1. Introduction
Demand of concrete has boomed up to 30 billion ton/year in the construction industry [
1]. Currently there is great interest in reducing the life cycle cost (LCC) of concrete. This can be achieved through the use of lightweight concrete, thus obtaining fewer material quantities for the building elements due to the lower dead weight of the concrete. The lightweight concrete has a number of advantages, including less unit weight, compressive strength ranging from 7 to 25 MPa, good insulation properties along with low handling and transportation cost, which ultimately reduces the construction cost. One of the approaches to produce lightweight concrete is by using lightweight aggregates, such as lightweight aggregates formed from plastic waste products.
On the other hand, the continuous production and use of plastic products have resulted in huge quantities of plastic (i.e., 6300 Mt in 2017) [
2], accumulated in stockpiles (only 9% recycled) causing negative health, environment and economic impacts. Beside plastic recycling, attempts were made to encourage the use of plastic waste in various application, such as in aggregate and concrete production. Incorporating plastic waste (i.e., polyethylene terephthalate—PET) either directly or indirectly as (coarse and fine) aggregate in concrete production not only reduces the burden on landfill dumping sites, but also aids economical concrete construction. The life cycle cost of plastic-based aggregate concrete is significantly reduced compared to that of normal concrete.
Previously, natural coarse aggregate (CA) and fine aggregate (FA) were replaced by light weight plastic aggregates for production of lightweight concrete (LWC). For instance, LWC produced by replacing 5% to 75% FA with shredded PET bottles [
3,
4,
5,
6,
7]. Similarly, CA and FA were replaced from 5% to 15% by shredded PET plastic for LWC production [
8,
9,
10,
11].
Besides decreasing the weight of the concrete, plastic aggregates significantly reduce the slump of concrete. For instance, a 95% concrete slump reduction was noticed by replacing 20% FA with plastic waste [
12]; while another study [
13] reported a decrease in the slump varied that from 42% to 73% at 20% replacement of FA with PET. This linked to the friction of the non-uniform and high surface area of shaped shredded plastic [
6,
12,
13,
14]. On the other hand, synthetic aggregates, for example, waste plastic lightweight aggregate (WPLA), having spherical shape, smooth texture and less water absorption properties increased the workability of concrete in the case of replacing FA with WPLA [
15,
16].
Moreover, fresh density of LWC containing plastic particles is governed by plastic proportions and type. For instance, 80% and 55% replacement of CA and total aggregates with expanded polystyrene (EPS) reduced the fresh density by 40% and 64%, respectively, [
17,
18]. Similarly, replacing 20% of FA with plastic waste reduced fresh density by 9% [
12]. The reduction in fresh density is attributed to the light weight of plastic as compared to natural aggregates [
6,
8,
10,
11,
13,
14,
19]. Likewise, the dry density of LWC produced by replacing 40% to 80% CA with plastic waste reduced dry density from 11% to 68% [
18,
20,
21,
22,
23,
24,
25]. As similarly reported in fresh density, the reduction in dry density is governed by the light weight of plastic, so the dry density decreases when plastic waste increases [
15,
16,
26,
27,
28].
The compressive strength of plastic-based lightweight concrete decrease significantly due to replacement of high compressive strength conventional CA and FA with plastic aggregates. For instance, 62% to 82% reduction in compressive strength was noticed by replacing 15% to 80% of CA with plastic (i.e., EPS, high density polyethylene-HDPE, PET, polyurethane—PUR foam and ethylene vinyl acetate—EVA) [
10,
18,
19,
20,
22,
23,
24,
25]. Similarly, 75% replacement of FA with plastic (i.e., WPLA) reduced compressive strength by 33%, 33% and 31% at water cement ratio W/C of 0.45, 0.49 and 0.53, respectively [
15,
16]. In case of using synthetic plastic-based aggregate, 100% replacement of CA with synthetic plastic (i.e., SLA) caused 43% to 72% reduction of compressive strength, as compared to control samples [
26,
28,
29]. The prime factor of reducing compressive strength is the weak interfacial transition zone (ITZ) between cement and the plastic particles [
5,
6,
13,
14,
20,
24,
30]. Moreover, the hydrophobic nature of plastic hinders the hydration process of LWC, resulting in compressive strength reduction [
13,
30].
In addition, previous studies reported a reduction in splitting tensile strength ranging from 26% to 38%, and from 37% to 40%, by replacing CA at 100% and FA at 75% with plastic-based aggregate (i.e., SLA or WPLA), respectively. Similarly, a reduction in flexural strength that varied between 50% to 59% was reported previously at different replacement levels, which ranged from 15% to 50% of CA with plastic particles [
10,
23]. The reduction in splitting and flexural strength can be explained by the differences in physical and mechanical properties, such as shape, texture and strength, of plastic aggregate and that of conventional aggregate [
6,
10,
30,
31]. In addition, to the weak bonding between plastic particles and cement paste, similarly as reported in compressive strength [
9,
30].
In the study conducted by Marzouk et al. [
32], scanning electron microscope (SEM) images of LWC produced by 30%, 50% and 100% replacement of FA with PET were presented. The study results showed that the composites were highly compacted up to 50% replacement, while it was porous at 100% replacement levels. The internal structure variations (i.e., high porosity) is linked with lower strength characteristics and bulk density of the mix [
32].
On the other hand, the durability and durability-related properties for the concrete made with plastic has received a little attention. According to Kou et al. (2009) [
30], the chloride permeability decreased by 36% and 43% at 45% and 37% replacement of FA and TA with polyvinyl chloride (PVC) and EPS, respectively.
Plastic-based lightweight aggregate concrete is predominately produced by direct incorporation; however, indirect replacement, where synthetic aggregate was produced prior to its implantation in concrete, is less focused. The synthetic aggregates, such as polyethylene terephthalate (PET) and fillers (i.e., fly ash), are less covered [
15,
16,
26,
28,
29,
33,
34,
35,
36]. For instance, PET with granulated blast furnace slag (GBFS) and river sand powder were used to develop the WPLA [
15,
16]. The authors reported that the WPLA particles possessed a rounded shape and smooth surface texture. Moreover, the produced WPLA had density and water absorption of 1390 kg/m
3 and 0%, respectively [
15,
16].
Furthermore, lightweight aggregates were produced by melting mixed plastic comprised of 85% PET and 15% polypropylene (PP) [
37]. The produced aggregate had a smooth surface texture with angular particle shape. Additionally, the density and water absorption of the produced aggregate was 1.2 g/cm
3 and 0% [
37]. In another study [
38], the synthetic aggregate (i.e., fly ash aggregate—FAA) was successfully produced by sintering fly ash. The particles of produced aggregate possess angularity in their shape and roughness in their surface texture. The water absorption of FAA was 60% lower than that of control mix [
38].
The effect of incorporating shredded or virgin plastic, as it is in concrete, has extensively been studied over the last few decades. However, less research work has been carried out to use plastic-produced synthetic aggregate in concrete. Thus, the current study focuses on the formulation of a plastic-based green lightweight aggregate (PGLA) manufactured using PET plastic waste and mineral additives. In addition, a green lightweight aggregate concrete (GLAC), comprising 100% manufactured synthetic plastic aggregates as a replacement of normal weight and lightweight aggregate, was also produced. Comparison of physical and mechanical properties for the produced green aggregates with the reference one was also made. Likewise, all the major properties including durability properties of GLACs were also examined and equated to the concrete made with conventional normal weight and lightweight aggregates.
3. Plastic-Based Green Lightweight Aggregate Production
Due to the fact that not all plastic can be recycled, along with the differences in the physical, thermal and microstructure properties of plastic such as melting temperature [
50,
51], a polyethylene terephthalate (PET) was chosen amongst various recycled plastics, based on its bulk availability in the plastic production and waste stream, as extensively reported in the literature. The plastic polyethylene terephthalate (PET) used was supplied by a local recycling supplier and shredded to the desired size (see
Figure 4) for the production of PGLA. The unit weight, water absorption and melting temperature of the PET used were 838 kg/m
3, 0.10–0.20% and 285 °C, respectively.
Similarly, three types of additives (see
Figure 5), namely dune dust additive (DDA), fly ash additive (FAA) and quarry dust additive (QDA), were selected according to their local availability and based on their similarity with the major components in concrete.
The specific gravity, absorption and unit weight of DDA and QDA were examined according to ASTM C128 and ASTM C29. Whereas ASTM C618 was followed to analyze specific gravity of FAA. The specific gravity of FAA, QDA and DDA was 2.3, 2.71 and 2.62, respectively. While the water absorption of QDA and DDA was 1.52% and 0.28%, respectively. The particle size distribution of additives was performed using the laser particle size analyzer. The median size of FAA, QDA and DDA was 6.14, 19.27 and 186.37 µm, respectively, which revealed that FAA was the finest, while DDA was the coarsest among these additives.
The compression press was used in the manufacturing process of the plastic-based green aggregates. A homogenized mixture was formed, compressed, heated, melted and then cooled before grinding to give the final shape of the produced aggregates. The steps used in the production of these aggregates have already been patented under [
52]. Applying the before mentioned method, three different aggregate types were produced using three different categories of additives (i.e., dune dust additive (DDA), flay ash additive (FAA) or quarry dust additive (QDA) with a ratio of 30:70 (plastic: additives) as shown in
Table 4.
Figure 6 shows the produced PGLA and it was observed that DDA, FAA and QDA produced brown, grey and yellow PGLA, respectively. The difference in the colour was principally due to the different colour of the additives added during the manufacturing of the different aggregate series.
In the subsequent sections, the physical, mechanical and morphology properties of the manufactured PGLA aggregates will be presented and also a comparison will be made with the NCA, VLA and LYA, along with similar studies conducted to produce plastic-based aggregates.
4. Plastic-Based Green Lightweight Aggregate Investigation
The developed PGLAs were investigated in terms of their physical, mechanical and morphological properties. The summary of the findings of these tests are presented in
Table 5.
The optical microscope was used to evaluate the shape and surface textures of PGLA particles. The shape of the PGLAs developed with DDA, FAA and QDA was more angular with sharp edges. This can be explained by the high degree of stiffness of PGLA, since it breaks in the crushing phase of aggregate production, rather than shredding, as reported in similar plastic-based aggregate produced in the literature.
In addition, the surface textures of the PGLAs (
Figure 6) varied from partially smooth to rough depending on the type of additives used. In general, the texture of PGLAs was smoother than those of plastic-based aggregate reported in the previous studies. This observation can be linked to the behavior of the PGLAs during the crushing, since other plastic-based aggregates were shredded instead of crushed, as mentioned in previous studies.
Therefore, it can be concluded that the PGLAs are comparable to that of NCA in terms of particle shape and texture. However, the PGLAs were entirely different to VLA and LYA. It was noticed that the additives type mainly governed the shape and texture of the plastic-based green lightweight aggregate. Furthermore, it was also observed that shape and texture of the SPGA series were not identical to each other, as well as that reported in the previous studies [
15,
16,
26]. The differences in the experimental procedures and raw materials used for the manufacturing of PGLAs series, as compared to the previous studies, are mainly responsible for the variation in the shape and texture of produced aggregates.
The voids between particles of aggregates are controlled by the size and shape of the particles. For example, poorly graded aggregate exhibits a higher void percentage [
53]. Consequently, more cement paste is needed, as the amount of cement paste influences by the percentage of voids [
54]. Therefore, sieve analysis was conducted for PGLAs to develop the particle size distribution curves plotted in
Figure 7, together with that for VLA and the lightweight standard limits. The particle size distribution curves of the PGLAs were also compared with those of NCA and the normal weight standard limits, as shown in
Figure 8. As it can be seen in these figures, the nominal maximum size of the PGLAs, VLA and NCA is 10 mm.
As it can be seen in
Figure 7, the tested aggregates were classified into the following categories as per the findings of the sieve analysis:
Category 1—contains aggregates that diverged from the maximum limits of [
42] for the lower sieve size (#4), as seen in PGLA2 and PGLA3. These aggregates deviated by 19% and 8%, respectively.
Category 2—which diverged from the minimum limits of [
42] for the upper sieve (#3/8). For instance, PGLA1 slightly deviated by 3.5%, while VLA significantly deviated by 39%.
However, comparing PGLAs with NCA (see
Figure 8) revealed that the grading of PGLAs is poor compared with that of NCA and it does not fulfil the minimum and maximum grading requirements for the normal weight coarse aggregate. This trend is expected as the developed aggregate is lightweight not a normal weight.
These results showed that the PGLAs made with DDA were the coarsest, followed by those PGLAs made with QDA and, finally, the finest were the PGLAs made with FAA. This is accredited to the sizes of the additive particles themselves, since the median size of DDA was the largest, whereas FAA had the smallest particles size among all the additives. The void percentage of the PGLAs (see
Table 5) was 13% to 34%, 34% to 50% and 15% to 36% lower than that of NCA, VLA and LYA, respectively. Additionally, the percentage of voids in the PGLAs is lower as compared to all reference aggregates, which proves that the aggregates’ particles in PGLAs have a better angularity, as reported earlier.
Moreover,
Table 5 showed no major impact of PGLAs on the fineness modulus values. Findings show a reduction in fineness modulus ranging between 1.5% and 13%, as compared to NCA and VLA; with exception of PGLA1, which showed a marginal increase by 0.7%. Moreover, the reference lightweight aggregates (i.e., VLA) showed the highest fineness modulus as compared to all the aggregate types. These results indicate that PGLA1 (i.e., containing DDA) was the coarsest, while PGLA2 (i.e., containing DDA) was the finest. This is associated with the large size of the DDA particles in contrast with the sizes of FAA.
The specific gravity or density is a crucial property, as it is used to determine the aggregates’ volume in the mix. Additionally, controlling the density of concrete as aggregate represents almost 70% of the total volume of concrete’s components.
Table 5 presents the bulk specific gravity of PGLAs in saturated surface dry (SSD) and oven dry (OD) states. As it can be seen in
Table 5, the maximum dry density/specific gravity was achieved by PGLA1 (i.e., 70% DDA), because the DDA had the largest density among all additives. Additionally, the specific gravity of PGLAs increased from 26% to 38% and from 24% to 35% with respect to those of VLA and LYA, respectively. However, it was 24% to 31% less than that of NCA. Similarly, the dry density of PGLAs was 61% to 81% and 26% to 42% higher than those of VLA and LYA, respectively; while it was 21% to 29% less than that of NCA. Although the developed PGLAs showed higher specific gravity/density compared to reference lightweight aggregates, these results are comparable to those plastic-based aggregates developed in previous studies [
26,
27,
55,
56,
57]. In those studies, the specific gravity ranged between 0.9 and 1.9 for different plastic-based aggregates.
The observed increase in specific gravity and density of PGLAs, as compared to that of VLA and LYA, is attributed to the least number of voids between PGLAs as compared to that in VLA and LYA (See
Table 5). However, the reduction in PGLAs in comparison with NCA is linked with the lighter weight of the plastic and additives together with the increase in void percentages of PGLAs, as confirmed by the SEM images.
Furthermore, the concretes’ quality depends on the precise calculation of aggregates’ water absorption, because some concrete properties, such as porosity, cement hydration and slump, were significantly affected with the extra or shortage of water in the mix [
57].
The results in
Table 5 revealed a significant reduction in PGLAs’ water absorption, ranging from 90% to 92%, from 90% to 93% and from 6% to 18%, as compared to LYA, VLA and NCA, respectively; except PGLA3, which had a 13% increase compared to NCA. This substantial reduction in the water absorption of the PGLAs is due to the water repellent feature of plastic in the PGLAs matrix. These results suggest that the water absorption of the developed aggregate is less compared with those reported in literature [
26,
28,
55,
56]. In those studies, the aggregates’ water absorption ranged from 4.2% to 19.3%.
Finally, the strength of aggregate has a direct impact on the concretes’ strength, as the stiffer aggregate provides lower impact value and vice versa [
58]. Accordingly, the impact values of NCA, VLA and LYA were 9.65%, 39% and 21.55%, respectively. The impact value of PGLAs was reduced from 42% to 49% and from 1% to 8%, as compared to VLA and LYA, respectively, except PGLA2, which exhibited a marginal increase of 5% as compared to LYA. The high strength of PGLAs (i.e., lower impact value) is likely attributed to good interlocking in the aggregates’ matrix with fewer voids, as shown in the optical microscopic images (see
Figure 9).
The optical microscopic images indicate a non-uniform distribution between the plastic waste/binder (PET) and additives (i.e., DDA, FAA, QDA). These images show better bonding between the binder and additives, with fewer voids as compared to that of VLA. Additionally, these images showed the presence of voids in the PGLAs containing FAA (
Figure 9c) compared with entrapped air bubbles formed in the VLA, as shown in
Figure 9a. Therefore, it is expected that the PGLA concretes would possess a higher strength in comparison with that made from VLA.
Although PGLAs demonstrated a significant reduction in the strength of aggregate (i.e., higher impact value) varying from 105 to 134% in comparison with NCA, the observed impact value of developed aggregate was less than the maximum limit (i.e., 30%) specified by [
59]. In addition, concrete made with PGLAs provides a higher strength than that made with VLA, which was also expected due to the better bonding in the aggregate matrix (see
Figure 9).