Introduction

Concrete is reported to be the second most consumed material, with about three tons of concrete used by each person in the world (Zheng et al., 2018). With its significant consumption, a vast amount of demolished waste is also produced from old concrete structures. Demolished waste generated from various construction projects and old structures are major concerns in many countries due to a lack of proper recycling technology and knowledge. Consequently, occupying landfill sites with these demolished wastes also seems to be an issue in some countries. Conversely, the growing demand for modern infrastructures requires a vast amount of aggregates as it is one of the major constituents of modern concrete. Therefore, alternative sources such as recycling the demolished concrete or aggregates like byproducts generated from different sources could play an important role in the smooth production of concrete. This drive has motivated researchers to research various properties of concrete produced from recycled waste materials (Jiang et al., 2020; Li, 2008). The use of recycled materials like aggregates in the production of concrete has numerous benefits in reducing CO2 emission, saving space for landfilling, preserving natural resources, reducing the cost of new concrete, etc. (Nedelikovic et al., 2021). Generally, recycled aggregates are cheaper than natural or virgin aggregates as they abound and often do not meet the requirements in many applications. One of the major challenges of using recycled aggregates is the old cement paste that adheres to their surface. Old cement paste in recycled aggregates increases the water demand and fine aggregate content in concrete, which may affect the mechanical and transport properties of concrete by forming many interfacial transition zones (ITZs) (Nedelikovic et al., 2021). Generally, the ITZs of mortar to aggregate ratio are linearly related to the concrete strength (Wang et al., 2020). However, if proper quality and optimum amount of recycled aggregates are achieved, they can also fulfill the requirements of many applications, including pavements, temporary infrastructures, etc. It has been found that concrete produced with recycled concrete aggregates (RCA) has more mechanical strength than concrete manufactured with natural aggregates when the proper water-to-cement ratio and superplasticizer content are used (Ait et al., 2021).

Recently, ceramic tiles are very popular as decorative materials for floors, walls, etc. However, at their end-of-life, they become waste (Ray et al., 2021). If recycled into aggregate, waste ceramic tiles aggregates (WCTAs) are relatively hard and have a substantial value of specific gravity, slightly lower water absorption, bulk density, and higher voids, crushing value, impact value, and abrasion value than the natural crushed stone (Senthamarai et al., 2005). Additionally, it is flat and has a smooth surface on both faces. Nevertheless, tile aggregates are durable, hard, and highly resistant to biological, chemical, and physical degradation forces (Senthamarai et al., 2005). As a result, the successful implementation of the ideal WCTA content levels in concrete manufacturing could reduce its environmental impact and establish a new recycling industry.

In a study, concrete mixes were prepared for WCTA and conventional aggregates for six different w/c of 0.35–0.60 in steps of 0.05 (Senthamarai et al., 2005). The physical properties of both aggregates were very much similar and comparable. In the case of mechanical properties, at 28 days of testing, it was found that the compressive, splitting tensile and flexural strength of ceramic coarse aggregate concrete only decreased by about 4, 18, and 6%, respectively, when compared to reference concrete made from conventional aggregates. Better performance of concrete with partial replacement of 15, 20, and 25% WCTA was also reported by Medina et al. (2012). For 25% of WCTA replacement, the compressive strength of concrete was found to be about 21, 11, and 7% higher than reference concrete at 7, 28, and 90 days of curing. The strength improvement was attributed to the compact ITZ between the cement paste and WCTA than the conventional aggregate. Conversely, the compressive and split tensile strength of concrete with ceramic tile aggregate reduced as the percentages of tile aggregate increased in the mix (Giridhar et al., 2015). For 20, 40, 60, 80, and 100% replacement of natural aggregate by tile aggregate, the concrete compressive strength was reduced by 1.2, 5.6, 6.4, 7.3, and 13.2%, respectively. Similarly, split tensile strength also reduced from 6.2% to 31.1% for the same replacement level of natural aggregate by tile aggregate. Giridhar et al. (2015) concluded that concrete's optimum ceramic tile aggregate is around 20%. Other researchers also made similar conclusions on the optimum content of ceramic tile in concrete (Ch et al., 2015; Javed et al., 2015; Umapathy et al., 2014).

The durability of concrete through a water absorption test of 100% coarse waste ceramic tile aggregate concrete was also studied by Senthamarai et al. (2005). For w/c ratios of 0.35–0.60 of ceramic concrete, water absorption was reported to be in a range of 3.74–7.21%. For the same w/c ratios, it was 3.1–6.52% in conventional concrete. The results showed that the ceramic aggregate contributed to higher water absorption in concrete. Finally, from the discussion above, it can be said that ceramic tile aggregate is suitable as an alternative to the natural aggregate in concrete. However, the source of ceramic tiles plays an important role in concrete properties. Typically, the smooth surface of ceramic tiles obtained from the plumbing fixtures forms weak bonds between the aggregates and cement paste, and thus lower strength can be achieved.

This study investigated the properties of concrete made from various percentages of WCTA as a substitute for natural coarse aggregate. Concrete mixes were prepared from 0, 10, 20, 30, 50, and 100% of WCTA (by weight of natural coarse aggregates), and properties such as compressive, split tensile strength, water absorption, permeable pore volume, and microstructural analysis were performed. The artificial neural network (ANN) was trained to predict the 28-day compressive strength of concrete containing different percentages of WCTA. Lately, researchers have used ANN to predict different behavior of various types of concrete composites, from the prediction of strength to structural analysis and optimization with reasonably minimal error (Kaveh & Iranmanesh, 1998; Kaveh & Khalegi, 1998; Kaveh & Khavaninzadeh, 2023; Paul et al., 2018). For reliability, the modulus of elasticity and ultrasonic pulse velocity (UPV) of WCTA mixed concrete were also predicted using the formulas provided in different standards and the literature. Finally, successful applications of WCTA in projects like road pavements and runways can reduce the removal of debris cost and the dependency on natural aggregate.

Materials and test methods

Table 1 shows the material compositions for six different concrete mixes fabricated with WCTA as coarse aggregates at six different replacement levels of 0, 10, 20, 30, 50, and 100% by weight of natural coarse aggregate. CEM II/A-M 42.5 N, known as Portland composite cement (PCC) blended with 60% clinker + 40% supplementary materials, was used as the main binder. The maximum coarse aggregate (natural crushed stone and WCTA) size was 19 mm. River sand was used as the fine aggregate, whose fineness modulus was 2.2. The WCTA is shown in Fig. 1a. It was collected from the local landfill sites, washed with normal water to eliminate contamination, and broken down mechanically to achieve the required size. Before mixing the composites, the coarse aggregate was prepared for saturated surface dry (SSD) conditions. Hence, the coarse aggregate was kept in the water before mixing the concrete for at least 24 h. After removing from the water, a wet towel was used to soak the extra water from the surfaces of the aggregate. Note that the SSD was only followed for coarse aggregate, while the fine aggregate was used in air-dry conditions.

Table 1 Material compositions for different concrete mixes
Full size table
Fig. 1
figure 1

a Waste ceramic tiles aggregates (WCTA), b disk sample used for absorption and pores test in the research work

Full size image

Right after mixing, the workability of fresh concrete was measured by slump test according to ASTM C143 (2014). For mechanical testing, cylinder samples with 100 mm diameter and 200 mm length were used for compressive and splitting tensile strength tests per ASTM C39 (2014) and ASTM C496 (1996). Both compression and split tensile tests are important parameters of concrete as they can play an important role in modern designs of concrete structures. These values also depend on many factors, such as paste volume, aggregate-to-paste ratio, curing age, compaction, quality of materials, etc. A minimum of four samples for each mix and age, i.e., a total of 96 samples, were prepared for testing at 7, 14, and 28 days of curing age. All the samples were cured in water at 21 ± 2 °C temperature until the day of testing to confirm a uniform wetting process. Microstructural analysis was also performed using a scanning electron microscopy (SEM) machine for selected samples.

The water absorption, permeable pores or voids and bulk density in percentages of concrete weight tests were performed according to the ASTM C642 (1996). For these tests, 100 mm diameter and 50 mm thickness disk size samples were used, as shown in Fig. 1b. The water absorption, density, and pores are important parameters as the durability of concrete depends on these. However, no direct relation exists between the concrete strength and pores in the samples, as the percentage of pores in concrete depends on various factors, including aggregate grading, water-to-cement ratio, compaction factor, etc. Equations 1–3 were used to calculate the water absorption, pore volume, and density of different concrete mixes investigated in this research.

$${\text{Absorption }}\left( \% \right) \, = \, \left[ {\left( {{\text{m}}_{{2}} - {\text{m}}_{{1}} } \right) \, /{\text{ m}}_{{1}} } \right]{\text{ x 1}}00,$$
(1)
$${\text{Pore space or voids }}\left( \% \right) \, = \, \left[ {\left( {{\text{m}}_{{3}} - {\text{m}}_{{1}} } \right) \, / \, \left( {{\text{m}}_{{3}} - {\text{m}}_{{4}} } \right)} \right]{\text{ x 1}}00,$$
(2)
$${\text{Bulk density }}\left( {{\text{Mg}}/{\text{m}}^{{3}} } \right) \, = \, \left[ {{\text{m}}_{{1}} / \, \left( {{\text{m}}_{{3}} - {\text{m}}_{{4}} } \right)} \right]{\text{ x}}\rho ,$$
(3)

where m1, m2, m3, and m4 (g), represent the sample mass in oven-dried, surface dried in the air after immersion, surface dried in the air after immersion and boiling, and the apparent mass of the sample in water after immersion and boiling, respectively. p denotes the density of water in mg/m3. The sample preparations and testing procedures are broadly discussed in ASTM C642 (1996).

Analysis of variance (ANOVA) single factor test was also performed to check the statistical significance that different percentages of WCTA have on the compressive strength, split tensile strength, water absorption, and pore space in the concrete samples. The probability or p-value was considered 0.05 as the level of significance for the hypothesis test.

For the purpose of ANN, in this research, gene expression programming (GEP) algorithms which are a combination of gene algorithms and genetic programming, were used to predict the compressive strength of the ceramic tile aggregate concrete. Six parameters were chosen as input parameters, while the 28-day compressive strength was the output parameter. The input parameters are the percentage of tile aggregate, water, cement, natural coarse aggregates, fine aggregate, and tile aggregate contents. A total of 87 data were chosen from the published literature (Anderson et al., 2016; Correia et al., 2006; Daniyal & Ahmad, 2015; Goyal et al., 2022; Medina et al., 2012; Umapathy et al., 2014; Varma & Pravalli, 2022; Yasin Mousavi et al., 2020) and from the current study. The data were randomly selected for training and testing set at 70% and 30% ratio. The details of the GEP and ANN used in this study can be found in Hossain et al. (2023).

Results and discussion

Slump of WCTA concrete

Figure 2 reports the slump of freshly mixed concrete of various compositions. The findings show a medium to high slump range when various percentages of WCTA were used instead of natural crushed stone aggregate. As shown in the figure, the concrete slump increases as the WCTA content increases in the mixtures. About 15, 31, 62, 53, and 103% higher concrete slump is observed for the WCTA content of 10, 20, 30, 50, and 100%, respectively. The higher slump can be attributed to the smooth surface of WCTA compared to the natural aggregate. As shown in Fig. 1a, at least one side of the WCTA is smooth, formed during the tile’s production. A smooth surface provides less friction among the particles (i.e., which creates poor cohesion with cement mortar paste). These properties could dramatically reduce the interparticle resistance among the WCTAs and other constituents of freshly mixed concrete. Therefore, it gains a smooth discharge generated by the nominal required energy to overcome frictional stress in the freshly mixed concrete, leading to a remarkably higher slump to the mix containing WCTAs than natural crushed aggregates. This behavior is better noticeable with the increasing percentages of WCTAs in the concrete mixes. Typically, ceramic tile particles have a lower absorption capacity, which could leave some free water (i.e., does not absorb free water from the cement mortar) in the concrete mix and assist in enhancing the interparticle movements in the mix and increase the slump rendered by the more promising ball-bearing effect. On the contrary, the reference concrete (M1) fabricated with natural aggregates are more angular and has rough surfaces, which significantly enhance the frictional stress in the mix and dramatically reduce the workability of concrete, as shown in Fig. 2.

Fig. 2
figure 2

Slump of WCTA mixed concrete

Full size image

Compressive and split tensile strength of WCTA concrete

The compressive strength development in the reference and WCTA concrete at different curing ages is shown in Fig. 3. For the reference concrete (i.e., M1 cast with 100% natural aggregates), 13 and 39% higher strength is recorded at 14 and 28 days, respectively, with reference to the strength at the 7 day. For M2, the differences are 21 and 55%, 15 and 21% for M5, and 6 and 12% for M6, respectively. The outcomes reveal that as the percentage of WCTA content increases in the concrete mixes, the strength development rate at different curing ages reduces. When comparing the strength among the different mixes tested at 28 days, the compressive strength of concrete decreases by about 12, 35, 29, 30, and 52%, respectively, for the inclusion of 10, 20, 30, 50, and 100% WCTA. Similar to the compressive strength, the split tensile strength of the concrete also reduced as the WCTA content increased in the concrete mixes, as reported in Fig. 4. At 28 days, compared to the M1, about 12, 0, 13, 15, and 29% lower strength is found for mix M2, M3, M4, M5, and M6, respectively.

Fig. 3
figure 3

Compressive strength values of different percentages of WCTA mixed concrete tested at different curing days

Full size image
Fig. 4
figure 4

Split tensile strength values of different percentages of WCTA mixed concrete tested at 28 days

Full size image

As shown in Fig. 1a and the visual observation reveal that the WCTAs are relatively flat and flaky. Thus, it causes inhomogeneity in their distribution in the concrete mix, leading to higher voids (i.e., higher porosity, as shown in Sect. 3.4) and permeability. It is known that the higher the porosity and permeability in the concrete, the lower the mechanical strength. On the contrary, the flat and flaky shape of WCTAs is easy to break/damage when subjected to mechanical loading, causing a greater slenderness effect. However, this is dependent on their orientation in the concrete mix.

Moreover, the higher smoothness of the two uncrushed surfaces of WCTAs can create a weak ITZ around the WCTAs and cement mortar (as discussed in Sect. 3.7). In addition, the high brittleness of WCTAs is another disadvantage, which could promote cracks quickly at the ITZ and allow for the distribution of the cracks more effortlessly due to easy detachment from the cement mortar induced by poor bond strength. These phenomena could cause failure at the ITZ (i.e., bond failure) rather than the combined failure (i.e., failure of cement mortar and WCTAs), which is more evident for the concrete mix fabricated with higher percentages of WCTAs. Hence, this behavior prevents the concrete from withstanding higher mechanical load, reducing compressive and tensile strength than natural stone aggregate concrete.

Effect of different waste coarse aggregates on concrete strength

The effect of various types of waste coarse aggregates in concrete strengths reported by the various researchers was also compared with the experimental results. Figure 5 shows the normalized compressive and split tensile strength of concrete made from different percentages of coarse glass aggregates (GA) (Omoding et al., 2021), plastic aggregates (PA) (Mohammadinia et al., 2019; Osei, 2014), crumb rubber aggregates (CR) (Medina et al., 2016), recycled concrete aggregates (RCA) (Grdic et al., 2010), and the experimental results of WCTA concrete. It is worth mentioning that the normalized strength was calculated from the strength of any percentages of aggregate divided by the reference concrete (i.e., concrete made from natural or virgin aggregates). It can be seen from the figure that the strength reduces for all percentages of all types of waste aggregates reported here. The highest reduction of strength is found for PA and CR mixed concrete. The best performance can be found concrete with RCA and GA. Nevertheless, the performance of concrete samples with WCTA is comparable to those made with PA and CR.

Fig. 5
figure 5

Effect of different waste coarse aggregates on concrete a compressive and b split tensile strength

Full size image

Figure 5 shows that the compressive strength dropped sharply up to the replacement level of 20%. Above that replacement, both compressive and tensile strength have maintained the strength of concrete containing 20% WCTAs (i.e., almost no reduction for the concrete fabricated with 30% and 50% WCTAs). The measured compressive strength of concrete containing 30–50% WCTAs is more than 15 MPa, which could be a suitable concrete strength for several low-rise building strength in the rural area of many countries in Southeast Asia, e.g., India, Bangladesh, etc. (Miah et al., 2022a). Hence, WCTA is recommended to use where the compressive design strength is not higher than 15 MPa, reducing the demand for new aggregate, contributing to the circular economy, solving dumping issues, and saving the environment by lowering CO2 emissions to the atmosphere (typically in countries like India and Bangladesh burnt clay bricks are used to produce coarse aggregate, which produces significantly higher CO2 (Miah et al., 2022b; Miah et al., 2021), which also satisfies sustainability.

Water absorption and pores in WCTA concrete

Figures 6a, b shows the percentages of water absorption and permeable pore volume in the different concrete mixes tested. Both absorption and pores significantly increase with the increase of WCTA in concrete. At 28 days, compared to the reference concrete, about 130, 142, 150, 174, and 221% higher absorption are found in concrete with WCTA content of 10, 20, 30, 50, and 100%, respectively. In the case of pores, these increments are 98%, 11, 100, 123, and 124%. The absorption and pores of each concrete mix were also related to their respective 28 days' compressive strength, as shown in Fig. 7. A linear relationship is found for both cases. From the figure, it can be seen that the higher absorption and pores contributed to the lower compressive strength of concrete.

Fig. 6
figure 6

a Water absorption and b pore volume (in percentages of concrete weight) of WCTA mixed concrete after 28 days of curing

Full size image
Fig. 7
figure 7

Relation among the water absorption, pore volume and compressive strength of WCTA mixed concrete after 28 days of curing

Full size image

Prediction of modulus of elasticity and ultrasonic pulse velocity of concrete

For reliability of using the waste coarse ceramic tile aggregate in concrete, the modulus of elasticity and ultrasonic pulse velocity (UPV) were calculated from the compressive strength of concrete as suggested in the different standards and empirical relationships of existing literature, as shown in Fig. 8 (ACI 318-08; ACI 363-R 1992; CSAA23 M94 1994; CEB-FIP 1990; AS-3600-09 2009) and Fig. 9 (Mahure et al., 2011; Nash’t et al., 2005; Ofuyatan et al., 2021; Trtnik et al., 2009). Generally, the modulus of elasticity defines the concrete ability to deform elastically (i.e., stiffness). It is very important for defining the serviceability of concrete structures. Typically, for 20 MPa concrete strength, the expected modulus of elasticity of concrete is around 15 GPa. The aggregate's volume, quality, and water-to-binder ratio can greatly influence the concrete modulus of elasticity (Vakhshouri & Nejadi, 2019). From Fig. 8, it can be seen that at all waste tile replacement levels, the concrete modulus of elasticity is above 15 GPa. Except for CEB-FIP, all other standards can predict the modulus of elasticity of concrete that is not significantly different from each other. Nevertheless, it depicts that incorporating WCTA in concrete can produce the required modulus of elasticity.

Fig. 8
figure 8

Relation between experimental compressive strength to predicted modulus of elasticity of tiles concrete

Full size image
Fig. 9
figure 9

Relation between experimental compressive strength to predicted ultrasonic pulse velocity of tiles concrete

Full size image

In addition, UPV is a non-destructive test method that can define the quality and integrity or homogeneity of the microstructure of concrete elements (Miah et al., 2023). The UPV value of concrete ranges from 3.5 km/s to 4.5 km/s, which is considered to be a good category (Miah et al., 2023). From Fig. 9, it can be seen that the concrete strength above 15 MPa, i.e., the WCTA replacement level up to 50% falls within the good category of UPV. Generally, UPV value increases as the concrete strengths increase. However, the empirical relationship recommended by CEB-FIP (1990) seems an outlier as it shows the opposite behavior than the relationships provided by the other researchers, as shown in Fig. 9.

Prediction of compressive strength using ANN

The regression lines from the comparison of experimental and predicted models for training, testing, validation and all samples are shown in Fig. 10. In all cases, the R2 values are over 0.95, indicating the accuracy of the predicted values, which proves a substantial relationship between training and testing samples. It also emphasizes that the ANN model used here can be used to predict the strength of concrete with different percentages of tile content accurately.

Fig. 10
figure 10

Comparison of predicted and target values of compressive strength (CS) at various testing database

Full size image

ANOVA results of WCTA concrete properties

Table 2 summarizes the ANOVA test results for different groups of WCTA concrete properties. All groups' probability or p-value is greater than the assumed value of 0.05 (for a 95% confidence level). Therefore, it can be said that statistically, different percentages of WCTA have significance in different concrete properties such as compressive, split tensile strength, water absorption, and pore spaces.

Table 2 ANOVA test for different parameters of WCTA mixed concrete
Full size table

Microstructural analysis using SEM

Figure 11 shows the scanning electron microscopic (SEM) images of concrete with 10 and 50% ceramic tile aggregate. As the percentage of tile aggregate increases, the ITZ between the natural and tile aggregate diminishes. Parallel cracks along the ITZ of tiles and natural aggregates are also visible, as shown in Fig. 11a. The weak ITZs formed by the different percentages of tile aggregates may also be the reason for lower mechanical strength, water absorption, and pores of concrete with tile aggregate. It is known that the ITZ of concrete can be enhanced by increasing the cement content and especially by incorporating supplementary cementitious (SCM) materials (e.g., fly ash, slag, silica fume, etc.) in the mix. It is noted that the studied concrete did not contain any SCMs. Therefore, further research is needed to be carried out by introducing the SCMs in the concrete containing WCTA, thus dramatically reducing the pore voids and connectivity left by cement particles and providing dense microstructures (lower porosity and permeability), leading to a robust ITZ and better mechanical strength and durability of the concrete.

Fig. 11
figure 11

Microstructural analysis of concrete with a 10% and b 50% ceramic titles

Full size image

Conclusions

Successful utilization of waste aggregates in concrete protects the environment and human health from different kinds of pollution and creates new business opportunities (Gouasmi et al., 2019). The effect of waste ceramic tile as a coarse aggregate replacement for natural coarse aggregate in concrete on workability, compressive strength, split tensile strength, water absorption, and pores has been investigated. The following conclusions can be made.

  • The workability of concrete increased as the percentage of waste ceramic tile (used as a coarse aggregate) in the concrete mix increased. A minimum of 15% to a maximum of 103% higher slump is found for concrete produced with 10 and 100% ceramic tile aggregate.

  • Both compressive and split tensile strength decreased gradually as the percentage of tile aggregate increased in the mix. Poor bonding between the cement paste and tile aggregate due to the smooth surface of tiles caused lower compressive and split tensile strength in concrete mixes. At 28 days, about 12 and 30% lower compressive strength were observed for the concrete mixes containing 10 and 100% tile aggregate. These reductions were about 11 and 29% for split tensile strength. Therefore, from the results, the optimum tile content in concrete is 10%.

  • Percentages of water absorption and pores in concrete are also increased as the content of the tiles increases. A linear relationship is also found between compressive strength to water absorption and pores. As the absorption and pores in concrete increased, compressive strength decreased gradually.

  • The ANN model can also be used to predict the compressive strength of tile aggregate concrete with good accuracy.

  • ANOVA test proved that statistically, the different percentages of WCTA have significance on the different properties of concrete.

  • Weak ITZs were formed between the tiles and cement paste, contributing to the lower strength of concrete with tile aggregate.

From the research experiment, it is clear that ceramic tile aggregate has a detrimental effect on the engineering properties of concrete. Therefore, applying this type of material in important structural concrete where strength is high may not be a good option. However, in other applications, such as temporary structures, concrete blocks, and pavement works where strength is not only the deciding factor, this type of waste aggregate may be considered for use. An example could be concrete pavement, where higher absorption and voids are important for rainwater percolation (e.g., pervious concrete). As the new aggregate, such as crust stone aggregate, is expensive and not environmentally friendly, this waste ceramic tiles could be an alternative to building low-rise buildings in the rural area where strength requirement is low, which will also reduce the cost of the project, reduce the demand for new aggregate, and reduce the dumping of waste tiles in landfills. Also, utilizing waste tiles in the concrete industry could save several metric tons of CO2 emissions and achieve the goal of sustainability, especially where the burnt clay brick aggregate is used in many structures as a coarse aggregate, e.g., in Bangladesh, India, Pakistan, and eco-friendly.