribution of pores between 0–25 mm below the eroded surface after 30, 180, and 360 d of NaCl-MgSO₄ erosion was examined (Figure 13a–e). In a solution containing only NaCl, with prolonged erosion time, the proportion of medium and small pores in the samples slightly increased and the cumulative mercury saturation gradually decreased. The proportion of pores from large to small was mesopores, small pores, and large pores. After NaCl-MgSO₄ erosion, the accumulated mercury saturation was significantly higher than that of only NaCl. In NaCl-MgSO₄ solutions, with increased erosion time and MgSO₄ concentration, the cumulative mercury saturation of concrete gradually increased, especially in high and high-concentration environments. With extended erosion time, the proportion of small pores in CS0.2-M samples first decreased and then increased, the proportion of large pores first increased and then decreased, and the proportion of medium pores did not significantly change. In CS0.5-M, the proportion of large pores increases rapidly and then decreases slowly, while the proportion of medium and small pores tends to decrease and then increase. In CS0.7-M, the proportion of small pores rapidly reduces first, and after the big pore proportion, increases slowly and then increase rapidly. In CS1-M, the proportion of small pores first rapid declines after a slow increase, pore ratio was increased gradually and slowly, and the big pore proportion basic did not change significantly. The proportion of 0.1–1 μm pores in CS1-M medium increased significantly.

The distribution of pores at different depths (0–25 and 25–50 mm) after being eroded by NaCl-MgSO₄ solution for 360 d showed that, when the NaCl-MgSO₄ molar ratio was <0.5, the proportion of small pores at 0–25 mm was significantly higher than at 25–50 mm and the proportions of macropores and mesopores at different positions exhibited opposite trends (Figure 14). As the concentration increased (molar ratio < 0.5), the gap between the proportions of mesopores and small pores decreased significantly. As the concentration further increased (molar ratio > 0.5), the proportion of medium and small pores at 0–25 mm in CS0.7-M was less than that at 25–50 mm, with large pores showing a downward trend. However, the proportion of macropores, mesopores, and small pores did not change significantly after the CS1-M solution treatment.

The results showed that, when the solution only contained NaCl salt, the effect on the pore structure in concrete was less after different times of erosion. With increased MgSO₄ concentration in the composite solution and erosion time, AFt, Gyp, Mg(OH)₂, and C–S–M gradually increased and filled into pores, cracks, or ITZs, changing the pore distribution, with large pores in the middle. As the MgSO₄ concentration continued to increase, crystal stress generated by corrosion products was higher than the tensile strength of the ITZ or matrix, resulting in the transformation of medium pores and small pores into large pores. Therefore, changes in pore proportions with different diameters were mainly due to the effects of ion erosion depth and corrosion products on pores and cracks, which further affected the concrete macroscopic properties.

4. Discussion

In this study, the damage trends and deterioration mechanism of concrete under erosion from different concentrations of NaCl-MgSO₄ (molar ratio: 0, 0.2, 0.5, 0.7, 1) were studied from multi-scale perspectives. At the same time, analyses using XRD, SEM, EDS, and YG-97A capacitive mercury piezometer revealed that different solutions led to different changes in concrete samples as well as their microstructure and pores.

During sample erosion with different NaCl-MgSO₄ concentrations, Mg²⁺, SO₄²⁻, and Cl⁻ gradually diffused into the concrete through pores, cracks, and ITZs. As the concentration gradient decreased poorly, pores, cracks, and ITZs were the main factors affecting ion diffusion. At the initial stage of erosion or when the molar ratio was < 0.5, with increased erosion time and MgSO₄ concentration, Mg²⁺, SO₄²⁻, and Cl⁻ reacted with C–S–H, AFm, and Ca(OH)₂ form Friedel’s salt, AFt, Gyp, C–S–M, and Mg(OH)₂ which filled into the matrix and interfacial transition regions. At this time, the expansion stress generated by the corrosion products was less than the tensile strength of the matrix, and the large pores within the matrix gradually shift to medium and small pores. The pore structure in the matrix is refined and the compactness of the concrete is improved. At this time, the diffusion rate of Mg²⁺, SO₄²⁻, and Cl⁻ was hindered and the mechanical properties of samples gradually increased. With increased salt concentration and prolonged erosion time, the corrosion products AFt, Gyp, C–S–M, and Mg(OH)₂ gradually increased and filled the interior of the pores. Meanwhile, Friedel’s salt showed a downward trend, but the volume of corrosion products produced in the pores was much higher than the capacity of the pores to hold them. At this time, the lateral stresses generated by corrosion products were higher than the matrix tensile strength such that micro-cracks and new cracks formed in the matrix, thus creating new channels and accelerated speed for ion diffusion. Concurrently, mesopores and small pores transitioned to medium pores, the proportion of medium pores increased, and sample compressive strength and DEM gradually began to decrease.

After high-concentration solution erosion (molar ratio > 0.7) after prolonged erosion, the corrosion products filled up new cracks and pores and the large pores transitioned to medium pores and small pores again. However, the compactness of the sample decreased, which was helpful for Mg²⁺, SO₄²⁻, and Cl⁻ diffusion. At this time, the concrete appearance showed serious spalling, quality, and compressive strength damage. Mainly due to the reaction of Mg²⁺, SO₄²⁻, and Cl⁻ with cement hydration products, the eroding concrete generated a large amount of AFt, Gyp, C–S–M, Mg(OH)₂ precipitates, thus consuming a large number of hydration products, such as Ca(OH)₂. The pH value in the matrix decreased and to maintain the pH in the cement, the C–S–H gel decomposes to form magnesium silicate hydrate (M–S–H) and Gyp, while M–S–H was a less viscous substance with lower strength, which in turn reduced the matrix strength and cohesion. Coupled with erosion by SO₄²⁻ and Cl⁻, a large number of corrosion products, including Friedel’s salt, AFt crystals, and Gyp crystals were produced inside the substrates. This led to the generation of new cracks and changes in the distribution of pores in the pores, ITZ, and matrix of the concrete. This led to accelerated damage to the concrete appearance, compressive strength, and relative DEM and created new channels for ion diffusion, thus ultimately reducing the service life of the concrete.

5. Conclusions

The damage trends of the macroscopic properties of concrete samples and changes in their microscopic structure after different times of erosion by different molar ratios of NaCl-MgSO₄ solution were revealed from multi-scale perspectives.

With increased erosion solution concentration and time, the mass change rate, compressive strength change rate, and relative DEM of samples showed a trend of first increasing and then decreasing, with the peak points different.

The binding ability of SO₄²⁻ and Mg²⁺ ions and hydration products in the solution was higher than that of Cl⁻.

The diffusion of corrosive ions into concrete samples was related to the solution MgSO₄ concentration. In early erosion, as the of MgSO₄ concentration increased, corrosion products were deposited in pores and cracks, which refined the pore structure and reduced ion diffusion rates, thus inducing deterioration of sample macroscopic properties.