Minerals Engineering xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Minerals Engineering

journal homepage: www.elsevier .com/locate /mineng

The effect of particle breakage mechanisms during regrindingon the subsequent cleaner flotation

http://dx.doi.org/10.1016/j.mineng.2014.04.0200892-6875/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding authors. Address: Julius Kruttschnitt Mineral Research Centre,The University of Queensland, Isles Road, Indooroopilly, Brisbane, QLD 4068,Australia (Y. Peng). Tel.: +61 7 3365 7156; fax: +61 7 3365 3888.

E-mail addresses: x.chen7@uq.edu.au (X. Chen), yongjun.peng@uq.edu.au(Y. Peng).

Please cite this article in press as: Chen, X., et al. The effect of particle breakage mechanisms during regrinding on the subsequent cleaner flotation.Eng. (2014), http://dx.doi.org/10.1016/j.mineng.2014.04.020

Xumeng Chen a,⇑, Yongjun Peng a,b,⇑, Dee Bradshaw a

a Julius Kruttschnitt Mineral Research Centre, The University of Queensland, Isles Road, Indooroopilly, Brisbane, QLD 4068, Australiab School of Chemical Engineering, The University of Queensland, St. Lucia, Brisbane, QLD 4072, Australia

a r t i c l e i n f o a b s t r a c t

Article history:Received 25 December 2013Revised 14 April 2014Accepted 24 April 2014Available online xxxx

Keywords:RegrindingTumbling millStirred millParticle breakage mechanismFlotationChalcocite

Stirred mills have been widely used for regrinding, and are acknowledged to be more energy efficientthan tumbling mills. These two types of mills present different particle breakage mechanisms duringgrinding. In this study, the effect of regrinding by both mills on surface properties and subsequent min-eral flotation was studied, using chalcocite as the mineral example. A rod mill and a stirred mill with thesame stainless steel media were used to regrind rougher flotation concentrates. Different chalcocite flo-tation recovery was achieved in the cleaner stage after regrinding in tumbling and stirred mills. The fac-tors contributing to the different recovery included particle size, the amount of created fresh surfaces,surface oxidation and the redistribution of collector carried from rougher flotation. All the factors wereexamined. It was determined that the predominating factor was the different distribution of collectorresulting from different particle breakage mechanisms in the stirred and tumbling mills, in line withToF-SIMS analysis. In the tumbling mill, the impact particle breakage mechanism predominates, causingthe collector to remain on the surface of newly produced particles. In the stirred mill, the attrition break-age removes collector from the surface, and decreases particle floatability. Furthermore, the type ofgrinding media in the stirred mill also influences the subsequent flotation, again due to the change of par-ticle breakage mechanisms. The results of this study demonstrate that the selection of regrinding millsand grinding media should not only depend on the required energy efficiency, but also on the propertiesof the surfaces produced for subsequent flotation.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Regrinding rougher flotation concentrates is a common practiceto improve the recovery and grade of valuable minerals as a resultof the need to process low grade and complex ores. For example, inNewcrest’s Telfer gold mine, the regrinding of copper rougherconcentrates significantly improves copper and gold recovery,especially when processing the West Dome ores containing coppersulphides (e.g., chalcocite and bornite etc.) with a small grain size(Seaman et al., 2012). The small grain size means finer grinding isneeded to provide sufficient mineral liberation. In many mineralprocessing plants treating low grade ores, mineral liberation canonly be achieved at a regrind size of less than 10 lm (Johnson,2006).

Grinding is the single largest energy-consuming process in min-eral processing plants, and the selection of energy efficient mills iscritically important for fine grinding. The stirred mills, recentlyintroduced to mineral processing, have proved to be more energyefficient than tumbling mills in terms of fine grinding, and havebeen extensively used in many mineral processing plants at theregrinding stage (Gao et al., 2002; Jankovic, 2003; Pease et al.,2006). In contrast to tumbling mills, where motion is imparted tothe charge via the rotating mill shell, stirred mills impart themotion to the charge by the movement of an internal stirrer, pro-viding different particle breakage mechanisms. In tumbling mills,impact breakage from the free-fall motion of grinding media isthe main breakage mechanism with some attrition existing at thebottom of the mill (Wills and Napier-Munn, 2006). In stirred mills,the movement of the stirrer through the ball bed and the sliding/rolling motion of the charge provides a solely attrition basedbreakage environment (Wills and Napier-Munn, 2006). Recentstudies suggest that impact breakage mechanisms may also existin stirred mills, and the proportion of impact and attrition variesfrom case to case (Kwade and Schwedes, 2002; Yue and Klein,

Miner.

2 X. Chen et al. / Minerals Engineering xxx (2014) xxx–xxx

2005; Roufail and Klein, 2010). The different breakage mechanismshave been proved to influence the particle size distribution (Kellyand Spottiswood, 1982; Gao and Forssberg, 1995; Hogg, 1999),particle shape (Andreatidis, 1995; Roufail and Klein, 2010) andmineral liberation (Andreatidis, 1995; Roufail and Klein, 2010).Furthermore, these parameters may play a role in the subsequentflotation.

The grinding chemistry, including the use of different types ofgrinding media and pulp chemistry (pH, Eh and dissolved oxygen),have a significant effect on the subsequent flotation behaviour(Johnson, 2002; Peng et al., 2003a,b; Bradshaw et al., 2006;Ekmekci et al., 2006; Bruckard et al., 2011; Chen et al., 2012a,b),and different grinding mills used may affect the subsequent flota-tion not only through changes in particle size and shape, but alsothrough the change of mineral surface properties. As demonstratedby Chen et al. (2013), there are two types of surfaces after regrind-ing: the remaining surfaces carried over from the regrind feed, andthe fresh surfaces generated during regrinding. In terms of regrind-ing rougher flotation concentrates, the surfaces carried from roughflotation concentrates are generally covered by collector resultingin a certain degree of residual floatability. The distribution of col-lector on the new surfaces is influenced by the particle breakagemechanisms. As shown in Fig. 1, if an impact or compressionbreakage is applied, surface collector may be evenly distributedon the new particle surfaces. If the attrition breakage is applied,the collector may be removed from the surfaces and distributedonto fine and ultra-fine particles. Ye et al. (2010a) found that thefloatability of coarse particles decreased, to a greater extent, afterregrinding by a stirred mill rather than a tumbling mill. This couldbe due to the greater contribution of the mechanism to size reduc-tion provided by the stirred mill. However, this phenomenon wasonly observed for the regrinding of relatively coarse particles i.e.from P80 of 80–60 lm. Ye et al. (2010a) proposed that breakagemechanisms were not influential for regrinding of finer products.

Besides the surfaces carried over from regrind feed, a largeamount of fresh surfaces may also be produced during regrinding.Chen et al. (2013) reported that the large amount of fresh surfacesafter regrinding strongly depressed pyrite flotation in the cleanerstage. However, it is important to note that sulphide mineralscan be easily oxidized during regrinding, and oxidation speciesformed on the surfaces can markedly change the mineral floatabil-ity (Smart, 1991; Gonçalves et al., 2003; Bicak and Ekmekci, 2012;Chen et al., 2014). Some species are hydrophobic, such as

Fig. 1. The three different proposed particle breakage mechanisms, and the resultantSpottiswood, 1982).

Please cite this article in press as: Chen, X., et al. The effect of particle breakageEng. (2014), http://dx.doi.org/10.1016/j.mineng.2014.04.020

metal-deficient sulphide and polysulphide, improving the particlefloatability, while some species are hydrophilic, such as metalhydroxide and sulphate species, decreasing the particle floatability.Therefore, the different types of grinding mills may affect mineralflotation through different oxidizing conditions on and above theeffects due to breakage mechanisms.

In this study, the effect of particle breakage mechanisms onchalcocite flotation was investigated. Two types of grinding mills,a tumbling mill and a stirred mill, were used to produce groundchalcocite via different breakage mechanisms. The same type ofgrinding media, stainless steel, was used in the two mills in orderto eliminate the potential influence of the type of grinding mediaon flotation. Furthermore, as ceramic media is widely used in stir-red mills commercially, it was also used in this study to compareresults obtained with stainless steel media.

2. Experimental

2.1. Materials and reagents

Chalcocite single mineral, supplied by GEO Discoveries, wascrushed through a jaw crusher and a roll crusher, and thenscreened to collect the – 3.35 + 0.71 mm particle size fraction.XRD analysis indicated that the chalcocite sample had a high puritywith a minor amount (<2%) of iron sulphide impurity. The pro-cessed feed samples were sealed in polyethylene bags and thenstored in a freezer at a temperature of �20 �C to avoid further sur-face oxidation.

Potassium amyl xanthate (PAX) and Interfroth 56 were used asa collector and frother, respectively. Collector and frother were ofindustry grade and were used as received. The pH was adjustedby the addition of a NaOH solution. De-ionized water was usedin all experiments. All chemical solutions were made fresh daily.

2.2. Grinding and flotation

The crushed chalcocite (100 g) was combined with 150 ml ofde-ionized water, and ground in a stainless steel rod mill (Length:260 mm, Diameter: 205 mm) for 8.3 min using 4 stainless steelrods (3750 g) to achieve a P80 of 75 lm. A certain amount of 2.5%w/v sodium hydroxide solution was added in the feed before grind-ing to achieve pH 9.0 in the mill discharge.

distribution of surface collector on broken particles. (Revised based on Kelly and

mechanisms during regrinding on the subsequent cleaner flotation. Miner.

Fig. 2. Chalcocite recovery as a function of flotation time in rougher flotation.

X. Chen et al. / Minerals Engineering xxx (2014) xxx–xxx 3

After grinding, the pulp was transferred to the flotation cell(1.5 dm3). A JK laboratory batch flotation cell with a bottom-drivenagitator was used in this study. Collector (160 g/t) and frother(200 g/t) were added and 2 min of conditioning time was allowedfor the conditioning of each reagent. During flotation, the pH wasfixed at 9.0 by adding a sodium hydroxide solution (2.5% w/v).The froth was scraped every 10 s, and four concentrates were col-lected after cumulative times of 0.5, 2.0, 4.0 and 8.0 min. The airflow rate was 1.5 dm3/min during the first 0.5 min, and thenincreased to 3.0 dm3/min thereafter. Flotation time was based oncomparable plant recovery that could be obtained in thelaboratory.

The four flotation concentrates were combined and mixed withadditional water to achieve a pulp density of 18%, and thenreground. A rod mill and stirred mill were used for the regrinding.The target particle size of the regrind product was P80 = 20 lm. Forregrinding by the rod mill, 10.3 kg stainless steel rods were used,and the desired particle size of P80 = 20 lm was achieved aftergrinding for 21 min. The stirred mill used in this study was a ver-tical bead mill with a disc-type agitator, and the volume of thegrinding chamber was 1.5 L. Two different types of grinding media,stainless steel beads and ceramic beads, were used. The diameterof both media was 2.5 mm, and 1 L of media was added duringregrinding. The rotating speed of the stirrer was 1200 rpm. Thegrinding time was 7 min and 6 min when ceramic media and stain-less steel media were used, respectively, to again produce theproduct particle size of P80 = 20 lm. When N2 was applied duringregrinding, the regrind feed was firstly purged by N2 to reducethe DO to zero before transferring it into the grind mill. Then themill was purged again by N2 for 10 min to expel the air. The sizedistribution was determined by Laser Diffraction with a MalvernMasterSizer (Malvern Instrument Ltd., U.K.). A sodium hydroxidesolution (2.5%) was used to maintain pH 9.0 in the regrindingdischarge.

After regrinding, the pulp was transferred to a 1.5 dm3 flotationcell. Frother (200 g/t) was added during the conditioning of 2 min.For some tests, more collector was added. The procedure was thesame as used in the rougher flotation after primary grinding.

2.3. XPS analysis

Samples for XPS surface analyses (about 10 ml of slurry) werecollected from selected mill discharges. The slurry samples werefrozen in liquid nitrogen to avoid surface oxidation.

XPS measurements were carried out with a KRATOS Axis Ultra(Kratos Analytical, Manchester, United Kingdom) with a mono-chromatic Al X-ray source operating at 15 kV and 10 mA(150 W). The analysis spot size was 300 � 700 lm. The frozenslurry samples were defrosted just prior to the analysis. The solidswere placed on a stainless steel bar and immediately loaded intothe introduction chamber of the spectrometer. The samples wereanalysed at a pressure of 9 � 10�10 Torr at room temperature. Eachanalysis started with a survey scan from 0 to 1200 eV using a passenergy of 160 eV at steps of 1 eV with one sweep. High resolutionspectra of O 1s, C 1s, S 2p, and Cu 2p were collected at 20 eV passenergy at steps of 100 meV with two or three sweeps. Bindingenergies were charge-corrected by referencing to adventitious car-bon at 284.8 eV.

2.4. ToF-SIMS analysis

Time of flight secondary ion mass spectrometry (ToF-SIMS) wasused to study the distribution of collector on different size frac-tions after regrinding. Samples for ToF-SIMS surface analysis werethe same as used for XPS analysis as described in Section 2.3. Theinstrument used in this work is a PHI TRIFT V nanoTOF equipped

Please cite this article in press as: Chen, X., et al. The effect of particle breakageEng. (2014), http://dx.doi.org/10.1016/j.mineng.2014.04.020

with a pulsed liquid metal 79+Au primary ion gun (LMIG), operatingat 30 kV energy. ‘‘Unbunched’’ beam settings were used to opti-mize spatial resolution. Surface analyses, in positive and negativeSIMS modes, were performed at a number of locations typicallyusing a 75 � 75 lm raster area. For the purposes of statistical inter-rogation, generally ca. 25 particles of interest were imaged persample to collect representative data. Region-of-interest analyseswere performed on the collected raw image data, which involvedthe extraction of mass spectra specifically from within the bound-aries of the particles of interest. Resulting spectra in each polaritywere calibrated using WincadenceN software (Physical ElectronicsInc.) and peaks were selected based upon previously identified spe-cies of interest. Integrated peak values of the selected ions werenormalised to the total selected secondary ion intensities, tocorrect for differences in total ion yield between analyses andsamples. The resulting data were then compared qualitatively bypreparing plots of average normalised counts (with 95% confidenceintervals) for the collector species.

3. Results and discussion

Fig. 2 shows chalcocite flotation recovery as a function of flota-tion time after primary grinding with 92 wt.% chalcocite recoveryachieved after 8 min. The rougher flotation concentrates werereground in different mills, as described earlier, and chalcocite flo-tation in the cleaner stage was then examined.

The rougher flotation concentrates (about 90 g chalcocite) werereground by two different grinding mills: a rod mill and a stirredmill. In the rod mill, stainless steel rods were used as grindingmedia, while in the stirred mill, both stainless steel and ceramicmedia were used. In the subsequent cleaner flotation, no additionalcollector was added. Flotation results are shown in Fig. 3. Chalco-cite cleaner recovery after regrinding in a rod mill was 59 wt.%after 8 min of flotation. It was only 22 wt.% and 30 wt.% respec-tively after regrinding by stirred mill with stainless steel and cera-mic media. Therefore, chalcocite flotation was strongly depressedafter regrinding, and regrinding conditions significantly influencedthe flotation behaviour. The rod mill produced better chalcociteflotation than the stirred mill, and when the stirred mill was used,ceramic media produced better chalcocite flotation than stainlesssteel media.

There are several factors that could affect chalcocite flotationafter regrinding in this study:

(1) Particle size. The particle size was reduced from P80 75 lmto 20 lm during regrinding, and different grinding condi-tions may generate different size distributions.

mechanisms during regrinding on the subsequent cleaner flotation. Miner.

Fig. 3. Chalcocite cleaner recovery as a function of flotation time after regrindingwith different grinding mills and grinding media.

Fig. 4. Particle size distribution after regrinding in rod mill and stirred mill.

Table 1Surface area of the regrind feed and product, and percentage of fresh surfaces.

Regrindfeed

Rod millproduct(stainlesssteel)

Stirred millproduct(stainlesssteel)

Stirred millproduct(ceramic)

Size (lm) 75 20 20 20Surface area (m2/g) 0.2 1.3 1.6 1.4Percentage of fresh surfaces (%) 85% 88% 86%

Fig. 5. S 2p XPS spectra of mineral surfaces of regrind feed and product.

Table 2S 2p quantification for the regrind feed and product.

Area contribution (%)

Regrindfeed

Rod mill product(stainless steel)

Stirred mill product(stainless steel)

Stirred millproduct (ceramic)

S2� 75.8 79.5 100 100Sn

2� 6.0 4.1 – –SO4

2� 13.2 11.4 – –

4 X. Chen et al. / Minerals Engineering xxx (2014) xxx–xxx

(2) Particle shape. Different particle shape may be producedafter regrinding in rod and stirred mills. However, asreported by Vizcarra et al. (2011), particle shape only influ-ences the flotation of minerals with slow flotation kinetics(e.g., with no collector addition, for gangue minerals). There-fore, this factor was not considered in this study.

(3) Fresh surfaces. After regrinding, a large amount of fresh sur-faces are produced, which may change particle floatability.

(4) Surface oxidation. The oxidation of mineral surfaces duringregrinding in different mills may be different and then influ-ence particle floatability.

Please cite this article in press as: Chen, X., et al. The effect of particle breakageEng. (2014), http://dx.doi.org/10.1016/j.mineng.2014.04.020

(5) The distribution of collectors. The collectors carried thoughfrom rougher flotation may have a different distribution onparticle surfaces after regrinding by rod and stirred millsand with different media. This may influence the subsequentflotation.

In the following sections, these four factors were examined.

3.1. Particle size

The particle size was reduced from P80 = 75 lm to P80 = 20 lmafter regrinding. It is known that fine particles show low floatabil-ity due to low bubble–particle collision efficiency (Trahar, 1981).However, a great number of studies indicate that particles at20 lm still show maximum flotation recovery given sufficient sur-face hydrophobicity (Grano, 2009; Pease et al., 2010). Therefore,the low chalcocite flotation after regrinding in this study may bemore to do with the altered surface properties resulting from theparticle size reduction process rather than with the reduced parti-cle size itself.

The particle size distributions after regrinding are shown inFig. 4. With the same target P80 = 20 lm, the size distributionswere similar. The stirred mill, especially with stainless steel media,

mechanisms during regrinding on the subsequent cleaner flotation. Miner.

Fig. 6. Cu 2p XPS spectra of mineral surfaces of regrind feed and product.

Table 3Cu 2p quantification for the regrind feed and product.

Area contribution (%)

Regrindfeed

Rod millproduct(stainless steel)

Stirred millproduct(stainless steel)

Stirred millproduct(ceramic)

Cu–S 89.3 91.1 92.0 91.4Cu–OOH 10.7 8.9 8.0 8.6

Fig. 7. S 2p XPS spectra of mineral surfaces of regrind product with nitrogenpurging during regrinding.

Fig. 8. Chalcocite recovery as a function of flotation time after regrinding in air(solid line) and in nitrogen (dashed line).

X. Chen et al. / Minerals Engineering xxx (2014) xxx–xxx 5

produced slightly more fine particles (�10 lm) than the rod mill,which may be due to the more attrition mechanism in the stirredmill (Hogg, 1999; Yue and Klein, 2005). It is unlikely that the slightdifference in size distribution produced by the two types of mills isthe major contributing factor to the significant difference in chal-cocite flotation.

3.2. Fresh surfaces

Two different types of particle surfaces are generated afterregrinding: the surfaces carried over from the grind feed, and thesurfaces freshly produced after regrinding (Chen et al., 2013). Thesurface area of the feed and products was measured using theBET method. As shown in Table. 1, the surface area of the regrindfeed was 0.2 m2/g, and increased to 1.3 m2/g after the regrindingby the rod mill, and 1.6 m2/g and 1.4 m2/g after regrinding by thestirred mill with stainless steel media and ceramic media, respec-tively. The percentage of fresh surfaces produced after regrindingwas calculated and is also shown in Table 1. The results indicatethat more than 80% of the surfaces of the regrind product are fresh.Since the fresh surfaces are presumably not covered by collector,the surface hydrophobicity was significantly diluted after regrind-ing, most likely resulting in the decreased flotation recovery ingeneral. Ye et al. (2010a) reported that the increase in surface areaafter regrinding strongly depressed pyrrhotite flotation. In linewith the particle size distribution in Fig. 4, the slightly larger per-centage of fresh surfaces after stirred milling may result in a lowerflotation recovery, but this should not be expected to be the major

Please cite this article in press as: Chen, X., et al. The effect of particle breakageEng. (2014), http://dx.doi.org/10.1016/j.mineng.2014.04.020

contributing factor to the different chalcocite flotation afterregrinding by rod and stirred mills due to the small differencesin fresh surfaces.

3.3. Oxidation of mineral surfaces

The oxidation species of chalcocite at alkaline conditionsinclude Cu(OH)2, metal-deficient sulphide or polysulphide, ele-mental sulphur, and sulphur species of higher oxidation state, suchas sulphate (Walker et al., 1984). Metal-deficient sulphide, elemen-tal sulphide and polysulphide are hydrophobic, and improve chal-cocite floatability. Other oxidation species such as Cu(OH)2 andsulphate are hydrophilic, reducing chalcocite flotation. To detectthe species on the chalcocite surface, selected regrinding feedand products were examined by XPS. The S 2p XPS spectra areshown in Fig. 5. These spectra were fitted using the 2p1/2 and

mechanisms during regrinding on the subsequent cleaner flotation. Miner.

Fig. 9. Chalcocite flotation recovery as a function of particle size distribution afterregrinding.

6 X. Chen et al. / Minerals Engineering xxx (2014) xxx–xxx

2p3/2 doublet with a fixed 1:2 intensity ratio and 1.18 eV energyseparation. The doublet at 161.8 eV is assigned to S2� from Cu2S(Smart et al., 1999). The doublet at 162.8 eV can be attributed tothe formation of polysulphide. The broad peak observed at167.7 eV is attributed to the sulphate SO4

2�. An energy loss (EL)peak was also fitted at 164.0 eV (Chen et al., 2014). The quantifica-tion of the regrinding feed and product is summarized in Table 2. Asignificant amount of polysulphide and sulphate species weredetected in both regrind feed and regrind product from the rodmill, which indicates oxidation on the surfaces. However, afterregrinding by the stirred mill, less oxidation species were detected.Apparently, the rod mill produced a more oxidizing condition thanthe stirred mill.

The surface copper species of the regrind feed and product werealso examined by XPS. The Cu 2p XPS spectra are shown in Fig. 6.The Cu+ Cu 2p3/2 component was identified at about 932.6 eV(McIntyre and Cook, 1975; Deroubaix and Marcus, 1992). A secondcomponent in the Cu 2p3/2 at about 934.2 eV is attributable toCu2+-hydroxyl species (Weisener and Gerson, 2000). Quantificationof the regrind feed and product surface species is summarized inTable 3. The compositions of copper species after regrinding by dif-ferent mills with different media vary only slightly. A certainamount of Cu2+-hydroxyl species was detected in all the samples,which indicates copper oxidation under all regrinding conditions.

Fig. 10. ToF-SIMS normalised intensity of collector fragments (potassium amyl xanthatemill with stainless steel media (a) and stirred mill with stainless steel media (b).

Please cite this article in press as: Chen, X., et al. The effect of particle breakageEng. (2014), http://dx.doi.org/10.1016/j.mineng.2014.04.020

To investigate further the effect of oxidation on chalcocite flota-tion, nitrogen purging was applied in both mills to provide a reduc-ing regrinding condition. The concentration of dissolved oxygen ofthe regrinding product dropped to zero after applying the nitrogenpurging. In addition, the surface change with nitrogen purging wasexamined by XPS analysis. Fig. 7 shows the S 2p spectra afterapplying nitrogen purging. It is obvious that oxidation was signif-icantly reduced during regrinding in the rod mill.

Cleaner flotation was conducted on reground product to inves-tigate the effect of nitrogen purging during regrinding on chalco-cite flotation. As shown in Fig. 8, flotation recovery was similarwith or without nitrogen purging during regrinding. It suggeststhat the different flotation behaviour after regrinding in differentmills is not primarily caused by oxidizing conditions.

3.4. The distribution of collectors on particle surfaces

As discussed in the previous sections, the particle size, freshsurface produced and surface oxidation only have a slight influenceon the subsequent flotation. It is more likely that the significantdifference in chalcocite flotation, after regrinding by rod and stir-red mills with different media, can be attributed to the re-distribu-tion of collector species. To further prove this hypothesis, theflotation recovery of the particles with different size fractions afterregrinding was evaluated. As shown in Fig. 9, the recovery of theparticles at different size fractions is similar after regrinding bythe rod mill, but overall much higher than the recovery afterregrinding by the stirred mill. In addition, after regrinding by thestirred mill, the floatability of particles decreased with an increasein particle size. The fine and ultra-fine particles displayed betterflotation than the intermediate particles.

The size-by-size flotation data suggests that the collector wasdistributed evenly on the particles at different size fractions afterregrinding by the rod mill, while, after regrinding by the stirredmill, more collector remained on the fine and ultra-fine particlesthan on the intermediate particles. The collector distribution maybe linked to particle breakage mechanisms. As shown in Fig. 1, ifthe impact breakage and compression breakage are applied, thecollector remains on the new particle surfaces. However, if theattrition breakage is applied, the collector is removed from the par-ticle surfaces, and distributed onto the fine and ultra-fine particles.The impact breakage is dominant during regrinding by the rod mill,therefore, the floatability of particles is still high at all size frac-tions. However, the attrition breakage predominates in the stirredmill, and removes the collector from particle surfaces, resulting in

) on particle surfaces of chalcocite at different size fractions after regrinding in rod

mechanisms during regrinding on the subsequent cleaner flotation. Miner.

X. Chen et al. / Minerals Engineering xxx (2014) xxx–xxx 7

reduced flotation recovery in general. Fine and ultra-fine particlesare produced by attrition, and carry the surface collector, resultingin better flotation recovery for them compared with the intermedi-ate particles.

ToF-SIMS was used in this study to investigate the collectordistribution on chalcocite particles after regrinding. Fig. 10(a)shows the normalised intensity of collector fragments (potassiumamyl xanthate) on particles of three different size fractions afterregrinding by rod mill with stainless steel media. There is no statis-tical difference between the intensity of collectors on particles ofthree size fractions, which proves the collector was evenly distrib-uted on the particles after regrinding by rod mill. However, thecollector intensity on the fine particles (<10 lm) is statisticallyhigher than that on the coarser particles (>20 lm) after regrindingby the stirred mill with stainless steel media. This is likely due tothe attrition breakage in the stirred mill, which removes the sur-face collectors from coarse particles as illustrated in Fig. 1. TheseToF-SIMS results are in accordance with the size by size flotationresults as shown in Fig. 9, further confirming that the differentparticle breakage imparted by rod and stirred mills influencesthe collector distribution.

It has been reported that grinding media in a stirred mill caninfluence the stress intensity through media density (Jankovic,2003). In this study, the recovery after regrinding with stainlesssteel media was lower than that with ceramic media. It seems thatmore attrition breakage was created when stainless steel mediawas used with higher stress intensity, corresponding to a slightlylarger particle surface area as shown in Table 1.

In summary, particle size, fresh surfaces and the surface oxida-tion do not play an important role in the different flotation behav-iour after regrinding in different milling conditions. The maincontributing factor appears to be collector distribution governedby the particle breakage mechanisms.

3.5. Flotation with additional collector

Since a large amount of fresh surfaces are produced afterregrinding, it is reasonable to expect that additional collector inthe cleaner flotation stage may improve chalcocite flotation recov-ery. Flotation results after adding more collector in the cleanerflotation are shown in Fig. 11. After regrinding in both stirredand rod mills, chalcocite recovery was increased with the increasedcollector dosage. The recovery reached the maximum at 160 g/tcollector after regrinding by the rod mill. More collector wasrequired after regrinding by the stirred mill to restore the recoveryto the same level. This may be associated with the removal of

Fig. 11. Chalcocite recovery after regrinding as a function of collector dosage.

Please cite this article in press as: Chen, X., et al. The effect of particle breakageEng. (2014), http://dx.doi.org/10.1016/j.mineng.2014.04.020

collector carried over from rougher flotation concentrate by attri-tion breakage. The recovery after regrinding by stirred mill withstainless steel media was always lower than that with ceramicmedia, which may be because more collector was removed byattrition when using high density stainless steel media.

Chalcocite recovery as a function of particle size after collectoraddition was further investigated. Results are shown in Fig. 12.After regrinding by the rod mill, chalcocite recovery was increasedwith collector addition in all size fractions, and the percentageincrease was similar. After regrinding in the stirred mill, the fineand ultra-fine particles still displayed higher recovery than

Fig. 12. Chalcocite flotation recovery as a function of particle size distribution inthe presence of additional amount of collectors after regrinding in rod mill withstainless steel media (a), in stirred mill with stainless steel media (b), and in stirredmill with ceramic media (c).

mechanisms during regrinding on the subsequent cleaner flotation. Miner.

8 X. Chen et al. / Minerals Engineering xxx (2014) xxx–xxx

intermediate particles at the same collector addition. The floatabil-ity of coarser particles was more difficult to be restored, probablydue to more denuded surfaces by attrition. The restoration ofrecovery was more difficult in all size fractions when using stain-less steel media in the stirred mill compared with using ceramicmedia.

4. Conclusions

A rod mill with stainless steel media and a stirred mill withstainless steel and ceramic media were used to regrind chalcociterougher flotation concentrates so as to examine chalcocite flotationrecovery in the cleaner stage. The factors which could influence theflotation were investigated, and these included: the particle size,the fresh surfaces produced, the oxidation of particle surfaces,and the distribution of collector carried over from rougherflotation.

The major contributing factor to different chalcocite flotationobserved was found to be the distribution of collector carried overfrom rougher flotation, which was in turn controlled by the particlebreakage mechanism. In the rod mill, the main breakage mecha-nism is impact, and the collector can remain on freshly producedsurfaces. In the stirred mill, the attrition breakage predominatesand tends to remove the collector from surfaces onto fine andultra-fine particles resulting in lower flotation recovery, especiallywith coarser size fractions. Furthermore, stainless steel and cera-mic media in the stirred mill produced different chalcocite flota-tion in the cleaner stage. This may be associated with a differentproportion of attrition breakage mechanism as well.

The chalcocite flotation recovery can be improved by the addi-tion of more collector in the cleaner stage. More collector additionis required after regrinding in the stirred mill than in the rod millto restore the flotation recovery to the same level.

Acknowledgments

The authors thank Dr. John Denman for his assistance with per-forming the ToF-SIMS analysis. The financial support fromNewcrest Mining Limited and the Australia Research Council isgratefully acknowledged. The discussion of this study with Dr.David Seaman, the technical expert from Newcrest Mining Limited,is greatly appreciated. The first author also thanks the scholarshipprovided by the University of Queensland to assist with undertak-ing this study.

References

Andreatidis, J.P., 1995. Breakage mechanisms and resulting mineral liberation in abead mill. MEngSc, Thesis, The University of Queensland.

Bicak, O., Ekmekci, Z., 2012. Prediction of flotation behavior of sulphide ores byoxidation index. Miner. Eng. 36–38, 279–283.

Bradshaw, D.J., Buswell, A.M., Harris, P.J., Ekmekci, Z., 2006. Interactive effects of thetype of milling media and copper sulphate addition on the flotationperformance of sulphide minerals from Merensky ore Part I: Pulp chemistry.Int. J. Miner. Process. 78, 153–163.

Bruckard, W.J., Sparrow, G.J., Woodcock, J.T., 2011. A review of the effects of thegrinding environment on the flotation of copper sulphides. Int. J. Miner. Process.100, 1–13.

Chen, X., Peng, Y., Bradshaw, D., 2012a. Effect of Regrinding Conditions on theRejection of Pyrite in the Cleaner Stage. In: 26th International MineralProcessing Congress, New Delhi, India.

Chen, X., Peng, Y., Bradshaw, D., 2012b. Effect of Regrinding Conditions on theFlotation of Chalcopyrite and its Separation from Pyrite in the Cleaner Stage. In:11th AusIMM Mill Operators’ Conference, Hobart, Tasmania.

Please cite this article in press as: Chen, X., et al. The effect of particle breakageEng. (2014), http://dx.doi.org/10.1016/j.mineng.2014.04.020

Chen, X., Peng, Y., Bradshaw, D., 2013. Effect of regrinding conditions on pyriteflotation in the presence of copper ions. Int. J. Miner. Process. 125, 129–136.

Chen, X., Peng, Y., Bradshaw, D., 2014. The separation of chalcopyrite and chalcocitefrom pyrite in cleaner flotation after regrinding. Miner. Eng. 58, 64–72.

Deroubaix, G., Marcus, P., 1992. X-ray photoelectron spectroscopy analysis ofcopper and zinc oxides and sulphides. Surf. Interface. Anal. 18, 39–46.

Ekmekci, Z., Bradshaw, D.J., Harris, P.J., Buswell, A.M., 2006. Interactive effects of thetype of milling media and CuSO4 addition on the flotation performance ofsulphide minerals from Merensky ore Part II: froth stability. Int. J. Miner.Process. 78, 164–174.

Gao, M., Forssberg, E., 1995. Prediction of product size distributions for a stirred ballmill. Powder Technol. 84, 101–106.

Gao, M., Young, M., Allum, P., 2002. IsaMill Fine Grinding Technology and itsIndustrial Applications at Mount Isa Mines. In: Proceedings 34th AnnualMeeting of the Canadian Mineral Processors, pp. 171–188.

Gonçalves, K.L.C., Andrade, V.L.L., Peres, A.E.C., 2003. The effect of grindingconditions on the flotation of a sulphide copper ore. Miner. Eng. 16, 1213–1216.

Grano, S., 2009. The critical importance of the grinding environment on fine particlerecovery in flotation. Miner. Eng. 22, 386–394.

Hogg, R., 1999. Breakage mechanisms and mill performance in ultrafine grinding.Powder Technol. 105, 135–140.

Jankovic, A., 2003. Variables affecting the fine grinding of minerals using stirredmills. Miner. Eng. 16, 337–345.

Johnson, N.W., 2002. Practical Aspects of the Effect of Electrochemical Conditions inGrinding Mills on the Flotation Process. Flotation and Flocculation-fromFundamentals to Applications, Kailua-Kona, Hawaii, USA.

Johnson, N.W., 2006. Liberated 0–10 lm particles from sulphide ores, theirproduction and separation-Recent developments and future needs. Miner.Eng. 19, 666–674.

Kelly, Z.G., Spottiswood, D.J., 1982. Introduction to Mineral Processing. Wiley, NewYork.

Kwade, A., Schwedes, J., 2002. Breaking characteristics of different materials andtheir effect on stress intensity and stress number in stirred media mills. PowderTechnol. 122, 109–121.

McIntyre, N.S., Cook, M.G., 1975. X-ray photoelectron studies on some oxides andhydroxides of cobalt, nickel, and copper. Anal. Chem. 47, 2208–2213.

Pease, J.D., Curry, D.C., Young, M.F., 2006. Designing flotation circuits for high finesrecovery. Miner. Eng. 19, 831–840.

Pease, J.D., Young, M.F., Curry, D., Johnson, N.W., 2010. Improving fines recovery bygrinding finer. Trans. Inst. Min. Metall., Sect. C: Miner. Process. Extr. Metall. 119,216–222.

Peng, Y., Grano, S., Fornasiero, D., Ralston, J., 2003a. Control of grinding conditions inthe flotation of chalcopyrite and its separation from pyrite. Int. J. Miner. Process.69, 87–100.

Peng, Y., Grano, S., Fornasiero, D., Ralston, J., 2003b. Control of grinding conditions inthe flotation of galena and its separation from pyrite. Int. J. Miner. Process. 70,67–82.

Roufail, R., Klein, B., 2010. Mineral liberation and particle breakage in stirred mills.Can. Metall. Quart. 49, 419–428.

Seaman, D.R., Burns, F., Adamson, B., Manton, P., 2012. Telfer Processing PlantUpgrade – The Implementation of Additional Cleaning Capacity and theRegrinding of Copper and Pyrite Concentrates. In: 11th AusIMM MillOperators’ Conference, Hobart, Tasmania.

Smart, R.S.C., 1991. Surface layers in base metal sulphide flotation. Miner. Eng. 4,891–909.

Smart, R.S.C., Skinner, W.M., Gerson, A.R., 1999. XPS of sulphide mineral surfaces:metal-deficient, polysulphides, defects and elemental sulphur. Surf. Interface.Anal. 28, 101–105.

Trahar, W.J., 1981. A rational interpretation of the role of particle size in flotation.Int. J. Miner. Process. 8, 289–327.

Vizcarra, T.G., Harmer, S.L., Wightman, E.M., Johnson, N.W., Manlapig, E.V., 2011.The influence of particle shape properties and associated surface chemistry onthe flotation kinetics of chalcopyrite. Miner. Eng. 24, 807–816.

Walker, G.W., Stout Iii, J.V., Richardson, P.E., 1984. Electrochemical flotation ofsulfides: reactions of chalcocite in aqueous solution. Int. J. Miner. Process. 12,55–72.

Weisener, C., Gerson, A., 2000. Cu(II) adsorption mechanism on pyrite: an XAFS andXPS study. Surf. Interface. Anal. 30, 454–458.

Wills, B.A., Napier-Munn, T.J., 2006. Wills’ Mineral Processing Technology: AnIntroduction to the Practical Aspects of ore Treatment and Mineral Recovery.Butterworth-Heinemann, Oxford.

Ye, X., Gredelj, S., Skinner, W., Grano, S.R., 2010a. Regrinding sulphide minerals –breakage mechanisms in milling and their influence on surface properties andflotation behaviour. Powder Technol. 203, 133–147.

Yue, J., Klein, B., 2005. Particle breakage kinetics in horizontal stirred mills. Miner.Eng. 18, 325–331.

mechanisms during regrinding on the subsequent cleaner flotation. Miner.