The results of clay body (CB) mineralogical analysis by XRD are shown in Fig. 1. The diffractogram reveals that CB is composed of, quartz (Q), mica (M), gibbsite (Gi) and less intense peaks related to some potassic feldspar (Fk).
Table 1 presents the CB chemical analysis. One should notice the predominance of SiO2, Al2O3 and Fe2O3, which reveals a typical composition of clay-based raw material for red ceramics [30, 31]. However, the lower percentage of Fe2O3 was responsible to provide a lighter red color after sintering. In addition, it is possible to observe a relatively high percentage of Al2O3, which tends to increase the CB refractoriness, when compared to other clays used to produce roofing tiles with recognized technical quality [32]. The loss on fire (LoF) comes mainly from the dehydroxylation of the clay minerals and the gibbsite [33].
Table 1
– Chemical composition of clayey body (% wt.)
| SiO2 | Al2O3 | Fe2O3 | K2O | TiO2 | MgO | Na2O | CaO | P2O5 | MnO | LoF1 |
| 52.20 | 28.07 | 3.46 | 1.31 | 1.20 | 0.64 | 0.37 | 0.30 | 0.11 | <0.05 | 12.33 |
1Loss on fire. |
Figure 2 - Granulometric distribution of clayey body on Winkler diagram
The apparent dry density (DD) of the clay ceramic specimens under different treatments is shown in Table 2. Statistically different variations (P≤0.05) in DD, are verified when evaluating the general averages for the heating rate (HR) and the sintering temperature (ST), as well as by comparing the interaction between these two factors. In the shaping by extrusion, which is the most used method for red clay ceramic bricks production in Brazil, it is not possible to directly control the specimens’ weight. The same thing happens in the shaping by pressing. Thus, variations in density may occur.
Table 2
Apparent dry density (DD - g/cm3) of specimens as a function of sintering temperature (ST) and heating rates (HR)
HR (°C/min) | ST (°C) | Mean |
800 | 900 | 1000 |
02 | 1.77 Cb (±0.02) | 1.77 Bb (±0.02) | 1.82 Aa (±0.02) | 1.79 C (±0.03) |
15 | 1.82 Aa (±0.02) | 1.83 Aa (±0.01) | 1.82 Aa (±0.02) | 1.83 A (±0.02) |
30 | 1.80 Bb (±0.02) | 1.81 Aab (±0.03) | 1.82 Aa (±0.02) | 1.81 B (±0.03) |
Mean | 1.80 b (±0.03) | 1.81 b (±0.03) | 1.82 a (±0.02) | 1.81 (±0.03) |
* Means followed by the same uppercase letters in the column and lower case letters in the row, do not differ according to the ANOVA and Tukey test (P ≤ 0.05). |
The water absorption (WA) values under the different treatments are shown in Table 3. As expected, an increase in ST resulted in WA reduction. However, promising results are observed when comparing the averages between the different heating rates. The increase in HR from 2 to 15ºC/min was able to reduce the WA, proportionally, by about 6.4%. The HR of 30ºC/min was also able to reduce the WA when compared to the conventional HR (2ºC/min). However, it revealed a higher value than the intermediate HR (15 ºC/min).
The apparent porosity (AP), which is a parameter associated with pores interconnected with the exterior of the structure, is markedly related to the WA. In fact, an apparent inverse relationship is observed between WA and AP when comparing the results in Tables 3 and 4. It is also possible to note the existence of this inverse relationship between the DD and WA, justifying the differences associated with distinct HRs. It can be seen that the averages values of WA for each HR follow the contrary behavior of the DD. Indeed, HR at 15ºC/min is related to the highest DD, followed by 30ºC/min and 2ºC/min. In the inverse way, the lowest WA was found at HR of 15°C/min, followed by 30°C/min and 2°C/min.
Table 3
Water absorption (WA - %) of specimens as a function of sintering temperature (TQ) and heating rates (TA)
HR (°C/min) | ST (°C) | Mean |
800 | 900 | 1000 |
02 | 22.43 Aa (±0.31) | 21.99 Ab (±0.12) | 20.52 Ac (±0.49) | 21.65 A (±0.89) |
15 | 21.51 Ba (±0.29) | 20.28 Bb (±0.28) | 19.25 Bc (±0.33) | 20.34 C (±0.98) |
30 | 22.23 Aa (±0.62) | 20.39 Bb (±0.56) | 19.19 Bc (±0.39) | 20.60 B (±1.37) |
Mean | 22.06 a (±0.58) | 20.89 b (±0.87) | 19.65 c (±0.74) | 20.86 (±1.23) |
*Means followed by the same uppercase letters in the column and lower case letters in the row, do not differ according to the ANOVA and Tukey test (P ≤ 0.05). |
Table 4
Apparent porosity (AP - %) of specimens as a function of sintering temperature (TQ) and heating rates (TA)
HR (°C/min) | ST (°C) | Mean |
800 | 900 | 1000 |
02 | 36.48 Ab (±0.53) | 3717 Aa (±0.15) | 35.10 Ac (±0.63) | 36.25 A (±0.99) |
15 | 35.78 Ba (±0.70) | 34.49 Bb (±0.31) | 33.27 Bc (±0.45) | 34.51 C (±1.15) |
30 | 36.73 Aa (±0.82) | 34.72 Bb (±0.60) | 33.26 Bc (±0.49) | 34.90 B (±1.57) |
Mean | 36.33 a (±0.78) | 35.46 b (±1.29) | 33.88 c (±1.02) | 35.22 (±1.46) |
*Means followed by the same uppercase letters in the column and lower case letters in the row, do not differ according to the ANOVA and Tukey test (P ≤ 0.05). |
However, when assessing the interaction between HR and ST, this trend is not observed by setting ST equal to 1000 ºC, which shows homogeneity in the DD, while the WA has undergone significant changes with the HR increasing. To make this information visually more clear, the data are shown in the graphs of Fig. 4, which contains the results of DDs and WAs for the different HRs and STs. It is important to mention that the density results shown in Fig. 4 (a) are relative to the apparent dry density and the specimens have not yet been sintered. However, this representation was found necessary in order to clarify the understanding in how the differences in DD may lead to changes in other properties, especially the WA (Fig. 4b).
Thus, considering only the ST at 1000°C, for the same DD of 1.82 ± 0.02 g/cm, the WA was significantly higher (P≤0.05). When the HR of 2°C/min was used (20.52 ± 0.49%) it compares to the rates of 15°C/min and 30°C/min, respectively, of 19.25 ± 0.33% and 19.19 ± 0.39%, that did not differ among themselves. At the 1000°C, by increasing the HR from 2°C/min to 15°C/min, an approximately 6% reduction in WA was achieved. However, no statistical variation was noted with increasing the HR from 15°C/min to 30°C/min.
In order to better understand the relationship between these properties, the Pearson (R) correlation was calculated for all the variables studied. The lower variance by the Tukey test (P <0.05) and correlation superior to 60% were stipulated, so that only the interaction between WA and AP, as well as WA and DD were chosen, and are represented in Table 4. For WA and AP correlation, they are already well described in the literature [22, 30, 31], showing that pores interconnected with the surface are those responsible for absorbing water from the environment.
The direct correlation between DD with the other properties, shown in Table 5, such as WA, AP and mechanical strength, makes DD a property of great interest, especially for the ceramic industry [22, 37–40]. Indeed, the higher the DD, the smaller the AP becomes and, as a consequence, the smaller the WA the greater the strength. It should be emphasized that packaging degree and DD are different concepts. However, as there is no variation in the grains density, it is correct to extend the effect of DD correlation to the other properties, as long as one takes care to use the same clay specimens to produce the specimens.
This discussion is pertinent for the correct results evaluation, since the differences found between HR from 2ºC/min to 15ºC/min and 30ºC/min, for ST of 800 and 900ºC, can be explained by differences in DD, which presented R ≥ |60%|. However, this fact does not explain the difference in ST at 1000°C (R = -22%), so it can be inferred that this difference was due to the increase of the HR.
Table 5
Significant (P<0.05) Pearson correlations (R) and greater than |60%| among the evaluated properties
ST (ºC) | Water Absorption |
Apparent Porosity | Apparent Dry Density |
800 | 0.67 | -0.60 |
900 | 0.99 | -0.73 |
1000 | 1.00 | -0.22 |
Another interesting parameter is the linear shrinkage (LS) with results presented in Table 6, which also indicates that the sintering occurred more efficiently with increasing HR. This property is correlated with the degree of sintering. In a simplified explanation, sintering is a process of mass transport in which material is displaced from its original positions to another, resulting in closing of empty space (porosity reduction) and consequent structural shrinkage [12]. In this way, (LS) is a parameter that helps the sintering degree evaluation, being related to the volumetric reduction of total porosity, i.e., both open and closed pores.
In this context, Table 6 shows that there is an increase in LS with an increase in HR from 2ºC/min to 15ºC/min and 30ºC/min. The same trend is observed when analyzing, separately, the interaction between the HR versus ST variables. In all cases, i.e., for all the STs evaluated, there is an increase in LS when comparing the conventional sintering cycle (2°C/min) with the fast-sintering cycle (15ºC/min and 30°C/min).
Table 6
– Sintering linear shrinkage (LS - %) of specimens as a function of sintering temperature (ST) and heating rates (HR)
HR (°C/min) | ST (°C) | Mean |
800 | 900 | 1000 |
02 | 0.52 Ac (±0.16) | 1.15 Bb (±0.35) | 1.79 Ba (±0.24) | 1.15 B (±0.58) |
15 | 0.60 Ac (±0.22) | 1.43 Bb (±0.28) | 2.15 Aa (±0.23) | 1.39 A (±0.68) |
30 | 0.43 Ac (±0.31) | 1.86 Ab (±0.57) | 2.19 Aa (±0.36) | 1.49 A (±0.88) |
Mean | 0.52 c (±0.24) | 1.48 b (±0.50) | 2.04 a (±0.33) | 1.35 (±0.73) |
*Means followed by the same uppercase letters in the column and lower case letters in the row, do not differ according to the ANOVA and Tukey test (P ≤ 0.05). |
With respect to the bending rupture stress results, shown in Table 7, it is revealed that the highest strength of 7.32 MPa, was obtained by the sintered specimens with HR of 15°C/min, followed by 6.03 MPa for TA at 2°C/min and 5.07 for TA at 30°C/min. At 30°C/min. this might be detrimental, due to the fact that the different crystalline phases, as well as the amorphous matrix, have distinct thermal expansion coefficients. Thus, an excessive increase in HR generates tensions that are responsible for surpassing the material mechanical resistance causing cracks and reducing its mechanical strength [9]. In addition, impairing in strength is also attributed to the volumetric expansion caused by the quartz allotropic transformation, in the sand, which at higher HRs is capable of generating stresses above those the products may support.
Table 7
– Bending rupture stress (MPa) of specimens as a function of sintering temperature (ST) and heating rates (HR)
HR (°C.min-¹) | ST (°C) | Mean |
800 | 900 | 1000 |
02 | 4.99 Ab (±0.25) | 5.84 Bb (±0.44) | 7.26 Ba (±1.10) | 6.03 B (±1.17) |
15 | 5.47 Ac (±0.35) | 7.70 Ab (±0.99) | 8.79 Aa (±0.97) | 7.32 A (±1.62) |
30 | 4.86 Aa (±0.94) | 4.80 Ca (±0.83) | 5.55 Ca (±1.85) | 5.07 C (±1.30) |
Mean | 5.10 c (±0.64) | 6.11 b (±1.44) | 7.2 a (±1.89) | 6.14 (±1.65) |
*Means followed by the same uppercase letters in the column and lower case letters in the row, do not differ according to the ANOVA and Tukey test (P ≤ 0.05). |
The aforementioned results clearly demonstrate that sintered samples at higher HRs exhibited greater densification rates, assessed indirectly through the presented properties, than samples sintered at lower HRs. This shows that the rate of material transport depends not only on the ST but also on the HR. Moreover, the HR has such a significant influence on the rate of material transport that it exceeds the effect of reducing the sintering time.
Dilatometry experiments were performed with the clay body using three sintering cycles with different HRs of 2°C/min, 15°C/min and 30°C/min. Corresponding dilatograms are shown in Fig. 5. Regarding this figure, it is important to mention that the dilatometry tests were repeated and the same behavior was observed.
The curves in Fig. 5 describe similar qualitative behavior. In fact, they correspond to the same events, but with different HR and ST intensities, so that some events occur at higher temperatures for higher heating rates. The shifting in temperature observed in Fig. 5 was expected, since the temperature sensor of the equipment indicates the environment temperature in which the sample is located. Due to the thermal gradient experienced by any material subjected to heating, the environment temperature does not necessarily match the temperature of the sample. Thus, HRs rates should correspond to higher temperature gradients.
After initial shrinkage, caused by elimination of free and adsorbed water, the sintering starts producing kaolinite dehydroxylation. At this moment, there is an increase of particles specific surface area [24], resulting consolidation of the sintered structure and an increase and crystallinity loss, which may favor mass transport. Simultaneously, the quartz allotropic transformation, indicated by the structure expansion, occurs at different temperatures as a function of HR, being 575, 598 and 618°C at 2°C/min, 15°C/min and 30°C/min respectively.
After the complete kaolinite dehydroxylation, the liquid phase begins to form through eutectics reaction involving the silica contained in the metacaulinite and the alkali oxides contained in the potassic feldspars [20]. At this stage there is the greatest intensity discrepancy of LS values displayed in Table 6, when comparing the HR of 2°C/min with 15°C/min and 30°C/min. The end of this stage can be identified by the change in the curve slope, initiating the spinel and/or primary mullite formation [41–44].
The dependence that the sintering rate has on the HR has already been reported in metals and ceramics. Silva et al. [45] found that the shrinkage rate of hard metal structures (WC-Co) shows a strong dependence on HR. Saleiro and Holanda [46] identified that a red clay body sintering at HRs of 1, 10 and 20°C/min culminated in similar properties, favoring energy savings. Santos et al. [47] sintered hard metal alloys via pulsed plasma with HRs above 50°C/min and found that the structure densification rate has a HR dependence. In fact, it is well known that pulsed plasma sintering can densify structures in short time intervals and one of the characteristics of this technique is the high HR employed.
Jones and Miles [48] and Harmer and Brook [49], cited by Johnson et al. [50], sintered alumina materials at high HRs, between 40°C/min and 200°C/min, and reported high densification rates under these conditions. In addition, they observed that grain sizes were significantly lower when compared to sintering at slower HRs. These authors attributed the high densification rates and the finer grain sizes to the suppression of surface diffusion occurring by the rapid passage through the temperature regime where surface diffusion rate is intense, avoiding the deleterious effects of this process [51, 52].
In the conventional sintering process, much of the time is spent in temperature ranges where the effects of grain growth, through surface diffusion, outweigh the densification produced by diffusion in grain boundaries and in the crystalline lattice itself. Other studies [53, 54] reaffirm that grain growth reduces the driving force for sintering. This is explained by the increase in the average distance required for diffusion, where mass transfer from the grain contours to the neck formation occurs, which is the mechanism responsible for porosity reduction [55, 56].