Engineering, 2013, 5, 316-321 Published Online March 2013 (
Effect of Gangue Mi nerals on Hydrophobic
Recovery of Gold
Anderson E. L. Mlaki, Jamidu H. Y. Katima, Henry T. Kimweri
Department of Chemical and Mining Engineering, College of Engineering and Technology,
University of Dar es Salaam, Dar es Salaam, Tanzania
Received February 23, 2011; revised February 18, 2013; accepted February 26, 2013
In this paper, the effect of gangue minerals on the hydrophobic recovery of gold is being investigated using ores ob-
tained from the active small scale gold mining sites in Tanzania. Gold ores of different gangue contents were tested.
The effects of silica gangue and high sulphide gangue on gold attachment were examined including the effect of surface
activators (potassium amyl xanthate) and the possibility of depressing the effects of gangue using reagents. The results
were evaluated in terms of gold recovery, volumes and grade of concentrates formed. There was no change in gold re-
coveries when the amount of oxide gangue (quartz) in the ore was increased, indicating absence of competition between
gold and quartz gangue. High sulphide contents in the ore above 6% reduced gold recoveries considerably. It was noted
that potassium amyl xanthate surfactants increased the attachment of both gold and the sulphide gangue. Using lime at
pH 10 it was possible to depress the sulphide gangue which is mainly pyrite and hence increased gold recoveries con-
Keywords: Aggl/Ore Ratio; Gangue Minerals; Hydrophobicity; Surfactants; Gangue Depressants; Oleophilicity
1. Introduction
The presence of gangue minerals in gold ores is among
the main reasons for the selection of appropriate gold
recovery method. The common gangue minerals which
are normally associated with gold ores include quartz,
fluorite, calcite, pyrites, chalcopyrites, galena and many
others in small amounts e.g. arsenopyrites, fluorites,
carbonates and chlorites [1,2].
The hydrophobic recovery of gold is dependent on
greater oleophilicity and hydrophobicity of gold com-
pared to many other minerals. However it is known that
some of the gangue minerals that are associated with
gold have oleophilic and hydrophobic properties and the
presence of such mineral gangue in the ore may be a
source of competition during the attachment to agglom-
erates. Such minerals include mainly the metallic sul-
phides gangue [3,4]. The presence of low amounts of
sulphides in the ore up to 5% chalcopyrite is reported to
have little effect on gold recoveries [5,6].
However, some of the analysed gold ore samples col-
lected from the active small scale gold mining sites in
Tanzania indicated the presence of high sulphides upto
12.9% as shown in Table 1. The presence of high sul-
phides in Tanzanian gold ores are also reported by [7].
This work is focused on the effects of high sulphides
and other gangue on gold recoveries using the hydropho-
bic recovery of gold. Potassium amyl xanthate was the
reagent used as surface activator in the tests. This work
was done as a contribution to the ongoing research on the
application of the hydrophobic recovery of gold as an
alternative to the hazardous use of mercury in small scale
gold mining in Tanzania [8].
Table 1. Gold ore mineralogy.
Mineral composition %
Mineral type Chunya
(Ore 1)
(Ore 2)
(Ore 3)
Chalcopyrite CuFeS2 0.21 1.16 1.09
Pyrite FeS2 7.06 9.44 11.53
Arsenopyrite AsFeS2 0 0.01 0
Galena PbS 0 0.25 0.3
Quartz SiO2 89 79.5 75.64
Liberated gold 1 1 1
Other minerals 2.73 8.64 9.8
Total Mineral 100 100 99.36
opyright © 2013 SciRes. ENG
A. E. L. MLAKI ET AL. 317
2. Experimental
2.1. Materials
a) Agglomerates
Agglomerate samples were prior formed using castor
oil as liquid hydrophobic phase and coal of particle size
–110 microns mixed at oil/coal ratio 0.3. The size of ag-
glomerates used in the tests was 100% below 2 mm.
b) Gold bearing materials
Three different gold ore samples from SSM areas in
Chunya (Ore 1), Geita Gold Mine (Ore 2) and Chocolate
reef-Buzwagi (Ore 3) were used.
Main gangue minerals in the three gold ore types were
determined as quartzite (SiO2), pyrite (FeS2), chalcopy-
rite (CuFeS2), Arsenopyrite (AsFeS2), and Galena (PbS).
The mineral percentages were determined as per major
element analysis which was done with the assistance of
the Southern and Eastern Africa Mineral Center labora-
tory (SEAMIC) as presented in Table 1. Gold contents in
each sample were upgraded to 1% Au, to easy recovery
calculations during experimentation. (The ore mineral-
ogy in Table 1 pertain to samples collected for experi-
mentation and do not necessarily represent the particular
c) Other mater i al s
Gold powder from small scale mining areas in Chunya.
Silica sand (–150 microns) originating from mineral
quartz obtained from the defunct Pugu kaolin plant lo-
cated 25 kilometers from Dar es Salaam. Potassium Amyl
Xanthate (KAX) water soluble, HCl, HNO3, and Lime.
Pyrite (FeS2) was used to provide incremental sulphide
contents in the ores during experimentation. The pyrite
was obtained as complement from Geology Department
2.2. Methods
2.2.1. Sampl e Pr eparation
The ore samples were ground to 100%, –150 microns.
The gold contents in each ore sample were determined by
fire assay then upgraded by addition of free gold (minus
150 microns) to 1% Au. The overall mineral percentage
in the ore is shown in Table 1.
2.2.2. Gold Attachment Tests
The gold attachment tests were carried out in a 1000 ml
beaker equipped with two baffles 180˚ apart. The con-
tents were agitated at 700 rpm by a four-bladed stainless
steel vertical impeller at 20% solids and aggl/ore ratio
0.4. The tests were carried out in the following proce-
dures: (i) Wetting 100 g of gold bearing material with
500 ml de-ionized water then surfactants such as potas-
sium amyl xanthate was added and/or depressants such
as lime was added then agitated for 3 - 5 minutes fol-
lowed by (ii) addition of the agglomerate sample and
stirring for a given time; (iii) Screening the mixture on a
0.325 mm sieve in water to separate the gold-loaded ag-
glomerates from the pulp; (v) Drying agglomerates in
open air for 6 hrs and weighing them then assessing the
recovery of gold by analyzing the amount of gold at-
tached to agglomerates.
2.2.3. Analysis of Gold Recovered in Agglomerates
The gold content in the dry agglomerates was analyzed
by roasting the dry agglomerates to ash at 500˚C in a
muffle furnace and treating the ash with aquaregia to
dissolve gold followed by analysis of the diluted gold
solution by atomic absorption spectrometry.
3. Results and Discussions
3.1. Effect of Silica Gangue on Gold Attachment
Four different sets of samples of increasing silica content
were made by addition of silica sand to Ore 3 (lowest
silica content). Each of the samples was treated for gold
recovery at 60 minute and the recoveries attained did not
show any particular relationship with the changes in sil-
ica content as shown in Table 2.
According to the results, silica content in the ore did
not interfere with gold attachment and hence does not
affect gold recovery. This result seemed to agree with the
theory that quartz has a strong covalent surface bonding
which exhibit high free energy values at the polar surface
[9] and therefore during agitation the polar surface of the
quartz could have reacted with the water dipoles and
making them hydrophilic thus preventing attachment of
quartz to agglomerates.
3.2. Effect of High Sulphide Ores on Gold
The effect of high sulphide contents in the ore, was
shown by using the three natural ores Ore 1 (7.27% S),
Ore 2 (10.6% S) and Ore 3 (12.9% S) with two additional
samples prepared to give a wide variation of sulphide
contents from 6% to 15% S as shown in Figure 1. The
contact time was varied from 30, 60 and 90 minutes. The
gold recoveries determined after each test showed a sig-
nificant decrease with increase in sulphide content. The
highest recovery of 85.1% was attained by the lower
sulphide sample (6% S) at 90 minutes while the high
Table 2. Recoveries attained at different amount of quartz
content in ore.
Description Sample 1Sample 2 Sample 3Sample 4
Quartz content %76 79 81 89
Recoveries % 75.4 77 75.8 76
Copyright © 2013 SciRes. ENG
Copyright © 2013 SciRes. ENG
sulphide sample (15% S) which had a low recovery of
only 72.5%. The same recovery trend (high sulphides
content—low gold recovery) was also noted at lower
contact time as shown in Figure 1.
The lower recoveries attained due to high sulphides
were therefore attributed to natural hydrophobicity and
oleophilicity of sulphides which enabled attachment to
agglomerate thus creating competition between gold and
3.3. Effect of Potassium Amyl Xanthate as
Surface Activator on Gold and Gangue
Generally there were higher recoveries attained when
surfactants were added. As compared to the recoveries
attained in the previous Section 3.2, at 6% sulphide, gold
recoveries increased from the previous 85.1% to 94.0%
at 90 minutes contact as shown in Figure 2.
However, although the recoveries were higher due to
the effect of surfactant, they continued to proportionally
decrease with increase in sulphide content in the ore. The
results indicated that, the action of surfactant increased
gold recoveries, but did not affect the attachment compe-
tition between gold and sulphide gangue.
3.4. (i) Effect of Surfactant on Gangue Recovery
Measured by Concentrate Weights
The results of two sample groups which were treated
with and without surfactant (Figure 3) and (Figure 4)
showed that samples treated with surfactant had higher
concentrate weights. The increase in concentrate weight
was attributed to increased attachment of both gold and
sulphide gangue due to the action of surfactant. The re-
sults agreed with previous results that, the action of xan-
thate surfactant could not decrease attachment compete-
tion between gold and gangue but only promoted at-
tachment of both gold and sulphide gangue as presented
in Figures 3 and 4. Higher concentrate weights also
showed a decrease in grade as discussed in Section 3.4
3.4. (ii) Effect of Concentrate Weight on Grade
The concentrate weights were plotted against grades as
shown in Figure 5. The grade decreased with increase in
concentrate weight indicating high gangue recovery, thus
confirming high competition between gold and sulphides.
It was therefore necessary to carryout further tests as an
attempt to successfully depress the effect of sulphide
gangue and reduce competition on gold recoveries as
presented in Section 3.5.
3.5. Depressing the Effects of Gangue on
Recovery Using Reagents
Due to reported effective performance of lime in selec-
tive depression of pyrite among other sulphides in copper
flotation [4,9,10], slaked lime solution Ca(OH)2 was
Figure 1. Gold recoveries from high sulphide ores without surfactants.
A. E. L. MLAKI ET AL. 319
Figure 2. Gold recoveries from high sulphide ores with addition of surfactants.
Figure 3. Concentrate weights attained due to gangue recovery (without surfactant).
Copyright © 2013 SciRes. ENG
Copyright © 2013 SciRes. ENG
Figure 4. Concentrate weights attained due to gangue recovery (with surfactant).
Figure 5. Weights of concentrates formed versus concentrate grade.
prepared and tested as the reagent of choice for gangue
depression in the experiments.
The results showed a significant increase in gold re-
coveries after lime addition to the gold bearing slurries.
According to Figure 6, lime showed effectiveness as
gangue depressant for all ore samples which contained
high pyrite above 6%. This was shown by the changes in
recovery between corresponding points on the two curves
in Figure 6 when sulphide concentrations were above
6%. It was noted that pyrite concentrations below 6%
had no effect on gold recoveries and hence the effect of
lime could not be noticed in the region. This results
agrees with a report by Calvez et al. [6] who reported
that sulphide contents higher than 5% in the ore may
have detrimental effects to gold recovery. According to
the results it was therefore possible to depress the effect
of sulphide gangue (above 6%) on gold recoveries by
usg lime at pH 10 as shown in Figure 6. in
A. E. L. MLAKI ET AL. 321
Figure 6. Curves showing gold recoveries before and after pyrite depression using lime at 90 minute contact time.
The depression of sulphide gangue (pyrite) could be
explained as due to possible reaction between pyrite sur-
face and lime which changed its natural hydrophobic and
oleophilic properties to hydrophilic, reducing its attach-
ment competition.
It is known that the oleophilicity of unoxidised sul-
phides is decreased due to formation of hydroxides on
the surface in alkaline conditions [10,11].
It was therefore necessary to do the attachment tests at
pH 10 which was considered alkaline enough for hy-
droxide formation on the pyrite surface.
4. Conclusions
1) The results have shown that the presence of certain
types of gangue in the ore may have detrimental effect to
gold recoveries in the hydrophobic recovery of gold, e.g.
high sulphides contents above 6%.
2) The use of surfactants such as potassium amyl xan-
thate increased the attachment of both gold and sulphide
gangue thus increasing attachment competition.
3) Using lime at pH 10 it was possible to depress the
sulphide gangue thus reducing competition on gold at-
tachment to agglomerates and raising gold recoveries
4) There was no change in gold recoveries when the
amount of quartz gangue in the ore was increased, sug-
gesting the existence of quartz particle hydrophilicity due
to reaction between the polar surface and the water di-
[1] M. John and H. Lain, “The Chemistry of Gold Extrac-
tion,” Ellis Horwood Ltd., Chichester, 1969.
[2] M. Hausen, “Characterizing the Textural Features of Gold
Ores for Optimizing Gold Extraction,” Journal of Mining,
Vol. 52, No. 4, 2000, pp. 14-16.
[3] D. Kocabag, G. H. Kelsall and H. L. Shergold, “Natural
Oleophilicity-Hydrophobicity of Sulphide Minerals,” Im-
perial College of Science and Technology, Mineral Re-
sources Engineering Department Publications, London,
[4] B. A. Wills and N. M. Tim, “Mineral Processing Tech-
nology; An Introduction to the Practical Aspects of Ore
Treatment and Mineral Recovery,” 7th Edition, Elsevier
Science & Technology Books, 2006, p. 269.
[5] A. Akcil, X. Q. Wu and E. Aksay, “'Coal-Gold Agglom-
eration; An Alternative Separation Process in Gold Re-
covery,” Separation & Purification Reviews, Vol. 38, No.
2, 2009, pp. 173-201. doi:10.1080/15422110902855043
[6] J. P. Calvez, M. J. Kim, P. L. Wong and T. Tran, “Use of
Coal-Oil Agglomerates For Particulate Gold Recovery,”
Minerals Engineering, Vol. 11, No. 9, 1998, pp. 803-812.
[7] R. Brett, “Re-Logging of Diamond Drill Core, Chocolate
Reef (Buzwagi), Nzega Greenbelt, North West Tanzania,”
Exploration Reports, Dodoma, 2001.
[8] A. E. Mlaki, J. H. Katima and H. T. Kimweri, “Gold Re-
covery from Gold Bearing Materials Using Bio-Diesel,
Vegetable Oils and Coal,” Engineering, Vol. 3, No. 5,
2011, pp. 555-560. doi:10.4236/eng.2011.35065
[9] D. C. Ronald, “Flotation Theory Reagents and Ore Test-
ing,” Pergamon Press, Oxford, 1992.
[10] T. Güler, S. Çetinkaya, Ü. Akdemir, T. Doğan and D.
Kocabağ, “Effect of Fe-Ions on Pyrite-Xanthate Interac-
tion in Chemically Manipulated Electrochemical Condi-
tions,” International Journal of Natural and Engineering
Sciences, Vol. 3, No. 3, 2009, pp. 1-7.
[11] C. Fan, R. Markuszewski and T. Wheelock, “Effect of
Multivalent Cations on Oil Agglomeration of Coal and
Pyrite,” International Journal of Coal Preparation and
Utilization, Vol. 11, No. 3, 1992, pp. 167-175.
Copyright © 2013 SciRes. ENG