Materials

The CH was provided by a local company (Cúcuta, Colombia) belonging to the National Federation of Colombian Coffee Growers. The other raw materials used to make the ceramic paste were four clay types (CL01, CL02, CL03, CL04), feldspar (Feld), limestone, raw, and fired break. A local ceramic company (Cúcuta, Colombia) provided these raw materials to conduct the study with the standard mixture formula, which has been designated as “P.” The feldspar was replaced with increasing amounts of CH and CHA.

Obtaining CHA

CH underwent two different treatments before being used in pulp. First, to obtain a uniform particle size, CH was milled in a rotary hammer mill and passed through a 74 μm mesh. Second, uncut kerosene was burned by imperfect combustion in a temperature-controlled oven (650 °C) for two and a half hours to remove the highest content of organic matter present in the material. The carbonized husks were brought back to 700 °C for two and a half hours to guarantee the total decomposition of organic matter [²¹]. After combustion, the CHA was crushed and passed through a 74 μm sieve.

The chemical composition of raw materials was determined using the X-Ray Fluorescence (XRF) technique and a sequential X-ray fluorescence spectrometer of 4 KW dispersive wavelength (Bruker S8 Tiger model). The main crystalline phases of the feldspar and CH were defined by X-ray diffraction (XRD) using a Bruker D8 Advance powder diffractometer with DaVinci geometry. Phase quantification was performed by Rietveld refinement using corundum as the reference phase for the calculation of the amorphous phase [²²].

Evaluation of CH as a Feldspar Substitute in the Manufacture of Ceramic Tiles

Blends were prepared with 78.5% clays (CL01, CL02, CL03, CL04), 16.5% feldspar, 2% limestone, 1% raw break, and 2% fired break (p0). Six more mixtures were prepared, in which the proportion of feldspar was replaced by 5, 10, and 100% CHA (CHA5, CHA10, CHA100) and cha (CHA5, CHA10, CHA100); the percentages for the other components remained constant. Each 400 g mixture was prepared by wet milling (36% by weight of water), using a Gabrielli fastball mill, with 400 g of alumina grinding bodies for 2 minutes and 45 seconds.

After grinding, the mixtures were dried at 110 °C for 24 hours and subsequently ground to diameters less than 1 mm in a Servitech rotary hammer mill. Ten specimens of each mixture were prepared using as a base 80 g of semi-dry powder (humidified at 5.6- 5.8% by weight) for each specimen; uniaxial pressing at 26.7 bar was performed on a GABTEC press using a steel mold (110 mm x 55 mm x 8 mm). The specimens were dried to constant weight at 110 °C. Subsequently, the drying shrinkage (DS) percentage was determined for its newly formed lengths. The specimens were then enameled using a bell-type application system to be subsequently baked in an industrial roller kiln at a maximum temperature of 1,120 °C with a full cycle of 34 min. For the ten fired specimens, losses by calcination, firing shrinkage, water absorption, bulk density, and mechanical resistance to bending were determined.

Losses of ignition (LOI) (physically adsorbed water still present, dehydroxylation, decomposition of carbonates, and organic matter) were determined by evaluating the mass difference of dry and fired test pieces. The firing shrinkage (FS) of the obtained specimens was calculated by determining the linear shrinkage of the specimens after firing them compared to the dry dimension. The water absorption percentage (E), the boiling in water method) and bulk density q (ρ) were established as outlined in the NTC 4321-3 standard [²³]. Afterward, all the specimens were dried at 110 °C at a constant weight to determine the mechanical flexural strength (S) and modulus of rupture (R) according to NTC 4321-4 [²⁴].

Characterization of Raw Materials

Table 1 shows the chemical composition of the raw materials determined by XRF. The results of microstructure analysis by XRD are presented in Table 2. From the chemical point of view, the clay materials that make up the paste (CL01, CL02, CL03, and CL04) are characterized by being rich in silicon and aluminum. The mineralogical composition allows establishing quartz as the predominant crystalline phase, followed by the presence of phyllosilicate minerals such as kaolinite and muscovite. These two phases are responsible for explaining the higher content of Al and Si. The potassium content is associated with the presence of muscovite and the microcline-type feldspar that is a minority phase. The iron content is associated with the presence of hematite in clay materials and possibly in muscovite. The presence of titanium correlates with the existence of anatase in the mineralogical composition of Table 2.

Due to its contents of Fe₂O₃, Na₂O, K₂O, MgO, and CaO, clay materials can be considered suitable for the manufacture of stoneware red, with some variations in their tones by the content of titanium oxide; these colors are characteristic of the region, especially the metropolitan area of Cúcuta [²⁵], [²⁶]. High silica content in the form of quartz (SiO₂) is observed. It acts as a degreaser in the paste, favoring the degassing of impurities and the elimination of water during the drying process and reducing linear shrinkage and deformation during firing [²⁷].

The content of alumina of the analyzed raw materials is approximately 21%, enough quantity to provide the clay with high melting temperature, low thermal conductivity, high resistance to corrosion, low expansion coefficient, and hardness [²⁸]. The results obtained from the feldspar show that it is an alkaline potassium-sodium type, with a low sodium proportion and traces of barium compared to potassium.

For CH, small traces of quartz (SiO₂₎ are identified as crystalline phases with XRD (Table 2 and Fig. 1), revealing that CH is a practically amorphous material. The CHA shows higher contents of iron and potassium oxides, which will be responsible for the reddish color of the fired bodies and the sintering at low temperature, respectively. The high content of K₂O allows inferring that this material will have a strong flux effect compared to feldspar, whose K₂O content is 5.58% by weight.

Source: Own elaboration.

Fig. 1 XRD pattern of raw materials. 

The presence of microcline and albite (sodium and potassium feldspar) in the feldspar used is compelling from the ceramic point of view since, if only potassium feldspar prevails, the vitreous mass will remain very viscous, even at high temperature. If sodium predominates, the decrease in viscosity with the increase in temperature will be faster.

Evaluation of the physical-mechanical behavior of mixtures

The use of CH and CHA in the manufacture of ceramic tiles affected the final properties of these products. Table 3 presents the results of the properties evaluated. The viscosity of the paste increased with the incorporation of larger quantities of CH. Therefore, the CH100 mixture, whose viscosity after wet milling was so high that it could not flow, was not tested further since it did not guarantee the adequate discharge of mills, the transport to storage tanks, and atomization at the industrial level [²⁹].

Table 3 indicates that in the ceramics obtained at 1,120 °C, as CHA replaced feldspar, there was an increase in the percentage of firing linear shrinkage, being very similar to the p mixture with the 10% CHA addition. On the contrary, for mixtures with CH, the linear firing shrinkage decreased at higher amounts of CH as feldspar replacement. The apparent density of the fired pieces decreased with the percentage of the addition of CH and CHA in replacement of feldspar, the CH10 and CHA100 mixtures having the lower apparent density, even inferior to the standard mixture (20.9% and 20.4% respectively).

For CH fired pieces, at higher additions of CH, the percentages of water absorption decreased significantly, even getting reductions of 16.8% and 18.4% for the CH5 and CH10 mixtures, respectively, in comparison to the p mixture. The mixtures with CHA presented a similar behavior with additions of up to 10%, whose values are very similar to that of the standard mixture. On the chemical composition of CHA, the literature reveals the presence of potassium oxides (23-46%), phosphorus (3-25%), magnesium (4-22%), calcium (10-18%), sodium (0.1-8%), and quartz (1-17%) [³⁰]-[³²]. This composition explains why the more significant the proportion of CHA, the higher the content of fluxes (K₂O, CaO), which increased the amount of liquid phase during sintering, decreased porosity, and helped to increase the mechanical strength. However, for the mixture with 100% feldspar replacement by CHA, water absorption increased up to 1.3 times compared to P.

Based on the results of e and the criteria in [³³], the fired specimens of mixtures CH5, CH10, CHA5, and CHA10 are classified within group BIIB (6% < E < 10%, S > 500 N), ceramic tiles of colored and porous support without application of glazes or decorations. Different surface defects in the enameled prototypes, probably due to the degassing process of the inorganic matter that mainly compose CH (99.6% by weight), are the principal causes of porosities that contribute to decreasing bending resistance and increasing water absorption, as suggested [³⁰]. Fig. 2 shows that mixes CH5, CHA10, and CHA5 presented the best bending resistance, exceeding the flexural strength of the pattern mix by 17.2%, 7.5%, 2.3%, respectively.

Source: Own elaboration.

Fig. 2 Flexural strength and modulus of rupture of fired specimens. 

Also, blackheart gradually increased as the CH and CHA content increased; its presence additionally caused swelling of the prototypes, deterioration of the technical and aesthetic characteristics of the enamel, and change of tone in the support and enamel materials, especially CHA100 prototypes. This behavior can be attributed to the rapid firing cycle handled by the roller kiln and the impossibility of gases (CO₂ and CO), which have developed inside the ceramic piece, to go outside because of the surface waterproofing condition.