Early-age carbon dioxide (CO₂) curing of concrete is an innovative carbon utilization technology, enabling resource-oriented utilization of carbon dioxide that allows carbon neutrality. Thus, it is a promising technology for the concrete industry. Enhancing carbonation reactivity contributes to the promotion of CO₂ curing technology. Among all the influencing factors, previous studies showed that water loss during pretreatment could be one of the most important parameters affecting carbonation degree and mechanical properties of concrete.

In a previous work, water loss was primarily evaluated based on the overall water loss of specimens. Note that water loss varies significantly at different depths due to the complex pore structure and heterogeneity of cement-based materials. This causes depth-dependent carbonation reaction, forming gradients in both water content and carbonation degree. However, the existing studies neglect water loss gradient and its impact on carbonation. This is mainly due to a lack of direct visualization and continuous tracking of moisture distribution and carbonation products inside concrete by conventional testing methods, making it difficult to accurately characterize water loss gradients and carbonation gradients. A more precise and efficient water loss and carbonation characterization method is required.

This study was to apply X-ray imaging technology to non-destructively and quantitatively track the formation of water loss gradients and carbonation gradients in cement-based materials at different pre-drying durations.

This study cast cube specimens with the size of 40 mm for compressive strength tests and specimens with the sizes of 10 mm × 20 mm × 40 mm for X-ray imaging tests. The raw materials consisted of water and P.O.42.5 cement at a water-to-binder ratio of 0.4. X-ray imaging technology was employed to capture projection images of the 20 mm × 40 mm face of the specimens before and after pre-drying and accelerated carbonation.

Image processing for water loss distribution and carbonation distribution included field non-uniformity correction, normalization, image registration, grayscale difference calculation, water loss distribution characterization, attenuation coefficient variation calculation, carbonation-water signal decoupling and carbonation distribution characterization.

Furthermore, a phenolphthalein alcohol solution was used to validate carbonation depths determined by X-ray imaging. The mechanism of different water loss gradients on carbonation behavior was analyzed by thermogravimetric analysis (TGA). The porosity changes of carbonated specimens under different water loss gradients was characterized by low-field nuclear magnetic resonance, thereby elucidating the formation mechanisms of water loss gradients and carbonation gradients. The 3-d compressive strength of the specimens was determined using a universal testing machine.

Based on the X-ray projection images, water loss mapping and water loss-depth curves are obtained, revealing a decrease of water loss with increasing depth. Besides, the overall water loss increases with prolonging pre-drying duration. A substantial water loss difference appears between the surface and internal sample core, forming a distinct water loss gradient. The peak water loss and the water loss gradient both increase with extended drying.

In addition, 2D mapping of attenuation coefficient variation caused by carbonation and the corresponding variation along the depth can be acquired, introducing carbonation gradient and total attenuation coefficient change as indicators for carbonation behavior. Among all the samples, the sample CC18-0 exhibits the shallowest carbonation depth, with the maximum carbonation degree appearing at the surface. Carbonation depths of sample CC18-1.5/3/6 show a slight increase with drying time, while the decline rate within the partially carbonated zone remains consistent. Finally, the sample CC18-24 achieves the maximum carbonation depth of approximately 6.6 mm, while its peak carbonation degree is the lowest. Moreover, prolonged drying time leads to a significant decrease in the carbonation gradient, dropping progressively from 0.0088 mm^-1 for the sample CC18-0 to 0.0061 mm^-1 for the sample CC18-24 (a maximum reduction of 30.7%). In addition, the total attenuation coefficient change within the carbonated zone also increases with drying time (from 0.093 to 0.138).

A comparison of carbonation depths determined by X-ray imaging, thermogravimetric analysis, and the phenolphthalein test indicates that X-ray imaging is a more precise carbonation testing method. Furthermore, the porosity of carbonated specimens decreases as water loss during pre-treatment increases. The sample CC18-24 has the lowest porosity of approximately 25.29%. Eventually, the early-age strength of carbonated cement pastes slightly improve from hydrated samples.

The design of water loss gradients in cement-based materials prior to carbonation can be achieved via controlling the pre-drying duration. In general, short pre-drying period results in a gentle water loss gradient due to the larger pores and lower resistance to moisture migration. The high surface water content impedes CO₂ diffusion, causing the carbonation reaction to be dominated by the transport rate of CO₂. As a result, carbonation mainly occurs near the surface, leading to a high carbonation degree that is limited to a shallow depth. In contrast, a prolonged pre-drying promotes a higher hydration degree and pore refinement, which hinders outward moisture migration. This leads to a sharp increase in water loss at the surface, forming a steep water loss gradient. The low surface water content broadens the pathways for CO₂ diffusion, facilitating a deeper carbonation penetration. Meanwhile, the presence of continuous water films in the internal regions enhances the dissolution efficiency of calcium ions, resulting in a gradual carbonation gradient. The steep water loss gradient drives CO₂ to diffuse inward, thereby increasing the overall carbonation degree, which is evidenced by higher CO₂ uptake and calcium carbonate content. The increased calcium carbonate content further refines the pore structure, significantly reducing the total porosity of the matrix, which in turn enhances the compressive strength.

The samples subjected to shorter drying durations exhibited less and more uniform moisture losses from surface to interior. The relatively high residual moisture at the surface favored in-situ carbonation but limited CO₂ transport inwards, resulting in a higher surface carbonation degree with a shallower carbonation depth. Conversely, the samples dried longer experienced a greater moisture loss and an increased surface drying. The lower surface moisture facilitated CO₂ diffusion and slowed down calcium ion dissolution, leading to a relatively lower surface carbonation degree yet a greater carbonation depth. It was indicated that a steep moisture loss gradient promoted CO₂ diffusion, enhancing the overall carbonation degree of cement pastes. This significantly reduced total porosity, which consequently improved the compressive strength.

The transport process of moisture in concrete is a physical phenomenon and a medium for the transmission of many harmful substances, having a far-reaching impact on the durability and structural safety of concrete. Moisture flows through pores and cracks in concrete that can provide penetration channels for harmful substances such as chloride ions, sulfate ions and carbon dioxide. Especially in humid environments, moisture promotes the migration of these harmful substances and accelerates the chemical processes such as chloride corrosion caused by chloride ions, thus significantly reducing the durability of concrete. The quantitative study of moisture transport can reveal the migration patterns of moisture in concrete, and identify/quantify various durability problems caused by moisture, providing a theoretical support for the maintenance, repair and life prediction of concrete structures.

In this study, the quantitative principle of moisture transport in mortar was derived based on Beer's law theory, and a triangular prism moisture layer sample mold was designed and conceptualized. Water could be directly infused into the prism, and in the process of X-ray transmission, a superposition model of dry mortar and water layer was formed. The relationship between known water content and X-ray transmittance was obtained by this calibration model, and then the calibration was performed through the change of X-ray transmittance of mortar sample before and after water absorption.

According to the standard GB 175—2007, China, mortar samples with the dimensions of 40 mm×20 mm×60 mm (h×b×d) were prepared with water, sand, and P·O·42.5 ordinary cement. The samples were the samples with different water-cement ratios of 0.4, 0.5, and 0.6 (W/C) (at a cement-to-sand ratio of 1:1), the mortar samples with different cement-sand ratios of 1:1, 1:2, and 1:3 (C/S)(at a water-cement ratio of 0.5), and the mortar samples with different sand particle sizes of 0.5–1.5 mm, 1.5–3.0 mm, and 3–5 mm (at a water-cement ratio of 0.5 and cement-sand ratio of 1:1), totaling 8 groups. Prior to the capillary water absorption test, the calibration model for each sample was measured by X-ray projection, and the X-ray radiography error was corrected for field inhomogeneity. A water calibration curve was then obtained, showing a relationship between X-ray transmittance (–ln(I_w/I_w0)) and moisture thickness, ranging from 0 to the maximum thickness d. Subsequently, a capillary water absorption experiment was conducted. After the absorption experiment, X-ray radiography and weighing were performed to obtain the X-ray transmittance (–ln(I_w/I_w0)) at different water absorption stages, with field inhomogeneity correction and normalization applied. Finally, the mass after water absorption was calibrated via the calibration curve and verified by the weighing results.

The relationship curve between water thickness and X-ray transmittance in mortar samples is obtained, having the fitting results (R²>0.99). During water absorption, the bottom of the sample begins to absorb water, forming a distinct water peak, indicating a one-dimensional water absorption process. As water absorption progresses, the water front gradually moves upward. However, the water distribution during capillary water absorption varies significantly due to the differences in composition and porosity among the mortar samples. The height of the water absorption front increases with time, though the rate of increase gradually slows. The water absorption front disappears when capillary water absorption approaches a saturation after a prolonged absorption period. For the total water absorption of each mortar, the maximum relative error between the calibration and weighing results is 4.8%.

A quantitative method for cement mortar was proposed. This method could allow for the quantification of moisture content in samples with different water-cement ratios, cement-sand ratios, and sand particle sizes through a single calibration process. The method could reduce both cost and time via eliminating a need to prepare different calibration samples. The method could accurately track a moisture transmission in mortar in real-time, capture the distribution and evolution of moisture in various mortars, and enable local quantification of moisture content in different regions of the samples. The experimental results showed that the smaller the C/S and sand particle size was, the better the water absorption uniformity would be. The seepage velocity and height of water surface peak increased with the increase of W/C and C/S. This method could obtain a reliable calibration for different mortar samples. Compared to the weighing method, the maximum relative error of all the samples during the water absorption process by this method could be less than 4.8%.