XRD-Rietveld Method for Evaluating the Leaching Characteristics of Hardened Cement Paste in Flowing Water

In order to study the leaching characteristics of hardened cement paste by flowing environmental water, indoor simulated leaching tests of Portland cement paste were carried out at different flow velocities with home-made flow field device and tap water as the erosion medium. By means of mechanical properties tests, X-ray diffraction (XRD), and Rietveld full-spectrum quantitative analysis of specimens after leaching at different ages, the change rules of the flexural strength, main phase composition, and mass fraction of the specimens after leaching were obtained. +e results show that, within 90 days, the flow rate in the test has no obvious effect on the leaching of hardened cement paste. With the increase in age, the leaching of hardened cement paste gets more obvious. Rietveld full-spectrum quantitative analysis shows that the leaching of hardened cement paste changes with the relative content of CaCO3 and Ca(OH)2. +e higher the relative content of CaCO3 is, the better compaction and leaching resistance the paste has. +e leaching resistance of hardened cement paste is also related to the gradient of Ca(OH)2. +e smaller the gradient is, the better the leaching resistance is. +e Rietveld full-spectrum quantitative analysis method can be used as a method to evaluate the leaching effect of hardened cement paste.


Introduction
Concrete structures such as dams, bridges, ports, wharves, and tunnels are vulnerable to erosion by flowing water [1,2], which affects the durability and service life of them [3,4]. Calcium ion is the main component of hydration products of Portland cement, and its content affects the stability of hydration products. Flowing water will cause the diffusion and leaching of Ca(OH) 2 in concrete and the decalcification of C-S-H gel [5,6]. e leaching of calcium ion in concrete is largely responsible for the reduction of concrete strength [7], and the dynamic leaching rate is higher than the static leaching rate. e leaching test methods include the direct method and indirect method. Direct method is usually used to test the leaching of specimens directly with pure water or deionized water [8]. e indirect method is to use chemical solution to accelerate the test, such as ammonium nitrate solution [9], ammonium chloride solution, and electrochemical theory. [10]. In terms of test efficiency, the leaching process of specimens in the direct method is relatively slow [11,12]. e indirect method has the characteristic of fast erosion speed [13,14], but the chemical reaction is complex and there are many influencing factors. e leaching of concrete is mainly caused by the leaching of hardened cement paste, accompanied by the change in crystal content of hydration products, such as Ca(OH) 2 , CaCO 3 , C-S-H, and SiO 2 .
ere are many parameters to evaluate the leaching effect, such as the permeability of concrete, the change in pH value of leachate, the quantity of dissolved chemical substances, and the loss of mass and strength. [15]. XRD can be used to identify and quantify the mineral composition of cement-based materials, discriminate hydration products, study the hydration process of cement, and determine the hydration rate of cement clinker minerals [9,16]. e method based on full-spectrum structure fitting (Rietveld) can be used to analyze the influence of hydration properties [17,18], phase quantification, and evolution of cement-based materials [19,20] and the influence of pozzolanic action [21,22]. e standard samples are not required in the method but only the structure information, so it has high accuracy. Actually, there are two kinds of quantitative methods for phase analysis by X-ray diffraction: one is the traditional quantitative method (external standard method and internal standard method); the second method is based on fullspectrum structure fitting (Rietveld diffraction spectrum analysis). e former method requires a standard sample when quantifying. e later, full-spectrum fitting quantitative analysis, has a high accuracy, which only needs to know the structure information of the phase, with no need of the standard sample, and can call the crystal structure information of the phase in the software.
In order to simulate the leaching characteristics of concrete in real-water environment, especially the leaching of hardened cement paste in underground concrete structure by groundwater of reservoir, the ordinary Portland hardened cement paste was taken as the research object and a homemade flow leaching device was used. e mechanical properties, main hydration products composition, and relative mass change in the hardened cement paste specimens leached with different flow rates at different ages were analyzed in order to obtain the leaching characteristics of hardened cement paste under the action of flowing tap water.

Materials and Mix Proportions.
e cement used in the test was ordinary Portland cement (OPC) and superfine cement (SMC) with 5% silica fume, and the strength grade was 42.5. e physical properties of the materials are shown in Tables 1 and 2. Ultrafine cement was obtained by grinding OPC. Its average particle size was 4.25 μm, and the specific surface area was 3770 m 2 /kg. Tap water was used as mixing water and leaching medium. e pH value of tap water was 7.5-8.0.

Specimens for Leaching and Mechanical Tests.
e specimens were prisms of size 10 mm × 10 mm × 60 mm. e water-binder ratio was 0.55. After the cement paste was molded, the standard curing time was 24 h, and then, the mold was removed and the leaching test was carried out.
ere were 3 specimens in each group in the flexural strength test. e strength value was taken as the average of the measured values of three specimens.

XRD Testing.
Sampling positions were 0-2 mm (surface), 2-4 mm (middle), and >4 mm (core) depth, respectively (shown in Figure 1). e samples were placed in an agate bowl and grinded manually until they passed through 200 meshes (0.074 mm) sieve, and then, 5 g samples were randomly taken. In order to prevent carbonization during grinding, anhydrous ethanol was added to the grinder. e samples were dried to constant weight at 40°C in the oven, and then, the XRD spectra were collected in X-ray diffractometer.

Tests and Methods.
After 24-hour standard curing, the specimens were dismantled and soaked in water tanks, respectively.
e experimental flow velocities were 0 mm/s, 2.8 mm/s, and 11.2 mm/s, respectively. e experimental flow velocities were calculated according to the groundwater flow field of a reservoir project. e actual average flow velocity was 2.8 mm/s and 0 mm/s meant that groundwater did not flow. e test ages were 3 d, 7 d, 28 d, 56 d, and 90 d, respectively.
As shown in Figure 2, the flow field test device consisted of an upper water tank, a lower water tank, a flow control valve, a flow tank, a support platform for the upper water tank, a support platform for the flow tank, a sand cushion, a water pump, and a water pipe. After the water in the upper tank flowed out from the flow control valve, it entered the flume to erode the specimens on the sand cushion. ere was an opening at the end of the flume. e water flowed out from the opening and entered the lower tank. e upper water tank was a high-strength plastic water tank with a capacity of 200 L. e water in the tank was controlled intelligently to maintain a uniform height. e flow tank was a PVC rectangular tank with a net width of 92 mm. Six specimens could be arranged evenly along the width direction. e specific operation steps were as follows: (1) Appropriate amount of water was injected into the upper water tank, and then, the flow control valve was opened; the flow rate was calibrated by measuring cylinder and stopwatch, and the opening degree of the valve was adjusted. (2) e test blocks were put into the flow tank, and a gap was kept between the test blocks for the flow to pass through, while a part of the test blocks was reserved for curing in still water under the same conditions. (3) When reaching the required erosion time, the specimens were tested for flexural strength and XRD. Six specimens in each group were tested, of which three were for flexural strength, and the other three for XRD. e testing age of XRD was 28 d, 56 d, and 90 d. Among them, the 56 d samples were also tested for XRD with different depths.

Leaching Resistance Coefficient
Tests. e ratio of the flexural strength of specimens leached with different flow rates to that of hydrostatically leached specimens at the same age was defined as the leaching resistance coefficient, and the leaching resistance coefficient was used to evaluate the leaching effect of cement-based materials. e larger the coefficient was, the better the leaching resistance of the material was. e flexural strength of the specimens was tested by three-point bending-tension test, and L was 40 mm: 2 Advances in Civil Engineering where K f is the leaching resistance coefficient; f e is the flexural strength of the test piece under flowing water erosion, MPa; and f 0 is the flexural strength of the test piece under static water, MPa.

X-Ray Powder Diffraction (XRD).
When the multiphase system is irradiated by monochrome X-ray, the diffraction patterns of each phase in the diffraction space superimpose each other to form a one-dimensional diffraction pattern. In the process of weighted overlapping addition of powder diffraction spectra of each phase, the positions of the diffraction lines of each phase will not change, and the intensity of the diffraction lines varies with the percentage of the phase in the mixture (volume or mass), the scattering force, and the absorption force of other phases, and the scale factor is the reflection of the intensity change. e obtained dry hydration samples were subjected to X-ray diffraction (XRD) analysis using a X-ray diffractometer (D8 ADVANCE, Bruker AXS Corporation, GER) employing Cu-Kα radiation (λ � 0.15418 nm, 40 kV, 50 mA) over scanning range 2θ � 15°∼ 65°at step width 2°per min.

Calculation and Analysis of Leaching Resistance
Coefficient.
e leaching and decomposition of hydration products occurred after the hardened cement paste was eroded by flowing water, which resulted in the pore structure becoming loose and compactness decreasing and ultimately    Advances in Civil Engineering the mechanical properties deteriorating [7,23]. e threepoint flexural tensile strength of specimens at different ages and velocities was tested. e flexural strength is shown in Figure 3(a), and the results of calculation of leaching resistance coefficient by formula (1) are shown in Figure 3(b). From Figure 3(a), it can be seen that the flexural strength of the specimens increased with the increase in age, the growth rate was relatively high before 28 days, and the OPC is more obvious. e flexural strength of OPC decreased slightly after 28 days, and the decrease in flexural strength of OPC was more obvious than that of SMC. As can be seen from Figure 3(b), the leaching resistance coefficient of OPC decreased gradually, and the boundary point was about 28 days later. e leaching resistance coefficient of SMC increased and approached 1 with the increase in time. e designed flow rate had little effect on the flexural strength within 90 days.
From the change in flexural strength, the leaching resistance of SMC was better than that of OPC.
is was related to the addition of silica fume. Active SiO 2 in silica fume could produce pozzolanic reaction and filling effect. Free Ca(OH) 2 from cement hydration reacted with active SiO 2 to form stable low alkalinity C-S-H gel. e C-S-H gel with low alkalinity was more compact and denser than that without silica fume, and its porosity was lower. Its strength was higher than that of Ca(OH) 2 crystal.
erefore, the addition of silica fume reduced the content of Ca(OH) 2 in the paste, and the macropore was replaced by the small pore, which reduced the most probable pore size and specific surface area, reduced the porosity, and improved the compactness of the structure. Macroscopically, the strength, impermeability, and frost resistance of cement with the addition of silica fume were improved [24].

Analysis of XRD-Rietveld.
Hydration of Portland cement mainly generates Ca(OH) 2 and C-S-H, as follows: However, in the water environment, Ca(OH) 2 in the hardened cement slurry will interact with CO 2 in the air to form CaCO 3 , as shown in the following equations: In addition, after the hydration reaction of cement, the relative concentration of hydration product reaches a certain value and keeps an equilibrium. Among them, Ca(OH) 2 has the highest limit concentration and is most easily dissolved by permeable water. Once the equilibrium is broken, the amount of Ca(OH) 2 will be adjusted, resulting in calcium ion precipitation. So, like the solution is continuously diluted, the hardened cement paste constantly dissolves out calcium hydroxide under the action of current, which can lead to the decalcification of hydrated calcium silicate and the decomposition of Ca/Si after lowering. e strength of hardened cement paste reduces, and its performance deteriorates.   Figure 7. Figure 4 shows that Ca(OH) 2 was the hydrated product of each sample at 28 d, and its diffraction peak was the highest. In addition, the diffraction characteristic peaks of CaCO 3 and SiO 2 can be seen, and the diffraction peaks of SiO 2 were weak. From Figure 4(a), it can be seen that the diffraction peak of Ca(OH) 2 and CaCO 3 on the surface of OPC increased with the increase in water velocity. From Figure 4(b), it can be seen that the Ca(OH) 2 diffraction peak of SMC decreased with the increase in water velocity, which indicates that less and less Ca ions can be leached. e diffraction peaks of CaCO 3 increased gradually, indicating that more and more stable CaCO 3 was formed by calcium ions.
As the curing age increased to 56 d, it can be seen from Figure 5 that the diffraction peaks of Ca(OH) 2 and CaCO 3 in OPC increased with the increase in flow rate. SMC specimens were the opposite. e diffraction peaks of SiO 2 in the samples were still weak, and there was little difference among the samples at different flow velocities.
When the age reached 90 days, the diffraction peaks of Ca(OH) 2 and CaCO 3 in the sample atlas decreased with the increase in flow rate, which indicates that the leaching of calcium ions tends to be stable and slow. e diffraction peak of SiO 2 was still weak.
According to the effect of flow rate on leaching in Figures 4-6, the effect of flow rate set in this experiment on erosion was not obvious. With the increase in age, the effect  Advances in Civil Engineering of flow velocity was greater, which indicates that, in the actual environment, the calcium leaching of concrete structure is also a slow process, and the longer the leaching time is, the more obvious the calcium leaching is. e diffraction peaks of Ca(OH) 2 and CaCO 3 in each sample increased first and then decreased gradually, indicating that the leaching of calcium ions was more before 56 days. e leaching of the specimens proceeded from the surface to the interior gradually. With the increase in the leaching depth, the diffusion and leaching rate of calcium ions slowed down. When the SMC sample was mixed with silica fume, the diffraction peak of SiO 2 had little change compared with that of OPC. e main component of silica fume was active SiO 2 , which had a pozzolanic reaction with cement hydrates, but it was consumed because of its small amount [26]. e decrease in Ca(OH) 2 diffraction peak in SMC should be attributed to the reaction with SiO 2 , and the conversion of hydration products to low alkaline hydration products was also the main reason for the leaching resistance of hardened cement paste with silica fume.
On the one hand, the increase in the diffraction peak of CaCO 3 indicates that Ca(OH) 2 dissolved into CaCO 3 gradually, while CaCO 3 was more stable and filled the pore; on the other hand, it would be beneficial to reduce the leaching of Ca ions. e leaching rate was related to the porosity of hardened cement paste. Combined with the change rule of flexural strength of specimens, this phenomenon was verified. e porosity of ordinary Portland cement paste was larger than that of superfine cement. So OPC was easier to erode than SMC.
XRD at different sampling depths for 56 d is shown in Figure 7. e diffraction peak values of Ca(OH) 2      Advances in Civil Engineering had little change, so the effect of flow rate on leaching was not significant. e intensity of phase diffraction peak of SMC varied little at different depths, which indicates that SMC has good leaching resistance. Figures 4∼7, it is not easy to quantify and compare the mass change laws of the main phases. Rietveld quantitative phase analysis is based on the relationship between the scaling factor and the reference intensity ratio to obtain the relationship between the relative content of the phase and the scaling factor [27,28]. Quantitative calculation of phase was carried out by the Rietveld refinement method with Jade software. Quantitative analysis of the phases of different age samples is shown in Figures 8 and 9. According to Figure 8(a), the relative mass fractions of Ca(OH) 2 and CaCO 3 in different depths of OPC specimens changed little with the increase in age, which indicates that the effect of flow rate set in the experiment on calcium leaching was not obvious. However, on the 90 th day, the larger the flow rate was, the smaller the amount of Ca(OH) 2 was, which indicates that more Ca ions leached or transformed into CaCO 3 , and the leaching phenomenon was more obvious. Figure 8(b) shows that, with the increase in age, the relative mass fraction of Ca(OH) 2 decreased gradually, while the relative mass fraction of CaCO 3 increased gradually, but the change was not obvious. is Advances in Civil Engineering 9

Rietveld Quantitative Analysis. From
indicates that the reaction between active SiO 2 and Ca(OH) 2 in silica fume was basically completed on the 28 th day. is is consistent with the conclusions of the literature [29,30]. With the increase in day, the leaching of calcium ions gradually slows down. is is related to CaCO 3 filling pore and reducing calcium ion diffusion. For 56 days of OPC, the relative mass fraction of Ca(OH) 2 and CaCO 3 was not affected by flow rate. With the increase in sampling depth, the relative mass of Ca(OH) 2 increased gradually, while the relative mass of CaCO 3 decreased gradually, as shown in Figure 9(a). is indicates that the calcium leaching of OPC was gradual from outside to inside. Figure 9(b) shows that the relative mass fractions of Ca(OH) 2 and CaCO 3 were not affected by the flow rate and sampling depth of SMC at 56 d. e reaction between active SiO 2 and Ca(OH) 2 in silica fume was basically completed before 56 days. SMC promoted the hydration of C 3 S due to the reduction of Ca 2+ and OHconcentration in the early stage of cement-water system, shortening the induction period. Fine silica fume particles also acted as crystallization nucleus of hydration products, thus accelerating the early hydration of cement and making the concrete with silica fume have early strength. is is consistent with the results of flexural strength of SMC mentioned above. In addition, due to the pozzolanic reaction, Ca(OH) 2 in cement paste was absorbed with a large amount, forming a low alkalinity C-S-H gel. e amount of Ca(OH) 2 decreased with the increase in silica fume and hydration age. From Figures 8 and 9, it can be seen that the relative content of SiO 2 in OPC and SMC was very small.
In fact, in running water environment, the calcium dissolution of hardened cement paste occurs at the same time with the hydration of the cement. After 28 days, the hydration rate of the cement slows down, and the dissolution is related to the ion concentration of the solution of the pore. As shown in Figure 3(a), the bending performance of hardened cement paste shows a decreasing trend at different ages and flow rates. From the perspective of XRD, the phase Ca(OH) 2 decreases and CaCO 3 increases, as shown in Figure 8, which is more obvious at different sampling depths, as shown in Figure 9(a). According to equations (2)∼(5), as the age increases, Ca(OH) 2 dissolves more and reacts with CO 2 to generate CaCO 3 . erefore, the leaching of hardened cement paste may be correlated with the mass gradient of Ca(OH) 2 .
In Figures 8 and 9, the mass fraction of the main phases determined by quantitative analysis is actually a relative quantity, and the mass fraction of Ca(OH) 2 , CaCO 3 , and SiO 2 are selected as variables. In order to facilitate the comparison with the change trend of mechanical indexes, the ratio of CaCO 3 to Ca(OH) 2 mass fraction was taken as the change quantity, and the change curve of this ratio was drawn, as shown in Figure 10. It can be seen from Figure 10(a) that the mass ratio of OPC samples does not change much in different ages and tends to increase, indicating that the relative content of CaCO 3 increases, the leaching of calcium is more obvious, and the greater the water velocity is, the greater its influence on Ca ions erosion is. For the SMC sample, the changing trend was similar to that of OPC, but the water velocity had little effect on the leaching effect. According to Figure 10(b), the leaching of the OPC sample took place from the outside in, while the leaching difference between the inside and outside of SMC sample was not significant, which shows that the leaching of hardened cement paste is related to the changes in the relative quantity of CaCO 3 and Ca(OH) 2 . e higher the relative content of CaCO 3 is, the denser the hardened cement paste is, and the better its erosion resistance is. e erosion resistance of hardened cement paste is related to the mass gradient of Ca(OH) 2 . e smaller the gradient is, the better the erosion resistance is.
Compared with mechanical properties test and XRD phase analysis, Rietveld phase quantitative calculation shows that the conclusions of three analytical methods are consistent.

Conclusions
Using tap water as the leaching medium, the leaching tests of three different hardened cement pastes with flow velocities of 0 mm/s, 2.8 mm/s, and 11.2 mm/s were carried out. e flexural strength, phase composition, and mass fraction of the paste after leaching were tested and analyzed. e following conclusions were drawn: (1) Within 90 days, the effect of flow rate on the leaching of hardened cement paste is not obvious. (2) Tap water acts as the erosion medium at test flow rate. Before 28 d, the leaching of OPC and SMC is not obvious, and the flexural strength increases gradually. After 28 d, the leaching of OPC is higher than that of SMC, and the flexural strength of OPC decreases, while the strength of SMC does not change much. (3) XRD-Rietveld quantitative method was used to analyze the leaching characteristics of hardened cement paste. It is found that the leaching of hardened cement paste is related to the increase in CaCO 3 content and the gradient of Ca(OH) 2 content. (4) By comparing the change in flexural strength and XRD diffraction intensity, the XRD-Rietveld method can be used to evaluate the leaching characteristics of hardened cement paste under flowing water.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest.