A Preliminary Laboratory Evaluation of Artificial Aggregates from Alkali-Activated Basalt Powder

Abstract

The widespread use of natural aggregates is one of the main causes of the depletion of natural resources, as aggregates are constituents of several construction materials. Alternatively, it is, today, proven to be feasible to use mining tailings, either natural or recycled materials, to produce artificial aggregates through specific processes. A possible way to produce artificial aggregate is through the alkali activation of the powdered material in a process called geopolymerization. This study proposes to use a basalt powder and two different metakaolins as precursors for the production of an alkali-activated artificial aggregate, with a specific shape and size achieved by using 3D-printed molds. The experimental aggregates were evaluated using traditional tests for natural aggregates, such as resistance to compression, specific density and resistance to abrasion and fragmentation. Furthermore, the material was chemically analyzed in order to evaluate the geopolymerization process promoted by the two adopted metakaolins. The physical tests showed that artificial aggregates do not perform well in terms of resistance to wear and fragmentation, which can be improved. However, they revealed promising results in terms of skid, polishing and micro-texture.

1. Introduction

Natural aggregates mainly consist of crushed rock and sand coming from crushing bedrock or possibly from unconsolidated sand and gravel [1]. The widespread use of natural aggregates is one of the main causes of the depletion of natural resources, considering that aggregates are the basic constituent of several construction materials, such as bituminous and cementitious concretes. According to estimates for the year 2015, around 48 billion tons of natural aggregates was used worldwide, with a predicted growth rate of 5% every five years [2]. Thus, finding alternatives to natural aggregates is becoming crucial in the current scenario of sustainable transition that involves the construction industry [3,4]. The increasing need for sustainable construction materials is driving the research towards the reuse and recycling of waste to produce “recycled”, “manufactured” and “artificial” aggregates [5]. The latest version of the EN 13242 standard [6] lists, in fact, recycled and manufactured aggregates among the aggregates that can be used in civil engineering work and road construction.

Recycled aggregates are classified as a result of the processing of inorganic or mineral material previously used in construction, while the manufactured ones come from an industrial process involving thermal or other modifications. To date, the most common example of recycled aggregates is represented by the Construction and Demolition Wastes (CDWs), which are processed to obtain recycled aggregates to partially or completely substitute the natural aggregates for the production of construction materials [7]. On the other hand, there are several examples of manufactured aggregates, such as various slags or materials produced, involving thermal treatment of pumice, volcanic ashes, clays or siliceous rocks [8,9]. In the last few years, in addition to these mineral-based manufactured aggregates, an increasing number of by-products or wastes have been processed and treated to produce artificial aggregates (AAs) [10]. Unlike the processing of mineral-based materials, most of the solid wastes or industrial by-products must be pre-treated in order to eliminate the presence of potentially harmful substances. As an example, Incinerator Bottom Ashes (IBAs) contain heavy metals, chlorides, sulfates and other pollutants and must be pre-treated accordingly to be safely recycled and used as AAs [11]. However, despite the relatively long and expensive pre-treatment, the production of AAs contributes to the reduction in the exploitation of raw materials and represents potential economic and functional values for by-products and wastes [12,13]. Several by-products from different fields have been studied and experienced for the production of AAs, from coal bottom ash to crumb rubber, plastic waste and waste glass [14,15,16,17,18,19], and many others are still under investigation.

In terms of production methods, AAs generally come from a two-step procedure: granulation and hardening process [10]. The granulation process is widely adopted in different sectors (construction materials, food industry, pharmaceutical, etc.) and consists of the agglomeration of powdery substances to form grains or granules. The granulation technology for AA production generally requires the addition of liquid (wet granulation process) to create bonds between particles strong enough to bind them together and so enlarge their sizes. The following hardening step is needed to solidify the agglomerations to achieve specific mechanical properties. Based on the type of powder and the final application of the granules produced, different hardening technologies can be adopted. According to the most widespread applications of AAs, the most common hardening procedures are: sintering, cold bonding and alkali activation [10]. In the first process, the AAs are produced through a thermal treatment of the fresh granules at very high temperature (up to 1000 °C). The material undergoes an expansion and vitrification process, which allows for the strengthening of the granule. With the sintering being a chemical and physical process, the quality of the final AA, in terms of mechanical properties, is directly related to the chemical composition of the base material [19]. As a downside, this method has a significant impact in terms of emissions and energy consumption, due to the high temperature needed to run the process. On the contrary, the cold-bonding hardening technology is widely adopted considering the possibility to cure the AAs in water or at constant humidity. In this case, the strength of the granule is given by a pozzolanic reaction [20]. As a consequence, not every material is suitable for this process, with the pozzolanic reaction directly affected by the reactivity of the base material. The chemical reaction is also the core of the third mentioned hardening method: alkali activation. This is a chemical synthesis that occurs between silica and alumina-rich-based materials and strong alkali solutions, which allows for the development of ceramic materials [21]. In the production of AAs through this method, alkaline solutions, such as blends of sodium hydroxide and sodium silicate, are added as a binding agent as water replacement during the granulation process. Considering the relatively easy process and the absence of high temperatures for curing, the alkali activation of AAs is gaining popularity. Moreover, several studies verified the possibility of achieving high-performance AAs in terms of mechanical properties [22,23]. The application of geopolymeric AAs in cement concrete, in total or partial substitution of natural aggregates, is a current practice for buildings and pavements [24,25,26,27]. However, several applications have been carried out with the aim to substitute not only natural aggregates in construction materials but also add extra value to the final product. Fang et al. verified the feasibility of porous alkali-activated AAs carried for Phase Changing Material (PCM) to produce cement concrete to improve the energy efficiency of buildings [28] or asphalt concrete for pavements [29].

Following this trend of research, in the present paper, a preliminary laboratory characterization of AAs produced through the alkali activation of a basalt powder is presented. Even if the use of basalt for the formation of alkali-activated material has already been studied by different authors [30,31], the novelty of the present research is represented by the production process of AAs. The proposed basalt alkali-activated AAs are, in fact, cured in specific molds created with a 3D printer in order to have defined shapes and sizes, conceived to improve friction and acoustic properties when applied in the construction of road pavement wearing courses.

2. Materials and Methods

The geopolymer material is obtained from a mix of activators and precursors. Activators are alkaline liquids responsible for dissolving the precursors’ structure and forming a brand-new structure. Precursors are usually aluminosilicate powders that, when dissolved in an alkaline solution, react and, depending on their origin, will provide different characteristics and proprieties to the final material. The geopolymer design might follow the trial-and-error method [32], mixing different precursors and activators’ ratios, when the materials used as precursors constitute a heterogeneous and complex system and the involved chemical reactions are hard to predict in advance. However, it is possible to use theoretical basis to have a specific mix that allows for the geopolymers’ production based on the chemical constitution of activators and precursors. In the following paragraphs, a basic characterization of the materials used in the experimental application is presented.

2.1. Precursors

Different precursors have been used in various research studies, among them fly ashes [33], metakaolin [34,35] and blast furnace slags [36]. The type of precursor has a direct impact on many mechanical and chemical properties of the geopolymer. Inert waste powder could act as a filler/aggregate in the geopolymer structure. On the other hand, a reactive amorphous waste could act as a precursor in the geopolymer matrix. Therefore, studying the precursors’ mineralogy is vital to understand its constituent species and possible reactivity [37]. The materials used in this research as precursors were metakaolin and basalt in the shape of powder.

2.1.1. Basalt

Basalt powder is a by-product in the mining industry that has some commercial value. It can be used as a filler material [38] in fiber production or as a reactive part of the geopolymer mix [39,40]. In this research, the basalt powder used has 100% passing on the 0.063 mm sieve, with the following mineralogical composition, obtained by XRD analysis (

Table 1. Chemical composition of the basalt powder (Reference Intensity Ratio (RIR) method).

Name	Composition	Percentage (%)
Leucite (L)	(K(AlSi2O6))	44
Augite (Au)	((Ca,Mg,Fe)2Si2O6)	22
Anorthite (An)	(Ca(Al2Si2O8))	11
Orthoclase (O)	(K(AlSi3O8))	5
Muscovite (M)	(KAl2(Si3Al)O10(OH)2)	5
Magnesiohornblendeferroan (Mh)	(Ca2(Mg4Fe3+) (Si7Al)O22(OH)2)	4
Magnetite (Mt)	(Fe2+Fe3+O2)	1

).

The XRD of the Basalt powder is shown in Figure 1, where it is possible to see the intensity of each material, as described in Table 1.

Several researchers used basalt as a geopolymer material [31,37,38]. The data presented here demonstrate that the basalt powder has a sufficient amount of aluminum and silica. These indicate that the precursor is adequate to generate a geopolymer, as these minerals are responsible for the final geopolymer structure [32,39].

2.1.2. Metakaolin

Calcined kaolinitic clays, otherwise known as metakaolin (MK), were one of the first precursors used in geopolymer research. MK initial applications were mainly in paper and plastic industries, where it was used as filler. The composition of metakaolin is primarily made of SiO₂ and Al₂O₃ with a small percentage of metal oxides [32]. When used in cement concrete applications, MK increases the compressive and flexural strength of the concrete, reduces its permeability, increases its resistance to chemical attack, enhances the workability and increases the durability of the concrete [40]. Geopolymers can benefit from the MK qualities, especially as it has a high Al₂O₃ content, being very reactive with the activators [41,42,43]. In the present work, two distinct commercial metakaolin powders were used.

Based on XRD analysis for both the metakaolins, MK2 was found to be more reactive if compared to MK1, denoting the presence of amorphous reactive phases [34] and, thus, suitable to be used. On the other hand, MK1, even though it has an amorphous phase, has a crystal formation in it, probably due to some impurity originated in the production process, more likely to be a less calcinated kaolin (quartz-like structure).

2.2. Activators

The mixture of sodium silicate (SS) and sodium hydroxide (SH) creates a very basic NaOH solution, which is called activating fluid that allows for the dissolution of the aluminosilicates. It increases the pH and it compensates for the electric charge of the aluminates in the mixture. A 10M SH was used in the present work.

2.3. Research Plan

The characterization of artificial alkali-activated material is generally based on the characterizations of the alkali-activated paste and the characterization of the physical and mechanical properties of the produced aggregates. Due to the lack of specific tests for artificial alkali-activated aggregates, it was necessary to use the same methodologies adopted for aggregates of natural origin.

The characterization for the MK1/basalt mixture (Aggregate B-MK1) was reported in a previous study by the authors [31]. In that study, the authors proposed a regression equation that had as input the variables in the mixture (e.g., MK/basalt ration, liquid to solid ratio) and as a result the predicted Unified Compression Resistance. This regression equation was applied for MK2/basalt mixture (Aggregate A-MK2). Results of compressive strength tests on cubic samples prepared with the alkali-activated paste, according to the EN10151-1 standard, are 35 MPa for the MK1 mix and 47 MPa for the MK2 one.

After the selection of the appropriate mixture design for both metakaolins, artificial-aggregates’ samples were prepared through casting into the molds and curing, as shown in Figure 2. The A-MK2 aggregate passed through a second curing process to evaluate if it was possible to improve some properties. In this case, the aggregates underwent a second curing process of 24 h at 80 °C. The micro-Deval test after the second curing process was only performed on A-MK2 aggregates.

The molds were designed and printed with a 3D SLA printer to cast aggregates with a specific shape and size, aiming to reduce noise and increase friction in surface asphalt courses.

The aggregates were produced as described in a previous study [31], using the same mix proportions, with the used metakaolin being the only variable. Aggregate A was made with MK2 and Aggregate B with MK1.

2.4. Artificial Aggregate Production

The artificial aggregate was produced using 3D-printed molds that guarantee a specific shape and size that, according to [31,44], could have an interesting behavior in friction and acoustic prospects. The mixtures were prepared as described in Figure 3.

The casting of the geopolymer mortar in the molds and the resulting aggregate are shown in Figure 4.

2.5. Artificial Aggregate Evaluation Methods

Both mixes were evaluated in terms of compressive strength in preliminary research [30] and subjected to X-Ray Diffraction (XRD) and Scanning Electron Microscopy (SEM). The produced aggregates were evaluated in terms of specific mass and water absorption following the EN 1097-6 Standard; Aggregate B-MK1 was submitted to EN 1097-2 for Los Angeles and EN1097-8 for Polished Stone Value (PSV), and both aggregates underwent EN-1097-1 for the micro-Deval.

Water absorption was determined according to the EN 1097-6 standard.

The Los Angeles (LA) test is an empirical measure of the resistance to fragmentation of mineral coarse aggregates used for pavement courses. A sample with a certain mass, depending on the particle size, is placed in a horizontal drum together with an appropriate number of steel balls. The drum is rotated for a total of 500 revolutions, during which the ball loads impact on the aggregates determining a fragmentation action, causing the formation of fine particles. The LA coefficient is obtained by the proportion of the fine particle mass created during the test and passing through a 1.6 mm sieve over the total mass of the sample [45].

The Polished Stone Value (PSV) test consists of evaluating the resistance of the aggregates against polishing action by a well-coded procedure. This test is carried out only for aggregates of specific dimension (7.2–10 mm), which are used only for asphalt wearing courses (EN 1097-8:2020). The polishing cycle has a duration of six hours, during which the polishing action is enhanced by using coarse abrasive sand and water for the first three hours and fine abrasive sand and water for the last hours. The specimens for PSV test consist of manually and randomly placing the aggregates into the mold, making sure that the flat surface of each aggregate is properly matched with the mold surface, obtaining a single layer (Figure 5a).

Interstices between particles are filled with fine sand (Figure 5b) and then a bicomponent fiberglass is poured into the mold, in order to obtain a rigid sample once the resin hardens so that it preserves the mold curvature (Figure 5c).

For the case study, only the Aggregate B-MK1 (ID: B) was used for this test and as prescribed by the standard, it was compared with the basalt aggregate (ID: C). It is worth noting that, given the lack of a control stone recognized by the standard, a suitable alternative control stone of established PSV value was used [46]. The PSV was calculated according to the following standard equation.

$$PSV=\mathrm{S}+\mathrm{X}-\mathrm{C}$$

(1)

where:

• S is the mean value for the aggregate test specimens (Aggregate B-MK1);
• X is the mean PSV for the source of the control stone (in this case, 49);
• C is the mean value for the control stone specimens (basalt).

Furthermore, a roughness analysis was conducted on PSV specimens by means of a laser profilometer based on conoscopic holography methodology ISO 13473-3:2002. For each sample, five profiles were analyzed, two longitudinal profiles in the direction of the PSV wheel (a and b) and three in the crosswise direction (c, d and e), identified through the use of a mask (Figure 6). The alignments marked on the mask and a reference system defined for each sample help to carry out the profilometric analysis of aggregate surface texture before and after the PSV test. This ensures that the same profile is detected so that it can be completely superimposable with a reasonable approximation. Profiles were captured by means of a lens with an objective focal length of 50 mm and a sampling resolution of 5 µm.

The estimated micro-texture indexes (see Table 2) are defined as follows:

• Ra: arithmetical mean deviation roughness of the profile, which represents an average of the profile deviations from a center line;
• Rq: root mean squared roughness;
• Rz: average peak-to-valley height, which is based on the five highest peaks and the lowest valleys over the entire length of the evaluation segment [47]

Table 2. Formulas and graphical explanation of the investigated roughness indicators.

Roughness Indicator	Ra	Rq	Rz
Formula	$R_a=\frac{1}{n}·{\displaystyle {\displaystyle \sum }_{i=1}^n}\left|y_i\right|$	$R_q=\sqrt{\frac{1}{n}·{\displaystyle {\displaystyle \sum }_{i=1}^n}{y}_i^2}$	$R_z=\frac{1}{5}·{\displaystyle {\displaystyle \sum }_{i=1}^5}{R}_{pi}-R_{vi}$
Graphical Explanation

Table 2. Formulas and graphical explanation of the investigated roughness indicators.

Roughness Indicator	Ra	Rq	Rz
Formula	$R_a=\frac{1}{n}·{\displaystyle {\displaystyle \sum }_{i=1}^n}\left|y_i\right|$	$R_q=\sqrt{\frac{1}{n}·{\displaystyle {\displaystyle \sum }_{i=1}^n}{y}_i^2}$	$R_z=\frac{1}{5}·{\displaystyle {\displaystyle \sum }_{i=1}^5}{R}_{pi}-R_{vi}$
Graphical Explanation

The last test carried out was the micro-Deval (EN 1097-1), which evaluates the resistance to wear of coarse aggregates. The test provides a measure of aggregate abrasion resistance by means of an abrasive action that occurs between the aggregates themselves and the aggregates and small steel balls in the presence of water. A 500 g sample is placed together with 2.5 liters of water and an abrasive charge, consisting of a set number of steel balls, inside a specific apparatus that is rotated for 12,000 revolutions. After the test, the sample is removed from the drum, washed and the oversize fraction retained on a 1.6 mm sieve is dried. The micro-Deval coefficient (MDE) is calculated by the difference between the initial mass and the oversized fraction mass compared to the total mass of the sample.

3. Results

3.1. Alkali-Activated Materials’ Chemical Characterization

The characterization of the alkali-activated materials was conducted using XRD and SEM, for both mixes. The analysis of SEM images allows for the identification of the formation of geopolymer gel. Furthermore, with the XRD, it is possible to identify the geopolymerization region and some of the crystals that did not react.

From the XRD image in Figure 7, it is possible to notice that a reaction occurred due to the alkaline activation of the precursors, given the hump between 20 and 30 2θ recorded on the spectra, compared to that of the constituent materials. This suggests the formation of geopolymeric structures, as observable in the SEM images. The crystalline phases detected in both samples derive from unreacted basalt and metakaolin. As expected, no new crystalline phases were formed.

Figure 8 shows, at 10,000 magnification, the internal part of Aggregate A and Aggregate B, with these mixes made with MK2 and MK1 metakaolin, respectively.

In both materials, it is possible to see the formation of a geopolymer gel (GP) and the presence of crystals of different mineral phases from unreacted precursors (C). The most notable difference is that in the MK2 mix, the geopolymer gel is more homogeneous and compact, and the unreacted crystalline phases from basalt are well incorporated into the geopolymer matrix. In the other mix, the geopolymer formation resulted in a less cohesive material, with typical nanoprecipitates formed by alkali activation, unevenly disposed around the unreactive phases of both precursors. The scarce reactivity of MK1, deriving from the presence of fewer readily reactive amorphous phases and its coarser particle size, resulted in a highly porous and less compacted material, likely having lower physical and mechanical properties.

3.2. Artificial Aggregate Characterization

Aggregate A and Aggregate B were characterized in terms of water content, specific mass, Los Angeles abrasion, Polished Stone Value and micro-Deval abrasion.

Table 3. Materials’ characteristics and comparison with other aggregates.

Aggregate	Water Content (%)	Specific Mass (g/cm3)	Los Angeles (%)	PSV	Micro-Deval (%)	Reference
Aggregate A	19.1	2.120	-	-	35 * (27)	-
Aggregate B	21.0	2.017	37	59	70	-
Basalt	<2.5	2.700	14–20	53	14	[48,49,50]
Calcined Bauxite	6.8	2.629	10–17	50–70	5	[48,51,52]
Fly ash AA	5.5	2.140	27	-	-	[53]
Steel slag	1.1–9.0	2.96–3.59	14–15	25–55	6–10	[47,54]

* Results after treating the material in the oven for 24 h at 80 °C, after previous curing.

shows the results for both artificial aggregate types and their direct comparison with a natural aggregate (basalt), two artificial aggregates that follow a different production method, (i.e., calcined Bauxite and steel slag) and an AA produced through the alkali activation of fly ashes.

It is important to note that both artificial aggregates produced in this research have a higher water content than the aggregates used as a reference. It is important to note that natural aggregates, such as basalt, typically have less water absorption. Regarding the calcined bauxite and steel slag, which are not geopolymer materials, the internal structure is different and, thus, can contribute to different water absorption. As for the fly ash AA, even though it is a geopolymer artificial aggregate, the production process is quite different. The different production process and use of fly ash can be contributors to the difference in the water absorption. The specific mass is comparable to the fly ash AA one.

The artificial aggregate B had high results in terms of Los Angeles and micro-Deval abrasion tests; however, promising results were obtained from PSV. Aggregate A was not tested in terms of Los Angeles and PSV. However, its micro-Deval was considerably better than the one for aggregate B, especially if the material was heat-treated in the oven after the initial curing.

3.3. Artificial Aggregate Micro-Texture Analysis

As for the micro-texture characterization, the analysis was carried out for the Aggregate B-MK1 and the control basalt. Data were obtained by analyzing micro-profiles of an average length of 0.7–1.05 mm, graphically identified on each single grain that composes a profile. For Aggregate B-MK1, a total of 119 micro-profiles with an average length of 909 µm was analyzed, whereas for the control basalt, 120 micro-profiles with an average length of 834 µm were recorded.

Table 4. Post-processing analysis of roughness data collected by laser profilometer: mean values before and after polishing and variation (%).

ID Sample:	Data	Before Polishing (µm)			After Polishing (µm)			Variation (%)
		Ra,bp	Rq,bp	Rz,bp	Ra,ap	Rq,ap	Rz,ap	ΔRa	ΔRq	ΔRz
Aggregate B-MK1 (B)	Longitudinal sections (mean)	12.6	16.8	50.9	19.6	24.4	69.1	+88.9	+78.0	+60.1
	Crosswise sections (mean)	9.5	12.6	40.9	25.4	30.9	79.1	+173.4	+151.6	+94.5
	Mean	10.7	14.3	45.0	23.1	28.3	75.1	+127.7	+111.1	+75.4
Control Basalt (C)	Longitudinal sections (mean)	24.8	30.6	75.7	26.3	32.5	84.7	+6.1	+6.5	+14.3
	Crosswise sections (mean)	22.3	27.5	67.9	27.0	32.9	80.0	+21.2	+20.0	+17.9
	Mean	23.3	28.7	71.0	26.7	32.8	81.9	+14.7	+14.0	+15.8

shows the results of the micro-texture analysis, where the collected data for each aggregate were averaged for longitudinal and crosswise profiles, for both polished and unpolished samples. The percentage variation in roughness data Ri (with i = a, q, z) registered before (Ri,bp) and after (Ri,ap) polishing is given by:

$$\Delta R_i=\frac{R_{i,ap}-R_{i,bp}}{R_{i,bp}} ·100$$

(2)

As it can be clearly seen from Table 4, Aggregate B-MK1 shows an important increase in the micro-texture indexes, almost doubling in value, whereas the basalt exhibits a lower variation, 15% on average, which can be considered not significant. This is in line with other studies carried out by means of a similar methodology, which recorded low variation in the basalt aggregates in terms of roughness [47,48].

In addition, profile tracking with the laser profilometer provided the basis to conduct further analysis on the samples in order to obtain more information about the wearing characteristics of the materials. Thus, the percentage of loss material was estimated starting with the calculation of the area underneath the profiles (Aap: area after polishing; Abp: area before polishing) by means of the trapezoid method, with a step of 5 µm, which is given by the following formula:

$$\Delta A=\frac{A_{ap}-A_{bp}}{A_{bp}} ·100$$

(3)

Table 5. Percentage estimation of the material loss occurring after the polishing action.

ID Sample: Aggregate B-MK1 (B)	ΔA Variation (%)	St. Dev (%)
Longitudinal sections (mean)	−16.9	±4.5
Crosswise sections (mean)	−17.4	±2.1
ID Sample: Control Basalt (C)	ΔA Variation (%)	St. Dev (%)
Longitudinal sections (mean)	−3.7	±1.4
Crosswise sections (mean)	−4.1	±1.0

summarizes the percentage of loss material estimated for the longitudinal and the crosswise sections after submitting the samples to the PSV abrasion. Figure 9 and Figure 10 show an example of the profile’s superposition for both the Aggregate B-MK1 (A6 sample) and the Control Basalt (C2 sample), respectively. It can be clearly seen that the artificial aggregates have undergone a significant loss in material due to the polishing, whereas the basalt profiles appear to be almost unaltered.

Skid resistance results obtained by means of the British Pendulum Test are shown in

Table 6. Results of the PSV test: Pendulum Test Value (PTV) before and after polishing, percentage variation and PSV values.

ID Sample	PTVb	PTVa	Variation (%)	PSV
Aggregate B-MK1 (B)	62	56	−9.9	59.0
Control Basalt (C)	56	46	−18.3	49.0

, where PTV_b is the British Pendulum value measured on unpolished samples, whereas PTV_a is the measurement after the polishing action. Aggregate B-MK1 kept high values of PTV, showing a variation of about −10% on average after the polishing, whereas the control basalt samples showed almost a doubled percentage variation (−18.3%).

Based on the results, it is possible to state that the basalt kept a similar value between pre and post-polishing (Table 3 and 4) on average and did not suffer the effects due to the polishing also in terms of loss of material. Instead, the Aggregate B-MK1 showed a significant increase in all the micro-texture indexes after the polishing. This might also be due to the initially smooth surface finishing, with the aggregates cast and cured in a mold, whereas the polishing action let the crystals (Figure 9) rise to the surface. Despite maintaining a high value of PTV after conditioning, the artificial aggregates exhibited a great loss in terms of material, which corresponded to a remodeling of the macro-texture of the samples, obtaining a quite different contact area between the surface of the sample and the pad of the British Pendulum.

4. Conclusions

This research aimed at the preliminary characterization of engineered artificial aggregates produced through the alkali activation of basalt and metakaolin powders.

The main goal was to conduct a preliminary evaluation of the artificially engineered aggregate. The AEA was evaluated in terms of resistance to abrasion, fragmentation, water absorption and polishing. The main conclusions are as follows:

• It was noted that the selected metakaolin played a very important role, as the MK2 one had better geopolymerization results.
• The basalt powder proved to be a suitable material.
• The AEA with MK2 had a higher resistance to wear than the one with MK1.
• The post-curing heat treatment improves the material characteristics.
• The PSV and BPN results are quite promising, showing that it might be used in replacement of natural aggregate or other artificial aggregates.
• The process must be improved in order to increase the overall quality of the AEA, especially in terms of fragmentation results.

As for the basalt action as a constituent, it is clear that it did not fully react. Part of it contributed to the geopolymerization and partly acted as a non-reactive material (filler), as evidenced by the crystals in Figure 9. However, more research is necessary to better understand its contribution to the geopolymerization process in a multi-material system along with metakaolin.

Regarding the physical characterization, the aggregate B-MK1 showed lower results in most tests compared to the reference materials (i.e., basalt, calcined Bauxite, fly ash AA, steel slags), except for the PSV. However, despite the fact that the artificial aggregate does not perform well in those tests, it revealed high PTV values and optimal micro-texture indexes.

By changing the metakaolin in aggregate A-MK2, the resistance to wear increased, halving the loss of material due to the abrasive action. It appears that the results improved even further in terms of micro-Deval coefficient when the artificial aggregates were heat-treated.

Therefore, although the mixture requires improvements in order to achieve a better behavior in terms of resistance to wear and fragmentation, the high PTV and PSV values, as well as the quality of micro-texture indexes, are proof of potential good performance, making the artificial aggregates suitable as a replacement of natural aggregates for wearing courses in road pavements.

All in all, the research shows that producing artificial geopolymer aggregates using 3D-printed molds to control their shape and size is a promising approach that could be scaled to larger industrial productions. It is necessary to evaluate the possibility to 3D print the aggregates and not only the molds, which could increase the production process, especially in an industrial context.

Future developments will foresee a complete characterization of aggregate A that will be used for the production of microsurfacing mixes to evaluate its behavior in friction and acoustic terms.