Upcycling of demolition material from concrete and brick for the production of cold-bound, alkali-activated lightweight aggregates

This paper deals with the production of artificial aggregates based on the recycled fine fraction (≤ 4 mm) from construction demolition waste. Concrete powder, brick powder and their combination were used to produce aggregates through pelletisation using sodium silicate solution as an activator. For all aggregates, efficiency, bulk crushing resistance, particle density, water absorption and loose bulk density were evaluated. The bulk crushing resistance was evaluated for samples cured with different methods. A higher proportion of concrete powder increased the strength and density. The aggregates were successfully formed with bulk crushing resistance and particle densities in the range of lightweight aggregates.


Introduction
Concrete, the most used building material in the world [1,2], does not only impact the environment by the high carbon dioxide emissions due to the cement production process [1] but also because of the use of limited natural resources such as coarse and fine aggregates [2,3].
The increasing global population is affecting the demand for housing and infrastructure [3,4]; henceforth, concrete consumption is continuously growing. Moreover, older brick and building structures are being demolished and replaced by more modern buildings, thus, generated construction demolition waste is often dumped at landfills [4,5]. Looking forward, a collective solution lies in producing new building materials from the wastes of demolished buildings [6,7], leading to a sustainable circular system [8].
Up to 70% of all concrete volume consists of natural aggregates and these natural resources are facing a shortage in several parts of the world [9, 10]. Therefore, the interest in using artificial and recycled aggregates has grown significantly in recent years. In Germany, the recycled construction and demolition wastes (CDW) from building demolition can be separated into two types depending on the size of the aggregates [11]: the coarse fraction (C 4 mm) and the fine fraction (B 4 mm). The coarse fraction is re-used in new concrete or as base courses in road construction. Additional restrictions and requirements for the recycled aggregates are published in DIN 4226 [12] and EN 1045-2 [13]. However, these requirements do not apply to the fine fraction (B 4 mm) of the recycled material. According to the standards and due to missing recycling applications, the fine fraction of the recycled material is mainly dumped in landfills [10,14,15].
Several publications analyse the physical, chemical and mineralogical properties of fine recycled material [16][17][18], but only a few publications study the fine recycled fraction as an alternative for natural fine aggregates in new concrete. The state-of-the-art paper by Nedeljković et al. [15] critically discusses the opportunities and challenges of the deployment of fine recycled material (B 4 mm). Lower CO 2 emissions, a reduction of waste in landfills, scarcity of raw materials and lower costs are addressed as drivers for the deployment of fine recycled aggregates. Moreover, depending on the composition of the original material, this fine fraction shows a wide range of physical and chemical properties, which cannot be easily standardised. Additional separating and sorting to extract the fine aggregates would increase the costs significantly, thus challenging the deployment of fine recycled aggregates. The review by Villagrán et al. [19] discusses similar challenges for valorising the fine recycled fraction but emphasises the feasibility. Evangelista et al. [20] provide not only an overview of the use of fine aggregates from CDW but also monitored the mechanical behaviour [21] and durability performance [22] of concrete made with fine recycled concrete aggregates. Special attention is given to the increasing water absorption, which highly affects workability [20,22]. However, it needs to be emphasised that lab-made materials were used in these studies. The consensus in the literature is that the production of concrete made of fine recycled aggregates is feasible but impacts its properties due to lower densities and compressive strengths.
The production of artificial aggregates commonly involves sintering or cold-bounding method and was extensively investigated by Qian et al. [23] and Ren et al. [24]. Sintering generally yields improved physical and mechanical properties, but it comes at the cost of higher energy consumption and increased CO 2 emissions [25]. Conversely, the cold-bound technique shows promise regarding economic and environmental factors, as it has the potential for reduced energy consumption [25,26]. Hence, in this study, the cold-bounding method is employed to explore its advantages.
Another topic addressed in the current work is alkali activation. Construction and demolition waste (CDW) is composed of silica and alumina, making it suitable for alkali activation. The research published until now focuses on utilising various solid precursors to create alkali-activated aggregates, as intensively studied in the review papers of Qian et al. [23] and Ren et al. [24]. As an example of previous research findings, Rehman et al. [6] generated alkali-activated aggregates by combining ground-granulated blast furnace slag and fly ash, resulting in loose bulk densities ranging from 764 to 876 kg/m 3 . In a separate study, Tian et al. [7] achieved loose bulk densities of 1008-1132 kg/m 3 for alkali-activated aggregates made from red mud and fly ash. Gomathi et al. [27] manufactured alkali-activated aggregates using fly ash and bentonite, yielding loose bulk densities of 950 kg/ m 3 . Shi et al. [14] used concrete waste powder to produce alkali-activated aggregates with loose bulk densities between 814 and 1010 kg/m 3 . Tang et al. [28] achieved loose bulk density of 980 kg/m 3 made of bottom ash fine particles. All these aggregates are lightweight aggregates based on their loose bulk densities. While some publications [29][30][31][32][33][34] have explored CDW as a precursor for alkali activation, none have investigated the production of artificial aggregates using CDW. However, the production of alkali-activated artificial aggregates utilising CDW remains an unexplored area of research. This publication aims to fill this research gap by investigating and addressing this specific study area.
This paper aims to provide a production path for lightweight aggregates consisting of recycled fine material from an actual recycling yard through the production of alkali-activated lightweight aggregates using fine brick and fine concrete recycling waste. The two recycling materials with B 4 mm were sourced from Berlin recycling yards, which are currently disposed due to the lack of recycling applications. Another motivation behind this research is the need to investigate alternative materials with focusing on sustainability, particularly those derived from secondary resources that can be consistently sourced in the long run. Given the enduring availability of this untapped material, it becomes logical to employ it purposefully. Therefore, these materials were investigated as potential precursors for alkali activation.
Sodium silicate solutions with four different silica moduli (SiO 2 /Na 2 O) in the range of 0.7-1.3 mol/mol and five different water-to-sodium ratios (H 2 O/ Na 2 O) in the range of 10-50 mol/mol were used as an alkaline activator. Out of the 40 tested compositions, one composition was chosen to produce the artificial aggregates via the pelletisation process, and the resultant aggregates were cured with three curing methods. Several tests regarding the physical and mechanical properties, such as water absorption, bulk crushing resistance, loose bulk density and particle density, were performed on the chosen aggregate samples according to the European standard of lightweight aggregates [35]. This paper evaluates alkali-activated aggregates made of fine recycled construction waste according to the European standard of lightweight aggregates.

Materials
In this section, the starting materials are introduced. The materials in this study include two different construction demolition wastes and 20 different activators.

Concrete and brick demolition waste
The concrete and brick demolition waste were obtained in the size fraction 0-4 mm. It is essential to highlight that these were no lab-made materials but mixed demolition materials from actual construction demolition sites. The material was crushed and sorted at the recycling yard by a combination sorting system with optical sensors, inductive sensors, and X-ray transmission to detect colours, shapes, metals and density differences, respectively. The colour sorting had an error rate of about 20%. Following this, the material was sorted by size, of which the smallest fraction B 4 mm was used for further research.

Alkaline activator
In this study, a combination of sodium hydroxide (NaOH) and sodium silicate solution was applied. A sodium-based water glass was selected due to its affordability and widespread availability [36]. Previous research, such as Provis et al. [36] and Bondar et al. [37], demonstrated successful outcomes using this type of water glass solutions. The sodium silicate solution (Betol 52 T) was obtained from Woellner GmbH (Germany). To prepare the activation solution, NaOH pellets of 99 wt.-% purity, obtained from VWR International GmbH, were initially dissolved in deionized water to prepare the NaOH solution. After that the NaOH and sodium silicate solution were mixed to give different silica moduli (SiO 2 /Na 2 O, mol/mol). The silica moduli studied in this paper were 0.7, 0.9, 1.1. and 1.3 mol/mol. Five different water-tosodium ratios (H 2 O/Na 2 O) in the range of 10-50 mol/mol were studied. Table 1 lists the chemical concentrations of all possible compositions.

Methods
In this section, the methods used are briefly presented. Mineralogical analysis was performed on the powder and the crushed aggregates. The physical and mechanical testing was only applied to the aggregates.

Mineralogical analysis
Blaine fineness, which describes the specific surface area of the powder, was determined by Blaine airpermeability method according to EN 196-6 [38]. Particle size distribution, including d 50 and d 90 of the powders, was measured by laser granulometry using Mastersizer 2000 from Malvern Instruments. The density of the powder was measured by a helium pycnometer. The chemical composition was determined by X-ray fluorescence (XRF) spectroscopy. X-ray diffraction (XRD) analysis was performed using an Empyrean diffractometer, PANalytical, Netherlands, equipped with CuKa radiation-emitting X-ray tube (k = 1.54 Å ). The samples were measured for 60 min in the range of 5°-65°with an automatic aperture. Thermogravimetric analysis (TGA) was performed using TG 209 F3 Tarsus (Netzsch-Gerätebau GmbH, Germany) under a nitrogen atmosphere. The analysis was carried out with 10 ± 1 mg of the samples and heated from 25 to 850°C at a heating rate of 10°C/min while the weight loss was logged. Lowvacuum scanning electron microscope (SEM) analysis was performed using a GeminiSEM500 microscope (Carl Zeiss Microscopy Deutschland GmbH, Germany). 100 Pa of gas pressure in the sample chamber and a voltage of 15 kV were adjusted to examine the samples. Additionally, elemental mapping was performed using energy-dispersive X-ray analysis (EDX). Before the investigation, the samples were manually broken and fixed by conductive carbon adhesive tape onto a sample carrier.

Physical and mechanical testing
Particle density, water absorption and loose bulk density were measured after 28 d of curing at 20°C in the air-conditioned laboratory. Oven dry particle density and water absorption were determined within one investigation according to EN 1097-6, annexe C [39]. For this investigation, the aggregates were stored in water in a pycnometer for 24 h. Afterwards, water absorption was determined by drying the material to constant mass in an oven at 105°C and weighing it before and after drying. Oven dry particle density was calculated by weighing the aggregates in different moisture levels according to EN 1097-6 [39]. Loose bulk density was determined by weighing aggregates in a cylinder of known volume according to EN 1097-3 [40]. Bulk crushing resistance was measured according to EN 13055, annexe B [35]. Overall, 7, 28 and 90-d bulk crushing resistance were determined with three different curing regimes.

Experimental plan
The experimental setup was divided into three parts. The first part describes the production of concrete powder and brick powder. In the second part, the reaction between the different sodium silicate solution and the two powders as precursors were tested on cubes. The compressive strength of alkali-activated cube samples was determined to find the appropriate composition of the sodium silicate solution for the recycling materials. The third part presents the preparation of the alkali-activated aggregates with the determined composition of the sodium silicate solution.

Production of concrete powder and brick powder
Both waste materials were milled to comparable Blaine fineness of around 3200 cm 2 /g. This fineness was achieved by milling the material in a vibratory disc mill (RS200, Retsch GmbH, Germany). The milling time differed to reach comparable finesses. Fineness of standard Portland Cement was used as reference value. 150 g of material was milled at a speed of 1200 rpm to obtain a reproducible fraction of the material. Brick demolition waste was milled for 30 s, whereas concrete demolition waste was milled for 20 s.

Production of alkali-activated cube samples
Alkali-activated cube samples made of sodium silicate solution with different silica moduli and recycling powder CP and BP were investigated in an iterative process. The literature shows that an optimum silica modulus should be determined to obtain the highest compressive strength of the pastes [41]. Therefore, 20 compositions of sodium silicate solution were tested for each material in this study. The mixtures were tested on small cubes with dimensions of 2 9 2 9 2 cm 3 . For the preparation of all cube compositions, 50 g of the solid (powder) and 10 g of the liquid (sodium silicate solution) were mixed with a laboratory hand mixer. After 60 s of mixing, the paste was 30 s at rest and then stirred again for 30 s. Afterwards, the paste was poured into the moulds and compacted for 60 s on a compaction table. The samples were sealcured at 20°C until the testing day.

Production of alkali-activated aggregates
The composition of the aggregates was based on the optimum composition of the cubes. The same silica modulus and liquid-to-solid ratio (l/s ratio) were chosen for the preparation. A pelletizer disc was used to produce the aggregates, as described in the research of Tajra et al. [25,42,43]. The aggregates were prepared by mixing each CP and/or BP (250 ± 20 g) with the sodium silicate solution (50 ± 5 g) for about 15 s at the highest stage of the mixer; this resulted in small granulates. When the surface of the granulates was homogeneous, the small granules were transferred to the pelletizer disc. The most effective parameters for the pelletisation process were based on Tajra et al. [25], i.e. an angle of 35°and a rotation speed of 30 rpm. In the pelletizer disc, the small granules were concurrently sprayed with sodium silicate solution (10 ± 2 g) and encased with the powder (50 ± 5 g). After 15 ± 3 min of pelletisation, the pellets reached their desired size of 6 mm in average diameter. Three different kinds of aggregates were produced: concrete powder aggregates (CPA), brick powder aggregates (BPA) and a combination of concrete and brick powder with 50% of each powder labelled as concrete brick powder aggregates (CBPA). For the aftertreatment, the pellets needed to be carefully taken from the pelletizer disc and stored in plastic bags in a climate room at 20 ± 1°C. Data obtained from previous studies by different authors [27,30,[44][45][46] showed that the time and temperature of the curing regime enhanced the properties of waste-derived alkali-activated products. From the literature, it emerged that heating at an appropriate temperature improved the quality of the alkali-activated materials by reducing porosity and enhancing its strength [34,47,48]. Lower temperatures hardly accelerated the reaction, whereas higher temperatures led to shrinkage, as the material dried out too quickly [47][48][49]. Therefore, based on the literature, 40°C was chosen for curing in this study. Additionally, it was reported that long curing periods of alkaliactivated systems led to the formation of further strength-giving products [50]. Jha and Tuladhar [30] determined a linear correlation between time and compressive strength for alkali-activated materials cured at 40°C. Therefore, two different curing times were investigated. One sample batch without heat treatment was used as a reference. Following the literature, three curing regimes were chosen: a. The aggregates were cured at 20°C (no heat treatment).
b. The aggregates were cured at 40°C for 24 h and then at 20°C until the testing day. c. The aggregates were cured at 40°C for 7d and then at 20°C until the testing day.
The three curing conditions and compositions of the samples are given in Table 2. The aggregates were stored in sealed bags until the testing day to avoid moisture loss.

Results
The result section is divided into three parts following the structure of the experimental plan: the properties of concrete and brick powder, alkali-activated cube samples and alkali-activated aggregates.

Properties of concrete and brick powder
The powders were evaluated regarding their physical, chemical and mineralogical properties. The physical properties, such as densities, Blaine fineness and laser granulometry, are shown in Table 3. The powders had comparable specific surface area measured by Blaine. Concrete powder was coarser in the particle size distribution, as shown in Table 3 and Fig. 1. Comparable Blaine Fineness was targeted, as it is not possible to mill the material to comparable particle size distribution. The raw materials exhibited varying degrees of grindability due to their different mineralogical composition, resulting in distinct particle size distributions. Within the fine fraction of concrete, for example, there was a combination of quartz particles that were challenging to grind and cement paste that was comparatively easier grindable, leading to variations in the fineness of specific components. Moreover, as the Blaine air-permeability method is commonly practiced in the cement industry, Blaine fineness was regarded as the desired target value. It should be, however, noted that in this scenario, the diverse particle size distribution could not be eliminated.
The chemical composition in Table 4 shows that both materials were rich in silicon. CP contained a minor amount of calcium, most probably from the old cement matrix [16]. The SO 3 content of CP with 0.2 wt.-% was within the limit values of recycling material type 1 [12]. BP additionally contained a higher amount of aluminium, which could be traced back to burned clay particles. The SO 3 content of BP with 0.5 wt.-% was within the limit values of recycling material type 2 [12].
XRD measurements were conducted with the powders. Figure 2 shows the diffractograms of CP and BP with the associated phases. Comparing the     The main component of the sample CP was quartz, which is conclusive with the high silicon content determined by XRF. The occurrence of quartz was attributed to natural aggregates. The second most common mineral was calcite. It potentially results from the carbonation of hydrated cement. In addition, hatrurite (alite) was possibly detected in CP. The dominant phases in BP were also quartz and calcite, as seen in Fig. 2. Quartz can be due to the presence of quartz in the original clay material or due to industrial sorting of the material. Minor phases such as hatrurite (alite) and hematite (iron oxide) were potentially detected. The presence of hematite agrees with the higher iron content in XRF results. It must be pointed out that important components, such as the strengthgiving C-S-H phases, are largely X-ray amorphous and, therefore, cannot be detected with XRD. For further investigations, TGA was performed on the powders. Figure 3A and B present TGA measurements with thermogravimetry (TG) and derivative thermogravimetry (DTG) results of CP and BP. For CP, the results indicated weight loss in three stages. The first stage was observed between 50 to 150°C. This stage was characteristic of the evaporation of physically bound water of cement stones hydrate phases [41,51,52]. In the second stage, weight loss was located from 400 to 500°C. This stage was typical for the thermal dehydroxylation of Ca(OH) 2 (portlandite) [41,52,53]. The last weight loss stage was located from 650 to 720°C. This peak was assigned to the decomposition of carbonates. The carbonates were either attributed to carbonated portlandite or the raw material cement and/or aggregates already contained limestone or limestone filler [52,53]. In total, a loss of about 9 wt.-% was detected.
For BP, the results show weight loss in two stages. The initial weight loss corresponded to the evaporation of physically bound water. The second weight loss from 650 to 720°C was attributed to the decomposition of carbonates. In total, a loss of about 6 wt.-% was detected. Figure 6A and C show SEM images of CP and BP, respectively. Additionally, combined elemental maps made with energy-dispersive X-ray analysis (EDX) of aluminium, silicon and calcium for CP and BP are shown in Fig. 5B and D. The SEM image of CP in Fig. 4A shows particles with irregular shapes and surfaces. These can presumably be identified as cement stone and natural aggregates due to the high calcium and silicon and low aluminium contents.
SEM image of BP in Fig. 4C shows particles of different sizes, assumingly clay-based particles. Burned clay particles are usually rich in aluminium, which explains the higher aluminium content in the image. The roughness on top of the particle in the middle suggested a sintered surface. Underneath the sintered surface, layers were recognized. This layered structure was presumably assigned to the layers of the clay structure. The calcium content was traced back to cement stone due to sorting.

Properties of alkali-activated cube samples
The cubes were classified regarding consistency, efflorescence and compressive strength, as documented in Fig. 5. Figure 5 shows the results of the 20 different compositions tested on two powders. The first step was to categorise the consistency of the paste. There were 5 categories: dry-dry, dry, good, moist and wet. When the consistency was identified as dry-dry or wet, the moulds could not be filled without gaps, or the paste was not reactive after many days of casting and, Compositions of sodium silicate solution tested on 2 9 2 9 2 cm 3 cubes with concrete and brick powder as precursors therefore, too plastic and could not be demoulded. In the dry, good and moist categories, the cube forms could be filled and demoulded after 7 d. After that the efflorescence on the cubes was evaluated. The degree of efflorescence was categorised as none, partial and areal efflorescence, as shown in Fig. 6.
In the case of areal and partial efflorescence, the compressive strength of the cubes could not be tested because the samples were very soft after the formation of efflorescence. The cubes had already disintegrated on their own due to the crystallisation pressure of salts inside; hence compressive strength of these cubes was below the detection limit of the testing device. The compressive strengths of the intact cubes showed that the compressive strength first increases with increasing silica moduli and then decreases again (Fig. 5). As reported in the literature [37,41,54], the evaluation supported that lower silica modulus and lower waterto-sodium ratios led to higher alkali concentration and caused efflorescence and brittleness. Thus, excess alkalis reduced compressive strength by causing efflorescence. In contrast, silica moduli that were too high again reduced the compressive strength, as seen in Fig. 5. Increasing the silica modulus to reach higher compressive strength was also not recommended by the literature [41].
The cubes without efflorescence were tested in quadruple repetition. The highest compressive strength was measured for specimens CP and BP with a silica modulus of 1.1 for SiO 2 /Na 2 O (mol/mol) and 20 for H 2 O/Na 2 O (mol/mol). The best composition was tested again for 7 and 28-d strength, showing the mean of quadruple repetition, as seen in Table 5

Properties of alkali-activated aggregates
The following investigations examined the main properties of aggregates made of recycled material. Figure 7A shows CPA and Fig. 7B shows BPA in their designated size. Various mechanical and mineralogical tests were performed on the aggregates, as explained below. The results are shown in Figs. 8 and 9.   Figure 8 presents the bulk crushing resistance for all aggregate samples. The diagram shows the mean values of three tests with the associated standard deviations. At this point, it should be emphasised that this method must be considered with care. Some trials, such as BPA_20 90 d, showed strong deviation with the same method. According to Thienel et al. [55] and EN 13055:2016-11 [56], this method was intended only for production control and internal comparison [56]. The reason for employing this method is that it aligns with recent literature and European standards.
Nevertheless, a notable trend of high standard deviation was observed in bulk crushing resistance of brick powder aggregates (BPA). This deviation aligns with the evident higher variations observed in loose bulk density and particle density, as illustrated in Fig. 9B, and C, respectively. These findings suggest that the microstructure of this aggregate group lacks stability and therefore displays a tendency towards being heterogeneous.
Although all nine samples showed the same trend of increasing strength with an advancing time of 7 d to 90 d, nearly all samples achieved about 70% of their bulk crushing resistance after 28 d. Only the samples CBPA_40-1 and CBPA_20 achieved less than 70% of their bulk crushing resistance after 28 d. In contrast to Jha and Tuladhar et al. [30], who observed a linear increase of compressive strength over time, the results shown in Fig. 8 show a more intense initial reaction that slowly continued over time.
Moreover, it emerged that the curing period had a particularly strong impact on strength. The highest strength in all tested combinations was achieved by the curing method iii. 40°C for 7 d. For example, CPA_40-7 achieved a mean bulk crushing resistance of 3.7 MPa after 28 d, whereas CPA_40-1 and CPA_20 reached 2.4 and 1.6 MPa, respectively. These results lead to the assumption that curing at 40°C does not hamper alkali activation.
The mean strength of BPA_40-7 after 28 d (2.3 MPa) was significantly lower than CPA_40-7 (3.7 MPa). The mixed sample CBPA_40-7 reached values between the pure concrete or brick powder aggregates (3.4 MPa). This indicates that the strength of CBPA is likely due to a larger share of the concrete powder. Figure 9A shows the production efficiency determined by two ratios for CPA, BPA and CBPA samples cured for 7 d at 40°C. One defines the mass of the output of aggregates related to the mass of raw material input. It was calculated as follows: The second ratio describes the output of aggregates with the desired diameter of 4-8 mm related to the input of raw material, as only 4-8 mm aggregates were used for further testing. It was calculated as follows: 4 À 8 mm efficiency % ð Þ ¼ Output of4 À 8 mm aggregates g ð Þ Input of raw material g ð Þ CPA achieved a 4-8 mm efficiency of 83%, BPA about 72% and CBPA around 77%. Following the literature [57,58], smaller size of particles in the material can be captured more easily by aggregates, because of the larger specific surface area. This trend was not confirmed by this material, as BP consisted of smaller particles (Table 3) but captured less material. The total efficiency of the aggregates produced through pelletisation was very similar for the three aggregate types and ranged between 92 and 96%. Figure 9B shows the loose bulk density of the aggregates. It was located between 1112 kg/m 3 for CPA, 990 kg/m 3 for BPA and 1008 kg/m 3 for CBPA. Therefore, the manufactured aggregates were classified as lightweight aggregates following EN 13055 [35], where loose bulk density is restricted to B 1200 kg/m 3 [55]. Even though, the density of BP is higher than of CP, as listed in Table 3, loose bulk density of BPA is lower than of CPA. A correlation between efficiency and loose bulk density can be found; hence if more powder material is consumed, a higher productivity and thus higher densities were achieved. It can be presumed that the higher efficiency results in Fig. 7 A: concrete powder aggregates and B: brick powder aggregates a denser structure. This also impacts the oven-dry particle density shown in Fig. 9C and the water absorption illustrated in Fig. 9D. CPA achieved values up to 1.93 g/cm 3 for particle density and 10 wt.-% water absorption; CBPA reached 1.79 g/cm 3 and BPA reached 1.72 g/cm 3 of particle density with a water absorption at 14 wt.-% and BPA at 19 wt.-%. Higher water absorption can be explained due to lower densities and, therefore, higher porosity, as confirmed by other researchers [25,26]. The restriction for particle density of lightweight aggregates was limited to B 2.0 g/cm 3 in EN 13055 [55]. Hence, the aggregates were classified as lightweight aggregates concerning particle density following regulations of [55]. The results of particle density and loose bulk density show the same trend. A significant correlation between the properties can be summarised as follows: reduced efficiency leads to a looser packing of particles, resulting in increased porosity. This higher porosity, in turn, leads to lower overall densities. Consequently, the lower particle and looser bulk densities of artificial aggregates contribute to a decrease in bulk crushing resistance. Furthermore, the elevated porosity and lower densities also result in higher water absorption. Any alteration to a single parameter in the production process of the aggregates affects all these properties collectively. Table 6 presents the literature results of the physical and mechanical properties of the aggregates. Only little research is available regarding the production of cold-bound alkali-activated aggregates made from construction demolition waste. Therefore, the results of using fly ash, blast furnace slag, red mud are also summarised for a comparison. The results show that the aggregates made of concrete and brick powder have comparable strengths and loose bulk densities to other precursors; but due to varying precursors and activators, it needs to be compared with care. As the chemical and mineralogical composition of the precursor largely impacts its reactivity and resultant physical and mechanical properties [41]. Only Jha and b Fig. 8  Based on the present findings, it is evident that concrete powder exhibits higher reactivity than brick powder, primarily due to its superior densities and resistance to bulk crushing. Surprisingly, despite concrete powder being coarser than brick powder (Fig. 1), the alkali activation proves more successful for concrete powder. This contradicts the common understanding in literature [57,58] that smaller particles generally possess a larger specific surface area for enhanced material capture. Nevertheless, in this case, the geopolymerisation of concrete powder surpasses that of brick powder. It can be assumed that the chemical and mineralogical composition of the raw material defines its reactivity. Concrete powder is more reactive than brick powder due to its composition as shown in Table 4, even when particle size is coarser than of brick powder.
According to this literature [49,[59][60][61], authors have investigated that phase composition of C-N-A-S-H type gel and/or zeolites have been formed in systems similar to the system investigated in this studies. These phases may primarily be responsible for the strength in the system. Further investigations, such as TGA and SEM are necessary to identify the strength giving components. However, the exact chemical composition of the gel phases is beyond the scope of this paper. b Fig. 9 Properties of manufactured concrete powder aggregates (CPA), brick powder aggregates (BPA) and concrete brick powder aggregates (CBPA) after 40-7 curing. A: efficiency, B: loose bulk density, C: particle density, D: water absorption

Conclusion
This study investigated the properties of recycled fine fractions from concrete and brick demolition waste provided by Berlin recycling yards. The materials were alkali-activated and processed into aggregates. The aggregates were studied regarding their properties. The studies showed that aggregates could be formed by composing recycled fine powder with sodium silicate solution with molar ratios of 1.1 for SiO 2 /Na 2 O (mol/mol) and 20 for H 2 O/Na 2 O (mol/mol). Loose bulk density of the aggregates achieved values of 990-1110 kg/m 3 . Oven-dried particle density ranged between 1.72 and 1.95 g/cm 3 . Therefore, the produced aggregates can be termed lightweight aggregates. Aggregates that contained concrete powder achieved higher densities than aggregates with brick powder. Water absorption ranged between 10 and 19%. Lower densities resulted in higher water absorption. Pelletisation efficiencies for 4-8 mm aggregates varied between 83 and 72% for concrete powder aggregates and brick powder aggregates, respectively. Higher efficiency was achieved for concrete powder aggregates even though the particle size (d 90 or d 50 ) of the raw concrete powder was coarser than that of brick powder. Bulk crushing resistance of the aggregates was in a range of 0.4-4.5 MPa after 28 d, with higher resistance for concrete powder aggregate than for concrete brick powder aggregates and only brick powder aggregates. A correlation between the content of concrete powder and the bulk crushing resistance has been found. Curing temperature and time impact the bulk crushing resistance. The highest resistance in all testes combinations was achieved with the curing of 40°C for 7 d.
The methodology shows that artificial lightweight aggregates can be produced from construction demolition waste. The material can be recycled and thus conserve the use of finite resources. Additional research is required to explore the durability of the aggregates in greater detail. In-depth investigations of the reaction kinetics of the materials should be conducted using different analytic methods on the aggregates. Furthermore, since the utilisation of alkali-activated aggregates in mortar or concrete is still limited, further research is necessary to examine their practical performance.