Keywords

1 Introduction

In the past decades, carbon nanomaterials, including carbon nanotubes (CNTs), graphene, graphene oxide and reduced graphene oxide, have attracted great attention as additives to reinforce the physical and mechanical properties of cementitious composites [1]. Due to their ultra-high specific surface area, superior mechanical properties, and the densification effect on the microstructure of the cement, CNTs have been proved to be capable of enhancing the tensile strength [2] and durability [3] of cement-based materials with a minimal mixing ratio (0.01–0.05% by weight of cement) [4]. Moreover, the latest research suggests that carbon nanomaterial reinforcement is a cost-effective strategy for reducing the environmental impact of cement usage and maintenance, providing an effective way to reduce CO2 emissions and energy consumption in the cement production and construction industries [5]. Hence, it is significant to carry out research on the development and application of carbon nanomaterial-modified cementitious composites and investigate the corresponding reinforcing mechanisms.

The micromorphological characteristics of the fracture surface of cementitious composites are closely related to the micromechanical damage evolution process of hardened cement-based materials. The micro-geometric features of the fracture surface record the irrecoverable deformation of the cementitious sample when it was broken, as well as micro-information from the initiation, propagation, penetration and nucleation of cracks to the final fracture and instability of the overall cement structure [6]. In scientific research and engineering practice, the relevant information recorded during the fracture damage process of cement materials can be obtained through the study of the micromorphology of the typical fracture surface of cement samples. Afterwards, mathematical statistics, induction and analysis of geometric features can be applied and then the mechanical mechanism of fracture could be reversed [7].

In the present study, we analyzed the reinforcing mechanisms of CNTs in cementitious composites by investigating the micromorphological features of the fracture surface of cement-based specimens under the Brazilian split test. First, we used 3D scanning technology to reconstruct the micromorphology of the cement-based specimens after the Brazil split test. Next, the micro-roughness characteristic parameters of the specimens were calculated and quantitatively analyzed. Considering the influential mechanism of micro-roughness on material fracture, a micro-dilatancy micro-element model of the fracture surface of the cement samples was constructed. The micro-dilation angle and the normal dilatation deformation of the CNT composite cement-based material after the Brazilian split test were calculated using the micro-roughness characteristic parameters. Finally, the influence of the CNTs on the fracture surfaces characteristics of cementitious composites under the Brazilian split test was discussed.

2 Methods

Ordinary Portland cement (PO. 32.5) was applied as the binder material and fly ash (FA), with a density of 2.4 g/cm3, was used as a partial replacement for cement powder in the cementitious composites. Multi-wall carbon nanotubes (MWCNTs), fabricated via chemical vapor deposition, were selected as the nano-reinforcing material to enhance the composites. Polycarboxylic acid water reducer (PC), an ionic surfactant, non-toxic and harmless, containing both hydrophobic and hydrophilic groups in the molecules, was used to improve the dispersion of CNTs in the suspensions. The CNTs dispersion and cement-based material pouring process were consistent with our previous report [8]. Three groups, Ref-group, FA-group and CNT-group, were marked in this study. A 0.4 water-to-cement ratio was applied in all three groups. For the FA-group, 20 wt% FA was used as a substitute to reduce cement usage and enhance the workability of the paste to generate a cost-effective material. For the CNT-group, 0.08 wt% CNTs was mixed into the cement–FA-based slurry to further optimize the pore structure of the hardened matrix and enhance the mechanical properties of the cementitious composite.

Then, the 28-day-old standard cured disc samples with a diameter of 50 mm and a thickness of 25 mm were selected for the Brazilian split tests. The loading rate was 0.10 mm/min. In order to reduce the influence of errors in the test and improve accuracy, three samples were selected for each group for testing. Afterwards, a VR 3000 3D contour scanner with high precision was used to scan the fracture surface of the tested specimens. The manual stitching mode was selected for the scanning process. The maximum scanning size was 90 × 160 mm, and the highest resolution was 1 μm.

3 Results

3.1 Tensile Properties of the Cementitious Composites

The mean and standard deviation of the tensile strength of the cementitious composites are shown in Fig. 1. The tensile strength of the plain cement specimen was ≈3.6 MPa. After mixing in FA, the tensile strength of the cementitious specimens was significantly reduced to 3.17 MPa, ≈12% lower than that of the Ref-group. The deterioration in tensile strength was mainly due to the decreased cement powder content, which led to a decrease in the hydration degree of the composite. The hydration in the sample was uneven and incomplete, and there were more microcracks and micropores in the specimens. In contrast, the mean tensile strength of the CNT-group specimen was significantly improved to 3.77 MPa with only mixing 0.08wt% CNTs into cementitious composites, which was 4.6% and 18.9% higher than the Ref-group and FA-group specimens, respectively. The reinforcing rate of the tensile properties of cement-based materials after mixing CNTs was highly consistent with previous studies [9]. According to our previous report [8], by their nucleation and crack bridging effects, CNTs effectively enhance the microstructure of cementitious composites, thereby strengthening their mechanical properties.

Fig. 1
A graph of sigma c. The values are as follows. Ref group 3.7. F A group sample 3.2. C N T group 4. The values are approximate.

Tensile strength of different groups of cementitious composite materials

Full size image

3.2 Analysis of the Mesoscopic Asperities of the Fracture Surface

The fracture method describes the surface of the fracture through three aspects: shape, undulation and roughness [10]. Typically, macro-geometric features are described by shape and relief, and micro-geometric features are described by roughness. The micro-roughness is the degree of unevenness of the asperities attached to the macro-relief surface in a small-scale range, as shown in Fig. 2a. Previous studies have proposed several qualitative and quantitative methods for evaluating the microscopic roughness of rock fracture surfaces [7]. For example, Barton et al. [11] proposed evaluating the roughness of fracture surfaces by visual inspection, comparing and referencing 10 standard joint roughness coefficient curves. Based on their theory, many researchers have further explored and developed related statistical and fractal methods to more accurately characterize the roughness of fracture surfaces [12, 13].

Fig. 2
Two illustrations. a. Concave body, undulation, and roughness. b. The three-dimensional and two-dimensional surfaces are depicted.

Schematic of microscopic fracture surface: a microscopic roughness of fracture; b slope angle of reference surface in different dimensions

Full size image

In this study, the statistical method for the microscopic roughness of the fracture surface of cement-based materials was based on the method proposed by Belem et al. [14]. As demonstrated in Fig. 2b, the 3D fracture point cloud data were discretized into equally spaced points and gridded into a series of fine-scale planes. For each microscopic plane, the local inclination angle (αij) is defined as its angle with the horizontal plane, which is the angle between the normal vector n and z-axis. The αij and the height of each minutiae plane are counted to obtain the maximum, minimum, mean and standard deviation. For microstructure characterization, a 3D contour scanner was employed to image the typical fracture surface of the cementitious specimen after tensile damage to obtain the fracture surface point cloud data. The fracture surface point cloud image was cropped by Geomagic Wrap software, with a cropping size of 40 mm in length and 20 mm in width. The cropped 3D mesh data were exported by PolyWorks software and 3D reconstruction was performed using MATLAB. Representative 3D reconstructed images are exhibited in Fig. 4a–c.

The fracture surface undulation height distribution for the three groups of cementitious specimens after the Brazilian split tests are shown in Fig. 3. It can be seen that the standard deviations of the undulation heights of the Ref-group, FA-group and CNT-group specimens are 0.83, 1.01 and 0.76 mm, respectively. The standard deviation of the undulation height of the FA-group specimens’ fracture surface was ≈21.7% larger than that of the Ref-group, because the incorporation of FA decreased the hydration reaction rate and reduced the hydration products, resulting in more microcracks and micropores in the specimens. The fundamental cause of the failure of the sample was the original microcracks and micropores inside the sample gradually extending outward and deeper, causing the damage. Hence, the more microcracks and micropores inside the FA-group specimens caused a more complex and tortuous extension of the primary fracture. The undulation of the fracture surface was larger, which eventually led to a more significant standard deviation of the undulation height of the fracture surface of the FA-group than that of the Ref-group. With the addition of CNTs, the micropores inside the cementitious composites were optimized due to the nucleation and pore infilling effects [8], and porosity was also reduced. As a result, the extension path of the primary fracture developed from fewer micropores and microcracks became single and smooth, with fewer undulations on the fracture surface. Finally, the standard deviation of the undulation height of the fracture surface of the CNT-group specimens was smaller than that of the Ref- and FA-groups.

Fig. 3
Three illustrations. The standard deviation of the undulation heights of the ref groups, F A, and C N T group specimens are depicted.

Three-dimensional reconstruction images of fracture surface of typical cementitious materials under tensile damage: a Ref-group, b FA-group and c CNT-group

Full size image
Fig. 4
3 graphs of frequency versus slope angle. 1. Minimum 73.71244, maximum 73.0506, mean negative 2.6875, S t D e v 7.5973. 2. Minimum negative 80.5366, maximum 71.9139, mean 0.5596, S t D e v 9.9002. 3. Minimum negative 75.9570, maximum 80.5827, mean 0.0397, S t D e v 11.4152.

Microscopic undulation height distribution of the fracture surface of cement-based samples under tensile failure condition: a Ref-group, b FA-group and c CNT-group

Full size image

The measured results of the microscopic relief angle of the fracture surface are presented in Fig. 4. The standard deviations of the microscopic relief angles of the CNT-group and FA-group specimens were more extensive than those in the Ref-group, indicating that the plane slope changed more in space, and the fracture surface was correspondingly rougher. Cementitious composites continuously absorb energy during the loading process. For the FA-group specimens, after FA replaced part of the cement powder, the internal hydration reaction of the slurry was not uniform, resulting in nonuniform energy distribution inside the loading samples. The crack propagation path was more complicated when the unevenly distributed energy began to be released into the stress concentration area at the tip of the microcrack. For CNT-group specimens, with the incorporation of CNTs, the nucleation effects promoted the hydration reaction, contributing to more and denser hydration products. The integrity of the specimens became higher and could absorb more energy during tensile loading. When the energy accumulated to a specific value, a large amount of energy was released instantaneously, and the stress was concentrated around the fracture surface, making the crack propagation path more complicated. In addition, the bridging and pull-out effects [9] of the CNTs effectively inhibited the development of cracks. More secondary microcracks were generated inside the sample, and the crack propagation path became more complicated, resulting in higher fracture surface roughness (Fig. 5).

Fig. 5
3 graphs of frequency versus slope angle. 1. Minimum negative 73.71244, maximum 73.0506, mean negative 2.6875, S t D e v 7.5973. 2. Minimum negative 80.5366, maximum 71.9139, mean 0.5596, S t D e v 9.9002. 3. Minimum negative 75.9570, maximum 80.5827, mean 0.0397, S t D e v 11.4152.

Microscopic slope angle distribution of the fracture surface of cement-based samples under tensile failure condition: a Ref-group, b FA-group and c CNT-group

Full size image

4 Discussion and Conclusions

In this work, CNTs were used as the additive in cementitious composites to assist FA in reducing cement usage and generating cost-effective, environmentally friendly, high-workability cement-based materials. The tensile properties of three groups of specimens were tested by the Brazilian split test. The fracture surfaces after tensile loading were scanned by a high-precision 3D scanning system. Based on fracture theory, the micro-enhancing mechanism of CNTs on cementitious composites was revealed.

Compared with plain cement, replacing the same cement mass in the slurry with 20 wt% FA resulted in a 12% deterioration in the tensile strength of the cement-based material. Nevertheless, adding only 0.08 wt% of CNTs into the cement–FA hybrid composites significantly increased the tensile strength by 18.9%. Compared with plain cement-based materials, the tensile strength of the hardened matrix was enhanced by 4.6%. This finding indicated that 0.08 wt% CNTs combined with 20 wt% FA could be a good substitute for cement in cementitious composites without affecting its mechanical properties.

The 3D scanning results of the fracture surfaces of the tested specimens demonstrated that the FA-group samples had a higher standard deviation of the macroscopic undulation height than the Ref-group at the macroscopic level. After mixing 20 wt% FA, more microcracks and micropores appeared in the cementitious matrix. As a result, the extension path of the primary fracture became tortuous and complicated in the FA-group samples. By contrast, due to the ultra-high specific surface area and bridging effects of CNTs, the formation of hydration products was promoted, internal microcracks and micropores were effectively reduced, and the regularity and integrity of the samples were also improved. Therefore, under tensile loading, the extension path of the final primary fracture was simple, and the standard deviation of the corresponding macro-fluctuation height was low.

At the microscopic level, energy was continuously absorbed during tensile loading of the specimens. Compared with the Ref-group specimens, the uneven hydration reaction inside the FA-group specimens made the energy distribution inside the sample uneven, resulting in a rougher fracture surface. The addition of CNTs promoted nucleation and had pore filling effects that optimized the pore structure and inhibited the development of microcracks in the cement matrix during loading. The CNT-group specimens could absorb more energy, resulting in more complex stress paths and rougher fracture surfaces. In conclusion, using 0.08 wt% CNTs combined with 20 wt% FA to replace the same cement mass in the slurry will reduce the microcracks and micropores in the hardened matrix, making the change rate of the macro-undulation height smaller. CNT-reinforced cementitious composites can absorb more energy during tensile loading. As the energy is released to the stress concentration zone at the tip of the microcrack, the stress path of the CNT-group specimens became more complicated, with a more extensive change rate of the microscopic inclination angle and rougher fracture surface.