Use of recycled carbon fibre as an additive in the manufacture of porous bricks more durable against salt crystallization

Ceramics International 50 (2024) 9682–9696 Available online 26 December 2023 0272-8842/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Use of recycled carbon fibre as an additive in the manufacture of porous bricks more durable against salt crystallization Laura Crespo-L´ opez a , * , Chiara Coletti b , Salvador Morales-Ruano a , c , Giuseppe Cultrone a a Department of Mineralogy and Petrology, Faculty of Sciences, University of Granada, Avda. Fuentenueva s/n, 18002, Granada, Spain b Department of Geosciences, Universit` a Degli Studi Di Padova, Via Giovanni Gradenigo, 6, 35131, Padova, Italy c Andalusian Institute of Earth Sciences (IACT), University of Granada-CSIC, Avenida de las Palmeras, 4, 18100, Armilla, Granada, Spain ARTICLE INFO Handling Editor: Dr P. Vincenzini Keywords: Bricks Carbon fibre Recycling Circular economy ABSTRACT Within the framework of the Sustainable Development Goals of the Agenda 2030, the circular economy is being promoted as a means of ensuring a sustainable use of resources and a reduction in the amount of waste produced. The aim is to reduce the demand for often scarce raw materials through the continuous reuse, recycling and regeneration of materials and products. This paper explores the use of carbon fibre from wind turbine blades as an additive in the production of new efficient bricks. Clay mixes with 0, 5 and 10 wt% additive were fired at three temperatures (800, 950 and 1100 ◦C) and the fired bricks were analysed from mineralogical and physical points of view to determine their suitability for use in the construction industry. The results show that carbon fibre improves the durability of the bricks, which became 16 % more porous as the firing temperature increased. However, the compressive strength of the bricks with 10 wt% carbon fibre was about 50 % lower than that of the control bricks made without additive. It is interesting to note that the distribution of the carbon fibres within the brick varies considerably and that they are shorter and wider in the core of the samples. These results could offer an alternative line for new product development in the brick industry. The bricks tested here are an example of a circular economy in which waste from one industrial process (wind turbine blades) is reused as an input in another (brick manufacture). The environmental benefits achieved are twofold: reduced demand for clay and recycling of decommissioned turbine blades, which are currently amassed in wind turbine graveyards. 1. Introduction The circular economy is a sustainable economic model that aims to reduce resource consumption and waste generation by promoting the long-life usage, recycling and regeneration of materials [1,2]. By adopting circular practices, businesses and society can achieve long-term environmental and economic benefits, increasing resource efficiency and creating new employment opportunities in green industries [3,4]. Great efforts are currently being made to achieve sustainable development, inspired by the challenges outlined in the United Nations 2030 Agenda (Sustainable Development GoalsSDG) (UN General Assembly, [5]) in spite of these efforts, the Circularity Gap Report 2023 [6] reveals that the global economy is still only 7.2 % circular. It is therefore necessary for each economic sector to assess how circularity can best be achieved. Within the construction sector, one option that is currently being explored is the use of alternative and/or local raw materials in the manufacture of building materials. To this end, Gencel et al. [7,8], Er et al. [9], Crespo-L´ opez et al. [10], and Erdogmus et al. [11], have recently studied the use of inorganic residues such as industrial slags, water treatment slag, household glass and metallurgical wastes as additives in brick production. These additives modify certain physical properties of the fired bricks (above all, porosity, mechanical strength and durability). The aim of this paper is to support the brick industry in its search for suitable alternative raw materials that can be used in the manufacture of bricks, so reducing the amount of clay required. In this case, the study focuses on the carbon fibre waste from decommissioned wind turbine blades. In line with the SDG challenges (in particular SDG 12, “Responsible Consumption and Production” which has the most targets specifically addressing waste management), there are two main objectives, namely to reduce the demand and therefore the extraction of natural resources and to reuse and recycle as raw materials the waste produced by other anthropogenic activities. Successful application of alternative raw materials of this kind would also have socioeconomic benefits in terms of cost savings. * Corresponding author. E-mail address: [email protected] (L. Crespo-L´ opez). Contents lists available at ScienceDirect Ceramics International journal homepage: www.elsevier.com/locate/ceramint https://doi.org/10.1016/j.ceramint.2023.12.287 Received 25 September 2023; Received in revised form 12 December 2023; Accepted 21 December 2023 Ceramics International 50 (2024) 9682–9696 9683 Wind is a competitive, clean energy source and its industry is booming. If the wind targets set out in REPowerEU (2022) [12], the EU’s energy security strategy, are successfully met, a massive 65 billion cubic meters of gas could be saved. Apart from reducing greenhouse gas emissions (GHG), this would also make Europe self-sufficient in terms of energy production and less reliant on imported fossil fuels. Wind is an increasingly stable form of power supply. New onshore wind farms now operate at capacity factors of 30–45 %, and new offshore wind farms at 50 % [13,14]. Wind power does not emit any CO 2 , SO x , NO x , or metal particles [15]. It emits ~95 % less CO 2 than gas-based electricity production and ~98 % less CO 2 than coal-based production [16] and consumes very little water. Wind meets 15 % of electricity demand in the EU as a whole and much more in certain countries: Denmark 44 %; Ireland 31 %; Portugal 26 %; Spain 24 %, and Germany 23 % [17–19]. The International Energy Agency (IEA [20]) expects wind to be the first source of power in Europe by 2027. The standard lifetime of a wind farm is around 20–25 years. When wind turbines reach the end of their life, project developers must take them down and restore the site to its pre-existing condition. According to WindEurope 2021 [21], the European Wind Energy Association (EWEA, 2013) [22] and Recycling International [23], around 14,000 blades may have to be dismantled across the continent in the next five years, which would create between 40,000 and 60,000 tonnes of waste. Wind-turbine blades are typically composed of several materials, including glass fibre, carbon fibre, carbon/glass (hybrid) fibre, reinforced polymer composites and foam. The exact combination of materials used in wind turbine blades may vary depending on the manufacturer and the wind turbine model [24]. Carbon fibre is the typical material used in the blades on the larger, higher-capacity turbines [25]. It is known for its high strength, stiffness and durability, making it an ideal material to support the load and stress experienced by blades during operation [26]. Another benefit is that it is lighter than other materials, so ensuring that the blades are lighter and more energy-efficient [27,28]. The use of carbon fibre also allows longer blades to be made, which increases the amount of energy that can be produced with each turn of the turbine [29]. Although carbon fibre is expensive, its use as the raw material in wind turbine blades can lead to increased efficiency and durability of the turbines, so reducing long-term maintenance and operating costs [30]. The wind industry has committed to recycling or recovering 100 % of decommissioned wind turbines by 2025 and work is currently underway to boost recycling through different technologies, e.g. mechanical recycling, which involves shredding materials so that they can be reused for example as filler material in building materials or plastics; thermal recycling, which entails incinerating the blades to break down the composites, producing energy in the process; or chemical recycling, which uses solvents and thermal processes to separate the resins from the fibres so that both materials can be reused [24,26,31]. In this regard, the construction industry can play an important role in the recycling of these blades, by promoting the production of new materials. Yazdanbakhsh et al. [32] investigated the addition of 5–10 % volume of slender elements from wind turbine blades (known as “needles”) to a concrete mix. Results revealed no negative effects on the steadiness or workability of the new concrete or on its tensile, compressive and flexural strength. This paper aims to offer the wind industry a new perspective on blade waste recycling by using this residue as an additive in the production of bricks. It could also provide an opportunity for the brick industry to test the feasibility of using carbon fibres in the manufacture of what could potentially be high-performance materials. Carbon fibre is very light, so transport costs would be lower than with traditional bricks, and its use in the brick production process would provide direct benefits for the environment by reducing the depletion of clayey soils, which are nonrenewable natural resources [33,34]. This paper breaks new ground in the brick industry sector, because although there have been studies that investigated the addition of carbon fibre to concrete for the construction of structural elements such as beams and columns [35–37], no specific research has been conducted on the reuse of carbon fibre from wind turbine blades in brick manufacture. With this in mind, the aim of this research is to produce a new type of brick which i) is environmentally friendly, ii) uses fewer raw materials, iii) contributes to the elimination of waste and iv) meets the relevant standards in terms of its physical-mechanical properties for use in the construction industry. 2. Materials and method 2.1. Geological context and supply of raw materials Bricks were made by mixing clayey materials from Viznar and Guadix (Granada, Spain). In geological terms, Viznar is part of the Granada basin, while Guadix belongs to the Guadix basin. Both basins are part of the central area of the Betic Cordillera. The clayey material from Guadix was deposited in the Middle-Late Pleistocene during the last stages of basin infilling, while the clay from Viznar was deposited during the late Turolian in a lacustrine environment with alluvial sediments (conglomerates and sands), deltaic sediments, small calcarenitic platforms and other minor materials from Sierra Nevada and Sierra Arana [38]. This work was carried out in partnership with the company Tesela, Materiales, Innovaci´ on y Patrimonio S.L. (Granada, Spain), who supplied carbon fibre powder (CF) from the blades of decommissioned wind turbines from wind-farms in Spain, and with the brick manufacturer Cer´ amica Castillo Siles (Granada, Spain), who supplied the clayey materials. 2.2. Traditional brick production In this study, the bricks were made using a mixture of clayey materials from Viznar and Guadix in proportions of 3/5 and 2/5 by weight, respectively. These proportions were chosen on the recommendation of the brick manufacturer who supplied the raw materials, because in his experience they produced the most workable mix and the best-quality end products. Fig. S1 shows the different steps in the brickmaking process. Several kilograms of the two clayey materials from Viznar and Guadix were removed from piles of clay in the brick factory using a shovel. All coarse rock fragments and plant roots were discarded. Once in the laboratory, the clayey materials were sieved and fragments of over 1.5 mm in size were removed. The two types of clay were then mixed together in the proportions mentioned above. With a view to using as much residue as possible without affecting the workability of the clay mixture, and as there was no previous bibliography to follow, it was decided to use 5 and 10 wt% of carbon fibre powder (CF). With higher amounts of carbon fibre, the clayey mass became less workable. The addition of carbon fibre means that less clay is required to produce the bricks (Table S1). The maximum amount of added CF was set at 10 wt%, given that as CF is a very lightweight material in relation to its volume, the addition of over 10 wt% CF could have prevented the residue from mixing properly with the clayey materials. The clay and CF mixtures were then mixed with water and moulded in moistened wooden moulds of 15 ×20 ×4 cm (Fig.S1 A and S1B). Table S1 shows the proportions of clayey material, carbon fibre powder (by weight percentage, wt.%) and water used in each mix design. A control group (CF0) of bricks with no added carbon fibre powder was also prepared. Moulds were removed after 1 h and the clayey pastes were cut into ~4 cm edge cubes using a stretched cotton thread and left to dry (Fig. S1C). Once the samples had dried, they were fired in a Herotec CR35 electric oven at 800, 950 and 1100 ◦C (Fig. S1C). These three temperatures were selected on the following basis. The intermediate temperature (950 ◦C) was chosen because it is one of the most commonly used in the brick industry (the brick manufacturer who provided the raw materials fires his bricks at this temperature). The other two were chosen to enable the analysis of the mineralogical, textural, physical and L. Crespo-L´ opez et al. Ceramics International 50 (2024) 9682–9696 9684 durability changes in the bricks with and without added CF over a range of 300 ◦C. Table S2 offers a summary of the samples analysed in this paper. The temperature inside the oven was initially kept constant for 1 h at 100 ◦C to eliminate any residual moisture in the samples. Then, the temperature was increased at a rate of 2 ◦C/min (heating). Once the desired temperature had been reached, the oven was kept at a constant temperature for 3 h (soaking). Finally, it was turned off and the samples were left to cool slowly (cooling). They were not removed from the oven until the next day. After removal from the oven, the bricks were immersed in water for about 1 h to prevent possible “lime blowing” due to the presence of lime grains [39]. 2.3. Analytical techniques 2.3.1. Chemistry, mineralogy and texture of the raw materials and fired bricks The granulometry of the raw material and of the CF was determined using a Galai CIS-1 laser gauge that uses laser diffraction to measure the range of the particles after 10 s sonication in water over a 0.02–1500 μ m range. The thermal decomposition of the clayey material up to 950 ◦C was carried out using a METTLER-TOLEDO TGA/DSC1 thermogravimetric analyser coupled with differential scanning calorimetry (TGDSC). About 20 mg of sample was deposited on an Al crucible and analysed in a flowing air atmosphere (50 ml/min) at a heating rate of 20 ◦C/min. The components of the carbon fibre in the IR spectrum were identified using a Fourier transform infrared spectroscopy with an attenuated total reflectance sample holder (ATR-FTIR), in a frequency range of 4000-400 cm −1 and with a step size of 0.50 cm −1 . X-ray fluorescence (XRF) was used to determine the major elements in the clayey material and in the CF with a PANalytical Zetium compact spectrometer with an Rh anode and a 4 kV X-ray generator. 5 g per sample were milled to powder in an agate mortar and then analysed. The Loss On Ignition (LOI) was determined by burning the samples at 1000 ◦C for 1 h. The mineralogical composition of the raw material and the fired bricks was determined by powder X-ray diffraction (PXRD) using a PANalytical X’Pert PRO diffractometer. The working conditions were as follows: CuK α radiation, 45 kV voltage, 40 mA current, 3–70◦2θ exploration range, 0.1 2θ s −1 goniometer speed. Mineral phase identification was first performed using the PANalytical X’pert Highscore Plus 3.0 software, matching the experimental diffraction peaks with those from the Joint Committee for Powder Diffraction Standards (JCPDS) PDF-2 database. Quantitative estimations were also performed by adding 10 wt% internal standard α -Al 2 O 3 to the powder samples and detecting the peaks using open-source Profex-BGMN software, which has dedicated fitting functions [40–43] for Rietveld refinements [44]. The goodness-of-fit of the refinements (GoF) were between 1.68 and 1.87. The petrographic features of the fired bricks were observed by means of polarized optical microscopy (POM). Observations under planeand cross-polarized light were carried out on polished thin sections using a Carl Zeiss Jenapol-U microscope equipped with a Nikon D7000 digital camera. Detailed observations of the morphology of the carbon fibres and of the texture, pore morphology and degree of vitrification of the fired bricks were made using a high-resolution field emission scanning electron microscope (FESEM) Carl Zeiss SMT (AURIGA series) coupled with energy dispersive X-ray analysis (EDS). The carbon fibres and brick fragments were carbon-coated prior to their observation under FESEM. 2.3.2. The pore system of the bricks The pore system of the fired bricks was investigated using hydric and porosimetric tests (Table S3). To this end, free (Ab, at atmospheric pressure [45]) and forced (Af, under vacuum) water absorption and drying tests (Di, [46]) were carried out (according to the relevant standards, see Table S3). These tests enabled us to determine the degree of pore interconnectivity (Ax), the capillary rise (C) [47], the saturation coefficient (S) [48], the apparent ( ρ a ) and real ( ρ r ) densities and the open porosity (Po) (Equations 1-9, Table S3) [49,50]. Hydric tests were performed under controlled thermo-hygrometric conditions (20 ◦C and 60 % RH) using deionized water. Three samples per brick (~4 cm edge cube) group were analysed. The pore system of the bricks within a range of 0.002–200 μ m was analysed by mercury intrusion porosimetry (MIP) using a Micromeritics Autopore V 9600 porosimeter with a mercury-brick contact angle of 130◦. Open porosity (Po MIP ) (Equations 6, Table S3) [49,50] and specific surface area (SSA) were calculated. 2.3.3. Durability of the bricks Fifteen salt crystallization cycles were performed to assess a theoretical degradation that could affect the lifetime of the bricks according to the UNE-EN 12370 standard [51]. This test reproduces the decay that the bricks may undergo due to the dissolution and recrystallization of soluble salts within their porous systems. Three samples per brick type were used for this test. 2.3.4. Mechanical behaviour (Cs) In the analysis of compressive strength (Cs), three samples measuring 1.5 ×1.5 ×6 cm were tested for each brick type (Equation 10, Table S3) [52]. The compressive strength of these samples was measured using a Controls Uniframe T1192 electromechanical universal tester, equipped with 100 kN and 25 kN load cells and a loading rate 0.6 of mm/min. This device was specially designed for measuring the compressive strength of small bricks. 2.3.5. The compactness of the bricks The ultrasound propagation velocity of compressional (V p ) and shear (V s ) waves was measured with a Panametrics HV Pulser/Receiver 5058 PR coupled with a Tektronix TDS 3012B oscilloscope. It was measured in accordance with the ASTM D2845 [53] (Equations 11, Table S3) standard for dry test samples using transducers of 1 MHz. These data were used to obtain information on the compactness of the bricks. A viscoelastic gel was applied to ensure good coupling between the transducers and the brick samples. V p and V s were measured in the three perpendicular directions on cubic samples. Once the ultrasound velocities had been determined, total anisotropy (ΔM), the Poisson coefficient ( ν ) [54, 55], and the Young (E) [54,55], Shear (G) [56] and Bulk (K) [54] moduli were calculated (Equations 12–15, Table S3). The Leeb Hardness (LH) tester PCE-2500 N was used to measure the surface hardness of the bricks using a rebound hammer. Ten linear measurements were made from one edge of each brick to the opposite side passing through the centre. The harder the surface of the material, the higher the rebound velocity [57]. 2.3.6. Colour Colour measurements were performed to quantify the lightness (L*) and chromatic coordinates (a* and b*) of the fired bricks [58]. A Konica Minolta CM-700d spectrophotometer was used. Illuminant D65, 10◦ observer angle and 8 mm measurement area were used. Nine measurements per sample were performed. The total colour variation (ΔE) between the control bricks and those with added CF was calculated (Equation 16, Table S3). 3. Results and discussion 3.1. Grain size distribution, thermogravimetry and chemistry of the clayey material and CF The granulometric analysis shows that the clayey material has a unimodal particle size distribution with a maximum peak at 23.2 μ m (Fig. 1A). As expected, the carbon fibre (CF) curve has a shorter size distribution compared to the clayey material. This curve is bimodal, although there is a dominant maximum peak at 14.1 μ m and a smaller one at 102.5 μ m. Thermal analysis of the clayey material (Fig. 1B) shows a weight loss of 5 % at 86 ◦C due to the loss of hygroscopic water [59]. At L. Crespo-L´ opez et al. Ceramics International 50 (2024) 9682–9696 9685 297 ◦C there is an inflection in the curve corresponding to the combustion of organic matter [60,61]. At ~500 ◦C, a gradual dehydroxylation of the phyllosilicates begins to occur, which continues up to 711 ◦C [62,63]. Between 711 ◦C and 800 ◦C, the main weight loss of around 13 % occurs. This is linked to the decomposition of carbonates and the release of CO 2 [62,64,65]. The clayey material starts losing weight again at ~840 ◦C, probably due to further dehydroxylation of the phyllosilicates [66–68]. Finally, an exothermic peak is observed at 925 ◦C due to the formation of new crystalline phases, which will later be identified using PXRD analysis. The IR spectrum obtained from the carbon fibre was compared with other spectra reported in the literature. Table 1 shows a summary of the bands identified, their attribution and references. The IR spectrum revealed that at some earlier stage the carbon fibre had been carbonised at a temperature of approximately 850 ◦C [69–71], probably in order to make the blades stronger and more durable. Fig. 2 illustrates the infrared spectra of the carbon fibres, where the larger and thinner bands in the range 1300–1800 cmˉ 1 show a chaotic trend due to the presence of several compounds together with the carbon fibre. A peak at 1730 cm −1 was evidence of the ester (arachidyl dodecanoate) functional group [72]. Ketone and carboxylic acid C═O peaks were centred at 1705 cm −1 [73]. The peak of benzoic acid was present at a lower wavenumber, 1640 cm −1 , probably because of the aromaticity of the compound [74]. This peak is quite high due to the carbonisation detected in the sample. Peaks at 2389 and 2667 cm −1 are related to C–H stretching of the monomer units [75]. Other compounds detected in the FC sample were: free water (H 2 O: 450–600 cm −1 ; 1250–2000 cm −1 ; 3500–4000 cm −1 ) [76], carbon dioxide (CO 2 : 3570–3750 cm −1 ; 2298 cm −1 ; 631–779 cm −1 ), carbon monoxide (CO: 2000–2100 cm −1 ) [80], carbonyl sulphide (COS: double peaks at 2100 and 2051 cm −1 ), ammonia (NH 3 : 965 cm −1 ) [77] and sulphur dioxide (SO 2 : 1300 cm −1 ) [76]. The degradation of thermoplastic products (mainly the peak of ε -caprolactam [79] at 1713 cm −1 ) and the release of phenol at 1510 cm −1 [78] were also observed. In relation to these compounds detected by FTIR, Cusid´ o et al. [80] and Cremades et al. [78] studied the hazardousness of the gases emitted during the firing of ceramics made with additives, recommending the use of technologies that minimize the emission of pollutants that are dangerous for human health. In this case, the addition of 5–10 wt% carbon fibre would increase CO 2 emissions during the firing process, as explained below. Table 2 shows the chemistry of the clayey material and the carbon fibre (CF). The raw material is rich in SiO 2 and has a high Al 2 O 3 content. The presence of CaO (9.23 %) and MgO (2.82 %) is a sign of its carbonate content. The carbon fibre is composed almost exclusively of carbon, which is consumed during calcination of the sample, so explaining the very high loss on ignition (LOI) of over 99 %. The calcination of carbon fibres results in the emission of CO 2 into the atmosphere, an issue that must be taken into account if this residue is to be used in the brick industry, as these factories are subject to carbon tax and CO 2 emission restrictions. Furthermore, the resin residues associated with the fibres will increase the total organic carbon in the flue gas [80]. However, the use of carbon fibres as an additive in brick production is more environmentally friendly in terms of air pollution than other possible methods for dealing with decommissioned turbine blades such us incineration, pyrolysis, or chemical recycling [81–83]. Came´ an 2011 [84] and Rodríguez Vazquez [85] have shown that if metals such as Fe, Ti and Al are present, even in small amounts, in the clayey material, they can promote the graphitization of carbon fibres during the firing process. 3.2. Mineralogy of the clayey material and fired bricks Rietveld refinement of PXRD analysis reveals that the clay mixture is rich in quartz (41 wt%) (Table 2), which confirms the high SiO 2 content detected by XRF (see Table 3). Other phases detected include carbonates Fig. 1. Grain size distribution of the raw materials (clayey material and carbon fibre, CF) (A) and TG-DSC analysis of the clayey material (B). A) Particle size (in μ m) vs. volume (in %). B) Temperature (in ◦C) vs. weight loss (right ordinate, in %) and differential scanning calorimetry (left ordinate, in mW). Table 1 FTIR bands identified for CF, their attribution and references. Bands identified (cm⁻ 1 ) Attribution References 450–600 free water [76] 965 ammonia [77] 1250–2000 free water [76] 1300 sulphur dioxide [76] 1510 phenol [78] 1713 ε - caprolactam [79] 1730 ester [72] 1705 ketone and carboxylic acid [73] 1640 benzoic acid [74] 2000–2100 carbon monoxide [80] 2100–2051 carbonyl sulphide [77] 2298 carbon dioxide [80] 2389–2667 monomer [51] 3500–4000 free water [76] 3570–3750 carbon dioxide [80] L. Crespo-L´ opez et al. Ceramics International 50 (2024) 9682–9696 9686 (calcite and dolomite, 11 wt%), K-feldspar (microcline 8 wt%), hematite (5 wt%) and phyllosilicates (smectite, illite, paragonite, kaolinite and chlorite, 21 wt%). The CaO and MgO content analysed by XRF (12 %, Table 2) is in accordance with the carbonate content determined by PXRD. Paragonite is a common phyllosilicate in the inner areas of the Betic Cordillera and is found in the raw material from Guadix [86,87]. For its part, the clay from Viznar is rich in carbonates from the different geological formations in the Granada Depression [88]. The amorphous phase content in the clay mixture is 13 wt%. Table 3 shows the changes in the mineralogy of the bricks after firing. These changes are more obvious at 1100 ◦C with the crystallization of high-temperature silicates, the increase in background noise due to the vitrification of the samples [67] and more extensive development of a graphite-type crystalline structure in the samples with added CF. The amorphous phase content (in wt.%) increases in line with the firing temperature and with the percentage of CF, reaching its highest value in the samples fired at 1100 ◦C and with 10 wt% CF. At the three firing temperatures studied (800, 950 and 1100 ◦C), quartz is the most abundant phase (ranging from 41 to 20 wt%), while hematite and feldspars s.l. are detected in all the samples. At 800 ◦C, all the Fig. 2. Infrared spectrum of the carbon fibre used in this study. The range of the spectrum from 2000 to 1000 cm −1 has been marked to highlight the presence of other components in the carbon fibre sample. Table 2 Chemical analysis of mayor oxides (in wt.%) in the raw material and CF. SiO₂ Al₂O₃ Fe₂O₃ MnO MgO CaO Na₂O K₂O TiO₂ P₂O₅ LOI Clayey material 48.78 15.97 5.69 0.07 2.80 9.43 1.27 2.70 0.83 0.16 12.40 CF 0.47 0.08 0.03 – 0.02 0.32 0.02 0.03 – 0.03 99.13 Table 3 Mineralogical characterization by PXRD of the clayey material and the fired bricks with or without added CF. Data is provided in weight %. Abbreviations of minerals according to Warr 2020 [89]: Sme: smectite; Ilt: illite; Pg: paragonite; Kln: kaolinite; Chl: chlorite; Mc: microcline; Qz: quartz; Cal: calcite; Dol: dolomite; Hem: hematite; Or: orthoclase; Sa: sanidine; Gh: gehlenite; An: anorthite; Mul: mullite; Wo: wollastonite; Di: diopside; Amph: amorphous phase. Sme Ilt Pg Kln Chl Mc Qz Cal Dol Hem Or Sa Gh An Mull Wo Di Amph ERROR Clay mat. 3 12 2 2 1 8 41 6 5 5 – – – – – – – 13 ±5 CF0/800 – 19 – – – – 40 4 – 4 4 – – – – – – 30 ±2 CF5/800 – 18 – – – – 39 4 – 2 2 – – – – – – 35 ±3 CF10/800 – 16 – – – – 32 5 – 1 3 – – – – – – 42 ±4 CF0/950 16 – – – – 31 – – 5 – 1 7 12 – – – 30 ±2 CF5/950 – 10 – – – – 32 – – 2 – 2 3 14 – – – 37 ±2 CF10/950 – 14 – – – – 20 – – 1 – 1 7 11 – – – 47 ±3 CF0/1100 – – – – – – 26 – – 5 – 2 4 19 3 2 9 33 ±1 CF5/1100 – – – – – – 23 – – 1 – 2 2 22 3 2 10 41 ±2 CF10/1100 – – – – – – 24 – – 0 – 3 2 22 4 3 12 48 ±3 L. Crespo-L´ opez et al. Ceramics International 50 (2024) 9682–9696 9687 phyllosilicates have disappeared with the exception of a dehydroxylated illite. The fired bricks have a higher illite content than the raw unfired bricks. This seems to be due to an overestimation of this phase due to the decomposition of the other phyllosilicates and their absence in the phase count (they accounted for 8 %, Table 3). Illite can still be detected at 950 ◦C, although the reflection (001) at 10 Å is lower than that observed in the clay mixture, suggesting a lower concentration of this phase. Dehydroxylation can influence the properties of the surrounding material, such as the strength and porosity of the fired brick [90]. Calcite is detected in very small amounts at 800 ◦C (~4 wt%) and disappears at higher temperatures, while dolomite has already disappeared at 800 ◦C (Table 2). Microcline detected in the clayey material converted into its polymorphs, i.e. orthoclase at 800 ◦C and sanidine at 950 ◦C and 1100 ◦C [91]. The hematite concentration decreases with increasing firing temperature in the samples made with added CF, reaching values of <1 wt% in CF10/1100 as compared to 5 wt% in CF0/1100 (Table 2). The decomposition of carbonates and their reactions with quartz and other silicates leads to the appearance of new Ca- (and Mg-) silicates such as gehlenite at 950 ◦C and wollastonite, anorthite and diopside at 1100 ◦C. Another new phase, mullite, is identified at 1100 ◦C. The appearance of these phases is manifested in the exothermic peak observed in TG-DSC analysis at 925 ◦C (Fig. 1B). 3.3. Texture and microtexture of the fired bricks The observations under MOP of the bricks made with and without carbon fibre revealed the presence of fragments of mica-schists, gneiss and quartz grains with undulose extinction with a maximum length of about 1 mm. At 800 ◦C these fragments are homogeneously distributed in an orange matrix (Fig. 3A) forming the temper of the bricks. With the addition of carbon fibre, the matrix becomes browner in colour and disoriented carbon fibres are observed as part of the matrix (Fig. 3B and C). These fibres probably encourage the development of a reducing environment in the oven and the formation of “black core”. At 800 ◦C, other minerals can also be observed. The carbonate grains are partially decomposed and have lost their typical high interference colour (Fig. 3B). Phyllosilicates appear unaltered and often have a preferential orientation due to the pressure exerted during kneading. Pores are mainly elongated and follow the same orientation as planar minerals. At 950 ◦C, the matrix is slightly darker (Fig. 3D) due to the gradual vitrification and even more so in the samples with added CF (Fig. 3E and F). The phyllosilicates become less birefringent, and carbonates are totally decomposed. In the bricks with carbon fibre (Fig. 3E and F), it seems that the fibres start to form clumps and orient themselves slightly. Many of the fibres were severely affected, and some even disappeared, leaving imprints in their place. At 1100 ◦C, due to the high firing temperature, the pores become ellipsoidal-to-rounded and the matrix becomes dark (Fig. 3G, H and I). This change is more noticeable in the samples made Fig. 3. Optical microscopy images of bricks with or without added carbon fibres fired at 800, 950 and 1100 ◦C, as observed in plane-polarized light. A) orange matrix of sample FC0/800; B) and C) brown matrix of the samples fired at 800 ◦C with added CF (CF5/800 and CF10/800, respectively). D) phyllosilicate grains with a preferential orientation in the FC0/950 sample; E) and F) the matrix is slightly darker due to the gradual vitrification and the presence of imprints of CF; G) dark matrix in the CF0/1100 sample; H) and I) rounded pores in the samples fired at 1100 ◦C with added CF (CF5/1100 and CF10/1100). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) L. Crespo-L´ opez et al. Ceramics International 50 (2024) 9682–9696 9688 with CF. This is probably because of the combustion of fibres (remember the high LOI of CF weighed at 1000 ◦C in Table 1), which brings more heat into the matrix. In these samples, the development of sandwiched structures, i.e. black cores, is more evident. The phyllosilicates have lost their birefringence and have a whitish interference colour. According to Pask & Tomsia [92] and Rodríguez Navarro et al. [93], the kaolinite and illite have probably been replaced by mullite, as also suggested by the PXRD results (Table 3). Fig. 4 offers a schematic image of the distribution and orientation of the imprints of the carbon fibres and pores in the matrix of sample CF10/ 1100. Three concentric areas (A, B and C) can be distinguished. In the outer area of the sample (A zone), which has a more orangey colour, a small amount of randomly arranged imprints can be observed under the microscope, together with small, rounded pores. In the middle area (B zone), the imprints are shorter and thicker than those observed in the outer area and tend to orient themselves. In the core of the sample (C zone), the imprints remain thick but increase in length forming oriented aggregates in bundles. The additional heat provided by the combustion of the fibres is probably responsible for these textural changes. According to Came´ an [85], it is possible that carbon fibres are oriented by the pressure applied during the preparation of handmade bricks prior to firing, an effect that becomes more intense as more carbon fibres are added. The FESEM observations were made on detached particles of carbon fibre before it was kneaded with the clayey material (Fig. 5) and on fragments of the bricks made with 0 wt% and 10 wt% CF fired at 800 ◦C (Fig. 6). The aim was to highlight the textural differences in the brick matrix caused by the addition of the highest amount of CF. Although this question was not analysed in our study, previous researchers argued that the microstructural changes resulting from the addition of CF could have improved the thermal properties of bricks, and in particular their thermal conductivity [94]. Fig. 6 shows a general view of the carbon fibres used to make the bricks. As expected, this waste is not only composed of elongated fibres but also contains rough and irregular aggregates (Fig. 5A–C). EDS analysis confirms the presence of C and O in the carbon fibres. These aggregates probably come from the reinforcing foam used in the construction of the blades [95–97]. They are very sensitive to increases in temperature, forming pores in the fired bricks. The carbon fibres range in length from 5 to 90 μ m and have a diameter of around 6–9 μ m. The surface of the fibres appears striated (Fig. 5D). The addition of carbon fibres to the clayey mix results in a partial change in the texture of the fired bricks and the matrix becomes quite heterogeneous (Fig. 6A, CF0/800 and Fig. 6C, CF10/800). In both groups of bricks (CF0 and CF10), the phyllosilicate grains maintain their lamellar habit, although the lamellae tend to separate along their basal plane due to dehydroxylation (Fig. 6B) [98]. Carbonate grains are Fig. 4. Left – Optical microscopy images in plane-polarized light. The fibres and their imprints are highlighted in yellow. Right – schematic view of a thin section of the CF10/1100 brick. Moving inwards from the outside to the core of the brick, three quite distinct zones can be observed: Zone A, the outermost zone, corresponds to MOP image A, zone B to image B and zone C, the innermost area, to image C. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) L. Crespo-L´ opez et al. Ceramics International 50 (2024) 9682–9696 9689 partially decomposed, confirming MOP observation and PXRD data. The carbon fibres in the fired bricks, when visible, show a change in their morphology compared to the unfired fibres (Fig. 6D), as a result of firing. The striae observed on the unfired fibres (Fig. 5D) have disappeared or thinned out. It is also possible to corroborate the changes in the size of the fibres noted during MOP observations. At 800 ◦C they have already started to thicken and shorten and by 1100 ◦C they are completely calcined. At 800 ◦C some fibres have disappeared due to combustion, leaving imprints in the matrix of the brick (Fig. 6D), as already observed under POM. At the same time, greater porosity and irregular-shaped pores can be distinguished in the bricks made with CF. This may be linked to the burning of carbon fibres and reinforcing foam observed under FESEM (Fig. 5B). The orientation of the CF occurs because they are oxygen-rich organic compounds that develop a structure with aliphatic crosslinks that provide great rigidity to the carbonaceous structure [99,100]. This structure is stable at 25 ◦C and 35 % HR (Fig. 7). When the temperature augments, the carbon fibre changes from a state of carbonisation (around 800 ◦C) to one of graphitization (over 900 ◦C) (Fig. 7), in which there are far fewer crosslinks. This allows the fibres to reorganise themselves and acquire a graphitic structure as the temperature increases [85]. Fig. 5. FESEM secondary electron images of carbon fibres added to the clayey material to make the bricks. A) General view of carbon fibres and other particulate matter from wind turbines; B) unidentified aggregates with irregular morphology together with fibres; C) detailed image of the cylindrical morphology of the carbon fibres; D) detail of the surface morphology of a carbon fibre highlighting the presence of striae on the surface. Fig. 6. FESEM secondary electron images.: A) General view of the texture and porosity of the sample CF0/800; B) dehydroxylation of a phyllosilicate grain along (001) planes in CF0/800 (the EDS of the phyllosilicate can be seen in the inset); C) general view of the texture and porosity of the sample CF10/800; D) imprints left by carbon fibres in sample CF10/800. L. Crespo-L´ opez et al. Ceramics International 50 (2024) 9682–9696 9690 Came´ an Martinez [84] and Rodríguez V´ azquez [85] demonstrated that temperatures close to 1800 ◦C are required for complete graphitization of the CF, although this process can be accelerated by the presence of resins and clays (as in this study), leading to further fibre orientation at 1100 ◦C as compared to 950 ◦C (Fig. 7). 3 4. Hydric behaviour and pore system of fired bricks The bricks fired at 800 and 950 ◦C show an increase in water absorption (Ab and Af) and open porosity (Po) values with the addition of carbon fibre, changes that can also be observed in line with increases in the firing temperature (Table 4). CF0 samples showed lower values than CF5 and CF10 samples. This indicates that at low firing temperatures, the presence and amount of carbon fibre have a greater influence on the hydric behaviour of the fired pieces than the firing temperature does. The opposite trend can be observed at 1100 ◦C, where the addition of carbon fibre seems to hinder water absorption. The addition of carbon fibre has a clear impact on the circulation of water in the pores and capillaries of the bricks. The bricks with no added CF achieve the best Ax values at all firing temperatures, although these values decline with the addition of CF and with increasing CF content (Table 4). Ax also worsens with the increase in the firing temperature, reaching the worst values with FC10/1100. Bricks reach high saturation indices (S, Table 4), except at 1100 ◦C when the extended vitrification of the ceramic mass makes them less absorbent [67]. As for drying, bricks fired at 800 ◦C dry faster than those fired at 950 ◦C, and these in turn dry faster than those fired at 1100 ◦C (Di, Table 4). This is logical given that pore interconnection worsens as firing temperature increases and the water finds it more difficult to move around the pore network. As regards porosity, two trends can be recognised: Po increases between 800 and 950 ◦C, but falls at 1100 ◦C (Table 4). In all cases, the presence and percentage of CF lead to higher porosity because of the imprints left by CF in the matrix of the bricks. The different behaviour at the different firing temperatures can be explained by the change in size of the carbon fibres, a process that began at 800 ◦C, as seen in POM and FESEM observations. This could have encouraged the development of small fissures around the fibres. Although at 1100 ◦C the carbon fibres underwent the same change as at 950 ◦C, the vitrification of the matrix was so pronounced that it hampered the development of fissures. Water absorption by capillarity tends to increase with the addition of CF and with the rise in the firing temperature (C, Table 4). No clear trend can be observed with densities. Both, ρ a and ρ r are lower when CF is added, and their values fall further as higher amounts of CF are added. This is logical given the lower density of carbon fibres, as found in previous research (1.75 g cm −3 ), compared to the silicate compounds present in the bricks. The MIP analysis confirms the results obtained by HT (Table 4) and shows how the addition of CF modifies the pore system of the bricks in terms of open porosity and pore size distribution. All bricks display a unimodal pore size distribution with a maximum peak at around 0.1 μ m (Fig. 8). The only exception is CF10/1100 which has a bimodal distribution, with a main peak between 0.1 and 1 μ m and a second family of pores at around 10 μ m (Fig. 8). When comparing the different types of bricks, the maximum peak always shifts towards the right of each diagram (i.e. towards larger pores), as the firing temperature increases. This causes a reduction in the Fig. 7. Influence of temperature on the reorientation of carbon fibres. Table 4 Hydric parameters of bricks: Ab: free water absorption (%); Af: forced water absorption (%); Ax: degree of pore interconnection (%); S: saturation coefficient (%); Di: drying index; Po: open porosity (%); C: capillarity coefficient (g/m 2 s 0.5 ); ρ a : apparent density (g cm −3 ); ρ r : real density (g cm −3 ); MIP: Po MIP : open porosity (%, determined by MIP); SSA: specific surface area (m 2 g −1 , determined by MIP). The standard deviation of each result is indicated in brackets. Hydric Test (HT) MIP Ab Af Ax S Di Po C ρ a ρ r SSA Po CF0/800 18.30 18.96 0.21 91.73 0.92 38.46 2.09 1.84 2.93 9.65 37.23 (4.34) (5.37) (1.11) (0.89) (0.003) (0.66) (1.34) (0.02) (0.03) (0.07) (1.12) CF5/800 26.63 26.79 1.46 93.30 0.90 39.31 2.47 1.71 2.88 2.46 38.47 (4.95) (3.96) (3.14) (2.15) (0.001) (1.54) (2.21) (0.01) (0.05) (0.01) (0.93) CF10/800 32.45 32.74 19.55 92.21 0.89 40.14 2.84 1.64 2.61 2.57 41.11 (5.84) (0.67) (3.22) (1.55) (0.012) (3.45) (3.44) (0.03) (0.05) (0.11) (1.25) CF0/950 21.46 21.97 3.87 91.08 0.89 41.89 2.05 1.73 2.74 8.37 43.89 (4.27) (5.90) (2.65) (2.31) (0.004) (2.01) (0.96) (0.01) (0.02) (1.45) (1.32) CF5/950 29.87 30.09 15.79 91.74 0.90 44.96 2.38 1.56 2.48 2.21 45.11 (5.33) (3.89) (5.11) (0.94) (0.009) (3.65) (3.69) (0.03) (0.02) (1.88) (2.54) CF10/950 36.22 36.69 27.27 92.64 0.90 46.82 2.64 1.48 2.22 1.64 46.38 (6.45) (5.71) (3.86) (2.64) (0.011) (4.09) (5.96) (0.03) (0.01) (1.13) (2.99) CF0/1100 24.66 24.81 3.61 89.65 0.88 37.12 1.99 1.59 2.61 7.20 39.67 (3.17) (0.56) (3.14) (0.99) (0.007) (1.34) (3.74) (0.02) (0.03) (2.74) (0.89) CF5/1100 15.48 15.86 30.59 82.03 0.88 39.94 1.69 1.42 2.54 1.38 40.97 (3.69) (0.43) (5.67) (3.44) (0.009) (4.61) (5.21) (0.04) (0.03) (1.71) (3.45) CF10/1100 18.01 18.44 37.34 84.40 0.87 42.97 1.82 1.25 2.36 0.74 42.27 (3.70) (2.52) (5.31) (2.89) (0.009) (6.33) (6.12) (0.04) (0.02) (0.09) (4.65) L. Crespo-L´ opez et al.