Mechanical Properties and Chloride Penetration Resistance of Copper Slag Aggregate Concrete

Abstract

The authors of this paper systematically studied the mechanical properties and durability of concrete prepared with copper slag instead of natural aggregates. An analysis index was used to assess compressive strength, and a statistical method was used to establish a mix proportion design theory of copper slag aggregate concrete. The analysis was used to quantify the effect of copper slag aggregate concrete on resistance to chloride ion migration. Combined with the morphological analysis of SEM images and fractal calculations, the tests were used to explain the improvement mechanism of copper slag as a fine aggregate on concrete’s mechanics and durability from the microscopic mechanism perspective. The results showed that replacing a natural sand fine aggregate with copper slag improved the compressive strength of concrete, and the optimum replacement rate was found to be 40%. The influence of the water–cement ratio on the strength of copper slag aggregate concrete was exceptionally conspicuous—the more significant the water–cement proportion was the lower the compressive strength of the concrete. The optimum dosage of the water-reducing agent was found to be 3.8 kg/m³. A rapid chloride ion migration test and potential corrosion analysis showed that copper slag aggregate concrete’s initial density and corrosion resistance were higher than those of natural aggregate concrete. Electrochemical impedance spectroscopy analysis results showed that the structural concrete comprising copper slag aggregate instead of natural sand had a better anticorrosion effect on embedded steel bars. SEM morphology and fractal dimension analyses showed that the incorporation of steel slag aggregate decreased the initial damage to the concrete internal section.

1. Introduction

Concrete, the leading engineering construction material, is widely used in water conservancy, transportation, construction, and other industries. Currently, the world annually consumes tens of billion tons of concrete, requiring 1.7–2.0 tons of sand aggregate for every 1 m³ of concrete produced [1]. Large quantities of natural resources must be mined for raw materials, such as natural aggregate and limestone. According to the literature, the world annually mines about 32 to 50 billion tons of sand, most of which is river sand. Ore extraction is mainly conducted in China, India, and Africa. The excessive exploitation of this kind of nonrenewable resource damages the ecological balance of mines and rivers, causing severe environmental problems [2].

According to the World Bureau of Metals Statistics data, the global refined copper output was 23.72 million tons in 2019, with China alone reaching 41.25% (9.784 million tons) [3]. In the future, the figure is expected to increase. China’s traditional copper smelting technology is fire smelting. One ton of refined copper smelting produces 2.2 t of water-quenched copper slag [4]. Copper slag is still mainly stacked in a centralized way that occupies land and causes many environmental problems [5]. Furthermore, glassy copper slag may lead to debris flow and other safety problems during geological disasters. In 1996, the United Nations Basel Convention on Transboundary Movement and Disposal of Hazardous Wastes described copper slag as a harmless substance that poses no threat to the environment and can be used as an alternative material resource [6]. An experimental study on the heavy metal content of copper slag confirmed that the rich metal content of copper slag is lower than the recommended limit composition [7]. An analysis of copper slag’s physical and chemical properties showed that copper slag can be classified as a nontoxic aggregate. There is no expansion phenomenon in its mortar rod, proving that copper slag does not contain active silicate [8]. The use of copper slag as an aggregate in concrete does not affect volume stability. Using copper slag as a fine concrete aggregate can eliminate dumping costs, reduce land occupation area, and reduce the natural damage and energy consumption of industrial fine sand production to realize the rational utilization of solid waste resources. Accordingly, the study of copper slag could have substantial economic benefits and far-reaching social influence.

Research data show that copper slag fine aggregate concrete (CSAC) and natural sand concrete (NSC) have similar physical and mechanical properties. The substitution rate of copper slag for fine aggregate is a crucial index affecting the performance of concrete. Most scholars believe that the threshold of substitution rate affecting the compressive strength of concrete is 40% volume fraction [9,10]. Some of the literature shows that the entry of substitution rate should be between 50 and 80% volume fraction [11,12]. The results of the threshold analysis of copper slag replacement rate differ significantly, and scholars agree that replacing fine aggregate with copper slag improves the workability and mechanical properties of concrete [13,14]. However, if copper slag is irregular and granular with similar properties to the glass surface and low hygroscopicity, then the fine aggregate volume fraction exceeds the threshold value, and the free water content in concrete increases, leading to the decline in concrete strength [15,16].

Concrete is a porous material. Its microstructure directly determines the durability of concrete, and the corrosion of reinforcement induced by chloride ions is one of the crucial reasons for the durability degradation of concrete. Harmful media, such as chloride ions, water, and carbon dioxide in the external environment, penetrate the material through the pores of the concrete. Chloride ion reaches the surface of the steel bar. It accumulates to a particular critical concentration, which causes corrosion and expansion of steel bar, resulting in concrete cracking, reducing the durability of reinforced concrete, and shortening its service life cycle. As a new composite material, it is of great significance to study the migration of chloride ions in copper slag concrete for its application and promotion. By mixing copper slag into concrete as fine aggregate, the chloride ion penetration test proves that the chloride ion penetration of copper slag concrete is permeable to natural aggregate concrete [17]. The addition of copper slag enhances the volcanic ash activity of concrete, thus forming a dense microstructure on the concrete surface, increasing strength and improving corrosion resistance.

The preparation of copper slag to more fully replace natural aggregate concrete requires further performance studies, and existing design specifications for raw aggregate concrete do not apply to copper slag aggregate concrete due to the quality differences of copper slag. Previous research on the durability of copper slag aggregate concrete is insufficient; therefore, the authors of this study used a statistical analysis method and developed an orthogonal test design based on the compressive strength index in order to figure out the optimal copper slag mix ratio. The authors of this work studied durability with a rapid chloride ion migration test and the corrosion potential analysis of copper slag aggregate concrete. Finally, combined with the morphological analysis of SEM images and fractal dimension calculations, the authors of this study were able to systematically explain the mechanism of improving the mechanical properties and durability of copper slag from the microscopic mechanism perspective.

2. Experimental Section

2.1. Materials

This study used P.O42.5 ordinary Portland cement produced by Shanshui Cement Co., Ltd. (Jinan, China).

Table 1. Physical properties of cement.

Index	Result
Unit weight (kg/m3)	1270
Specific surface area (g/cm3)	3850
Solid density (g/cm3)	3.14
Poriness %	83.71
Moisture content %	0.41
Fineness %	7.2
Initial setting time min	≥45
Final setting time h	≤12

shows the performance indicators of cement.

The fine aggregate was natural sand with gradation. According to GB/T14684-2011(EN1097) [18], the fineness modulus (sieving method), apparent density (volumetric flask method), bulk density (density cylinder method), mud content (washing method), and the porosity of natural sand were tested.

Table 2. Detection indexes of natural sand.

Index	Fineness Modulus	Performance Density (g/cm3)	Stacking Density (g/cm3)	Water Absorption (%)	Poriness (%)
Technical standard	>2.5	>1.35	>3.0	<1.0	<47
Testing result	2.83	2.68	3.36	0.79	42.33

shows the test results, and all indexes of natural sand met the requirements.

Coarse aggregate is natural limestone gravel. Using a linear vibrating screen for particle size screening, the crude aggregate mix ratio of 5~10 mm gravel 30%: 10~20 mm stone 70% was determined. According to GB/T14685-2001(BS EN 933-5-1998) [19], the apparent density, bulk density, mud content, porosity, and water absorption of graded gravel were detected, and the results are shown in

Table 3. Detection indexes of limestone gravel.

Index	Performance Density (g/cm3)	Stacking Density (g/cm3)	Water Absorption (%)	Silt Content (%)	Poriness (%)
Technical standard	>2.5	>1.35	<1.0	<0.5	<47
Testing result	2.68	1.60	0.62	0.34	45.2

. All the indexes of natural coarse aggregate met the technical requirements.

Copper slag, solid waste from Yantai mine, was black with the particle size of 0.5~2 mm. This study used copper slag as fine aggregate instead of natural sand. The sampling and test method of copper slag used in the test was under the current industry standard JGJ52-2006(EN1097) [20], and

Table 4. Chemical composition of copper slag (wt/%).

Element	SiO2	MgO	SO3	CaO	CuO	Fe2O3	Al2O3	Total
Percentage	47.92	11.14	2.25	13.58	2.44	12.74	9.21	99.28

shows the chemical composition. The test determined the physical properties of copper slag, water absorption, hardness, bulk density, specific gravity, fine set modulus, and free water content.

Table 5. Physical properties of copper slag.

Properties	Value	Standard
Water absorption (%)	0.46	ASTM C127
Hardness (mohs)	6–7	/
Bulk density (kg/m3)	1850	ASTM C29
Specific gravity	3.57	ASTM C127
Fineness modulus	2.8	/
Free water content (%)	<0.5	/

shows the physical properties test results. Copper slag had the advantages of high hardness, high wear resistance, and high strength, which could improve the comprehensive properties of concrete. The surface characteristics of copper slag were rough, angular, and porous. In terms of physical properties, the proportion of copper slag was higher than that of natural sand, and the density of concrete prepared as a substitute for sand increased. Copper slag had a low-water absorption rate compared with natural sand, and the amount of water needed for mixing concrete was less than that of the natural sand concrete mixture. With increased copper slag content, the free water content in a concrete matrix will increase [21].

Henan Construction Road and Bridge Building Materials Co., Ltd. (Zhengzhou, China). produced the polycarboxylate superplasticizer with the solid concentration of 25%. The mixed water was tap water.

2.2. Mix Design

The mixing ratio test based on compressive strength adopted an orthogonal test design, and the test ratio had the characteristics of uniform dispersion and consistent comparability. We selected the orthogonal test table with three factors and four levels. The essential difference between copper slag and natural sand was water absorption. The three factors were:

• Substitution rate of copper slag;
• Water cement ratio;
• Water reducing agent.

Each element was divided into four levels, as shown in

Table 6. Orthogonal factor level table.

Level	A-Substitution Rate of Copper Slag (%)	B-Water Cement Ratio	C-Dosage of Water Reducing Agent (%)
1	0	0.3	0
2	15	0.35	0.5
3	30	0.4	1
4	40	0.45	1.5

.

According to the orthogonal test table, 16 groups of tests are shown in

Table 7. Experimental design table.

Specimen Number	Cement (kg/m3)	Natural Sand (kg/m3)	Copper Slag	Coarse Aggregate (kg/m3)	Water Cement Ratio	Water Reducer (kg/m3)
1	380	872	0	1027	0.3	0
2	380	872	0	1027	0.35	1.9
3	380	872	0	1027	0.4	3.8
4	380	872	0	1027	0.45	5.7
5	380	741.2	130.8	1027	0.3	1.9
6	380	741.2	130.8	1027	0.35	0
7	380	741.2	130.8	1027	0.4	5.7
8	380	741.2	130.8	1027	0.45	3.8
9	380	610.4	261.6	1027	0.3	3.8
10	380	610.4	261.6	1027	0.35	5.7
11	380	610.4	261.6	1027	0.4	0
12	380	610.4	261.6	1027	0.45	1.9
13	380	523.2	348.8	1027	0.3	5.7
14	380	523.2	348.8	1027	0.35	3.8
15	380	523.2	348.8	1027	0.4	1.9
16	380	523.2	348.8	1027	0.45	0

.

2.3. Specimen Preparation

Concrete compressive strength test specimens were cubes of 150 mm × 150 mm × 150 mm in size. Concrete was entirely mixed and poured into the test mold according to the proportions in Table 7 and then vibrated and smoothed. The specimens were placed in an indoor environment 24 h after pouring and demolding. The models were placed in a standard box with a relative humidity of (95 ± 5)% and a close temperature of (20 ± 2) °C for 28 d.

Before the chloride ion erosion test, we removed the original corrosion products of steel bars. We configured 1000 mL 38% hydrochloric acid solution, added 20 g Sb₂O₃, 50 g SnO₂, and mixed. The temperature was maintained at 20~25 °C, and the steel bar was immersed in the solution for 30 min. At the end of immersion, the surface impurities were removed with a soft brush, and ultrasonic cleaning was used to observe the removal of surface impurities. We repeated this process until eradicating the surface corrosion, cleaning the steel bar thoroughly, and drying in the oven for 24 h. The specimens used for the rapid chloride ion migration test were prisms of 50 mm × 100 mm × 400 mm. In the middle of the prisms, two-building steel bars (HRB335) with a diameter of 10 mm were embedded and labeled as steel bars No. 1 and No. 2, respectively. The thickness of the protective layer was 20 mm. The exposed area of the middle reinforcement was about 31.4 cm², and epoxy resin was used to seal both ends. Figure 1 shows the test process.

SEM test determined the sample of a specified age, terminated its hydration with anhydrous ethanol, dried it to constant weight, and produced the sample with a size of about 1 cm and a smooth surface.

2.4. Test Process

The compressive strength test was performed according to the relevant provisions of GB/T50081-2002(EN 12390-3:2019) [22]. The test used a WAW-600C microcomputer control electro-hydraulic servo universal testing machine and determined the average value of three fast specimens of each mixing ratio as the compressive strength of the mix ratio. If the maximum or minimum value exceeded 15% of the average value, it was considered an invalid specimen, and a new specimen was selected for testing. Rapid chloride migration (RCM) was an unsteady method. The test was according to GB/T50082-2009(ASTM C876-09) [23] standard. The equipment was a multifunctional chloride ion durability tester for concrete. Absolute voltage U = 60 V.

2.5. Test Methods

The experiment used Galvanostat/Potentiostat 283 tester for the electrochemical test. The AC voltage was ten mV, and the frequency was 0.01–105 Hz at different penetration times. A wet sponge ensured effective contact between the specimen and the electrode plate. Additionally, the data of the electrochemical impedance spectrum was analyzed with Zsimpwin v3.30 software (Ann Arbor, MI, USA), created by Bruno Yeum, Ph.D. The test used the corrosion potential method and electrochemical impedance spectroscopy (EIS).

The SEM testing equipment adopted the Quanta250 scanning electron microscope produced by FEI Company in the United States of Oregon (Hillsboro, OR, USA), which can observe the morphology and structure of polymer materials. The magnification was 20 × 10⁵~30 × 10⁵ resolution.

3. Results and Discussion

3.1. Mix Analysis

Table 8. Compressive strength at 28 d.

Group	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
Compressive strength (MPa)	52	50	49	48	54	52	51	49	56	50	49	49	55	54	51	50

shows the compressive strength test results of 16 mix-ratio groups obtained according to the orthogonal test design described in Table 7. First, range analysis was used to mark the rules of data in the 16 test groups and to determine the optimal level of each factor. Then, the variance of the test data was analyzed using the single element and multivariate analysis module of IBM SPSS Statistics R26.0.0.0, created by Stanford University, development of IBM (Armonk, NY, USA). The analysis results were used to systematically evaluate the substitution effect of copper slag.

A range was the difference between maximum and minimum values of sum results in the test at each level of a single factor. The range values in

Table 9. Range analysis results of compressive strength.

Serial Number	A	B	C
Ⅰ1	200.1	219	204.7
Ⅱ2	208.7	208.2	206.7
Ⅲ3	206.4	202.3	210.6
Ⅳ4	212.5	198.2	205.7
K	4	4	4
Ⅰ1/K	50.01	54.76	51.19
Ⅱ2/K	52.17	52.05	51.66
Ⅲ3/K	51.6	50.57	52.65
Ⅳ4/K	53.14	49.54	51.42
Range	200.1	219	204.7

were obtained from the compression test results in Table 8, showing how this factor affected the test results. The greater the range, the greater the influence of this factor on the test results. Ultimately, the effect was found to be negligible.

The range analysis of factor A (copper slag replacement rate) in Table 9 shows that the compressive strength of CSAC gradually increased with the increase in the copper slag replacement rate, indicating that the addition of copper slag is beneficial to improving the compressive strength of concrete. These results parallelled the research conducted by [1,4,24]. The edge of a copper slag aggregate was sharper than that of ordinary concrete, and as the cohesion between an aggregate and a concrete matrix increased the contact area between the aggregate rose and the strength of the concrete improved [25].

The range analysis results of factor B in Table 9 show that the compressive strength of CSAC decreased with the water–cement ratio increase. These results were similar to the research findings of [26], and they occurred because the water absorption rate of copper slag aggregate is significantly lower than that of ordinary concrete aggregate. The same water–cement ratio contained more free water than copper slag concrete, which was detri-mental to the strength of concrete. To solve this problem, a smaller water–cement per-centage can be used [27].

The range analysis results of factor C in Table 9 show that CSAC had the highest compressive strength when the content of the superplasticizer was 1%. A superplasticizer content exceeding 1% was found to reduce the compressive strength. These results were nearly identical to the findings of [28,29], and the main reason for this situation was the addition of a high-efficiency, water-reducing agent to concrete. Knowledge of the particle size distribution of powder can be used to optimize powder size distribution. Fully dispersed powder particles can fill the gap in a slurry, which is conducive to improving the strength of concrete. An excessive superplasticizer content can cause significant fluidity and increase the initial damage to matrix concrete, which are not conducive to improving compressive strength [30].

The calculation of range analysis was simple, and the conclusion was intuitive and easy to understand. However, the causes of errors in range analysis was difficult to distinguish. The data fluctuations caused by the operator’s mistakes were difficult to eliminate. Therefore, this study further conducted variance analysis on the compressive strength test results, and

Table 10. The results of variance analysis for impermeability performance test.

Variance Source	Type III Sum of Squares	Degree of Freedom	Mean Square	F-Value	Sig.
A	1007.202	3	335.734	5.002	0.045
B	3018.419	3	1006.140	14.990	0.003
C	241.336	3	80.445	1.198	0.387

shows the results.

The variance analysis results in Table 10 show that the Sig value of factor A, copper slag replacement, was rate = 0.045 < 0.05, and the F test was significant at the 95% confidence level. The results show that the substitution rate of copper slag had a significant effect on the compressive strength of CSAC. Sig value of factor B, water–cement, was ratio = 0.003, far less than 0.05; the F test was very significant at 95% confidence level. The water–cement ratio had an extremely significant effect on the compressive strength of CSAC. Sig value of factor C, plasticizer, was content = 0.387 > 0.05; the F test was significant at 95% confidence level. The water-reducer dosage has no significant effect on CSAC compressive strength.

The residual scatter diagram of ANOVA is shown in Figure 2.

Figure 2 shows the scatter diagram of the measured value, predicted value, and standardized residual of the dependent variable of 28 d compressive strength. An apparent correlation between the expected and measured values exists, and the scatter distribution is close to a linear trend. The standardized residual was randomly distributed near 0 and evenly distributed, indicating that the relevant result was good and the conclusion of the variance analysis was highly reliable. Based on the above findings,

Table 11. Concrete batching table kg/m³.

Materials	Cement	Fine Aggregate	Copper Slag	Coarse Aggregate	Water	Water Reducing Agent
NAC	380	872	0	1027	114	3.8
CSAC	380	523.2	348.8	1027	114	3.8

shows the material consumption per cubic meter of copper slag concrete.

3.2. Corrosion Potential Analysis

Figure 3 shows the time-varying curves of the corrosion potential and polarization resistance of rebar during the 34 d electric migration process. NAC-1 and NAC-2 represent two reinforcing bars embedded in concrete. The corrosion trends of the two reinforcing bars in the same specimen were similar, with minor fluctuations.

According to Figure 3, the E_corr changes in steel bars in the NAC concrete during electromigration could be divided into four stages:

• Ascent stage: From 0 to 8 d, the E_corr of steel bars in the NAC concrete slightly rose in the initial phase of electromigration because the initial density of NAC was small; therefore, there were a lot of pores and microcracks. Corrosion products filled in the initial defects, and the compactness of the NAC concrete was accordingly improved.
• Descent stage: From 8 d to 20 d, E_corr began to rapidly decline. The E_corr value of NAC-1 started to fall below −276 mV (relative to saturated calomel electrode) at 16 d. The E_corr value of NAC-2 started to fall below −276 mV (relative to saturated calomel electrode) at 18 d. According to the recommended standard of ASTMC876, the corrosion probability of the steel bar was 90% at this time.
• Stabilization stage: From 20 d to 26 d of corrosion, E_corr slightly fluctuated and remained stable. At this stage, the reinforcement began to rust, and the categories and quantities of corrosion products increased, which positively affected the compactness.
• Rapid descent stage: After 26 d, E_corr again rapidly descended; at 34 d, E_corr had descended to close to −700 mV. At this stage, with the increase in corrosion time, the reinforcement corrosion in the NAC concrete was severe.

Figure 4 shows the time-varying curves of corrosion potential and polarization resistance of steel bars in copper slag aggregate concrete during the 34 d electric migration process. The corrosion trends of the two reinforcing bars in the same specimen were the same.

According to Figure 4, changes in the E_corr of steel bars in the CSAC concrete during electromigration could be divided into three stages:

• Ascent stage: During the initial stage (0–8 d) of the CSAC concrete, E_corr slightly increased, though the trend was insignificant. The initial compactness of CSAC was greater than that of NAC, and the filling effect of corrosion products was not apparent.
• Descent stage: From 8 d~26 d, the E_corr value steadily declined. At about 22 d, the E_corr of CSAC began to be lower than −276 mV (relative to saturated calomel electrode), at which time the corrosion probability of steel bar was 90%. At this stage, the reinforcement in the CSAC concrete gradually corroded, but the corrosion process was gentler than that in NAC because of the high initial compactness of concrete.
• Rapid descent stage: From 26 d to 34 d, E_corr rapidly declined, reaching about −450 on day 34. The concrete at this stage began to significantly rust.

Figure 5 compares the corrosion potential of NAC and CSAC, showing that the corrosion of the reinforcement in CSAC was smaller than that in NAC over 0–7 d of CSAC corrosion time. The corruption of support in CSAC was more extensive than that in NAC during 7–18 d of corrosion time. After 18 days, the corrosion of reinforcement in NAC gradually became more significant than that in CSAC. The initial density of CSAC was greater than that of NAC, and the erosion of the natural aggregate concrete in phase I was more serious. In the second stage of corrosion, many corrosion products filled the concrete pores, and the compactness of concrete was improved. The porosity of the natural aggregate concrete was large, and the filling effect was more pronounced; therefore, the corrosion potential of the second stage increased. In the third stage of corrosion, the corrosion degree of the concrete sharply increased, the corrosion rate of natural aggregate concrete was high, the decline slope of the corrosion degree curve was large, and the decline speed was fast. The corrosion degree of the copper slag aggregate concrete was gentle, and the decline rate was slow. The degree of corrosion and the rate of corrosion increase for CSAC was lower than those of NAC, which confirmed that the compactness of CSAC was higher than that of the NAC concrete. Copper slag was also shown to help increase the compactness of concrete. These observations follow the findings of [9,17].

3.3. Electrochemical Impedance Spectroscopy Analysis

The Nyquist diagram of the NAC and CSAC specimens (Figure 6) measured with the EIS method shows that impedance varied with corrosion time. The abscissa in the diagram represents the genuine part of impedance Z, the vertical coordinate represents the imaginary part of impedance Z, and each point in the figure represents a different frequency. The left side is the high-frequency region, and the right side is the low-frequency region, which are shown as the high-frequency capacitive arc and the low-frequency capacitive arc, respectively, on the Nyquist curve. The high-frequency volumetric reactance arc represents the characteristics at the interface between the concrete and reinforcement. The intersection position for high- and low-frequency arcs reflects the resistance of ions to the surface of the mount through concrete, which can be used to characterize concrete compactness.

Figure 6 compares the Nyquist curves of the No. 1 steel bar before and after NAC and CSAC electrical transfer for 34 days.

The steel bar (NAC-0) before electromigration showed a complete high-frequency capacitive arc while the low-frequency capacitive angle showed an upward curve. After 30 days of electromigration, the Nyquist curve shape of the steel bar (NAC-34) remained intact. The diameter of the high-frequency capacitive arc decreased, which means the resistivity of the concrete layer outside the steel bar decreased, and the substantial compactness decreased. The resistivity of a concrete coating was inversely proportional to the corrosion rate of reinforcement. The shrinkage of high-frequency capacitive arc indirectly indicated the increase in the corrosion rate of support. The position of the intersection of arc in the high-frequency area and low-frequency area moved in a negative direction from day 0 to day 34, indicating the resistance of electric ions to the surface of reinforcement through concrete, indicating that concrete corroded and the compactness decreased.

The steel bar (CSAC-0) before electromigration showed a complete high-frequency capacitive reactance arc, and the low-frequency capacitive reactance arc showed an upward curve. After 30 days of electromigration, the Nyquist curve of the steel bar (CSAC-34) remained intact. The diameter of the high-frequency capacitive arc obviously decreased, indicating that the resistivity of the concrete layer outside the steel bar decreased. Corrosion led to a decline in concrete compactness. The resistivity of the concrete layer was inversely proportional to the corrosion rate of the reinforcement. The shrinkage of the high-frequency capacitive arc indirectly indicated an increase in the corrosion rate of support. The position of the intersection of arcs in the high-frequency and low-frequency areas moved in the negative direction from day 0 to day 34, indicating the resistance of electric ions to the surface of reinforcement through concrete, the erosion of concrete, and the decrease in compactness.

The Figure 6 shows that, before electromigration, the arc radius of the high-frequency reactance of the No. 1 steel bar in CSAC is more significant than that in NAC, indicating that the initial compactness of CSAC is greater than that of NAC. After 34 d of electromigration, the high-frequency capacitance arc radius of the No. 1 steel bar in CSAC is more significant than that of the No. 1 steel bar in NAC, proving that CSAC has better corrosion resistance than NAC. The addition of copper slag can improve concrete’s chloride ion erosion resistance. This result is consistent with the studies of [2,31]. The concrete protective layer made of copper slag, instead of natural sand aggregate, has a better protective effect on the embedded reinforcement in the structure.

4. Microscopic Mechanism Analysis

Composite cement-based materials are porous, multiphase, and quasi-brittle, and their mesoscopic characteristics are closely related to their macroscopic chloride penetration resistance. This study of the mesoscopic characteristics of copper slag aggregate concrete helps to reveal the influence mechanism of copper slag aggregate on the chloride penetration resistance of concrete, a critical durability index.

4.1. Microscopic Morphology Analysis

Figure 7 is the SEM comparison diagram of recycled aggregate concrete and reference concrete after curing for 28 days.

Figure 7a shows the microscopic morphological analysis results of the combination of a natural fine aggregate and a cement matrix. The compatibility between the natural fine aggregate and cement paste was poor and formed a relatively obvious boundary area on the contact surface after cement hydration. There were apparent coalescence cracks between the aggregate and the cement matrix, and the bonding was not dense. Figure 7b shows the microscopic morphological analysis results of the combination of the copper slag aggregate and a cement matrix. The combination of the slag aggregate and the cement matrix was more compact than the natural fine aggregate combination. The weak area of concrete strength at the interface between the aggregate and the cement matrix reduced the overall continuity of the material such that the concrete could quickly form a high-stress area at the initial stage of loading, and cracks could rapidly expand. The cracks ultimately ran through the entire specimen, damaging the concrete. The existence of the interface led to the formation of a rapid transmission path around the aggregate and accelerated the diffusion of chloride ions. The stable interface between the copper slag aggregate and the cement matrix delayed the damage to the concrete and improved the strength and toughness of the concrete matrix. The authors of [5,32] drew the same conclusions: the density of the transition zone between the cement paste and slag aggregate, compressive strength, and splitting tensile strength of concrete containing a copper aggregate are higher than those of concrete containing a limestone aggregate.

The SEM results show that the fine aggregate of copper slag influenced the surface morphology and structure of the concrete. Figure 7c shows that there was more ettringite in the hydration products of the internal system of the ordinary concrete, presenting a slender needle bar shape and less contribution to the strength and compactness of the concrete. The concrete was internally loose, with more pores, which are not conducive to compressive strength and resistance to chloride ion erosion. Figure 7d shows that after adding 40% copper slag, the porosity of the concrete significantly decreased. The apparent morphology evolved from needlelike and complex morphology to a dense and orderly morphology. The copper slag aggregate was able to effectively improve the low compactness of the concrete.

4.2. Analysis of Toughening Mechanism

There are three cracking modes for quasi-brittle materials: opening, sliding, and tearing. In low cycle and under continuous load less than the peak value, cracks of quasi-brittle materials extend mainly by introduction. The following formula can obtain spreading force:

$$K_{IC}=\sqrt{\frac{2E\gamma }{1-{\mu }^2}}$$

(1)

where K_IC is fracture toughness, reflecting the resistance ability of cracked quasi-brittle material to external action. The power to prevent crack propagation is an inherent property of the material; E is the elastic modulus; γ is the fracture surface energy; and the two coefficients are nonstructural sensitive. µ is a coefficient that has a great relationship with the material’s microstructure; therefore, K_IC is structurally sensitive.

According to Equation (1), the change in the internal cracking path of ordinary aggregate concrete and copper slag aggregate concrete can be seen by applying the peak load of 80% numerical stress on concrete specimens and continuing for 36 h. The crack path and morphology are shown in Figure 8.

Figure 8a shows the microscopic morphology of the crack cracking path of natural aggregate concrete 36 h after 80% peak load is applied. The figure shows that the approach is short and straight with few branches, which easily causes direct cracking of the material and shows prominent quasi-brittle material characteristics. Figure 8b shows the microscopic morphology of CSAC. The internal microstructure of concrete made of copper slag aggregate has significantly changed, and the mode of crack propagation under stress also changed. The figure shows many tiny microcracks were generated inside the sample, which increased the total length of the crack propagation path and the number of branches. The microcracks and the second phase particles formed the diffusion toughening. The crack propagation force K_IC is reduced. Thus, the durability of the sample is improved. The results provide a theoretical basis for the potential corrosion analysis.

4.3. FRACTAL Characteristics of SEM

SEM images show that, when the surface of concrete fracture is scaly, the surface is very uneven with pores developing, and the broken blocks are mostly flaky and scaly. It is difficult to describe the characteristics of concrete fracture surfaces quantitatively by directly observing the image. Fractal dimension is an index to measure image texture roughness, which can explain the roughness of the image surface and quantitatively describe the differences in human visual perception of images. The commonly used methods to calculate fractal dimension in various disciplines include gray interpolation, fractal Brownian motion self-similar simulation method, differential box dimension method, and carpet covering method [33,34]. Box dimension method is easy to determine the fractal dimension and is convenient for programmatic design. This study uses the box dimension method for quantitative image analysis [35,36].

The image must be binarized first to calculate fractal dimension by the box dimension method [37]. The target image only consists of black and white; then, using the box with the side length of K, the target image and the number of regions with black (white) parts recorded, which is usually N(k), were divided. The point of view of the image is side-length, and to divide the image, we use 1, 2, 4…, the size of 2i pixel points, to finally obtain the number of boxes N(1), N(2), N(3), N(4). When k approaches 0, Equation (2) obtains the fractal dimension of the box dimension method.

$$D=\underset{k\to 0}{\lim }\frac{\lg N\left(k\right)}{\lg k}$$

(2)

According to the above equation, the smaller k is the more accurate the fractal dimension, but the minimum value of value is a single pixel point. Generally, fit N(k) corresponding to different k, by using the least square method, can obtain the following Equation (3):

$$\lg N\left(k\right)=D\cdot \lg k+h$$

(3)

The slope of the line drawn by the equation is the fractal dimension, usually negative, so the fractal dimension is −D.

Figure 9 shows select 28 d SEM images of NAC and CSAC with standard curing for two magnification levels of 10,000 times and 20,000 times, respectively. Grayscale images of section morphology and staining images of section depth were obtained by scanning electron microscopy.

MATLAB R2021b created by Cleve Moler, development of MathWorks (Natick, The United States of America) for binarization is the threshold value of image binarization. The threshold value directly affects the image quality, leading to different proportions of black-and-white parts in the image. Therefore, an essential parameter setting [38] directly affects the results of fractal dimension calculation. Otsu method, using the clustering algorithm, divides the number of gray level images into two parts according to the gray level. For the most significant gray value differences between the two parts, each part of the grayscale difference, from minimum value through the variance calculation, finds a suitable gray level by using simple math, image fault points, and the advantages of low threshold [39]. This study uses the Otsu method to determine the binarization threshold. Figure 10 shows the image after binarization of the SEM gray image, and Table 11 shows the fractal dimension calculation and fitting accuracy.

Figure 11 shows the fractal dimension accuracy of SEM images. R² is high and can better reflect the nonlinear degree of images.

The fractal dimension can be used as a parameter to characterize the roughness of concrete cross-sections. Accordingly, the correlation coefficients shown in

Table 12. Results of fractal dimension.

Species	Binarization Threshold	Fractal Dimension	R2
NAC-10,000	0.459	1.830	0.996
CSAC-10,000	0.388	1.663	0.999
NAC-20,000	0.412	1.811	0.999
CSAC-20,000	0.433	1.746	0.998

are all above 0.99, indicating that the cross-section of the concrete SEM images had fractal characteristics. These results are nearly identical to the findings of [40,41]. Fractal dimension calculations have shown that the roughness of specimens increases with increases in the fractal dimension at the same amplification ratio [42]. In this study, the fractal dimension of the SEM section of the copper slag aggregate concrete was smaller than that of the natural aggregate concrete, indicating that the addition of the copper slag aggregate promoted decreases in the initial damage to the internal section of the concrete. These results lead to the conclusion that the strength and durability of copper slag aggregate concrete are superior to those of natural aggregate concrete.

5. Conclusions

• The highest compressive strength of concrete was reached when a fine natural aggregate was replaced with 40% copper slag. The influence of the water–cement ratio on the compressive strength of concrete was found to be significant. When the water–cement ratio exceeded the threshold, the compressive strength decreased. The effect of the water-reducing agent on the compressive strength of CSAC was not found to be significant.
• The rebar corrosion of CSAC over 0–7 d was less than that in NAC, and the rebar deterioration in CSAC was more significant than that in NAC over 7 d~18 d. After 18 days, the rebar corrosion in NAC was greater than that in CSAC. The addition of copper slag could be used to improve the concrete’s chloride ion erosion resistance.
• A comparison of NAC and CSAC Nyquist curves after electromigration for 34 days showed that the initial compactness of CSAC was greater than that of NAC. CSAC had a stronger corrosion resistance than NAC; therefore, copper slag aggregate concrete could be beneficial in improving the corrosion resistance of embedded steel bars.
• The box dimension method was used to calculate the fractal dimension of damage caused to the concrete section. The fractal dimension of the SEM section of the copper slag aggregate concrete was shown to be smaller than that of the natural aggregate concrete, indicating that the addition of a copper slag aggregate could promote decreases in the initial damage to the internal sections of concrete.