1. Introduction

This paper describes the way diverse iron ore types behave during desliming and flotation stages. According to Clemer[1], desliming prior to the reverse cationic flotation of iron ores started with the USBM process. Araujo et al.[2] stated that efficient desliming requires an adequate dispersion degree of the particles in the pulp. The dispersion degree is strongly affected by the slurry pH.

The presence of slimes in a flotation feed may be quantified by the pass (the by-pass represents the fine particles that follow the water flow in a classification process, in a way that they may be considered as incorporated to the fluid)

Queiroz at al.[3] observed that for certain itabirite ores attrition increases the mass recovery to the slimes and flotation concentrate fractions enhancing the selectivity.

The main goal of the investigation was evaluating how different slimes contents (by-pass) affect the flotation behavior of different iron ore types.

2. Methodology

Channel sampling on the bench face was utilized. The samples preparation flow sheet is illustrated in figure 1.

The –150µm fraction of each ore type was employed in the experiments.

Figure 1. Samples preparation and processing flow sheet

The thermodynamic stability level of the suspensions was assessed via the determination of the slimes dispersion degree in a sedimentation tube (2.5% solids in 500mL volume), described by Peres and Silva[4] and Martins et al.[5]. The pH of the slurry was varied, with caustic soda acting as dispersing agent, in the range from 5.5 to 10.5, at 0.5 intervals. After adjusting the pH, the slurry was stirred under constant speed inside a cylindrical device with orifices placed in order to remove settled and non-settled solids. Stirring was achieved by a magnetic bar for 5 minutes. Settling took place during one minute after which an overflow and an underflow were separated and weighed. The dispersion degree (DD%) was calculated by equation (1):

(1)

where: W_of and W_uf refer, respectively, to the weights of the overflow and the underflow.

The test for the by-pass determination was performed in a 4000mL volume PVC vessel. The slurry at 50% solids weight/weight was stirred, homogenized, and then left under rest for 5 minutes prior to several desliming stages until the desired by-pass level was achieved. The by-pass (BP) is quantified by equation 2:

(2)

where: V_ore = volume of ore in the slurry; 4000mL = volume occupied by the slurry in the PVC vessel; 1600mL = limit established in the vessel for collecting the supernatant (overflow or slimes); n = number of desliming stages.

Flotation tests, under standard procedure, were performed as close as possible to five by-pass levels: 33%, 10%, 4%, 1% and 0.33%.

Additional flotation tests were performed with the -45µm deslimed fraction of iron ores presenting different mineralogical compositions (specular hematite and goethite / porous hematite).

3. Results and Discussion

The following results are presented and discussed:

Size analyses, chemical and mineralogical characterizations, dispersion degree as a function of pH and flotation results for different levels of by-pass.

Tables 1, 2 and 3 show, respectively, the chemical composition, particle size distribution and mineralogical characterization of the samples under investigation.

The dispersion degree as a function of pH is illustrated in figure 2.

Sample 940 FN presents the highest values of dispersion degree. The dispersion degree increases only slightly with the pH. The dispersion degree of sample 1080 L4 is not affected by the pH. Most of the samples show an increase in the dispersion degree in the pH range between 7 and 8 followed by a region in which the pH does not affect the dispersion degree. Sample 910 E6 presents the lowest levels of dispersion degree, in the full pH range, among the samples under investigation.

Table 1. Chemical composition of the samples

Table 2. Size distribution of the samples (cumulative % passing) Sample 0,150mm 0,105mm 0,075mm 0,053mm 0,044mm 1280 E2 100,0 86,1 72,5 53,5 38,9 1290 E2 100,0 83,6 70,5 53,0 39,6 1090 E3 100,0 91,2 75,0 53,9 42,7 1040 E4 100,0 81,8 70,2 40,9 31,4 1000 E5 100,0 80,3 69,7 44,7 34,5 910 E6 100,0 77,9 64,7 34,2 24,6 1040 L3 100,0 90,2 84,2 66,5 57,6 1080 L4 100,0 84,6 68,8 49,0 37,7 940 FN 100,0 86,6 78,7 61,0 54,1

Table 3. Mineralogycal characterisation of the samples

Figure 2. Dispersion degree as a function of pH

Tables 4 to 12 illustrate the flotation behavior of the samples under different by-pass levels. SI stands for Gaudin’s selectivity index.

Table 4. Flotation results (sample 1280 E2) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 68.02 0.95 0.024 80 15 10.00 68.00 0.96 0.024 81 15 4.00 67.79 0.97 0.023 79 14 1.00 67.67 0.75 0.026 80 15 0.33 68.10 0.73 0.25 81 16

Table 5. Flotation results (sample 1290 E2) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 68.47 0.72 0.028 71 12 10.00 68.32 0.98 0.028 73 11 4.00 68.40 0.93 0.029 74 11 1.00 68.84 0.67 0.030 73 13 0.33 68.94 0.67 0.029 84 17

Table 6. Flotation results (sample 1090 E3) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 62.74 7.87 0.060 49 4 10.00 66.42 2.80 0.058 56 8 4.00 66.03 2.67 0.056 71 10 1.00 66.31 2.54 0.050 71 10 0.33 67.25 1.16 0.054 73 15

Table 7. Flotation results (sample 1040 E4) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 63.69 7.51 0.044 50 3 10.00 64.55 6.40 0.041 52 4 4.00 64.93 5.93 0.042 55 4 1.00 66.78 3.27 0.039 61 7 0.33 67.47 1.82 0.041 63 8

Table 8. Flotation results (sample 1000 E5) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 68.65 1.34 0.037 70 9 10.00 69.18 0.56 0.039 72 14 4.00 69.25 0.64 0.038 71 13 1.00 69.27 0.51 0.038 75 16 0.33 69.17 0.48 0.040 76 16

Table 9. Flotation results (sample 910 E6) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 65.60 4.70 0.032 76 5 10.00 65.95 4.57 0.036 76 5 4.00 67.16 2.44 0.033 81 8 1.00 69.02 0.51 0.034 91 27 0.33 68.87 0.65 0.033 91 23

Table 10. Flotation results (sample 1040 L3) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 67.00 1.56 0.018 73 7 10.00 68.84 1.04 0.017 87 15 4.00 69.28 0.65 0.019 72 14 1.00 69.14 0.61 0.017 80 17 0.33 69.04 0.67 0.017 82 17

Table 11. Flotation results (sample 1080 L4) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 67.98 0.75 0.054 74 12 10.00 67.66 0.71 0.051 73 12 4.00 68.01 0.70 0.050 73 12 1,00 68.27 0.68 0.048 72 12 0.33 68.29 0.70 0.046 73 13

Table 12. Flotation results (sample 940 FN) By-pass (%) Concentrate grades (%) Fe distribution % in concentrate SI Fe SiO2 P 33.00 68.47 0.80 0.054 75 10 10.00 68.96 0.47 0.049 78 16 4.00 68.97 0.47 0.046 79 17 1.00 68.70 0.48 0.041 76 14 0.33 68.91 0.58 0.037 81 15

The results on tables 4 to 12 show that, in general, the by-pass affects the flotation behavior. The flotation selectivities of samples 1280 E2 and 1080 L4 are not affected by the by-pass. The high dispersion degree of sample 1080 L4 might be an explanation for this behaviour. The by-pass level interferes with the flotation selectivity of the other samples. For instance, in the case of sample 910 E6 the selectivity index seemed not to be affected by the presence of slimes in the first desliming stages, but the last stage caused a sudden increase in the selectivity index from 8 to 27 when the by-pass was reduced from 4% to 1% (4th desliming stage). For this sample, low by-pass levels yield extremely low silica grades in the concentrate. Sample 1040 E3 did not present an adequate flotation performance even after 5 desliming stages (by-pass 0.33%). A general trend that is also observed is a significant reduction in phosphorus content in the concentrate for tests at low by-pass levels.

The different behaviors observed for samples with different by-pass levels did not show any correlation with size distribution, chemical analyses nor mineralogical composition. The samples used in the investigation up to this point were not available any more when evidences from further studies with different iron ore types suggested that differences in specific surface area might be the explanation for the differences observed in the flotation performances. The results of additional flotation tests, performed with the -45µm deslimed fraction of iron ores presenting different mineralogical compositions (specular hematite and goethite / porous hematite), are shown on table 13.

Table 13. Effect of the specific surface on the flotation behavior of iron ores presenting different mineralogical compositions Mineralogical composition By-pass (%) Specific surface are (cm2/g) Fe distribution in the concentrate SiO2 in the concentrate Specular 4 1000 69,7 1,2 hematite 6 3000 67,0 2,7 Goethite 2 1295 92,9 1,0 / porous 4 1556 92,5 1,1 hematite 8 1887 90,7 1,1

Table 13 indicates that by pass-levels from 2% to 8% did not affect the flotation performance of iron ore samples with predominance of goethite and porous hematite, possibly because the specific surface area did not present a significant increase for higher by pass levels.

On the other side, the specular hematite sample showed a very high increase in the specific surface area of the fraction -45µm of the flotation feed with the increase of the by-pass from 4 to 6. In this case, the SiO₂ content in the concentrate increased from 1.2% to 2.7% when the specific surface of the -45µm fraction increased from 1,000cm²/g to 3,000cm²/g, corresponding to a small increase of the by-pass level (4% to 6%).

4. Conclusions

The flotation behavior of different iron ore types depends on the size distribution, chemical and mineralogical characteristics. The dispersion degrees of the samples correlate with the by-pass levels. The methodology developed represents a simple and reliable tool for plant practice control. The industrial selectivity of an ore type may be predicted from dispersion degree determinations and a few desliming and flotation laboratory scale experiments.

Most samples showed a maximum dispersion degree and the effect of increasing the pH level above 8.0 is not significant. Concerning plant practice, this fact provides a guideline for the desliming operation, for an effective removal of slimes may be carried out without the need of adding high caustic soda dosages. In addition to the reduction in caustic soda consumption, a better control of the subsequent slimes thickening operations is achieved, for the dispersion state may be easily reversed in this stage with lower dosages of coagulants.

The design of industrial desliming circuits must take into account the water partition to the underflow. An industrial hydrocyclones circuit, capable of effectively discarding the slimes, must be designed to provide a maximum of 4% by-pass in order that the SiO₂ content in the concentrates from most samples reaches the usual specifications.

The specific surface area can be used in future investigations to explain the flotation behavior of different iron ore types under distinct by-pass levels.