Influence of Size and Content of Recycled Aggregate on Mechanical Properties of Concrete

Abstract

To promote the recycling and reuse of waste concrete, this study investigated the comprehensive impact of recycled aggregate (RA) content and particle size on the mechanical properties of concrete. A novel equivalence parameter (λ_eq) of RA was developed to consider the influence of RA content and size on the mechanical properties of concrete. Empirical equations were developed using linear regression to describe the test results and predict the impact of content and size of RA on the mechanical properties of concrete. The results showed that the comprehensive impact on the mechanical strength of recycled concrete shows a certain regularity when the content and particle size of RA change simultaneously. The measured mechanical properties and regression equations provided a reference and basis for engineering applications, such as the processing of RA in a crushing plant, the design of mix proportions in concrete using RA, and the rapid assessment of mechanical properties on-site. This study provides a design method and technical path for green construction.

1. Introduction

In concrete, coarse aggregate is the component that has the largest proportion of mass and volume. Recycling waste concrete by using it as coarse aggregates (hereinafter referred to as “recycled aggregate”, RA) in new concrete (hereinafter referred to as “recycled aggregate concrete”, RAC) provides an effective approach for utilizing construction and demolition waste, thus helping to protect natural resources [1,2,3,4,5,6,7,8,9]. The high content and physical properties of RA directly affect the mechanical properties of RAC and have been the primary focus of previous studies [10,11,12,13,14]. The compressive strength of RAC decreases as RA content increases [15]. Gandel, Jerabek et al.found that the use of recycled aggregate has a limited impact on the performance of concrete. Although its compressive strength decreases with an increase in replacement degree, its flexural strength remains stable [16]. However, no influence can be observed for replacement ratios up to 30% by weight [17]. Typically, the compressive strength, splitting strength, and elastic modulus of RAC are lower than those of natural aggregate concrete (NAC) by 5–30%, 0–10%, and 10–33%, respectively [18,19,20,21,22,23,24,25]. The percentage loss in the compressive or tensile strength of RAC is more significant when RA is derived from low-strength waste concrete than from high-strength waste concrete [5]. Recycled concrete aggregates were capable of producing higher-strength concrete than recycled aggregates from crushed ceramic bricks [26]. However, the RA obtained from red ceramic exerted a larger influence than the other materials in reducing the RAC elastic modulus due to the lower density of the ceramic. Similarly, a higher air content in the RAC mix may result in lower strength [25]. Various treatments have been used to improve the performance of RA. For example, RA has been treated using a ball mill to remove the old cement paste from the aggregate and by impregnation of a silica fume solution and ultrasonic cleaning [27]. Togay used RA with maximum particle sizes of 7 mm and 12 mm to prepare concrete and studied the effect of RA content and size on the mechanical properties and durability of RAC. The test results showed that the RA particle size is a favorable factor for the elastic modulus, which is unfavorable for the flexural and splitting tensile strengths. Owing to the higher water absorption of the coarser recycled aggregates, mixes prepared with coarser recycled aggregates had a higher drying shrinkage (i.e., up to 49% at 70 days) and water absorption (i.e., up to 30% at 28 days) than those of companion mixes prepared with finer recycled aggregates [28]. Liu, Guo et al.added copper slag (CS), abrasive blast furnace slag (GGBS), and basalt fiber (BF) to recycled aggregate concrete, which effectively improved its compressive strength, splitting tensile strength, and flexural strength [29]. Regarding the mechanical properties of RAC, most previous studies have focused on content, physical properties, treatment of RA, and air content. However, due to the mixed crushing of concrete, bricks, and ceramics at demolition sites, there is a lack of efficient grading and desizing methods. There are a few reports on the influence of the RA particle size. The range (5–40 mm) of coarse aggregate size in concrete is wide, which may affect the physical properties of RA. Although a large number of studies have discussed the influence of recycled aggregate replacement rate or single particle size on strength, a simplified model that can simultaneously quantify the coupling effect of particle size and content has not been established. Therefore, based on the method of Togay and the same distribution curve of coarse aggregates in concrete, in this study, a series of mechanical property tests with a wider range of particle sizes and more groups were conducted to examine the comprehensive influence of the size and content of RA on compressive strength, splitting tensile strength, and uniaxial compressive strength of RAC, to promote the development and use of RA in practical applications of concrete.

2. Materials and Methods

2.1. Specimen Preparation

2.1.1. Raw Material

The cement used in the experiments was ordinary Portland cement produced by the Nanjing Cement Corporation, with a specific area of 340 m²/kg and a fineness of 0.65. According to “The Methods of Cement and Concrete for Highway Engineering” (JTG E30-2005), the measured 28-day compressive strength was 43.5 MPa [30]. River sand, classified as medium sand with a density of 2580 kg/m³, was used as the fine aggregate in the concrete mixtures. Tap water was used as the mixing water. The coarse aggregates used were natural (i.e., virgin) coarse aggregate (NA) and RA derived from waste concrete acquired from a demolition site in Nanjing, PR China. The strength of the parent (waste) concrete was tested and classified as 37.3 MPa. After fragmentation, purgation, and classification, the RA was ready for use in the experiments. The grading curve for RA was determined prior to the production of RAC according to the specification of “construction with pebbles, gravel” (GB/T 14685-2022) [31]. The grading curve of RA was similar to that of NA and fell within the required range of the current Chinese codes, as shown in Figure 1. The particle sizes of NA and RA ranged from 5 to 31.5 mm. The size grouping is shown in

Table 1. Different particle size grouping.

Group	a	b	c
Particle size range	20–31.5 mm	16–20 mm	5–16 mm

.

2.1.2. Mix Proportion Design and Specimen Grouping

To compare the effect of RA size on the mechanical properties of the RAC, three water-cement ratios (0.39, 0.45, and 0.51) were evaluated corresponding to the mix proportion of ordinary concrete (Grade C30) of specification JGJ55-2011(CABR) [32]. The volume of coarse aggregates accounted for the same proportion of the total volume in all samples. Each sample included equal amounts of coarse aggregates from the three size groups (a, b, and c) such that the material from each group comprised 1/3 of the total aggregate. The composition of each coarse aggregate group is shown in

Table 2. Composition of coarse aggregate groups.

Group	Types in Aggregate	The Proportion of a Aggregate %	The Proportion of b Aggregate %	The Proportion of c Aggregate %
N	NA	100	100	00
	RA	0	0	0
0.5A	NA	50	100	100
	RA	50	0	0
0.5B	NA	100	50	100
	RA	0	50	0
0.5C	NA	100	100	50
	RA	0	0	50
A	NA	0	100	100
	RA	100	0	0
B	NA	100	0	100
	RA	0	100	0
C	NA	100	100	0
	RA	0	0	100
AB	NA	50	50	100
	RA	50	50	0
BC	NA	100	50	50
	RA	0	50	50

.

Because the water absorption of RA is much greater than that of NA, RAC requires the addition of more water than conventional concrete to achieve the same workability [1]. Therefore, the amount of water used in the mixing of RAC included a reference amount and an additional amount. The reference amount was the amount required for conventional concrete, and the additional amount was the quantity absorbed by RA that must be added in order to achieve the same workability as that of ordinary concrete.

The composition of the specimens with a water-cement ratio of 0.51 is shown in Figure 2. The letters A, B, and C in (b) refer to the size categories a, b, and c of the polymers.

Due to the porous nature of the recycled aggregate, the premixing process can fill some pores and cracks, thereby producing denser concrete and improving the interface area around the recycled aggregate, thus having higher strength than the traditional mixing method. In this study, a two-stage mixing approach was adopted because Tam et al. (2005) reported that it improves the compressive strength of recycled aggregate concrete and hence lowers its strength variability [33]. First, NA, RA, and sand were poured into a horizontal agitator and dry-stirred for 60 s. Half of the required water was added and mixed for 60 s. Cement was then poured into the agitator and mixed for 30 s. The remaining water was added and mixed for 120 s. When the machine stopped, the slump of fresh concrete was measured. Finally, fresh concrete was poured into steel molds, which were placed on a shaking table to vibrate. Specimens were cured in a standard curing room at a controlled temperature of 20 ± 2 °C and humidity of 95 ± 2% until the testing age was reached. According to GB/T 50081-2002(CABR) [34], as shown in Figure 3, six cube specimens with a side length of 150 mm were used for compressive and splitting tensile tests for each concrete mixture. As shown in Figure 4, the stress-strain relationship and elastic modulus were measured using three prism specimens with dimensions of 150 mm × 150 mm × 300 mm.

2.2. Test Scheme and Test Method

According to the ‘Ordinary Concrete Mechanical Properties Test Method Standard’ (GB/T50081-2002) [34], the compressive strength, splitting tensile strength, elastic modulus, and axial compressive strength of each group of test blocks were tested, and their average values were obtained.

2.2.1. Compressive Strength Tests

As shown in Figure 5. The concrete compressive strength test equipment was a hydraulic universal testing machine produced by Changchun Machinery Co., Ltd., Changchun City, Jilin Province, China. After the specimen was cured for 28 days, it was removed, wiped dry, and the side length was measured and weighed. Subsequently, it was placed at the center of the spherical support of the 3000 kN microcomputer-controlled electro-hydraulic servo universal testing machine and continuously loaded to failure at a constant rate of 0.5 MPa/s. The maximum load Fmax was recorded, and the compressive strength was calculated. Six blocks in each group were averaged after eliminating abnormal values.

2.2.2. Splitting Tensile Strength Test

As shown in Figure 6, the splitting tensile strength test was completed using a cube test block. After 28 days of curing, a 15 mm wide steel pressure bar and a 3 mm thick wooden padding were placed in the center of the upper and lower forming surfaces of the specimen to make the pressure bar coincide with the center line of the specimen. Then, it was loaded to the splitting failure at a rate of 0.05 MPa/s on the same test machine, the load Fmax was read, and the splitting tensile strength was calculated. Six blocks in each group were removed from the abnormal values, and the average value was calculated.

2.2.3. Axial Compressive Strength Test

After leveling the two ends of the specimen with high-strength gypsum, two LVDTs (standard distance of 150 mm) were symmetrically installed on both sides. First, the load was preloaded to 40% of the estimated failure load using the 0.6 MPa/s force control method and cycled three times before being formally loaded. The load-deformation curve was measured in the range of 0.5 MPa to 1/3 of the estimated axial strength, and the elastic modulus Ec was calculated. The specimens were loaded until failure, and the peak load Fmax was recorded to calculate the axial compressive strength. The axial compressive device used for the prism test block is shown in Figure 7.

2.2.4. Elastic Modulus Measurement

After the two ends of the specimen were leveled with high-strength gypsum, they were placed at the center of the bearing plate of the 3000 kN microcomputer-controlled electro-hydraulic servo universal testing machine. Two LVDTs with a stroke of ±2 mm and an accuracy of 0.001 mm were symmetrically installed on both sides of the specimen, and the gauge distance was 150 mm. First, the force control rate of 0.6 MPa/s was preloaded to 40% of the estimated failure load and then unloaded to 0.5 MPa, which was repeated three times to eliminate the system gap. During the fourth loading, the axial compressive strength σa was continuously and uniformly loaded from 0.5 MPa to 1/3 of the estimated axial compressive strength σa, and load and deformation data were collected synchronously. The elastic modulus was the secant modulus of the stress-strain curve in this interval. The stress-strain curve test device is shown in Figure 8.

3. Results and Discussion

3.1. Mechanical Properties and Failure Modes of RAC

The compressive strength, splitting tensile strength, and axial tensile strength of the samples with various sizes and contents of RA are shown in Figure 9. A contrast group for the least RA content and the largest RA size (i.e., “0.5A”), and the greatest RA content and the smallest RA size (i.e., “BC”), were added to (a2), (b2), and (c2) for comparison. Each specimen (e.g., 0.5A) was tested at three different w/c ratios. Because the results for the three w/c ratios were close to each other, for convenience, the mean of the test results is used in this figure. Figure 9a1,b1,c1 show that the compressive strength, splitting tensile strength, and axial tensile strength of the RAC decreased with increasing RA content, as demonstrated elsewhere [10]. The most notable relationship is that for a given content of RA, whether the aggregate size groups of A, B, C in the “N” sample were partially or completely replaced with RA, the compressive strength, splitting tensile strength and axial compressive strength of RAC with large RA size (Group A) were greater than those of concrete using small RA size (group C). Figure 9a1,b1,c1 also shows that when the content of RA increased by one time, except for the individual cases of compressive strength, the strength of the larger RA size was lower than that of the smallest RA size in the lower content, forming a general downward trend. For example, at a water-cement ratio of 0.39 in Figure 9a1, the “A” group had a lower strength than the “0.5C” group. Similarly, the splitting tensile strength of the “A” group was lower than that of the “0.5C” group at various water-cement ratios, as shown in Figure 9b1. As shown in Figure 9a2,b2,c2. The compressive strength of RAC with the largest RA size (Group A) was 11–18% higher than that of the RAC with the smallest RA size (Group C) for a given RA content; likewise, the splitting tensile strength and axial compressive strength were 6–9% and 5–12% higher, respectively. The maximum difference caused by RA particle size did not exceed 20%. However, when both content and size of RA were taken into account, the maximum differences in compressive strength, splitting tensile strength, and axial tensile strength were 55%, 39% and 79%, respectively. Thus, when the content and size of RA changed simultaneously, the mechanical strength of RAC was affected considerably.

The elastic modulus of each sample in the axial compression test is shown in Figure 10. Figure 10a shows the variation trend of the elastic modulus of recycled aggregate concrete (RAC) with the type of specimen (reflecting the particle size and content of recycled aggregate (RA)) under different water-cement ratios (w/c = 0.39,0.45,0.51). Based on the all-natural aggregate specimen (N), its elastic modulus was the highest at each water-cement ratio. In the low RA content stage (0.5A), (0.5B), and (0.5C), the elastic modulus continued to decrease with an increase in RA content and a decrease in particle size. In the stage of high RA content (A), (B), (C), (AB), and (BC)), in the single-particle-size RA, the elastic modulus of large-particle-size RA (A) is higher than that of RA (B) and RA (C), and the elastic modulus of multi-particle mixed RA still decreases with the decrease of particle size. In contrast, the elastic modulus in the axial compression test (Figure 10a) did not show a trend similar to those illustrated in Figure 9a1,b1,c1, because the elastic modulus was affected mainly by the size of RA.

The typical failure patterns of NAC and RAC under compressive, splitting tensile, and uniaxial compression tests are shown in Figure 11. At the initial stage of loading, very few cracks were observed in either NAC or RAC in the mechanical tests on the samples. Figure 11a,b are photographs of cross-sections of a sample, and Figure 11c is a photograph of the side view of the sample. The dashed circle in b2 denotes a broken RA. Differences between the compressive failure patterns of RAC and NAC were not obvious; both materials exhibited failure in an inverted pyramid pattern (Figure 11a1,a2). Some RA was split on the failure surface of RCA under the uniaxial compression test. Figure 11c shows that as the loading stress continued to increase in the uniaxial compression test, the micro-cracks expanded in coverage and extended in length, even forming in the interior of the specimen. The failure behavior of both NAC and RAC was the destruction caused by shearing strength; NAC mainly produced narrow and densely distributed cracks, and the maximum width at failure was less than 0.1 mm. These cracks propagate along the interface between natural aggregate (NA) and cement matrix. RAC forms wide and sparsely distributed cracks during uniaxial compression, and the maximum width of failure is more than 0.2 mm. Accordingly, it can be considered that NAC experienced more uniform stress than RAC, and narrower cracks were formed. The three types of mechanical tests showed that the failure modes and characteristics of RAC containing various sizes of RA were very similar to those of NAC. However, RAC exhibited some broken aggregate in the splitting tensile test, and a few wide cracks developed in the uniaxial compression test. The failure patterns were consistent with the results showing that the three strength behaviors of RAC were inferior to those of NAC.

3.2. Influence of Size and Content of RA on Mechanical Properties of Concrete

The mechanical properties of RAC were studied along with the physical properties of RA, and a regression equation was developed between the mechanical behaviors of RAC and the size and content of RA. The physical appearance and several mechanical properties of NAC and RAC are shown in Figure 12. A large amount of old cement or mortar was attached to the RA, whereas the NA appeared very fresh and clean (Figure 12a,b. This observation was consistent with that of Safiuddin (2013), who reported that RA contained old concrete and hardened cement or mortar attached to the surface of the original coarse aggregate [35]. DRC et al. (1992) reported that the volume percentage of old cement mortar in RA can be as high as 20–30% [36].

The bulk and apparent densities of RA were lower than those of NA, while the water absorption and crush index of RA were significantly higher than those of NA, as shown in (Figure 12c–f. The lower density of RA is due to the fact that the attached cement and mortar have a lower density and higher porosity compared to the crushed stone. Furthermore, as a result of the crushing process, a large number of internal micro-cracks were inevitably generated in the RA, resulting in a significantly higher water absorption and crush index. As the particle size of RA increased, the specific surface area and the content of attached cement and mortar per unit volume decreased; therefore, the apparent density of RA increased, and the water absorption and crush index decreased (Figure 12c–f). In other words, compared with the smaller-sized RA (Group C), the larger-sized RA (Group A) had a lower proportion of attached mortar and a smaller number of micro-cracks. Therefore, the mechanical properties of concrete prepared using large-size RA (Group A) with better physical properties were superior to those of concrete prepared using small-size RA (Group C). This was a complete explanation of the decrease in various mechanical strengths as the size of RA decreased, as shown in Figure 4a and Figure 9a1,b1,c1. Accordingly, the crushing plant was advised to process the waste concrete into large-size RA to obtain higher-strength concrete. If finer aggregate was used, the composition or percentage of other materials was adjusted to improve the strength of RAC.

Although the influence of RA size on the mechanical properties of RAC was clarified, the specimen designations along the x-axis of the left column of the panels in Figure 2 actually integrated two parameters: RA particle size and RA content. According to the experimental design, moving from left to right in Figure 9a1,b1, and c1, as the specimens changed from 0.5A, 0.5B, and 0.5C to A, B, and C, and finally to AB and BC, there was an exact decrease in each strength behavior. To further investigate the influence of RA size and content on the three strength indicators, an “aggregate equivalent” parameter of RA (λ_eq) was developed, which considered the range of RA content and RA size in a sample. The equivalent parameter of RA is formulated as shown in Equation (1).

$${\lambda }_{\mathrm{eq}}={\xi }_{\mathrm{r}}{\displaystyle \frac{{\mathrm{r}}_{av}}{{\mathrm{r}}_{\max }}}+{\xi }_{\mathrm{m}}\left(1-\mathrm{m}\right),$$

(1)

In which, λ_eq represents the equivalent size of RA in the sample, mm; r_av represents the average size of RA in the sample, mm; r_max represents the maximum size of coarse aggregate in the sample, mm; m represents the content of RA in the sample; ξr and ξm are the coefficients used to quantify the weight of the influence of particle size and content of recycled aggregate (RA) on the mechanical properties of concrete, and satisfy ξr + ξm = 1. The physical significance is to clarify the relative contribution ratio of the two to the mechanical properties of concrete in the combined action: ξr represents the proportion of the influence of RA particle size. The larger the value is, the more significant the influence of particle size on the mechanical properties is. For example, ξr = 0.9 is set for the elastic modulus, which means that 90% of the change of elastic modulus is determined by particle size. ξm represents the proportion of the influence of RA content. The larger the value, the more prominent the influence of the content. For example, ξm = 0.8 was set for the compressive strength, splitting tensile strength, and axial compressive strength, that is, 80% of the change in these strength indexes was determined by the content.

Scatter plots of various mechanical properties and the equivalent size of RA (λ_eq) were obtained by substituting λ_eq for the specimen designations in Figure 9 and Figure 10; these are shown in Figure 13. Simple linear regression equations were determined to describe the test results for each water-to-cement ratio, as shown in Figure 13. As mentioned previously, three mechanical strength behaviors were examined (that is ξ_r = 0.2 and ξ_m = 0.8 in Figure 13a–c) as was the elastic modulus, (that is ξ_r = 0.9 and ξ_m = 0.1 in Figure 13d). In other words, using linear regression, we determined that the three mechanical strength indicators were mainly (80%) affected by RA content and less affected (20%) by RA size. In contrast, the modulus of elasticity was less affected (10%) by RA content and significantly affected (90%) by the RA size.

For the mechanical properties of RAC described in Section 3.1, linear regression not only provided a qualitative explanation but also provided a quantitative measure of the influence of RA content and size on the mechanical properties of RAC. As a result, the compressive strength, splitting tensile strength, axial compressive strength, and elastic modulus of recycled concrete at various water-to-cement ratios can be preliminarily predicted using the regressions for a given size and content of RA. Alternatively, the regressions provide a scientific basis for determining the required content and size of RA in a RAC mixture to achieve specific mechanical properties. This approach has useful applications and practical significance in engineering.

It should be noted that ξ_r = 0.9 and ξ_m = 0.1 for the elastic modulus, and a near-perfect linear relationship was observed between the elastic modulus and equivalent particle size (Figure 13d). Because ξr was as large as 0.9 in the elastic modulus regressions, the effect of RA content on the elastic modulus was negligible. Furthermore, the modulus values for various water-to-cement ratios were almost vertically distributed for a given aggregate equivalent size (i.e., RA content) (Figure 13d). This result shows that Equation (1) has no practical significance for describing the elastic modulus as a function of equivalent particle size λeq. The elastic modulus was primarily affected by RA particle size (Figure 11a). For this reason, linear regression analysis was used to evaluate the relationship between the elastic modulus and the average size of RA in the sample. The relationship between the elastic modulus and mean aggregate size was highly linear (Figure 14). Therefore, in terms of the elastic modulus, the regression equations shown in Figure 8 should be used instead of those in Figure 13d for predictions or designs in engineering.

4. Conclusions

The effects of RA size and RA content on the compressive, splitting tensile, and uniaxial compressive mechanical properties, as well as the elastic modulus, of concrete were investigated in this paper. The following conclusions are drawn:

(1)

Consistent with previous research results, the compressive strength, splitting tensile strength, and axial compressive strength of recycled aggregate concrete (RAC) decreased with an increase in recycled aggregate (RA) content. In addition, the mechanical properties of recycled aggregate concrete are typically lower than those of natural aggregate concrete (NAC). The elastic modulus of recycled aggregate concrete is affected by the characteristics of recycled aggregate. Larger recycled aggregate particles can alleviate the decrease in the elastic modulus compared to smaller particles.

(2)

The compressive strength, splitting tensile strength, and axial compressive strength of RAC containing large-size RA were greater than those of RAC containing small-size RA for any given RA content and coarse aggregate grading curves. There was a uniform decrease in the mechanical strength of RAC as the proportion of large-size RA decreased and the proportion of small-size RA increased. The combined effect of the RA size and RA content on the mechanical strength of RAC exhibited a certain regularity.

(3)

The elastic modulus in the axial compression test is mainly affected by the RA size. Crushing plants that produce RA should process waste concrete into large-sized RA to obtain higher-strength concrete. When greater amounts of fine aggregates are used in an RAC mixture, the composition or percentage of other materials must be adjusted to increase the strength of the RAC.

(4)

A novel equivalent particle size parameter (λeq) for RA facilitates the development of simple linear regressions that can describe the mechanical strength of RAC indicators (compressive, splitting tensile, and uniaxial compressive strengths) as a function of RA content and RA size. The elastic modulus of RAC can be described by linear regression as a function of mean RA size. For practical applications similar to those investigated in this study, the regressions can provide estimates of the compressive strength, splitting tensile strength, axial compressive strength, and elastic modulus of RAC for a given combination of RA size and RA content. The regressions provide a scientific basis for determining the content and size of RA required to achieve the desired mechanical properties of RAC. As such, the regressions have a useful application value and practical significance in engineering.

(5)

The proposed equivalent parameter λeq and regression equation can be used to quickly predict the performance of concrete under different RA content and particle size combinations. This study provides direct support for the optimization of the recycled aggregate crushing process, the mix design of recycled aggregate concrete, and the rapid evaluation of mechanical properties in practical engineering.