1 Introduction

Beneficial reuse of waste materials in various civil engineering applications has attracted broad interest, as can be inferred from the amount of research dedicated to the subject in the past few decades. This may attribute to the large amount of waste produced every year and the challenges due to landfill space.

In South Africa and globally, the phenomenon of rapid urbanisation and ageing of some existing infrastructure led to considerable refurbishment and demolition operations. The latter is said to generate massive quantities of Construction and Demolition Waste (C&DW), which results in a high environmental cost if C&DW is inappropriately managed.

According to [1,2,3], C&DW is described as waste that originates from construction, renovation and demolition activities. This may include additional damaged products and materials generated by the construction work and temporary on-site activities. Gauteng Provincial Government [4] issued guidelined in this regard, which identifies concrete, earth, rock and wood as the main elements of C&DW waste, while Yuan [5] named concrete, asphalt, wood, metals, gypsum, wallboard and plastic as materials making the waste. These definitions reveal a common understanding of the source of C&DW (i.e. processes of construction and demolition) and the reasonable agreement in terms of the materials making the waste. The former statement is affirmed by Shen et al. [2], , who stated that C&DW emanates from different sources in the life cycle of the construction projects, from the start to construction and demolition.

There has been some discrepancy in South Africa regarding whether to classify the C&DW as a waste or a resource. The legislation for Environmental Conservative Act (ECA), Act 73 [6], defines waste as unwanted or surplus material, while the National Water Act, Act 36 [7] defines waste as a material with the potential of creating pollution and hazard. In Japan, C&DW is considered a by-product instead of waste and emphasis is placed on recycling [8]. In the authors’ opinion, the classification of C&DW as waste, by-product or resource may depend on how the material is managed or reused.

It is relatively hard to obtain accurate information on the quantities of C&DW in South Africa. The construction industry is said to produce about 5 to 8 million tons of C&DW annually [9], which is significant and indicative of the size of the problem related to this type of waste. It has been suggested that the lack of accuracy in C&DW statistics emanates from (among others) lack of consistency in defining C&DW, absence of a national information centre for C&DW, the intermingling of C&DW constituents, illegal dumping, lack of weighbridges, and no recording of C&DW in some of the dumping sites [10]. Data specific to the amounts of recycled concrete aggregate (RCA), the subject of this paper, is even more challenging to obtain or does not exist. This may attribute to the additional process required to subdivide the C&DW into its constituents, including the RCA.

RCA is an international matter in many states because of a high consumption of limited natural resources by the construction industry. The main alternative for natural aggregates in unbound pavement layers is RCA and recycled crushed clay (RCM). RMC usage is limited and only used when blended with RCA in most countries. Some European specification such as Finland permit a maximum of 30% of RMC blended with RCA and Australia allows a maximum of 20% [11].

The properties and performance comparison of RCA to natural aggregates is well documented in European Union (EU) [12]. The ranges for RCA are: specific gravity 2.05–2.85, flakiness index 6.0–21.0%, Los Angles (LA) abrasion loss 18.0–42%, CBR 19–215%, optimum moisture content (OMC) 8.6–15.8%, and maximum dry density (MDD) 1743.7–2151.6 kg/m3. Whereas for natural aggregates the ranges are: specific gravity 2.42–3.11, flakiness index 8.0–18.0%, LA 13.1–30.1%, CBR 36–170%, OMC 5.2–7.1%, and MDD 1835.5–2375.9 kg/m3.

RCA contain natural aggregates coated with residual mortar (i.e., cement paste) which makes them less dense, more porous and inhomogeneous than natural aggregates. These prorperties bring challenges such as inferior elastic modulus and strength, lower specific gravity and higher water absorption. The use of RCA may therefore have a negative effect on durability. Past research by Kuo et al. [13], suggested treatment strategies to improve properties of RCA such as strengthening residual mortar or removing it. The residual mortar can be strengthened by carbonation which uses carborn dioxide (CO2). In addition, the CO2 treatment was found to reduce water absorption by 10 − 32%, improved porosity by 18 − 21%, increase the specific gravity by 0.47 − 0.56% and reduce the aggregate crushing value by 7.6 − 9.1%. The aggregate treatment and durability fall out of the scope of this study.

Literature review reveals limited information and technical publication on the use of recycled concrete aggregates (RCA) as pavement layers in South Africa. However, the study conducted by Macozoma [14] points out the successes achieved in other countries in terms of RCA use as a base in pavement layers. The research findings reported in the following sections are expected to contribute to existing technical knowledge on reuse of C&DW. Furthermore, the result can perhaps contribute to a future South African specifications or guidelines on using RCA. The absence of local specifications can slow down the wide use of material construction practitioners.

The focus of this study is to investigate the potential reuse of recycled concrete aggregates (RCA) in pavement structure. Results indicate that the material meets the minimum requirements of unbound materials for base and subbase application as per South African specifications, while further testing is required to evaluate the cement-treated properties of the material.

1.1 Objectives

The aim of this study is to investigate the potential use of recycled concrete aggregates (RCA) produced from Portland cement concrete in pavement structures through laboratory testing. The main objectives are to:

  • Compare RCA materials to natural aggregates used for road pavement applications.

  • Evaluate the pertinence of RCA to the minimum aggregate standards prescribed for unbound subbase and /or base layers.

  • Assess the stabilisation requirements to improve the RCA to meet the minimum standards for cement treated base or subbase layers.

1.2 Significance

The findings of this research will assist the construction industry with technical information required to characterise RCA and therefore encourage its incorporation in pavement infrastructure. The information obtained is expected to contribute towards reducing the environmental cost involved in managing the C&DW by consuming large amounts of it in the construction of pavements layers. Furthermore, additional environment-specific benefits can be conceded by lowering the ever-increasing demand on natural aggregates.

2 Experimental Programme

The study followed a comparative approach using natural aggregates (i.e. used as reference material) often used for pavement layers and recycled concrete aggregate (RCA). Comprehensive bench-scale tests were conducted to classify the materials as per the South African road materials specifications.

The RCA was obtained from returned conventional concrete at a typical dry batch Ready-Mix Concrete plant on the Northern side of Bloemfontein in the Free State Province, South Africa. The plant mainly produces 25 MPa conventional concrete. The ready-mix plant generates about 4115 m3 of returned concrete rubble per year which is transported to the aggregates quarry for dumping. RCA aggregate was produced by a jaw crusher and stockpiled thereafter.

The experimental programme is divided into two sets of tests to determine the properties of natural aggregates and RCA intended for use in base or subbase layers of pavement structures. The first set involves tests required to characterise unbound aggregate materials. The second set is devoted to the extra tests required to characterise cement-treated aggregates (i.e. bound aggregate). The COLTO standard is used as a reference specification [15].

3 Results and Discussion

3.1 Unbound Aggregate Materials Characterisation

3.1.1 Particle Size Distribution

Three samples of RCA and one natural aggregates sample were tested as per South African National Standards, SANS 3001- GR2 [16] and SANS 3001 - GR5 [17]. Figure 1 shows the grading curves for natural aggregate, RCA and the grading requirements for G4 materials despite being processed to meet G5 requirements after crushing. The G4 and G5 follow the South African classification system for soil materials.

Fig. 1
figure 1

Particle size distribution for RCA, natural aggregates, and G4 envelopes

Full size image

The particle size distribution calculations yield a coefficient of uniformity (Cu) greater than 4 and a coefficient of curvature (Cc) in the range of 1 to 3 for natural aggregate (Cu = 27 and Cc = 1.44) and RCA (Cu = 29.5 and Cc = 1.57), which indicate a well-graded distribution as per the Unified Soil Classification System (USCS). The Grading Modulus, GM, (P 2.00 mm + P 0.425 mm + P 0.075 mm) / 100, where: P denotes the percentage retained on a particular sieve size) is determined as 2.56 and 2.48 for natural aggregate and RCA, respectively. These values suggest that both materials can form a dense matrix that correlates well with the well-graded particle distribution revealed by the sieve analysis test on the two tested materials. This agrees well with the results obtained by Robinson et al. [18], who concluded that the crushing characteristics of hardened concrete (i.e. the parent material for RCA) are similar to those of natural rock and are not influenced by the strength of original concrete. In addition, Akentuna [19] stated that the gradation of the RCA will be much dependent on the crushing process.

Compacted bulk density (CBD) and loose bulk density (LBD) tests were also conducted, according to the methods of SANS 5845 [20], on the reference natural aggregates and RCA. CBD and LBD values of 1761 and 1574 kg/m3 were measured for natural aggregate, while 1758 and 1568 kg/m3 were measured for RCA aggregate, which suggests similar particle distribution and the likelihood of dense matrix if materials are used for pavement layers.

The COLTO standard [15] recommends a grading envelop (i.e. grading curves for upper and lower limits) for materials classified as G1, G2, G3 are unbound, graded crushed sound rock while G4 is an unbound natural gravel material. These are often used as pavement basecourse materials. It’s worth pointing out that materials G1, G2 and G3 are excluded in this study since they represent high-quality materials produced from rock crushing. Figure 1 shows that natural aggregate and RCA meet the particle size distribution for G4 which is specified by maximum flakiness index of 35% on each of the − 26.5 mm + 19 mm fraction and − 19 mm + 13.2 mm fraction, used as a sub-base layer or filling.

Instead of a grading envelop, the particle size requirements of G5, ubound natural gravel, is specified by the percentage passing 2.0 mm sieve (i.e. 70% ≥ passing 2.0 mm sieve ≥ 20) and the Grading Modulus, GM (i.e. 2.5 ≥ GM ≥ 1.5). The sieve analysis indicated that 30% and 31% of natural aggregates and RCA passed the 2.0 mm sieve. The GM results indicate that natural aggregates fall slightly out of the range specified, whereas RCA meets the minimum requirement.

3.1.2 Atterberg Limits

Tests were conducted in accordance with SANS 3001-GR10 [21] to determine the Atterberg limits. Table 1 shows the results of natural aggregates and RCA samples.

Table 1 Atterberg limits for natural aggregates and RCA
Full size table

The results show that the reference natural aggregates pertain to the G5 specifications but do not meet G4 requirements. On the other hand, the RCA meets the requirements for both G4 and G5 materials.

Plasticity of aggregate materials used for pavement layers is limited, so materials of lower PI are favoured. RCA was non-plastic, possibly due to the presence of cementitious material that reacts with the clay leading to reduced or non-plasticity. In general, besides its influence on strength properties, non-plastic aggregate materials usually suggest insignificant volumetric changes due to moisture variations when the material is used to construct a road layer. Hence, RCA is favourable over natural aggregates in terms of plasticity requirements for a pavement layer.

3.1.3 Modified AASHTO Compaction

In road construction, pavement material is compacted at optimum moisture content (OMC) to get to maximum dry density (MDD), which correlates to the pavement’s performance under in-service loading. Compaction test trials were conducted with natural aggregates and RCA in accordance with the SANS 3001-GR51 [22], method is adopted from the AASHTO standard. The results are summarised in Table 2.

Table 2 MDD and OMC for natural aggregates and RCA
Full size table

COLTO specifications [15] do not provide minimum requirements in terms of MDD or OMC, instead, the ratio between field density and laboratory density is specified. The compliance of RCA to a specific aggregate class can only be verified if the material is compacted in an actual field layer, which falls out of the scope of this study. However, the MDD for RCA is found to be comparable (only about 7% less) to that of the natural aggregate, which is indicative of the material meets the compaction requirements of G5 materials. The lesser MDD can be explained by the fact that RCA, the hardened cement paste coating the aggregates, breaks down during the compaction leading to more finer materials which could influence the final packing of the compacted sample.

The OMC of the RCA is found to be higher by about 12.5% as compared to natural aggregate. This was expected because of the hardened cement paste coating the particles of RCA with high porosity and surface area, which demands more water as compared to solid larger aggregates.

Water absorption tests were also conducted in accordance with the SANS 3001-AG21 [23]. The results indicate comparable water absorption properties of 7.4% and 7.3% for natural aggregate and RCA aggregate. However, these results do not correlate with the findings regarding the OMC. The water absorption results indicate natural aggregate requires more water than RCA which is not reciprocated on OMC. This miss alignment invites for thoughts or further testing.

The results in terms of RCA higher absorption capacity is affirmed by Edil [24] who studied RCA and reclaimed asphalt pavements. Furthermore, the study found MDD in the range of 1900 to 2000 kg/m3 (units in the original text are reported in kN/m3) which reasonably correlates to the results reported in Table 2 above. However, the average OMC of RCA (10%) as reported in [24] is way less than the kind of values shown in Table 2. This discrepancy could be due to the cement paste content of RCA materials tested here, which indicates the variability of RCA generated by different sources of parent concrete.

3.1.4 California Bearing Ratio (CBR)

CBR tests were conducted for natural aggregate and RCA samples to compare the bearing capacity of the two materials and assess the materials’ compliance to the G4 and G5 specifications. The tests were done according to Technical Methods for Highways (TMH) - Method A8 [25], which is a localised South African test method adopted from AASHTO. The CBR results at a percentage Mod. AASHTO MDD ranging from 90 to 100% are reported in Table 3.

Table 3 CBR for natural aggregates and RCA
Full size table

Within the tested range of sample densities (i.e. 90 to 100% of MDD as per Mod. AASHTO), the results indicate that the CBR of RCA is lower than that of the natural aggregates. The lower values obtained for the RCA can be attributed to the higher OMC and due to the presence of cement paste coating individual stones, which cause loss of stiffness when samples undergo compaction. However, both materials were found to meet the minimum CBR requirement for G4 and G5.

Swell readings were taken during the four days while the CBR samples were soaking in water. The average swelling percentages were determined and indicated in Table 4.

Table 4 Swelling properties of natural aggregate and RCA
Full size table

Results show that both RCA and natural aggregate are stable in relation to swelling potential. Moreover, the materials were found to meet the G4 and G5 standards. Contrary to intuition, the RCA sample (non-plastic) have shown a higher swelling potential to natural aggregate (PI = 8). Obtained swelling (0.05) is negligible and way below the maximum allowable for G4 (≤ 0.2) and G5 (≤ 0.5). As speculation, a pavement layer constructed using RCA would possess a better drainage capacity as compared to natural aggregate and therefore yields improved performance. The expected improved drainage performance due to high voids in RCA may be achieved provided voids are interconnected and the grading permits.

A study was conducted by Behring et al. [26], to evaluate the recementation reactivity of RCA fines (passing 75 μm) which may impair the drainage performance. The evaluation encompassed compressive strength test and petrographic assessment to ascertain the recementation reactivity of RCA fines. The compressive strength test results indicated that the recementation reactivity effect was modest. The chemical composition and microstructure results from the petrographic test indicated that RCA fines recementation reactivity was negligible.

3.1.5 Chemical Tests

Tests were conducted on the natural aggregate and RCA samples to assess the clay content, water soluble chloride content, sulphate content and organic impurities that may present.

Methylene blue tests were performed to determine the presence of clay in natural aggregates and RCA. The tests were done according to SANS 6243 [27]. Water-soluble chlorides tests were conducted according to SANS 5831 [28]. Sulphates content was determined according using SANS 5850-2 [29] method. Organic impurities in the aggregates were determined according to SANS 5832 [30]. Table 5 shows the results obtained.

Table 5 Chemical tests on natural aggregates and RCA
Full size table

The result of the Methylene Blue content for natural aggregate and RCA were way below the threshold specified by the SANS 1083 (i.e. 0.7 g per 100 g of the aggregate) [31]. Therefore, both materials did not have excessive active clay. The presence of active clay in aggregates is unfavoured as it results in deterioration of the pavement as a result of expansion and shrinking due to a change in moisture content. The results also indicate that RCA has less clay as compared to natural aggregate, which confirm the findings in terms of plasticity index discussed earlier under Atterberg limits. The clay content of RCA is expected to depend on how the parent concrete is stored and processed to the crusher.

The water-soluble chloride content for natural aggregate and RCA were 0.001 and 0.002, respectively. These values are significantly below the permissible chloride content allowed for aggregate and, therefore, will not affect pavement layers. The presence of excessive chloride in the aggregate can lead to damage of the overlaying bituminous layer. It’s worth noting that the presence of chloride in the RCA is not only due to the contents naturally existing in the aggregates but also due to some additives, such as plasticisers, sometimes used in the production of Portland cement concrete, the parent material for RCA.

The sulphate content for natural aggregate was 0.058% as compared to 0.62% for RCA, which is well below the values usually acceptable for aggregates used in Portland cement concrete. The sulphate content of RCA aggregates was approximately ten (10) times high than that of natural aggregates due to the presence of cementitious material in the RCA. The presence of high sulphate content can lead to swelling and expansion, which have an adverse effect on the performance of a pavement layer.

As far as organic impurities are considered, both natural aggregate and RCA were found to have no organic impurities. Organic matter such as vegetation has detrimental effects on road pavements, primarily when it decays; it leads to the formation of voids which negatively affect the load-bearing capacity of pavement. It must be stressed that the impurities on RCA will mostly depend on the storage conditions of the parent concrete.

3.2 Bound Aggregate Materials Characterisation

Cement type Cem II/B-M (V-S) 32.5 N, which conforms to SANS 50,197 specifications [32], was used for stabilisation of natural aggregates and RCA with the aim to improve the materials’ properties A series of pH tests were performed to determine the initial consumption of cement (ICC) for both materials according to SANS 5854 [33]. The test results indicated ICC of 3% and 2% for natural aggregates and RCA, respectively. The lesser ICC found for RCA is attributed to the presence of un-hydrated cement in the aggregate material.

As per COLTO [15], materials designated C3 and C4 are obtained by cement treating unbound materials G5 and G6, respectively. Therefore, the standard specifications for C3 and C4 are used as a reference. Additional Modified AASHTO compaction tests were conducted, following the method discussed in Sect. 3.1.3 of this paper, to quantify the MDD and OMC for natural aggregate and RCA mixtures containing varying percentages of cement.

3.2.1 Unconfined Compressive Strength

Unconfined compressive strength (UCS) tests were carried out to establish the bearing capacity of cement stabilised pavement layer. The test was done according to SANS 3001-GR53 [34] and the samples were prepared and cured according to SANS 3001-GR50 [35]. Samples of natural aggregate and RCA, with minimum diameter of 38 mm, a height of 76 mm, the largest particle size not exceeding 1/8 of the specimen diameter and moisture content of 15%, were generated by adding 2%, 2.5% and 3% of cement and compacted to 100% Mod. AASHTO density. The test results are indicated in Table 6.

Table 6 UCS results of cement-treated natural aggregate and RCA
Full size table

The UCS results indicate that the cement-treated natural aggregate significantly exceeds the C3 (cement treated G5) and C4 (cement treated G6) specifications at cement dosage of ≥ 2%. On the other hand, RCA meets the C4 requirements at cement content of 2% or more while it meets the C3 requirements only at a cement content of ≥ 2.5%.

3.2.2 Indirect Tensile Strength

Indirect tensile strength (ITS) test was done to determine the performance and durability of a pavement layer when stabilised. The test was done according to SANS 3001-GR54 [36]. Samples of natural aggregate and RCA were prepared by adding 2%, 2.5% and 3% of cement and compacted to 100% Mod. AASHTO density. The test results are indicated in Table 7.

Table 7 ITS results of cement-treated natural aggregates and RCA
Full size table

Table 7 indicates that cement-treated natural aggregates with ≥ 2% cement content meet the ITS requirement of C4 and with ≥ 2.5 meet the C3 specifications. However, RCA at cement contents up to 3% does not meet the COLTO requirement for ITS. The ITS increases for both materials with the increase in cement quantity as expected, which indicates that ITS minimum requirement can be achieved with a further increase in cement content (i.e. >3%).

With 3% cement, the ITS strength gap is approximately 19% to meet the C4 requirements. The trend shown in Fig. 2, though with few data results to extrapolate from, reveals that the ITS minimum requirement of 0.2 MPa can be achieved at approximately 3.4% cement content. Further research is required to explore the effect of cement content ≥ 3.5%. However, care should be taken to avoid block cracks usually associated with cement-treated pavement layers containing a relatively high cement dosage. The actual amount of cement required to achieve the desired UCS and ITS and mitigation of block cracking formation risk can be a subject for further study.

Fig. 2
figure 2

The relationship between ITS and cement content for RCA

Full size image

4 Conclusions

  • RCA is found to be comparable to natural aggregate of G5 class and meet the minimum standards of G4 and G5 materials as per the South African road pavement material classification. Subsequently, RCA is applicable to unbound base or subbase layers.

  • The use of RCA in other pavement layers, such as selected subgrade and pavement fills, which requires aggregate materials that is inferior in quality to G4, should not be prevented if such a use is economically justifiable.

  • RCA treated with 2.5% cement were found to meet the minimum UCS requirements for C3 and C4, while it did not meet either C3 or C4 requirements in terms of ITS. However, the potential exists that the C4 standards can be achieved at 3.5% cement. The latter is a subject of further investigation.

5 Recommendations

  • The findings of this study suggest that road pavement structure can provide leeway in terms of C&DW management, hence, alleviating a portion of the environmental cost associated with material disposal.

  • The RCA results obtained herein relate to a parent concrete of approximately 25 MPa cube compressive strength. Future research is required to assess the properties of RCA obtained from parent concretes that have a wide range of strengths. Also, further testing is required to assess the RCA resilient modulus to generate design values that can be utilised in case the mechanistic approach is adopted.

  • Further investigation is recommended to assess the durability properties of unbound RCA basecourse and the risk of block cracking in cement treated RCA.

  • Trial sections using RCA as pavement layers are recommended to provide more data in terms of the material performance under traffic loading.