Influence of Recycled Concrete Aggregates on the California Bearing Ratio (CBR) of Granular Sub-bases

Recycled concrete aggregates (RCAs) have a high potential to be used as replacements for natural aggregates in the construction of pavements with similar or even better characteristics. This paper evaluates the influence of RCA as a partial replacement in a granular sub-base of 37.5 mm nominal maximum size (GSB38). Specifically, two replacement percentages of RCA by weight of the coarse aggregate of the GSB38 were evaluated: 10 and 15. To estimate the influence on the mechanical behavior, abrasion tests were carried out in the Los Angeles (LA) abrasion and California bearing ratio (CBR) methods. For GSB38, the average abrasion results were 30.54%. The CBR for the same material at a penetration of 0.254 cm was 26.90, 61.00 and 76.20% and at a penetration of 0.508 cm was 33.00, 65.70 and 82.00% for 10, 25 and 56 blows per layer, respectively. For GSB38 with 10% RCA, similar and even better behavior than natural GSB38 was found in terms of LA abrasion, with an average result of 38.68%. The CBR for the same alternative at a penetration of 0.254 cm was 13.20, 52.80 and 57.70% and at a penetration of 0.508 cm was 18.00, 61.30 and 68.90% for 10, 25 and 56 blows per layer, respectively. But GSB38 with 15% RCA showed an average abrasion of 45.51%, indicating a decrease compared with natural GSB38 and GSB38 with 10% RCA. The CBR for the same alternative also shows a decreasing trend.


Introduction
Road construction is a sector that consumes amount of non-renewable natural resources, mainly due to the large quantities of virgin materials required for construction and the impacts associated with the production of binders such as Portland cement and asphalt [1].About 85% of asphalt is used in the construction of pavements, and its production is linked to significant environmental impacts [2].Construction accounts for approximately 70% of the world's annual production of natural aggregates, resulting in large emissions of various environmental impacts [3].In particular, the road transport sector alone accounts for 74% of greenhouse gas emissions (GHG) [4].
Consequently, the pavement industry has seeking sustainable alternatives to reduce its environmental impacts.One alternative is the use of asphalt mixes produced at lower temperatures [4][5][6], while another is the use of recycled aggregates as a replacement for the entire pavement structure [7][8][9].Recycled concrete aggregates (RCAs) have been extensively studied as a partial and total replacement material in the different layers of the pavement infrastructure [10,11].This recycled material not only reduces the consumption of natural materials but can also be a solution for their final disposal and can represent economic and environmental savings [3].However, its performance is related to its origin, the crushing process and the characteristics of the mortar cement adhered to the surface of the material [8,12].
The use of RCA in granular layers has been analyzed over a wide range of replacement percentages, and similar or even better performance has been found when compared with natural material [8,13].The strength, in terms of California bearing ratio (CBR), has been found to be influenced by the size of the replacement, the origin of the RCA and the amount of replacement.As the RCA content increases, the CBR tends to decrease, and replacement in the fine size may not improve strength [9,14,15].Abrasion in the Los Angles (LA) machine tends to increase with increasing RCA content [9], although improvements in this property have been found when compared with the natural material [8].Table 1 summarizes the LA abrasion and CBR results reported by various studies that have evaluated the use of RCA in granular pavement layers, at different replacements percentages.
According to Kolay and Kumar [14], their study established a decreasing trend in CBR as the RCA content in the mix increased.It should be mentioning that this research evaluated natural aggregate mixes with 20, 40, 60 and 80% RCA [14].For their part, Jayakody, Gallege and Ramanujam [13] evaluated the use of RCA as unbound pavement material and found that the CBR ranged between 90 and 95%.Toka and Olgun [9] in their research evaluated three replacement percentages (25, 50 and 75) and three types of RCA mixed with natural aggregates (NA) to be used as granular sub-base (GSB) and granular base (GB).The trend of their results shows that the higher the compressive strength of the RCA, the better the performance of the mixes; however, this does not follow for all CBR and LA abrasion specimens [9].Specifically, for CBR, RCA shows better results than NA and the higher the RCA content, the higher the CBR on average [9].In LA abrasion, the same behavior is observed, in such a way that the higher the RCA content, the higher the abrasion of the sample [9].Aytekin and Mardani-Aghabaglou [15] conducted a literature review paper and concluded that RCA has an equivalent or better performance than NA.The LA abrasion is reported to be around 30% for blends with 20, 50 and 80% RCA, and CBR varies between 67 and 190% for RCA [15].It is highlighted that the variation in CBR is related to its origin.Finally, in a previous investigation [8], 10 and 15% RCA replacements were evaluated, and it was found that with 15% RCA, the abrasion is higher.Samples with 10% RCA showed the best abrasion resistance, even better than that of the natural material [8].
In view of the above, the studies reported to date have primarily focused on evaluating large RCA replacements.Those evaluating replacements in small proportions are limited.For this reason, it is important to study the CBR and abrasion behavior of a soil with replacement percentages lower than 20% of RCA [9,13,14], to better understand how this artificial recycled material affects the physical and mechanical properties of the NA.With a better knowledge of the research area, better decisions can be made in construction works, considering a possible reduction or increase in economic costs, execution time and environmental impact [2][3][4].
Thus, this research aims to investigate the influence of RCA on the CBR of the GSB.Considering the literature review and the gaps found in the literature, two RCA substitution percentages of 10 and 15% of the total weight of coarse aggregates were proposed.The substitution was carried out on the material exceeding 3/8 and retained at No. 4 sieves.Physical properties such as natural moisture, particle size, plasticity and proctor as well as mechanical properties such as LA abrasion and CBR were evaluated.All the laboratory tests were performed according to the corresponding American Society for Testing and Materials (ASTM) standards and compared with the minimum technical specifications of the GSB according to the Colombian standards.Finally, a brief environmental analysis of the alternatives with RCA was performed and compared to 100% natural GSB38.Atterberg limits and plasticity ASTM D4318 [20] Proctor-standard compaction ASTM D698 [21] LA abrasion ASTM C131 [22] CBR ASTM D1883 [23] 2

Materials and Methodology
Although the research focuses on the influence of RCA on the CBR of a GSB, other physical and mechanical characterization tests were performed.All tests were developed on both the 100% natural GSB and the material with RCA replacements.Two replacement percentages were evaluated: 10 and 15% coarse RCA.The replacement was carried out considering the total weight of the coarse aggregates in the material passing 3/8" and retained in No. 4 sieves.In this way, we want to observe and analyze the behavior of RCA with nominal size equivalent to a gravel, since these could contribute significantly to the mechanical strength of GSB38 [16].
The granular sub-base (GSB) comes from Arroyo de Piedra, a small town in the municipality of Luruaco in the Department of Atlántico, Colombia.This type of soil corresponds to a marine deposit of the shallow to inner marine platform, which constitutes fossiliferous limestones and calcareous mudstones with layers of sandstones [17].
The RCA came from the demolition of a local road located in Barranquilla, Colombia.A laboratory jaw crusher was used for its crushing process.The GSB came from a certified quarry in the area and corresponds to a GSB with a nominal maximum size of 37.5 mm (GSB38).To follow a standard norm and obtain comparable results with other studies, the corresponding ASTM standards were followed for each test [18][19][20][21][22][23].In addition, each property was evaluated with at least three samples of the material and compared to the corresponding minimum specification for GSB38 [24].Table 2 contains the tests performed for each type of material and the respective standard followed.
The analysis was complemented with the calculation of the energy consumption associated with the construction of a 100% natural GSB38, a GSB38 with 10% RCA and a GSB38 with 15% RCA.For this environmental analysis, the PaLATE 2.0 tool was used, and a functional unit was established to compare the results obtained for the base scenario and the alternatives.Figure 1 summarizes the general methodology implemented for the development of the research.

Results and Discussion
According to the proposed methodology, tests were performed on the GSB38 as well as on the GSB38 with 10% and 15% RCA replacement.The results obtained for each test are shown in the following.

Natural Moisture Test
The natural moisture content of three samples of GSB38 was measured according to the ASTM D2216 standard [18].An average natural moisture content of 6.4% was obtained, and the results are shown in Fig. 2.

Granulometry Test
The particle size distribution of the GSB was analyzed in accordance with ASTM D6913 standard [19].Three samples were collected for GSB38, and the results were compared with the limits established for that material (shown as dashed lines) [24], as illustrated in Fig. 3.
The particle size test was carried out using sieves from 1 1/2 to No.200.It was found that all three samples evaluated meet the specifications established for GSB38.Therefore, the

Atterberg Limits and Plasticity Test
For liquid limit (LL) and plastic limit (PL), three samples of GSB38 were tested following ASTM D4318 [20].The results showed an average LL of 24% and PL of 21%, approximately.As a result, the average plasticity index (PI) was determined to be 3%, indicating that GSB38 is slightly plastic.
The technical specification for this property sets a maximum of 25% for LL and 6% for PI [24].It can be concluded that the evaluated samples meet the corresponding specifications.

Characterization of the RCA
The RCA used in this study was obtained from the demolition of a road in Barranquilla-Colombia, and a laboratory jaw crusher was used to crush it.It should be noted that the material underwent a first crushing process to reduce it to sizes that the laboratory crusher could handle.Tests were conducted on the RCA to determine its Atterberg limits and plasticity, specific gravity, density, absorption and LA abrasion.The results are presented in Table 3, and all values are the average of at least three samples.
According to the results shown in Table 3, the values obtained are similar to those reported in other studies.The specific gravity of the RCA has been reported in a range between 2.14 and 2.65 [15].In the case of densities, OD is in a range between 2.20 and 2.40, SSD between 2.31 and 2.68 and apparent between 2.42 and 2.70 [8,15,27].The absorption of the coarse fraction of RCA varies between 1.43 and 8.05% [8,15].The reported LA abrasion shows a range between 27.3 and 39.0% for the RCA [8,15].Finally, the Atterberg limits and PI show a slightly plastic behavior, which is consistent with the literature [8,15].

Proctor-Standard Test
The compaction characteristics were evaluated following the standard Proctor methodology (ASTM D698 [21]) with a compaction energy of 600 kJ/m 3 .Three samples with three different moisture contents were tested, and the results for the materials evaluated are shown in Fig. 4.
As shown in Fig. 4, a maximum dry density of 2.16 g/cm 3 and an optimum moisture content of 7.8% were obtained for 100% GSB38.For GSB38 with 10% RCA, the maximum dry density was 2.10 g/cm 3 and the optimum moisture content was 8.0%.For GSB38 with 15% RCA, the maximum dry density was 1.95 g/cm 3 and the optimum moisture content was 9.5%.The results agree with what was expected, according to the characteristics of the RCA.RCA typically has a higher absorption than natural aggregate [4,8,13].This trend is evident in the results since the higher the RCA content, the higher the optimum moisture content obtained.On the other hand, considering that the density of RCA is lower than that of the natural mineral skeleton [13], it was expected that the dry density of samples with RCA replacements would decrease relative to natural GSB38.

LA Abrasion Test
The abrasion loss resistance was evaluated according to the ASTM C131 [22].Three samples of each material were evaluated, and the results were compared with the minimum specifications for GSB38, according to the corresponding standard [24].
Figure 5 presents the results of the samples evaluated and compared with the maximum limit for GSB38 at low (NT1), medium (NT2) and high (NT3) traffic levels.On average, abrasion loss resistances of 30.54, 38.68 and 45.51% were obtained for natural GSB38, GSB38 with 10% RCA and Fig. 4 Average compaction curves for samples of 100% natural GSB38, 90% GSB38 + 10% RCA and 85% GSB38 + 15% RCA GSB38 with 15% RCA, respectively.All the alternatives, on average, comply with the 50% limit for the three traffic levels.The results are consistent with the literature where abrasion resistances between 24.0 and 33.76% have been established for natural aggregates [8,9,15] and between 25.70 and 46.66% for GSB with RCA [9,15].This decrease in abrasion resistance may be related to the origin of the RCA, the crushing process and the strength of the concrete from which the recycle is sourced [9].In this research, it comes from the demolition of a road with a useful life greater than 20 years and has gone through different crushing processes to reach the required size.Based on the above, it is possible to justify the decrease in the abrasion resistance of the mixes with RCA replacement and the increase in this abrasion as the recycled content increases when compared with natural GBS38.

California Bearing Ratio (CBR) Test
The resistance of the materials was evaluated by considering their CBR at 10, 25 and 56 blows per compaction layer (five layers), following ASTM D1883 [23].The samples were prepared using the optimum moisture content obtained in the corresponding standard proctor compaction test.The results were compared with the minimum specifications for GSB38 according to the standard [24].
Figures 6 and 7 present the penetration versus stress plots to determine the CBR of the 100% natural GSB38 and the GSB38 with 10 and 15% RCA.The plots showing lower stress are those compacted at a lower number of blows.It is worth highlighted that the behaviors of the 100% GSB38 and GSB38 with 10% RCA are similar, with the 100% natural material being better.For the natural GSB38 with a 0.508cm penetration (Fig. 6a), the stress values of 3416.8, 6793.9 and 8476.2 kN/m 2 are obtained for 10, 25 and 56 number of blows, respectively, whereas for a GSB38 with 10% RCA replacement (Fig. 6b), stress values of 1865.6, 6344.0 and 7121.7 kN/m 2 are obtained for the same number of blows, respectively.When comparing the stress values generated for a 0.254-cm penetration, it is found that for the 100% natural GSB38 (Fig. 6a), the stress values are 1857.2,4205.9 and 5253.7 kN/m 2 for 10, 25 and 56 number of blows, respectively, while for GSB38 with 10% RCA (Fig. 6b    The decrease in the effort required to generate a 0.254-cm and 0.508-cm penetration is closely linked to the type and age of the structure from which the RCA originated [8,15].Considering that the RCA used in this research comes from the demolition of a road that had already exceeded 20 years of useful life, a decrease in stress could be expected with the replacement.Based on the above, it is possible to attribute the decrease in stress of the alternatives with RCA replacements to the low strength of the recycled material.This is also the reason for the similar trend in the CBR results shown below. Figure 8 shows the results of the samples evaluated at 10, 25 and 56 blows per compaction layer for the 100% natural GSB38, as well as 10 and 15% RCA mixtures.The results are compared with the minimum limit for GSB38 at low (NT1), medium (NT2) and high (NT3) traffic levels, which require a minimum of 30, 30 and 40% CBR, respectively, according to the standard [24].
The maximum CBR values obtained at 56 blows per layer for a 0.254-cm penetration were 76.20, 57.70 and 15.40% for natural GSB38, GSB38 with 10% RCA and GSB38 with 15% RCA, respectively, while for a 0.508-cm penetration the maximum CBR values were 82.00, 68.90 and 17.80% for natural GSB38, GSB38 with 10% RCA and GSB38 with 15% RCA, respectively.
Comparing the maximum results with those reported in the literature, it can be concluded that they fall within the established range.For the natural GSB, CBR values have been reported to range between 33.42 and 75.14% [9,14], while for GSB with RCA the range is between 13.09 and 83.00% [9,14].
When comparing the results with the minimum specification for GSB38, it can be concluded that addition of 15% RCA significantly affects the CBR of the material, to the point of not meeting with the minimum requirements established for the different traffic levels.

Environmental Results
The energy consumption related to the construction of a 100% natural GSB38 (base scenario), a GSB38 with 10% RCA and a GSB38 with 15% RCA (alternative scenarios) was evaluated.The PaLATE 2.0 tool was used for this environmental analysis, including the production of materials and equipment used for construction.
To establish a functional unit in the proposed scenarios, the usual design and construction conditions for the area were considered [8].The functional unit was defined as the construction of a GSB for a 3.5-m wide lane, 1 km long and 20 cm thick.This indicates a functional unit of 700 m 3 for the base scenario and the two alternative scenarios.
The results presented in Table 4 suggest that the use of RCAs may represent an increase in energy consumption.GSB38 with 10% of RCA shows a percentage difference of 50% in increasing energy consumption compared with the natural GSB38.For GSB38 with 15%, the difference is 75% in increasing energy consumption.RCA production used to be similar to that of the NA [8].However, mechanical or thermo-mechanical processes have been included for the production of higher quality RCA so that their characteristics are similar or superior to those of NA [28].Mechanical processes may include combinations of shredders, and thermo-mechanical processes apply high temperatures and shredders in the processing [28].In this research, the processing of the material was modeled as mechanical, and it has been reported that this type of processing trends to increase energy consumption by three to four times that of conventional [28].Considering the above, it can be concluded that the increase in energy consumption shown in the results of GSB with RCA may be related to the crushing (production) process of the recycled material [8,29].

Summary and Conclusions
This study evaluates the influence of replacing RCA as coarse aggregate in a GSB.The research was conducted in Colombia, using a GSB (from a local quarry) with a nominal maximum size of 37.5 mm-GSB38.Samples of GSB38 with 10% and 15% replacement of coarse RCA were prepared and compared with 100% natural GSB38.Physical and mechanical properties test was performed according to ASTM standards and compared with the minimum required according to Colombian standards.Additionally, a short analysis of the energy consumption related to the construction of a natural GSB38 was performed and compared with the consumption for the GSB38 with 10 and 15% RCA.To make the comparison, a functional unit was established according to the typical design conditions of the study area.
The physical characterization tests show the results in accordance with the material evaluated, with the granulometry and Atterberg limits complying with the maximum values established for GSB38.The LA abrasion results of the natural material and the ones with RCA replacements meet the maximum established for GSB38 on average.The CBR of GSB38 with 10% RCA shows a similar behavior to that of the 100% natural GSB38, while the GSB38 with 15% RCA shows a significant decrease in CBR that does not meet the minimum requirements for different traffic levels.
Comparing the energy consumption generated by the construction of a natural GSB38 versus the GSB38 with replacements, the results show that using RCA may increase energy consumption.GSB38 with 10% and 15% RCA showed an increase of 50% and 75%, respectively, which could be related to the crushing processes of the recycled material.
Future research should explore variations that the age and type of structure where the RCA comes from can generate in the behavior of the material where it is being reused.

Fig. 1
Fig. 1 General methodology adopted in this research

Table 1
Literature results of California bearing ratio (CBR) and Los Angeles (LA) abrasion resistance * Literature review paper

Table 2
Test, standard and material evaluated

Table 3
RCA characterization results