Abrasion resistance of concrete made with recycled aggregates

Abstract

Global sustainable development will require using every resource to its maximum potential. The potential of construction and demolition waste's inert fraction can be maximised if it is reused as recycled aggregates (RA) for concrete production. However, there are technical risks involved in this technique, particularly due to lack of knowledge in terms of the long-term behaviour of concrete made with these aggregates.

Abrasion resistance is one of the least known properties of hardened concrete but nevertheless one of the most important for certain applications (hydraulic structures and floors/pavements). Hardly any research has been published on this property for concrete made with RA.

In this paper, the results of four experimental research projects supervised by the author are put together to draw conclusions on this property. It is concluded that there is no loss of abrasion resistance with the incorporation of RAs and that there are reasons to believe that it even improves, even though the quantification of that improvement needs further investigation to be accurate.

Keywords:

• construction and demolition waste
• recycled aggregates
• concrete
• abrasion resistance

1. Introduction

Sustainable development is a consensual concept nowadays, meaning that the fulfillment of today's generation's needs should not impair those of future generations. The construction industry, as a fundamental part of development, is known for its important environmental, social and economic impacts: the first is usually negative, while the second and third are positive. Concrete is the material that is most widely used in construction (in fact, and according to the European Commission ECOserve Thematic Network, at 1.5–3 tonne per capita per year in the industrialised world, it is the second most widely used material in the sum of human activities, after water) and the impacts of its production are huge.

According to average compositions, each tonne of current concrete contains roughly 300 kg of cement (the material with the highest environmental impact, mostly due to the energy used in its production), 950 kg of coarse aggregates (crushed natural stone) and 700 kg of fine aggregates (river/sea sand or crushed natural stone), among other components. Stone quarries underpin a lucrative industry (3 million tonnes per year in the EU, according to Arab (2006)) and provide jobs but have several negative impacts on the soil, water and air and also on neighbouring populations (noise and sound pollution). In addition, the extraction of sand from river and sea beds may sometimes coincide with the need to dredge their bottom but on the whole this practice is harmful to the environment.

Another consequence of the construction activities is the build-up of construction and demolition waste (CDW; around half a tonne per capita per year in Europe, according to the Symonds report (1999)), to which refuse from the materials industry (e.g. pre-cast concrete and ceramics) should be added. Even though in the most developed countries dumping of this waste has turned almost completely into materials' reclamation, nearly all the reuse of these materials (according to Angulo (1998), more than 80% of which are inert, e.g. stone, concrete, bricks, mortars) can be classified as downcycling, i.e. reuse below its potential. As a matter of fact, CDW is mostly used as aggregate in roads, streets and floors of buildings, the landscaping of spent mining infrastructures or drainage layers of landfills.

Though part of the problem, concrete can also be part of the solution. In fact, CDW can be used either in cement production, as replacement of coarse and fine aggregate [so-called recycled aggregate (RA)] and as filler. Given the huge quantities of concrete produced, the potential of these uses of CDW is particularly promising and at least in the case of RA, it does not fit into the definition of downcycling.

2. Concrete and abrasion resistance

Concrete abrasion resistance is particularly important in check dams (designed to protect the downstream riverbed from hydraulic scouring and erosion), spillways and other hydraulic structures due to the phenomena of erosion and cavitation that lead to surface peel-offs, concrete disintegration and rebar exposure. It is also important in smaller-scale elements such as pavements/floors and staircase steps subjected to very intensive wear from pedestrians and vehicles. Finally, abrasion resistance is relevant to any use of concrete where rubbing, scraping, skidding or sliding of objects on the surface commonly occur (Zeng-Qiang and Chung 1997).

2.1 Conventional concrete

Concrete performance is usually measured in mechanical terms by properties such as compressive strength and modulus of elasticity, and less frequently, tensile strength, shrinkage and creep. In durability terms, other properties are used, such as carbonation and chloride penetration resistance, and less frequently, water absorption and air/oxygen permeability. Abrasion resistance is very rarely cited and therefore the studies on this property are few by comparison and include those listed below.

Liu (1981) examined the influence of aggregate type, water-cement ratio and hence the compressive strength of the concrete and surface treatment of the concrete test specimens on abrasion resistance and provided some guidelines.

Fernandez and Malhotra (1990) concluded that the abrasion resistance of concrete containing ground granulated blast-furnace slag is inferior to that of the control concrete without slag.

According to Dhir et al. (1991), abrasion resistance in concrete depends on various factors, such as water/cement ratio (W/C), curing, workability, maximum aggregate size and the constituent materials. The same authors examined, among others, compressive strength, rebound hardness and permeation tests (initial surface absorption, intrinsic permeability and vapour diffusivity) for their ability to predict the abrasion resistance of concrete, and recommended the initial surface absorption test for this purpose (Dhir et al. 1986).

Laplante et al. (1991) concluded that: (i) while silica fume can slightly improve concrete abrasion resistance, coarse aggregate nature is more significant; (ii) as long as W/C is maintained at about 0.30, the concrete's abrasion resistance almost equals that of the coarse-aggregate source rock; and (iii) abrasion resistance in high-strength concretes with very low W/C is largely determined by that of the coarse aggregate.

Sustersic et al. (1991) concluded that the erosion-abrasion resistance is improved by increased compressive strength and by an increase in fibre content. It can be correlated to improvements in abrasion resistance from the Bohme test method but only at constant W/C and for different fibre content.

Naik et al. (1995) evaluated the abrasion resistance of concrete with different levels of cement replacements with Class C fly ash and also a reference concrete (RC) without fly ash. Test results showed that the abrasion resistance of concrete with cement replacement up to 30% was comparable to the RC. Beyond 30% cement replacement, fly ash concrete exhibited slightly lower resistance.

Zeng-Qiang and Chung (1997) found the abrasion resistance of mortar to be significantly improved by the addition of latex, and further improved by an addition of short carbon fibres. Both effects relate to the increase in tensile strength. The abrasion resistance was also improved by the addition of silica fume, due to the increase in tensile modulus. The abrasion resistance was better for mortar with silica fume than mortar with latex, but was worse for mortar with silica fume than for mortar with latex and carbon fibres.

Yu-Wen et al. (2006), using a full-scale field test, concluded that low-cement, high-performance concrete (with admixtures such as fly ash and slag) has a significantly higher erosive resistance than that of normal concrete, as measured by the averaged abraded depths.

The most relevant conclusion drawn from various researchers, notably Laplante et al. (1991), for the present paper is that, besides the finishing conditions of the surface and the compressive strength of the concrete, the aggregates (and the coarse ones in particular) significantly affect the concrete's abrasion resistance. When using aggregates from recycled CDW in the production of concrete a reasonable doubt arises as to the effect that this replacement will have on the abrasion resistance of the resulting concrete.

2.2 Concrete with RAs

International research on the performance of concrete made with RAs, from here on called RA concrete (RAC), is abundant and it has yielded such references as: Olorunsogo and Padayachee (2001), Soberón (2002), Chen et al. (2003), Hansen and Narud (2003), Kou et al. (2004), Muller (2004), Xiao et al. (2006). But abrasion resistance does not usually feature in this research on RACs performance, with only a few exceptions: Olorunsogo (1999), Limbachiya et al. (2000), Sagoe-Crentsil et al. (2001) and Latterza and Machado (2003).

Olorunsogo (1999), using coarse RA from crushed concrete, stated that no clear relationship could be found between the abrasion resistance of RAC and the proportion of RA included in the mixes.

Limbachiya et al. (2000), also using coarse RA from crushed concrete, concluded that the differences between concrete made with only natural aggregates (NA) and that with 100% coarse CA at design strengths of 50 and 60 N/ram² were 0.03 and 0.04 mm, respectively (in terms of abrasion depth). This suggests that, at a given design strength, RAC mixes have similar abrasion resistance to the corresponding concrete with NA, regardless of the proportion of RA used in the mix.

Sagoe-Crentsil et al. (2001) concluded that the abrasion loss of RCA made with ordinary Portland cement increased by about 12% compared to conventional concrete without coarse RA (crushed concrete). They also concluded that the fresh and hardened RCA properties can be enhanced by using industrially-produced aggregates instead of aggregates derived from laboratory-crushed concrete, due to improved grading and quality achievable in plant crushing operations.

Latterza and Machado (2003) did not perform any specific test to directly assess the abrasion resistance of RAC and deduced that it was not affected by the incorporation of coarse RA (crushed concrete) by comparing the results of the surface hardness measured with a sclerometer.

3. Abrasion resistance in RAC

3.1 Research on RAC at Instituto Superior Técnico

At the Department of Civil Engineering and Architecture of the Instituto Superior Técnico (IST), Technical University of Lisbon, Portugal, a long-term research programme on the use of RAs in the production of RAC and mortar mixes started around 10 years ago and has resulted in 16 Masters dissertations and four PhD theses (most supervised by the author and some still in progress), as well as some international publications cited below. These works have focused on the following themes: single and multiple recycling of coarse concrete aggregates (Santos et al. 2004); recycling of coarse ceramic aggregates (de Brito et al. 2005, Correia et al. 2006); influence of admixtures on the performance of RAC with coarse concrete aggregates; recycling of stone slurry (Almeida et al. 2007); recycling of fine concrete aggregates (Evangelista and de Brito 2007); recycling of fine ceramic aggregates (Silva et al. 2008, 2009); recycling of coarse concrete, ceramic and mortars aggregates (Gomes and de Brito 2008); susceptibility of RAC to alkali-silica reactions; influence of the pre-saturation process of coarse concrete aggregates; survey of existing codes on RAC; surveys of national and international research on RAC (de Brito and Robles 2008); susceptibility of RAC to curing conditions; recycling of coarse and fine rubber crumb aggregates; and methodology for the early prediction of long-term properties of RAC.

3.2 Research on RAC's abrasion resistance at IST

Most of the works cited did not concern the abrasion resistance of hardened concrete, however. The only ones which report final results on this concrete property are:

• Rosa (2002), tested replacement rates of 1/3, 2/3 and 3/3 of coarse limestone aggregates with coarse recycled ceramic aggregates to determine the compressive and tensile strength, abrasion resistance, water absorption by capillarity and immersion of hardened concrete.

• Matias and de Brito (2005), tested replacement rates of 25, 50 and 100% of fine limestone aggregates with fine recycled concrete aggregates (with a focus on the use of super-plasticisers) to determine the compressive and tensile strength, abrasion resistance (only the 0 and 100% replacement rates' mixes were tested for this property), shrinkage, water absorption by capillarity and immersion, carbonation and chloride penetration of hardened concrete.

• Evangelista (2007), tested replacement rates of 10, 20, 30, 50 and 100% of fine limestone aggregates with fine recycled concrete aggregates to determine the compressive and tensile strength, modulus of elasticity, abrasion resistance (only the 0, 33 and 100% replacement rates' mixes were tested for this property), shrinkage, water absorption by capillarity and immersion, carbonation and chloride penetration of hardened concrete.

• Fonseca (2009), tested replacement rates of 20, 50 and 100% of coarse limestone aggregates with coarse recycled concrete aggregates to determine the influence of the curing conditions (OEC, outer environment; LCC, laboratory conditions; WCC, wet chamber and WIC, water immersion) on the compressive and tensile strength, modulus of elasticity and abrasion resistance of hardened concrete.

Some common procedures were used in all these experimental campaigns in order not to hinder the detection of the influence of the incorporation of RA on the different properties of the RAC mixes with other parameters. Therefore, the following factors were kept constant since, when they change, almost every concrete property can be affected (de Brito 2009).

• Effective W/C ratio (distinguishing between the total amount of water introduced into the mix and that which effectively contributes to the hydration of the cement and the workability of fresh concrete); this property has a strong influence on the concrete's mechanical performance, but even more on its durability; in order to keep this factor constant, the aggregates were pre-wetted in the Rosa, Matias and de Brito campaigns, and extra water was added to the RAC mixes to compensate for the higher water absorption of the RA used in the campaigns of Evangelista and Fonseca; in these last two campaigns the standard mixing method (kept constant within each campaign for every concrete mix) had to be adapted to the need of the RA to absorb the extra water inside the mixer.

• Workability (this property must be maintained by using plasticisers or increasing the total amount of water without increasing the effective W/C ratio, for example, by pre-saturating the RAs); for practical purposes, concrete mixes with different workability levels may not have the same range of applications and therefore should not be directly compared.

• Grading curve (size distribution) of the aggregates (when replacing NA with RA this curve should be kept exactly constant because any change leads to uncontrolled shifts in almost every relevant property of concrete).

• In the case of RA from recycled concrete, the origin/nature of the original NA should not differ drastically from that of the NA used to make the RAC and the RC without RA.

There is no consensus on the most suitable test to measure the erosion/abrasion/wear resistance of hardened concrete. Depending on the country of origin of the authors, it is possible to find research on abrasion resistance using the procedures set forth in ASTM C779-82 (procedure C – Ball bearing), ASTM C944-90a (rotating-cutter method), CRD-C 63-80, the Bohme test method, BS 8204 (part 2: 1999), and others. Given the apparatus in IST's laboratory, the standard used in all the campaigns described here was DIN 52108, which establishes the following steps (Figure 1): before the test, three specimens for each concrete mix (obtained by sawing pieces off bigger specimens from other tests, leading to prismatic specimens with a square base approximately 70 mm wide and a 50 mm height) are kept in a humid chamber and then dried until their mass stabilises; then they are measured, the testing machine is cleaned and a known quantity of abrasive sand is placed in the disk; the specimen is positioned over it, with and under a calibrated weight; after four cycles of a pre-determined number of rotations, the dimensions of the specimen are measured and they are weighed.

Figure 1 Two steps of the abrasion resistance test.

The results of each individual campaign are presented next in Figure 2, where the average loss of thickness by abrasion Δl is plotted against the replacement ratio (%) of the NA with the different RA.

Figure 2 Average loss of thickness by abrasion Δl (cm) vs. replacement ratio (%). (a) Rosa (2002), (b) Matias and de Brito (2005), (c) Evangelista (2007) and (d) Fonseca (2009). OEC, outer environment; LCC, laboratory conditions; WCC, wet chamber; WIC, water immersion.

The results of the first three campaigns show nearly linear correlation between the two parameters with a positive trend of most interest to every use of concrete where abrasion resistance is important. This can be explained by the better adhesion between the mortar paste and the RAs, caused by their greater porosity as compared with the limestone aggregates (for ceramics this is an intrinsic characteristic; for fine and coarse crushed concrete particles this is due to the paste that adheres to the original stone aggregates). Abrasion resistance is controlled fundamentally by the wear of the matrix and its bonding with the coarse aggregates, as opposed to the wear of the coarse aggregates, which explains the results of Rosa (coarse ceramics) and Matias and de Brito (coarse crushed concrete). As for the results of Evangelista (fine crushed concrete), they may be explained by the fact that abrasion resistance is also dependent on the bond of the cement paste with the fine aggregates, which is better when RAs are used, because of their greater roughness.

The results of Fonseca's study on the sensitivity of RAC (with coarse concrete RA) to curing conditions are not as straightforward as the other ones and they need a finer interpretation. Since curing conditions strongly affect concrete's surface layer, it is noted that the test specimens (71 × 71 × 50 mm³) were obtained by sawing pieces off larger concrete cubes (100 mm edge) after curing. This was done so that the concrete's surface finish would not be a variable in the test. Thus, the test surface is the cut surface itself, i.e. an internal plane of the concrete element, composed of aggregates and binder mix, and not an outer surface. The irregular variation of abrasion resistance values, in all curing conditions, does not allow the establishment of a clear relationship between this property and the incorporation of coarse concrete RA. Excluding the mix with 50% replacement in WCC, which reveals a 10% higher wear than the RC, the abrasion resistance of all other types of concrete differs by not more than 5.4% in relation to the RC, which is not statistically significant from an experimental point of view. But it must be noted that all specimens from concrete mixes with 100% replacement ratio exhibit the lowest loss of thickness and subsequently higher abrasion resistance. It can therefore be concluded that the incorporation of coarse concrete RA leads to a slightly better performance in terms of abrasion resistance. This can be explained by the better connections formed between the binder and the recycled concrete coarse aggregate, in view of their higher porosity, and so the results are consistent with those of the researchers mentioned above. Regarding the influence of the curing conditions, no clear conclusion can be reached, although the lower variation values suggest that RAC does not appear to be affected any differently from conventional concrete.

3.3 Results comparative analysis

In order to allow an early prediction of long-term characteristics, properties and performance of RAC, a methodology has been put forward (de Brito and Robles 2008) that depends on the weighed value of the density and water absorption of all the aggregates in the mixture. This value depends on two factors: the density/water absorption of the individual aggregates (depending on their origin), and the proportion of each type of aggregate used in the mixture to produce the concrete. To calculate the weighed density value used, where the percentage of each type of aggregate is multiplied by the corresponding density, the following equation was used:

where D _mix, weighed density of the mixture of aggregates in the concrete mix; FA, percentage of fine aggregates used in the mix; subst_FRA, replacement ratio of fine RA by fine NA; subst_CRA, replacement ratio of coarse RA by coarse NA; D _FRA, density of the fine RA; D _FNA, density of the fine NA; D _CRA, density of the coarse RA and D _CNA, density of the coarse NA.

To calculate the water absorption of the aggregates used in the mixture, a similar equation was adopted where the density values were replaced by the water absorption values for each of the aggregates used. Different density and water absorption values for the aggregates in the mixture were obtained for each substitution rate and type of RA. The individual values of density and water absorption for the various aggregates used in the experimental campaigns are given in Table 1.

Table 1 Densities (D) and water absorptions (WA) of the aggregates used.

	Rosa		Matias and de Brito		Evangelista		Fonseca
	D (kg/m^3)	WA (%)	D (kg/m^3)	WA (%)	D (kg/m^3)	WA (%)	D (kg/m^3)	WA (%)
FNA	2727	0.9	2544	0.8	2544	0.8	2.540–2590	0.4–05
FRA	–	–	–	–	1913	13.1	–	–
CNA	2690	1.0	2632	0.8	2566–2611	0.7–13.1	2510–2570	1.3–1.7
CRA	2019	12.0	2355	4.1	–	–	2310	6.1
Notes: FNA, fine natural aggregates; FRA, fine RAs; CAN, coarse natural aggregates and CRA, coarse RAs.

This methodology allowed comparison of the results (in terms of abrasion resistance) of all the experimental campaigns presented here, regardless of the type of RA used. Figure 3 summarises the results obtained.

Figure 3 Ratio of concrete abrasion mass losses (Δl _RAC/Δl _RC) between RAC and RC vs. the ratio of the densities D (a) and water absorptions WA, (b) of the mixture of aggregates (Fonseca 2009).

The correlation coefficients obtained are low, contrary to what occurs with other properties (de Brito 2009), even though there is a clear descending trend in terms of mass loss as the density of the aggregates decreases and their water absorption increases. Therefore, even though abrasion resistance seems to increase consistently as the replacement ratio of RA by NA increases, the rate at which it does so varies considerably and its determination needs further investigation.

4. Conclusions

The following conclusions can be drawn from the study presented in this paper:

• Reusing CDW as RAs in concrete production is an environmentally sound and viable process and a good alternative to the currently prevalent downcycling.

• There are abundant research results on conventional concrete's abrasion resistance, but almost nothing has been reported for concrete made with RAs.

• Abrasion resistance of concrete with RAs is at least as good as the corresponding RC without RAs.

• The rate at which the abrasion resistance of concrete made with RAs changes with the incorporation rate of these aggregates depends not only on their origin but on other factors, too (e.g. curing conditions), and further research is needed on the subject.

Acknowledgements

The author is grateful for the support of the ICIST Research Institute of IST, Technical University of Lisbon and of the FCT (Foundation for Science and Technology).