Abstract
This study compares the mechanisms of organic (Hydroxypropyl Methyl Cellulose, HPMC) and inorganic (bentonite) thickeners in regulating the bleeding behavior and surface homogeneity of cement pastes. In situ low-field nuclear magnetic resonance (LF-NMR) was employed to monitor water migration, while X-ray diffraction (XRD), scanning electron microscopy (SEM), and carbonation tests were conducted to evaluate the property disparities between the top surface and bottom layers. Results indicate fundamentally different working modes: HPMC reduces bleeding by swelling to block capillary channels, exhibiting a saturation threshold at 0.2% dosage. Beyond this point, as the primary transport channels are effectively sealed, additional HPMC merely densifies the polymer “plugs” without further suppressing the bleeding rate. XRD and SEM analyses reveal that despite the reduction in total bleeding, HPMC-modified pastes still exhibit significant stratification; the top layer retains a loose, granular morphology with higher carbonation susceptibility compared to the dense bottom layer. In contrast, bentonite mitigates bleeding through a volume-filling mechanism and thixotropic structuring, demonstrating a continuous, dosage-dependent efficacy up to 1.2%. At a 0.6% dosage, bentonite effectively eliminates microstructural disparities, yielding a top surface with a dense matrix and hydration product distribution nearly identical to the bottom layer. These findings demonstrate that the specific inorganic thickener (bentonite) utilized in this work is more effective in restoring surface homogeneity and enhancing carbonation resistance than the evaluated organic polymer (HPMC).
1. Introduction
Cement-based materials, including cement paste, mortar, and concrete, are fundamentally thermodynamically unstable suspensions immediately after mixing [1]. Driven by the density difference between solid particles and the mixing water, the solid phases tend to settle under gravity, forcing the lighter free water to migrate upward. This phenomenon, known as bleeding, is a complex two-phase flow process governed by sedimentation and consolidation theories [2,3]. While minor bleeding is sometimes deemed beneficial for preventing plastic shrinkage cracking, excessive bleeding poses severe threats to the engineering quality of concrete structures. The upward-migrating water creates continuous capillary channels and accumulates at the surface, forming a porous layer with an extremely high water-to-cement ratio (w/c), commonly referred to as the laitance layer or top surface layer [4,5]. This layer is structurally weak and exhibits significantly higher permeability compared to the bulk paste. Since the surface serves as the first line of defense against environmental attacks, a porous top layer accelerates the diffusion of aggressive agents such as carbon dioxide (CO2) and chloride ions. Specifically, the rapid carbonation of the laitance layer leads to a drop in pH, surface dusting, and, eventually, the corrosion of reinforcing steel, thereby compromising the long-term service life of the infrastructure [3,6]. Therefore, controlling bleeding is not merely a rheological requirement but a critical durability imperative.
To mitigate bleeding and enhance the stability of fresh cementitious mixtures, the incorporation of thickeners (or Viscosity Modifying Agents, VMAs) has become a standard practice in modern concrete technology [7,8,9]. Thickeners function by modifying the rheology of the pore solution or the flocculation state of solid particles to increase the yield stress and plastic viscosity of the paste, thereby resisting particle sedimentation. Based on their chemical nature, thickeners are generally classified into organic polymers and inorganic minerals, each operating under distinct mechanisms [10].
Organic thickeners, with cellulose ethers (e.g., Hydroxypropyl Methyl Cellulose, HPMC) being the most prominent representatives, are widely used due to their high efficiency [11,12,13]. HPMC molecules contain hydrophilic groups (hydroxyl and ether groups) that form hydrogen bonds with water molecules, significantly increasing the viscosity of the liquid phase. Furthermore, the long polymer chains adsorb onto cement particles, creating a steric hindrance effect and forming an intertwined three-dimensional network that entraps water and solid particles. Despite their effectiveness, organic thickeners often come with drawbacks, such as high cost, potential retardation of cement hydration, and air-entrainment issues, which may negatively affect the mechanical strength if not carefully controlled [11].
In contrast, inorganic thickeners, such as bentonite (a montmorillonite-based clay), offer a cost-effective and environmentally friendly alternative [14,15,16,17,18,19]. Bentonite particles are characterized by a layered silicate structure with a high specific surface area. Upon hydration, water molecules penetrate between the clay platelets, causing them to swell and disperse. The mechanism of bentonite relies primarily on electrostatic interactions; the negatively charged faces and positively charged edges of the platelets interact to form a thixotropic “house-of-cards” structure [10]. This structural network provides a volume-filling effect that entraps free water and resists sedimentation, which is the primary mechanism under consideration for enhancing the stability of the investigated cement pastes. Unlike organic polymers, inorganic thickeners usually exhibit better compatibility with the cement hydration process, although their dosage requirements are typically higher to achieve comparable viscosity enhancement [20].
The influence of thickeners on the stability of cementitious systems has been a focal point of recent research [21,22,23,24,25,26,27]. For instance, Palacios et al. [21] demonstrated that cellulose ethers like HPMC significantly enhance water retention by modifying the rheological behavior of the pore solution. Similarly, Brumaud et al. [23] reported that the incorporation of HPMC improves the stability and microstructure of cemented backfills. In contrast, the use of inorganic modifiers has also gained traction. Sanchez et al. [25] identified that bentonite clay alters fresh mortar rheology primarily through particle-to-particle interactions, while Kawashima et al. [26] emphasized the rapid thixotropic rebuilding of cement pastes modified with nano-clays. While these studies individually highlight the benefits of organic or inorganic VMAs, comparative investigations—particularly regarding their impact on vertical water migration and the resulting structural gradients—remain limited. Most existing literature focuses on bulk rheological parameters such as yield stress and plastic viscosity, rather than the dynamic evolution of the bleeding process. Current industry standards typically characterize bleeding using macroscopic volumetric methods (e.g., measuring the volume of bleed water in a graduated cylinder). These methods provide a single value representing the total bleed water but fail to capture the dynamic migration process of water within the paste or the distribution gradient of water along the sample height. Crucially, existing studies often model or analyze the hardened specimen as a homogeneous bulk, neglecting the critical microstructural disparity between the “top surface layer” and the “bottom bulk layer” [28,29,30]. It is this structural stratification—specifically the quality of the laitance layer—that dictates the surface scaling resistance and carbonation rate in practical applications. The lack of a comparative understanding of how organic versus inorganic thickeners differentially shape the microstructure of this critical surface layer limits our ability to design concrete with superior surface durability.
To address these deficiencies, this study aims to establish a comprehensive “plastic-to-hardened” framework for understanding the role of thickeners. A systematic comparison was conducted between the performance of HPMC and bentonite in regulating the bleeding kinetics and microstructural evolution of cement pastes. Unlike existing studies that primarily focus on bulk rheology or macroscopic bleeding volume, this research investigates the microstructural stratification induced by bleeding, specifically elucidating the physicochemical disparities between the top surface layer and the underlying bulk matrix. A micro-kinetic perspective on vertical water migration is established through in situ low-field nuclear magnetic resonance (LF-NMR), offering insights into transport mechanisms that remain inaccessible to traditional macroscopic methods. This approach facilitates a clear differentiation between the ‘capillary-blocking’ mechanism of organic polymers (HPMC) and the ‘volume-filling’ mechanism of inorganic minerals (bentonite), particularly regarding their respective roles in restoring surface homogeneity. Furthermore, this study establishes a comprehensive framework to link these dynamic water states to the final mineralogical and morphological disparities between the top surface and bottom layers. This approach allows for a first-time disclosure of how organic and inorganic thickeners differentially restore the longitudinal homogeneity and surface durability of cementitious materials.
2. Materials and Methods
2.1. Raw Materials
P.O 42.5 Portland cement complying with the GB 175–2023 [31] standard was used in this study. The inorganic thickener, bentonite, was purchased from Changxing Leixiang Chemical Co., Ltd. (Huzhou, China) The chemical compositions of both the cement and bentonite were determined via X-ray fluorescence (XRF) using a wavelength-dispersive spectrometer (PANalytical Axios, Almelo, The Netherlands). Prior to testing, the samples were dried at 105 °C to a constant weight, and then prepared using the fused bead method, where the sample powder was mixed with a lithium tetraborate flux at a 1:10 ratio and fused in a platinum crucible at 1100 °C. The particle size distributions were measured using a laser particle size analyzer (Malvern Mastersizer 3000, Malvern Instruments, Malvern, UK) with ethanol as the dispersant. The chemical compositions and particle size distributions of both the cement and bentonite were presented in Table 1 and Figure 1, respectively. The hydroxypropyl methylcellulose (HPMC) utilized in this study was a commercial-grade cellulose ether (Macklin, Shanghai, China) characterized by a methoxy content of 19.0–24.0%, a hydroxypropyl content of 4.0–12.0%, and a nominal viscosity of 80,000 mPa·s (measured in a 2% aqueous solution at 20 °C).
Table 1.
Chemical compositions of cement and bentonite (wt.%).
Figure 1.
Particle size distribution of cement and bentonite.
2.2. Preparation of Cement Paste
Cement pastes were prepared using a standard NJ-160 planetary mixer (Wuxi Jianyi Instrument & Machinery Co., Ltd., Wuxi, China) with a water-to-cement ratio (w/c) of 0.5, following the guidelines of GB/T 8077-2012 [32]. Initially, the HPMC and bentonite powders were placed in the mixing pot. Upon the addition of cement and mixing water, the mixture was stirred at a low speed (62 rpm) for 2 min. After a 15 s interval, mixing resumed at a high speed (125 rpm) for another 2 min. The fresh cement paste was subjected to immediate testing. The specific dosages of HPMC and bentonite, calculated by the weight of cement (bwoc), are detailed in Table 2.
Table 2.
The dosage of HPMC and bentonite.
2.3. Test Methods
2.3.1. Low-Field Nuclear Magnetic Resonance (LF-NMR)
Low-field nuclear magnetic resonance (LF-NMR) measurements were conducted using a PQ-001 analyzer (Shanghai Niumag Analytical Instrument Corporation, Shanghai, China). The instrument operates with a permanent magnet strength of 0.49 T and a proton resonance frequency of 21 MHz. The magnet temperature was maintained constantly at 32.00 °C. For the measurement, the fresh cement paste was kept at a room temperature of 25 °C. At specific time intervals, the sample tube (Φ 25 mm, containing 10 g of cement paste) was placed into the coil of the NMR analyzer; after a 3 min thermal equilibration period, the transverse relaxation time (T2) was measured using the Carr-Purcell-Meiboom-Gill (CPMG) [33] pulse sequence. The acquired relaxation decay curves, which represent a superposition of exponential components, were inverted using the Simultaneous Iterative Reconstruction Technique (SIRT) [34] algorithm to obtain the T2 distribution curves.
According to the classification by Ji et al. [35], water molecules in fresh cement paste can be categorized into three states: pore water, water in diluted grouts, and bleeding water. The transverse relaxation time (T2) of these water populations exhibits a progressive increase as the physical constraint decreases from the pore structure to the surface. In this study, to accurately quantify the bleeding process, the relaxation signals corresponding to both “water in diluted grouts” and “bleeding water” were integrated and collectively designated as the total bleeding water for the calculation of the bleeding rate (. As previously defined, the signal area of the main peak () represents the pore water, while the integrated area of the secondary peaks (), including water in diluted grouts and supernatant water, represents the total bleeding water. The bleeding rate is calculated using Equation (1):
where and are the integrated areas of peak P1 and peak P2, respectively, obtained via the SIRT inversion algorithm. This method allows for a non-destructive and quantitative assessment of the bleeding process at specific time intervals.
2.3.2. Accelerated Carbonation Test
The accelerated carbonation test was conducted in accordance with the Chinese standard GB/T 50082-2009 [36] (Standard for test methods of long-term performance and durability of ordinary concrete). A concrete carbonation test chamber (Model CCB-70F, Beijing Shuzhi Yilong Instrument Co., Ltd., Beijing, China) was employed. Cubic specimens with dimensions of 100 mm × 100 mm × 100 mm were used.
The pre-conditioned specimens were placed on the racks within the chamber, maintaining a clear spacing of at least 50 mm between samples to ensure air circulation. The gas convection system was activated, and CO2 was gradually introduced. The environmental conditions were strictly maintained at a CO2 concentration of (20 ± 2)%, a temperature of (20 ± 2) °C, and a relative humidity of (70 ± 2)%.
At carbonation ages of 3, 7, 14, and 28 d, the specimens were retrieved and split open using a compression testing machine (Wuxi Jianyi Instrument & Machinery Co., Ltd., Wuxi, China) (via the splitting tensile method) to obtain a fresh cross-section perpendicular to the casting surface. Residual powder on the fractured surface was removed with a brush, followed by the immediate application of a 5% phenolphthalein alcohol solution. After approximately 30 s, the carbonation depth (the uncolored region) was measured using a vernier caliper with a precision of 0.02 mm at intervals of 10 mm along the edge. The average carbonation depth for each age was calculated using Equation (2), and the relationship between carbonation depth and time was subsequently plotted.
where
is the average carbonation depth (mm) after t d of carbonation, precise to 0.02 mm;
is the carbonation depth at each measurement point;
n is the total number of measurement points.
2.3.3. X-Ray Diffraction (XRD) Analysis
To establish a link between surface durability and microstructure, samples for XRD and SEM were extracted from the top surface layer of the 28 d cured 100 mm cubic specimens. XRD analysis was conducted to determine the phase compositions and their relative contents in the specimens. To investigate the mineralogical variations between the top surface layer (bleeding layer) and the bottom layer (molded surface), sampling was performed after curing to predetermined ages. Materials within a depth of 1 mm from both the top and bottom surfaces were carefully scraped off using a spatula. The collected samples were immediately immersed in ethanol and sealed to stop hydration. Prior to testing, the samples were ground into fine powders and oven-dried at 40 °C for 24 h. The XRD patterns were acquired in continuous scanning mode with a 2θ scanning range of 5–80° and a scanning speed of 5°/min.
To quantify the phase compositions, Rietveld refinement was performed on the diffraction data using HighScore Plus software (Version 3.0). The scale factors, lattice parameters, and peak profiles were refined to obtain the weight fractions of the crystalline phases, with a specific focus on Portlandite (CH) and Calcium Carbonate (CaCO3). Furthermore, to assess the carbonation resistance of the bleeding surface, the carbonation ratio ()—representing the conversion degree of Portlandite to Calcium Carbonate—was calculated based on the quantitative results of the CH and CaCO3 phases. The carbonation ratio was determined using the following equation (Equation (3)):
where and are the weight fractions of Calcium Carbonate and Portlandite obtained from Rietveld refinement, and and are their respective molar masses.
2.3.4. Scanning Electron Microscopy (SEM) Test
The SEM was further employed to investigate the microstructural variations between the top surface layer (bleeding surface) and the bottom layer (molded surface) of cement pastes modified with different types and dosages of thickeners. Upon reaching the designated curing ages, small rectangular specimens with dimensions of approximately 10 mm × 10 mm × 5 mm were sectioned from both the top and bottom surfaces of the paste samples using a precision diamond saw (SYJ-150, Shenyang Kejing Auto-instrument Co., Ltd., Shenyang, China). To preserve the microstructure, the cut samples were immediately immersed in anhydrous ethanol to arrest hydration, followed by drying in an oven at 50 °C. Prior to imaging, the sample surfaces were sputter-coated with a thin layer of gold to enhance conductivity and prevent charging. The morphological observations were conducted using a field-emission scanning electron microscope (ZEISS Gemini 300, Oberkochen, Germany) operating at an accelerating voltage of 3–15 kV.
3. Result and Discussion
3.1. Evolution of Water Distribution via LF-NMR Analysis
Figure 2 illustrates the transverse relaxation time (T2) distribution curves of cement pastes incorporating different dosages of HPMC at rest times ranging from 0 to 180 min. Generally, the water in the fresh paste can be categorized into two distinct populations based on the relaxation time: the main peak located at approximately 1–10 ms (P1), representing the physically bound water within the flocculation structures or immature pores; and the secondary peak distributed beyond 100 ms (P2), corresponding to the semi-free water in diluted grouts or the supernatant bleeding water [35,37]. The evolution of these peaks provides dynamic insights into the water migration and stability of the paste.
Figure 2.
Evolution of T2 distribution curves of cement pastes with different HPMC dosages at (a) 0, (b) 10, (c) 30, (d) 60, (e) 90, and (f) 180 min.
For the Control sample, a typical solid–liquid separation process was observed. As the standing time extended from 0 to 180 min, the amplitude of the P1 peak exhibited a continuous decline, accompanied by a slight shift towards shorter relaxation times. Concurrently, the P2 peak emerged as early as 10 min and grew significantly in area over time. Notably, the center of the P2 peak shifted towards approximately 1000 ms at 180 min, a value close to that of bulk water. This indicates that without thickeners, the cement particles settled rapidly under gravity, squeezing the pore water upwards to form a layer of pure bleeding water on the surface.
The sample with 0.05% HPMC (H-5) displayed a unique “dispersion-then-segregation” behavior. In the initial stage (0–30 min), the P1 peak of H-5 shifted to a longer relaxation time compared to the Control. This suggests that the trace amount of HPMC acted as a dispersant, breaking the initial flocculation of cement particles and releasing the entrapped water, thereby increasing the degree of freedom of the pore water. However, due to the insufficient polymer concentration to form a supporting network, this dispersion effect accelerated the subsequent particle settling. Consequently, at 180 min, the P1 peak shifted drastically to the left (indicating compaction), and a prominent P2 peak appeared. Interestingly, the T2 value of this bleeding peak (~200 ms) was significantly shorter than that of the Control. This implies that the supernatant liquid in the H-5 group was not pure water but a viscous solution containing dissolved HPMC, indicating a segregation phenomenon rather than simple bleeding.
When the dosage reached 0.2% (H-20) and above, the bleeding behavior changed fundamentally. Although P2 peaks were still observable at 180 min (Figure 2d–f), their signal areas were significantly reduced compared to the Control. Critically, the integrated areas of the P2 peaks for H-20, H-35, and H-50 were comparable at 180 min, despite differences in their peak positions. This indicates that the bleeding rate reduction reached a saturation plateau at a dosage of 0.2%.
This threshold phenomenon can be attributed to the “pore-blocking” mechanism of HPMC [38]. Upon hydration, HPMC molecules absorb water and swell significantly, occupying the interstitial spaces between cement particles. These swollen polymer domains act as plugs that effectively clog the capillary channels responsible for water upward migration. At a dosage of 0.2%, the quantity of swollen HPMC is sufficient to block the majority of these bleeding channels. Consequently, further increasing the dosage (to 0.35% or 0.5%) may densify the plugs or increase the viscosity of the pore solution (altering the P2 position), but it contributes little to further reducing the volume of bleeding water, as the transport paths are already effectively sealed.
Figure 3 presents the time-dependent bleeding rates of cement pastes, calculated based on the quantitative integration of LF-NMR signals. The Control sample exhibited a continuous accumulation of bleeding water throughout the testing period, reaching a maximum rate of approximately 33% at 180 min. This reflects a progressive sedimentation process where solid particles settled under gravity, displacing pore water to the surface. In contrast, the H-5 group (0.05%) displayed a distinctive “delayed segregation” behavior. During the initial 0–30 min, the bleeding rate was negligible (<2%), corroborating the dispersion effect observed in the T2 spectra. However, a sharp surge occurred after 30 min, with the final rate rising to ~29%, approaching that of the Control. This confirms that trace HPMC temporarily suspends particles but fails to form a stable network to sustain long-term stability.
Figure 3.
Time-dependent bleeding rates of cement pastes with different HPMC dosages.
For dosages of 0.2% (H-20) and above, the bleeding rates were significantly reduced, stabilizing at approximately 24–25%. Crucially, the curves for H-20, H-35, and H-50 were closely clustered, showing minimal variance despite the increase in dosage. At a threshold of 0.2%, swollen HPMC domains effectively clog the capillary transport channels. Consequently, further increasing the dosage densifies the plugs but does not further reduce the bleeding volume, indicating that 0.2% is the critical saturation dosage for this system. It should be noted, however, that while a dosage of 0.2% HPMC effectively inhibited bleeding, the precise critical saturation threshold may reside between 0.05% and 0.2%. A more granular investigation of this specific dosage range was beyond the scope of the present study, which focused on establishing the fundamental mechanistic differences between HPMC and bentonite. Future research dedicated to precise industrial mix design could benefit from a narrower dosage gradient to identify the exact economic and technical optimum.
Figure 4 presents the T2 distribution curves of cement pastes incorporating bentonite dosages ranging from 0.3% to 1.2%. Consistent with the observations in the HPMC group, the water population is clearly divided into pore water (P1) and bleeding water (P2). However, the evolution of the P2 peak reveals a fundamentally different regulatory mechanism. In contrast to the “saturation threshold” observed in the HPMC group at 0.2%, the bentonite group demonstrated a continuous, dosage-dependent suppression of bleeding water. At a dosage of 0.3% (B-30), the bleeding signal was already significantly reduced compared to the Control. As the dosage progressively increased to 1.2% (B-120), the P2 peak diminished monotonically until it nearly vanished. This linear trend suggests that the water-retention capacity scales directly with the volume of bentonite added, functioning via a volume-filling mechanism where the swelling clay particles and their thixotropic “house-of-cards” structure physically accommodate the excess water.
Figure 4.
Evolution of T2 distribution curves of cement pastes with different bentonite dosages at (a) 0, (b) 10, (c) 30, (d) 60, (e) 90, and (f) 180 min.
Furthermore, a critical distinction lies in the position (T2 relaxation time) of the bleeding peak (P2), which reflects the physical state and viscosity of the water in the laitance layer. As previously analyzed, the P2 peak for the HPMC group shifted to a shorter relaxation time (~200 ms) compared to the Control, indicating that the bleeding fluid is a viscous, polymer-rich solution. Conversely, the P2 peaks for the bentonite group were centered at longer relaxation times (>1000 ms), even higher than that of the Control group. This suggests that the residual bleeding water in bentonite samples behaves as bulk free water with minimal physical restriction. This implies that bentonite primarily immobilizes water physically without significantly altering the rheology of the escaped supernatant water.
This contrast in signal characteristics highlights a divergence in how surface durability is enhanced by the two thickeners. HPMC improves the surface properties by modifying the rheology of the bleeding fluid, effectively “densifying” the laitance layer (evidenced by the lower T2), thereby creating a more cohesive surface structure despite some segregation. On the other hand, bentonite improves surface quality primarily by “eliminating” the volume of the bleeding layer. The clay network effectively traps solid particles and water; while any escaped water remains pure (high T2), the total volume of this water is drastically minimized or eliminated at high dosages, thus fundamentally preventing the formation of a porous laitance layer.
Figure 5 quantitatively illustrates the time-dependent bleeding rates of cement pastes incorporating bentonite, calculated from the integrated LF-NMR signals. The evolution of bleeding rates in the bentonite group exhibits a distinct behavior compared to the HPMC group, characterized by a monotonic dosage-dependent reduction and superior absolute efficacy.
Figure 5.
Time-dependent bleeding rates of cement pastes with different bentonite dosages.
Specifically, the Control group reached a maximum bleeding rate of approximately 33% at 180 min. Upon the addition of 0.3% bentonite (B-30), the final bleeding rate dropped sharply to ~18%. As the dosage increased from 0.3% to 1.2% (B-60, B-90, and B-120), the bleeding rate decreased linearly, eventually reaching a minimum of ~6.5% for the B-120 group. Unlike the HPMC group, where the bleeding curves for dosages above 0.2% were clustered and indistinguishable, the bentonite curves remained clearly separated, demonstrating that the water-retention capacity of the paste continued to improve with higher clay content without reaching a saturation plateau.
A critical comparison between the two thickeners reveals that bentonite is more effective in reducing the total volume of bleeding water in this system. As previously analyzed, the HPMC group exhibited a “saturation threshold” at 0.2%, where the bleeding rate stabilized at approximately 24–25% and could not be further reduced by increasing the dosage. In contrast, even the lowest dosage of bentonite (0.3%, B-30) achieved a bleeding rate of ~18%, which is significantly lower than the “best-case scenario” of the HPMC group.
This disparity reinforces the fundamental mechanistic difference: Since HPMC functions by swelling to clog existing capillary channels, once these primary transport paths are effectively sealed, additional polymer dosage yields diminishing returns in bleeding reduction. In contrast, bentonite operates as a ‘water-holding sink’ through its volume-filling and thixotropic ‘house-of-cards’ structure. Its efficacy is dictated by its total absorption capacity, which scales linearly with its volume within the tested range (up to 1.2%). Any theoretical saturation for bentonite would likely only occur when the volume of clay platelets is sufficient to immobilize the entire volume of free mixing water, a state that would likely render the paste unworkable due to excessive yield stress. Therefore, the absence of a plateau in the bentonite series highlights its broader adjustable range and superior capacity for total water entrapment compared to the pore-blocking mechanism of HPMC.
3.2. Carbonation Resistance of the Top Surface Layer
Figure 6 presents the carbonation depths of the top surface layer (bleeding surface) for cement pastes modified with HPMC (Figure 6a) and bentonite (Figure 6b) over a curing period of 3 to 28 d. The results clearly demonstrate that mitigating bleeding is a prerequisite for enhancing the surface durability of cementitious materials.
Figure 6.
Average carbonation depth of the top surface layer (bleeding surface) of cement pastes modified with (a) HPMC and (b) bentonite at different ages.
The Control group exhibited the poorest resistance to CO2 ingress, with the carbonation depth reaching approximately 12.0 mm at 28 d. This high vulnerability is directly attributed to the severe bleeding observed in the previous sections. The upward migration of water resulted in a locally elevated water-to-cement ratio (w/c) at the surface. Upon evaporation, this excess water left behind a network of interconnected capillaries, providing rapid diffusion channels for CO2 to penetrate and neutralize the matrix.
For the HPMC-modified series (Figure 6a), the evolution of carbonation resistance mirrored the trend observed in the bleeding rate analysis. The low-dosage group (H-5) showed only marginal improvement (reducing the 28 d depth to ~10 mm), consistent with its inability to prevent segregation. Conversely, increasing the dosage to 0.2% (H-20) and above resulted in a significant reduction in carbonation depth, stabilizing at approximately 8.0–8.5 mm at 28 d. Crucially, no significant difference was observed between H-35 and H-50, exhibiting a saturation plateau. This aligns perfectly with the “pore-blocking” mechanism: once the HPMC dosage reached the threshold (0.2%) to block water migration, the bleeding rate stabilized. Consequently, the porosity of the top surface layer was fixed at a certain level, limiting any further enhancement in carbonation resistance despite higher polymer dosages.
In the bentonite series (Figure 6b), the carbonation depth decreased monotonically with increasing dosage. Even at a low dosage of 0.3% (B-30), the carbonation resistance was notably improved compared to the Control. At higher dosages (B-90 and B-120), the 28d carbonation depth was reduced to ~8.0 mm, comparable to the optimized HPMC groups. This continuous improvement validates the “volume-filling” mechanism. By progressively absorbing free water and reducing the total volume of bleeding, bentonite effectively lowered the local w/c ratio of the top layer. Unlike HPMC, which relies on modifying the viscosity of the bleeding fluid, bentonite physically minimizes the formation of the porous laitance layer itself, creating a denser surface structure that linearly retards the advancement of the carbonation front.
3.3. Mineralogical Composition and Surface Carbonation Degree
To further investigate the microstructural consequences of bleeding and subsequent carbonation of different thickeners, the H-20 and B-60 samples were selected for detailed XRD and SEM characterization. This selection was based on the previous bleeding analysis: H-20 (0.2%) represents the critical saturation threshold for HPMC, where its pore-blocking efficacy reaches a plateau, while B-60 (0.6%) was identified as the characteristic dosage for bentonite that effectively bridges the mineralogical gap between the surface and the bulk matrix. Comparing these two samples provides insights into the maximum potential of organic versus inorganic thickeners in restoring longitudinal homogeneity. The XRD patterns of the top surface layer (bleeding surface) and the bottom layer (molded surface) for the Control, H-20, and B-60 groups at 28 d are presented in Figure 7.
Figure 7.
XRD patterns of the top surface layer and bottom layer of cement pastes at 28 d: (a) Control, (b) H-20, and (c) B-60.
As shown in Figure 7a, the Control group exhibits a drastic mineralogical disparity between the top and bottom layers. In the bottom layer, the characteristic peak of portlandite (CH) is dominant, indicating a normal hydration product composition. However, on the top surface, the intensity of the CH peak diminishes significantly, while the characteristic peak of calcite (CaCO3) becomes exceptionally strong. This qualitative inversion confirms that the porous laitance layer, formed by bleeding, has undergone severe carbonation, consuming the majority of the surface CH.
In the HPMC-modified group (Figure 7b, H-20), the intensity of the CaCO3 peak on the top surface is noticeably reduced compared to the Control, and the CH peak is partially preserved. This suggests that the HPMC network hindered the formation of a loose bleeding layer to some extent. In contrast, the bentonite-modified group (Figure 7c, B-60) displays the highest degree of homogeneity. The diffraction patterns of the top and bottom layers are remarkably similar, with comparable intensities for both CH and CaCO3 peaks. This visual evidence implies that the volume-filling effect of bentonite effectively eliminated the mineralogical stratification.
To quantitatively evaluate this stratification, the carbonation ratio of CH was calculated, as shown in Figure 8. The Control group exhibits the poorest homogeneity, with a massive gap between the top and bottom layers. The carbonation ratio of the top surface reached as high as ~56%, whereas the bottom was only ~27%. This huge difference corroborates the mechanism that the high w/c ratio bleeding layer creates open channels for CO2 diffusion, accelerating the consumption of portlandite.
Figure 8.
Carbonation ratio of CH in the top surface layer and bottom layer calculated via Rietveld quantitative analysis.
For the H-20 group, the carbonation ratio of the top surface decreased to ~44%. Although this represents an improvement over the Control, a distinct gap still exists compared to its bottom layer (~28%). This indicates that while HPMC densified the top layer (as proven by the lower T2 in NMR results), it did not fully eliminate the existence of a chemically distinct surface layer. Most notably, the B-60 group achieved optimal longitudinal homogeneity. The carbonation ratio of the top surface dropped significantly to ~26%, which is statistically almost identical to that of the bottom layer (~25%). The negligible difference strongly supports the conclusion that bentonite, by physically retaining water and preventing the accumulation of free water at the surface, fundamentally inhibited the formation of a vulnerable laitance layer. Consequently, the top surface possesses the same microstructural integrity and carbonation resistance as the internal matrix.
3.4. Micro-Structure Analysis of the Top and Bottom Surface
To intuitively visualize the impact of thickeners on the structural stratification and carbonation behavior, the microstructures of the top surface layer (bleeding surface) and bottom layer (molded surface) for the Control, H-20, and B-60 groups were characterized via SEM, as shown in Figure 9.
Figure 9.
SEM images of the top surface layer (a,c,e) and bottom layer (b,d,f) of cement pastes at 28 d: (a,b) Control group; (c,d) H-20 group; (e,f) B-60 group.
The SEM observations in Figure 9 further substantiate the mineralogical gradients identified by XRD. In the Control-Top samples (Figure 9a,b), a characteristic loose and granular morphology is observed, dominated by large, well-developed portlandite (CH) plates and sporadic needle-like ettringite. This microstructural ‘looseness’ is a direct consequence of the high local w/c ratio in the laitance layer caused by unchecked bleeding.
In contrast, the B-60-Top sample (Figure 9e) exhibits a significantly more refined and continuous matrix. The C-S-H gel appears more integrated, effectively encapsulating the hydration products and reducing the visible inter-particle porosity. Crucially, the microstructural features of B-60-Top are nearly indistinguishable from those of its bottom counterpart (Figure 9f), confirming the longitudinal homogeneity promoted by the volume-filling effect of bentonite. While the physical volume-filling and thixotropic effects are the primary drivers for the observed stability, the potential chemical contribution of bentonite should not be entirely overlooked. As a clay mineral, bentonite may exhibit a limited degree of pozzolanic reactivity with CH over time. However, considering that the bentonite used in this research was uncalcined and the dosages were relatively low, such chemical effects are expected to be secondary to the immediate physical immobilization of water. The rapid suppression of bleeding channels—as evidenced by the early-age LF-NMR results—confirms that the restoration of surface homogeneity is predominantly a result of the inorganic thickener’s ability to physically structure the suspension and prevent particle–water separation.
For the H-20 series (Figure 9c), although the top layer appears denser than the Control-Top due to the reduction in total bleeding, it still lacks the structural maturity seen in the bottom layer. The ‘pore-blocking’ effect of HPMC-swollen particles, while effective at a macroscopic level, results in a micro-domain distribution that is still ‘patchy’ compared to the uniform ‘house-of-cards’ scaffolding provided by bentonite. These observations confirm that while both thickeners improve surface quality, only the inorganic modifier (bentonite) restores the surface microstructure to its bulk-level integrity.
It is also noteworthy that while high dosages of HPMC and bentonite yield similar absolute carbonation depths, the underlying microstructural quality differs fundamentally. The carbonation resistance of HPMC-modified pastes is limited by the inherent stratification; the top layer remains a ‘modified laitance,’ which, despite being densified by polymer residues, lacks the mineralogical stability of the bulk matrix. In contrast, bentonite-modified pastes demonstrate superior longitudinal homogeneity. By physically entrapping water and preventing particle–water separation, bentonite ensures that the surface properties are a true reflection of the bulk material. The nearly identical carbonation ratios and SEM morphologies between the top and bottom layers of the B-60 group suggest that inorganic thickeners not only ‘fix’ the surface, but also prevent the surface from becoming a weak link in the first place.
4. Conclusions
This study presented a comparative investigation into the effects of organic (HPMC) and inorganic (bentonite) thickeners on the bleeding behavior, water migration dynamics, and surface durability of cement pastes. Based on the in situ LF-NMR monitoring and multi-scale characterization of the hardened “top surface layer” versus the “bottom layer,” the following conclusions can be drawn:
- (1)
- HPMC and bentonite exhibit fundamentally different regulatory modes on water migration. HPMC functions via a “pore-blocking” mechanism with a distinct saturation threshold. A dosage of 0.2% is critical; below this (e.g., 0.05%), HPMC acts as a dispersant, inducing delayed segregation; above this, the bleeding rate stabilizes (at ~25%), and further dosage increases yield minimal additional reduction. In contrast, bentonite operates through a “volume-filling” mechanism. It exhibits a continuous, dosage-dependent efficacy, where the bleeding rate decreases monotonically with increasing dosage, achieving a significantly lower bleeding capacity (~6.5%) at 1.2% dosage compared to the saturation state of HPMC.
- (2)
- LF-NMR analysis revealed that the two thickeners alter the physical state of the bleeding fluid differently. HPMC increases the viscosity of the pore solution, resulting in a shorter transverse relaxation time (T2) for the bleeding peak, which implies a “densified” supernatant. Conversely, bentonite physically entraps water within its thixotropic “house-of-cards” structure without significantly altering the rheology of the residual free water (longer T2), thereby primarily reducing the volume of the laitance layer rather than its viscosity.
- (3)
- Bentonite demonstrates superior capability in maintaining the longitudinal homogeneity of the hardened paste. XRD and SEM analyses confirmed that while the Control and low-dosage HPMC groups suffered from severe mineralogical stratification (porous calcite-rich top layer vs. dense portlandite-rich bottom layer), the high-dosage bentonite group (B-60) effectively eliminated this disparity. The top surface of the B-60 group exhibited a dense microstructure and phase composition indistinguishable from the bottom matrix. HPMC improved surface density but failed to fully eliminate the compositional gap between the surface and the bulk.
- (4)
- Both thickeners significantly enhanced the carbonation resistance of the bleeding surface. However, their efficiency profiles differ. HPMC efficacy saturates after a 0.2% dosage, reducing the 28 d carbonation depth to a plateau of ~8.5 mm. Bentonite provides a progressive enhancement, with high dosages (0.9–1.2%) reducing the carbonation depth to ~8.0 mm and the surface carbonation ratio to a level comparable to the internal matrix.
In summary, while HPMC is a highly efficient thickener for general stability, bentonite offers superior performance in eliminating structural stratification and ensuring the uniformity of surface durability, provided that sufficient dosage is applied. This suggests that for engineering applications sensitive to surface quality (e.g., abrasion or carbonation resistance), inorganic volume-filling thickeners may offer distinct advantages over pore-blocking organic polymers.
Author Contributions
Conceptualization, J.Y. and Z.S. Investigation, S.Z. and A.G. Data curation, S.Z. Writing—original draft, J.Y. and S.Z. Writing—review and editing, Z.S. Supervision, J.Y. and Z.S. Project administration, Z.S. Funding acquisition, J.Y. and Z.S. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by the National Natural Science Foundation of China, Grant No. 52408284 and Grant No. 52402033.
Data Availability Statement
The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding authors.
Conflicts of Interest
The authors declare no conflicts of interest.
References
- Wainwright, P.; Ait-Aider, H. The influence of cement source and slag additions on the bleeding of concrete. Cem. Concr. Res. 1995, 25, 1445–1456. [Google Scholar] [CrossRef] [Scilit]
- Josserand, L.; Coussy, O.; De Larrard, F. Bleeding of concrete as an ageing consolidation process. Cem. Concr. Res. 2006, 36, 1603–1608. [Google Scholar] [CrossRef] [Scilit]
- Han, J.; Wang, K. Influence of bleeding on properties and microstructure of fresh and hydrated Portland cement paste. Constr. Build. Mater. 2016, 115, 240–246. [Google Scholar] [CrossRef] [Scilit]
- Torney, C.; Forster, A.M.; Banfill, P.F.G.; Szadurski, E.M. The effects of site practice on the physical properties of proprietary stone restoration mortar. Constr. Build. Mater. 2015, 75, 359–367. [Google Scholar] [CrossRef] [Scilit]
- Wang, J.; Xu, H.; Wang, B.; Zou, J.; Liu, H. Accurate detection technology of super long bored cast-in-place pile concrete pouring thickness based on ultrasonic inclined plane reflection measurement. Measurement 2022, 198, 111314. [Google Scholar] [CrossRef] [Scilit]
- Topçu, İ.B.; Elgün, V.B. Influence of concrete properties on bleeding and evaporation. Cem. Concr. Res. 2004, 34, 275–281. [Google Scholar] [CrossRef] [Scilit]
- Paiva, H.; Silva, L.M.; Labrincha, J.A.; Ferreira, V. Effects of a water-retaining agent on the rheological behaviour of a single-coat render mortar. Cem. Concr. Res. 2006, 36, 1257–1262. [Google Scholar] [CrossRef] [Scilit]
- Cappellari, M.; Daubresse, A.; Chaouche, M. Influence of organic thickening admixtures on the rheological properties of mortars: Relationship with water-retention. Constr. Build. Mater. 2013, 38, 950–961. [Google Scholar] [CrossRef] [Scilit]
- Tan, O.; Zaimoglu, A.S.; Hinislioglu, S.; Altun, S. Taguchi approach for optimization of the bleeding on cement-based grouts. Tunn. Undergr. Space Technol. 2005, 20, 167–173. [Google Scholar] [CrossRef] [Scilit]
- Marchon, D.; Kawashima, S.; Bessaies-Bey, H.; Mantellato, S.; Ng, S. Hydration and rheology control of concrete for digital fabrication: Potential admixtures and cement chemistry. Cem. Concr. Res. 2018, 112, 96–110. [Google Scholar] [CrossRef] [Scilit]
- Zhang, Q.; Qiu, J.; Jiang, H.; Kong, X.; Guo, Z.; Xiang, J.; Sun, X. Effect of hydroxypropyl methyl cellulose on coarse tailings cemented backfill: Rheology, stability, strength and microstructure. Constr. Build. Mater. 2024, 425, 136042. [Google Scholar] [CrossRef] [Scilit]
- Zhang, Y.; Zhao, Q.; Liu, C.; Zhou, M. Properties comparison of mortars with welan gum or cellulose ether. Constr. Build. Mater. 2016, 102, 648–653. [Google Scholar] [CrossRef] [Scilit]
- Ma, B.; Peng, Y.; Tan, H.; Jian, S.; Zhi, Z.; Guo, Y.; Qi, H.; Zhang, T.; He, X. Effect of hydroxypropyl-methyl cellulose ether on rheology of cement paste plasticized by polycarboxylate superplasticizer. Constr. Build. Mater. 2018, 160, 341–350. [Google Scholar] [CrossRef] [Scilit]
- Kaci, A.; Chaouche, M.; Andréani, P.A. Influence of bentonite clay on the rheological behaviour of fresh mortars. Cem. Concr. Res. 2011, 41, 373–379. [Google Scholar] [CrossRef] [Scilit]
- Lin, C.Y.; Huang, W.T. Effect of mineral viscosity-enhancing admixtures on the solidification of evaporator concentrates. J. Hazard. Mater. 2015, 298, 294–302. [Google Scholar] [CrossRef] [Scilit]
- Jiang, J.; Lu, Z.; Li, J.; Xie, Y.; Luo, K.; Niu, Y. Preparation and properties of nanopore-rich lightweight cement paste based on swelled bentonite. Constr. Build. Mater. 2019, 199, 72–81. [Google Scholar] [CrossRef] [Scilit]
- Huertas, F.J.; Hidalgo, A.; Rozalén, M.L.; Pellicione, S.; Domingo, C.; Garciagonzalez, C.; Andrade, C.; Alonso, C. Interaction of bentonite with supercritically carbonated concrete. Appl. Clay Sci. 2009, 42, 488–496. [Google Scholar] [CrossRef] [Scilit]
- Liu, Y.; Han, J.; Li, M.; Yan, P. Effect of a nanoscale viscosity modifier on rheological properties of cement pastes and mechanical properties of mortars. Constr. Build. Mater. 2018, 190, 255–264. [Google Scholar] [CrossRef] [Scilit]
- Liu, J.; Li, Y.; Zhang, G.; Liu, Y. Effects of cementitious grout components on rheological properties. Constr. Build. Mater. 2019, 227, 116654. [Google Scholar] [CrossRef] [Scilit]
- Huang, J.; Wang, C.; Jian, S.; Tan, H.; Li, X.; Huang, J.; Lv, Y.; Gao, X.; Peng, B.; Tu, L. Feasibility of applying attapulgite, sodium bentonite and nano-silica as a viscosity modifier admixture for 3D printing of gypsum-based materials. J. Build. Eng. 2025, 107, 112745. [Google Scholar] [CrossRef] [Scilit]
- Palacios, M.; Flatt, R.J. Working mechanism of viscosity-modifying admixtures. In Science and Technology of Concrete Admixtures; Woodhead Publishing: Cambridge, UK, 2016; pp. 415–432. [Google Scholar]
- Bessaies-Bey, H.; Baumann, R.; Schmitz, M.; Radler, M.; Roussel, N. Effect of polyacrylamide on rheology of fresh cement pastes. Cem. Concr. Res. 2015, 76, 98–106. [Google Scholar] [CrossRef] [Scilit]
- Brumaud, C.; Baumann, R.; Schmitz, M.; Radler, M.; Roussel, N. Cellulose ethers and yield stress of cement pastes. Cem. Concr. Res. 2014, 55, 14–21. [Google Scholar] [CrossRef] [Scilit]
- Quanji, Z.; Lomboy, G.R.; Wang, K. Influence of nano-sized highly purified magnesium alumino silicate clay on thixotropic behavior of fresh cement pastes. Constr. Build. Mater. 2014, 69, 295–300. [Google Scholar] [CrossRef] [Scilit]
- Sanchez, F.; Sobolev, K. Nanotechnology in concrete—A review. Constr. Build. Mater. 2010, 24, 2060–2071. [Google Scholar] [CrossRef] [Scilit]
- Kawashima, S.; Chaouche, M.; Corr, D.J.; Shah, S.P. Rate of thixotropic rebuilding of cement pastes modified with highly purified attapulgite clays. Cem. Concr. Res. 2013, 53, 112–118. [Google Scholar] [CrossRef] [Scilit]
- Gu, L.; Liu, T.; Wu, K.; Yang, Z.; Wen, Z.; Zhang, Z.; De Schutter, G.; Li, H. Comparative study on the role of PAM and PANA on the property of fresh cement paste. Cem. Concr. Compos. 2022, 133, 104701. [Google Scholar] [CrossRef] [Scilit]
- McDonald, P.J.; Mitchell, J.; Mulheron, M.; Aptaker, P.S.; Korb, J.-P.; Monteilhet, L. Two-dimensional correlation relaxometry studies of cement pastes performed using a new one-sided NMR magnet. Cem. Concr. Res. 2007, 37, 303–309. [Google Scholar] [CrossRef] [Scilit]
- Peng, Y.; Lauten, R.A.; Reknes, K.; Jacobsen, S. Bleeding and sedimentation of cement paste measured by hydrostatic pressure and Turbiscan. Cem. Concr. Compos. 2017, 76, 25–38. [Google Scholar] [CrossRef] [Scilit]
- Ji, Y.; Pel, L.; Sun, Z. The microstructure development during bleeding of cement paste: An NMR study. Cem. Concr. Res. 2019, 125, 105866. [Google Scholar] [CrossRef] [Scilit]
- GB 175-2023; Common Portland Cement. Standards Press of China: Beijing, China, 2023.
- GB/T 8077-2012; Methods for Testing Uniformity of Concrete Admixture. Standards Press of China: Beijing, China, 2012.
- Meiboom, S.; Gill, D. Modified Spin-Echo Method for Measuring Nuclear Relaxation Times. Rev. Sci. Instrum. 1958, 29, 688–691. [Google Scholar] [CrossRef] [Scilit]
- Gilbert, P. Iterative methods for the three-dimensional reconstruction of an object from projections. J. Theor. Biol. 1972, 36, 105–117. [Google Scholar] [CrossRef] [Scilit]
- Ji, Y.; Sun, Z.; Yang, J.; Pel, L.; Raja, A.J.; Ge, H. NMR study on bleeding properties of the fresh cement pastes mixed with polycarboxylate (PCE) superplasticizers. Constr. Build. Mater. 2020, 240, 117938. [Google Scholar] [CrossRef] [Scilit]
- GB/T 50082-2009; Standard for Test Methods of Long-Term Performance and Durability of Ordinary Concrete. Standards Press of China: Beijing, China, 2009.
- Ji, Y.; Sun, Z.; Yang, X.; Li, C.; Tang, X. Assessment and mechanism study of bleeding process in cement paste by 1H low-field NMR. Constr. Build. Mater. 2015, 100, 255–261. [Google Scholar] [CrossRef] [Scilit]
- Gu, L.; Liu, T.; Wu, K.; Yang, Z.; Wen, Z.; Zhang, Z.; Li, H. Temperature-dependent water retention and bleeding of fresh cement-based materials with VEAs. Cem. Concr. Compos. 2022, 130, 104539. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.










