Introduction

In recent years, the problem of carbon emission has been gradually focused on, and low-carbon green building materials have become the focus of research in the industry1. With the increasing demand for building materials, river sand begins to show the phenomenon of running out or even depleting2, and the massive mining of river sand has also brought severe environmental problems such as undercutting riverbeds and loss of sand and gravel3. Furthermore, there is a large number of desert areas around the world, which are formed by weathering of parent rocks; the sand in these areas has a low fineness modulus and high water demand, making it challenging to apply desert sand to cement4. In addition to alleviating the scarcity of river sand, the efficient utilization of desert sand will provide a reference basis for desert management in local areas.

The critical reason for limiting the application of desert sand resources is the low fineness modulus, high water requirement, and concrete bleeding problem. Most studies have concluded that desert sands are finer-grained and play a good role in filling and improving the density of the paste5,6,7,8, which can improve the mechanical properties of concrete in a reasonable mix ratio9. Mixing mineral admixtures10,11 and fiber reinforcement12,13,14 with desert sand concrete (DSC) has also been found to improve performance, which provides a practical reference to overcoming the DSC’s shortcomings. On the other hand, the properties of desert sand are different from river sand; it contains small amounts of alkali metal oxides such as K2O and Na2O, and active elements such as Si, Al, and Ca, which will produce potential hydration products to enhance the concrete properties15. However, the chemical composition and fineness modulus of desert sands are different in different regions16, affecting the universal law of desert sand materials17. Meanwhile, acoustic emission techniques have started to use in the damage detection of concrete, but concrete damage monitoring of compounds with steel fibers and desert sands has been relatively underreported.

According to previous research, it can be known that concrete has poor tensile properties, and in order to increase its ductility, fibers are added to concrete for reinforcement18. According to available research19,20,21,22,23, adding fibers can inhibit the development of microcracks and restrain the tensile strain after cracking; which is an effective way to improve the mechanical (strength and shrinkage) properties of concrete, but researchers24,25 also claim there was a decrease in compressive strength after incorporation of fibers. Meanwhile, fiber distribution influences the strength of concrete26, which is in turn impacted by the matrix rheological properties25,27. Thus, the matrix’s quality and the fibers’ properties (type, volume, aspect ratio, etc.) will impact concrete performance28. As fiber reinforcement has been effective at improving the load-bearing capacity of concrete, it could be a feasible way to enhance the performance of desert sand concrete by adding fiber as well as improving the utilization of desert sand in the local area.

Therefore, considering the convenience of obtaining materials in Xinjiang and the urgent need for environmental management in sandy areas, Taklamakan desert sand is selected in this paper to prepare samples and test the mechanical properties of 14 groups with different desert sand replacement rates (Rds) and steel fiber contents (Vsf). The concrete axial compression process is characterized by the acoustic emission (AE) technique to analyze the fitting of total AE hits to stress levels. As a final step, SEM and XRD were performed to provide analytical verification of the microstructure. This study aims to provide an experimental basis for using desert sand resources as construction materials, which will also serve as a valuable guide for better understanding the characteristics of fiber and desert sand synergy, improving the effective utilization of desert sand resources, and facilitating local use in desert areas.

Materials and methods

Materials

P.O 42.5R portland cement was selected for all concrete mixtures. The desert sand was collected from the Taklamakan desert, with an average particle size of 0.963 mm, and fineness modulus of 0.855; its chemical composition is shown in Table 1. The medium sand in Xinjiang, with a fineness modulus of 2.99 and an apparent density of 2487.5 kg/m3, was chosen as a natural fine aggregate. The coarse aggregate was made of natural pebbles with a continuous gradation of 5–20 mm, and its bulk density is 2700 kg/m3. Figure 1 shows the particle distribution of the aggregates. The hooked steel fibers are produced in Hebei, China, and the specific parameters are shown in Table 2. The retarder is a Q8081 polycarboxylate superplasticizer, and the water reduction rate is ≥ 25%.

Table 1 Chemical composition of materials (mass fraction)/%.
Full size table
Figure 1
figure 1

Particle distribution of coarse and fine aggregate.

Full size image
Table 2 Basic parameters of hooked steel fiber.
Full size table

Preparation of concrete

Based on the previous research29,30,31, this paper focused on the effect of desert sand replacement rate (Rds) and steel fiber content (Vsf) on the performance of desert sand concrete (DSC), fiber content by volume and concrete design strength of C40. With a detailed mix design shown in Table 3, the superplasticizer dosage was adjusted based on fluidity during trial formulation to meet slump requirements. The samples were prepared according to the procedures specified in Chinese standards (GB/T 50081-2019)32 and (CECS 13-2009)33.

Table 3 Mix design of steel fiber reinforced desert sand concrete/(kg m−3).
Full size table

According to Table 3, the water-to-binder ratio (w/b) was fixed at 0.45, the sand-to-aggregate rate was 40%, and the superplasticizer dosage was about 0.5% of the cementing materials. In the mixture, desert sand partially replaces natural sand, and steel fiber is added by volume content. As is shown in Fig. 2, after weighing coarse aggregate, fiber, fine aggregate, and cement, they were mixed for 90 s to ensure uniform dispersion, then retarder and water were added and mixed for another 180 s; finally, the mixtures were poured under vibration into 150 mm * 150 mm * 150 mm cube and 150 mm * 150 mm * 300 mm prismatic molds, and were demolded after keeping 24 h at 90% RH/20 ℃ and final curing for 28 days.

Figure 2
figure 2

Test flowchart of steel fiber desert sand concrete.

Full size image

Methods

After being cured to the age of 28d, samples were removed from the curing room and cleaned of surface debris and water before being tested for mechanical properties. The test parameters include compressive strength, splitting tensile strength, axial compressive strength and stress–strain curves. All steps were performed conforming to the Chinese standards norms (GB/T 50081-2019). The compressive strength of the DSC was tested by the compressive machine (WHY-3000) with a loading speed was 0.06 MPa/s. The axial compressive strength was tested by the compressive machine (306A-3000) with the method of displacement controlling, and the loading speed was 0.1 mm/min; the axial compression process using the acoustic emission acquisition instrument (SAEU2S) for the measurement and acquisition of its internal damage data. In all of test samples, the maximum and minimum values shall not exceed 15% of the average. Figure 3 shows the test schematic and AE probe arrangement. A steel collar is used to fix the prismatic sample, and the axial displacement is measured by a micrometer. Then AE probe and test point contact parts are coated with petroleum jelly and taped after wiping the surface of the samples.

Figure 3
figure 3

Schematic diagram of acoustic emission test.

Full size image

Results and discussions

Mechanical properties

Figure 4 shows the mechanical properties of each group at different Rds and Vsf. Overall, the performance of Rds at 40% is better, and there are slight differences in the performance between samples at 20% and 60% replacement rates, but all their strength has exceeded JZ-0. The addition of steel fibers showed a more significant improvement in splitting tensile strength; this may benefit from the fibers’ bridging effect, which effectively limits the development of inside microcracks during the damage process. The fiber content, however, was not found to improve compressive strength significantly, and Xu et al. also drew similar conclusions in their study34. In the following sections, we will discuss the mechanical properties separately.

Figure 4
figure 4

Mechanical properties of concrete with different Rds and Vsf.

Full size image

Compressive strength

Figure 5 demonstrates the effect of different Rds and Vsf on the compressive strength of DSC. It can be found that the highest strength of all groups was JZ-5 at Vsf = 0.5%/Rds = 40%, and was an increase of 27.6% compared to the JZ-0 compressive strength that was increased from 46 to 58.7 MPa. At the same Vsf, the compressive strength showed a trend of increasing and then decreasing with increased of Rds. In the group at Vsf = 1.0%, only the JZ-2 reached the highest strength of 56.8 MPa, which was increased by 23.5% compared to JZ-0, and other groups reached their peak at Rds = 40%. The results showed that Rds = 40% could present better strength properties, which is consistent with the findings of Lv et al.35. As an extra fine aggregate with a low fineness modulus, desert sand can fill the internal pores of the concrete and improve its compactness. Although desert sand has a large surface area and high water demand, it might absorb part of the water and play a role in dispersing colloids, which may positively impact mechanical properties11. On the other hand, even though the desert sand from the weathered parent rock has a more stable crystalline nature, its composition contains a certain amount of CaO and reactive Al2O3, which still shows some volcanic ash reactivity in the hydration reaction, promoting the generation of the hydration product C–S–H gel and improving the compressive properties of the matrix36. However, the high replacement perhaps reduces concrete workability. To achieve the slump requirement, we may need to increase water content, which may cause water secretion or not be easy to vibrate densely, thereby reducing performance.

Figure 5
figure 5

Effect of different Rds and Vsf on compressive strength.

Full size image

It can be seen from Fig. 5b that considering the effect of Vsf, the compressive strengths of JZ-5 to JZ-8 are 58.7 MPa, 55.6 MPa, 53.8 MPa and 52.9 MPa, respectively; for the same Rds, the compressive strength gradually decreases with Vsf, and the strength was increased by 23.8%, 17.3%, 13.5% and 11.6% respectively compared with 47.4 MPa of JZ-13. However, an anomalous set of data was also observed, where the JZ-2 group showed an increase in strength when the Vsf was increased to 1.0%, then decreased in line with the trend, and the rest of the groups showed a decrease in strength with the increase in fiber content. It can be concluded that at lower Vsf, it is favourable to the compressive strength, but after exceeding 1.0%, the interfacial transition zone (ITZ) with weak mechanical properties will be formed on the surface of the fiber-to-paste and fiber-to-aggregate. Under stress, these parts will crack rapidly and gradually connect to form damaged cracks, resulting in a certain degree of deterioration of compressive strength37.

Splitting tensile strength

As shown in Fig. 6, the splitting strength of the specimens roughly showed an increasing trend in the first and then a decreasing trend for all fiber content, and only for Vsf = 0.5%, the splitting strength was enhanced with the increase of Rds. For all groups of fiber content, when Rds = 20%, the strength was increased by 25.2%, 57.5%, 88.6% and 94.5% compared to JZ-0, respectively, and the overall strength improvement trend increased with Vsf. In terms of the highest strength set of curves (Vsf = 2.0%), as the Rds increased, the splitting strength of the specimens reached 8.19 MPa, 8.92 MPa, and 7.30 MPa successively, and all groups increased at other Vsf levels. The matrix still contributes to the tensile strength of concrete even though its tensile properties are weak. During the test, uniformly distributed tensile stresses are generated in the compressed centerline, and the bearing capacity increases as the strength of the matrix increases. Nevertheless, special data points were observed in which the Vsf increased from 0.5 to 1.5% when Rds = 60% and the error limits of the data were larger; the strengths of the three groups were relatively close to each other without an obvious pattern. This phenomenon is attributed to the ultrafine properties of desert sand. Based on previous studies38, it is considered that desert sand within a certain replacement rate can ensure stable development of the strength of concrete; when Rds is too high (> 50%), the mixture cannot form a uniform particle gradation, which may lead to instability in the strength of the samples.

Figure 6
figure 6

Effect of different Rds and Vsf on splitting tensile strength.

Full size image

Unlike the aforementioned compressive strength law, the enhancing effect of fibers on the splitting tensile strength is very significant, which can effectively act as “micro reinforcement” to disperse the stresses during the tensile cracking process as the steel fibers are distributed in a 3D disorderly manner within the matrix, thus increasing its contribution to the bearing capacity19. For the tensile strength enhancement, after matrix cracking, the fibers can effectively bear the tensile stress and better inhibit the development of microcracks39,40. Some studies have pointed out that the distribution of fibers may be affected by the yield stress, and the concrete performance may decrease due to an inadequate orientation of fibers25. By contrast, the hooked fibers improve the contact resistance between the two parts, and this form of embedding is similar to the anchorage of a reinforcement bar and greatly enhances the integrity of the fibers and the concrete.

Axial compressive strength

The trend of axial compressive strength is similar to that of cubes, but the effect of factors is different due to the change in sample size. Figure 7 shows the results of axial compressive strength, and it is easy to see that with the increase of Rds, the axial compressive strength also shows the trend of increasing first and then decreasing and better mechanical properties when Rds = 40%. The strength of each group reached 42.3–43.8 MPa without much fluctuation when Rds = 20%. When the replacement rate was increased to 40%, the trend of each group showed unconformity, and the strength of JZ-5 at Vsf = 0.5% was only 45.3 MPa, which was 3.4% higher than 43.8 MPa of JZ-1, while with an increase of 15.7% and 16.1% the strength of JZ-7 and JZ-8 were more obviously increased to 49.5 MPa and 49.1 MPa, respectively. When the Rds was increased to 60%, only the strength of the JZ-9 still increased to 47.4 MPa, which was 4.6% higher than that of JZ-5, while a decrease in strength was observed in other groups. The results show that the fibers and the matrix resist stress together during the concrete compression process, but the lower tensile bearing capacity of the concrete during the tension process leads to tiny cracks first; at the same time, the stress is transferred to the fibers, thus limiting the further development of the cracks. This can lead to lower matrix performance and affect the axial compressive strength at higher Rds. Due to the fine particle size and high specific surface area of desert sand, a higher admixture will result in weaker parts of the matrix, decreasing mechanical properties41, which is consistent with the compression strength law. On the other hand, among the different loading methods in the uniaxial compression test, the displacement control, makes the internal damage have enough time to develop. In addition, the "confinement effect" during the loading process constrains the sample along the direction of tensile stress.

Figure 7
figure 7

Effect of different Rds and Vsf on axial compressive strength.

Full size image

The fiber has a negative effect on the axial compressive strength at the high volume dose, but unlike the cube compressive strength law, it shows an increase in the axial compressive strength at Vsf = 1.5%. However, this pattern may be related to the randomness of fiber distribution. Perpendicular distribution of fibers to the compression direction can effectively disperse the tensile stress and limit the further development of microcracks by their bridging action42. It can be seen from Fig. 6b that the highest axial compressive strength was obtained when Rds = 40%/Vsf = 1.5%, followed by Rds = 60% and Rds = 20%. Among the three curves, the 40% replacement rate shows a gradual increase in strength with increasing Vsf, and the strength starts to invert after the Vsf exceeds 1.5%, which is related to the ITZ inside the matrix. The remaining two curves largely show a negative correlation trend, where the strength at Rds = 20% is lower, its strength development is influenced by the concrete, and the compressive strength bearing capacity contribution is mainly based on the cement matrix. At the same time, the fibers mainly play a crack-arresting role and have little contribution in enhancing the compressive bearing capacity43.

Stress–strain relationship

Figure 8 shows the axial compressive stress–strain curve of steel fiber-desert sand concrete. It can be seen that the Vsf does not show an obvious effect on the peak stress at the same Rds, but it improves the concrete ductility and the falling section after the peak stress tends to be moderate. In Fig. 8d, it is observed that the Vsf = 0.5% has an abnormal curve in the descending section. After the curve reaches the peak stress, it drops briefly and then appears as a slight rebound; there is a significant fluctuation in the stress–strain curve; the strain drops steeply when ε reaches 10 × 10–3, and local damage occurs in a very short time with a large amount of skin peeling off the surface and rapid crack occur. This situation indicates the unstable performance of DSC under a high replacement rate, the desert sand cannot play a suitable filling effect after the change of particle gradation, and the matrix compactness has defects. The highest peak stress was found in all samples when Rds = 40%/Vsf = 1.5%, reaching 49.5 MPa; compared with other samples, the average value of peak strain of fiber-reinforcement DSC was 0.0033, which was greater than the peak strain of 0.0026 for Ordinary Portland cement concrete (OPC); the average value of ultimate strain was 0.0052, which was greater than the peak strain 0.0032 (Table 4). The specimens also demonstrated a greater destabilization strain than OPC, and the fibers bridged the stresses and delayed crack damage during crack development, resulting in better ductility. According to the above conclusion, after incorporating steel fibers, DSC has a better deformation capacity than OPC.

Figure 8
figure 8

Stress–strain curves of SFDSC under uniaxial compression.

Full size image
Table 4 Characteristic value statistics of the stress–strain curve.
Full size table

Acoustic emission damage analysis

Damage appearance

Figure 9 shows the splitting tensile test damage; OPC will break into two pieces under ultimate load, while steel fiber can guarantee the concrete's integrity; cracks appear and gradually penetrate under load but will not brake suddenly. Figure 10 shows the axial compressive test damage; as it was shown, fiber improves strength because the tensile properties of steel fiber are much higher than concrete. In the process of the crack appearing, the concrete cracked in tension fast, and then the steel fiber began to bear the tensile stress, forming the crack area near the stress redistribution. So the load continues to increase, and the steel fiber limits the development of the main crack, in its surroundings occur small and dense microcracks after concrete begins to appear skin spalling. As mentioned above, steel fibers improve the tensile properties of concrete and increase their ductility, which is extremely important for structural safety.

Figure 9
figure 9

Photos of concrete splitting tensile strength test.

Full size image
Figure 10
figure 10

Photos of concrete axial compressive strength test.

Full size image

Acoustic emission

In addition to the visual inspection, this paper adopts acoustic emission technology to monitor the process of concrete axial compression. When damage occurs, the acoustic emission event is statistics once exceeding the threshold number, which is one AE hit to reflect the change of damage inside the concrete. Figure 11 shows the time-AE hits statistics and is divided into three stages: initial stage (I), rapid rise stage (II) and stable stage (III). Among these three stages, the samples in the initial stage are still in the stage of online elastic change and elastoplastic change, and the internal stresses are not enormous under small displacements, triggering fewer AE hits44. With the increase of displacement, the sample began to appear extrusion inside to compress the tiny pores and increase the degree of matrix compactness, the internal microcracks gradually began to develop under the action of stress, and the AE hits gradually increased. In the rapid rise stage, the steel fibers began to occur shear deformation, the concrete appeared to happen local crushing or surface spalling, and the damage intensified; at the same time, the AE hits appeared to rise steeply, at which time the sample was in a state close to the ultimate bearing capacity. As the displacement continues to increase, the damage occurs after the sample exceeds the ultimate bearing capacity, the stress level decreases, but the displacement increases rapidly; before the matrix loses bearing capacity and integrity, the AE hits begin to decrease, and the curve tends to level off. As the stress reaches its peak, the total AE hits show rapid growth, and the overall curve almost resembles an "S" shape, while the AE hits roughly follow a normal distribution with time, which aligns with the curve trend for stress–strain relationships.

Figure 11
figure 11

AE hits and time correlation statistics of each group.

Full size image

As shown in Fig. 11, when the Rds increase from 0 to 40%, the pores are easily filled by the extra fine aggregate, and matrix compactness will be improved so that damage is slight. However, when Rds was further increased to 60%, the aggregate gradation was weakened and the internal structure deteriorated, resulting in the growth of total AE hits. Therefore, due to adding fiber, stress is transferred to the fiber during crack development to prevent further damage. Because of the influence of fiber distribution, the rapid rise stage (II) occurs between 500 and 2000 s, and AE hits in this stage will be jointly affected by various factors. The AE event record of crossing the threshold value is random, so individual samples will appear as regular anomalies, but it does not affect the general law.

According to the relationship between AE hits and stress, two types of functions were selected for fitting. Figure 12 shows the fitting results of exponential and cubic polynomials, and it can be found that the cubic polynomial has higher fitting variance and higher accuracy. However, it is also found that as the stress increases, the total AE hits start to show dispersion; this is because that the total AE hits are affected by the mix design and fiber content, and only AE events above the threshold value can be recorded so that the response of the different group to the load varies at higher stresses45. In Fig. 13, the cubic polynomial fitting functions for each group are compared; specifically, it is found that the equations for a single group show a high fitting effect with R2 above 0.96. It can be concluded that the total AE hits show a positive correlation with the stress level, and as the stress continues to increase, the internal damage intensifies, and more AE events are triggered at this time.

Figure 12
figure 12

Fitting curve of total AE hits and stress.

Full size image
Figure 13
figure 13

The fitting curve of total AE hits and stress in each group.

Full size image

Microstructure

Figure 14 shows the XRD of the samples at different Rds. In the DSC, the main silica phase is still the quartz, and it can be seen that the prominent peak of the quartz reaches the highest at 2θ = 26.77. With the increase of Rds, a more obvious change in the peak of quartz is observed, but this change does not have much effect on the hydration products. The silica content of the desert sand is lower than that of the normal sand, so the decrease in the quartz phase mainly results from increased Rds. In addition to the silica phase, the prominent diffraction peaks were plagioclase zeolite, calcite and calcium hydroxide stone, and a certain amount of aragonite was also observed, indicating that the samples had undergone a slight carbonation reaction while being cured to the test age. The DSC may have similar problems in terms of durability as OPC, which needs further study to explain.

Figure 14
figure 14

XRD patterns of different desert sand replacement rates.

Full size image

Figure 15 shows the SEM image of the DSC, and it can be seen that the physical phase structure is consistent with the XRD analysis; needle-like calcite, aragonite and other calcium carbonate phase products can be observed in the microcracks and broken surfaces. The cement hydration products are mainly flocculent C–S–H gel, the main source of matrix strength and somewhat determines the slurry properties. Hexagonal prismatic calcium hydroxide crystals, which make the unhydrated calcium hydroxide, will be a potential durability risk for carbonation and sulfate attack, whether DSC or OPC46.

Figure 15
figure 15

Scanning electron microscopy (SEM) image of desert sand concrete.

Full size image

Conclusions

In this paper, the mechanical properties of desert sand concrete (DSC) with different desert sand replacement rates (Rds) and steel fiber contents (Vsf) were investigated experimentally, and the process of concrete damage under pressure was monitored by the acoustic emission detection (AE) method. Combined with scanning electron microscopy (SEM) and X-ray diffraction analysis (XRD), the microscopic morphology of concrete was observed, which led to the following conclusions:

  1. 1.

    With the increase of Rds, the overall mechanical properties of the concrete show a trend of first increasing and then decreasing, and the reasonable Rds can play a “micro-aggregate effect” to optimize the pore structure and improve the compactness of the matrix, realizing the best performance according to the test results Rds = 40%.

  2. 2.

    Steel fibers have a significant effect on the splitting tensile strength of DSC, which is positively correlated with the Vsf due to the crack resistance effect of the fibers. However, the cement matrix contributes most of the concrete’s compressive strength, while fibers have a minor impact on the compressive strength.

  3. 3.

    After fiber incorporation, steel fiber desert sand concrete (SFDSC) exhibits a higher ductility than Ordinary Portland cement concrete (OPC), and the ultimate strain of concrete is increased, effectively increasing its bearing capacity. The falling section curve of fiber concrete slows down after the peak strain, and the strains at the damage time are all greater than those of OPC.

  4. 4.

    The SFDSC can meet the strength requirement of C40 concrete with the highest 28d compressive strength of 58.7 MPa. The optimum mix ratio is the JZ-7 group (Rds = 40%/Vsf = 1.5%), and the compressive strength, splitting tensile strength, and axial compressive strength is 53.8 MPa, 8.26 MPa and 49.5 MPa, respectively.

  5. 5.

    Acoustic emission can provide a more intuitive description of the damage process. Total AE hits versus time curves reflect the damage process of concrete under stress, and AE thresholds are triggered as stress increases and damage occurs rapidly, with AE events tending to smooth out after peak stress conditions. There is a reliable positive correlation between total AE hits and stress level, the goodness of polynomial fitting can be more than 0.96.