Introduction

Around 100 billion tons of construction and demolition waste (CDW) are generated globally. It is estimated that by 2060, the demand for materials such as steel, concrete, and cement, the main contributors to greenhouse gas emissions, will be double that of today (UNEP, 2022), indicating that the construction industry needs immediate measures to decarbonize the sector by 2050 (GCCA, 2021). To boost sustainability in the construction chain, it is essential to adopt the circular economy model, making it possible to reduce the consumption of natural resources, thus minimizing waste and greenhouse gas emissions (Benachio; Freitas; Tavares, 2020).

In this perspective, CDW may be processed and used as coarse and fine aggregate in producing mortars and concretes, with its use regulated in several countries (Kenai, 2018) and Brazil (ABNT, 2021). However, fine particles are generated (<0.15 mm) during the production process of these aggregates (Martin et al., 2023; Quattrone; Angulo; John, 2014). Around 15 to 19% of the processed material corresponds to CDW dust (Oliveira; Dezen; Possan, 2020; Zhang et al.; 2022), which, as it has no commercial value, becomes a liability for recycling companies (Lampris; Lupo; Cheeseman, 2009).

The RCP contains CaCO₃, SiO₂, AFt (ettringite), AFm (monosulfate), CH (portlandite), C-S-H (hydrated calcium silicate) and non-hydrated cement particles (anhydrous) (Oliveira et al., 2023; Wang et al.; 2024), it has low reactivity and high Absorption (Duan et al., 2020a; Mehdizadeh et al., 2021), acting as a filler material (Tan et al., 2020), where the particles fill the space between the cement grains, modifying their granular packing, which implies a change in the initial porosity of the paste (Irassar et al., 2015), and which, depending on the proportion and granulometry, can densify the microstructure, resulting in better mechanical performance (Awoyera; Adesina; Gobinath, 2019; Oliveira et al.,2024).

In recent years, the literature has shown that recycled concrete powder (RCP) may be used as supplementary cementitious materials (SCMs) (Aquino Rocha; Toledo Filho, 2023; Martin et al., 2023; Tang et al., 2020; Tokareva; Kaassamani; Waldmann, 2023), which, in addition to acting as a raw material with added value, may reduce the emissions associated with Portland cement production (Scrivener; John; Gartner, 2018).

Some treatments such as thermal activation (Chen et al., 2024; Vashistha et al., 2023), carbonation (Mehdizadeh et al.,2021; Ouyang et al.,2020; Qin; Gao, 2019), and mechanoactivation are used to improve the characteristics of these particles. Mechanoactivation occurs by comminution (grinding), which can modify the crystalline structure, provided by the higher SiO₂ content and lower CaO content, reducing the particle size and increasing the specific surface area (SSA), culminating in greater reactivity of the RCP due to the exposure of the anhydrous Portland cement grains (Bu et al., 2023; Oliveira et al., 2023; Zhang et al., 2022).

Although particle size decreases with increasing grinding time, there is an exponential reduction in the grinding efficiency of recycled CDW powders with increasing time (Aquino Rocha; Toledo Filho, 2023; Sun et al., 2021; Xu et al., 2021), justified by the tendency to aggregation due to Van der Waals forces (interparticle attraction). The magnitude of these forces increases during dry mechanoactivation (Bernardes, 2006) as the electric charges on the fractured surfaces are particularly negative in C₃S and C₂S crystals and positive in C₃A and C₄AF crystals (Aïtcin; Mindess, 2011).

In this way, to improve the efficiency of milling processes, producing greater fineness with lower energy consumption and consequently lower carbon dioxide emissions (Prziwara; Kwade, 2020; Zan; Ishak, 2023), the forces of repulsion between the particles must outweigh the forces of attraction (De Castro; Pandolfelli, 2009). Grinding aids may be used during this process, the most common being amines, glycols, alcohols, and phenols (Engelsen, 2009; Katsioti et al., 2009; Toprak; Altun; Benzer, 2018).

High-polarity grinding aid groups (-COOR, -NH₂, -OH, -SO₃) are generally composed of long-chain, electrically charged organic molecules that prevent agglomeration, causing a propensity for adsorption on electrostatic surfaces formed by breaking the covalent bonds of Al-O, Si-O, and Ca-O (Collepardi, 2005; Toprak; Altun; Benzer, 2018). The effect of using a grinding aid on a clinker is shown in Figure 1.

The concentration range of grinding aids in cement production is between 0.01 and 0.1%, and up to 0.5% may be used (Jeknavorian; Barry; Serafin, 1998; Teoreanu; Guslicov, 1999). The most commonly used grinding aids are propylene glycol (PG), triethanolamine (TEA), triethanolamine acetate, and polyglycol phenol ether (Assaad; Vachon, 2021; Engelsen, 2009; Katsioti et al., 2009; Toprak; Altun; Benzer, 2018). Sodium hexametaphosphate (SHMP) is also used as a particle dispersant (Rodrigues; Evangelista; Brito, 2013).

In the search for process efficiency (increased production capacity, reduced specific energy consumption, or finer particles), several studies are carried out related to the grinding of raw materials such as quartz, feldspar, limestone and calcite, and the performance of the final product. (Portland cement) by controlling the specific surface area (SSA) using the Blaine method (Cayirli et al., 2023; Katsioti et al.,2009; Kaya et al., 2023; Prziwara; Kwade, 2020). However, the grinding aid is commonly adopted in studies with recycled powder materials, without prior compatibility or efficiency studies.

In this sense, considering that, depending on the water/cement ratio, the content of the C3A phase (3CaO.Al₂O₃), as well as the amount of sulfates (the increase in the sulfate content reduces the adsorption of the additive if it is used during grinding) (Kapeluszna; Kotwica, 2022; Kujawa; Olewnik-Kruszkowska; Nowaczyk, 2021) and that the Blaine method does not take into account the difference between the particle size distribution curves of the powders (Assaad; Vachon, 2021; Delagrammatikas; Tsimas, 2004; Ferraris et al., 2002), this study aimed to evaluate the effect of mechanoactivation in the presence of grinding aids on the physical characteristics (d50 and SSA-BET) of recycled concrete powder for use as a supplementary cementitious material.

Materials and methods

Materials

The recycled concrete powder (RCP) used in this study originated from the comminution process of concrete specimens with different compositions supplied by a concrete and precast plant. During this process, the Los Angeles Abrasion machine was filled with 30 kg of material and used 24 stainless steel balls with a 48 mm diameter and a mass ranging between 390 and 445 g, and a 30 rpm speed was set for two hours. The fraction obtained was sieved, and the particles that passed the Mesh 100 sieve (<0.15 mm) were called RCP. Table 1 shows the raw materials’ chemical composition obtained by X-ray fluorescence spectrometry (XRF) using the Rigaku model ZSX Primus IV instrument equipped with a Rh tube.

It was found that the main oxide present was silica (SiO₂). Its origin is mainly associated with the natural aggregates present in the concrete and secondarily with calcium oxide (CaO), which is related to the binder (hardened cement paste, lime, gypsum, etc.) (Ulsen et al., 2010; Vashistha et al., 2023; Wang et al., 2022). The chemical requirements were compatible with those established by NBR 12653 (ABNT, 2014) for use as a pozzolanic material.

Propylene glycol (PG), sodium hexametaphosphate (SHMP), and triethanolamine (TEA) are commonly used in the literature for grinding cement (Chipakwe et al.; Nthiga Njiru et al., 2023), whose characteristics are shown in Table 2, were used as grinding aids (GAs).

Grinding study

The recycled concrete powder (RCP) was mechanoactivated in a PM 100 planetary ball mill (Retsch), equipped with a 250 cm³ agate grinding jar, rotating at 500 rpm following the protocol: filling with 60 grams of powder, 50 agate balls with approximately 10 mm diameter, using an additive on the powder (Prziwara et al., 2018) and automated rotation direction reversal every 15 minutes with a 1-minute stop to cool down.

Propylene glycol was used firstly (1^st stage) as grinding aid (GA) (Costa; Gonçalves, 2022) at 0.05 and 0.1% of the RCP mass, quantities provided in the literature for cement grinding (Chipakwe et al.,2020; Nthiga Njiru et al., 2023), and grinding was carried out at 60, 120, 180 and 240 minutes. However, powder agglomeration was observed as a reduction on the specific surface area (SSA) occurred.

For this reason, two other grinding aids were included (2^nd stage): Sodium hexametaphosphate (Flores et al., 2017) at a 12.5 g/l concentration and Triethanolamine (Katsioti et al., 2009). The contents were increased to 0.5 and 1% of the RCP mass, and the grinding time was reduced to 30 and 60 minutes, according to the experimental design shown in Figure 2.

The RCPs from this study were characterized by mean particle size (d50) and specific surface area (SSA). The powders’ granulometry was assessed by analyzing the particle size distribution by laser diffraction using a granulometer (CILAS 1190) for the grain reading ranging between 0.04 and 2500 µm, in a liquid medium (distilled water), without the presence of a dispersant, and according to the Fraunhofer analysis method, with an approximately 25% obscuration. The methodology used to evaluate the Specific Surface Area (SSA) was nitrogen adsorption (BET - Brunauer-Emmett-Teller) in Quantachrome equipment, NOVA 3200 model. Before the test, the samples were degassed (using a vacuum at 40 ºC for 16 hours) (Scrivener; John; Gartner, 2018).

To analyze the grinding process’ efficiency, in the 2^nd stage the RCP’s grindability index (K) was calculated based on the specific surface area (cm²/g) generated and the energy consumed for 1 ton of material (kWh/t) ratio (Costa; Gonçalves, 2022; Katsioti et al., 2009; Von Krüger, 2004), according to Equation 1. The planetary mill’s power for this calculation was 750W.

Where:

K: Grindability Index given in (cm²/g)/ (kWh/t);

SSA: specific surface area (BET) in cm²/g; and

E: Energy consumed in Wh to generate 1 ton of recycled concrete powder.

Potential use of recycled mechanoactivated concrete powders as SCM

Given the time and additive content chosen in the grinding study, the following powders (Table 3) were selected for the study of their potential use as SCM.

Initially, the powders were characterized according to the tests described below and compared with CPV ARI, a high initial strength Portland cement equivalent to Portland cement type III (ASTM, 2016). CPV ARI was used because it is the one commercially available with the lowest addition content (<10% carbonate material) (ABNT, 2018). The specific mass was determined using a gas pycnometer (He), Ultrapyc 5000 model, from Anton Paar, with a 10 psi operating pressure.

The pH was assessed using a digital bench pH meter in a solution containing around 10± 0.1 g of each powder/CPVARI and 10 ± 0.1 g of distilled water, kept in a glass beaker and left to stand for one hour (shaken every 10 to 15 minutes) to stabilize the Ph (GEB, 2015).

Thermogravimetric analysis (TGA) and X-ray diffraction (XRD) verified the elemental and mineralogical composition. Thermogravimetric analysis was carried out using TA Instruments equipment, model SDT Q600, an alumina crucible with a sample mass of around 10 mg, a 30 ml.min^-1nitrogen flow, a 100 ºC to 900 ºC heating interval, and a 20 ºC.min^-1 equipment heating rate.

X-ray diffraction (XRD) was used to identify the RCP mineralogical composition after grinding. The samples were analyzed in a Panalytical diffractometer with Cu Kα λ = 1.54 incident radiation operating at 40 kV / 20 mA and a fixed divergent slit. The measuring range was from 5^o to 100^o, with an angular step of 0.026° (2θ) and 247 seconds per step. The diffraction patterns and Rietveld refinement were analyzed with Panalytical X-Pert HighScore Plus software (version 4.9) with Crystallography Open Database and Inorganic Crystal Structure Database (ICSD) files used to quantify the mineralogical phases using the following refinement parameters: scale factor, zero shift error, unit cell and peak shape parameters (W, V and U) using the pseudo-Voigt function. Finally, using the spherical harmonic function, the preferred orientation was refined for compounds that tend to orient themselves along the same crystallographic plane.

Pastes were produced with a water/fines ratio of 0.48 and replacement of 10% of Portland cement - CPV ARI with recycled concrete powders to evaluate consistency and compressive strength at 7 days, following the recommendations of standard NBR 7215 (ABNT, 2019). The materials were mixed in a mechanical mixer with a 1600 rpm rotation speed for 2 minutes (Figure 3a), and the consistency was evaluated by the mini-slump test (Kantro, 1980) (Figure 3b). Six specimens with 25 mm (diameter) and 50 mm (height) were molded (Figure 3c), unmolded, and placed in submerged curing saturated in lime (3g/l) until the test ages. On the scheduled date, the specimens were ground and broken using a Contenco I-3025-B hydraulic press with a 100-ton capacity (Figure 3d).

Results and discussions

Grinding study

The mechanoactivation effect on RCP’s d50 and SSA-BET with propylene glycol grinding aid in the 1^st stage is seen in Figure 4.

As seen in Figure 4a, mechanoactivation reduced the average particle size of RCP (d50=20.99 µm) as the grinding time increased, with values ranging from 5 to 12 µm, with the greatest particle reduction observed between 1h and 2 h using 0.05% propylene glycol. Overall, a d50 close to or smaller than that of Portland cement (17.83 µm) has positive effects on the cementitious materials’ mechanical properties (Aquino Rocha; Toledo Filho, 2023; Horsakulthai, 2021) and generally the finer it is, the more reactive it is, as the increase in specific surface area increases the reaction area and the availability of a high number of active atoms that easily bond to other atoms (Duan et al., 2020b; Tang et al., 2020). Figure 4b showed that although the grinding aid favored the SSA-BET, even when milled for 4 hours, all values were lower than the RCP (6.23 m²/g), suggesting that the material agglomerated and the propylene glycol content used was inefficient.

Therefore, to reduce energy consumption during grinding and ensure a higher SSA-BET than the RCP, higher levels of grinding aids (0.5 and 1%) and shorter grinding times (30 and 60 minutes) were evaluated. Two other grinding aids were also included (Sodium Hexametaphosphate and Triethalonamine), and the results for d50 and SSA-BET for the milled powders in the grinding study 2^nd stage are shown in Figure 5 (5a, 5b, and 5c).

Figures 5 (5a, 5b, and 5c) showed that for all the additives with a 0.5% GA content and grinding for 30 and 60 minutes, there was a simultaneous d50 reduction and an SSA-BET increase when compared to RCP (20.99 µm and 6.23 m²/g). There was no trend for the grinding aid content (Figure 5d), i.e., the d50 and SSA-BET characteristics depend on the combination of mechanoactivation time and GA content. It was noted that the contents and GAs used in the 2^nd stage contributed to particle dispersion and greater efficiency in the mechanoactivation process, given the higher SSA-BET values compared to powders without GAs. To complement this analysis, Figure 6 shows the grindability indices for 1 ton of recycled concrete powder, considering different mechanoactivation times, contents, and GA types.

It can be seen that for the two mechanoactivation times with the use of grinding aids, the process was efficient, with values higher than 8.13 cm²/g/ (kWh/t) for the 30-minute period and 4.37 cm²/g/ (kWh/t) for the 60-minute period. Taking into account that the higher the grindability index, the greater the process efficiency, the best results were obtained with 1% PG/30 min < 1% TEA/30 min and 0.5% PG/30 min. Other scenarios may be evaluated to improve the powders’ grindability index, such as using more balls or changing the mechanoactivation equipment. However, energy consumption must be considered as it directly impacts the RCP processing viability due to carbon emissions (Costa; Gonçalves, 2022).

In this sense, corroborating the findings of other authors (Aquino Rocha; Toledo Filho, 2023; Duan et al., 2020a, 2020b; Xu et al., 2021), it was considered that mechanoactivation for 30 minutes using 0.5% content for all grinding aids concomitantly guaranteed a lower d50 and a higher BET for RCP, with values close to CPV ARI (12.21 µm and 3.1 m²/g) and were therefore analyzed regarding their potential for use as SCMs.

Analysis of the potential for using RCP as SCM

The specific mass, pH, and a summary of the physical characteristics (d10, d50, d90, and SSA-BET) of the CPV ARI and the powders with 0.5% GA mechanoactivated for 30 minutes are shown in Table 4.

It was noted that the mechanoactivated powders had a specific mass ranging between 2.68 and 2.47 kg.m^-3, values lower than the CPV ARI, consistent with the values found in the literature and used as SCMs (Kim et al., 2023; Oliveira et al., 2023). The pH in the aqueous solution ranged between 12.27 and 12.43, indicating their high degree of alkalinity.

The raw materials mineralogical composition was identified by TGA/ Differential Thermal Analysis (DTA) and XRD/Rietveld (Figure 7). TG/DTA (Figure 7a) showed that the CPV ARI loss of mass was attributed to gypsum (110-145 °C), while in the recycled concrete powders, the hydrated phases decomposition corresponds to the tobermorite dehydration (not necessarily in crystalline form) and ettringite (50-200 °C); calcium sulfate dihydrate dehydration (110-145 °C); calcium hydroxide dehydration (380-500 °C) and the characteristic peak of portlandite and calcium carbonate decomposition (600-800 °C) (Deng et al., 2021; Dweck et al.,2000; Nobre, 2016). The loss on ignition (LOI) of recycled concrete powder mechanoactivated for 30 minutes was 8.54%, within limits for cement containing mineral additions (ABNT, 2018). The highest residual mass was found when the SHMP additive was used, with 90.57%.

Analyzing the positions of the Bragg peaks of the crystalline phases used and the reliability factor of Rietveld refinement (Rwp) and Goodness of fit (GOF) (Figure 7b), it is observed that mechanoactivation without and with grinding aids did not enhance the particles’ amorphism.

The Consistency index (CI) of the pastes (Figure 8 a) and the compressive strength of the pastes containing 10% replacement of CPV ARI by RCP, and by mechanoactivated powders for 30 minutes without and with grinding aids at 0.5% are shown in Figure 8b. The Z-Score was used to detect outliers, and the result presented is the average of 4 determinations. The data were analyzed using the SISVAR Software (Ferreira, 2011) and were subjected to the Tukey Test (p<0.05).

Observing Figures 8 a and b, it can be seen that the replacement of CPV ARI with recycled concrete powders caused an increase in compressive strength at 7 days of 22% when using RCP and 5% when using RCP-30 min-No GAs. The modification of the particle size distribution due to the presence of RCP may have reduced the water demand by replacing the water in the voids, and this additional water reflected in the improvement of workability (CI= 46.6 mm) (Hawkins; Tennis; Detwiler, 2005), and compressive strength at 7 days (46.6 MPa), corroborating the findings of Chen et al. (2021), who found no adverse effects on the pore structure of the cement paste when replacing up to 10% of the cement with RCP.

When analyzing mechanoactivation in the planetary ball mill, it was observed that the paste containing RCP-30min-No GAs presented an IC= 47 mm and a fc7=20.49 MPa, a behavior that suggests that the dilution effect that occurs when the cement content is reduced, it has been maximized (Wang et al., 2024).

Taking into account that when RCP particles are finer than cement, due to water absorption, a decrease in the compressive strength of cementitious composites may occur (Zhang et al., 2023), it was verified from the Loss on ignition (LOI), which means the presence of grinding aids in the particles, may have compensated for the increase in fineness and specific surface area (SSA) in comparison to CPV ARI, providing an increase in the compressive strength of the pastes, when compared the reference (CPV ARI). It is worth highlighting that the paste containing RCP-30 min-0.5%PG was the grinding aid with the lowest LOI than others, and the particle with the highest SSA-BET (7.50 m²/g) had a direct impact on the consistency of the paste (CI=45.8 mm).

Statistically, the additives influenced the compressive strength of the pastes (p<0.05), as indicated by the letters a1 and a2 (Figure 8b).

Finally, taking into account that the water/fine ratio was set at 0.48, adjusting water according to consistency can be a strategy to enhance the use of recycled concrete powders as supplementary cementitious materials, as already pointed out by John et al. (2019).

Conclusions

Based on the results, the following conclusions were drawn:

1. the use of a planetary ball mill combined with grinding aids contributed to the mechanoactivation of recycled concrete powders for use as supplementary cementitious material since, after 30 minutes, the Mean particle size (d50) was reduced, and the specific surface area (SSA) - BET increased simultaneously. However, there was no change in the material crystallinity;

2. the grindability index (K) as a function of the specific surface area (SSA) may be considered a way of evaluating the recycled concrete powders’ mechanoactivation efficiency with grinding aids; the higher the k, the greater the efficiency of the process. Thus, the best results were obtained with 1% Propylene glycol (PG), followed by 1% Triethanolamine (TEA) and 0.5% Propylene glycol (PG), both lasting 30 minutes;

3. considering the shorter time and the smaller amount of grinding aid, grinding for 30 minutes with 0.5% Propylene glycol (PG) was suggested for the recycled concrete powder mechanoactivation since the RCP with a d50 of 20.99 and an SSA of 6.23 m²/g had a 42% d50 reduction and a 20% SSA increase; and

4. the results in pastes indicated that grinding aids favoured the use of recycled concrete powder as MSC, having a statistically significant influence on the compressive strength at 7 days. Using mechanoactivated RCP for 30 minutes and adding 0.5% SHMP showed an increase of around 3 MPa compared to RCP (<0.15 mm). However, for RCP-30min-0.5%PG, due to its high specific surface area (SSA) - BET, there was a reduction of around 3.5 MPa caused by the increase in water demand.

Suggested future studies:

1. carry out rheology studies on pastes containing recycled powders from mechano-activated concrete with different types and levels of grinding aids;

2. evaluate the effect of grinding aids on the setting of cements containing recycled concrete powders as additional mechano-activated cementitious materials; and

3. produce pastes with a standardized consistency index and use superplasticizing additives to compensate for the adverse effect of the finer particle size of recycled mechano-activated concrete powders, enhancing their use in higher levels as supplementary cementitious materials.

Acknowledgements

To the Federal University of Latin American Integration (UNILA), the Itaipu Binacional Concrete Technology Laboratory and Mineromix concretos e pré-fabricados for supporting the experimental project. To Juliana Abatti Stopassoli for her laboratory support. To Thiago Ricardo Santos Nobre for his support in the Rietveld analysis; To the National Council for Scientific and Technological Development - CNPq - CNPq/MCTI/FNDCT Call No. 18/2021 - Band B - Consolidated Groups. To the Dean of Research and Postgraduate Studies PRPPG/UNILA for supporting the research.

References

Edited by

Publication Dates

History