Minerals Engineering 16 (2003) 11971203 This article is also available online at: www.elsevier.

com/locate/mineng

Froth recovery measurement in plant scale otation cells

a,*

, J.P. Franzidis

b,1

, E.V. Manlapig

b,1

JKTech, University of Queensland, Isles Road, Indooroopilly, Qld 4068, Australia Julius Kruttschnitt Mineral Research Centre, University of Queensland, Isles Road, Indooroopilly, Qld 4068, Australia Received 27 May 2003; accepted 14 July 2003

Abstract Over recent years a number of techniques have been developed as part of the AMIRA 2 P9 project to represent mathematically or measure directly sub-processes occurring in otation cells. These include the determination of: a machine parameter (bubble surface area ux, Sb ); a froth parameter (froth recovery, Rf ); entrainment and drainage parameters (n and d); and an ore oatability parameter (Pi ).

A methodology for estimating froth recovery in large scale plant otation cells is proposed. The primary aim of the technique is to estimate the froth recovery in a way that is simple, inexpensive and non-intrusive. Plant engineers can use the estimate of froth recovery in day to day trouble-shooting and circuit optimisation. The current denition of froth recovery is given, along with a detailed description of the proposed methodology, equipment required, assumptions and calculations. Advantages and limitations of the methodology are described and possible applications of the froth recovery estimate in current otation circuit modeling are presented. 2003 Elsevier Ltd. All rights reserved.
Keywords: Flotation froths; Modelling; Process optimisation

1. Introduction Recent advances in the modelling of the otation process in industrial scale circuits have been based on the development of sub-process models of bubble surface area ux (Gorain et al., 1997; Deglon et al., 1999), entrainment (Johnson, 1972; Savassi et al., 1998), ore oatability (Harris, 1997; Runge et al., 1997; Alexander and Morrison, 1998) and froth recovery (Vera et al., 1999; Savassi et al., 1997; Mathe et al., 2000). However, all the above sub-process models are in their industrial scale infancy and extensive industrial trials are required to collect data that will validate each sub-process model.

Once validated, these sub-process models when combined will provide the plant engineer with a tool for optimising existing otation circuits. A new method for measuring froth recovery in industrial otation cells is proposed. The technique allows measurement of the froth performance in a otation cell directly, with minimum eect on cell operation. It is based on a combination of two published methods to measure froth recovery; viz., Savassi et al. (1997) and Vera et al. (2000). This combination provides plant engineers a method that will allow froth performance measurements to be made easily in large scale otation cells. This paper outlines the previous work in this area and describes the new experimental procedure to estimate froth recovery.

Corresponding author. Tel.: +617-3365-5984; fax: +617-33655900. E-mail addresses: [email protected] (D.J. Alexander), [email protected] (J.P. Franzidis), [email protected] (E.V. Manlapig). 1 Tel.: +617-3365-5888; fax: +617-3365-5999. 2 Australian Mineral Industries Research Association. 0892-6875/$ - see front matter 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.mineng.2003.07.016

2. Background Flotation is usually modelled as a single separation process with an overall rst order rate constant. In these

1198

D.J. Alexander et al. / Minerals Engineering 16 (2003) 11971203

generalised approaches, it is often dicult to correlate numerous commonly used operating parameters (examples include air owrate, froth depth, etc.) with a otation kinetic model. To overcome this problem, several authors have suggested that the otation process should be investigated by splitting the process into two distinct phases; the pulp phase and the froth phase. Recent studies into modelling the pulp phase in conventional industrial otation cells have indicated that a machine parameter (namely bubble surface area ux) and distributed oatability components are a useful modeling approach. A number of techniques have also been proposed to model the froth phase, including direct estimation (Falutsu and Dobby, 1989; Savassi et al., 1997), mass balancing (Contini et al., 1988) and model tting (Moys, 1978; Yianatos et al., 1988). The majority of these techniques were developed for batch otation cells or column cells and are not applicable for continuously operated industrial scale sub-aeration otation cells. In a detailed critique of froth phase modelling, Mathe et al. (2000) concluded that the most practical approach to modelling the froth phase in an industrial mechanical cell is to determine a froth recovery parameter. This approach, developed for column cell otation, has been used by a number of authors (Feteris et al., 1987; Laplante et al., 1989; Finch and Dobby (1990); Mular and Musara, 1991; Vera et al., 1999). Finch and Dobby (1990) dene the froth recovery parameter as Rf k kc 1

First Order Rate Constant (min-1)

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0 10 20 30 Froth Depth (cm)

y = -0.0019x + 0.113 2 R = 0.9991

40

50

Fig. 1. Linear relationship between overall rst order rate constant and froth depth.

depth; i.e. y intercept. The collection zone rst order rate constant can then be used to calculate the froth recovery parameter at any froth depth using Eq. (1). The advantages of this technique are that it is simple and applicable to various cell sizes, types and duties. However, the following factors have been found to limit the applicability of this technique in most plant situations: the method is intrusive on the overall cell performance; the method disturbs the performance of downstream cells; the method generates a large number of samples; the method assumes that entrainment has a minimal impact on the general form of the relationship. Savassi et al. (1997) proposed a method to estimate the froth recovery in large otation cells by collecting samples of pulp just beneath the pulpfroth interface and the concentrate, and mass balancing the ow of attached particles over the froth phase. The advantages of this technique are that it is non-intrusive on the operation of the otation cell, removes the eect of entrainment, and downstream cells are not aected. However, Savassi et al. (1997) noted that the main limitation of the method was that it was only applicable to rougher cells. This was likely due to: the diculty in estimating the attached particle grade; the low mass of samples taken which may have high sampling errors; and, the dierence between the attached and entrained grade being typically too small in cells operated in anything other than a rougher application, to allow use of the method with any degree of condence. In this study, a new technique is proposed that improves the estimate of froth recovery in a cell by com-

where Rf is the froth recovery (%), k the overall rst order rate constant (min1 ) and kc the collection zone rst order rate constant (min1 ). In AMIRA P9 otation testwork, the froth recovery parameter is currently estimated using the technique rst developed by Feteris et al. (1987) and later used by Laplante et al. (1989), Mular and Musara (1991) and Vera et al. (1999). The technique involves varying the froth depth in a cell and taking samples of the feed, concentrate and tailings. These samples are assayed and the information used to calculate the rst order rate constant for each froth depth, from the equation: k R s 1 R 2

where R is the overall mineral recovery, and s the cell residence time (min). Note that Eq. (2) assumes that the system behaves as a perfect mixer. Previous authors have validated this assumption in large industrial otation cells (Burgess, 1997; Alexander et al., 2000). The overall rst order rate constant is then plotted against froth depth to produce a linear relationship as shown in Fig. 1. The collection zone rst order rate constant (kc ) is estimated by extrapolation to zero froth

D.J. Alexander et al. / Minerals Engineering 16 (2003) 11971203

1199

bining the methods of both Savassi et al. (1997) and Vera et al. (1999). 3. Experimental technique The following procedure is used to estimate froth recovery in industrial cells: Measure supercial gas velocity (Jg ) and air hold-up (eg ) at various locations across the cell; Take samples of feed, concentrate, tail, air hold-up and top of the froth using appropriate sampling equipment (see below); Measure slurry weight, solids weight and percent solids, and assay all samples (by size if required); Conduct a mass balance over the entire cell with the appropriate sample standard deviations; Conduct a mass balance over the concentrate stream with the appropriate sample standard deviations; Conduct a mass balance over the air hold-up with the appropriate sample standard deviations; Calculate the ow of attached particles entering and leaving the froth (the ratio of these is dened as the froth recovery). Cell characterisation measurements (i.e. the supercial gas velocity and air hold-up measurements) are made at several locations across the otation cell. The experimental techniques required are well documented in Power et al. (2000). Feed, concentrate and tail samples of individual cells are taken using appropriate sampling equipment. These samples are assayed and percent solids noted. The air hold-up is dened as the ratio of the volumetric fraction of air to the overall volumetric fraction of air and pulp. Fig. 2 illustrates the measuring equipment used (note that the air hold-up equipment is operated vertically). Briey, the device comprises a cylindrical tube with two air operated pinch valves at each end. The device is lowered to about 10 cm below the pulp/froth interface with the pinch valves closed. The pinch valves are then opened and air and pulp travels through the device for approximately 30 s. After this period, the pinch valves are closed and the volume of pulp is measured. This sample is analysed for percent solids and assayed. Fig. 3
Rubber tube Rubber tube

Fig. 3. Plant demonstration of the air hold-up probe.

demonstrates the use of the air hold-up probe in taking a sample 10 cm below the pulpfroth interface. Vera et al. (2000) proposed that a sample taken at the very top of the froth and well away from the lip of the cell should represent the grade of attached particles in the pulp. This top of froth sample has been incorporated into this method for measuring froth recovery. Mass balances over the cell, concentrate and pulp (air hold-up) are required to estimate the mass owrate of attached particles entering the froth and the mass owrate of attached particles in the concentrate. Each mass balance uses a least squares minimization technique to estimate balanced ows and assays. 4. Theory The technique is based on three key assumptions: The grade of the froth sample represents the grade of particles attached to bubbles; The average grade of particles attached to bubbles does not change over the froth phase; The grade of the tail sample is equivalent to the grade of particles that are entrained in the water. All three assumptions can be tested on a cell, bank or circuit basis. Varying the froth level in a cell and determining the maximum concentrate grade produced from the cell can test the rst assumption. This is illustrated in Fig. 4 where the maximum grade in this industrial rougher cell is approximately 9.2%. This gure compared very well to the grade of the top of froth sample of 9.3%. In other cases, this result may not be valid and hence this should be tested on a regular basis. Recently, several authors have suggested that the froth is non-selective in terms of attached particles, i.e., that particles do not break their bonds with air bubbles in the froth phase due to their hydrophobicity but rather that the dropback of particles to the pulp phase is caused by bubble breakage and coalescence. A simple

Plastic cylindrical tube

Compressed air Three-way valve

Air line Pinch valve 2

The atmosphere

Pinch valve 1
Fig. 2. Air hold-up sampling device.

1200
10 9 8 7 6 5 4 3 2 1 0 0 100 200

D.J. Alexander et al. / Minerals Engineering 16 (2003) 11971203

The three assumptions outlined above allow a set of mass balance equations to be formed over the froth phase. The air hold-up sample taken directly below the pulpfroth interface and the supercial gas velocity measurement are used to estimate the mass owrate of attached particles entering the froth phase, based on the following equation:
Pulp MATT 3:6 QAir BPulp Pulp MATT

Grade (%)

300

400

500

Froth Depth (mm)

test to determine the validity of this assumption is to plot the froth recovery of one mineral against another mineral under the same conditions. This is illustrated in Fig. 5 from data collected using the change in froth depths method described previously. A 45 line through the origin and passing through the experimental results supports the hypothesis of non-selective drop-back in this case. Selectivity to mineral type in the froth phase would be indicated by deviation of the data from the 45 line. There is evidence in the literature that this assumption may not always be valid for some mineral systems. The third assumption can be tested using the entrainment analyser developed by Savassi et al. (1998) or by determining the mixing characteristics (i.e. plug ow of perfect mixing) of the cell using a soluble tracer. The entrainment analyser sample is the preferred method because it is simple, direct and requires one extra sample. The grade from the entrainment sample can be compared directly to the grade of the tail sample and if there is a discrepancy the entrainment sample grade should be used.
1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10
0.00

where is the mass owrate of attached particles in the pulp zone (t/h), QAir the volumetric owrate of air (m3 /s) and BPulp the bubble load in the pulp phase (grams of solids/litre of air). The volumetric owrate of air can be measured directly in each cell or estimated using the supercial gas velocity measurements across the cell cross sectional area: QAir A Jg 4 where Jg is the supercial gas velocity (cm/s) and A the open area of cell (m2 ). The bubble load is dened as the mass of attached solids per volume of air. Numerous authors have attempted to measure bubble load in laboratory environments with limited applicability to industrial scale performance. Savassi et al. (1997) proposed a method to estimate the bubble load in industrial otation cells. In this method, samples are collected just below the pulp/ froth interface using a funnel connected to a vacuum (see Fig. 6). This technique was shown by Savassi et al. (1997) to be very good for obtaining reliable bubble load

C P V

R = 0.9967

(a)

Rf Mineral 1

180

0.00

0.20

0.40

0.60

0.80

1.00

(b)
Fig. 6. (a) Sketch of the bubble load analysergeneral layout (after Savassi et al., 1997). (b) Sketch of the bubble load analyserfunnel layout (after Savassi et al., 1997).

Rf Mineral 2
Fig. 5. Typical plant data showing that the froth is non-selective within experimental error.

D.J. Alexander et al. / Minerals Engineering 16 (2003) 11971203

1201

measurements in industrial cells. However, Savassi et al. (1997) also showed that the technique was not accurate in low bubble load environments (scavengers) or in high grade pulp environments (cleaners and recleaners). This limited the application of the technique to the rougher sections of otation circuits. In addition, the methodology was dicult to use in most plant applications since it was time consuming and the equipment was not particularly robust. A new method of obtaining bubble load is proposed which is based on the Savassi technique, but uses more robust equipment and a more accurate method. It involves obtaining a sample using the air hold-up probe and using the assumptions outlined above. Fig. 7 illustrates that air bubbles (with particles attached) and pulp (with particles entrained) pass through the air hold-up sample in the open position. When the air hold-up probe is closed, the attached particles become detached from the air bubbles and are included with the pulp containing entrained particles and water. The volume of air and the mass and grade of the pulp sample are noted. The bubble load can then be calculated using: matt 1000 5 V where matt is the mass of attached particles in the air hold-up probe (g), V the volume of air measured in the air hold-up probe (ml). The mass of attached particles in the air hold-up probe is estimated using the following mass balance equation: BPulp matt mtotal Gtotal Gent Gatt Gent 6

grade of entrained particles (Gent ) is taken to be equivalent to the grade of the tailings. This is in accordance with the assumptions made at the beginning of this section. The estimate of the mass of attached particles in the air hold-up sample obtained from Eq. (6) can then be used to calculate the mass owrate of attached particles into the froth phase (Eqs. (3) and (5)). The mass owrate of attached particles in the concentrate can be calculated using a similar mass balance equation to Eq. (6) applied to the concentrate sample: mconc att mConc Gtotal Gent total Gatt Gent 7

where mConc is the mass owrate of attached particles in att the concentrate stream (t/h), mConc the mass owrate of total all particles in the concentrate stream (t/h), Gent the grade of entrained particles in the concentrate, Gtotal the concentrate grade and Gatt the grade of attached particles in the concentrate. The froth recovery of attached particles can then be calculated using the following equation: Rf mConc att mPulp att 8

where Rf is the froth recovery of attached particles, mConc att the mass owrate of attached particles in the concentrate stream (t/h) and mPulp the mass owrate of attached att particles in the pulp (t/h). 4.1. Case studycentury zinc concentrator, Australia The new froth recovery technique was used to measure froth recovery in a rougher and scavenger application. In both cases, the otation cells were Outokumpu 100 m3 tank cells in the zinc circuit. Samples of feed, concentrate, tail, air hold-up and top of froth were taken from each of the cells and analysed for percent solids, lead, zinc, iron and silica. The supercial gas velocity was measured at several locations across the cell with the average supercial gas velocity calculated at 1.43 and 1.30 cm/s in the rougher and scavenger, respectively. Measurements of the average air hold-up were estimated at 12% and 15% in the rougher and scavenger, respectively. Froth depths of 50 cm in the rougher and 20 cm in the scavenger were also recorded. The bubble load was calculated using Eqs. (5) and (6). Bubble load estimates of 46.8 and 29.2 g/l were calculated in the rougher and scavenger, respectively. The rst point to note from these results is that the magnitude of the bubble loads are in a similar range to that reported by Savassi et al. (1997). Secondly, the bubble load attained in the rougher was signicantly higher than the scavenger. This follows the logical argument that more attached particles would be present in the rougher section of the circuit than the scavenger. Finally,

where mtotal is the total mass of particles in the air holdup probe (g), Gent the grade of entrained particles in the air hold-up probe, Gtotal the grade of all particles in the air hold-up probe and Gatt the grade of attached particles in the air hold-up probe. In order to calculate the mass of attached particles (matt ), the grade of attached particles (Gatt ) is taken as that of the grade of the top of froth sample, while the

Fig. 7. Simple illustration of the use of the air hold-up probes in obtaining fast, robust and reliable bubble load estimates in large scale otation cells.

1202

D.J. Alexander et al. / Minerals Engineering 16 (2003) 11971203

the measurement was easily attained and time required for calculation was dependent only on the assay turn around time. The bubble load was used to calculate the mass owrate of attached particles in the pulp phase (rougher 68.1 t/h; scavenger 38.7 t/h). In addition, a mass balance of the concentrate sample was conducted to estimate the mass owrate of attached particles in the concentrate (rougher 27.8 t/h; scavenger 35.5 t/h). The froth recovery was then calculated using Eq. (8). The estimate of the froth recovery in the rougher cell was 40.8% and 91.6% in the scavenger cell. The large dierence in the froth recovery values measured was due to the dierence in the froth depths in each of the cells (rougher 50 cm; scavenger 20 cm). Gorain et al. (1997) showed that strong relationships between froth recovery and the froth residence time of air and slurry existed. Examples of these relationships in a 100 m3 plant operating cell are shown in Figs. 8 and 9 from work previously conducted by the authors at another site. These relationships show that small dierences in froth depth can have a large impact on the froth recovery and is the likely reason for the dierence in froth recovery between the two cells. The study at Pasminco Century shows that the new technique provides froth recovery and bubble load values which are similar to reported values in similar applications in the literature. Extensive testwork is planned

to validate the proposed method at numerous sites and applications. Included in these studies will be a statistical analysis to estimate the statistical reliability of these measurements. 4.2. Proposed areas of froth recovery technique use The proposed froth recovery technique can be used in the following areas: For plant engineers to estimate froth recovery in operating plant cells, to assist with operational trouble shooting; For project engineers to estimate froth recovery in operating plant cells, to assist with plant diagnosis and future project areas; For engineering consultants to measure froth recovery in plant cells, to provide better models and simulators of otation circuits. 4.3. Future work The following work is proposed by the authors to improve the reliability of the froth recovery technique: Collect froth recovery measurements in dierent applications to better dene the main areas of use. These applications should include various cell types and sizes, duties (i.e. rougher, cleaner scavenger, etc.) and ore types (zinc, copper, nickel, etc.); Test the three main assumptions to determine the situations where they do not apply; Conduct an error and statistical analysis on the froth recovery technique by  Multiple measurements at similar conditions;  Using a Monte Carlo simulation. Investigate the usefulness of the entrainment and bubble load estimates in determining the impact of various operating variables including reagent effects.

Measured Froth Recovery (%)

100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0%

y = 0.6862e-0.0046x 2 R = 0.6843

0.0%
0.00 50.00 100.00 150.00 200.00 250.00 300.00 350.00

Froth Residence Time of Slurry (s)

Fig. 8. Typical relationship between measured froth recovery and the froth residence time of slurry in a 100 m3 Outukumpu tank cell.

5. Conclusion
Measured Froth Recovery (%)
100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% 0.00

y = 1.4357e 2 R = 0.9079

-0.0519x

A practical methodology for measuring the froth performance in plant scale otation cells has been proposed. The technique was developed to improve the practicality of two previous methods of measuring froth recovery, namely, Savassi et al. (1997) and Vera et al. (1999). The improvements were based on:
50.00

10.00

20.00

30.00

40.00

Froth Residence Time of Air (s)

Fig. 9. Typical relationship between measured froth recovery and the froth residence time of air in a 100 m3 Outukumpu tank cell.

Measuring the froth performance in a sub-aerated plant scale otation cell directly, with minimum eect on cell, bank or plant performance; Reducing sample numbers and time required to take measurement.

D.J. Alexander et al. / Minerals Engineering 16 (2003) 11971203

1203

The technique requires three assumptions which can be easily tested in each application of the technique. Future work will be conducted to conrm the validity of these assumptions under various conditions and cell duties. The froth recovery technique was used to measure the froth recovery in a 100 m3 rougher and scavenger cell at Pasminco Century lead/zinc mine in North Queensland. The results of this application show that the technique can be applied to both a zinc rougher and zinc scavenger duty. Both the bubble load and froth recovery estimates were in a similar range to previous results reported by other authors. Further studies at various sites are planned to test the validity of the technique in other duties.

Acknowledgements The assistance of sta from Pasminco Century Mine, in the development of this technique is gratefully acknowledged. The authors would also like to acknowledge the members of the AMIRA P9 otation group especially Kym Runge, Marco Vera and Robert Coleman for their contribution in developing this technique. References
Alexander, D.J., Morrison, R.D., 1998. Rapid estimation of oatability components in industrial otation plants. Minerals Engineering 11, 133143. Alexander, D.J., Runge, K.C., Franzidis, J.P., Manlapig, E.V., 2000. The application of multi-component oatability models to full scale otation circuits, AusIMM Seventh Mill Operators Conference, Kalgoorlie, Australia, October, 2000, pp. 167178. Burgess, F.L., 1997. OK 100 tank cell operation at PasmincoBroken Hill. Minerals Engineering 10, 723741. Contini, N.J., Wilson, S.W., Dobby, G.S., 1988. Measurement of rate data in otation columns, Column 88, AIME, pp. 8189. Deglon, D.A., Sawyerr, F., OConnor, C.T., 1999. A model to relate the otation rate constant and the bubble surface area ux in mechanical otation cells. Minerals Engineering 12, 599608. Falutsu, M., Dobby, G.S., 1989. Direct measurement of froth drop back and collection zone recovery in a laboratory otation column. Minerals Engineering 2, 377386. Feteris, S.M., Frew, J.A., Jowett, A., 1987. Modelling the eect of froth depth in otation, 1987. International Journal of Mineral Processing 20, 121135.

Finch, J.A., Dobby, G.S., 1990. Column Flotation. Pergamon Press, Oxford. Chapter 3. Gorain, B.K., Franzidis, J.P., Manlapig, E.V., 1997. Studies on impeller type, impeller speed and air ow rate in an industrial otation cellPart 4. Eect of bubble surface area ux on otation performance. Minerals Engineering 10, 367379. Harris, M.C., 1997. A practical framework for otation circuit modelling and simulation, paper presented at SAIChE 97, the Eighth National Meeting of the South African Institute of Chemical Engineers, Cape Town, 1618 April, 1997. Johnson, N.W., 1972. The otation behaviour of some chalcopyrite ores, PhD Thesis, University of Queensland. Laplante, A.R., Kaya, M., Smith, H.W., 1989. The eect of froth on otation kineticsa mass transfer approach. In: Frothing in Flotation. Gordon and Breach Science Publishers, New York, pp. 7799. Mathe, Z.T., Harris, M.C., OConnor, C.T., 2000. A review of methods to model the froth phase in non-steady state otation systems. Minerals Engineering 13, 127140. Moys, M.H., 1978. A study of a plug ow model for otation froth behavior. International Journal of Mineral Processing 5, 221238. Mular, A.L., Musara, W.T., 1991. Batch column otation: rate data measurements. In: Proceedings of an International Conference on Column Flotation, Sudbury, Ontario, Canada, pp. 6374. Power, A., Franzidis, J.P., Manlapig, E.V., 2000. The characterization of hydrodynamic conditions in industrial otation cells. In: AusIMM Seventh Mill Operators Conference, Kalgoorlie, Australia, October 2000, pp. 243255. Runge, K.C., Frew, J., Harris, M.C., Manlapig, E.V., 1997. Floatability of streams around the Cominco Red Dog Lead Cleaning Circuit. In: AusIMM Sixth Mill Operators Conference, Madang, Papua New Guinea, October 1997, pp. 157163. Savassi, O.N., Alexander, D.J., Johnson, N.W., Manlapig, E.V., Franzidis, E.V., 1997. Measurement of froth recovery of attached particles in industrial cells. In: AusIMM Sixth Mill Operators Conference, Madang, Papua New Guinea, October 1997, pp. 149 155. Savassi, O.N., Alexander, D.J., Manlapig, E.V., Franzidis, J.P., 1998. An empirical model for entrainment in industrial otation plants. Minerals Engineering 11, 243256. Vera, M.A., Manlapig, E.V., Franzidis, J.P., 1999. Simultaneous determination of collection zone rate constant and froth zone recovery in a mechanical otation environment. Minerals Engineering 12, 11631176. Vera, M.A., Manlapig, E.V., Franzidis, J.P., 2000. The locus of otation performance. In: Massacci, P. (Eds.), Proceedings of the XXI International Mineral Processing Congress. Developments in Mineral Processing, vol. 13. Elsevier, Hilton, Rome, Italy, pp. A3:7482. Yianatos, J.B., Finch, J.A., Laplante, A.R., 1988. Selectivity in column otation froths. International Journal of Mineral Processing 23, 279292.