Study on the Sulfuration Mechanism of Concrete: Microstructure and Product Analysis

This paper presents an experimental investigation of the sulfuration mechanism of concrete. The microstructure, mineral phase composition, substance content, and pH of the concrete were determined using scanning electron microscopy, X-ray diffraction, comprehensive thermal analysis, and pore-solution pH test. It was observed that light-grey spots appeared on the surface of the specimen, and a large amount of powdery precipitated substances appeared. At the initial stage of sulfuration reaction, the formation of ettringite blocked the concrete pores and densified its cracks and voids. Subsequently, ettringite reacted with H+ to form gypsum, and the continuous increase in gypsum in the pores increased the number of cracks and broadened their width. Gypsum was the final product of the sulfuration reaction, and the mass percentage of gypsum in the powdery precipitated substances at different water–cement ratios was more than 50%. When the water–cement ratios was 0.37, 0.47, and 0.57, the highest Ca(OH)2 content was found for the lowest water–cement ratio. As the water–cement ratios increased, the amount of powdery precipitated substances decreased and the CaCO3 content and pH increased.


Introduction
Coal is one of the most important energy sources in the world. The global coal consumption in 2017 was 3731 Mtoe, which accounted for 28% of the total energy consumption [1]. In China, the reserves of coal are much higher than those of oil and natural gas [2]. The total coal consumption in China in 2018 was 2.74 billion tons, which accounted for 59% of disposable energy [3]. The consumption of sulfur-containing coal led to a large amount of SO 2 emission from industrial production [4,5]. From 2000 to 2006, the total SO 2 emission in China increased by 53% from 21.7 to 33.2 million tons [6]. According to the report of the State of the Environment in China, the total SO 2 emission was almost 20 million tons as of 2014 [7].
Because of the special production process, the SO 2 concentration in industrial environment is much higher than that in general atmospheric environment. The SO 2 concentration in Wuhan Iron and Steel Plant was monitored. The SO 2 concentration of the crystallization-tank section in the coking ammonium sulfate workshop was 6.65 mg/m 3 [8], and the maximum SO 2 concentration in Wuhan No. 2 steelmaking plant was 1.29 mg/m 3 [9]. SO 2 diffuses into concrete and reacts with calcium-containing substances, which reduces the alkalinity of pore solution and expands the solid-phase volume, that is, concrete sulfuration [10]. On the one hand, the expansion failure of concrete from outside to inside leads to spalling and damage

Theoretical Chemical Reaction
SO 2 diffuses into concrete through micro-pores and micro-cracks, and continuously dissolves in water to form H 2 SO 3 . On the one hand, H 2 SO 3 ionizes into H + and SO 2− 3 . On the other hand, H 2 SO 3 reacts with O 2 in the air to form H 2 SO 4 , which ionizes into H + and SO 2− 4 [26].
Gypsum reacts with calcium aluminate in concrete to form calcium sulfoaluminate. When the gypsum content is low, gypsum and calcium aluminate first form AFm.
With the continuous formation of gypsum, AFm is transformed into AFt.
When the temperature exceeds a certain temperature, AFt decomposes into AFm, and calcium sulfate. When the temperature dropped to a room temperature, AFm could transform to AFt again in moist environment [23].

Materials
P.O. 42.5 Ordinary Portland cement, in compliance with Chinese Standard GB 175-2007 [28], was used in this experiment. The mass percent of cement chemical composition determined by X-ray fluorescence spectroscopy (EDX8000, Shimadzu, Kyoto, Japan) is listed in Table 1. The fine aggregate used in this study had a fineness modulus of 2.34. The coarse aggregate used granite macadam with continuous grading from 5 to 25 mm. A polycarboxylic-based superplasticizer was used in the experiment with water reducing rate of 35%, solid content of 40%, and alkali content of 6.5%. The amount of water reducer was 0.5% of the cementitious material.

Specimen Preparation
The experimental specimens with dimension of 100 mm × 100 mm × 100 mm were prepared using three mixture proportions. The specimens were prepared in three mixtures, namely C40, C30, and C20, with three water-cement ratios of 0.37, 0.47, and 0.57, respectively. The specimens were moisture-cured at 20 ± 2 • C and 95% relative humidity for 28 days and then dry-cured until 90 days. The mixture proportion of concrete are listed in Table 2. According to Chinese Standard CB/T 50081-2019 [29], three specimens of each water-cement ratio were used for compressive strength at the age of 28 days. The compressive strength and its standard deviation are shown in Table 2.

Sulfuration Test
SO 2 is mainly generated by combustion, and it is emitted after desulfurization treatment through the chimney. The wet-desulfurization technology without Gas Gas Heater (GGH) is usually adopted in China. The SO 2 concentration and temperature of the flue gas decreased, and the humidity increased. The average temperature of the flue gas was approximately 50 • C, and the humidity at the tail device reached saturation. Therefore, the temperature in this experiment was 50 • C, and the relative humidity was 98%. There is no generic standard of concrete sulfuration test method. According to the concrete accelerated carbonation test method in accordance with Chinese Standard GB/T 50082-2009 [30], the accelerated sulfuration test method was established. The self-designed concrete sulfuration test chamber was used in this test, as shown in Figure 1. The test procedure was performed as follows: • The specimens were taken out and dried at 60 • C for 48 h. When the specimen temperature dropped to room temperature, the specimen surface was sealed with epoxy resin except for the two opposite sides. The purpose is to ensure that the transmission and reaction of SO 2 in the concrete is one-dimensional.

•
The specimens were placed on a bracket in the concrete sulfuration test chamber. The distance between adjacent specimens was not less than 50 mm.

•
The ambient temperature in the sulfuration test chamber was 50 • C, the relative humidity was 98%, and the SO 2 concentration was 9000 ppm.

•
The experiments were performed for four different test ages, including 2, 5, 10, and 20 days, and the specimens were taken out from the sulfuration test chamber. The powdery precipitated substances on the specimens surface were collected for 20 days.
Materials 2020, 13, x FOR PEER REVIEW 5 of 15 accelerated sulfuration test method was established. The self-designed concrete sulfuration test chamber was used in this test, as shown in Figure 1. The test procedure was performed as follows: • The specimens were taken out and dried at 60 ℃ for 48 h. When the specimen temperature dropped to room temperature, the specimen surface was sealed with epoxy resin except for the two opposite sides. The purpose is to ensure that the transmission and reaction of SO2 in the concrete is one-dimensional.

•
The specimens were placed on a bracket in the concrete sulfuration test chamber. The distance between adjacent specimens was not less than 50 mm.

•
The ambient temperature in the sulfuration test chamber was 50 ℃, the relative humidity was 98%, and the SO2 concentration was 9000 ppm.

•
The experiments were performed for four different test ages, including 2, 5, 10, and 20 days, and the specimens were taken out from the sulfuration test chamber. The powdery precipitated substances on the specimens surface were collected for 20 days.

SEM
SEM (Quanta 200, FEI, Hillsboro, OR, USA) test was employed to observe the morphology of concrete after the sulfuration reaction. Single specimen was used at one test age for each watercement ratio. The specimen was cut into slices of 1-2 mm thickness from the surface, and then the samples with the diameter about 5 mm were obtained by pliers and tweezers. A few samples were taken out at different test ages and were immersed in ethanol to terminate all the reactions, including the hydration reaction and carbonation reaction. Before the test, the samples were removed from ethanol, air dried at room temperature for 10 min, and dried at 50 ℃ for 24 h. A platinum coating was subsequently applied onto the surface of the specimens for enhanced electrical conductivity. The measurements were carried out with the secondary electron detector. The accelerating voltage was 20 kV and the working distance was 10 mm.

Pore-Solution pH Test
Three specimens were used for each water-cement ratio. The powdery precipitated substances were gathered from the exposed surface of specimens at the age of 20 days. The powdery precipitated substances of three specimens were crushed and mixed evenly.
Pore-solution pH test was performed to determine the pH of the powdery precipitated substances, and it was conducted three times for each water-cement ratio. The experimental method was performed as follows: • The powdery precipitated substances of sulfated concrete were crushed and mixed, and the powder was then passed through a 0.16-mm sieve.

•
The powder was placed in a weighing bottle and dried at 50 ℃ for 24 h. Then, the weighing bottle was sealed and transferred in a drying and cooled down to room temperature.

•
Three grams of dried powder were placed in a triangular flask. A pipette was used to accurately

SEM
SEM (Quanta 200, FEI, Hillsboro, OR, USA) test was employed to observe the morphology of concrete after the sulfuration reaction. Single specimen was used at one test age for each water-cement ratio. The specimen was cut into slices of 1-2 mm thickness from the surface, and then the samples with the diameter about 5 mm were obtained by pliers and tweezers. A few samples were taken out at different test ages and were immersed in ethanol to terminate all the reactions, including the hydration reaction and carbonation reaction. Before the test, the samples were removed from ethanol, air dried at room temperature for 10 min, and dried at 50 • C for 24 h. A platinum coating was subsequently applied onto the surface of the specimens for enhanced electrical conductivity. The measurements were carried out with the secondary electron detector. The accelerating voltage was 20 kV and the working distance was 10 mm.

Pore-Solution pH Test
Three specimens were used for each water-cement ratio. The powdery precipitated substances were gathered from the exposed surface of specimens at the age of 20 days. The powdery precipitated substances of three specimens were crushed and mixed evenly.
Pore-solution pH test was performed to determine the pH of the powdery precipitated substances, and it was conducted three times for each water-cement ratio. The experimental method was performed as follows: • The powdery precipitated substances of sulfated concrete were crushed and mixed, and the powder was then passed through a 0.16-mm sieve.
• The powder was placed in a weighing bottle and dried at 50 • C for 24 h. Then, the weighing bottle was sealed and transferred in a drying and cooled down to room temperature. • Three grams of dried powder were placed in a triangular flask. A pipette was used to accurately measure 60 mL of distilled water. Next, the distilled water was poured into the triangular flask. The triangular flask was shaken in an oscillator (Xinbode, Tianjin, China) for 10 min. After standing for 24 h, the solution in the triangular flask was filtered using a fast quantitative filter paper.

XRD
XRD (D/Max 2200, Rigaku, Tokyo, Japan) analysis was used to examine the mineral-phase composition of the surface powdery precipitated substances. The powdery precipitated substances were first pulverized using an agate mortar, and then passed through an 80-µm sieve. The powder was dried at 50 Rigaku for 24 h. The diffraction used a Cu-Kα source. Scanning was performed from 5 • to 45 • at a step of 0.02 • .

TG-DSC
The TG-DSC (TGA/DSC 2, Mettler Toledo, Zurich, Switzerland) method was used for qualitative and quantitative investigation of the surface powdery precipitated substances. The preparation method of the powder was the same as that of the powder used in the XRD method. The samples were heated in air atmosphere from 30 to 900 • C at a heating rate of 10 • C/min.

Appearance Analysis
The appearance of the C30 specimens for sulfuration ages of 0, 2, 5, 10, and 20 days is shown in Figure 2. The changes in the concrete appearance mainly underwent four stages. Firstly, the surface of the hardened concrete was dark grey and smooth, and some capillary pores were visible to the naked eyes ( Figure 2a). Then, light-grey spots began to appear on the concrete surface at the initial stage of sulfuration reaction ( Figure 2b). The light-grey area on the surface increased with the progress in the sulfuration reaction, and the surface became smoother and denser ( Figure 2c). Next, all the areas on the specimen surface became light-grey, and the corners began to peel off ( Figure 2d). Finally, a large amount of white powdery precipitated substances appeared on the specimen surface, and the surface became rough (Figure 2e).
Materials 2020, 13, x FOR PEER REVIEW 6 of 15 standing for 24 h, the solution in the triangular flask was filtered using a fast quantitative filter paper.

XRD
XRD (D/Max 2200, Rigaku, Tokyo, Japan) analysis was used to examine the mineral-phase composition of the surface powdery precipitated substances. The powdery precipitated substances were first pulverized using an agate mortar, and then passed through an 80-μm sieve. The powder was dried at 50 Rigaku for 24 h. The diffraction used a Cu-Kα source. Scanning was performed from 5° to 45° at a step of 0.02°.

TG-DSC
The TG-DSC (TGA/DSC 2, Mettler Toledo, Zurich, Switzerland) method was used for qualitative and quantitative investigation of the surface powdery precipitated substances. The preparation method of the powder was the same as that of the powder used in the XRD method. The samples were heated in air atmosphere from 30 to 900 ℃ at a heating rate of 10 ℃/min.

Appearance Analysis
The appearance of the C30 specimens for sulfuration ages of 0, 2, 5, 10, and 20 days is shown in Figure 2. The changes in the concrete appearance mainly underwent four stages. Firstly, the surface of the hardened concrete was dark grey and smooth, and some capillary pores were visible to the naked eyes ( Figure 2a). Then, light-grey spots began to appear on the concrete surface at the initial stage of sulfuration reaction (Figure 2b). The light-grey area on the surface increased with the progress in the sulfuration reaction, and the surface became smoother and denser ( Figure 2c). Next, all the areas on the specimen surface became light-grey, and the corners began to peel off ( Figure 2d). Finally, a large amount of white powdery precipitated substances appeared on the specimen surface, and the surface became rough (Figure 2e).
The appearance of light-grey spots on the concrete surface was due to the formation of sulfated products. These products gradually increased as the sulfuration progressed, and, therefore, the lightgrey area on the concrete surface increased. Because the sulfated products blocked the pores in the concrete, the surface initially became denser. When the sulfated products exceeded the pore capacity, cracks formed on the concrete surface. Subsequently, powdery precipitated substances appeared.  The appearance of light-grey spots on the concrete surface was due to the formation of sulfated products. These products gradually increased as the sulfuration progressed, and, therefore, the light-grey area on the concrete surface increased. Because the sulfated products blocked the pores in the concrete, the surface initially became denser. When the sulfated products exceeded the pore capacity, cracks formed on the concrete surface. Subsequently, powdery precipitated substances appeared.
The appearance of concrete at water-cement ratios of 0.37, 0.47, and 0.57 in the 20 days sulfuration is shown in Figure 3a. Only a thin layer of powdery substances appeared on the surface of the C20 specimen, and powdery precipitated substances were produced in part of the C30 specimen. The C40 specimen produced a large amount of powdery precipitated substances throughout. The corrosion layer of the C40 specimen was the thickest, and the expansion volume was the largest.
The appearance of concrete at water-cement ratios of 0.37, 0.47, and 0.57 in the 20 days sulfuration is shown in Figure 3a. Only a thin layer of powdery substances appeared on the surface of the C20 specimen, and powdery precipitated substances were produced in part of the C30 specimen. The C40 specimen produced a large amount of powdery precipitated substances throughout. The corrosion layer of the C40 specimen was the thickest, and the expansion volume was the largest.
After cleaning the powdery precipitated substances, the appearance of sulfated concrete for 20 days is shown in Figure 3b. Compared with that shown in Figure 3a, the appearance of the C20 specimen remained almost unchanged. However, the visible pores in the surface of C30 and C40 specimens were larger than that of C20 specimen, and the surface was rough and honeycomb after sulfuration reaction. The change of specimens' appearance increased with the decrease of watercement ratios.
The cement content per unit volume of concrete increased with the decrease in the water-cement ratio, thus the content of the hydration products Ca(OH)2 increased. Increasing the alkali content can accelerate the reaction rate of concrete [31]. Therefore, the change in the appearance of concrete with the low water-cement ratio was more serious.

Microscopic Fracture Surface Observation
The crack development trend of the C30 specimen at different sulfuration ages is shown in Figure 4. An initial damage existed in concrete due to the few cracks and voids (Figure 4a). A large amount of white flocculent substances remained, and the initial cracks were obviously reduced after sulfuration for 2 days (Figure 4b). After sulfuration for 5 days, many white substances were generated inside the concrete (Figure 4c). Figure 4d shows the enlargement of the rectangular area shown in the Figure 4c, and a large amount of needle-like materials can be seen. When the sulfuration ages increased to 10 days, the needle-like products disappeared, and new micro-cracks appeared ( Figure  4e). After sulfuration for 20 days, cracks appeared in many locations, and the crack width increased (Figure 4f).
After sulfuration for 10 and 20 days, a circle of white substances appeared around the cracks, indicating that the cracks were caused by the formation of a large amount of solid sulfated products, which exceeded the capacity of the pore structures. Sulfated products were generated inside the concrete; thus, the solid-phase volume in the cracks and voids continuously increased. The cracks and voids were significantly reduced. Finally, a large amount of sulfated products blocked the pores, resulting in excessive internal stress [16]. Therefore, the concrete cracked, and the products precipitated on the concrete surface. After cleaning the powdery precipitated substances, the appearance of sulfated concrete for 20 days is shown in Figure 3b. Compared with that shown in Figure 3a, the appearance of the C20 specimen remained almost unchanged. However, the visible pores in the surface of C30 and C40 specimens were larger than that of C20 specimen, and the surface was rough and honeycomb after sulfuration reaction. The change of specimens' appearance increased with the decrease of water-cement ratios.
The cement content per unit volume of concrete increased with the decrease in the water-cement ratio, thus the content of the hydration products Ca(OH) 2 increased. Increasing the alkali content can accelerate the reaction rate of concrete [31]. Therefore, the change in the appearance of concrete with the low water-cement ratio was more serious.

Microscopic Fracture Surface Observation
The crack development trend of the C30 specimen at different sulfuration ages is shown in Figure 4. An initial damage existed in concrete due to the few cracks and voids (Figure 4a). A large amount of white flocculent substances remained, and the initial cracks were obviously reduced after sulfuration for 2 days (Figure 4b). After sulfuration for 5 days, many white substances were generated inside the concrete (Figure 4c). Figure 4d shows the enlargement of the rectangular area shown in the Figure 4c, and a large amount of needle-like materials can be seen. When the sulfuration ages increased to 10 days, the needle-like products disappeared, and new micro-cracks appeared (Figure 4e). After sulfuration for 20 days, cracks appeared in many locations, and the crack width increased (Figure 4f).  (Figure 5b). The main constituent elements of the needle-like substance were Ca, S, Al, O, and Si; therefore, the needle-like substance was AFt. With the progress in the sulfuration reaction, an increasing amount of needle-like products appeared near the hydration products (Figure 5c), and AFt was continuously formed.
C-S-H reacted with H + dissolved in water to form non-gelling silica gels, which were stacked with one another (Figure 5d). The cementitious materials in the specimen ran off, and Ca(OH)2 was gradually consumed. Then, the pH of the pore solution rapidly decreased, resulting in the decomposition of the AFt crystal. Subsequently, the number of large pores increased (Figure 5e). At the same time, AFt was decomposed by H + to form a plate-like substance (Figure 5f). The main elements were S, Ca, and O, indicating that this type plate crystal was gypsum. Finally, gypsum was produced in large quantity (Figure 5g). After sulfuration for 10 and 20 days, a circle of white substances appeared around the cracks, indicating that the cracks were caused by the formation of a large amount of solid sulfated products, which exceeded the capacity of the pore structures. Sulfated products were generated inside the concrete; thus, the solid-phase volume in the cracks and voids continuously increased. The cracks and voids were significantly reduced. Finally, a large amount of sulfated products blocked the pores, resulting in excessive internal stress [16]. Therefore, the concrete cracked, and the products precipitated on the concrete surface.
The mineral-phase composition of the C30 specimen at different ages is shown in Figure 5. Many flocculent hydration products appeared in the hardened concrete (Figure 5a). With the diffusion and dissolution of SO 2 into the concrete, H + , SO 2− 4 , and SO 2− 3 reacted with the hydration products at the cracks and pores to form a needle-like substance (Figure 5b). The main constituent elements of the needle-like substance were Ca, S, Al, O, and Si; therefore, the needle-like substance was AFt. With the progress in the sulfuration reaction, an increasing amount of needle-like products appeared near the hydration products (Figure 5c), and AFt was continuously formed.

Powdery Precipitated Substances Analysis
To qualitatively and quantitatively study the degree of damage with different water-cement ratios, we determined the mass, composition, and pH of powdery precipitated substances. Figure 6 shows the mass of the powdery precipitated substances at different water-cement ratios. The average mass of powdery precipitated substances in the C20, C30 and C40 specimens was 3.67, 15.40, and 41.03 g, respectively. Concrete with a high water-cement ratio had larger and more pores, which played the role of a capacity to absorb expansion caused by the sulfated products [32]. Therefore, the specimen with lower water-cement ratio was more likely to crack. Cracks can further C-S-H reacted with H + dissolved in water to form non-gelling silica gels, which were stacked with one another (Figure 5d). The cementitious materials in the specimen ran off, and Ca(OH) 2 was gradually consumed. Then, the pH of the pore solution rapidly decreased, resulting in the decomposition of the AFt crystal. Subsequently, the number of large pores increased (Figure 5e). At the same time, AFt was decomposed by H + to form a plate-like substance (Figure 5f). The main elements were S, Ca, and O, indicating that this type plate crystal was gypsum. Finally, gypsum was produced in large quantity (Figure 5g).

Powdery Precipitated Substances Analysis
To qualitatively and quantitatively study the degree of damage with different water-cement ratios, we determined the mass, composition, and pH of powdery precipitated substances. Figure 6 shows the mass of the powdery precipitated substances at different water-cement ratios. The average mass of powdery precipitated substances in the C20, C30 and C40 specimens was 3.67, 15.40, and 41.03 g, respectively. Concrete with a high water-cement ratio had larger and more pores, which played the role of a capacity to absorb expansion caused by the sulfated products [32]. Therefore, the specimen with lower water-cement ratio was more likely to crack. Cracks can further accelerate the diffusion and reaction of SO 2 . Therefore, the mass of powdery precipitated substances increased with the decrease of water-cement ratios. Materials 2020, 13, x FOR PEER REVIEW 10 of 15 Figure 6. The mass of the powdery precipitated substances at different water-cement ratios. Three samples were tested for each water-cement ratio.

Composition
The composition of the powdery precipitated substances was analyzed by XRD, as shown in Figure 7. Two types of obvious peaks exist in the XRD pattern as is seen through the comparative analysis of the corresponding characteristic angles. The characteristic angles of gypsum are 11.62°, 20.72°, 23.36°, 29.10°, 31.10°, 33.34°, 35.96°, 40.62°, and 43.34°. The characteristic angles of quartz (SiO2) are 26.62°, 36.54°, and 39.44°. Therefore, the sulfated product of the powdery precipitated substances was mainly gypsum.  Figure 6. The mass of the powdery precipitated substances at different water-cement ratios. Three samples were tested for each water-cement ratio.

Composition
The composition of the powdery precipitated substances was analyzed by XRD, as shown in  Figure 8 shows the thermal analysis curves of the powdery precipitated substances at three water-cement ratios. Three obvious endothermic peaks were observed before 900 • C, and they represent the dehydration decomposition of three substances. The temperatures of the endothermic peaks were mainly 130-150, 500-520, and 740-770 • C. Related studies have shown that the three temperature ranges correspond to gypsum, Ca(OH) 2 , and CaCO 3 , respectively. Figure 7. Two types of obvious peaks exist in the XRD pattern as is seen through the comparative analysis of the corresponding characteristic angles. The characteristic angles of gypsum are 11.62°, 20.72°, 23.36°, 29.10°, 31.10°, 33.34°, 35.96°, 40.62°, and 43.34°. The characteristic angles of quartz (SiO2) are 26.62°, 36.54°, and 39.44°. Therefore, the sulfated product of the powdery precipitated substances was mainly gypsum.  Figure 8 shows the thermal analysis curves of the powdery precipitated substances at three water-cement ratios. Three obvious endothermic peaks were observed before 900 ℃, and they represent the dehydration decomposition of three substances. The temperatures of the endothermic peaks were mainly 130-150, 500-520, and 740-770 ℃. Related studies have shown that the three temperature ranges correspond to gypsum, Ca(OH)2, and CaCO3, respectively. The endothermic peak of gypsum was the most obvious, indicating that the content of gypsum in the powdery precipitated substances was the highest. Gypsum dehydrates and decomposes into hemihydrate gypsum (CaSO4·0.5H2O) at 150 ℃, and hemihydrate gypsum becomes anhydrous gypsum (CaSO4) at 170 ℃. The content of gypsum was obtained by chemical reaction, i.e., Equation  Figure 8 shows that the gypsum of the powdery precipitated substances at water-cement ratios of 0.57, 0.47, and 0.37 was decomposed at 101.37-170.76, 106.90-173.76, and 111.23-174.76 ℃, respectively. The mass-loss rates of concrete were 12.41%, 11.51% and 11.83%, respectively. The mass percentage of gypsum was calculated using Equation (18), and the gypsum contents in the C20, C30 and C40 specimens were 59.29%, 55.01%, and 56.53%, respectively.  The endothermic peak of gypsum was the most obvious, indicating that the content of gypsum in the powdery precipitated substances was the highest. Gypsum dehydrates and decomposes into hemihydrate gypsum (CaSO 4 ·0.5H 2 O) at 150 • C, and hemihydrate gypsum becomes anhydrous gypsum (CaSO 4 ) at 170 • C. The content of gypsum was obtained by chemical reaction, i.e., Equation (17). Figure 8 shows that the gypsum of the powdery precipitated substances at water-cement ratios of 0.57, 0.47, and 0.37 was decomposed at 101.37-170.76, 106.90-173.76, and 111. .76 • C, respectively. The mass-loss rates of concrete were 12.41%, 11.51% and 11.83%, respectively. The mass percentage of gypsum was calculated using Equation (18), and the gypsum contents in the C20, C30 and C40 specimens were 59.29%, 55.01%, and 56.53%, respectively. Based on the reported results, the highest gypsum content was found for the highest water-cement ratio. The content of cement per unit concrete decreased with the increase of water-cement ratios. Therefore, the content of reaction product gypsum was likely to decrease. However, the powdery precipitated substances were mainly composed of gypsum and fine aggregate, and the fine aggregate of the powdery precipitated substances in the C20 specimen was the lowest because of the minimal surface damage. Therefore, the relative gypsum content in the C20 specimen was the highest.
The endothermic peak of Ca(OH) 2 in the C40 specimens was most obvious, indicating that the content of Ca(OH) 2 was the highest. Ca(OH) 2 dehydrates and decomposes into CaO, as shown in Equation (19). The contents of Ca(OH) 2 in the C20, C30 and C40 specimens were 9.85%, 8.81%, and 11.42%, respectively.
According to the research results, the highest Ca(OH) 2 content was found for the lowest water-cement ratio. The cement content per unit volume decreased with the increase of the water-cement ratios. Therefore, the content of Ca(OH) 2 per unit concrete decreased. However, fine aggregate of the C20 specimen was less than that of C30 specimen. Therefore, the Ca(OH) 2 content of C20 specimen was slightly larger.
CaCO 3 was observed in the C20 and C30 specimens, and the endothermic peak of CaCO 3 could not be obviously detected in the C40 specimens. CaCO 3 decomposes into CaO and CO 2 , as shown in Equation (20). The content of CaCO 3 in the C20 and C30 specimens was 16.68% and 8.25%, respectively. The carbonation of concrete occurred in air, and the larger was the water-cement ratio, the easier was the carbonation reaction. Therefore, the content of CaCO 3 increased with the increase of water-cement ratios.
CaCO 3 (s) → CaO(s) + CO 2 (g) (20) The main components of the powdery precipitated substances were gypsum, SiO 2 , Ca(OH) 2 , and CaCO 3 . SiO 2 is the main component of sand and granite, and Ca(OH) 2 is the hydration product of concrete. CaCO 3 is the carbonation product formed by the reaction of Ca(OH) 2 with CO 2 . Therefore, the final sulfated product was gypsum. Figure 9 shows the pH of the powdery precipitated substances at different water-cement ratios. The average pH of the powdery precipitated substances in C20, C30, and C40 specimens was 4.20, 3.25, and 3.19, respectively. The specimen appearance at a lower water-cement ratio was rougher and more honeycomb, thus it had a larger specific surface area and could absorb more H + [33]. Therefore, the pH of concrete with a lower water-cement ratio was likely to be lower. However, it can be seen in Section 4.3.2 that the alkaline substances Ca(OH) 2 were the most in C40 specimens. Therefore, the pH of C20 specimen was much higher than that of C30 and C40 specimens. The pH of C40 specimen was just a little lower than that of C30 specimen.

Pore-Solution pH
3.25, and 3.19, respectively. The specimen appearance at a lower water-cement ratio was rougher and more honeycomb, thus it had a larger specific surface area and could absorb more H + [33]. Therefore, the pH of concrete with a lower water-cement ratio was likely to be lower. However, it can be seen in Section 4.3.2 that the alkaline substances Ca(OH)2 were the most in C40 specimens. Therefore, the pH of C20 specimen was much higher than that of C30 and C40 specimens. The pH of C40 specimen was just a little lower than that of C30 specimen. Figure 9. The pH of the powdery precipitated substances at different water-cement ratios. Three samples were tested for each water-cement ratio.
The pH of the powdery precipitated substances at different water-cement ratios was acidic; thus, it was difficult for calcium sulfoaluminate to exist. Therefore, obtaining AFt and AFm in the powdery precipitated substances was impossible.

Conclusions
In this study, the variations in the appearance, development trend of a micro-crack, producttype changes, and powdery precipitated substances of sulfated concrete were evaluated. The main conclusions are as follows: (1) The surface of hardened concrete was dark grey and smooth, and light-grey spots appeared on the surface at the initial stage of sulfuration reaction. Then, the light-grey area increased, and the surface increasingly became denser. Next, all areas on the specimen surface became light grey, and  Figure 9. The pH of the powdery precipitated substances at different water-cement ratios. Three samples were tested for each water-cement ratio.
The pH of the powdery precipitated substances at different water-cement ratios was acidic; thus, it was difficult for calcium sulfoaluminate to exist. Therefore, obtaining AFt and AFm in the powdery precipitated substances was impossible.

Conclusions
In this study, the variations in the appearance, development trend of a micro-crack, product-type changes, and powdery precipitated substances of sulfated concrete were evaluated. The main conclusions are as follows: (1) The surface of hardened concrete was dark grey and smooth, and light-grey spots appeared on the surface at the initial stage of sulfuration reaction. Then, the light-grey area increased, and the surface increasingly became denser. Next, all areas on the specimen surface became light grey, and the corners of the concrete began to fall off. Finally, a large amount of powdery precipitated substances appeared. The content of the powdery precipitated substances decreased with the increase in the water-cement ratio.
(2) The initial cracks in the hardened concrete were obviously reduced because of the formation of sulfated products. The number of cracks initially decreased with the sulfuration age. Following the sulfuration-reaction process, the products exceeded the capacity of the pores in the concrete; thus, new cracks appeared. In the later stage of sulfuration reaction, the continuous increase in solid products in the pores increased the number of cracks and broadened their width.
(3) AFt was formed at the initial stage of sulfuration reaction. The hydration products in the concrete decomposed by H + . C-S-H was decomposed into a non-gelling silica gel. Ca(OH) 2 constantly dissolved and reacted; thus, the alkalinity of the concrete decreased. Finally, AFt decomposed into gypsum.
(4) The average mass of powdery precipitated substances in the C20, C30, and C40 specimens was 3.67, 15.40, and 41.03 g, respectively. The pH increased with the increase in the water-cement ratio, and the pH were between 3.19 and 4.20. The final product of the powdery precipitated substances was gypsum, and the mass percentages of gypsum at different water-cement ratios were more than 50%. The highest Ca(OH) 2 content was found for the lowest water-cement ratio, and the CaCO 3 content increased with the increase in the water-cement ratios.