Influence of Crack on Concrete Damage in Salt-Freezing Environment

+e damage development trend of concrete with cracks in salt-freezing environment is systematically studied. +e cracks are also tested in intact concrete for comparison, and crack characterization is introduced. +e mass loss, the relative dynamic elastic modulus, and the change of crack width are analyzed. Results show that the crack width increases as the salt-freezing cycle progresses. Following the development trend of the cracks, concrete cracks can be divided into three categories: 0–40, 40–100, and 100–150 μm. +e mass loss increases significantly, and the change of relative dynamic elastic modulus decreases in concrete with an initial crack compared with the intact concrete. When the crack width is 80 μm, a maximum mass loss rate of 0.19% and a minimum relative dynamic elastic modulus of 75.81% can be obtained. +ese test results prove that crack and freeze-thaw coupling can influence each other and accelerate the failure of concrete. Overall, this study can serve as a basis for the durability design and life improvement of concrete structures.


Introduction
With the construction of cross-sea bridges, concrete has been increasingly used in marine environments. us, the improvement of the durability of concrete material is a significant scientific issue for civil structure design [1]. In some cold areas, such as in northern China, the temperature is usually less than 0°, leading to the rapid degradation of the concrete. Frost attack on concrete is distress resulting from the phase transformation associated with the freezing of internal moisture. is phenomenon directly leads to the reduction of the cover and rusting of steel, affecting the service life of the concrete. In environments characterized by the presence of chloride ions, such as coastal environments and areas where deicing salts are used, the combination of freeze-thaw cycles and deicer salt application can seriously damage concrete [2,3]. e damage mechanism of saltfreezing is different from freezing and thawing damage. Deicing salt, employed to reduce the freezing point of water, alters the degree of saturation [4,5] and reacts with the hydrated ordinary Portland cement (OPC), creating expansive reaction products that lead to cracking and distress [6][7][8][9][10][11]. e deterioration due to the erosive effect of deicing salt can develop rather rapidly than the destructive effect of normal freezing and thawing cycle on concrete. Consequently, the performance degradation of concrete is more serious under salt-freezing circumstances. e extensive use of deicing salts can result in the deterioration of concrete structures, such as pop-outs, exfoliations, and cracks, significantly affecting the service life of structures [12][13][14][15].
For concrete, cracks can form in various ways, including applied external force, temperature changes leading to volumetric mismatch, and internal chemical reactions usually caused by the ingress of aggressive chemicals [16][17][18][19]. Once cracks take form, they can provide a way for the chloride transport to the concrete [20][21][22]. Ingress of steel corrosive agents, such as water and air, will also accelerate, leading to higher chances of initiating corrosion [23,24]. us, many scholars investigated the effect of cracks on chloride ion transport. Jang et al. [25] studied the influence of cracks on chloride ion diffusion behavior in concrete. ey found that the diffusion coefficient of chloride ion would not increase with the increase of the crack width until the so-called critical crack width is reached. When the crack width is larger than the critical value, the diffusion coefficient increases with the increase of the width, and the critical crack width obtained by the test varies from 55 μm to 80 μm. Wang et al. [26] showed that when the crack width is between 100 and 400 μm, the diffusion coefficient of chloride ions in the crack increases rapidly with the increase of the crack width. Otherwise, the crack width has no obvious effects on chloride diffusion. Djerbi et al. [27] compared the influence of cracks on ordinary concrete and high-performance concrete by simulating chloride ion penetration in concrete. ey proved that the diffusion coefficient of chloride ions in the crack increases with the increase of the crack width. When the crack width is greater than 80 μm, the diffusion coefficient of chloride ion in the crack is equal to the diffusion coefficient of chloride ion in water, which is independent of the concrete mix ratio. e above studies were only a few of the increasing investigations on the salt frost damage on concrete. However, concrete with cracks has not yet been considered, and the factors leading to its rapid degradation in a marine environment remain unclear. erefore, the frost damage resistance of concrete under cracking conditions requires further examination, which is helpful for the material's durability design and life prediction. e present study aims to determine how the development process of salt frost will damage concrete with crack and how the cracks influence salt frost damage. Specifically, the salt-freeze cycle is applied to the concrete specimens with initial cracks. After analyzing the mass loss of concrete, relative dynamic elastic modulus, and the change of crack width, the development trend of concrete damage under the coupled action of crack and the salt-freezing cycle is obtained.

Preparation of Test Specimens and Applied
Test Methods 2.1. Materials. Type II52.5 OPC provided by Jiangnan-Xiaoyetian Cement Co., Ltd., was used in the experiment. Fly ash (FA) and ground slag (SL) were also used to replace part of the cement as supplementary cementitious materials. FA and SL had replacement rates of 15% and 35%, respectively. e chemical compositions were determined using X-ray fluorescence (XRF; ermo Fisher Scientific ARL QUANTX), as shown in Table 1.
Fine aggregate was natural sand with a fineness modulus of 2.6 and an apparent density of 2640 kg/m 3 . e basalt with particle size ranging from 5 mm to 15 mm and an apparent density of 2800 kg/m 3 was used as coarse aggregate. A 20% solid superplasticizer was applied to adjust the workability of fresh concrete. e air-entraining agent was also used to regulate 5%-6% of the air content. e above materials were provided by Jiangsu Sobute New Materials Co. Ltd.
Poremaster GT-60 produced by Quantachrome Instruments was used to test concrete porosity. Quanta 250 scanning electron microscopy (SEM) manufactured by FEI Company was employed to conduct microscopic analysis of the samples obtained after the salt-freezing cycle.

Sample Preparation.
e w/b used for concrete was 0.35. e concrete mix proportion is shown in Table 2.
e concrete was mixed as follows. e cementitious materials and aggregate were added and mixed for 1 min. en, the superplasticizer and air-entraining agent were placed and stirred continuously for 2 mins. After stirring, the concrete samples with dimensions of 100 × 100 × 100 mm and φ of 100 × 200 were cast. e cube samples were used to test the compressive strength, and the cylindrical samples were examined. e samples were removed from their mold on the next day. ey were cured for another 28 days in 20 ± 2°C, under 95% RH following the GB/T50081-2019. Subsequently, the compressive strength of the sample was determined using the pressure testing machine. e compressive strength of concrete was 56.9 MPa. e cylindrical samples were finally tested.

Preparation and Characterization of Cracks.
Various methods can be used to prepare concrete cracks, such as the Brazilian splitting test, wedge splitting test, and three-point bending mode. Here, the method proposed by Toshiki on the basis of the Brazilian splitting test was employed. e Brazilian splitting test is simple and convenient, and thus, it is accepted by the majority [25,28,29]. However, due to the heterogeneity of concrete, controlling the formation and development of cracks in the loading process would be difficult, and the specimen might easily develop a fracture. In the present experiment, the method would use the PVC pipes to prevent specimen fracture.
After curing for 28 days, the samples with dry surfaces were put in PVC pipes whose one end was closed, as presented in Figures 1(a) and 1(b). en, a low viscosity epoxy resin was used to fill the gap between PVC pipes and samples. A sample slice with a height of 50 mm was obtained for crack preparation. e details of the production process are presented in Figure 1. On the basis of the Brazilian splitting test, a linear variable differential transformer (LVDT) was used to monitor the crack width during the whole process, as shown in Figure 1(e). When the width reached the design value, the sample was removed, and the fracture parameters were characterized after 15 min [30] as the crack width would shrink after unloading.
After completing the crack preparation, the exact crack width was tested as follows: first, a glass slide with a size of 24.5 × 76 mm was chosen as a reference. en, a photograph of the crack was taken with the reference, as shown in Figure 2(a). Next, the reference and the cracking area were selected and binarized. Finally, the number of pixels of the selected area was determined. In the same photograph, the ratio of the cracking area was equal to the ratio of pixels. e pixel of the crack and reference and the area of the reference were obtained. Accordingly, the area of the reference was calculated. For simplicity, the length of the crack was assumed to be the diameter of the specimen, which was 100 mm. us, the width of the crack was equal to the area divided by the length.
On the basis of the above method, the final crack width was divided into seven groups: 20, 40, 60, 80, 100, 130, and 150 μm. In the subsequent test, the width of each group was considered to be the actual crack width. e width error of each group was 5 μm. e detailed grouping is shown in Table 3.  Figure 1: Preparation process of cracks in concrete specimens. ese tests were run following the GB/T50082-2009 Chinese standard. In this case, the sodium chloride (NaCl) solution had a similar mass concentration of 3% as the freeze-thaw solution. e test surfaces of the samples were placed in 3% NaCl solution for seven days for solution uptake by capillary suction. e temperature in the box varied between +20°C and −20°C. A cycle of freezing and thawing lasted for 12 h, and the test only started when the temperature in the box was 20°C. In the first cycle, the samples were cooled at 10°C per hour for 4 h until the temperature was −20°C.
is setup was maintained for 3 h. en, they were heated up at a rate of 10°C per hour. Finally, the temperature was kept at 20°C for 1 h to complete the cycle. e temperature change during a cycle is shown in Figure 3. e specimens were characterized by the crack width and number of freeze-thaw cycles. For example, D28C8 indicates that the specimens had a crack width of 80 μm and were exposed to freeze-thaw cycles 28 times.
Following the GB/T50082-2009 standard, the related parameters were tested every four cycles (two days). e damage of specimens was reflected by the relative dynamic elastic modulus and scaling mass. After four cycles, the flaking material was collected, dried, and weighed. e measurement of ultrasonic transit time was a standardized test method used to determine the variation in the relative dynamic elastic modulus with time. Ultrasonic transducers were positioned to arrange the axis of the ultrasonic transit path perpendicular to the specimen surface and parallel to the crack depth direction. e transit time in the coupling medium was calculated from the velocity of the ultrasonic signal and the transit length in the coupling medium. In this experiment, ZBL-U510 nonmetallic ultrasonic detector was used to test the ultrasonic transmission time.
e relative dynamic elastic modulus was determined using the following equations: where E d is the relative dynamic elastic modulus of the specimen (%), τ n is the ultrasonic wave relative propagation  Advances in Materials Science and Engineering time of the specimen (%), t 0 is the total propagation time of ultrasonic wave in the specimen and coupling agent (μs), t n is the total propagation time of ultrasonic wave in the specimen and coupling agent after n cycles (μs), and t c is the propagation time of ultrasonic waves in the coupling agent. At the same time, a camera was used to take photos of the specimen. e photographs were processed to observe the changes in the crack width and sampling for porosity and microscopic testing.

Crack Width.
e effect of salt-freezing cycles on the crack width of concrete is presented in Figure 4. e initial crack width was roughly 60 μm, as shown in Figure 4(a). After the 8th, 16th, and 28th cycles, the crack widths indicated in Figure 4 were increased to 164.51, 193.58, and 268.01 μm. In other words, the crack width increased with the extension of the cycle period. e damage development of the crack width with the increase in salt-freezing cycle periods is shown in Figure 5.
e initial crack width of less than 40 μm gradually increased at the early and late stages, with the increase of the cycle period. In the initial stage, due to the small crack widths, considering, self-healing and other factors, the increase of crack width was relatively slow. However, with the continuous increase of the crack width, the self-healing ability was weakened. e bigger the crack width, the easier it for the solution to be absorbed by the concrete, leading to internally and externally salt-induced freezing damage and a significant increase in the crack width.
In the later period of salt-freezing cycle, concrete cracks tended to shrink. e identification and detection of cracks were mainly based on the software. With the development of salt-freezing cycle, on the one hand, flaking led to particles blocking cracks. On the other hand, the damage range of the specimen generally extended from the fracture area to both sides, making it difficult to identify fractures. Consequently, the crack width obtained by the test was reduced. For specimens with crack width less than 40 μm, the initial crack width was small, and the damage caused by salt-freezing was not serious, so this phenomenon did not occur at all.
Following the development trend of cracks, concrete cracks could be concluded as three categories. For the crack width of less than 40 μm, the salt-freezing cycle almost did not affect the crack development at the early stage. However, it caused a rapid growth for the crack width at the later stage. For the crack widths ranging from 40 μm to 100 μm, the saltfreezing cycle resulted in rapid growth at the first 20 cycles. e crack width remained stable or gradually decreased after 20 cycles. For the crack widths greater than 100 μm, the development trend of the cracks was similar, and the width was almost identical. To sum up, no obvious effect was observed on the crack development when the crack width was more than 100 μm. Moreover, for the cracks at different widths to reach the maximum growth rate of the crack widths, the number of salt-freezing cycles required is gradually reduced. For the crack widths of 40, 80, and 130 μm, the number of cycles required to achieve the maximum growth rate of the crack width was 28, 20, and 16 times, respectively.

Mass Loss Rate.
e mass loss rate of the specimen after 28 salt-freezing cycles is shown in Figure 6. At the stage where the crack width was less than 20 μm, no obvious effect was observed on its mass loss rate. When the crack width was ranging from 20 μm to 80 μm, the mass loss rate sharply increased. e maximum mass loss rate was 0.19%, which was obtained when the crack width was 80 μm. Afterward, the curve slowly decreased. e mass loss of surface undamaged concrete (crack width of 0 μm) was minimal, indicating that the surface cracks could significantly affect the salt-freezing resistance and reduce the durability of the specimen. Moreover, the increase of the crack width and contact parts between the specimen surface and salt solution increased the scaling mass. e mass loss of the specimens with different crack widths along with the salt-freezing cycle is shown in Figure 7. With the increase of the crack width, the mass loss gradually increased. At the early stage of the salt-freezing cycle, the mass scaling of concrete specimens was not obvious, and the mass scaling rate was low. With the increase of the salt-freezing cycle period, the quality loss of the specimen increased gradually. e mass loss also increased when the freeze-thaw cycle was more than eight times. With the continuous salt-freezing cycle, the peeling quality gradually increased, and the mass loss rate increased. e reason might be the gradual accumulation of the damage of the specimen. Given the existence of cracks, the salt solution could enter into the concrete through the cracks and cause damage. At the early stage of the salt-freezing cycle, no significant differences were observed in the scaling quality between specimens with different fracture widths, indicating that the effect of fracture width on the scaling quality was not obvious at this time. As the number of salt-freezing cycles went on, the effect of the fracture width on the scaling quality gradually increased.

Relative Dynamic Modulus of Elasticity.
e propagation of ultrasonic waves in concrete was affected by the matrix compactness and microcracks. e damage of concrete was determined by testing the change of ultrasonic wave transmission time in concrete. Ultrasonic testing of concrete was carried out in a direction perpendicular to the crack, as shown in Figure 8. e change of the relative dynamic elastic modulus of the specimen with the salt-freezing cycle under different fracture conditions is displayed in Figure 9.
When the crack width was less than 20 μm, the relative dynamic elastic modulus of concrete decreased slightly, and the damage occurred only in the later stages of the saltfreezing cycle. In Figures 9(c) and 9(d), the relative dynamic elastic modulus of concrete dropped sharply, whose crack width was 60 and 80 μm by 75.92% and 75.81%. However, when the crack width was 100 and 130 μm, the relative dynamic modulus was only slightly reduced. By contrast, when the crack width was 80 μm, the relative dynamic Advances in Materials Science and Engineering modulus did not continue to reduce as the crack width increases. Given the gradual damage accumulation, the relative dynamic modulus of all the specimens did not excessively change. e maximum change was only 92.96% when the width was 60 μm during the first eight salt-freezing cycles. In the early stage of the salt-freezing cycle, the microcracks were produced in the sample, and the relative dynamic elastic modulus did not change significantly. As the cycle progresses, the damage of the sample and the peeling were intensified. Macroperformances decreased in relative dynamic modulus of elasticity. e relative dynamic elastic modulus of the sample after completion of all salt-freezing cycles under different initial crack widths is shown in Figure 10. Without any produced cracks, the relative dynamic modulus of the specimen was 93% after 28 salt-freezing cycles. However, the relative       of the internal damage degree. When the crack width was less than 80 μm, the larger the crack width, the more obvious the decrease of the relative dynamic modulus of the specimen. e specimen would be more easily damaged in the salt-freezing cycle. However, at the stage where the crack width was greater than 80 μm, the relative dynamic elastic modulus exhibited the opposite trend, except when the crack width was 100 μm.
For the mass loss, obvious damage was produced in concrete specimens with an initial crack after eight saltfreezing cycles. When the crack width was 80 μm, the concrete was most likely to be damaged, and the destruction was the most serious in a salt-freezing environment.

Microstructure and Porosity.
e SEM images of the samples with the initial crack after 28 salt-freezing cycles are shown in Figure 11. Although cracks were observed in the concrete that did not undergo the salt-freezing cycle, their number was small, and the CSH gel was relatively complete and dense. After completing the salt-freezing cycle, the integrity of the specimen was damaged, and the peeling of small particles led to internal potholes and increased mass loss. When the crack width was 20 μm, the CSH gel was pitted, and the integrity was damaged. As the crack width increases, the damage to the specimen increases. For example, in the stage where the crack width was ranging from 60 μm to 80 μm, the microcracks inside the specimen significantly increased, and the crack expansion caused the specimen to peel. e damage of the specimen was the most serious when the crack width was 130 μm, and the CSH gel produced microholes. e damage caused by freezing and thawing resulted in many microcracks in the specimen. ese damages led to the scaling of the specimen. As a new transmission channel, microcracks accelerated the entry of chloride ions and water into the specimen. us, they could aggravate the damage degree of the next freeze-thaw cycle. e larger the crack width, the more serious was the microstructure damage and the more evident the damaging effect would be. Given the damage, more solutions entered the pores, further leading to the degradation of concrete.
is coupling effect was also one of the reasons why the damage of cracked concrete was worse than that of intact concrete.
e SEM images of the specimens with the same crack width at the 8th, 16th, and 28th salt-freezing cycles are shown in Figure 12. e damage of concrete increased gradually with the salt-freezing cycle. Similarly, the degree of surface scaling aggravation and the number of microcracks increased. e increase in the salt-freezing cycle periods gradually accumulated damage to the specimens aggravated. e changes in the void fraction during the salt-freezing cycle are shown in Figure 13. e samples with a crack width of 100 μm had the most probable pore radius increase and porosity diminution after completing all the salt-freezing cycles compared with only eight cycle periods. Salt-freezing caused the generation of tiny cracks. Many small damages, whether newly generated or originally in the specimen, increased gradually. Consequently, the most probable pore radius also increased. e voids grew and communicated with one another to form large damage, causing a decrease in porosity. For the specimen with a crack width of 60 μm, the most probable pore radius increased significantly after 28 cycles, but the change in the void fraction was not obvious compared with the specimens with a crack width of 100 μm. Yet, the porosity and most probable pore radius of D28C10 slightly changed in contrast to sample D8C10. is result proved that the influence of the crack width on concrete was greater than that of the freeze-thaw cycle, which was more likely to cause pore size increase and concrete damage.

Conclusion
In this study, the damage development trend of concrete with initial cracks in a salt-freezing environment was systematically examined. e development of cracks in concrete, the mass loss, the change of relative dynamic elastic modulus, and the change of microstructure and void ratio were tested. On the basis of the test results, the following conclusions can be drawn: (1) e crack width increases with salt-freezing. e larger the initial crack width, the larger the crack width will be. For the specimens with a crack width of less than 40 μm, the crack width develops slowly in the initial stage but rapidly at the later stage. e crack widths ranging from 40 to 100 μm have exhibited rapid growth. However, for the specimens with a crack width greater than 100 μm, under the same salt-freezing cycle, the crack width is almost the same, indicating that the influence of the initial crack width can be ignored.
(2) Under cracking conditions, the salt-free cycle exacerbates the damage of concrete. e mass loss and the relative dynamic elastic modulus have obvious changes compared with the intact concrete. In addition, the influence of the crack width on concrete damage is greater, and the acceleration effect on the concrete failure is more obvious than the salt cycles. (3) e SEM results show that the damage of concrete increases gradually with a continuous salt-freezing cycle. e peeling of small particles leads to internal potholes and mass loss. e larger the crack width is, the more serious the microstructure damage and the more evident the damaging effect. Salt-freezing also leads to the most probable pore radius increase and porosity diminution. (4) e concrete with an initial crack has developed internal and surface damage compared with the intact concrete. erefore, the surface and internal cracks and the pores of the specimen tend to fall off and increase, respectively. In other words, corrosive substances, such as chloride ions, can diffuse into the concrete through the surface and pass through the cracks. Consequently, substances can easily enter the concrete, causing durability problems, such as steel corrosion, and affecting the service life.

Data Availability
e data and charts used to support the findings of this study are available within in the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.