Samples of clay collected in Boane district—Maputo Province (OUA clay) and SPL from MOZAL were dried and pulverized, then stored in sealed plastic bags at room temperature.

Chemical composition of the SPL and clay samples was determined semi-quantitatively by X-ray Fluorescence (XRF) and the mineralogical analysis was done using X-ray Diffraction (XRD) to identify the phases present in SPL samples. The Rietveld method was used to estimate the present phases. The thermal behaviour was determined using thermogravimetric analysis (TGA).

The clay specimens were prepared using a prismatic mould (15 × 3 × 2.5 cm). The specimens were prepared by adding SPL in contents varying between 0 and 20% (Table 1). The specimens were first dried in air at room temperature for eight days, changing the exposed side each day. The specimens were transferred to an oven, where they were dried at 105˚C for 24 hours. Firing test pieces were carried out in an electric kiln at a rate of 5˚C/min to reach the maximum temperatures of 850˚C, 900˚C and 950˚C, for a period of 15, 30 and 45 minutes and cooled to a room temperature. Five specimens were tested for each condition and the average was recorded. The electric kiln was previously calibrated, and all specimens were prepared by the same operator.

Table 1. Nominal SPL content (in wt-%) of prepared samples.

Optimization of the process was based on results of the effect of firing temperature (850˚C to 950˚C), firing time (15 to 45 minutes) and SPL content (0 to 20%) on mechanical properties of fired masses, particularly linear shrinkage, weight loss, compressive strength, bulk density, and water absorption.

Determination of Linear Shrinkage (LS)

To determine the linear shrinkage of dried and fired samples, the methodology by [21] was used. The length of the specimens was measured right after moulding (C₀), after drying in an oven (C_s) and after firing (C_q) in an electric kiln. Equations (1) and (2) were used to calculate the Linear Drying Shrinkage (LDS) and the Linear Firing Shrinkage (LFS), respectively. Length measurements were performed using a digital Vernier calliper (resolution ± 0.001 mm).

Determination of Weight Loss (WL)

To determine the WL, the methodology used by [22] was followed. The specimen was weighed after drying in an oven (m₂) and after firing (m₃). Equation (3) was used to determine Weight Loss (WL).

Determination of Bulk Density (BD)

To determine the BD, the methodology by [23] was used. The specimen was weighed after firing (m₃, in g) and the specimen volume was determined from the dimensions of the specimen after firing (V, in cm³) with the aid of a digital Vernier calliper (resolution ± 0.001 mm). Equation (4) was used to calculate the bulk density (AD).

Determination of Water Absorption (WA)

To determine the WA, the methodology used by [24] was followed. The fired specimen was weighed after drying in an oven at 105˚C for 24 hours (m₃) and it was weighed after resting the specimen in water for 24 hours (m₄) after removing excess surface water with a cloth.

Determination of Compressive Strength (CS)

To determine the CS, the methodology used by [22] was followed. First, with the aid of a digital vernier caliper the edge of the specimen (a) was measured in cm and the area of the specimen section (A) in cm² was calculated and the compressive strength tests were performed in a Press (brand: AMSLER, model: SCHAFFHOUSE-SUISSE 699/365) with loading rate of 0,6 MPa/s. Equation (6) was applied to determine the CS.

Table 2 presents results of the chemical composition of OUA clay and SPL, while Table 3 and Table 4 present mineralogical compositions of OUA clay and SPL. Chemical analyses were performed by XRF, which is insensitive to light elements such as carbon and fluorine; therefore, the results of the chemical composition of SPL in Table 2 do not include coal content.

Clay is composed essentially of silica and alumina or magnesia and water, but iron replaces aluminium and/or magnesium in several positions, and appreciable amounts of potassium, sodium and calcium are often present. Expectedly, these results revealed high Si, Al and Fe contents, consistent with the fact that clay minerals are phyllosilicates. This clay sample presented the higher content of iron. This element provides reddish tones after firing the ceramic mass [25][26].

Table 2. Chemical composition (%) of OUA clay and spent pot lining.

The results of XRD revealed that OUA clay is predominantly composed of quartz, plagioclase, corundum, and muscovite as main phases (Table 3). These phases are related to the high amounts of the elements silicon, aluminium and iron.

Table 3. Mineralogical composition of OUA Clay.

The XRD results showed that the SPL sample has expectedly high contents of graphite, cryolite and alumina (Table 4). Apart from the carbon and fluorine content the chemical composition of SPL showed presence of aluminium. Iron and sulphur, with relatively small amounts of calcium, silicon and nickel were also present [4][16].

Table 4. Mineralogical composition of Spent pot lining.

Figure 1 and Figure 2 show the thermal behaviour of OUA clay and SPL, respectively. TGA was used to characterize the thermal behaviour of the clay samples. The TGA was carried out with a TA Instrument SDT-Q600 Simultaneous/DSC. Samples weighing approximately 15 - 20 mg were heated from ambient temperature to 900˚C at a heating rate of 10 K/min under an inert atmosphere (nitrogen) and air at a flowing rate of 100 mL/min. Diagram in Figure 1 show the typical water loss of a clay (starting at temperature below 100˚C) with a further weight loss extending up to 600˚C, which includes the firing of organic matter. Figure 2 shows a large weight loss peak in the interval 500˚C - 800˚C associated with carbon combustion in air atmosphere (Sun et al., 2019; Sun et al., 2021). Carbon combustion easily occurred in the oxidative atmosphere. Similar results were also reported by [20], comparing 70% CO₂/30% O₂ and 70% N₂/30% O₂ atmospheres. In the inert atmosphere, the mass loss was less than 5%, which may result from the decomposition of cryolite [20].

Figure 1. Thermogravimetric analysis of the OUA clay.

Figure 2. Thermogravimetric analysis of the spent pot lining.

Firing temperature

Figure 3 presents results of the effect of firing temperature on properties of fired masses with 10% SPL fired for 30 minutes. Increasing the firing temperature, the linear shrinkage increases up to 900˚C; after this temperature, it drops. It would be expected that with increasing temperature the shrinkage would increase because of the removal of volatile or organic matter. But it is important to consider that the increase in temperature can favour the dilation of the pores, improving the exposure of the organic material to air, thus allowing it to burn. This burning will produce gases that will increase porosity, which increases the volume of the specimens.

Figure 3. Effect of firing temperature on properties of OUA clay with 10% spent pot lining fired for 30 minutes. (LS: Linear Shrinkage, WL: Weight Loss, BD: Bulk Density, WA: Water Absorption and CS: Compressive Strength).

This observation is confirmed by the appearance of the specimens shown in Figure 4; specimens fired at 950˚C have a high porosity and greater volume compared to those produced at lower temperatures. This behaviour was also observed by [22]. Weight loss increases with the increase in firing temperature. Reference [27] found similar results. According to the authors, weight loss increases due to the increase in water loss associated with hydroxides, clay mineral water, loss of volatile material and burning of organic material during the firing process.

Figure 4. Visual appearance of specimens fired at (a) 850, (b) 900 and (c) 950˚C for 30 minutes with 10% spent pot lining.

From the results in Figure 3, a firing temperature of 900˚C was selected as the optimal one and was used in further experiments. Samples fired at 900˚C show the highest compressive strength values and the lowest water absorption. Firing of specimens at temperatures higher than 900˚C introduces deformation of fired material (Figure 4), which may have contributed to the reduction of compressive strength and the increase of water absorption values and linear shrinkage.

Firing time

Effect of time of firing on specimens with 10% SPL, carried out at 900˚C, is presented in Figure 5. Linear shrinkage increased with increasing burning time; a fact also observed by [27].

Although firing time of 45 minutes gave best values of the compressive strength, optimal time selected was 30 minutes, since specimens fired at 45 minutes causes cracks and warping (Figure 6). Increased weight loss and improved water absorption and compressive strength are a result of a better completion of the transformations taking part during firing of the material, including a higher degree of

Figure 5. Effect of firing time on properties of clay with 10% spent pot lining, fired at 900˚C. (LS: Linear Shrinkage, WL: Weight Loss, BD: Bulk Density, WA: Water Absorption and CS: Compressive Strength).

Figure 6. Visual appearance of specimens fired at 900˚C for 45 minutes with 10% spent pot lining.

vitrification of fired clay [27].

SPL content

Figure 7 presents properties of fired clay masses with different amounts of spent pot lining in the range of 0 to 20%, fired for 30 minutes at 900˚C. This temperature and time were observed to be those that produce specimens with the best properties. The replacement of clay by SPL caused a reduction in shrinkage; this behaviour was also observed by [28] in the incorporation of waste clay ceramics from a water treatment plant. The continuous addition of SPL resulted in an increase in weight loss. Increased weight loss is explained hereby, particularly by the combustion of organic matter present in the SPL [22][27]. With increasing percentage of SPL, the bulk density and water absorption decreased to 10%, then increased. This property confirms the behaviour of bulk density because, according to [29], the higher the density, the lower the water absorption.

Figure 7. Properties of fired clay masses with increasing amounts of spent pot lining, fired for 30 minutes at 900˚C. (LS: Linear Shrinkage, WL: Weight Loss, BD: Bulk Density, WA: Water Absorption and CS: Compressive Strength).

Water absorption and compressive strength registered an optimal value at 10% spent pot lining addition. Figure 8 shows the comparison between the minimum values established by the [30] standard to produce solid bricks and the compressive strength values in the partial replacement of clay by the SPL residue. Clay masses with 5% and 10% SPL can be used in the production of solid bricks of all classes (A, B and C), while clay masses with 15% can be used for the production of solid bricks of category B and C. Clay masses with 20% cannot be used for the production of solid bricks as it does not have strength within the acceptable limit.

Figure 8. Comparison of compressive strength values with values for solid bricks (norm NM 127 2009).