Portland clinker with civil construction waste: influence of pellet geometry on the formation of crystalline phases

Abstract Efforts to reduce greenhouse gas emissions in the context of sustainable development have intensified, with the development of research aimed at the production of new materials and binders for construction. This article analyzes the influence of pellet geometry in the production of clinkers, with the incorporation of construction waste (CCW). Procedures adapted from the method proposed by Brazilian Portland Cement Association were adopted in studies of laboratory clinkers, in an attempt to simulate the stages of the industrial process. Pellets were prepared with the same formulation, however, with four different geometries: spherical, with diameters of 1 cm, 2 cm and 3 cm, with manual molding, and semi-spherical, with a diameter of 2 cm, using molds of PLA (polylactic acid) printed on a 3D printer to facilitate the molding of the clinkers in a standardized way. Clinkers were characterized mineralogically by x-ray diffraction (XRD) and the Rietveld method was used to quantify the phases. Variations in the quantities of the alite and belite phases were observed depending on the geometry of the pellets, although the same calcination conditions were used. This is probably due to the variation in the surface area (exposure area) and the gradients of the cooling rate.


Introduction
In recent years, there has been a global effort to find solutions to reduce greenhouse gas emissions exploiting natural resources and inadequate waste disposal, with the development of new materials and binders for construction. In Brazil, in 2018, approximately 43.3 million tons of solid urban waste (59.5% of the total collected) was disposed of in landfills (ASSOCIAÇÃO…, 2020). The rest (40.5%) was dumped in inappropriate places, such as dumps or controlled landfills, with a high potential for environmental pollution and negative impacts on health (ASSOCIAÇÃO…, 2020).
Civil construction waste (CCW), which represents about 56% of urban solid waste (ASSOCIAÇÃO…, 2020), offers many opportunities for reuse. Several studies have reported positive results indicating the technical feasibility of using construction waste in various ways, mainly to replace natural aggregates (GUO et al., 2018;BUI;SATOMI;TAKAHASHI, 2018;AMORIM JÚNIOR et al., 2019;BOGAS;CARRIÇO;PEREIRA, 2019) and as a complementary cementitious material CHOI, 2012;ASENSIO et al., 2016;LETELIER et al., 2017;CANTERO et al., 2019) in concretes and mortars. These wastes have been tested in the manufacture of cement, replacing part of the conventional raw materials (limestone and clay) of Portland clinker (GALBENIS; TSIMAS, 2006;PUERTAS et al., 2008;SCHOON et al., 2015;TAE;KIM, 2018)  Thus, in the various proposals for alternatives, the incorporation of industrial wastes and/or by-products in the cement industry stands out as promising to minimize environmental impacts, such as reducing CO2 generation, energy consumption and extraction of natural material.
Portland cement is a hydraulic binder resulting from the grinding of Portland clinker, product of the calcination of a mixture of limestone (CaCO3) and clay, in a rotary kiln, with a temperature of around 1450 ºC, and with the addition of calcium sulfate to regulate the time beginning of hydration of cement components (WINTER, 2012). Portland clinker contains four main minerals, also called crystalline phases: alite (3CaO.SiO2), belite (2CaO.SiO2), tricalcium aluminate (3CaO.Al2O3) and tetracalcium ferroaluminate (4CaO.Al2O3.Fe2O3), complemented with a variable number of minor phases (LUDWIG; ZHANG, 2015;SNELLINGS, 2016).
CCW appears as an alternative source of the main chemical elements of Portland clinker (CaO, SiO2, Al2O3 and Fe2O3), allowing the introduction of CaO in the raw mix formulation, partially replacing the usual raw materials and, thus, reducing CO2 emissions during clinker (GALBENIS; TSIMAS, 2006;PUERTAS et al., 2006;SCHOON et al., 2015). Thus, the modification of the CCW by beneficiation processes and / or its use as an integral part of the raw mix in the cement industry stands out as an interesting alternative, since the CCW undergoes transformations from the mineralogical point of view, during clinker, together with other raw materials, and gives rise to a cementitious product with properties similar to Portland clinker.
Previous studies have been developed aiming at the manufacture of Portland cement clinker using construction waste as an alternative raw material. It is observed that in the highlighted works, different conditions of composition of the CCW were considered, that is, in some studies only recycled concrete wastes were used (GALBENIS; TSIMAS, 2006;NOBRE;GUERREIRO;KIRCHHEIM, 2015), in other wastes of ceramic tiles (PUERTAS et al., 2006) and, still, wastes of ceramic tiles (PUERTAS et al., 2008). That is, the surveys adopted CCW selected according to their research interest. In this work, a different approach was given to the use of CCW when considering all parts of the constituents of class A civil construction waste of CONAMA Resolution No. 307 (CONSELHO…, 2002), through the use of its thin part.
Since cement is a material with high consumption worldwide and with projections of growing demand, the interest in the theme and the initiatives for studies and research is still very high (SCRIVENER; HOHN; SUI, 2018;REN et al., 2020). Estimates indicate that cement production globally returned to 4.1 gigaton in 2018, an increase of 1% after annual drops of 1% in the period 2014-2017 (UNITED…, 2019).
Thus, the use of solid industrial waste in the production of cement induces the development of research and laboratory experiments that require rigorous criteria, most of the time with techniques on a small scale, to simulate the stages of industrial production, that is, the use of experimental techniques with miniature samples. Beaudoin (2001) cites a number of advantages of adopting small-scale tests, such as: material savings, especially when material is unavailable, cost reduction with equipment and supplies, space and time savings required to obtain the results, safety for handling smaller samples, in addition to less waste generation with the tests. Also, test preparation and execution procedures may have fewer steps or be simpler.
Several researchers have developed clinker studies, using different experimental methods, in an attempt to simulate industrial processing (ANDRADE; KIHARA; MARINGOLO, 2003;COSTA et al., 2013;MARIANI et al., 2019). The production parameters used include the geometry and mass of the clinker pellets. Table 1 presents information about some research developed in the last two decades, using different wastes to produce clinker. It highlights the shape of the molding and the geometry of the pellets in the different studies. There is variability in the pellet production methods used, which makes it difficult to compare the different clinkers produced in different ways.
However, there is an absence of studies that evaluate how the geometric characteristics of the pellet (size and format) influence the formation of crystalline phases on the clinker, since the variation in the surface area and mass influences the efficiency of the process parameters, such as the firing rate and cooling rate gradient. In addition, many published studies do not present details about clinkers conformation, omitting important information (TREZZA; SCIAN, 2007;SIMÃO et al., 2018). This work analyzed the influence of the pellet geometry on the production of Portland clinkers with the incorporation of CCW, according to crystalline phases formation.

Materials
For the development of this research and production of clinkers, samples of CCW class A were used, that is, reusable or recyclable waste in the form of aggregates (CONSELHO…, 2002). These wastes, used in the present study as they are marketed, come from the Waste Recovery Unit Grajau S.A. (WRUG), located in the city of São Paulo (Brazil). Limestone and clay samples were also used, provided by a cement company located in the state of Sergipe.  (2015) Lab Concrete with ladle furnace slag and agate waste Manual Spherical ( = 1 -1.5 cm) Costa et al. (2013) Ladle furnace slag Manual Spherical ( = 1 cm), m = 3g Costa (2013) Aluminum anodizing Mariani et al. (2019) Waste generated by production of (a) (b) Figure 1 shows a scheme of the waste processing system and a partial view of the equipment at the plant. Upon receiving the CCW, WRUG performs a pre-screening, in order to remove unwanted materials, such as metals, plaster, paper and other materials of dimensions close to or greater than the hopper's feed limit. In this stage, predominantly wood and materials of large dimensions are removed, and subsequently stored in Roll-on/Roll-off buckets. After pre-screening, the CCW goes to a magnetic separation unit and a low-density Portland clinker with civil construction waste: influence of pellet geometry on the formation of crystalline phases 209 material removal unit using airflow, and the material is then poured into a primary hopper (AMORIM JÚNIOR et al., 2019).
CCW beneficiation process continues through a vibrating sieve, in which, due to particle size differences, the waste is separated into two streams, one for particles larger than 10 mm and the other for particles smaller than 10 mm, called scalp, containing predominantly soil in its composition. Particles with diameters greater than 10 mm are deposited directly in the primary crusher for processing. At that time, a powdery material, known as fine civil construction waste, is generated, which is separated from the rest of the material by means of a cyclone filter, taking it to storage bays.
The remaining material undergoes a new screening, in order to obtain the CCW with dimensions less than 40 mm. The material that is retained in this screening process is crushed and sieved again. Finally, the material with dimensions less than 40 mm, when falling on the third sieve, is separated into three fractions, of different dimensions: between 0 and 10 mm, 10 and 20 mm and 20 and 40 mm (SILVA, 2017).
In this research, two samples of fine fractions of the construction waste (CCW) from WRUG were used, one from the initial stage of waste handling, separated from the other fractions by the action of a cyclone filter, being identified as CCW-1 (powdery material), and another sample from the grinding of CCW grains passing through 40 mm sieves and retained in the 20 mm sieve, and identified as CCW20-40. After manual separation and weighing of the different constituents in a representative sample, the CCW20-40 fractions were composed by concrete (57%), mortar (34%), rock (7%), and ceramic (2%).

Preparation and characterization of raw materials
Four different lots of each type of CCW were homogenized and had samples selected by means of quarting, obtaining about 50 kg of material. Then, the grains were ground, using a jaw mill available at the Structures Laboratory of the Federal University of Bahia (Timoshenko) and subsequently in a horizontal ceramic ball mill, until the desired granulometry was reached, namely less than 75 μm. The limestone and clay samples were also ground into particles below 75 μm.
Raw materials were characterized physically, chemically and mineralogically. The physical characterization of the CCW included the evaluation of the specific mass (Micromeritics AccuPyc II 1340 helium pycnometer), the specific BET surface area (Gemini VII 2390a Micromeritics) and the particle size distribution (Mastersizer 2000 Malvern Panalytical laser granulometer, available at Construction Materials Laboratory of the State University of Feira de Santana).
The determination of the chemical composition of the CCW, clay and limestone samples were carried out using the x-ray fluorescence spectroscopy (XRF) technique with a Bruker spectrophotometer, model S2 Ranger, available at the Materials Characterization Laboratory of the Federal Institute of Bahia.
CCW, clay and limestone were characterized mineralogically by x-ray diffraction (XRD), using a D2 Phaser diffractometer (Bruker), with a copper target tube (Kα radiation with a wavelength of 1.54060 Å) of 30 kV and 10 mA, without filtering system with secondary monochromator. The diffraction spectra were obtained for 2θ between 5 ° and 90 °, using the scan in increments of 0.02 ° and collection time of 0.2 s / step. The phases present in the samples were identified with the aid of the DIFFRAC plus -EVA software, with a database centered on the COD system (Crystallography Open Database).

Dosing and Production of Clinkers
From the results of characterization of the samples, three dosages of clinker flours (raw mix) were carried out. Two clinker dosages incorporated CCW, replacing the clay and part of the limestone. Additionally, a clinker was dosed with traditional raw materials, without CCW, to be used as a reference.
The main procedures adopted for the production of clinkers were used by other researchers (ANDRADE; KIHARA; MARINGOLO, 2003;COSTA et al., 2013COSTA et al., , 2016MARIANI et al., 2019), according to the methodology proposed by the Brazilian Portland Cement Association (BPCA), in an attempt to simulate the stages of the industrial process. This involves: grinding and homogenization of raw materials, precalcination of raw mix, calcination of clinker and sudden cooling.
Raw mixes with different compositions were dosed, one containing CCW and the other reference, containing only clay and limestone, using the lime saturation factor (LSF) (Equation Eq. 3 LSF relates the concentration of CaO present in the clinker raw mix with the other compounds with which it reacts. Expressed in percentage, this indicator determines the maximum amount of calcium oxide that can react with the other constituent oxides as well as the content of the clinker alite (C3S) (WINTER, 2012). This property is very important, since the degree of clinker burning is indicated by the amount of CaO that did not react, that is, free lime (AY; ÜNAL, 2000). The increase in the amount of C3S improves the initial strength of the cement and, therefore, the LSF needs to be high enough to produce cement with a high C3S content (WALENTA; FÜLLMANN, 2004).
LSF value of 98 was adopted, according to recent research (RODRÍGUEZ et al., 2013;BURUBERRI;SEABRA;LABRINCHA, 2015;VILAPLANA et al., 2015;MARIANI et al., 2019), in order to obtain clinkers with higher content of alite and lower content of free lime and the incorporation of CCW, replacing all the clay and part of the limestone. With the adoption of fixed LSF, the values of AM and SM were calculated in each of the dosages.
The SM relates the SiO2 concentration to the sum of the Al2O3 and Fe2O3 concentrations, that is, it is the ratio of non-melting materials (silicates) and melting materials. The increase in SM generates an increase in the proportion of C2S and C3S and a decrease in the amount of C3A and C4AF, while low SM provides an increase in the liquid phase, benefiting the burning (FAURE et al., 2019).
AM represents the relationship between the concentration of Al2O3 and Fe2O3, that is, it is the relationship between the two main fluxing materials. The low AM makes the liquid phase less viscous, facilitates reaction kinetics and improves clinker granulation. On the other hand, the high AM increases the proportion of C3A and the viscosity of the liquid phase, hindering the kinetics of chemical reactions (KIM; TAE; KIM, 2018).
Pellets were molded in four different configurations, however, with the same dosage of the raw materials and with the addition of 20% (by mass) of water, in order to facilitate the molding of these spheres. These are: (a) spherical pellet with 1 cm in diameter; (b) spherical pellet with 2 cm in diameter; (c) spherical pellet with 3 cm in diameter; and (d) semi-spherical pellet with 2 cm in diameter.
The spherical pellets with a diameter of 1 cm, 2 cm and 3 cm were molded manually, using a template with the dimensions printed on a sheet of paper as support, as shown in Figure 2a. For the molding of the semispherical pellets of 2 cm in diameter, PLA (polylactic acid) filament molds manufactured in a 3D printer were used ( Figure 2b).
The pellets were dried in a kiln at 100 ± 5 °C and placed in alumina crucibles, and taken for calcination at 1450 ºC in a laboratory kiln model LHT 02/17 LB, from the Nabertherm brand. Because of a motorized lifting table, clinkers are automatically exposed to sudden, forced cooling. This step is essential to maintain the mineralogical phases and guarantee the desired properties in the experimental clinker (WINTER, 2012).
The calcination routine of the clinker pellets is shown in Figure 3. The pellets were subjected to a heating rate of 5 °C/min, until reaching the temperature of 900 °C, remaining at this temperature for 30 minutes for the lime decarbonation. Continuing the second firing stage, with the same heating rate, the samples remained in the kiln until reaching the final temperature of 1450 ºC, remaining at that level for 15 minutes before being cooling.
In this research, 12 clinkers were produced, using 3 dosages (reference clinkers, clinkers with CCW-1 and clinkers with CCW20-40), and 4 pellets geometries (spherical pellets with diameter of 1 cm, 2 cm and 3 cm; and semi-spherical pellets with diameter of 2 cm). Each type of clinker was produced in triplicate and measurements of diameters and masses were carried out.
Portland clinker with civil construction waste: influence of pellet geometry on the formation of crystalline phases 211

Clinker evaluation
The experimental clinkers were evaluated for mineralogical composition (XRD), using the same diffractometer and conditions used in the characterization of raw materials, except for the amplitude of the diffraction spectra that, for clinkers, were obtained in the range of 2θ of 5° at 70°. The crystalline phases present in the clinkers were quantified with the TOPAS Academic software version 4.2, which is based on the Rietveld method and the CIF system (Crystallography Information File). Table 2 presents the crystallographic files used for the quantification of phases by the Rietveld method, also adopted by Mariani et al. (2019).

Physical characterization of raw materials
The values of density and specific surface area BET of the CCW, clay and limestone samples are shown in Table 3. It can be seen that the clay has a specific surface area BET higher than that of the wastes, however, the two CCW samples have a surface area greater than limestone, which can be beneficial at the time of clinker. This parameter indicates the fineness or a higher surface roughness of the material, which is directly proportional to the reactivity (GOMES et al., 2019). In addition, all the materials used have a high surface area and small D50 values, which shows the efficiency of the grinding process. This is so that problems such as low reactivity and formation of agglomerates of belite and free lime will probably not occur (PUERTAS et al., 2008).
The particle size distributions of the limestone, clay and CCW samples, obtained by laser granulometry, are shown in Figure 4. From the analysis of the granulometric distribution curves, it can be seen that the grains of both benefited samples from the CCW and the limestone and clay have an average diameter (D50) below 75 µm, as expected, which is important to facilitate the adequate reaction of these materials during the clinker process to produce Portland cement (PUERTAS et al., 2008).

Chemical characterization of raw materials
The chemical compositions of the CCW, besides clay and limestone, the materials traditionally used in the manufacture of Portland cement, were obtained by x-ray fluorescence (XRF), in the form of oxides, and losses on ignition, and the results are shown in Table 4. It is said that construction waste is rich in silicon, calcium, aluminum, and iron, besides other elements in a smaller proportion, such as magnesium, sulfur, sodium, potassium, and titanium.
CCW is a source of calcium oxide, containing 17.5 to 26.7% of CaO, and is used as an alternative raw material to limestone, traditionally used in flour, to produce clinker. This calcium oxide present in the CCW is an environmentally adequate source of calcium, without needing decarbonation process and consequent CO2 generation, such as occurs with limestone, responsible for the high loss on ignition (LOI about 37.5% to sample used). Limestone decarbonation (calcite decomposition) occurs at an approximate temperature of 900 °C, generating CO2 as a by-product (CaCO3  CaO + CO2). For each 1,000 kg of calcined calcite, 560 kg of CaO and 440 kg of CO2 are generated. Thus, the loss on ignition from high-purity limestone can reach 44% (BRANDŠTETR; HAVLICA; ODLER, 1997; ALI; SAIDUR; HOSSAIN, 2011).
CCW samples can also be used as a source of silica alternative to clay, as they have 28.8 to 42.4% SiO2. Thus, it is clear that this CCW is an important alternative source of calcium and silicon, and can replace part of the clay-limestone system, traditional raw materials used in the production of Portland cement.
CCW present about 2.8% SO3, a value ten times higher than that found in clay and limestone, added together. The effect of smaller components on clinker formation has been studied and cited as beneficial due to the ability of some of them to act as mineralizing agents (KOLOVOS et al., 2001;MARIANI et al., 2019). These components, such as fluorine, sulfur and titanium, can decrease the temperature of the first formation of the liquid phase and / or speed up the rate of reactions that occur in the solid state, in the liquid phase or in the liquid-solid interface (mineralizers), changing the viscosity and surface tension of the molten material and affecting crystal growth and morphology. The technological consequences are clear: changes in the reactivity and burning capacity of raw materials, formation of new phases in altered quantities, differentiation of hydraulic activity and the properties of the cements produced, opportunities for energy Portland clinker with civil construction waste: influence of pellet geometry on the formation of crystalline phases 213 conservation. However, there are limitations because of the restrictions related to the SO3 content in cement and its volatility, which can cause it to accumulate in the ovens and preheaters in the cement factories. Figure 5 shows the x-ray diffractograms of the CCW samples used. It is observed that clay is the main source of kaolinite [Si2Al2O5(OH)4], quartz (SiO2), and muscovite (Al3KSi3O12) and limestone, source of calcite (CaCO3), dolomite [CaMg(CO3)2], and quartz (SiO2).

Mineralogical characterization of raw materials
The quantitative assessment of the phases showed that the largest source of silica is clay, (42.5%,) with limestone comprising only 4% silica. Limestone is the main source of calcium, making up the mineral calcite, with approximately 92%, as expected and according to the results of the chemical composition, shown in Table 4.
In relation to the CCW, the various existing crystalline phases show its heterogeneous character, composed of hydrated cementitious materials (mortars and concrete), ceramic, and concrete blocks, ceramic tiles and tiles, clay minerals of soils, ornamental rocks (granites and marbles), aggregates, among other materials. The CCW used is composed mainly of quartz (SiO2) Table 5 shows the percentages of each raw material used for the manufacturing of the experimental clinkers with incorporating CCW and reference clinker (produced with limestone and clay). It is observed that it was possible to completely replace the clay with CCW, incorporating approximately 9.5 to 14.3% of CCW into the mixture in the raw flour (raw mix).      Intensity (a.u.) Portland clinker with civil construction waste: influence of pellet geometry on the formation of crystalline phases 215 incorporating the CCW into the clinkering process, reincorporating this waste into the production chain and offering an adequate and sustainable final destination alternative. In addition, the use of CCW implies less environmental impact, by decreasing the emission of carbon dioxide (CO2). However, since the CCW is heterogeneous, the content of incorporation of the waste in the raw mix may be higher, depending on its chemical constitution.

Dosing and production of Portland clinkers containing CCW
Besides the environmental benefits, incorporating CCW brings benefits to the clinkering process. Kolovos et al. (2001) researched the effect of smaller elements of the raw mix in the CaO -SiO2 -Al2O3 -Fe2O3 system and found that sulfur-containing compounds improve the reactivity of the mixture and speed up the CaO combination, especially during the final clinker stage. Another study developed by Raina and Janakiraman (1998) concluded that SO3 speeds up the decomposition of CaCO3 at a lower temperature, thus reducing heat consumption or fuel intake. The growth of C3S crystals is also favored to form a liquid phase at a lower temperature. SO3 can stabilize C3S, preventing it from becoming C2S and favoring the development of resistance in cements.
In the dosage incorporating the CCW in question, the amount of sulfur present in the waste is within the amount considered adequate for incorporation into the raw mix. The maximum content of incorporated CCW was 9.5%, which represents a content of 0.27% SO3, considering the chemical composition of the CCW shown in Table 4.
In Brazil, NBR 16697 (ABNT, 2018) establishes that the content of SO3 (% by mass), from clinker in soluble and insoluble form and calcium sulfate (gypsum) added to cement, must be less than 4.5% for all types of Portland cement.

Evaluation of Portland clinkers in four different molding configurations
Clinkers in the three different pellet configurations were evaluated for mineralogical composition in order to verify the formation of the expected crystalline phases. The diffractograms of the reference clinkers, produced with limestone and clay, and of clinkers with CCW incorporation are shown in Figures 6, 7 and 8. Intensity (a.u.)

C -FREE LIME P -PERICLASE
CL-CCW 20-40 (1 cm)  Intensity (a.u.)  Intensity (a.u.)  Intensity (a.u.) Portland clinker with civil construction waste: influence of pellet geometry on the formation of crystalline phases 217 It can be observed that all the clinkers produced with the four different pellet configurations presented the expected crystalline phases, mainly of alite and belite, phases that should be the majority. Regardless of the pellet configuration, the minor clinker phases, C3A and C4AF, were also formed, as can be seen in the diffractograms.
The mineralogical results show that all the clinkers produced with the three different pellet configurations presented the expected crystalline phases, mainly of alite and belite, must be majority. Regardless of the pellet configuration, the minor clinker phases, C3A and C4AF, were also formed, as can be seen in the diffractograms. The quantification of the identified phases is shown in Tables 6 and 7.  Note: *statistical indicator that provides the relationship between expected and calculated residues: GOF (goodness of fit) = RWP (weighted profile R factor) / Rexp (expected R factor).

218
From the quantification results of the formed phases, it appears that the alite phase was formed in greater quantity in all clinkers, regardless of the pellet configurations and the presence of the CCW, indicating that the wastes adequately replaced part of the limestone-clay system. For the production of Portland cement, the alite phase must be the major constituent of clinker, because among the other phases, this is the one that has the greatest influence on the hardening of cement, being responsible for the development of mechanical strengths in the early ages (YOUNG; YANG, 2019). In a research using granite rocks fines, waste generated in the production process of aggregates for civil construction to produce clinker, similar promising results were found, resulting in higher amounts of alite when compared to the reference clinker, when incorporating 1.2 to 5.5% of the waste (GOMES et al., 2019).
Among the four pellet configurations used, clinkers with spheres of 1 cm in diameter were the ones that had the highest content of alite, both in reference clinkers and in clinkers with CCW. The clinkers produced with semi-spherical pellets with a diameter of 2 cm, presented intermediate contents of alite and, as expected, the clinkers produced with spheres of 3 cm in diameter were the ones that formed the lowest contents of alite. In industry, the raw mix is burned in a rotating kiln, with the powder gradually being heated inside the kiln by rotating movement (CHATTERJEE, 2011) resulting in nodular material with varying diameters. The formation of nodules (nodulization) occurs through the union of solid particles by liquid and requires an adequate proportion of liquid, favored by the low viscosity and high surface tension of the liquid and the small particle size of the solids. Alite formation and loss of volatiles affect nodulization, and the compaction that the latter entails must affect the kinetics of the former processes. Therefore, alite formation occurs in a changing environment inside the furnace (TAYLOR, 1997). As already reported, in this laboratory research using a static electric kiln, the previously shaped pellets are calcined and the clinker process occurs with each pellet in the previously defined position inside the kiln, that is, nodulation is forced during molding, contrary to what occurs in industrial kiln. The decrease in the size of the pellets increases their surface area and, with the reduction in associated mass, there is a greater energy associated with these pellets, facilitating the formation of alite, which needs more energy to be formed, from the belite. It is known that alite (formed from 1260 ºC), the main cement phase, is formed from the reaction of belite (formed from 800 ºC) with CaO (TAYLOR, 1997). In other words, during clinkering, previously molded pellets with larger dimensions, or smaller specific surface area, require greater energy for the reactions to occur properly, as they do in an industrial rotary kiln. Thus, pellets with a greater specific surface area are more favorable in forming higher alite content.
The results obtained were evaluated statistically by means of analysis of variance (ANOVA), presented in Tables 8, 9 and 10. The analyzes were made by comparing the results of the alite phases in clinkers produced from pellets with 1 cm in diameter in relation to the other pellet configurations and indicated that the differences in the percentages of alite formed between the different pellet configurations are statistically significant, except for clinkers with CCW-1 incorporation, between 1 cm pellets and 2 cm diameter semisphere (Table 10). Ambiente Construído, Porto Alegre, v. 20, n. 4, p. 205-223, out./dez. 2020. Portland clinker with civil construction waste: influence of pellet geometry on the formation of crystalline phases 219  The specific surface area of 1 cm diameter pellets is, on average, approximately 2.57 cm /g, about 1.4 times greater than the specific surface area of 2 cm diameter pellets (on average, 1.90 cm /g), 2.1 times greater than the specific surface area of pellets with a diameter of 3 cm (on average, 1.23 cm /g) and 10% greater than that of semi-spherical pellets (on average 2.32 cm /g). Therefore, pellets with 1 cm diameter presents a better clinkering, due to their higher specific surface area, indicating that they are more reactive than pellets with other geometries and sizes. Specific surface area values are directly proportional to reactivity of clinkers (DINGER, 2005).
The average volume of pellets with 3 cm in diameter is 13.468 cm , the average volume of pellets with 2 cm in diameter is 3.495 cm , while the volume of pellets with 1 cm in diameter is 0.529 cm , 25.5 and 6.6 times higher, respectively, which results in a higher heat demand for the same calcination conditions.
Thus, in pellets with 2 cm and 3 cm in diameter there is a greater difficulty of heat transfer during clinker under the same burning conditions when compared to pellets with 1 cm in diameter and half-sphere shape. This also affects the cooling of the clinkers, influencing the formation of crystalline phases, in these different pellet configurations.
Analyzing the amount of C3A formed, it is clear that in clinkers with CCW-1 incorporation there was a lower formation of cubic C3A (contents of 6.7 to 7.5%), when compared to reference clinkers and with CCW20-40. (contents of 9.3 to 11.6%), probably because CCW-1 has a lower content of Al2O3 compared to CCW20-40 (7.5% against 11.67, as shown in Table 4) and clay, used in the clinker of reference (23.55% Al2O3).
Tricalcium aluminate (C3A) is the clinker phase responsible for the initial hydration reactions. In the absence of gypsum, the formation of hydrated aluminates, both in the cubic and orthorhombic crystalline system, occurs immediately after mixing with water, however, the cubic C3A is more reactive and releases more heat (KIRCHHEIM et al., 2010). Thus, the incorporation of CCW-1 may have a greater influence on cement hydration as is less reactive and may have a delayed setting time, in relation to the other types of clinker analyzed in this research. The content of orthorhombic C3A varied slightly between the different clinkers and conformations of the spheres.
From the analysis of the average contents of the crystalline phases alite, belite and free lime (Tables 6 and 7) in the experimental clinkers of reference and in the clinkers with the two types of CCW it was noticed that the main phases of the clinker were formed in similar contents to those indicated in the specialized literature to classify them as Portland clinker (50 to 70% alite) (TAYLOR, 1997). This was even with the different conformations of the pellets which it did not negatively affect its mineralogical characteristics. The C3S content of clinkers is the most important indicator for controlling the development of the main properties of Portland cement (SCRIVENER; JUILLAND; MONTEIRO, 2015).
The amount of free lime in all dosages and in different conformations remained below 1%, indicating that both the burning and the cooling were not affected when there was a change in the geometry of the pellets. It can also be noticed that the clinkers produced with the incorporation of CCW in total replacement of clay and partial of limestone had a higher alite content, when compared to the clinkers produced with traditional raw materials (limestone and clay). This demonstrates that the replacement is technically feasible, enabling lower CO2 emissions and less exploitation of natural resources.

Conclusions
From the results obtained in the present work, it can be concluded that: (a) the incorporation of the CCW enabled the reduction in the use of up to 14% of limestone in the raw mix for the production of Portland clinker. Therefore, there would be a reduction in the extraction of this natural material; (b) the specific surface area of the shaped pellets influences the formation of the alite and belite phases, due to differences in the exposure area, during sintering and cooling. Thus, we propose standardization of the size of the pellets for analysis purposes involving the production of clinker on a laboratory scale; (c) although there are differences in the contents of the crystalline phases formed, due to the different conformations of the pellets, this did not negatively affect the mineralogical characteristics. This is because the content of the main clinker phases, remained within the values indicated in the specialized literature. to classify them as Portland clinker (50 to 70% alite); (d) the use of molds standardizes and facilitates the making of clinkers, increasing the homogeneity of the material produced; and (e) clinkers containing CCW in total replacement for clay and partial replacement for limestone presented a mineralogical composition similar to reference clinkers.