Assessing the Suitability of Phosphate Waste Rock as a Construction Aggregate

Abstract

Phosphate waste rock (PWR) is gaining attention as a potential alternative aggregate for concrete. Its valorization could reduce the environmental impacts of quarrying natural resources and stockpiling phosphate mining waste. This study comprehensively investigated the properties of fine and coarse aggregates produced from three rock types selected from PWR in Morocco: Flint, Phosflint, and Dolomite. A range of techniques was used to study their characteristics, including microstructural observations up to the microscale and X-ray computed tomography (X-CT), mineralogical and chemical compositions, physical and geotechnical properties such as Los Angeles (LA), micro-Deval (M_DE), flexural strength, real dry density, and total porosity. The results showed that the coarse fractions of Flint, Phosflint, and Dolomite are code A or B of NF P 18-545 and exhibit good shape, density, and water absorption properties. Flint aggregates had the highest wear and fragmentation resistance with the lowest and finest porosity. They contained mainly quartz but also small proportions of Dolomite and fluorapatite. Phosflint aggregates had high resistance, shown by code A in LA and M_DE values, and flexural strength equal to 17.1 MPa. They contained phosphate microfacies with a Ca/P atomic ratio equal to 1.8, cemented by cryptocrystalline silica. Dolomite aggregates’ mineralogical make-up consisted mainly of dolomite with the presence of quartz particles in addition to impurities. They also displayed significant total porosity (10–12%), as confirmed by X-CT. These findings were discussed to develop insights for the use of three types of PWR as alternative aggregates for concrete production. This investigation contributes to unveiling the properties of PWR as concrete aggregates and encourages circularity between the mining and construction sectors.

1. Introduction

The construction industry is the largest consumer of aggregates, such as crushed stone, sand, and gravel. They are the main constituents of construction materials because they occupy 60–80% of the volume of Portland cement concrete [1], 60–75% of the volume of mortar [2], and between 75 and 85% of the volume of asphalt mixtures [3]. Aggregates are a vital resource, and they are the most extracted raw material in the world, with a quantity equal to 40–50 Gt [4,5]. Quarrying aggregates negatively impacts the ecosystem quality by increasing land occupation and contributing to energy consumption and greenhouse gas emissions from crushing and transportation operations [6]. The enormous quantity of aggregates quarried is expected to increase, aggravating the scarcity of natural resources. This underscores the necessity to explore alternative resources to natural aggregates [7].

Many by-products and waste materials have been characterized for use as aggregates, such as recycled concrete aggregates (RCAs) [8,9] and slag aggregates [10]. Evangelista et al. [8] studied the detailed properties of different fractions of fine RCAs, unveiling specific features encouraging the reuse of aggregates. The physical and strength properties of 13 types of alternative coarse aggregates were found to affect the rheological, mechanical, and durability properties of concrete [11]. Understanding the properties of alternative aggregates could help determine an optimal substitution ratio and improve the performance of concrete.

Researchers have reviewed current practices of phosphate waste management, the types and characteristics of waste streams, and their potential for reuse [12]. The current practice of phosphate mine waste rock (PWR) management consists of stockpiling a significant quantity of waste rock, creating a large land footprint and other environmental impacts. This management is driven today towards a more environmentally friendly practice in the framework of a circular economy.

The transition towards a circular economy in mining requires us to move away from the linear “take-make-use-dispose” model towards circular “reduce-reuse-recycle” thinking. One important step is to recognize the potential of PWR as a valuable and readily available resource for construction materials. However, economic, social, regulatory, and technical challenges are slowing this shift, including the characterization of these wastes and the assessment of their suitability for use as construction materials [13].

PWR was used as a coarse aggregate to produce ordinary concrete with 28 d compressive strength equal to 24 MPa in Egypt [14] and with 28 d compressive and 28 d splitting strength equal to 29 MPa and 5.4 MPa, respectively, in Jordan [15], equal to 29 MPa and 2.6 MPa, respectively in Morocco [16]. PWR of the Ben Guerir mine in Morocco has also been used for road embankment construction. Their application was confirmed based on the geotechnical properties using standards and based on slope stability analysis, which showed that a safe embankment height is between 5 m and 10 m. An economic evaluation of the transportation costs showed that these wastes can be an economically viable solution up to a profitability radius equal to 28 km [17].

The nearest viable coarse and fine aggregate quarry to Ben Guerir city is Oued Tensift, located in Marrakech, ~90 km away. This quarry consists of washed river gravel of mixed composition. On the other hand, the phosphate mine of Ben Guerir, located just 10–20 km away, offers several rock types as an alternative resource. The use of PWR wastes could provide a nearby source of aggregate and reduce the transportation distance and the associated energy consumption. Furthermore, the volume of PWR is equal to 15 Mm³ and contains ~12% of crushed hard rocks which is an enormous quantity [18,19].

Despite these promising results, research on the use of PWR as a construction aggregate is still in its early stages. The properties of the crushed hard rocks and their suitability for use in concrete requires further investigation. To the best the author’s knowledge, a comprehensive investigation into the properties of PWR for use as a concrete aggregate has not been undertaken.

This study aims to investigate the microstructural, mineralogical, and geotechnical properties of alternative aggregates from PWR. The by-products were processed into coarse and fine aggregates and comprehensive characterizations were conducted on different size fractions. Their properties were evaluated to achieve better reuse as concrete aggregates.

2. Materials and Methods

2.1. Materials

Three rock types—Flint (F), Phosflint (PF), and Dolomite (D)—were initially obtained by the authors by sorting and stocking a 25-ton sample of PWR [19]. These rock types compose the intercalation layers and waste rock of the phosphate mine of Ben Guerir in Morocco. Figure 1 presents the three rock types and the process of their recovery.

About 5 tons of the selected PWR was used to produce aggregate using a crushing and sizing system. The production system consisted of a single-stage jaw crusher with a feed rate equal to 200 kg per hour, a maximal feed size equal to 150 mm, and a three-level screen. A parametric study was executed on the production system, and the calibration of the jaw crusher shows that the 5–20 mm opening of the jaw crusher was the optimum to obtain the best combination of size and yield. Figure 2 presents the preparation techniques of the aggregates, including the different stages, products, and their yield. The produced aggregates consist of 4 fractions: Sand (S), Gravel 1 (G₁), Gravel 2 (G₂), and Gravel 3 (G₃), as shown in Figure 3.

2.2. Methods

The total quantity of the 12 aggregates was reduced using the recommendations of the NF EN 932-2 standard [20]. Geotechnical properties were executed as per the standards in NF P 18-545 [21]. A portion of the representative samples was used for mineralogical and chemical characterizations. This portion of the samples was subjected to two stages of grinding and pulverization to obtain powders finer than 50 µm. Five to eight particles of the three aggregate types were selected for microstructural observations. Figure 4 presents the characterization techniques. Supplementary data (Figure S1) present the preparation flowsheet of representative samples for testing.

2.2.1. Microscopic Observations

Four types of microscopes were used to determine the petrography and the microstructure, including the surface morphology, texture, composition, and arrangement of grains and cement: (i) Binocular microscopy: a Leica Binocular Stereoscopic microscope was used where samples were observed as fragments. (ii) Optical microscopy: a Leica DM2700 standard optical microscope was used under cross-polarizing light. Samples were prepared in thin sections and Flint as a polished section; (iii) scanning electron microscopy: Quanta (FEI^®) SEM operated at 20 kV accelerating voltage with backscatter electron (BSE). Samples were prepared as polished sections; (iv) transmission electron microscopy (TEM): investigations were performed using a JEM–ARM 200F Cold FEG. It operated at 200 kV and was equipped with a spherical aberration (Cs) probe and image. Energy dispersive X-ray spectroscopy (EDS) analyses were executed in the scanning transmission electron microscopy (STEM) mode using the Centurio Silicon Drift Detector (SDD) with a probe diameter equal to 0.5 nm. The quantification of the composition was performed using software from the Jeol and Cliff–Lorimer method. Samples were prepared as powders dispersed in ethanol.

2.2.2. Chemical and Mineralogical Compositions

The composition of oxides was measured with wavelength dispersive X-ray fluorescence (WD-XRF) using RIGAKU ZSX Primus IV spectrometer. Crystalline phases were determined using a D8 Advance X-ray Diffractometer (XRD) from Bruker coupled with the LynxEye detector and operated using CuKa (1.54 Å) radiation, with a scanning range [2θ] between 10° and 60° while counting 92 s per [2θ] at a step size equal to 0.02. The patterns of minerals were obtained from the International Centre for Diffraction Data (ICDD) PDF 4-2023. Peak identification was conducted using DiffracEVA V6 software [22].

2.2.3. Geotechnical Properties

The quality of aggregates is determined by partitioners and researchers alike by measuring geotechnical properties that were executed per NF EN 12620 standards [23]. The tests include the particle size distribution (PSD), the shape by the flakiness index (A), the real dry density (ρ_rd), and the coefficient of water absorption after 24 h of imbibition (WA₂₄). The mechanical resistance to fragmentation and wear was obtained using the Los Angeles (LA) and micro-Deval (M_DE) tests, respectively. The intergranular porosity (v) was measured by pouring samples into a 100Ø200 mm cylinder. Absolute gravity (ρ_a) was measured with a Micrometrics Helium Pycnometer AccuPyc II 1340, and the average of three tests was retained. The flexural strength of Phosflint and Dolomite was tested on three prisms of 40 × 40 × 160 mm³. The samples were prepared from an intact rock specimen by cutting with a diamond saw. Flint prisms could not be prepared because of their important hardness. The quality of fines in sands is indicated by testing the sand equivalent (SE) and methylene blue value (MB) on the 0/2 mm fraction. Table 1 presents the tests to determine the geotechnical properties.

2.2.4. X-ray Computed Tomography

X-ray computed tomography (X-CT) scans and image analysis were executed to determine the internal structure of rocks. X-CT is a technology used for the non-destructive testing of the structure of materials by obtaining high-quality images. The greyscale images represent X-ray penetration intensity in the material, and darker regions represent lower density, such as pores, while lighter regions represent materials with higher density [24]. Three micro cylinders (Ø~2.5 mm) were obtained from drilling into Flint, Phosflint, and Dolomite rocks and were mounted on a sample holder with wax and exposed to two testing scans. The first test was performed on Flint and Dolomite, and the second on Phosflint and Dolomite, both under a current of 40 µA and an X-ray peak energy of 60 kV. The voxel size was equal to 4.7 µm³. The tests resulted in two sets of projection data. After acquisition and reconstruction, one 2D slice image of each fraction of Flint, Phosflint, and Dolomite was selected to represent the internal core of the micro cylinder. The image analysis was carried out using Fiji (ImageJ) open source software version 1.54h [25]. First, the scale was set, and the images were cropped to a Region of Interest of 1200 µm × 1200 µm inside the sample to avoid the effect of borders; the images were then converted to 16-bit. The preprocessing was performed by adjusting the brightness and contrast to one default value; after that, the thresholding segmentation was carried out on the Region of Interest (ROI) using a Default algorithm. Threshold limits were set, and particles were analyzed for the surface area.

Table 1. Tests to determine the geotechnical properties.

Properties	Test	Method	Equations		Testing measures
Geometric properties	Flakiness coefficient	NF EN 933-3 [26]	$\mathrm{A}={\displaystyle \frac{{\mathrm{M}}_2}{{\mathrm{M}}_1}}\times 100$	(1)	A: global coefficient of flakiness (%). M2: sum of weight of aggregates passing through grids of Di/2 (g). M1: sum of weight of fraction of aggregates di/Di (g).
	Intergranular porosity	NF EN 1097-3 [27]	${\mathsf{\rho }}_{\mathrm{b}}={\displaystyle \frac{{\mathrm{M}}_2-{\mathrm{M}}_1}{\mathrm{V}}}$	(2)	${\mathsf{\rho }}_{\mathrm{b}}$: bulk density (t/m3). M1: weight of empty container (kg); M2: weight of container and sample (kg). V: volume of container (liter). v: intergranular porosity. ${\mathsf{\rho }}_{\mathrm{b}}$: bulk density (Mg/m3). ${\mathsf{\rho }}_{\mathrm{p}}$: real density (Mg/m3) according to NF EN 1097-6.
			$\mathrm{v}=1-{\displaystyle \frac{{\mathsf{\rho }}_{\mathrm{p}}-{\mathsf{\rho }}_{\mathrm{b}}}{{\mathsf{\rho }}_{\mathrm{p}}}}$	(3)
Physical properties	Real dry density	NF EN 1097-6 [28]	${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}={\displaystyle \frac{{\mathrm{M}}_4}{({\mathrm{M}}_1-({\mathrm{M}}_2-{\mathrm{M}}_3)}}$	(4)	${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}$: real dry density (t/m3). M1: weight of saturated surface dry aggregates in air (g). M2: weight of pycnometer with sample of saturated aggregates (g). M3: weight of pycnometer full of water only (g). M4: weight of sample of oven-dried aggregates in air (g).
	Coefficient of water absorption 1	NF EN 1097-6 [28]	${\mathrm{W}\mathrm{A}}_{24}=100\times {\displaystyle \frac{{\mathrm{M}}_1-{\mathrm{M}}_4}{{\mathrm{M}}_4}}$	(5)	${\mathrm{W}\mathrm{A}}_{24}$: coefficient of water absorption after imbibtion for 24 h (%). M1: weight of saturated surface dry aggregates in air (g). M4: weight of sample of aggregates dried in oven in air (g).
	Absolute density	Helium pycnometer	${\mathsf{\rho }}_{\mathrm{a}}={\displaystyle \frac{\mathrm{m}}{\mathrm{V}}}$	(6)	${\mathsf{\rho }}_{\mathrm{a}}$: absolute density (t/m3). m: weight of sample (g); V: volume of sample (cm3).
	Total porosity		$\mathrm{p}=1-{\displaystyle \frac{{\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}}{{\mathsf{\rho }}_{\mathrm{a}}}}$	(7)	p: total porosity.
Mechanical properties	Los Angeles resistance to fragmentation	NF EN 1097-2 [29]	$\mathrm{L}\mathrm{A}={\displaystyle \frac{5000-\mathrm{m}}{50}}$	(8)	LA: Los Angeles coefficient executed on G1 fraction (%). m: weight of sample retained by the 1.6 mm sieve after testing (g).
	Micro-Deval resistance to wear	NF EN 1097-1 [30]	${\mathrm{M}}_{\mathrm{D}\mathrm{E}}\left(\mathrm{\%}\right)={\displaystyle \frac{500-\mathrm{m}}{5}}$	(9)	MDE: micro-Deval coefficient with water executed on G1 fraction (%). m: weight of sample retained by the 1.6 mm sieve after testing (g).
	Flexural strength	Prismatic rock specimen	${\mathsf{\sigma }}_{\mathrm{f}}={\displaystyle \frac{3\mathrm{F}}{2\mathrm{b}{\mathrm{h}}^2}}$	(10)	${\mathsf{\sigma }}_{\mathrm{f}}$: flexural strength (MPa). F: maximum force applied during the bending test (N). b: width of the specimen (m); h: height of the specimen (m).
Sand cleanliness	Sand equivalent	NF EN 933-8 [31]	${\mathrm{S}\mathrm{E}}_{10}\left(\mathrm{\%}\right)={\displaystyle \frac{{\mathrm{h}}_2}{{\mathrm{h}}_1}}\times 100$	(11)	h2: total height of suspension (mm). h1: height of sand sample (mm).
	Methylene blue test	NF EN 933-9 [32]	$\mathrm{M}\mathrm{B}={\displaystyle \frac{{\mathrm{V}}_1}{{\mathrm{M}}_1}}\times 10$	(12)	MB: value of methylene blue (g/kg). V1: total volume of injected solution (mL). M1: weight of test sample (g).

¹ The sand specific gravity and water absorption were made with visual inspection instead of the cone test to determine the saturated surface dry test.

Table 1. Tests to determine the geotechnical properties.

Properties	Test	Method	Equations		Testing measures
Geometric properties	Flakiness coefficient	NF EN 933-3 [26]	$\mathrm{A}={\displaystyle \frac{{\mathrm{M}}_2}{{\mathrm{M}}_1}}\times 100$	(1)	A: global coefficient of flakiness (%). M2: sum of weight of aggregates passing through grids of Di/2 (g). M1: sum of weight of fraction of aggregates di/Di (g).
	Intergranular porosity	NF EN 1097-3 [27]	${\mathsf{\rho }}_{\mathrm{b}}={\displaystyle \frac{{\mathrm{M}}_2-{\mathrm{M}}_1}{\mathrm{V}}}$	(2)	${\mathsf{\rho }}_{\mathrm{b}}$: bulk density (t/m3). M1: weight of empty container (kg); M2: weight of container and sample (kg). V: volume of container (liter). v: intergranular porosity. ${\mathsf{\rho }}_{\mathrm{b}}$: bulk density (Mg/m3). ${\mathsf{\rho }}_{\mathrm{p}}$: real density (Mg/m3) according to NF EN 1097-6.
			$\mathrm{v}=1-{\displaystyle \frac{{\mathsf{\rho }}_{\mathrm{p}}-{\mathsf{\rho }}_{\mathrm{b}}}{{\mathsf{\rho }}_{\mathrm{p}}}}$	(3)
Physical properties	Real dry density	NF EN 1097-6 [28]	${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}={\displaystyle \frac{{\mathrm{M}}_4}{({\mathrm{M}}_1-({\mathrm{M}}_2-{\mathrm{M}}_3)}}$	(4)	${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}$: real dry density (t/m3). M1: weight of saturated surface dry aggregates in air (g). M2: weight of pycnometer with sample of saturated aggregates (g). M3: weight of pycnometer full of water only (g). M4: weight of sample of oven-dried aggregates in air (g).
	Coefficient of water absorption 1	NF EN 1097-6 [28]	${\mathrm{W}\mathrm{A}}_{24}=100\times {\displaystyle \frac{{\mathrm{M}}_1-{\mathrm{M}}_4}{{\mathrm{M}}_4}}$	(5)	${\mathrm{W}\mathrm{A}}_{24}$: coefficient of water absorption after imbibtion for 24 h (%). M1: weight of saturated surface dry aggregates in air (g). M4: weight of sample of aggregates dried in oven in air (g).
	Absolute density	Helium pycnometer	${\mathsf{\rho }}_{\mathrm{a}}={\displaystyle \frac{\mathrm{m}}{\mathrm{V}}}$	(6)	${\mathsf{\rho }}_{\mathrm{a}}$: absolute density (t/m3). m: weight of sample (g); V: volume of sample (cm3).
	Total porosity		$\mathrm{p}=1-{\displaystyle \frac{{\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}}{{\mathsf{\rho }}_{\mathrm{a}}}}$	(7)	p: total porosity.
Mechanical properties	Los Angeles resistance to fragmentation	NF EN 1097-2 [29]	$\mathrm{L}\mathrm{A}={\displaystyle \frac{5000-\mathrm{m}}{50}}$	(8)	LA: Los Angeles coefficient executed on G1 fraction (%). m: weight of sample retained by the 1.6 mm sieve after testing (g).
	Micro-Deval resistance to wear	NF EN 1097-1 [30]	${\mathrm{M}}_{\mathrm{D}\mathrm{E}}\left(\mathrm{\%}\right)={\displaystyle \frac{500-\mathrm{m}}{5}}$	(9)	MDE: micro-Deval coefficient with water executed on G1 fraction (%). m: weight of sample retained by the 1.6 mm sieve after testing (g).
	Flexural strength	Prismatic rock specimen	${\mathsf{\sigma }}_{\mathrm{f}}={\displaystyle \frac{3\mathrm{F}}{2\mathrm{b}{\mathrm{h}}^2}}$	(10)	${\mathsf{\sigma }}_{\mathrm{f}}$: flexural strength (MPa). F: maximum force applied during the bending test (N). b: width of the specimen (m); h: height of the specimen (m).
Sand cleanliness	Sand equivalent	NF EN 933-8 [31]	${\mathrm{S}\mathrm{E}}_{10}\left(\mathrm{\%}\right)={\displaystyle \frac{{\mathrm{h}}_2}{{\mathrm{h}}_1}}\times 100$	(11)	h2: total height of suspension (mm). h1: height of sand sample (mm).
	Methylene blue test	NF EN 933-9 [32]	$\mathrm{M}\mathrm{B}={\displaystyle \frac{{\mathrm{V}}_1}{{\mathrm{M}}_1}}\times 10$	(12)	MB: value of methylene blue (g/kg). V1: total volume of injected solution (mL). M1: weight of test sample (g).

¹ The sand specific gravity and water absorption were made with visual inspection instead of the cone test to determine the saturated surface dry test.

3. Results

3.1. Microstructural Observations

Figure 5 presents the micrographs that showcase the texture, composition, arrangement, and microstructure. They are described per aggregate type by stereo, optical, scanning electron, and transmission electron microscopy.

Flint particles observed under stereo microscopy appear as thin, brown, opaque chips with a smooth, glass-like texture and matte reflectivity. The particles were flaky, elongated, and highly angular. They had a conchoidal shape and lacked cleavage. Under optical microscopy, silica cement was micro- to crypto-crystalline. Boujo [33] examined the petrography of Flint from the same provenance using polarizing microscopy and identified two forms of silica, opal, and calcedony, primarily as microcrystalline or fibrous structures, with microcrystalline quartz appearing rare. Similarly, El Haddi [34] investigated the petrography of siliceous facies in other phosphate deposits in Morocco. Their findings indicated that Flint comprised microcrystalline quartz, calcedony, and opal-CT. Under SEM, the microstructure is fine and microcrystalline and displays fine pores in the form of veins. Quartz particles have angular edges under TEM.

Phosflint fresh fragments resembling “nougat” contain light brown particles with silica cement of dark brown color. The aspect is angular and rough because phosphate grains create irregular surfaces. Optical microscopy reveals phosphate microfacies, such as apatite grains, bioclasts, and coprolite, which are cemented with cryptocrystalline silica. Boujo [33] identified that the siliceous cement of Phosflint was in the form of calcedony and was associated with microcrystalline quartz or fibrous structures. Under SEM, large pores were identified inside phosphate particles. Under TEM, fluorapatite minerals appear as very small particles that are grouped into larger agglomerates.

Dolomite particles are light brown to greyish, angular, and have a rough surface texture. Under optical microscopy, the thin section presents a homogeneous and fine micritic carbonate cement in addition to the presence of dispersed quartz grains that appear darker under cross-polarizing light. Under SEM, Dolomite presents large pores that are uniformly distributed in the section and occupy an important surface area. The Dolomite mineral presents soft, rounded edges under TEM.

3.2. Mineralogical and Chemical Compositions

The chemical composition and the diffractograms are presented in

Table 2. Chemical composition in oxides according to XRF analysis (wt.% of oxides).

Aggregates	SiO2	CaO	MgO	P2O5	Na2O	K2O	Al2O3	Fe2O3	LOI
F-S	63.2	18.5	1.7	10.1	0.5	0.1	0.5	0.5	4.9
F-G1	79.7	8.0	2.0	4.1	0.3	0.1	0.3	0.8	4.6
F-G2	82.6	6.6	1.6	3.3	0.3	0.1	0.3	1.2	4.1
F-G3	83.1	6.3	1.7	3.3	0.3	0.1	0.2	1.2	3.8
PF-S	44.7	29.6	1.4	16.7	0.6	0.1	0.3	0.4	6.3
PF-G1	49.4	26.4	1.3	14.2	0.5	0.1	0.3	0.4	7.4
PF-G2	49.4	26.2	1.7	14.1	0.5	0.1	0.2	0.3	7.5
PF-G3	51.9	24.5	2.1	12.8	0.5	0.1	0.2	0.6	7.2
D-S	7.3	33.1	18.5	3.7	0.3	0.1	0.4	0.2	36.4
D-G1	6.7	33.0	18.4	2.1	0.2	0.1	0.4	0.2	38.9
D-G2	7.3	33.1	18.3	3.5	0.3	0.1	0.4	0.2	36.8
D-G3	6.9	32.3	18.2	2.2	0.3	0.1	0.3	0.2	39.5

and Figure 6, respectively. The EDS spectra of one particle using STEM are presented in Figure 7.

According to XRF analysis, Flint aggregates are mainly composed of silicon oxide (63–83%), which corroborates with an XRD composition that shows quartz to be the main crystalline phase. STEM analysis shows that Flint contains essentially siliceous crystalline particles composed of quartz. The Flint sand fraction contains a higher amount of CaO (18%) and P₂O₅ (10%). This is due to the fine fraction containing impurities from fragile rocks, such as phosphate, which is a characteristic of the heterogeneous nature of PWR.

Phosflint is composed of SiO₂ (45–52%), a great amount of P₂O₅ (12–17%), CaO (24–30%), and an LOI in the range of 6.3–7.5%. The mineralogical composition of different fractions (Figure 6B) corroborates these results with the presence of mainly quartz and fluorapatite. Fluorapatite has an atomic ratio of Ca/P equal to 1.8, which was calculated as an average of 57 particles using STEM-EDS analysis. In Figure 7E, the phosphor, calcium, and fluor contents are 10.2 atom%, 17.5 atom%, and 3.7 atom%, respectively.

Dolomite aggregates are composed of CaO (~33%) and MgO (~18%), and an LOI in the range of 36–39% with the presence of SiO₂ (~7%) and a quantity of P₂O₅ (2.1–3.5%). The XRD pattern shows that it is mainly composed of Dolomite minerals associated with small peaks of quartz and fluorapatite. According to STEM (Figure 7F), the Ca and Mg composition is 16.2 atom% and 14.4 atom% on average; the ratio is equal to 1.12. The presence of aluminosilicates is indicated by an Al₂O₃ content in the chemical analysis. Needle-like particles of aluminosilicates were identified with TEM in Dolomite powders and were observed as webs in the pores of Dolomite-polished sections with SEM, as shown in Figure 8. Aluminosilicate clays, such as palygorskite and montmorillonite, are linked to different lithologies in the phosphate series [33,34].

3.3. Geotechnical Properties

Figure 9 presents the PSD where similar gradings for Flint, Phosflint, and Dolomite coarse aggregate were obtained.

Table 3. Geotechnical properties.

Aggregates Fraction	Geometric Properties		Physical Properties				Mechanical Properties			Cleanliness Properties
	Flakiness Index, %	Intergranular Porosity	Absolute Density, t/m3	Real Dry Density, kg/m3	Water Absorption, %	Total Porosity, %	Los Angeles, %	Micro-Deval, %	Flexural Strength, MPa	Sand Equivalent, %	Methylene Blue Value, g/kg
F-S	–	0.526	2.71	2440	4.8	10.0			–	83	0.75
F-G1	43	0.573	2.66	2423	3.1	8.9	25	11
F-G2	25	0.534	2.67	2467	2.3	7.6
F-G3	21	0.493	2.68	2525	1.8	5.8
PF-S	–	0.484	2.80	2347	6.1	16.2			17.1	83	0.75
PF-G1	40	0.579	2.78	2467	3.8	11.3	28	15
PF-G2	19	0.544	2.78	2511	3.1	9.7
PF-G3	17	0.496	2.78	2586	1.7	7.0
D-S	–	0.429	2.89	2157	7.1	25.4			16.1	78	1.25
D-G1	32	0.541	2.89	2539	3.8	12.1	32	23
D-G2	17	0.530	2.89	2570	3.0	11.1
D-G3	20	0.524	2.89	2611	2.3	9.7

presents the results of geotechnical properties. The strength of the G₁ fraction is an LA value equal to 25%, 28%, and 32%, and an M_DE value equal to 11%, 15%, and 23% for Flint, Phosflint, and Dolomite, respectively. The quality based on the (LA + M_DE) value is Flint > Phosflint > Dolomite. The LA value for natural construction aggregates varies in the range of 18–43% [35], and the M_DE value varies in the range of 5–22% [36]. The flexural strength of Phosflint is equal to 17.1 MPa, and Dolomite is equal to 16.1 MPa (Figure 10). A previous study showed that the compressive strength on rock cylinders of 60Ø120 mm of Phosflint was equal to 120 MPa, and for Dolomite, was equal to 70 MPa [19]. The strength of aggregates depends on the mineralogy and microstructure of the parent rock [37]. The important strengths of Flint and Phosflint are caused by the significant amount of silica (Table 2), and the strength of Dolomite is attributed to its compact microstructure and the fine arrangement of Dolomite rhombs (Figure 5).

Figure 11 presents the values of the real dry density and the water absorption of G₁, G₂, and G₃ fractions. Figure 12 presents the relationship between the total porosity and water absorption and the total porosity and real dry density. The linear regression shows a strong correlation with R² superior to 0.90. For Flint aggregates, the real dry density varies in the range of 2423–2525 kg/m³; for Phosflint aggregates. real dry density varies in the range of 2467–2586 kg/m³; and for Dolomite aggregates, it varies in the range of 2539–2611 kg/m³. The ranking of real dry density is ${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}$-D > ${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}$-PF > ${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}$-F, and the absolute density shows a similar ranking, ${\mathsf{\rho }}_{\mathrm{a}}$-D (2.89 t/m³) > ${\mathsf{\rho }}_{\mathrm{a}}$-PF (2.78 t/m³) > ${\mathsf{\rho }}_{\mathrm{a}}$-F (2.67 t/m³), because the density of minerals impacts the density of aggregates. The coefficient of water absorption of Flint aggregates is in the range of 1.8–3.1%; for Phosflint aggregates, it is in the range of 1.7–3.8%; and for Dolomite aggregates, it is in the range of 2.3–3.8%. The coefficient of water absorption shows the following ranking: WA₂₄-D > WA₂₄-PF > WA₂₄-F; the total porosity follows a similar trend: p-D (10–12%) > p-PF (7–11%) > p-F (6–9%). The porosity can be observed under SEM in Figure 5. The porosity of natural Flint can be as high as 16% [38], with specific gravity and water absorption in the range of 2.4–2.7 t/m³ and 0.4–7.6%, respectively [22,39]. For the three aggregate types, the real dry density increases with the increase in the size fraction, ${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}$-G₃ > ${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}$-G₂ > ${\mathsf{\rho }}_{\mathrm{r}\mathrm{d}}$-G₁, and there is an inverse ranking of the coefficient of water absorption: WA₂₄-G₁ > WA₂₄-G₂ > WA₂₄-G₃. Fine aggregates have a higher coefficient of water absorption than coarse aggregates, which is explained by their larger specific surface area.

The flakiness index of Flint aggregates is in the range of 21–43%, Phosflint aggregates have a flakiness index in the range of 17–40%, and Dolomite aggregates have a range of 20–32%. The flakiness index of the G₁ fraction is equal to 43% for Flint, 40% for Phosflint, and 32% for Dolomite. The high flakiness index of the G₁ fraction is explained by the brittleness of these rocks and the higher fragmentation that happens to the smaller fractions. Flint aggregates are known to break into a conchoidal shape, while Phosflint and Dolomite exhibit brittle behavior, as shown by the sudden decrease in stress after attaining the ultimate flexural stress in Figure 10. The intergranular porosity of Flint aggregates is in the range of 0.493–0.573, of Phosflint aggregates have intergranular porosity in the range of 0.496–0.579, and for Dolomite aggregates, it is in the range of 0.524–0.541. For the three aggregate types, intergranular porosity is ranked in relation to the size fraction, v-G₁ > v-G₂ > v-G₃, and there is a similar ranking for the flakiness index: A-G₁ > A-G₂ ~ A-G₃. This is explained by the irregular shape increasing the porosity between aggregate particles.

Sand fraction has a sand equivalent value in the range of 78–83% and a methylene blue value in the range of 0.75–1.25 g/kg. The methylene blue value (1.25 g/kg) and sand equivalent (78%) show the presence of impurities that lower the cleanliness.

3.4. X-ray Computed Tomography

Figure 13A presents the Flint sample analysis, which shows that the microstructure is dense; however, there is the potential existence of different phases with similar greyscale and distribution, but they are overlapping, which makes it difficult to separate. Figure 13B shows that Phosflint is composed of two phases with different greyscales and denser phosphate grains are represented by lighter particles, while silica is less dense and is represented by darker cement. Porosity was not identified in Flint and in the silica matrix of Phosflint at the 4.7 µm spatial resolution, which indicates that the pores are smaller than ~5 µm. In Figure 13C, Dolomite shows the presence of large pores that are identified as darker spots. The segmentation determined that the porosity of Dolomite is significant and equal to 7.8% of the surface area. The particle size analysis was able to determine that the average pore surface area is 135 µm². The pores were separated and formed a closed porosity, which explains the difference between the total porosity and water absorption values in Table 3.

The porosity identified by X-CT analysis is lower than the total porosity of the 10–12% value, which could be due to the voxel size being equal to ~5 µm in image acquisition. These results corroborate SEM observations (Figure 5): the existence of relatively homogeneous silica microstructures in Flint while the pores are below 5 µm was identified. Phosflint is composed of two phases: phosphate grains containing large pores and silica cement, while Dolomite presents the most prominent porosity and has a larger pore size.

4. Discussion

The three PWR aggregates exhibited favorable geotechnical properties to be used as concrete aggregates according to the European standard NF EN 12620: “Granulats pour béton” [23] and the French standard NF P 18-545: “Granulats“, article 10: Aggregates for Hydraulic Concrete [21]. The LA fragmentation and M_DE wear strength for the G₁ fraction of Flint and Phosflint are represented by code A, and for Dolomite, it is code B. These codes attest to the potential use of Flint, Phosflint, and Dolomite as aggregates for ordinary concrete and for concrete with a compressive strength superior to 35 MPa. Their mechanical resistance encourages the investigation of their suitability as alternative aggregates for road construction.

The code for the flakiness index and the water absorption for different fractions is presented in Figure 14. The main concern with using aggregates with high water absorption and a highly flaky shape is their negative impact on workability [40]. Higher water absorption is expected to demonstrate a decrease in the slump of concrete because of a higher water demand, which might also reduce the compressive strength. This can be overcome using a water reducer and mineral fillers.

The adherence of aggregates is one of the most important factors that influences the strength of concrete [41], such as limestone aggregates giving higher strength while crushed flint gives lower strength [42]. Flint has a smooth surface texture, which can lead to lower adherence and produces a weaker interfacial transition zone between the aggregate and cement paste, causing a debonding fracture under compressive strength testing, which can reduce its strength. Dolomite has a rough surface texture, which is more favorable for enhanced adherence and can improve the compressive strength of concrete. The adherence of Phosflint can also be beneficial because irregular phosphate grains create a rough surface texture.

5. Conclusions

The use of PWR as a concrete aggregate could encourage a circular approach by improving the management of PWR and the quarrying of natural aggregates. Flint, Phosflint, and Dolomite, sourced from PWR, were evaluated in terms of geotechnical properties, mineralogical composition, and microstructure, indicating their suitability for concrete aggregates. The porosity was described by microscopic observations, calculated with geotechnical tests, and confirmed by X-ray computed tomography analysis. Impurities were revealed by chemical characterization and were identified by microscopic observations.

The main conclusions are as follows:

• Flint aggregates break into a conchoidal shape, which causes a very high flakiness index in smaller fractions (equal to 43% in the G₁ fraction). Their irregular shape causes a very high intergranular porosity equal to 0.526. The total porosity is equal to 5.8% for the G₃ fraction, which is similar to natural Flint aggregates. It is caused by very fine and vein-like pores that can be observed under SEM. Flint aggregates are very strong, as attested with an LA and M_DE equal to 25% and 11%, respectively, placing them in code A for both these properties; they are suitable for the production of concrete with higher compressive strength. The surface texture is smooth and glassy, and the water absorption of the G₁ fraction is quite high (equal to 3.1%); this potentially causes adherence issues and the high water demand for concrete.
• Phosflint aggregates are a virgin aggregate type constituted of phosphatic particles and encapsulated in silica-rich cement. They have received little interest for use as concrete aggregates because the P₂O₅ content is superior to 12%. However, the use of Phosflint in concrete is possible considering its geotechnical properties: the G₃ fraction is represented by code A for water absorption, the flakiness index, and LA and M_DE values. The flakiness index is high for the G₁ fraction because of the brittleness of the rock. Even though it contains cryptocrystalline silica, the SiO₂ content is inferior to Flint, and the surface texture is rougher than Flint, making it a better candidate.
• Dolomite aggregates have a higher real dry density equal to 2611 kg/m³ for the G₃ fraction. They also possess important intrinsic strength with resistance to fragmentation LA that is equal to 32% and flexural strength on rock specimens equal to 16.1 MPa because they have a compact arrangement of Dolomite rhombs. Dolomite aggregates showcase a more regular shape and a lower flakiness index compared to other aggregates (flakiness index equal to 32% for the G₁ fraction) and a rough surface texture. These properties are more favorable for concrete aggregates because they improve adherence with the cement paste and packing density. However, Dolomite possesses an important porosity, up to 12% of the total volume caused by closed pores between Dolomite rhombs, with an average surface area equal to 135 µm². Dolomite aggregates possess the most favorable properties for use as concrete aggregates compared to Flint and to Phosflint.

The shape characteristics of aggregates are highly dependent on the crushing and sizing techniques. The use of a cone or an impact crusher could improve the shape properties [43,44] and improve the suitability of PWR aggregates. To assess the suitability of PWR for road construction, it is recommended to include more tests, such as the aggregate crushing test, the aggregate impact test, and the soundness test.

Concerning durability, the alkali–aggregate reaction (AAR) is an important issue in concrete, which causes the degradation of structures and reduces serviceability [45]. Microscopic observations showed that the presence of cryptocrystalline silica was significant in Flint and Phosflint, and quartz particles were found in Dolomite. These constituents are potentially reactive towards AARs, and they may cause alkali–silica or alkali–carbonate reactions [46]. Chloride, alkalis, and sulfur were not identified in significant proportions (STEM, SEM, XRF, and XRD); therefore, durability-related tests concerning the solubility of sulfate, chloride, alkalis, and silica (AAR) need further investigation using standard tests.