The low density of lightweight concrete (LWC) makes it suitable for the rehabilitation of buildings, especially those located in seismic areas.
Most of the codes define LWC as those having a density between 1200 and 2000 kg/m3
. Such density is achieved by replacing natural aggregates by lightweight aggregates (LWA). However, LWCs tend to show lower compressive strength and less elastic modulus than traditional normal concrete (NC) made with natural aggregate. Besides, LWC tend to have lower durability than NC [1
]. One of the reasons is a weaker and more porous interfacial transition zone (ITZ) [4
Higher cement content can be used in order to improve the mechanical properties and durability of LWC; however, this carries high environmental costs. In fact, over 80% of the CO2
emissions to the atmosphere in the production of concrete are caused by the production of clinker. The cement industry emits about two billion tons of CO2
, which represents between 6 and 7% of the total world emissions [7
Eco-concrete can be defined as a concrete that reduces the amount of cement and aggregates by using waste materials, such as fly ash (FA), ash from sludge incineration, admixtures such as nanosilica (Ns), or recycled aggregates.
Ns is a nanocomposite technology. The nano-scale size of the particles can act as nucleation points in some phases of cement hydration due to their high reactivity, as well as nano-reinforcement or nano-filler to densify the microstructure, thus improving the ITZ between paste and aggregates [8
]. It is used in concentrations between 0.5 and 10% by weight of cement (bwoc), as at higher concentrations the heat of hydration rises, causing increased autogenous shrinkage. It is usually commercialized as colloidal Ns, that is, a nucleus of amorphous silicon dioxide with hydroxylated surface. It has a high surface area in relation to its volume, which makes it highly chemically reactive. It reacts with the calcium hydroxide crystals (Ca(OH)2
) found in the ITZ, producing hydrated calcium silicate (CSH gel). Due to this reaction, the amount of calcium hydroxide crystals decreases and the strength of the cement paste increases [9
]. Ns does not work only at a chemical level, but also at a physical one: it is 100 times smaller than cement, so it can fill the minuscule holes left by the partially hydrated cement particles [10
]. This leads to an increase of the density and a reduction of the porosity.
In this work natural aggregate is partially replaced by recycled polypropylene and cork. Plastic has become an indispensable material in our life due to its innumerable uses and possibilities; the counterpart, however, is the amount of waste that it causes. The generation of waste products involves storage or disposal issues, hence the need to reuse them as a recycled material. Polypropylene (PP) is one of the most common plastics. In Spain, it accounts for 22% of the total amount of plastic used according to the Spanish Center of Plastics [11
], and yet, despite the amount of literature on recycled plastic aggregates, [1
], very little has been published on PP, since most studies refer to PET or polyurethane foam [18
]. Although plastic is always a valid aggregate, it is not always suitable for structural concrete. The most usual percentage of substitution of conventional aggregate with plastic ranges between 15% and 50% in volume, while the strength reached ranges between 10 and 50 MPa. The most commonly used plastic is crushed PET bottles used as fine aggregate [13
A number of studies have analyzed the ITZ of plastic aggregates [22
] and there is a consensus that the link between cement and aggregate is weaker compared to traditional aggregates. This is because plastic being a hydrophobic material, it does not absorb free water in the ITZ. This free water leaves pores and holes on its way out. It also causes a lack of hydration in the areas around the aggregate whereas with traditional aggregates the correct hydration of such zones is improved by the slow release of the water that had been absorbed by the aggregate. The shape of the plastic used as aggregate also affects the ITZ; a smooth and rounded shape has a negative impact on the interface, producing what Ben Fraj, Kismi, and Mounanga [20
] call “wall effect”. To avoid such effects, some researchers suggest pre-treatment with sodium hydroxide [25
], use of plastic in the form of foam [20
], coating with blast furnace slag [32
], or sand [21
Cork was also assayed in this research work due to its low density. Cork is a plant tissue that is obtained by extracting cork oak bark (Quercus suber
L.). Its peculiar structure gives it a low density, between 140 and 200 kg/m3
, and a compressive strength over 90 MPa [33
]. The cork industry occupies a forest area of 22,000 km2
and produces 340,000 tons per year [34
], 68,000–85,000 tons of which are estimated to be discarded [35
]. Studies on cork as an LWA obtained very low densities, between 1000 and 2000 kg/m3
, with percentages of substitution between 2 and 18%; however, the strength was also low, below 20 MPa [36
]. Its capacity for absorption of energy, impact strength and behavior as thermal and acoustic insulation are also to be pointed out.
Nevertheless, the effects of replacing cement by supplementary cementing materials (SCM) and replacing natural aggregates by recycled ones are difficult to predict, thus there are conflicting views on this issue. This is the reason why in the current study two different dosages of LWC were assayed, replacing a certain percentage in volume of the traditional aggregate. The aim of this work was to assay and study LWC with recycled LWA capable of replacing NC in building structures, thus no special means were used for mixing or curing.
The durability behavior was studied together with density and compressive strength to determine whether or not the mix meets the requirements for a structural LWC. To analyze the durability, mercury intrusion porosimetry, sorptivity, penetration of water under pressure, depth of carbonation, and accelerated chloride ion penetration were studied.
3. Results and Discussion
The results obtained in the tests are shown in Table 2
3.1. Compressive Strength
Test results regarding compressive strength are shown in Figure 2
. The increase in the amount of cement from 550 to 700 kg/m3
did not carry almost any improvement in the strength (4.5% average), probably due to the to traditional curing by immersion at 7, 28, and 90 days. Numerous studies on reactive powder cement (RPC) achieved mixtures with high cement content (between 800 and 1000 kg/m3
) and very high strength (between 200 and 800 MPa), but using special curing such as steam, high temperature or autoclave. E.Freyssinet [39
] already, in the 1930s, applied pressure to fresh concrete getting to increase its strength, later on during the 1970s, Roy and Gouda [40
] obtained concrete with a strength of 650 MPa under pressure and temperature. The aim of our study, however, was to obtain a concrete capable of replacing NC for in situ usage, which prevented us to subject it to such treatments.
Generally, LWCs reach almost the total (85%) of their compressive strength by 7 days. Thorenfeldt [41
] pointed out that LWC reaches around 80% of its strength by the 7th day, possibly due to the limit of the strength of LWA. As it was to be expected, the use of the LWA that was selected had a negative impact on the development of strength, reducing it by 68%, this is due to many parameters, among others, an irregular graining of the fine aggregates caused by the high regularity in the shape and size of recycled aggregates, as Katzer and Kobaka [42
] asseverated, an stablished optimal field of graining of fine aggregate allows to achieve the most advantageous strength parameters of the concrete produced. In Table 2
a reduction in strength can be observed, which also means a substantial decrease in density. Even if the strength remains within the ranges of structural usage, a linear relationship between compressive strength and density can be observed.
3.2. Elastic Modulus
shows the evolution of the results of the tests over time. The concretes assayed reach high elastic modulus values. The fewer voids the concrete contains, the higher its modulus will be [43
]. High compactness in the microstructure restricts the movement of the shrinkage in the concrete, which causes high internal tension. Such pressures lead to high elastic modulus values.
The elastic modulus is a characteristic of concrete that is related to strength, shrinkage and microstructure. Neville [44
] stated that LWC usually bear great shrinkage due to their low elastic modulus and the substantial proportion of fine aggregate that originates many voids. The relation between shrinkage and modulus seemed to be so close that Reichard [45
] proposed a correlation between both. Even if such correlation is not valid for LWC, it does suggest that at low modulus should expect large shrinkages. Li and Li [46
] explored this issue and proposed a model to determine autogenous shrinkage of SCC based on the knowledge of the critical capillary diameter, the Poisson’s ratio and the elastic modulus. This was also observed by Beushausen and Gillmer [43
], who obtained values similar to those of our LWC in their mortars with low w/b ratio, both in elastic modulus and in shrinkage.
3.3. Mercury Intrusion Porosimetry (MIP)
shows the results of MIP. It is observed that total porosities of both HPCs are very similar and lower than those of the LWCs. The porosity of ECO HPC 700 is 24% lower than that of ECO HPC 550. Although the cement content does not influence compressive strength, it does modify the porous structure of the material, especially in concrete with a high volume of paste. As for capillary porosity ranges, HPC 700 has 18% less capillary porosity than HPC 550; and ECO HPC 700 30% lower than ECO HPC 550. Regarding the substitution of LWA by natural aggregate, MIP shows that HPC 700 has 25% less capillary porosity than ECO HPC 700 and HPC 550, 37% less than ECO HPC 550.
The dimensional distribution of the pore size is shown in Figure 3
It may be observed that HPCs have low porosity when pore diameter is higher than 0.070 μm; porosity rises sharply between 0.050 μm and 0.030 μm diameter then porosity decreases from 0.03 to the smaller pores where is consistently low. The lines of both HPCs overlap, which means that they do not show substantial differences in their porosity ranges. This suggests that, as far as total porosity is concerned, the amount of cement or Ns used has no effect.
On the other hand, the LWCs have 3 peaks, at diameter 6 μm, 0.5 μm and 0.045 μm. The last two are within the range of capillary porosity, the latter being the most pronounced. The main difference is in the pores under 0.045 μm: at this peak ECO HPC 550 shows higher porosity than ECO HPC700; besides, it decreases more slowly. This situation remains unvaried at diameters below 0.045.
Therefore, in LWC with plastic and cork aggregates replacing 50% of the natural aggregates, increasing the amount of cement and Ns reduces the porosity at pore diameters under 0.045 μm due to the improvement in the ITZ between LWA and cement paste, a fact that does not affect the ITZ between natural aggregate and cement paste.
compares the concretes assayed in this study with others in the scientific literature. It can be can observed that porosities are among the lowest. Bermejo Núñez, et al. [47
] compare concretes obtained using different SCMs (pozzolans and slags): for the mixtures with SCMs curves are similar to the LWCs of the present work, while for the mixture with cement only, the curve obtained is similar to HPCs. The authors attribute this to lower durability of mixtures with pozzolans.
shows the water absorption data in function of time, distinguishing between initial sorption—up to 6 hours, and secondary sorption—after 24 hours. As observed in Table 2
, the absorption during the first hours is very similar for all four concretes, while the differences stand out starting at 24 hours. HPC concretes have very similar levels in the early phases of secondary sorptivity with a tendency to increase over time. Differences between LWCs are more constant over time; the differences in LWC absorption are 57% higher than the differences between HPCs. This may be due to high levels of cement and Ns in HPCs, which may cause the water to be absorbed more slowly over time.
In initial sorptivity the data are very similar, ECO HPC 550 being the only one that shows absorption levels higher than the rest. In secondary sorptivity, the data show greater differences among the mixtures and results are very similar to those of porosimetry by MIP, HPC 700 being statistically equal (2% difference), to HPC 550.
LWCs absorb more water, the difference between ECO HPC 700 and 550 being greater than for HPC (12%). The impact of replacing natural aggregates with lightweight ones is low in the case of HPC 700 and ECO HPC 700—the HPC absorbs 9% less than the LWC. In 550 cement concrete the difference is greater: HPC 550’s secondary sorptivity is 19% lower than that of ECO HPC 550. These data only confirm those of porosimetry and that, in fact, pore ranges within the limits of capillary porosity are directly related to the absorption of water by capillarity.
The results of secondary sorption with compressive strength are plotted in Figure 4
, in which the sorptivity–density relationship of the mixtures assayed in this work is compared with other studies. Du, Du, and Liu [5
] analyze the effect of Ns on several properties of LWC. They conclude that, Ns in a concrete with type I cement, improves compressive strength, sorptivity, water penetration, porosity and chloride attack. Nibudey, et al. [48
], study the compressive strength and the sorptivity of concrete with recycled plastic fibers. Karahan, et al. [49
] analyze the effect of Metakaolin in self-compacting concrete, finding it has a positive effect, and Panesar [50
] tests LWC with different types of foaming agents and analyzes their effects.
3.5. Water Penetration under Pressure
The results of the test are given in Table 2
. As in the MIP and sorptivity tests, the results show that HPC 700 shows the best performance given its lower porosity. On the other hand, HPC 550 allows the same penetration as its equivalent in LWC. As for LWCs, ECO HPC 700 has higher water penetration values than ECO HPC 550. There is a relation between density and water penetration.
In any case, despite the differences obtained, the penetration values are much lower than the maximum values permitted by law. In case of aggressive environmental exposure, the EHE-08 Spanish standard requires to comply with the specifications for the water penetration test, which are 50 mm maximum depth and 20 mm average depth, that is, 54% more than the concrete with highest penetration among those who have been assayed. This implies that although some values do not correspond directly to the sorption and MIP tests, overall they are so low that the differences between the four concretes are minimal and all of them have markedly low values.
compares the values obtained for average penetration with other studies. The results of the test match those obtained by Liu, et al. [51
] that developed LWC with high strength to penetration of water and chlorides and obtained concrete with penetrations as low as those of the present study. For this they designed a concrete with 600 kg/m3
of cement, that is, a content in cement higher than the 500 they used in the rest of mixtures, adding silica fume (not used in the rest), a carefully selected granulometry and several different LAWs, expanded glass among others. In both cases the high cement content is the same, as well as the use of SCMs with silicon and the combination of several types of aggregate.
] points out that water penetration depth is 55% lower in Ns concretes compared to mixtures with no Ns. Besides, the analysis of their microstructure justifies this improvement insofar as Ns absorbs Ca(OH)2
crystals reducing its quantity, which densifies the interstitial zone. In addition, it can fill the gaps of CSH gel.
Therefore, the high cement content and the presence of Ns explain the reduced levels of penetration of the samples, while the low accessibility of the porosity due to its tortuosity explains the differences with the porosity tests.
3.6. Carbonation Depth
One of the main reasons for the deterioration of concrete constructions is carbonation. The reaction of atmospheric carbon dioxide (CO2
) with the products of cement hydration (such as Ca(OH)2
), produce calcium carbonate (CaC03
). This reaction results in a reduction of the PH of the concrete from 12.6 to less than 9, which can destroy the passive oxide layer that protects the reinforcement from corrosion [52
shows the carbonation depth of the concrete object of the study. The values achieved are very low, even zero carbonation after 90 days in the chamber in the case of HPC 700. This is explained by the fact that when CO2
enters the concrete, it reacts first with the available calcium hydroxide (Ca(OH)2
) in the pore network and, once it is used up, it reacts with CSH [54
]. The depth of carbonation of concrete is conditioned by the porous structure; Valcuende and Parra [55
] showed that the thinner and more tortuous the porous structure, the slower the process. For the concretes assayed in this work, MIP, sorptivity and water penetration tests show very low porosity in all the mixtures.
The incorporation of Ns produces a small reduction in PH and the content of alkali ions in the porous structure of the concrete. High pH and alkaline content accelerate the chemical reaction of carbonation [56
]. In addition, Ns increases the chemical stability of the products of cement hydration, which also improves their resistance to carbonation [57
The concrete curing method also has an impact. Concretes cured in a humid chamber usually present higher carbonation than those that are immersion cured. This is the case of the concretes of this study [58
]. This is possibly due to the slight superficial drying necessary to convey the water to the interior, which increases the capillary porosity of the concrete surface.
shows how the carbonation depths of the concretes assayed are in the lower ranges of the studies analyzed. In the work of Shi, et al. [59
], there are concretes with similar levels to those analyzed that have high amounts of cement (550 kg/m3
) and low water/cement ratio.
There is consensus on the direct relation between low w/c ratio and low carbonation [60
], but not on the influence of the amount of cement. Wassermann, Katz, and Bentur [60
] found that concrete with different amounts of cement but equal relation w/c showed the same depths of carbonation; however, maintaining the amount of cement but varying the a/c ratio, carbonation was indeed affected. On the other hand, in that study the maximum amount of cement used was 450 kg/m3, and in studies with amounts of cement above 500 kg/m3
the penetration of carbonation was very low or zero. This is the aforementioned case of Shi, Xu, and Zhou [59
] and others [62
]. In scientific literature, therefore, it has been stated that an amount of cement greater than 500 kg/m3
and w/b > 0.4 ratio lead to very low or even zero carbonation. The tests of accelerated carbonation at 56 and 90 days confirm this.
Regarding the influence of aggregates—plastic and cork in this case, according to Tragardh [65
] or Tam, et al. [66
], the ITZ is the weakest area of cementitious compounds. Therefore, if the aggregates get an ITZ resistant with the paste, the carbonation will be higher. Thus, for LWC, Rodhe and X-Betong. [67
] compared three types of LWA obtaining different depths of carbonation independently of the strength, which in their opinion is due to the improvement in the porous structure of the ITZ. The use of Ns, given that its positive impact on the ITZ of the aggregate-paste has been demonstrated, is another factor of improvement in the penetration of carbonation.
3.7. Chloride Ion Penetration
Concretes exposed to marine environments are attacked by salts such as sodium chloride, magnesium sulfate, and bicarbonates. Such salts dissociate and migrate by capillarity to the concrete; the attack is particularly aggressive on the armor of reinforced concrete.
There are different methods to measure the resistance to attack by chlorides in concrete. These, in turn, lead to studies to be able to compare the results or analyze their validity [68
]. In this case the immersion method has been chosen to simulate the attack and the colorimeter process to measure it.
The penetration was measured by spraying the specimens with silver nitrate solution (AgNo3
). The depth was determined by the whitish color due to silver chloride precipitation. The concentration of chlorides needed for the color change to occur is 0.07 N in concrete with Portland cement. According to Otsuki, et al. [69
], the color change becomes visible when the chloride content is greater than 0.15% of the weight of the cement.
The results of the tests carried out after 28 and 56 days of exposure are shown in Table 2
. The four concretes show low penetration, three being statistically similar, while HPC 550 is slightly lower than the others. As have been seen, in the present work porosity of the concretes is reduced and low porosity hinders the advance of the attack by free chlorides [51
Other authors have studied mixtures in which LWCs were more resistant to such attack than NC. Nyame [70
], for example, analyzed mortars with fine LWA, with higher porosity than its equivalent with traditional aggregate, but with lower chloride ion penetration. This was attributed to the reduced filtration in the ITZ and to the better internal hydration of the cement by the aggregate (Thomas [71
] and Liu, Chia, and Zhang [4
Regarding the use of SCMs, Hooton and Titherington [72
] and Detwiler, et al. [73
] studied their effects on chloride ion penetration, and found that their use had a positive impact against this attack.
Ns also plays an important role in resistance to this attack. Said, et al. [74
], for example, analyzed the properties of concrete with Ns, also using the colorimetric method for the penetration of chlorides. They found that the samples without Ns showed penetrations of 8 and 10 mm while those with Ns obtained levels close to 5 mm, similar to those of the concrete of this work. The use of FA replacing the cement was also analyzed in the previous work, resulting in similar levels in the samples regardless of the substitution ratio.
The work of Zahedi, et al. [75
] analyzes mortars with high content in cement and Ns, being the mixtures containing Ns more resistant to chlorides. The variation in cement content (between 825 and 1000 kg/m3
) does not seem to have an effect.