Tomographic images and electrical resistivity measurements
The total number of defects in anodes was determined with the software developed by the UQAC carbon group to analyze the numerous 2D images taken during tomography. An example is given in Fig. 2a for anodes 3 and 4 (Table 3). Images showing all defects along the length are presented in Fig. 2a for both green and baked anodes produced with different butt contents. Darker regions correspond to higher number of internal defects. The resistivity maps between the small side surfaces of anodes, which were measured with the equipment developed at UQAC (SERMA), were also determined [24, 25]. These for the same cases are shown in Fig. 2 (b). Darker regions have higher resistivities. Figure 2c shows a lab anode and the direction of measurement. It should be noted that the green anode electrical resistivities are about two orders of magnitude greater (above 1500 μΩ·m) than those of the baked anodes. Small variations in pitch concentration in green anodes could give significant differences in the resistivity values. The baked anode resistivities, on the other hand, directly indicate the presence of pores and cracks. These figures show that the internal number of defects is higher for both green and baked anodes containing greater butt percentage compared to the standard anode. The electrical resistivities also show a similar trend. From these figures, the agreement between the tomography and SERMA results can be seen: the zones that contain more defects (determined by the tomography) have higher resistivities (determined by SERMA). Also, the comparison of the baked and green anode resistivities shows that generally a high resistivity zone in a green anode results in a high resistivity zone after baking.
These analyses were carried out for all the cases in Table 3. Then, the overall results are determined and presented in the following section.
Effect of raw materials
In this part, the effects of anode recipe and butt and pitch contents on the cracking and anode properties were investigated (Table 3).
Two anodes (anodes 1 and 2, Table 3) were produced using different particle size distributions. In anode 1, a standard anode recipe was used whereas the anode 2 was produced with a modified recipe. The recipe modification and its effect on anode properties were previously studied . In this study, its effect on crack (defect) formation is studied. In the modified recipe, the fractions of medium and coarse particles were readjusted. The ultra-fine fractions (ball mill product (BMP) and filter dust (FD)) were kept similar to that of the standard recipe.
Comparing the electrical resistivity of the two green anodes (Fig. 3a) shows that the resistivity decreased when the recipe is modified; however, the number of defects remained practically the same. This indicates that the difference in resistivity values is due to slight differences in pitch distribution. On the other hand, after baking (Fig. 3b), the anode made with the modified recipe is found to have a higher resistivity compared to the anode prepared with the standard recipe. This shows that the modified recipe is likely to cause the creation of somewhat more cracks during baking. The internal defects of the baked anode made using modified recipe is slightly higher compared to the standard anode.
A good anode density is generally considered as a sign of a good quality anode. Figure 3c presents the apparent and optical density of two green anodes made with different recipes and Fig. 3d presents the densities of the same anodes after baking. It can be seen that the apparent densities of both the green and the baked anodes were slightly improved when the anode recipe was modified. The optical densities of the anodes also show the same trend. The optical density, which is obtained from the tomographic analysis, indicates the relationship between solid and void space in materials. Higher optical density was obtained for the modified aggregate which means that the solid content is higher than the void space in this anode compared to the standard anode both before and after baking. The results of the tomographic analysis results agree with the direct measurements of the densities. The surface crack densities of both anodes are similar (Fig. 3e). The surface cracks are not good indicators of anode quality since an anode with surface cracks can have an interior structure of acceptable quality.
Anodes 3 and 4 have different butt contents (25% and 35%, Table 3). Figure 4a and b show that increasing butt content did not influence the internal defect percentage in green anodes but increased it slightly in baked anodes. The electrical resistivity increased as the butt content increased both for green and baked anodes. The impact of higher butt content may be explained with the low wettability of butt with pitch matrix  and the difference in thermal expansion coefficients of coke and butt [2,3,4, 26]. This difference might create stress, consequently, anodes might crack during baking. It might also be due to the insufficient pitch penetration into the butt pores. The results also show that increasing butt content decreased the anode apparent and optical densities both for green and baked anodes (Fig. 4c and d). It should be noted that the percentage of pitch was not adjusted when the percentage of butt was changed. In general, the effect of butts on anode properties depends on the butt quality. Visual inspection of the baked anodes indicated that the surface cracks decreased when the butt content was increased (Fig. 4e). As it was mentioned previously, surface cracks do not necessarily represent the internal anode quality.
Three different pitch percentages (13%, 15%, and 18%) were used to study the effect of this parameter on the quality of the anodes (anodes 3, 5, and 6, Table 3). The results show that the increase in the percentage of pitch decreases the internal defect percentage for green anodes (Fig. 5a) as expected. When there is not enough pitch, pores of coke and the interparticle spaces are not completely filled, resulting in a porous anode .
Electrical resistivity of green anodes decreased with increasing pitch content up to a certain pitch percentage (15%), which was close to optimum pitch content. However, the electrical resistivity increased with further increase in pitch percent. When pitch fills the voids between the coke particles and the pores in the particles of green anodes, the electrical resistivity decreases. Nevertheless, if the pitch percentage is increased more, the accumulation of excess pitch between the particles increases the electrical resistivity [3, 29]. This can be seen in Fig. 5a where the anode produced using 15% pitch had lower electrical resistivity than the other two anodes containing 13% and 18% pitch. For baked anodes, increasing the percentage of pitch decreased the resistivity as shown in Fig. 5b. Internal defects were slightly higher for the baked anode containing the highest pitch percentage, which was over-pitched to some extent. Over-pitching can cause crack formation during baking due to the higher quantity of released volatiles (Fig. 5b).
As it can be seen in Fig. 5c, both the optical and apparent densities increased with increasing pitch percent for green anodes. The baked anode optical and apparent densities increased appreciably when the pitch content is increased from 13 to 15%. However, further increase in pitch percent did not have a significant effect on the densities (Fig. 5d). The apparent density increased slightly whereas the optical density remained the same.
The results of the visual inspection (surface cracks from Fig. 5e) show that increasing the amount of pitch can cause many surface cracks, but this does not reveal the internal quality. The anode with a high percentage of pitch has a low electrical resistivity and low internal defects compared to those of the anode made with 13% pitch (Fig. 5b), but it has high surface crack density (Fig. 5e). This again shows that the quality of anodes cannot be evaluated correctly only with the visual inspection of the surface.
Green anode fabrication parameters
Two anodes (anodes 3 and 7) were produced using vibration times of 60 and 72 s (Table 3). Figure 6 shows the effect of vibration time on internal defects, electrical resistivity, apparent and optical densities, and specific surface crack density.
The green and baked anodes produced using lower vibration time has lower resistivities than those of the anodes produced using a higher vibration time. This result indicates that the anode was over-compacted when 72 s was used. Over-compaction (too high a vibration time) causes more stress accumulation in the green anode. Both green anodes have similar amounts of defects. However, baked anode produced using a lower vibration time has a slightly lower defect percentage. Accumulation of stress during green anode formation results in defect formation during baking. The electrical resistivity is lower for the anode produced using a lower vibration time for both green and baked anodes (Fig. 6a and b). However, there is a limit for decreasing the vibration time. If the anodes are not compacted enough (under-compaction), this also increases the internal defects and the resistivity. It is important to find the optimum vibration time necessary for a given anode .
The results show that both green and baked anodes produced using the low vibration time has a higher apparent and optical densities than those of the anodes made with the higher compaction time (Fig. 6c and d).
The anode manufactured using a low vibration time has more surface cracks than that manufactured with a high compaction time (Fig. 6e). Visual inspection does not indicate the internal anode quality. The internal defect analysis show that the quality of this anode is better than the over-compacted anode (Fig. 6a and b) even if it has more surface cracks.
Top-former bellow pressure
Two anodes (anodes 3 and 8) were made using a top-former bellow pressure of 41 and 30 psi (Table 3). The electrical resistivity decreased as the pressure increased both for green and baked anodes (Fig. 7a and b). Under-compaction (in this case, due to an insufficient pressure) led to poor anode quality. Less pitch penetrated between the particles and in the particle pores, increasing the electrical resistivity. The percentage of internal defects for the two green anodes were similar (Fig. 7a) whereas those of the baked anodes were clearly different. The baked anode manufactured with the low compaction pressure had more internal defects than that manufactured using a higher pressure. Under-compaction makes the matrix weak and facilitates the formation of pores and cracks (Fig. 7b). If the anode has more defects, its electrical resistivity also increases as pores and cracks form a barrier to the passage of an electric current (discontinuous solid medium).
As expected, the anode manufactured under a low compaction pressure has a lower density (both apparent and optical) than the anode manufactured at the higher compaction pressure (Fig. 7c and d). The baked anode compacted at higher pressure had more surface cracks which did not again reflect the overall quality of the anode (Fig. 7e).
Green anode cooling medium
Three green anodes (anodes 3, 9 and 10, Table 3), which were produced under the same conditions, were cooled differently: free convection in air, forced convection in air, and forced convection in a water bath. Cooling was stopped when the measured temperature reached the temperature of the cooling medium. The percentage of internal defects was the lowest for the anode cooled by forced convection in air and the highest for the one cooled by free convection in air both for green and baked anodes. The electrical resistivity was lowest when the anode was cooled by immersing in water bath for green anodes, and the resistivities of the green anodes cooled with forced and free convection in air were similar. The resistivity of a green anode is generally influenced by pores/defects as well as by pitch distribution. Non-carbonized pitch has a high electrical resistivity. If there are high pitch regions, an increase in resistivity occurs in those regions [3, 26]. Also, the cooling rate is lower in air, which leads to a more porous anode due to the spring-back effect. For baked anodes, the resistivity and the number of internal defects were the lowest for the anode cooled by forced convection in air and the highest for the one cooled by free convection in air (Fig. 8a and b). In addition, the apparent and optical densities of baked anodes, which were cooled using forced convection in air after green anode is formed, were the highest. For the green anode, the apparent density was highest for the anode cooled with free convection in air whereas the optical density was highest for the anode cooled with forced convection in air. The presence of impurities in the raw material could also affect the optical density (Fig. 8c and d). The surface cracks were also lowest for the anode cooled with forced convection in air (Fig. 8e). In general, the results indicate that the forced convection in air seems to be the best option for cooling green anodes among the three options tested in this study. It must be noted that there are other parameters that need to be studied such as water and air temperatures, air flow rate for the forced convection, combination of water/air cooling media, etc. before the most suitable cooling medium could be selected.
Heating rate used during baking
Eight anodes, four without butts (anodes 14, 15, 16, and 17) and four containing butts (anodes 3, 11, 12, and 13), were baked using low (7 °C/h), medium (11 °C/h), high (15 °C/h), and combination (15–7–15 °C/h) heating rates (Table 3). Concerning the combination heating rate, after initially using a high heating rate at low temperatures, a low heating rate was used during the pitch devolatilization period since most of the cracks form during this period. The rest of the baking was carried out at the highest heating rate. Baking time depends on the heating rate used. Usually, the lowest heating rate is the best since the volatiles are released slowly during pitch carbonization, and this prevents the formation of high internal pressure in the anode during volatile release and consequently reduces the cracks formation . However, this increases the anode production time and cost, and decreases the production rate. The combination heating rate scheme was proposed since the baking during the critical period was carried out at a low heating rate without increasing the total production time compared to that of standard heating rate (medium) (Table 3).
Anodes without butts
The electrical resistivities and apparent densities of the anodes before baking (green anodes) are given in Fig. 9a and c. As it can be seen from these figures, there are some differences in these properties of the green anodes although they are produced under the same conditions. This is expected since the anode raw materials are non-homogeneous, resulting in small differences in properties. The effect of heating rate on the internal cracks and electrical resistivity are presented in Fig. 9b whereas their effect on the apparent and optical densities are given in Fig. 9d for baked anodes. Figure 9e presents the surface crack density of the baked anodes.
The results show that the internal defect percentage and electrical resistivity of anodes baked at the lowest and combination heating rates are similar and lower than those of the anodes baked using medium and high heating rates. In addition, the anodes baked at the lowest and combination heating rates have higher apparent and optical densities than the other two anodes baked at medium and high heating rates. The surface crack density also decreased with decreasing heating rate and the anode baked at the combination heating rate had a similar surface crack density to the anode baked at the lowest heating rate (Fig. 9e). It seems that the better-quality anodes are produced using the lowest and combination heating rates; however, it must be noted that their corresponding green anodes had better quality before baking. Therefore, more testing is needed to confirm the positive impact of low heating rate on anode quality.
Anodes with butts
Four other anodes, this time with the butt addition, were produced and baked using the four heating rates explained in the previous section. The properties of green anodes used are given in Fig. 10 a and c. The heating rates shown in these figures indicate the heating rate used when these green anodes were baked. It is difficult to make green anodes with the same properties due to the non-homogeneity of the raw materials. Although, the green anode, which was later baked using the combination heating rate, had the highest internal crack before baking, the corresponding baked anode had the lowest internal cracks among the anodes tested (Fig. 10b). In addition, this anode also had high anode apparent and optical densities after baking even if it had lower green anode densities compared to that of the green anode baked at the lowest heating rate and similar to that of the green anode baked at the medium heating rate (Fig. 10b). The anode baked with the combination heating rate also had the lowest resistivity. As it can be seen from the Fig. 10a and b, although the anode baked at the medium heating rate had similar electrical resistivity to those baked using the low and combination heating rates in green state, it had higher resistivity after baking when compared to those of the same anodes in baked state. The properties of the anode baked using the combination heating rate were similar to those baked at the lowest heating rate. Its surface crack density was also the lowest (Fig. 10e). After, these tests, it now possible to state that the lowest and combination heating rates give better quality anodes. The combination heating rate has the advantages of shorter production time (similar to that baked at medium heating rate, which is the average heating rate usually observed in industry) compared to the anode produced using the low heating rate.
Correlation between anode properties and anode quality
The Fig. 11a and b show the effect of internal defect (cracks/pores) percentage on the electrical resistivity for green and baked anodes, respectively. Electrical resistivity increased as the internal defect percentage increased for both anodes. This means that the anode quality decreases with increasing internal defects since higher resistivity indicates higher energy consumption to produce the same quantity of aluminum. The high resistivity indicates the presence of pore/cracks as well as the local high pitch regions in green anodes since pitch has high resistivity before carbonization (baking); and the high resistivity is due to defects/pores in baked anodes.
Similarly, it can be seen from the optical and apparent density vs. internal defect percentage data that the anode density, hence the anode quality, decreased with increasing internal defects both for green (Fig. 11c and d) and baked anodes (Fig. 11e and f). This is in agreement with the resistivity data. These results show the tendencies expected with respect to the relation between various parameters.