Flotation of A Pyritic Gold Ore PDF
Flotation of A Pyritic Gold Ore PDF
Flotation of A Pyritic Gold Ore PDF
a r t i c l e i n f o a b s t r a c t
Article history: Kinetics of semi-batch flotation behavior of a gold ore from North-Western Quebec, Canada was investigated with
Received 9 June 2010 respect to fineness of grind (P80). The gold was mostly associated with pyrite in native form as inclusions and was
Received in revised form 11 October 2010 also in contact with pyrite along the boundaries of non-sulphide gangue. Gold recoveries were found to increase
Accepted 16 October 2010
from 91.8% to 95.8% as the particles size decreased from an F80 of 205 μm to 53 μm. Data treatment according to
Available online 19 November 2010
the graphical method of kinetic analysis indicated a much better fit to second order kinetics for both gold and
Keywords:
pyrite with correlation coefficients higher than 0.998 compared to first order kinetics with correlation coefficients
Gold ore flotation of less than 0.95. Variation and implication of flotation rate constants are discussed with respect to chemical
Pyrite flotation conditions and particle size.
Flotation kinetics © 2010 Elsevier B.V. All rights reserved.
PAX
IPX
Particle size
0301-7516/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.minpro.2010.10.005
E. Yalcin, S. Kelebek / International Journal of Mineral Processing 98 (2011) 48–54 49
2. Experimental Table 1
Chemical analysis of ore sample.
Fig. 1. Photomicrographs of the mode of occurrence of gold at the Barry deposit. (A) Photomicrograph of free gold grain in siliceous gangue and pyrite (grey areas are finely disseminated
illmenite and magnetite). The photo on the right is a close up view of the same gold grain (50 μm), (B) Photomicrograph of gold occurring in contact with pyrite as free gold in gangue and
strong biotite alteration. The photo on the right (b) is a close up view of the 10 gold grains of 1–10 μm (Lariviere, 1997).
50 E. Yalcin, S. Kelebek / International Journal of Mineral Processing 98 (2011) 48–54
mill for each test. The transfer of slurry into the cell for each test was kept (1957). The following can be written for formation of gold ethyl
constant at 5 min. xanthate:
þ − 29
2.4. Reagents Au + X ↔ AuXs K = 1:67*10 ð3Þ
The pH regulator used in grinding at a single dosage was soda ash. The interaction of xanthates with iron species is of interest with
The collector used in flotation tests was a mixture of two xanthates, respect to flotation of pyrite. The chemical interaction between ferrous
namely PAX (i.e., potassium amyl xanthate) and SIPX (sodium isopropyl iron and ethyl xanthate has a much smaller stability constant
xanthate). Before use in the experimental program, both xanthates were (Kakovsky, 1957):
purified according to standard procedures in the literature (Rao, 1971)
2 + − 7
and kept under petroleum ether in a dark glass bottle. When needed, an Fe + 2X ↔ FeðX2 Þs K = 1:58*10 ð4Þ
appropriate amount of xanthate was pipetted out onto a watch glass to
allow evaporation of the ether in a fume hood, then weighed and Interaction of ferric iron species is more important since flotation
dissolved in deionized water to prepare a fresh aqueous solution at 0.1% often takes place in an oxidizing environment. Critchley and Hunter
wt strength. Frother was MIBC (methyl iso butyl carbinol), which was (1986) reported the solubility products of several ferrix xanthates,
used at its full strength in the form of droplets. One third of xanthate was based on which the following is valid for ferric ethyl xanthate:
added at the grinding stage. The remaining dosage of xanthates and
3 + − 24
MIBC was equally distributed before each stage. The overall dosage Fe + 3X ↔ FeðX3 Þs K = 6:31*10 ð5Þ
levels of the reagents are shown in Table 2.
It can be seen that reactivity of xanthate with gold particles has a
2.5. Flotation tests stronger thermodynamic driving force compared to iron species. Since
pyrite flotation is practiced in alkaline slurries, the interaction of
Batch tests were carried out using standard Denver laboratory rod hydroxylated ferric species with xanthates is also important. For
mill and Denver flotation machine with a 2-L cell. With the amount of example, in an alkaline medium applicable to flotation circuits, a ferric
ore used, the pulp density of the flotation feed became about 38% solids dihydroxy ethyl xanthate can form.
by weight. Following the grinding using 300 g/tonne of soda ash the
3 + − − 34
flotation feed had an initial pH of 8.7 to 9.0. No new addition of soda ash Fe + 2OH + X ↔ FeðOHÞ2 Xs K = 3:98*10 ð6Þ
was made during flotation. Pulp pH levels were monitored during each
test using a portable unit (model 3000) from VWR Scientific, which was Corresponding K value for the ferric dihydroxy amyl xanthate is
equipped with a combination electrode and probe for temperature 3.98 × 1034. Dixanthogen formation on top of such a hydroxy xanthate
compensation. layer on pyrite is also chiefly responsible for the hydrophobicity and
Flotation gas rate was based on the natural suction of air into pulp floatability of this mineral (Wang and Forssberg, 1991).
through the shaft of impeller set at 1200 rpm. Total flotation time was It can be concluded from above that the flotation mechanisms of gold
fixed at 7 min. A total of four concentrates were skimmed off at 0.5, 1.5, 2 and pyrite with xanthates have important common features. Branched
and 3 min with a froth removal frequency of 30 strokes per minute. xanthates are considered to be somewhat better suited for development
Solid products from all separation tests were dried and weighed for of hydrophobicity than the straight chain xanthates, and joint action of
the construction of mass balance tables. Analyses of gold in all products two collectors (adopted in the present work) is generally recognized to
were carried out using the fire assay procedure developed for sulphur be more efficient (Klassen and Mokrousov, 1963). The latter aspect has
bearing ores. In cases where the amount of solids was insufficient for an been demonstrated for a collector mixture of PAX and di-isobutyl
assay ton (i.e., 29.17 g), a concentrate sample was pre-diluted with a gold- dithiophosphate in gold flotation through a surface analytical technique
free granite sample, which was taken into account in mass balancing. (Chryssoulis and Dimov, 2004).
3.1. Interaction of xanthates with gold and pyrite Flotation recoveries of gold obtained are shown in Fig. 2 as a function of
time, which indicates recoveries between 91.8% and 95.8% within 7 min
Xanthates are common surfactants used as collectors in flotation depending on the grind size (P80). These results have been obtained using
recovery of pyrite and other sulphides that are associated with gold. a fixed dosage of soda ash to provide a pH of around 9 that is known to
Interaction of xanthates and gold and pyrite has a strong electrochem- allow high recoveries for both gold and pyrite unless the latter is heavily
ical basis that involves adsorption of xanthate for the formation of a oxidized (Monte et al., 2000). Mass recoveries are also given in this figure.
hydrophobic layer as a first step to be followed by formation of In general, such a low mass recovery is attributable to a low level of
dixanthogen (X2) as an outer hydrophobic layer that is even stronger sulphides in the ore corresponding to about 0.8% (Table 1) as well as the
than the initial adsorbed xanthate layer. strong selectivity of xanthates against the non-sulphide components. The
flotation test at a P80 of 205 μm yielded a mass recovery of about 3.66%,
−
X ↔ Xads + e ð1Þ and corresponded to a gold recovery of 91.8%. This lower recovery may in
general be attributed to floatability problems of coarse particles. The
−
Xads + X ↔ X2ads + e ð2Þ floatability by particle size curves for sulphides indicate significant drops
around 200 μm (Trahar, 1981). However, considering that gold is 4–5
The chemisorption results in formation of a hydrophobic surface times heavier than a sulphide-bearing mineral, a greater decrease in gold
product on gold and iron-bearing species. In a recent study of cycling recovery would be expected at this grind size. That the recovery has not
voltammetry and surface force measurements with gold treated with dropped more significantly is indicative of a strong possibility that the
potassium amyl xanthate (Wang and Yoon, 2009), hydrophobicity of gold particles after grinding are mostly recovered together with the pyrite
gold reached a maximum level corresponding to a contact angle of 94°, component of the ore. Another point is that the samples used contain an
which was attributed to formation of di-amyl dixanthogen on top of the insignificant amount of liberated native gold particles that were too coarse
chemisorbed layer of amyl xanthate. Solubility products for the for the recovery to drop more drastically. In order to examine the first
interaction of gold with various xanthates are given by Kakovsky point, pyrite content of the flotation products were estimated from the
E. Yalcin, S. Kelebek / International Journal of Mineral Processing 98 (2011) 48–54 51
100 6 Table 2
Reagents used in testwork.
70
90
Au Recovery (%)
80
70
60
P80: 53 µm
P80: 92 µm
50
P80: 159 µm
P80: 205 µm
40
40 50 60 70 80 90 100
Pyrite Recovery (%)
Fig. 4. Experimental data plotted according to 1st order flotation kinetics (using two R∞
Fig. 3. Gold recoveries versus estimated pyrite recoveries. values, a and b explained in the text).
52 E. Yalcin, S. Kelebek / International Journal of Mineral Processing 98 (2011) 48–54
Table 3 8.0
Parameters for the first order flotation kinetics of gold in the sample used. P80: 53 µm
7.0 P80: 92 µm
2 2
P80 Rovr R∞ (a) k1 r R∞ (b) k1 r
(μm) (%) (%) (min−1) (%) (min−1) P80: 159 µm
6.0
P80: 205 µm
53 95.8 95.9 1.07 0.9498 98.4 0.64 0.5779
92 94.4 94.5 1.03 0.9399 97.1 0.63 0.5772 5.0
159 92.4 92.5 1.06 0.9447 95.6 0.61 0.5633
t/R
205 91.8 92.0 0.96 0.9400 95.6 0.57 0.6015
4.0
3.0
constant for the sample tested appears to be strikingly small. One factor for
this might be related to a liberal dosage of collector used compared to 2.0
some other studies on gold flotation (Chryssoulis et al., 2003). In general, a
1.0
high dosage of collector tends to even out floatability of particles with
respect to their size (Torem et al., 2006). This is presumably related to 0.0
improvement of the recovery of coarse particles that are known to be 0 1 2 3 4 5 6 7 8
highly sensitive to collector concentration (Trahar, 1981). Industrial Flotation Time, t (minutes)
practice on recovery of precious metals tends to use higher dosages in lieu
of "insurance" given the relatively small cost of reagent. Another factor for Fig. 5. Experimental data plotted according to 2nd order flotation kinetics.
this behavior may be related to use of mixed xanthates, i.e., normal chain,
PAX and branched chain, SIPX, and perhaps more importantly, 30% of their where k2 is the second order rate constant. Re-arranging Eq. (10) one
addition into the grinding mill, where soda ash has been used as the pH can get
regulator during size reduction. Grinding with soda ash as a pH regulator
is beneficial to promote bulk flotation since lime as an alternate is known t t 1
to have some detrimental effect on flotation of pyrite (Gaudin, 1957; = + 2 ð11Þ
R R∞ R∞ k2
Klassen and Mokrousov, 1963) and gold (Dunne, 2005). Use of lime
increases the concentration of calcium ions in the flotation slurry, which
tend to block active sites for collector. In contrast, soda ash can act as a If the second order equation is valid, a plot of “t/R, flotation time/
dispersant (hence helpful for selectivity against gangue) and is able to recovery” against time t (cumulatively for both cases) should result in a
combat adverse effect of excess calcium ions by causing their precipita- straight line. The results plotted accordingly are shown in Fig. 5. As may
tion, which, by itself, might cause a problem in some cases. A detailed be noted, the same data now result in a very good linearity as compared
bench scale comparison on the effects of soda ash and lime is reported with the results of the first order kinetics (Fig. 4). The R∞ values obtained
elsewhere (Kelebek, 2009). First order kinetic analysis was also carried out from slopes and second order rate constants, k2 obtained from the
for pyrite. Results are summarized in Table 4, which indicate similar trends intercept values for each case are tabulated in Table 5. The correlation
to the case reported for gold (Table 3). However, although the rate coefficient (r2) obtained for each case is greater than 0.999 compared to
constants are of the same order of magnitude, there are some 0.949 or less for the first order case (Table 3). Due to this excellent data
inconsistencies, e.g., the rate constant for P80 of 159 μm in the first series fit, the alternate ultimate recoveries for the first order kinetics (i.e., R∞(2)
(k1 = 0.78 min−1) is significantly smaller than that of 205 μm in Table 3) were the R∞ values obtained from the second order model
(k1 =1.16 min−1). In the second series, the rate constants are the same (Table 5) demonstrating sensitivity of k1 and r2 values to variation in the
for the smaller two (0.68 min−1) and for the larger two grind sizes R∞ values. The second order rate constants, k2 appear to vary more
(0.58 min−1), which makes validity of the first order kinetics for this systematically with respect to particle size. The second order flotation
sample questionable. kinetic parameters were also computed for the pyrite component of the
ore sample studied. These results are given in Table 6. Model fit of the
pyrite data to second order kinetics is also much better than to that of
3.4. Second order kinetics
first order kinetics. It is interesting that the recovery of pyrite at smaller
grind sizes (i.e., P80 53–92 μm) is consistently lower than that of gold.
Since treatment of data according to the first order kinetics has
However, the second order flotation rate constants for pyrite at these
resulted in a significant degree of scatter and apparent inconsistencies,
grind sizes are significantly higher than those for gold. In fact, this trend
applicability of the data for second order flotation kinetics needed to be
may also be noted from a comparison of their first order rate constants
tested. Eq. (7), when integrated for the second order flotation kinetics
(Tables 3 and 4), although the difference is somewhat smaller in this
respectively yields the following (Arbiter and Harris, 1962):
particular case. Despite higher rate constants, the overall recoveries of
pyrite are somewhat lower than those of gold. This can be related to
R∞2 k2 t oxidizable nature of sulphides. Data by Monte et al. (2000) show that
R¼ ð10Þ flotation of oxidized pyrite in the alkaline pH range goes down
1 + R∞ k 2 t
significantly compared to both gold and fresh pyrite. A similar behavior
was found for oxidized pentlandite in samples from aged ore stockpile
(Kelebek and Nanthakumar, 2007).
Table 4
Parameters for the first order flotation kinetics of pyrite in the sample used. Table 5
Parameters for the second order flotation kinetics of gold in the sample used.
P80 Rovr R∞ (1) k1 r2 R∞ (2) k1 r2
(μm) (%) (%) (min−1) (%) (min−1) P80 (μm) Rovr (%) R∞ (1) (%) k2 (g−1 min−1) r2
53 94.8 94.9 1.23 0.9296 97.0 0.68 0.5180 53 95.8 98.4 5.61 0.9996
92 94.1 94.2 1.15 0.9317 96.3 0.68 0.5044 92 94.4 97.1 5.41 0.9999
159 92.3 92.3 0.78 0.7599 95.8 0.58 0.5825 159 92.4 95.6 4.89 0.9997
205 91.8 91.8 1.16 0.9392 95.8 0.58 0.5130 205 91.8 95.6 4.07 0.9999
E. Yalcin, S. Kelebek / International Journal of Mineral Processing 98 (2011) 48–54 53
Rao, S.R., 1971. Xathates and related compounds. Marcel Dekker, Inc., New York. 15–18. Wang, J., Yoon, R.-H., 2009. Surface force measurements between gold surfaces in xanthate
Torem, M.L., Bravo, S.V.C., Monte, M.B.M., Dutra, A.J.B., Tondo, L.A., 2006. Effect of thio solutions. In: Gomez, C.O., Nesset, J.E., Rao, S.R. (Eds.), Advances in Mineral Processing
collectors and feed particle size distribution on flotation of gold bearing sulphide ore. and Technology. : Advances in Mineral Processing and Technology. Sudbury-COM
Miner. Process. Extr. Metall. 115 (2), 101–106. 2009, Canada, pp. 13–28.
Trahar, W.J., 1981. A rational interpretation of the role of particle size in flotation. Int. J. Xu, M., Stratton-Crawley, R., 1996. Rate Constant and Selectivity Measurements in a
Miner. Process. 8, 289–327. Flotation Column. In: Gomez, C.O., Finch, J.A. (Eds.), In Column '96. Montreal, Canada,
Wang, X.-H., Forssberg, K.S.E., 1991. Mechanisms of pyrite flotation with xanthates. Int. J. pp. 81–93.
Miner. Process. 33, 275–290.