Influence of the graphene oxide nanosheet on tensile behavior and failure characteristics of the cement composites after high-temperature treatment

2.1 Materials and instrumentation

In this research, ordinary Portland cement (OPC, type PO.32.5 [47,48,49]) was used. The chemical compositions are shown in Table 1. The industrial-grade GO was supplied by Nanjing XFNANO Materials Technology Co. (Nanjing, Jiangsu, China). The physical properties are shown in Table 2. To improve the dispersibility of GO in the solution, a polycarboxylate superplasticizer (PS) was incorporated as an external additive.

2.2 Preparation of GO-reinforced cement composites

The preparation of GO-reinforced cement composites involves two steps: dispersing GO suspension and mixing with cement powder. As per prior research [35], GO and polycarboxylate superplasticizer were added to the suspension at 0.08 and 0.64 wt%, respectively. The detailed procedure is as follows: First, GO and PS powders were precisely weighed and dissolved in deionized water. The mixture was magnetically stirred for 180 s before ultrasonic treatment using a VCX-500W device (SONICS, USA). Following the authors’ previous protocol [48], ultrasonic dispersion was conducted at 60 W for 10 min with a 3s-on/3s-off pulse mode to avoid overheating. The dispersed GO suspension was mixed with cement (W/C = 0.4). The cementitious paste was mechanically mixed at 180 rpm for 3 min, poured into 50 mm × 100 mm cylindrical molds, and cured in a standard environment (20 ± 2°C, ≥95% relative humidity) for 28 days. After curing, each cylindrical specimen was cut into three Brazilian disk specimens (50 mm diameter, 25 mm thickness). A control group of pure cement pastes without GO was prepared simultaneously for comparative analysis (Table 3).

The cement composite specimens were subjected to high-temperature treatment using a SX2-8-10N box-type resistance furnace (dimensions: 250 mm × 400 mm× 160 mm, maximum working temperature: 1,200°C). Specimens were first preheated at 75 V for 10 min to prevent thermal shock, after which the voltage was increased to 100 V and the furnace was heated to target temperatures (200, 400, 600, and 800°C) at a rate of 10°C/min. Each target temperature was held isothermally for 120 min, followed by natural cooling to room temperature (25°C).

2.3 Experimental methods

3.1 Tensile properties of the GO-reinforced cement composites

As shown in Figure 1, it can be found that the tensile strength of the hardened cement composite specimens after high-temperature treatment tends to rise first and reaches the maximum tensile strength of 4.78 MPa at 200°C. Nevertheless, as the temperature rises, the tensile strength decreases gradually and hits the lowest value of 1.18 MPa at 800°C. This is mainly due to the fact that at 200°C treatment, no obvious material changes occurred in the cement. Mainly, the evaporation of free water and some bound water took place, and the main hydration products of the cement material did not decompose. Meanwhile, the free water and physically bound water in the cement composites begin to evaporate, forming an autoclave environment within the matrix. In autoclaved environments, unhydrated cement particles in the composites produce additional hydration reactions, enhancing the tensile properties of the hardened specimens [49]. Elevated temperatures induced progressive dehydration of hydration products and decomposition of the C–S–H skeleton in cement composites, leading to continuous deterioration of mechanical properties [50]. This degradation culminated in minimal tensile strength and elastic modulus at 800°C.

Figure 1

Tensile strength of each sample at different temperatures.

With GO addition, the tensile properties of cement composites after high-temperature treatment were notably enhanced, as illustrated in Figure 1. Across all tested temperatures, both the tensile strength and elastic modulus of GO-reinforced specimens exceeded those of the control group. Similar to control specimens, the tensile strength of GO-modified samples reached its peak at 200°C and lowest value at 800°C. According to the test results, incorporating GO can increase the tensile strength of the cement composites from 6.5 to 46.8%. With the increase in temperature, the enhancement efficiency showed an increasing trend. When the temperature is 400°C, the increasing rate of the GO-reinforced specimen achieved about 31%. By contrast, for 600 and 800°C, the increasing rates were rising to approximately 44–45% compared with the control group. Studies demonstrate that GO enhances cement composites via dual mechanisms: nucleation sites promotion [51] and pore structure refinement [52]. The oxygen-rich functional groups on GO surfaces facilitate calcium ion adsorption, enabling denser C–S–H formation through localized hydration compared to plain cement [53]. Under thermal exposure, mobile GO nanosheets establish supplementary chemical bonds with C–S–H networks, elevating thermal stability by increasing the energy barrier for structural disintegration. Microstructurally, GO acts as both bridging reinforcement within hydrated phases and a template for creating “wall-like” dense configurations through inter-sheet interactions upon hardening [54]. During mechanical loading, this hierarchical architecture generates interfacial friction between GO and matrix, effectively delaying failure initiation [53,55].

3.2 TGA

TGA in Figure 2 highlights the enhanced thermal stability of GO-modified cement composites. Three key degradation stages were observed. Below 200°C, both samples experienced approximately 50% mass loss primarily due to evaporation of capillary and bound water, with GO composites exhibiting slower dehydration rates compared to the reference group [11]. Between 200°C and 250°C, the decomposition of oxygen-containing groups in GO sample produces substances including carbon dioxide and water. This resulted in 23% less mass loss than that of plain cement, indicating structural stabilization through the selective elimination of labile components. At 650°C, a sharp increase in the DTG curve signaled accelerated C–S–H decomposition, yet GO-modified specimens demonstrated delayed degradation compared to carbon nanotube-reinforced systems that rapidly decomposed at 500–540°C [56]. This thermal resilience stems from GO’s dual mechanisms: chemically, the sacrificial release of oxygen groups strengthens residual graphene frameworks, while structurally, the reinforced GO-C–S–H interfaces require 18% higher energy for bond dissociation. Additionally, the preserved GO network physically inhibits crack propagation under thermal stress, synergistically improving high-temperature performance.

Figure 2

TG/DTG of sample.

3.3 Failure event distribution of AE

To investigate the influence of internal crack evolution on the tensile fracture behavior of cement composites after high-temperature degradation, an AE system was employed for real-time monitoring of the entire Brazilian splitting test. During loading, internal cracking and damage generate elastic waves, which are captured by the AE equipment [57,58]. The sample with the smallest error from the average was selected for analysis, as shown in Figure 3. During the initial stage, AE event rates are predominantly at low levels. As loading increases, a higher event rate emerges near the peak stress. This phenomenon occurs because hardened cementitious specimens contain inherent primary pores: the compaction stage generates fewer AE events, while the abrupt failure after the elastic stage leads to a surge in event rates [59]. Regarding temperature effects, the number of AE events first increases and then decreases as the treatment temperature rises. Previous research indicates that when the temperature surpasses approximately 105°C, free water within cement composites start to evaporate [60,61]. Between 120 and 150°C, the C–S–H phase releases water [62]. This water loss creates more pores, resulting in a higher AE event count during loading. As the temperature increases to 350°C, the C–S–H phase begins to degrade. At 450–570°C, the decomposition of calcium hydroxide (CH) and C–S–H structures promotes the interconnection of pores and cracks in the matrix, reducing the frequency of AE events.

Figure 3

Acoustic emission counting during loading process. (a)–(e) The trend of changes in acoustic emission counts and cumulative counts in the Ref group. (f)–(j) The trend of changes in acoustic emission counts and cumulative counts in the GO group.

Furthermore, the addition of GO promoted the transformation of the failure mode of the cement sample from a violent failure mode to a large number of small-scale failures. This change directly affected the change in the distribution of AE event numbers during the entire loading process of the sample until failure. As demonstrated in Figure 3(a) and (f), after mixing GO nanosheets, several small amplitude AE signals are generated before the peak, while the number of AE events of the control group specimens is concentrated near the peak value. The AE event rate generated at the peak stage under different temperature conditions is calculated, as exhibited in Figure 4(a). By contrast, the event rate of the sample in the GO group is significantly lower, with a reduction range of 29–30%. The increase in temperature also leads to an increase in the rate of AE events before the peak phase, which can be attributed to the damage caused by high-temperature degradation. Taking GO-4 and Ref-4 as an example, both have a small range of damage before the peak stage, and the frequency of GO-4 is smaller than that of Ref-4. Figure 4(b) shows the statistics of the cumulative AE rate before the peak stage. Overall, the number of AE events in the GO group is smaller than that in the Ref. group, with a reduction in the range of 27–76%.

Figure 4

(a) AE count during peak phase. (b) Accumulated AE count before peak stage.

Experimental results indicate that the addition of GO contributes to the closure and reduction in pores and cracks in cement composites, thereby decreasing the number of AE events during their loading process. Such effects are primarily attributed to GO’s promotion of the hydration reaction, which generates more C–S–H gel and refines the pore structure. As temperature rises, GO alleviates thermal stress and hinders the initiation of thermal cracks [63]. The presence of GO nanosheets prevents the rapid propagation of single cracks at high temperatures [64], leading to a lower occurrence of AE event rates.

3.4 Energy dissipation of AE

When the sample is damaged, strain energy and dissipative energy are generated, and the signal strength and absolute energy could reflect the intensity of the failure inside the hardened cement composites. As shown in Figure 5, both signal strength and absolute energy values remain low, indicating minimal damage intensity during this period before the peak stage. As the load increases, internal energy in cement composites accumulates gradually, leading to sample cracking at peak strength. At this point, the material experiences high-intensity failure, causing a sudden surge in signal strength that reflects substantial energy release. After peak strength, the sample undergoes large-scale structural failure, resulting in a significant drop in overall strength. Subsequent failure events primarily involve small-scale crack propagation, requiring less energy and producing values lower than those released during peak-strength failure. As shown in Figure 6(a) and (b), the signal strength and absolute energy exhibit a trend of increasing first and then decreasing as temperature rises. In the Ref. group, the signal strength and absolute energy reached the maximum value at 200°C, which were 1.8 × 10⁸ pVs and 1.9 × 10⁸ aJ, respectively. This is due to the high-temperature sterilization environment formed inside the cement composites at this temperature condition, which promotes the additional hydration of cement particles and enhances the tensile strength of the hardened matrix [65]. When the temperature rises to 800°C, the signal intensity and absolute energy reach the minimum values, which are 1.1 × 10⁷ pVs and 1.2 × 10⁷ aJ, dropping by 93.9 and 93.6% compared with 200°C, respectively. This phenomenon indicates that the high-temperature degradation causes severe damage inside the cement composites and dramatically reduces the ability of the cement composites to resist the load.

Figure 5

The dissipation of energy in AE signals emission. (a)–(e) The trend of changes in acoustic emission Signal Strength and Absolute Energy in the Ref group. (f)–(j) The trend of changes in acoustic emission Signal Strength and Absolute Energy in the GO group.

Figure 6

(a) Signal strength during peak phase. (b) Absolute energy during peak phase.

As shown in Figure 6, the signal strength and absolute energy of the hardened cement composite specimens in the GO group are lower than those in the control group on the whole. This phenomenon indicates that the destruction of the sample after the incorporation of GO presents a pattern of multiple small-scale destruction, and the energy released in a certain period is reduced. At 200°C, the signal strength and absolute energy reach the maximum, achieving 6.1 × 10⁷ pVs and 1.0 × 10⁸ aJ, respectively. When the temperature rises to 800°C, the signal intensity and absolute energy reach the minimum values, hitting 6.4 × 10⁶ pVs and 6.4 × 10⁶ aJ and decreasing by 89.5 and 93.6% compared to that at 200°C. Compared with the Ref. group, the decline in signal strength was reduced, suggesting that GO reduced the damage of cement composites under high-temperature treatment.

3.5 Analysis of destruction modes characterized by AE signals

Tensile and shear are the two primary forms of microcracks in the process of specimen failure [66]. In order to display the changes in ratio of rise time to amplitude-average frequency: ratio of AE count to signal duration in different loading stages, different colors are used to represent them, as shown in Figure 7. In the Ref. group, the damage of samples is mainly tensile. The number of tensile failures increases with the treated temperature until 400°C. This is because after high-temperature degradation, there are more hot cracks inside the sample. When the temperature rises to 600–800°C, the number of tensile failures decreases obviously, and the pattern of mixed failure increases. It can be seen that under these conditions, severe damage occurs inside the cement composites, and the microcracks extend and expand to form significant defects. Take Ref-3 and GO-3 as examples, as shown in Figure 7(e) and (f). At 400°C, cement composites will produce severe damage, at which point the C–S–H skeleton decomposes, significantly reduce the carrying capacity. The Ref. group produced more tension failure from the failure mode at this temperature. In contrast, the GO group significantly decreased the proportion of tension failure and was more inclined to mixed failure mode. This may be due to the bridge action of GO and the promotion of hydration, which inhibits the development of a single tensile fracture along the cement micropores and reduces tensile failure [35].

Figure 7

Distribution of AF and RA values of acoustic emission signals. (a, c, e, i, j) Distribution of AF and RA values of acoustic emission signals in Ref group. (b, d, f, h, j) Distribution of AF and RA values of acoustic emission signals in GO group.

The AE monitoring results show a high degree of consistency with the change in the tensile strength of cement paste. Overall, the AE event rate, signal intensity, and absolute energy reach their maximum values during the peak load stage, reflecting the failure process of crack cooperative propagation and concentrated energy release. At different temperatures, the AE response characteristics of the material are closely related to its tensile strength. Among them, the AE signal of the specimen under the treatment condition of 200°C is the strongest, the energy release is the largest, and at the same time, the tensile strength reaches its peak, indicating that the material has good load-bearing and energy accumulation capabilities. Cement paste exhibits the best mechanical properties at 200°C. This is mainly because this temperature can promote the secondary hydration of incompletely hydrated particles, generating more C–S–H gel, filling pores, and improving compactness. At the same time, the main cementing phase does not decompose, and the structure is stable. The corresponding AE parameters also show intense energy release, which is a direct manifestation of its enhanced performance. The introduction of GO changes the failure process from a single violent release to multiple small-energy releases. Although the signal intensity is relatively low at 200°C, the tensile strength is significantly improved. This indicates that GO improves the stability and toughness of the failure process, effectively delays the unstable propagation of cracks, and improves the structural integrity and load-bearing capacity of the material at high temperatures.

3.6 Crack development process

As shown in Figure 8, the stress curve for the tensile properties of cement composites are divided into four stages [67]. Stage I corresponds to the elastic strain phase, where stress distribution in hardened cement composites remains relatively uniform. Only minor strain areas appear at the periphery, indicating the sample is undergoing elastic deformation without permanent damage. Stage II represents the plastic stage. For instance, in Ref-1, as stress increases, a distinct strip-shaped tensile strain zone emerges. Green pixels denoting tensile strain show a dark, dense central region and lighter, thinner ends within this zone, reflecting localized deformation as the material transitions into plastic behavior. Stage III is the peak strain stage, during which crack propagation accelerates rapidly when the load reaches peak strength. The specimen loses its load-bearing capacity abruptly, causing a sharp stress drop as catastrophic failure occurs. Stage IV refers to the residual strain phase, where existing cracks continue to expand and damage accumulates progressively. The material’s carrying capacity diminishes until it is completely lost, marking the final stage of tensile failure.

Figure 8

(a)–(j) Destructive equivalent strain field cloud map.

For pure cement specimens, taking Ref-1 in Stage I as an example, the strain distribution of the sample is relatively uniform, indicating that the sample was in the stage of elastic deformation. At stage II, with the increase in load, tensile strain distribution appears in the local range of the sample. After that, the tensile strain expands to form cracks during stage III. In stage IV, the damage is further aggravated, the crack penetrates, and the specimen loses its bearing capacity. With the increased treated temperature, the area of tensile strain in a small range began to increase before stage II. By comparing Sample Ref-2 and Ref-3, it can be found that the strain in a small area began to appear at stage I due to the increase in micro-cracks and the intensification of regional damage in the sample caused by high temperature.

In the GO group, the overall failure mode of GO-reinforced specimens is similar to that of pure cement composite specimens. However, taking GO-4 as an example, more small-range tensile strain regions appeared and presented a network-like intermittent distribution in stages I and II. When the first major tensile strain zone appears, the presence of GO improves the performance of the weak zone and inhibits the continuous development of cracks along micro-cracks. At the same time, due to the better stress transfer of GO, the sample reached the failure critical state before the defects in other area were further developed into cracks, delaying the failure of the sample. In comparison to conventional reinforcement materials [68], GO serves as a more effective reinforcement material in cement composites due to its superior capability in regulating the initiation and expansion of cracks at the nanoscale [69].

3.7 3D scanning

This study employs the microscopic morphology statistical method proposed by Belem et al. [70], which involves two key steps: (1) discretizing and meshing the 3D fracture surface scanning point cloud data into a series of meso-planes; (2) defining the local inclination angle (αij) of each meso-plane as the angle between the plane and the horizontal plane, followed by statistical calculations of the maximum, minimum, average values, standard deviation of αij, and the height of plane corners for meso-plane.

The fracture surface was reconstructed using MATLAB software, and the corresponding three-dimensional reconstruction rendering is shown in Figure 9. Then, the point cloud data obtained by three-dimensional scanning is discretized into a grid, and the meso-roughness height of each grid point is calculated. Figure 10 shows the mesoscopic roughness height distribution results of the fracture surface, with sample GO-4 taken as a typical example, and the corresponding standard deviation is calculated. Additionally, the standard deviation results of the roughness height for each specimen are plotted as a line chart following the aforementioned procedure. Figure 10(a) and (b) shows that across all temperatures, the standard deviation of fracture surface fluctuation height in the Ref. group is higher than that in the GO group. Therefore, the macro fracture surface fluctuation in the Ref. group is higher than that in the GO group, indicating that the formation of fracture surface in Ref group overcomes more surface energy and causes the crack expansion path to be more complex and tortuous.

Figure 9

(a)–(j) 3D reconstruction of fracture surface after tensile failure.

Figure 10

(a) Roughness height distribution of fracture surface in sample GO-4. (b) Local inclination angle of sample GO-4 fracture surface. (c) Standard deviation of fracture surface roughness height for each group of samples. (d) Standard deviation of local inclination angle of fracture surface of each group of samples.

Owing to the presence of micro-cracks and pores in Ref. group specimens, under constant loading, internal stored energy is released to stress-concentrated areas at micro-crack tips, causing cracks to propagate along preferential directions and forming more tortuous paths. In contrast, GO sheets in the GO group fill pores and enhance cement hydration, promoting additional C–S–H gel formation that effectively improves specimen integrity. Thus, the accumulated energy was too late to release under the action of load, resulting in smaller macro-fluctuation of the fracture surface. In addition, according to the research of Win et al. [32,33,71,72], GO plays a bridging role in the cement matrix. When subjected to tensile stress, the inhibitory effect of GO on crack propagation is more pronounced.

It can be seen from Figure 10(c) and (d) that the standard deviation of fluctuation height in the Ref. group increases significantly at 400 and 800°C, indicating that the fracture surfaces formed at these temperatures are more complex and irregular. This trend is closely related to the thermal degradation and phase transitions of cement hydration products within this temperature range. Specifically, around 400°C, the decomposition of C–S–H gel begins, and chemically bound water is gradually released, leading to the formation of microcracks and increased porosity. Under loading, these defects promote irregular crack propagation and result in a rougher fracture morphology. At 800°C, the degradation becomes much more severe. The C–S–H skeleton is almost completely decomposed, while CH has already undergone significant decomposition starting around 450–550°C. Additionally, calcium carbonate (CaCO₃), which may be present as a secondary product, begins to decompose at temperatures above 700°C. These transformations result in a highly porous, structurally compromised matrix with poor cohesion, which explains the large fluctuations in fracture surface height at this stage. In contrast, the GO group exhibits more stable behavior across all temperatures. At 400°C, the fluctuation height remains relatively low, indicating a more uniform and less damaged fracture surface. From 600 to 800°C, only a gradual increase in roughness is observed. This improved thermal resistance can be attributed to the presence of GO nanosheets, which undergo partial thermal reduction above 200°C, forming nano-defects. These defects enable silicate species to interact with the GO structure, reducing the oxygen exchange rate between GO and C–S–H, and promoting the formation of thermally stable S–O–Si and C–O–Si bonds. These new chemical linkages contribute to enhancing the integrity of the C–S–H network and improving the thermal stability of the composite [73]. Although C–S–H begins to decompose at 400°C, due to the improved thermal stability, the samples in GO group can maintain a relatively complete C–S–H gel skeleton, and slow down the decomposition of C–S–H at 600 and 800°C. The mathematical statistical results of the meso-plane local inclination angle in fracture surface actually reflect the meso-plane roughness characteristics of the fracture surface. As shown in Figure 10(b), the standard deviation of the GO group is lower than Ref. group, indicating that the fracture surface is coarser.

3.8 SEM simulation

In this study, SEM was used to examine the fracture surfaces of samples. Fractal theory is widely used in studying porous materials, as specimen failure under external load is essentially attributed to the nonlinear superposition of internal microstructural deformations – a statistically significant process. Employ the fractal box dimension to analyze the fracture surfaces. SEM images are processed using graythresh(I) function with Otsu method in MATLAB, which determines threshold T by minimizing within-class variance of grayscale pixels in image (I). As shown in Figure 11, the fractal process refers to the method proposed [31].

Figure 11

(a) Image binarization and grayscale values. (b) Calculation process of fractal dimension.

Figure 12 shows the microstructure of the fracture surfaces of the Ref. and GO groups. The calculated fractal dimension results are presented in Figure 13. Previous studies have shown that a larger fractal dimension indicates a higher degree of surface damage in the sample. The fracture surface analysis results closely align with the findings from mechanical testing and AE characterization. The fractal dimension values of GO-group samples remain consistently lower than those of the reference group. This indicates that the addition of GO enhances sample structural integrity and that GO maintains its reinforcing capability even under high-temperature conditions. Notably, between 20 and 200°C, a downward trend in the fractal dimension of GO-group samples was observed, which correlates with the improved tensile strength of these samples at 200°C discussed in Section 3. This behavior may be attributed to the further hydration of unreacted cement particles under internal autoclaving conditions, implying that GO promotes the hydration reaction of cement within the 20–200°C range, thereby enhancing the integrity and strength of GO-group samples. At 400°C, an increase in fractal dimension and intensified fracture surface damage were observed, likely due to partial decomposition of CH or C–S–H phases. The fractal dimension continued to rise at 600 and 800°C, indicating progressive sample damage. This is because C–S–H begins to decompose around 560°C, with significant decomposition occurring above 600°C, culminating in complete breakdown of C–S–H at 750°C.

Figure 12

SEM test image.

Figure 13

Typing dimension of each group of samples.

These results offer meaningful implications for real-world engineering applications involving fire exposure or high-temperature environments. The observed reduction in fractal dimension and improved fracture integrity at temperatures up to 200°C suggest that GO-modified cement materials may be particularly well-suited for structures subjected to moderate thermal loads – such as tunnel linings, underground facilities, and industrial floors where internal temperature rises may occur during service. The ability of GO to maintain structural continuity and suppress crack complexity even under severe thermal degradation (up to 800°C) indicates its potential for enhancing the fire resistance and post-fire mechanical recoverability of cementitious structures. This suggests a feasible pathway for developing high-performance, thermally durable concrete formulations for fire-prone infrastructures, offering enhanced resilience, safety, and service life in extreme conditions.