Growth mechanism of carbonated tricalcium silicate (C3S) under the high concentration of CO2: A novel research for CCUS wells

Under the engineering background of the carbon dioxide capture and geological storage technology (CCUS) cementing project, an experiment on the generation of cement single-phase tricalcium silicate CO2 carbonization products were carried out. Combining the phase diffraction pattern data and the thermogravimetric experiment, a relative crystallinity algorithm is proposed, which combines the quantitative results of the carbonized products with the relative crystallinity (RCP) results of each component. The growth and development mechanism of tricalcium silicate carbide crystal products under high temperature, high pressure and high concentration CO2 environment is deduced. The experimental results show that under the conditions of early gas phase carbonization, the carbonization rate of C3S first increases and then decreases as the carbonized crystal product grows. Under the conditions of early liquid phase carbonization, the carbonization rate of C3S first decreases and then increases with the generation and fragmentation of the hydration barrier layer. It provides a research basis and a new perspective for the subsequent analysis of the changes in the microstructure of the cement paste in the carbonization process under the CCUS engineering background.


Introduction
The continuous consumption of fossil energy has led to a continuous increase in manmade CO 2 emissions. The International Energy Agency (IEA) predicts that by 2035, global CO 2 emissions will reach 35.4 billion tons (Chang et al., 2012). In the past 10 years, in order to effectively alleviate greenhouse gas emissions, carbon dioxide capture and geological storage technology has become the main way to reduce CO 2 emissions and the greenhouse effect, and has received more and more attention (Bachu, 2008;Gleick et al., 2010;Leung et al., 2014). Portland cement is usually used in CCUS cementing projects (Kutchko et al., 2008;Duguid, 2009;Brandão et al., 2017). Cementing involves injecting cement slurry into the annulus between the casing and the formation, or between successive casings, and waiting for it to solidify to form cement paste to shield and seal the formation fluid. However, the cement-based materials in the well need to withstand the harsh service conditions of high temperature, high pressure, and high CO 2 concentration. Under such harsh conditions, cementbased materials will be in a process of accelerating carbonation. The formation of carbonized products and the growth and dissolution of crystals will cause changes in the pore structure of the cement matrix, and ultimately lead to cement performance failure (Groves et al., 1991;Dong et al., 2019). This will extremely seriously affect the integrity and safety of the long-term seal of the carbon dioxide storage technology. Therefore, it is necessary to conduct in-depth research on the carbonization mechanism of cement in the strongly acidic downhole environment to accurately predict the carbonization failure process of cement and provide a basis for the research and development of acidresistant cement.
Portland cement mainly contains four single-phases: tricalcium silicate (C 3 S), dicalcium silicate (C 2 S), tricalcium aluminate (C 3 A), and tetracalcium aluminoferrite (C 4 AF) (Tadros et al., 1976;Sánchez Herrero et al., 2016;Cuesta et al., 2018). According to the American Petroleum Institute (API) standard, it can be known that C 3 S is the main mineral component (50%-80%) in common oil well cement (G-class/ H-class) (Sorrentino, 2008). In addition, the hydration and carbonization reaction of C 3 S determines the development of the microstructure and structural strength of the cement paste (Groves et al., 1991;Mei et al., 2018). Therefore, to accurately predict the carbonization failure process of cement paste, it can be start with the single-phase tricalcium silicate, which has the greatest impact on the strength development of Portland cement (Kjellsen and Justnes, 2004). In the current research on the carbide carbonization process of tricalcium silicate, through conventional qualitative analysis (transmission electron microscope combined energy spectrometer (TEM-EDS), X-ray diffraction (XRD) and 29Si magic angle rotating nuclear magnetic resonance ( 29 Si MAS) -NMR). It has been proved that the product of fully carbonated C 3 S is amorphous silicon dioxide (SiO 2 ) and calcite crystals and there will be no other polymorphic calcium carbonate (CaCO 3 ) (Li et al., 2018). In addition, a variety of technical methods (Rietveld refinement, thermogravimetric analysis, Fourier transform infrared spectroscopy, field emission scanning electron microscope and micro-Raman spectroscopy) have been used to study the growth process of calcite crystals produced after tricalcium silicate carbonization from the aspects of crystallite size, particle size, chemical bonds and thermodynamic stability (Wang et al., 2020;Klimavicius et al., 2021). Through these qualitative analyses, it can be known that the growth of calcite crystals formed by the carbonization of tricalcium silicate has a great influence on the strength of cement.
These existing studies focus on the qualitative analysis of the carbonization products of tricalcium silicate and the effect of the microstructure after complete carbonization on the changes in the macroscopic mechanical strength properties. However, the crystal growth of the carbonized product during the carbonization process of tricalcium silicate in a high-concentration CO 2 environment will seriously affect the development of its internal microstructure and macro-mechanical strength of cement. Due to insufficient research on the crystallization laws of single-phase cement crystal products after CO 2 carbonization, there is insufficient understanding of the changes in the microstructure of cement after CO 2 carbonization. There are few studies on the orientation of the cement crystal phase after CO 2 carbonization, or the crystallization and crystal transformation laws in the long-term carbonization process, and there is no quantitative or semi-quantitative method to describe the growth process of carbonized crystalline products. In this paper, according to the crystal change law of single-phase cement during CO 2 carbonization, the C 3 S hydration products and carbonization products were analyzed by X-ray diffraction (XRD) analysis, and the relative crystallinity was introduced according to the experimental results for semi-quantitative analysis. The changes of amorphous products during the hydration and carbonization of C 3 S were investigated by thermogravimetric (TG/DTG) analysis, and the relative crystallinity change law of C 3 S single-phase cement CO 2 carbonization crystal products was studied. It is expected to provide new ideas and research basis for the study of microstructure changes caused by crystal changes.

Experiment methods
The water-cement ratio provided by the API standard is 0.38-0.44. The effect of the water-cement ratio on cement particle hydration was mainly reflected in the cement particle spacing and in the final degree of hydration reaction, owing to the rapid hydration of the single phase, in order to ensure that the sample has a certain fluidity for moulding, a water-cement ratio of 0.6 was selected to prepare single-phase cement samples with different carbonization time intervals at 90°C (Mei et al., 2018).
In general oil and gas well conditions, the corrosive medium CO 2 usually exists in two states: one is in the form of a solute dissolved in formation water, and the other is in the form of supercritical gaseous CO 2 in a humid environment. To study the influence of these two different water-humidity environments on the carbonation process of the samples, the prepared samples were placed in a high-temperature and high-pressure carbonation reactor, as shown in Figure 1. The upper sample is in a gas-phase environment as a dry gas carbonization environment, and the lower sample is in a liquid phase environment as an acid solution environment. CO 2 gas with high pressure of 5.0 MPa (3.0 MPa for air and 8.0 MPa for total pressure) was injected to ensure saturation conditions, thereby simulating the underground environment (Mei et al., 2021).
During the experiment, load the prepared single-phase cement slurry into the kettle body, the assembled kettle body was immersed in a constant-temperature water-bath curing box to maintain the experimental temperature of the sample. The single-ore cement carbonization curing environment, temperature, and age are listed in Table 1. The selection of experimental temperature conditions is based on the formation depth-temperature relationship to simulate the temperature conditions under common well depth environments.

Characterization
To evaluate the carbonization products of the CO 2 carbonized samples in the early stage of the experiment and the microstructure changes after carbonization, the reaction of the samples was stopped by soaking them in isopropanol. The immersion time of the carbonized sample was defined as from the completion of the carbonization corrosion experiment to immediately before the inspection, and the following tests were carried out: 1) Grind all CO 2 carbonized samples (hand-grind with a grinder), and then use an X-ray diffractometer for phase analysis. The equipment model is DXJ-2000, and the place of production is China's Fangyuan Instrument Co., Ltd. When the equipment is performing phase analysis, the current is 20 mA and the voltage is 30 kV. The sample passes in 2θ at a rate of 0.04°/s in the range of 5°-70°. 2) A Thermogravimetric analysis (DTG) (TGA/SDTA851e, Mettler Toledo, Switzerland, TA Q20) was carried out to quantitatively investigate the carbonation products over a temperature range of 40°C-1,000°C at a heating rate of 10°C/ min. Additionally, the powder samples were tested using nitrogen gas as the working gas.

FIGURE 1
Photograph and schematic representation of HTHP carbonation kettle and sample loading.

Material property
The X-ray diffraction (XRD) peaks of these experiments with the standard powder diffraction pattern (PDF# 49-0442) are consistent with Figure 2. The particle size distribution of the powder was measured using a laser particle size analyser (Mastersizer 2000; Malvern Panalytical, United Kingdom), and its main physical properties are shown in Figure 3. The particle size was mainly concentrated in the range of 1-100 μm, the specific surface area was 1.09 m 2 /g, and the median size was 6.365 μm.

Phase analysis
For the XRD diffractograms shown in Figure 4, carbonation C 3 S samples (GA-1, 3, 7, 14, 28 and LA-1, 3, 7, 14, 28) with different curing times under liquid and gas phase carbonization conditions (CH: #44-1,481 and Calcite: #47-1743). It can be seen from the figure that calcite CaCO 3 (abbreviated as C C) is the main carbonization product of tricalcium silicate, and the hydrated product of tricalcium silicate, calcium hydroxide (abbreviated as CH), gradually decreases and disappears with the extension of the carbonization curing time.
The amorphous gel phase calcium silicate hydrate (abbreviated as C-S-H) is one of the products of cement hydration process, but the XRD test cannot effectively characterize the amorphous gel phase. Therefore, it is necessary to combine the decomposition temperature of the phase at different temperatures to characterize the phase, where 400°C-500°C is the decomposition interval of the hydration product CH, and 600°C-900°C is the decomposition interval of the CO 2 carbonization product (Gomez-Villalba et al., 2012). The decomposition temperature range and decomposition stage of single-phase C 3 S cement hydration and carbonization products are shown in Table 2.
The quality loss in the range of 120°C and 460°C are produced by the decomposition of C-S-H and CH, respectively. At the same time, the CH produced by C 3 S hydration is consumed in large quantities due to the carbonization reaction, and a large amount of CaCO 3 is produced. The quality loss in the range of 750°C is produced by C C decomposition (Wei et al., 2022). Through synergistic analysis with the thermogravimetric Figure 5 TG characterisation results, compared with the XRD calibration sample, it can be seen that the carbonization products of C 3 S are all calcite calcium carbonate.

Relative crystallinity analysis
The main product after full hydration reaction of tricalcium silicate is composed of gel phase C-S-H and CH phase with different crystallinity (Cuesta et al., 2018). These hydration products are carbonized under the conditions of high concentration of carbon dioxide. The carbonized crystal products produced by the carbonization reaction will have a  Frontiers in Materials frontiersin.org 05 significant impact on the structure of the tricalcium silicate single-phase cement slurry. Therefore, studying the dependence of crystalline hydration products on carbonized crystalline products can enable people to understand the properties of cement paste structure carbonization corrosion damage. After long-term carbonization, the hydration products of cement continue to decrease, and the carbonization products continue to increase. In order to explore the changes in the crystallization law of the product crystals during the CO 2 carbonization process, the relative crystallinity can be introduced to conduct a semi-quantitative analysis of the development process of the carbonized crystallization products, so as to compare the crystal phases in the C 3 S carbonization reaction products of cement single phase and determine the law.
By measuring the cumulative diffraction intensity of the crystalline phases, a semi-quantitative analysis of the crystallinity changes of the phases was carried out, and the law of changes in the crystal structure was obtained (Dai et al., 2017;Dai et al., 2018). The crystallization process of tricalcium silicate carbonized products is related to the carbonization environment temperature and carbonization age. With regard to the phase growth process of the crystalline product, there is no perfect crystal owing to the influence of the external environment. Therefore, the crystallinity mentioned in this article is all relative crystallinity (Capron et al., 1987;Kontoyannis and Vagenas, 2000;Lee et al., 2008;Jiang et al., 2018), where the relative crystallinity is related to the crystallinity of the reference phase taken. The crystallinity of the phase is represented by the main diffraction peak area, I, corresponding to the object in the diffraction test result. The relative crystallinity of a single crystalline phase relative to all crystalline phases in the sample (RCS) and relative crystallinity of the phase (RCP) of a single carbonized crystalline phase under certain curing conditions were used to understand the crystal change process of carbonization products in a single cement phase. The calculation process of the RCS and RCP of the crystalline phase is shown in Eqs 1-4.
I n-phase is the peak area of the n-phase diffraction peak corresponding to a certain phase; I SUM-phase is the sum of the integrated areas of the three strongest peaks in the diffraction peak of a certain phase, and I SUM is the sum of the integrated area of the diffraction peaks of the crystal phase in a certain sample. RCP = 100% indicates the maximum integral area of the diffraction peak of hydration or carbonization products during the carbonization age. This suggests that the crystal structure of the crystalline phase in the sample is relatively complete. Jade software (version 6.0) was used to subtract the background peaks of the amorphous phase in the test results, fit each peak of the main crystalline product to obtain the three characteristic peaks with the strongest phase intensity, and calculate the integral area. The RCS and RCP values of the crystalline phase were calculated using Eqs 1-4, and the specific sample analysis results are as follows.

Analysis of relative crystallinity change under gas phase carbonization
The phases of C 3 S after gas-phase carbonization at 90°C are shown in Figure 4A. The main hydration and carbonization phases were CH and C C ; and the intensity of the diffraction peaks changed continuously with the increase in carbonization age. For CH, which is the phase consumed in the carbonization process, its diffraction peaks disappeared after 14 days; the increasing peak of C C indicates that the amount of crystals and relative crystallinity also increased. The specific results are shown in Figure 6. The RCP of CH was highest after 1 day of reacting, then continuously decreased to only 5.34% at 14 days, disappearing at 28 days. The RCP for the corresponding carbonization product C C increased continuously from 23.89% at 1 day, reaching a maximum at 28 days. The RCS changes in C 3 S were consistent with the RCP results. The RCS ratio of C C continuously increased, reaching its highest value at 28 days, and the RCS of CH was almost reduced to zero at 14 days and was completely consumed. Clearly, the crystals formed by the hydration product CH continued to decrease in the sample, indicating that although the carbonization reaction did not completely consume the hydration product, the carbonization product formed crystals. The crystalline phase change in C 3 S is mainly due to the change from CH to C C. Since the main carbonization product C C is a high-strength phase, it also has greater brittleness (Marius-George et al., 2021;Poudyal et al., 2021). The formation of brittle phase will adversely affect the mechanical properties of cement.
Because both the CO 2 carbonization and hydration processes react with Ca 2+ to produce the corresponding products, the sum of the C C and CH products represents the degree of C 3 S hydration and C 3 S carbonization reaction. CO 2 carbonization product quantitative test DTG curve and content of each component of vapour phase carbonization samples are shown in Figure 7 (thermogravimetric analysis). As the carbonization age increased, the quantity of carbonization products continued to increase, to 59.19% at 28 days. The hydration products had been decreasing since the start of carbonization, and 4.03% remained at 28 days; at the same time, according to the image, under this carbonization condition, the carbonization rate was high at first and then decreased. In addition, when the carbonization age was 1 day, the sample already contained 33.93% carbonization products, indicating that the hydration process was affected by CO 2 carbonization. The quantitative  Figure 4B shows the XRD test results of the liquid-phase carbonized sample and the RCP and RCS results of the crystalline product. The phases contained in the C 3 S liquidphase carbonization sample at 90°C are mainly CH and C C. It can also be seen from Figure 4B that the diffraction peak of CH disappeared in 14 days, which indicates that the CH crystals in the sample at this time were lower than the minimum XRD detection amount. The intensity of the diffraction peak corresponding to the carbonization product C C increased with the carbonization age. From the change rule of RCP in Figure 8A, the RCP of CH continued to decrease from 1 to 7 days, and 76.92% remained at 7 days, but CH disappeared at 14 days. The corresponding product C C changed little at 1-3 days, then continued to increase, and the RCP reached a maximum at 28 days. As shown in Figure 8B, the RCS of the CH and C C crystals in the sample remained unchanged after 1-3 days. Subsequently, CH was greatly consumed by the corrosive medium CO 2 . At 7 days, the RCS of CH in the sample was only 21.51%, and the experiment after 14 days of carbonization did not contain the crystalline CH phase. The TG results also show that the amount of CH in the samples with a The DTG curve (A) and the content of each component in the quantitative tests of CO 2 carbonization products (B) of the gas phase carbonization sample (GA-1, 3, 7, 14, 28). Frontiers in Materials frontiersin.org carbonization age of 14 days is already very small and can be ignored.

Analysis of relative crystallinity change under liquid phase carbonization
Combining the results of thermogravimetric analysis to verify the carbonization law shown by the relative crystallinity is shown in Figure 9. The hydration product CH was at a maximum of 15.41% at 1 day, and then continued to decrease to only 3.69% at 28 days, so the decomposition peak of CH on the DTG curve disappeared. For C C , the increase was slower at 1-7 days, from 13.70% to 35.47%, and it increased rapidly at 14-28 days, which was much greater than the consumption of CH, and its content was 59.44% at 28 days. From Sum-Ca (Calcium-containing phases participating in the reaction), 39.60% at 14 days is far less than 63.13% at 28 days, so the C 3 S reaction at 14 days was not complete. This shows that because of carbonization, the hydration process of C 3 S is greatly affected, and the consumption rate of hydration products is slower than the rate of the carbonization reaction. The quantitative analysis results of the carbonization products are consistent with the above RCS analysis results of liquid carbonization.

Analysis of carbonization product crystal change process
The foregoing is the law of influence on the crystal products formed in the CO 2 carbonization reaction of C 3 S mono-ore cement under different corrosive environments, including CH generated by the hydration reaction. Among the carbonization products, C C was the main crystalline phase. Crystallisation and crystal changes have a greater impact on the internal structure. On the one hand, because the RCS of the crystalline phase is based on the relative crystallisation between the phases in the sample, the explanation of the change in the crystalline phase with time is weak; on the other hand, the RCP of the crystalline phase is based on the carbonization age. The phase with the highest crystallinity was used as a reference, and the relative crystallinity of the crystal phase changed with time. Therefore, the RCP of the main carbonization product, C C , is used to discuss its crystallisation law.
The main crystalline phase of C 3 S formed by CO 2 carbonization is C C , and no other crystal-type carbonization products are formed. As shown in Figure 10 in a gas-phase CO 2 carbonization environment, the initial RCP value of C C is low at 1-3 days. At 7-14 days, the carbonization product crystallisation speed increases, and then during carbonization at 14-28 days the

FIGURE 9
The DTG curve (A) and the content of each component in the quantitative test of CO 2 carbonization products (B) of the liquid phase carbonization sample 3,7,14,28).

FIGURE 10
RCP variation law of the carbonization crystallization product C C of GA-LA sample.
Frontiers in Materials frontiersin.org crystallisation rate of the product slows down. Combined with previous studies (Niel and Eurybiades, 1982;Palmer et al., 1988;Zuddas et al., 2003;Therese et al., 2011;Der, 2014), This is because in the early stage of gas phase carbonization, there is a mass space between the C 3 S particles and no binding phase is formed, and the moist CO 2 gas can penetrate into the sample unimpededly and the gas permeation rate is fast. In the later stage of gas-phase carbonization, with the formation of carbonized crystalline products, the void ratio of the C3S single ore sample decreases, which reduces the chance of contact between C 3 S and CO 2 .
In the liquid phase CO 2 carbonization environment, the RCP value of the carbonization products decreased from 1 to 3 days, then the liquid phase carbonization reaction speed continued to increase. Combined with the quantitative results, the hydration products at this stage increased, indicating that delayed hydration destroys the integrity of the carbonization product crystals, resulting in a decrease in the RCP. According to previous studies (Goodbrake et al., 1979), this is due to the faster hydration of C 3 S in the liquid phase environment, and more CH phase and C-S-H are formed at the beginning of the carbonization reaction. In the subsequent reaction, the presence of a large amount of free Ca 2+ during the hydration process is conducive to the reaction with CO 3 2− . It can still form a stable carbonization product even in an environment with a higher concentration of CO 2 , and there are only C C crystals in C 3 S. After the carbonization product crystals nucleate, they continue to grow as the carbonization medium continues to diffuse and react. When CH is carbonated, the volume will increase by about 11.23% (Mei et al., 2018), which will severely expand the cracks along with the carbonation reaction. The crack will continue to expand until the crystal breaks. This led to the acceleration of the later liquid phase carbonization.
For C-S-H, although the gel phase does not have a fixed molecular formula, many studies (Rostami et al., 2012;Morandeau et al., 2014;Shen et al., 2016) have shown that as C-S-H is corroded by CO 2 , when the amount of C-S-H is small, the volume of C-S-H will increase, and the volume effect of C-S-H accounts for 70% of the total volume change. When the amount of C-S-H is large, the volume of C-S-H decreases after carbonization, resulting in sample shrinkage. Combined with the previous quantitative analysis of the reaction products, the amount of C-S-H is relatively low; when it is corroded by CO 2 , it is mainly due to volume expansion, and a hydration barrier layer (the reduction of porosity and the dense growth of CHS under the early liquid phase environment carbonization) is formed on the surface of the cement paste to slow down the carbonization process. Numerous studies (Fabbri et al., 2012;Cheshire et al., 2017;Jimoh et al., 2018;Shah et al., 2018) have shown that the change in pore structure affects the penetration process of the CO 2 corrosive medium to a certain extent, thus delaying the carbonization process. Therefore, the RCP value of vapour-phase carbonization at 14 days is much larger than that of liquid-phase carbonization. In summary, the entire CO 2 cement single-ore carbonization process (Mei et al., 2018) is shown in the model in Figure 11.

Conclusion
In this study, XRD and relative crystallinity algorithms were used to analyse and study the relative crystallinity change law of C 3 S single-phase cement during CO 2 carbonization. The main conclusions are as follows.

FIGURE 11
Schematic diagram of CO 2 carbonization process of C 3 S single-phase cement.
Frontiers in Materials frontiersin.org 1) Compared with gas phase carbonization, the formation rate of liquid phase carbonization crystallization products is slower. The increase of carbonization time will accelerate the formation of crystallization products. For subsequent cement failure, the liquid phase environment is more dangerous. 2) Under the assumption that all carbonization products are crystals, the quantitative results of the carbonization products and the relative crystallinity (RCS) results of the components are combined to deduce the crystal change process of the carbonization products during the CO 2 carbonization process. Supported by previous research results, this semianalytical method is feasible and more intuitive. In the CCUS well condition, the carbonization products of cement will dissolve in the later stage, and the development trend of the later carbonization products can be predicted by the calculation results of the relative crystallinity.
3) The method of calculating relative crystallinity combined with thermogravimetric analysis can realize semi-quantitative analysis of the pre-development process of carbonized product crystals in the process of tricalcium silicate carbonization. The effect of this analysis method is intuitive, it provides a good analysis method to explore the evolution process of the crystal growth of carbonized corrosion products.
The above results provide a research basis for analysing the microstructural changes in cement paste caused by CO 2 carbonization and provides a new analysis perspective of XRD combined with relative crystallinity algorithm for carbonization research.

Data availability statement
The raw data supporting the conclusion of this article will be made available by the authors, without undue reservation.

Funding
This study was financially supported by 111 Project, No. D18016. The Sichuan Science and Technology Program (2021YFQ0045 and 2021YFSY0056) and the Cnooc's preliminary project: Feasibility study of Wenchang 9-7 oilfield (2021PFS-05).