Viability of Morisca Powder Tailings for Ceramic Applications

Abstract

In Juazeiro de Piauí and Castelo do Piauí, two cities in northeast Brazil, mined morisca rock is widely used in civil construction for external coating applications. Aiming at a sustainable practice for its extraction, which includes the economical use of the tailings from its mining, this work presents a solution for the use of discarded materials from processing morisca from Piauí. In this context, an analysis of the results of using morisca as a substitute for clay materials for applications in red ceramics are presented. For the analyses, XRF, XRD, microgranulometry, limits of consistency, TG, DTG, and a set of technological tests inherent to ceramic materials were used. It was concluded that morisca powder is viable as a clay substitute for red ceramic applications.

1. Introduction

The Brazilian lithological industry is still promising, because even with the economic crisis, in terms of ornamental and coating stones, Brazil exported a total of USD 1,014,293 and 2,165,405 t in 2019, with a positive variation in billing but a contraction in physical volume, when compared with 2018 [1].

In the state of Piauí, in Brazil, specifically in the regions that include the cities of Castelo do Piauí and Juazeiro do Piauí which are located in the centre of the northern area of the state, the extraction and processing of rock called arcosian sandstone (extracted quartzite)—commercially known as morisca stone—takes place [2]. It reached a production of 100,000 t in 2019, along with the slate produced in the state [1].

Arcosian sandstones are detrital sedimentary rock formations composed of solid particles such as sand and gravel, originating from parent material by mechanical and chemical weathering, composed mainly of quartz and feldspars [3]. In the region mentioned above, this ornamental stone has classic chromatological characteristics, with macroscopic colour tones from yellowish to slightly reddish, grey with a tendency to black, and multicolour [4]. Mafic minerals cause such colours, but morisca is not classified as mafic, as it has a high silicon content and is a rock that occurs in the most superficial layer of the earth (extrusive). For simplification of nomenclature, from now on we will call it morisca.

In general, its extraction still occurs without the implementation of technological resources, carried out in the open (Figure 1A) after cleaning the area chosen for mining, followed by squaring the structural plans of morisca through manual and rudimentary actions, with the use of suitable metal tools to obtain it. At this stage, it is already possible to detect many tailings (Figure 1B) accumulated near the quarry areas [5]. With the blades removed, these are taken for processing; they can be ground and polished, although they can be found commercially without any processing. More waste is generated during the grinding and polishing process, namely, morisca powder (Figure 1C), along with a large volume of used water. Around 20,000 litres of water is used per day, if all the machines are on, coming out of two storage tanks (75,000 litres and 150,000 litres) filled with decantation water, as all the water is reused in morisca processing procedures.

From this perspective, the environmental liabilities of this mineral prospect have grown each year, requiring the development of technological strategies for the reuse of the generated tailings, those strategies preferably being economically viable since only then can investment be undertaken to remediate this environmental problem.

In this scenario, as the traditional ceramic industry based on clay–silica–feldspar consumes a large amount of natural raw materials, its raw materials consisting of a diversified composition, tolerating fluctuations in its composites and the addition of different types of waste, morisca presents itself as a candidate to provide a solid waste reuse stream [6].

To exemplify that the statement above has been consolidated, we identified several types of mining tailings tested for construction materials in the literature: incorporation of quartzite fines for the production of ceramic materials [7,8]; copper–nickel slag aggregation in a multi-component formulation for the production of building ceramics, with low percentages of nickel (0.14% of NiO) and copper (0.06% of CuO) in the slag [9]; application of copper–nickel ore treatment tailings (0.09% and 0.20%), nepheline apatite and iron quartzite in clay-free construction ceramics, by means of compression molding [10]; association of clayey materials with the primary sludge from the red ceramic paper industry [11]; use of clay from gypsum mining and gypsum residue for the production of red ceramics [12]; addition to the ceramic mass of tailings rich in quartz and heavy metals from mining [13]; use of ash from the solid waste and tailings factories of a Greenland mine to produce clay-based bricks [14]; use of different levels of marble powder to improve the properties of ceramic bricks [15]; use of granite and marble tailings as raw materials to be aggregated with clay in the production of red ceramics [16,17]; and, viability of soapstone and soapstone residues soaked in oil, added to the clay mass for the manufacture of red ceramics [18].

Mining tailings are often incorporated into clayey masses to produce ceramics. Therefore, starting from the premise that the ceramic industry is a receptive sector to the incorporation of residues for the production of its products [19], as well as morisca powder manifesting an inherent characteristic of clayey materials, this objective work studies the technical feasibility of using this material in the red ceramics industry, following Brazilian technical norms, aiming to replace the raw material of this economic segment (clay) with morisca powder derived from processing waste.

2. Materials and Methods

2.1. Feedstock

The morisca powder used was purchased from ECB Rochas Ornamentais do Brasil Ltd., derived between the municipalities of Castelo do Piauí and Juazeiro do Piauí (Brazil) (5°12′37.2″; S 41°42′;12.5″ W), and was the result of the extraction, thinning, and polishing of the rock sheets in the beneficiation process. The sample was submitted to NBR 6457 [20], with the sample dried until achieving hygroscopic humidity and breaking up of the clods. The material was deagglomerated in a ball mill and dried in a laboratory oven at 110 °C ± 10 °C.

2.2. Chemical, Mineralogical, and Thermal Analysis Tests

Samples were passed through a #200 sieve (75 μm) for chemical, mineralogical and thermal tests. The semi-quantitative chemical analysis of the material was performed by the fluorescence technique of X-rays (FRX), using Shimadzu EDX-720 equipment, with a detection limit from Na (11) to U (92) (>0.1%).

The mineralogy was studied by X-ray diffraction (XRD) using a Shimadzu diffractometer, XRD 6000 model with CuKα (λ = 0.15406 nm), collecting diffraction data in 2θ from 5° to 75°, and a power of 1.5 kW with a step size of 2° min⁻¹. To better investigate the morisca peak attributions, another mineralogy study was repeated by X-ray diffraction (XRD) using a step size of 0.02° min⁻¹ with the same other parameters. The identification of the crystalline phases was carried out through a search in the database with the X’Pert system, HighScore Plus.

Thermogravimetric analysis (TGA) was performed on a TA Instrument SDT Q600 in a nitrogen atmosphere, with a heating rate of 10 °C/min between room temperature (25 °C) and 1200 °C. The amount of sample used was approximately 10.73 mg in a platinum crucible. For the thermodifferential analysis (DTA), the Calorimeter RB 3000–20 was used, with a heating rate of 10 °C/min from room temperature to 1150 °C.

2.3. Particle Analysis

After the milling process, the material was passed through a #40 sieve and then subjected to a particle size test, performed with a SediGraph III Plus V1.00 using the methodology of sedimentation by gravity monitored by X-rays, and the calculations were conducted based on Stokes’ and Beer’s laws. It was also possible to determine the material’s actual density of 2.65 g/cm³.

2.4. Consistency Limits

A portion of the material was passed through a #80 sieve, and distilled water was added and mixed to the presumed moisture point of the liquid limit to determine the Atterberg limits. Then, the prepared material was placed in a plastic bag so as not to lose moisture, and left to rest for 24 h, after which it was thrown into the liquid limit apparatus to determine the points of the humidity/number of blows ratio; as the temperature was increased, the humidity determined the liquidity limit at 25 strokes [21]. Portions consisting of 60 g of this material were used to determine the moisture content (h), and 10 g portions were used to determine the plasticity limit. Therefore, it was possible to calculate the plasticity index by the liquidity limit difference minus the plasticity limit [22].

2.5. Specimens and Technological Tests

With the material atomised and passed through a #80 sieve, the individual mass was determined for the dry and wet specimens, which were shaped by uniaxial pressing with an approximate load of 55.1 MPa, in a laboratory hydraulic press, through a prismatic matrix box sized 8 cm × 1 cm (Figure 2A), with an average height after pressing of 1 cm (Figure 2B). Dimensional measurements were carried out with a 0.05 mm resolution caliper. The specimens were weighed with a 0.01 g digital scale and then placed in an oven for 24 h after drying in the shade, to carry out the dimensional measurements and weighing of the specimens again after drying.

The firings were carried out at 800 °C, 900 °C, 1000 °C, 1100 °C, and 1150 °C in a muffle furnace, with a heating rate of 4 °C/min. New dimensional measurements and weight measurements were performed. After all these steps, it was possible to calculate the water absorption by Equation (1), the porosity by Equation (2), and the apparent specific mass by Equation (3). The linear drying shrinkage and the linear burning shrinkage were calculated by Equation (4). In contrast, the total shrinkage was derived by Equation (5). The equations cited are found in

Table 1. Equations of technological tests.

Variables	Equations		Description
Water Absorption ($WA$)	$WA=\frac{ww-dw}{dw}\times 100$	Equation (1)	$dw$ = dry weight $ww$ = wet weight $av$ = apparent volume $L_i$ = initial length $L_f$ = final length
Apparent Porosity ($AP$)	$AP=\frac{ww-dw}{av}\times 100$	Equation (2)
Apparent Specific Mass ($ASM$)	$ASM=\frac{ dw}{av}$	Equation (3)
Drying and Firing Linear Shrinkage ($S_{D/F}$)	$S_{D/F}\left(\%\right)=\frac{L_i-L_f}{L_f}\times 100$	Equation (4)
Total Retraction	$R\left(\%\right)=S_{Drying}+S_{Firing}$	Equation (5)

.

As for the flexural strength, the results were obtained by the three-point bending test, with a speed of 0.5 mm/min, using a EMIC deflectometer, DL 20,000 model, and with a load cell of 200 kN. We emphasise that the procedures adopted for the technological tests had the theoretical contributions of NBR 15270-2 [23] and NBR ISO 10545 [24] standards, as well as the literature on ceramic tests [25].

3. Results and Discussion

3.1. Raw Material Characterisation (before Firing)

Table 2. Result of the semi-quantitative chemical analysis of the morisca sample by XRF.

Oxides	SiO₂	Al₂O₃	Fe₂O₃	K₂O	TiO₂	MgO	CaO	P₂O5
(%)	70.55	14.24	9.10	2.49	1.70	1.40	0.33	0.33

shows the weight percentage results of the semi-quantitative chemical analysis of the morisca sample. XRF identified a high content of 70.55% silicon oxide (SiO₂), 14.24% aluminium oxide (Al₂O₃), 9.10 % iron oxide (Fe₂O₃), 2.49% potassium oxide, and other oxides with percentages below 2%. Given the SiO₂ content, a component that reduced the plasticity of the sample, the Al₂O₃ contributed to the plastic characteristic of morisca and collaborated with water retention [26], favouring the moulding capacity of pieces made with morisca [27], so the sample had an indication of medium plasticity, confirmed with the Atterberg tests shown in

Table 3. Granulometric composition in the percentage of the morisca sample.

Sample	Fine Quarry Materials (%)	Silt (%)	Sand (%)	Average Diameter
	v% ≤ 2 μm	2 μm < v% ≤ 20 μm	v% > 20 μm	μm
Morisca	60	37.4	2.6	1.34

. Fe₂O₃ occurred in a combined form and was responsible for the reddish tones of the pieces when fired; in addition, it negatively interferes with plasticity and mechanical strength, so its occurrence rate should be less than 10% [28]; thus, the morisca sample met the suggested limit.

The mineralogical analysis results are presented in the morisca diffractogram (Figure 3). A comparison of the diffractogram peaks with the diffraction pattern base identified compatibility with the following phases: quartz (SiO₂—PDF-01-086-1630), anorthite (CaAl₂Si₂O₈—PDF-01-086-1706), magnesium titanium oxide (Mg₂TiO₄—PDF-001-079-0830), iron titanium oxide (FeTi₂O₅—PDF-01-076-2372), potassium aluminum silicate (CaAl₂Si₂O₃—PDF-000-014-0081), aluminum silicate (K₅₇Si₁₃₅Al₅₇O₃₈₄—PDF-01-073-0491), and three suspiciously recognizable diffraction peaks, that we placed as unidentified, which disappeared after the first burn. Apart from quartz and tridymite, which are phases associated with silicon oxide, the other phases composed the fluxes that incorporated the plastic characteristics manifested by the morisca powder [29].

Despite being a plasticity-reducing agent, we emphasise that quartz is desirable in ceramic materials since it guarantees structural integrity during firing, controlling shrinkage and favouring the pieces’ drying. However, it is essential to be careful with the burning rate because at around 573 °C alpha quartz undergoes an exceptionally rapid transformation to beta quartz, causing volume increase and, as a consequence, cracks in the burned parts [30]. Therefore, the diffractogram corroborates the semi-quantitative chemical analysis presented in Table 2.

The micro granulometric characterisation of morisca is presented in Table 3 and Figure 4. We analysed the particle size distribution through the classification proposed by the International Society of Soil Science, which states that granulometric fractions smaller than 2 μm are defined as clay, granulometry between 2 and 20 μm are silt, and granulometric fractions larger than 20 μm are classified as sand [31]. Thus, the sample comprised 60% excellent quarry materials, 37.4% silt, and 2.6% sand, and an average diameter of 1.34 μm. These granulometric characteristics of morisca, mostly smaller than 2 μm, give this material the plastic characteristic of the morisca and water system, making it possible to create products similar to the clays used in the red ceramic industry [32].

Based on the results of the Atterberg tests (

Table 4. Plasticity index of the morisca sample.

Sample	Liquid Limit (LL)	Plastic Limit (PL)	Plasticity Index (PI)
	%	%	%
Morisca	38.24	24.61	13.63

), the sample may be classified, in terms of plasticity, as weakly plastic if the plasticity index (PI) is between 1 and 7, moderately plastic if the PI is between 7 and 15, and highly plastic if the PI is greater than 15 [33]. It is worth mentioning that these tests make it possible to identify the plasticity limit (PL) and the liquidity limit (LL), which, together with the PI, indicate the amount of water needed to carry out the moulding of the parts [34]. A value of 24.61% was found for the LP and 38.24% for the LL; these are the minimum and maximum limits in the percentage of water that can be added to the material to keep it in a plastic state. We obtained, for the morisca, a value of 13.63% for the PI, which allowed us to classify the material as moderately plastic. This characteristic can be explained by the high silica content identified in the semi-quantitative chemical analysis (Table 2). However, it was an excellent result because it would be difficult to dry if it had high plasticity, which could produce cracks in the parts.

The Atterberg limits tests also facilitate an extrusion prognosis, that is, to identify whether a material is in a more favourable region for the conformation of ceramic pieces by the extrusion method. Although the specimens of this work did not use the extrusion method, which is widely used in the ceramic industry, it was still essential to show that it is possible to use morisca utilising this methodology, as at an industrial level it impacts productivity and workability with the green pieces produced [35]. In this context, Figure 5 presents the result of the morisca extrusion prognosis, which was classified as acceptable for forming parts using the extrusion method even though it was not in an optimal region.

The lost mass identified was 0.484 mg (Figure 6), equivalent to 4.51% of the sample mass. It was a slight loss, a desirable characteristic for ceramic applications, and must be related to the elimination of water and evaporation of impurities that may exist in the material.

Figure 7 shows the thermodifferential analysis, where two endothermic peaks were identified. The first peak (482 °C) was associated with the dehydroxylation of clay minerals. The second peak (530 °C) was attributed to the transition from alpha quartz to beta quartz. In this case, it is crucial to maintain a low burn rate in the temperature range that composes the second endothermic peak, avoiding problems that may occur with the quartz transition.

3.2. Characterisation after Burning

Figure 8 compares the morisca diffractograms with morisca after firing at 800 °C, 900 °C, 1000 °C, 1100 °C, and 1150 °C. The crystalline phases identified in the firing at 800 °C were quartz (PDF 01-078-1422), anorthite (PDF 00-041-1486), and potassium aluminium silicate (PDF 00-046-0741); at 900 °C they were quartz (PDF 01-00-033-1161), anorthite (PDF 01-073-00-041-1486), and potassium aluminium silicate (PDF 00-046-0741).

Upon firing at 1000 °C, quartz (PDF 01-078-1422), anorthite (PDF 00-041-1486), and potassium aluminium silicate (PDF 01-079-1823) were observed in the same phases. As for the firing at 1100 °C, quartz (PDF 01-083-0539), anorthite (PDF 00-041-1486), and new phases microcline (KAlSi₃O₈ PDF 00-022-0675) and hematite (Fe₂O₃ PDF 001-073-0603), were observed.

Quartz peaks existed in all diffractograms, including silicon oxide, an isomorph of quartz, identified in the burning at 800 °C. However, at 1150 °C, only three phases were observed: quartz (PDF 01-079-1910), anorthite (PDF 00-041-1486), and hematite (Fe₂O₃ PDF 01-079-0007). The quartz peaks were smaller, and could be associated with the partial dissolution of quartz to a dense phase, favouring the reduction in porosity in the pieces. The hematite that emerged at this temperature brought the chromatology of the fired pieces to a darker reddish colour.

The results of the apparent porosity (AP), water absorption (WA), and specific mass (SM) tests are shown in Figure 9, which allows us to state that the apparent porosity and water absorption decreased with increasing temperature while the specific mass increased. This behaviour is expected for materials with clayey characteristics after firing.

Apparent porosity is closely linked to densification and sintering resulting from the material firing process [36], making it an essential parameter for ceramic applications. In this sense, porosity was reduced from 31.37% at 800 °C, to 4.56% at 1150 °C, which affected the other mechanical properties of the parts. The behaviour of water absorption, which depends on the porosity of the specimens, maintained the same downward trend in the apparent porosity. Regarding values, the water absorption at 800 °C and 900 °C was above the values allowed in the ABNT standard, which defines that ceramic products, specifically for masonry, must have a water absorption index between 8% and 22% [37]. However, for the production of tiles, the norm only determines the maximum index allowed of 20%, emphasising that these are values for applications in temperate or tropical climates [38]. For masonry, at 1000 °C and 1100 °C, the specimens met the normative specifications, while at 1150 °C, the result was below the lower limit allowed but in regulatory compliance for the production of tiles. The amortisation in the porosity performance directly impacted the specific mass, as the reduction in void spaces in the tested bodies promoted more significant densification of the material, and consequently, an increase in the specific mass due to the decrease in porosity, starting from 1.31 g/cm³ at 800 °C, to 1.88 g/cm³ at 1150 °C [39].

The variables AP, WA, and SM (Figure 9) directly impacted the flexural strength of the specimens shown in Figure 10. As for porosity, flexural strength is subject to the normative support of ABNT [32], which advocates a minimum strength of 2 MPa (20 kgf/cm²) for ceramic products. It can be seen in Figure 10 that the increase in temperature produced an increase in the resistance values of the specimens, but this behaviour did not meet the norm up to 1000 °C. However, at 1100 and 1150 °C, it was possible to identify compliance with the standard. The gain in strength above 1000 °C was strongly linked to the sintering of the clay minerals that emerged—calcium aluminium silicate, microcline, and potassium iron silicate—as predicted in the mineralogical analysis (Figure 8). Although calcium aluminium silicate and microcline were insufficient to place the strength at 1000 °C in the acceptable standardisation zone, their effects were added to the potassium iron silicate, raising the flexural strength for ceramic applications to a favourable condition at 1100 °C. As for the resistance at 1150 °C, where the beginning of the sintering of quartz was recorded, although it helped reduce porosity with its dense phase, in excess, it produced a slight loss in resistance, which was noticeable in our results.

The result of the total linear shrinkage test was used to estimate the dimensioning of the parts produced in the face of the temperatures, shown in Figure 11. It is known that, generally, shrinkage is the result of the elimination of water and decomposition of organic matter, as the granulometric composition of the material strongly influences it, since the smaller the granulometry, the greater the results measured for the retraction in the burned bodies [25].

Although the microgranulometry ensured that more than 60% of the material was composed of fine particles smaller than 2 μm (Table 3 and Figure 3), it was observed that up to 1000 °C, the manifested shrinkage was minimal, around 0.25%. However, there was an increase to 6.24% at 1100 °C, and 10.57% at 1150 °C. These results resonated perfectly with the porosity decay results (Figure 9). Therefore, we can say that the reduction in porosity with the increase in temperature promoted a decrease in void spaces, thus, having a positive impact on the reduction in water absorption and enlargement of the specific mass, allowing a desirable increase in the mechanical resistance of the tested bodies, maintaining the shrinkage suffered by the bodies within acceptable limits.

As the most relevant results were obtained with the samples sintered at 1000 °C, 1100 °C, and 1150 °C, the images of the scanning micrograph of the fracture surfaces of the samples are presented for the temperatures mentioned above (Figure 12). Figure 12a shows a significant amount of pores, the uniformity of morisca grains and quartz grains, and a low level of sintering of clay minerals. However, it is evident in Figure 12b that there was an increase in the level of sintering of the clay minerals, as the porosity reduction was noticeable and significant, as fluxes filled the spaces, improving the accommodation of the particles, improving the densification in the pieces and, therefore, contributing to the increase in mechanical strength. In Figure 12c, the porosity reduction is noticeable, as the level of filling of the spaces between the particles intensified, with a significant restriction on the visualisation of the grains, denoting that the dense phase of quartz was incorporated into the fluxes, making the microstructure of the pieces more compact.

4. Conclusions

The work described the characteristics of the morisca rock powder to ascertain its technical viability as a substitute material for clays for applications in the red ceramic industry. As a result of this investigation, we confirmed that this material is mainly comprised of silicon oxide, presenting plasticity, which allows the formation of ceramic pieces by pressing or extrusion, as well as having traditional melting clay minerals in its composition, capable of providing the mechanical resistance necessary for ceramic products. We verified with burning tests that using morisca powder requires a temperature of around 1100 °C for applications in masonry, as well as caution with the burning rate, as the material has a considerable concentration of quartz.

Conjecturing the economic viability inherent to ceramic materials, although it was not the focus of this work, we can recall the most common parameters that affect the cost of ceramic production, such as transport of the raw material, the preparation process of the materials involved, the type of energy that will be used for burning (electricity, coal or wood), and oven technology, among others. Murmu [40] clarified that for each type of kiln, there is a different energy consumption, and Zang [41] informed that to produce a burnt brick, an average of 2.0 kWh of energy is spent (parameters were not mentioned). This information demonstrates that many variables interfere with economic analysis. Nevertheless, for morisca, we can guarantee that it will incur the cost of transport, referring to the 50 km distance to the nearest ceramic pole, and the cost of grinding if demand exceeds the disposal of the settling tanks. However, in Brazil, industries that incorporate environmental liabilities into their operations receive tax incentives from the government; tax reductions are an example of these actions.

Given the above, we conclude that the environmental liability of morisca powder has been technically shown to have the necessary potential for its use as a raw material in the red ceramic industry.