University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
ADE | 4. ORE MINERALOGY AND PETROGRAPHY
16000
isation (figure 4.48). Other elements are de-
14000
tected at maximum concentrations of 236ppm R² = 0.9695
12000
for Se (median 40ppm) and 7ppb for Te (me-
10000
dian 2ppm). Both elements feature a clear en- 8000
richmenttrendsimilartoBi(figure4.48). Thal- 6000
lium is commonly present at concentrations in 44000000
the order of 4ppm (median) but is detected as 2000
high as 82ppm in one sample (CAF-1LS-1-1) 0
0 2000 4000 6000
from the upper most part of the orebody (fig-
ure4.48). Tinconcentrationsarelowandrange 111Cd intensity [cps]
between24and115ppm, followingtheenrich- Figure 4.49: Intensity cross-plot between 111Cd
ment trend defined by Bi, Se and Te (figure and 114Cd for galena analyses. Intensities (cps) of
both Cd isotopes feature a very good correlation
4.48). Similar to sphalerite, trace element con- (R2>0.95,n=4),thus,nointerferecebetweenbe-
centrations of one sample (NP950-3) is not in tween114Cdand114Snisindicated.
accordwiththeobserveddepthtrends. Theob-
served concentrations of this stringer ore sam-
ple,however,arewellwithinthevariabilityde- inasimilarwayasinsphaleriteforZn.
finedbytheentiredataset.
Tin is consistently present in galena with a
Cadmium, although at rather low concentra- median concentration of 52ppm and as high
tionlevels,iscommonlydetectedbetween8and as 115ppm. Time-resolved intensity profiles
21ppm,featuringplateau-liketime-resolvedin- feature well-defined plateaus. Many studies
tensityprofiles. Nosignificantelementconcen- reported Sn detected in galena analyses (e.g.
tration or counts per second intensity correla- BlackburnandSchwendeman,1977;Foordand
tionswereidentified(appendixAonpage379). Shawe, 1989; Sharp and Buseck, 1993). Most
If sphalerite micro impurities are the cause for authors suggest the presence of sub-micron,
elevated Cd concentrations in galena, correla- evenly distributed micro-impurities as the most
tions between Cd and Ga would have to be likely cause for elevated Sn. Foord and
present(seediagramBfigure4.57). Thosecor- Shawe (1989) reported the occurrence of “tel-
relations are lacking, thus contamination with lurian canfieldite” (Ag SnS Te ), Sharp and
8 4 2
sphaleritemicro-impuritiesisunlikely. Because Buseck (1993) observed inclusions of franck-
interference between 114Sn and 114Cd is possi- eite (Pb,Sn) Sn Sb FeS . The former may
6 2 2 14
ble, an additional isotope, i.e. 111Cd, has been be the cause for Sn in galena analysed in this
monitored. Thisisotopedoesnotinterferewith study, the latter not, because of the lack of cor-
any other stable isotope. Intensities (cps) of relation with Fe and Sb. In fact, none of the
both Cd isotopes feature a very good correla- element concentration or counts per second in-
tion (R2>0.95, n=4, figure 4.49). Thus, in- tensity correlations between Sn and other mea-
terference is not indicated. It is reported in sured elements is significant (appendix A on
the literature that Cd may occur in galena (e.g. page 379). Neither 118Sn nor 119Sn, both mea-
Blackburn and Schwendeman, 1977), whether suredinthisstudy, areaffectedbyanyinterfer-
as substitution or micro-impurities is unclear, encewithotherstableisotopes,i.e. interference
although the latter is the preferred explanation. isnotafeasiblecauseforelevatedSningalena.
BothCdandPbaredivalentandhaveionicradii TheionicradiusofSniswith1.12Åevencloser
of 0.97Å and 1.20Å, respectively. It is well to Pb as it is to Cd. Bonding characteristics
known and documented that Cd substitutes for of Pb and Sn, i.e. percentage of ionic bond-
Zn in sphalerite, which has an ionic radius of ing with S, are with 36.4% versus 37.0% very
0.74Å.ThedifferenceinionicradiibetweenCd similar(BlackburnandSchwendeman,1977). It
and Zn/Pb is equal to 0.23Å. Therefore, it ap- appearsthatincorporationofSnvialatticesub-
pearsfeasibleofCddirectlysubstitutingforPb stitution is at least up to minor concentrations
116
]spc[
ytisnetni
dC441111 |
ADE | 4. ORE MINERALOGY AND PETROGRAPHY
4.4 Fluid inclusion study in
sphalerite
SEVERAL scientists have studied fluid inclu-
sions hosted in various mineral phases of
Eluraorebody,interaliaSeccombe(1990);Sun
(2000); Sun and Seccombe (2000); Jiang et al.
(2000). Most studies investigated fluid inclu-
sions trapped in quartz and carbonate phases.
Ideally,fluidinclusions,undoubtedlyassociated
withthemainstageofbasemetalsulphidemin-
eralisation, are used to gain information about
the temperature conditions during ore forma-
tion. Sphaleriteistheonlysulphidephasehost-
ing fluid inclusions that may be studied due
to its translucent nature. Fluid inclusions con-
tainedinthisparticularmineralphasehavebeen
investigated in the past but almost exclusively
viacrushandleachtechniques. Temperaturees-
timates were subsequently calculated based on
observed fluid composition. Temperature es-
timates based on homogenisation temperatures
observed via micro-thermometry is sparse (i.e.
Sun (2000) reported 5 measurements of fluid
inclusion in sphalerite in 1 sample). Fluid in-
clusions hosted by sphalerite were investigated
in the course of this study in order to gain ad-
ditional temperature estimates of ore formation
basedonmicro-thermometry.
Ten samples were selected, covering the en-
tire depth extent of the Elura orebody (see
appendix A from page 258 for sample loca-
∗ Figure 4.58: Imagesoffluidinclusionshostedby
tions), and prepared as polished thick section .
sphalerite (sample 775z1-A1; z1 ore zone). Photos
Thesectionswereinvestigatedonatransmitted weretakenundertransmittedinfra-redlight.
lightmicroscopeequippedwithaninfraredlight
sourceandavideocamerasensitiveto1100nm
tigationofhomogenisationtemperaturesviami-
wavelength.
crothermometry. In places, sphalerite contains
The identified fluid inclusions are variable abundantmicrofracturespotentiallyintersecting
in shape, mostly angular to subangular and, in inclusions. Ore microscopy showed that sul-
places,rounded. Theyarerelativeabundantand phide minerals are commonly affected by de-
generally small with sizes in the order of 5 to formation (see section 4.1 on page 63). Spha-
10 µm. Figure 4.58 shows representative ex- lerite is relative brittle compared to chalcopy-
amples of fluid inclusions hosted by sphalerite rite, galena and pyrrhotite. It is highly likely
(sample 775z1-A1; z1 ore zone). All inclu- thatboth,liquidandgasphaseswerelostdueto
sions appear to be either single phase, e.g. liq-
uid or gas, or negative inclusions (i.e. empty). ∗.Sample preparation was undertaken at the Depart-
ment of Applied Geosciences and Geophysics; chair of
Thelackofvapourphaseseliminatestheinves- ResourceMineralogy;UniversityofLeoben,Austria.
126 |
ADE | 4.4 Fluid inclusion study in sphalerite
50000
lerite features primary characteristic vibrations
45000
at around 300 cm−1, with shoulder signals ex-
40000
tendingtoapproximately460cm−1. Secondary
35000
vibrations occur between ~580 and 700 cm−1.
30000
The remaining frequency band is free of sig-
25000
20000 nificant vibrations. A moderate increase in in-
15000 tensityisobservedintheupperfrequencyband
100 300 500 700 1500 2500 3500
10000 causedbyfluorescence. Severalinclusionscon-
5000 tain calcite with a primary characteristic vibra-
0 tion at ~1085 cm−1 (diagram A and C figure
0 500 1000 1500 2000 2500 3000 3500 4000
Wavenumber [cm-1]
4.61). The occurrence of carbonaceous matter
Figure 4.60: Raman spectra of seven sphalerite is indicated by signals at approximately ~1330
analyses.
and~1600cm−1(diagramBandCfigure4.61).
No pure graphite, characterised by a single pri-
mary characteristic vibration at approximately
disintegration of their containment as a conse-
1600cm−1,hasbeenidentified. Insomeinclu-
quence of deformation. A weak internal black
sions, signals were observed in the upper fre-
coatingisobservedonseveralinclusions.
quency band above 2750 cm−1 (diagram B fig-
Laser Raman micro-spectroscopy has been
ure 4.61). The observed broad vibration peaks
utilised in order to investigate the occurrence
stretching from ~2800 to ~3000 cm−1 is char-
of liquid and/or gas phases potentially missed
acteristic for long chain n-alkanes (Mazur and
during microscopy, to identify solid phases
Fanconi, 1979; Lawrie et al., 1999). Very mi-
within inclusions and to study the cause for
nor methane may be indicated by a sharp peak
the observed black coating. Fluid inclusions
at ~2900 cm−1 (Atamas et al., 2004) although
were analysed with a Jobin Yvon LABRAM
longer chain alkanes may feature similar peaks
confocal-Raman spectrometer at the Depart-
(Atamas et al., 2004). In places, an OH-bond
ment of Applied Geosciences and Geophysics;
vibrationat~3620cm−1 hasbeenobservedand
chair of Resource Mineralogy; University of
ismostlikelycausedbyminorchloriteormus-
Leoben, Austria. The system is equipped with
covite. No CO (characteristic vibration spec-
2
a frequency-doubled Nd-YAG laser (100 mW,
tra shown in figure 4.59) or other gas phases,
532.2 nm). Detection is realised with a Peltier-
nor NaCl brines (characteristic vibration spec-
cooled,slow-scan,CCDmatrix-detector.
tra shown in figure 4.59) or other liquid phases
Fluid inclusion analyses were compared to
weredetectedviaRaman.
sphalerite spectra. Figure 4.60 shows Raman
spectra of seven sphalerite analyses. Spha-
CO (cid:24)aCl brine OH-bond in chlorite
2
500 1000 1500 2000 2500 3000 3500 4000 3000 3500 4000
Wavenumber [cm-1] Wavenumber [cm-1] Wavenumber [cm-1]
Figure 4.59: Raman spectra of reference materials and showing their primary char-
acteristicvibrations.
127
].u.a[
ytisnetnI
].u.a[
ytisnetnI |
ADE | 4. ORE MINERALOGY AND PETROGRAPHY
4.6 Summary of findings
eter were unsuccessful and gave meaningless
results, confirming microscopic observations
The integration of observations made during of ore textures indicating that sphalerite and
microscopic and mineral chemical investiga- pyrrhotite are not in equilibrium with pyrite.
tions highlighted the pronounced heterogene- The systematic shift from low-Fe sphalerite in
ity in respect to grain sizes, texture and min- the upper main lode zone to consistent high-Fe
eral composition of Elura’s sulphide ore. Sul- sphalerite in lower zones of the orebody is in-
phide paragenesis is simple, comprising pyrite, dependentofwhetherpyriteorpyrrhotiteisthe
pyrrhotite, sphalerite andgalena as major, mar- prevailing iron sulphide species. The only fea-
casite, chalcopyrite and arsenopyrite as minor, sible explanation seems a change in the chem-
as well as tetrahedrite, native silver and mag- istry of the mineralising fluid as the hydrother-
netite as trace mineral phases. Grain sizes mal system evolved. The precipitation of sig-
of the most important sulphide phases, i.e. nificant pyrrhotite from the hydrothermal fluid
sphalerite,galena,chalcopyriteandtetrahedrite, may have caused a decrease of its FeS activity,
vary strongly over a large size range between a subsequentlylessFeSisavailablefortheincor-
few microns and up to several millimetres for porationwithsphalerite.
sphaleriteandgalena. Manganese, an important smelter penalty el-
ement for zinc concentrate, was detected at
Sphalerite is the most abundant base metal
trace concentration levels (maximum 300ppm
sulphide phase, followed by galena. Chalcopy-
based on LA-ICP-MS) and has no negative ef-
riteabundanceisgenerallylowandvariablebut
fect on the product quality. Cadmium concen-
slightly enriched towards increasing depth be-
trations are relatively low with a maximum de-
lowsurface. Inpyrrhotite-dominatedorezones,
tectedvalueof0.26wt%. ResultsofEMPAsug-
sphalerite and galena are commonly intimately
gested significant Hg and Bi concentrations in
intergrown with pyrrhotite, whereas in pyritic
excess of 1000ppm incorporated in sphalerite.
zones they predominantly occur interstitial to
LA-ICP-MSanalyses,however,onlyconfirmed
andfillfracturesorvughsofpyrite,partiallyre-
Hg in sphalerite but at much lower concentra-
placing it. Replacement of pyrrhotite by base
tion levels commonly below 100ppm (median
metalsulphideissubordinate. Sphaleriteisim-
24ppm). Other elements detected at low con-
pure and contains abundant inclusions of sul-
centrationsinsolid-solutioninsphaleriteareCu
phide as well as non-sulphide gangue (NSG)
(median of 93ppm) and very minor Ga (maxi-
phases.
mum of 18ppm), In (maximum of 5ppm), and
Sphalerite composition varies throughout the Sn(maximumof26ppm).
orebody. Low iron concentrations with a min- Investigations of potential fluid inclusions
imum detected value of 2.41wt% are observed hostedbysphaleriteonlyidentifiednegative,i.e.
in sphalerite from the upper main lode ore empty, inclusions. Most, if not all parts of the
zone and in peripheral parts of the orebody. orebody were affected by deformation. Former
Pyrrhotite dominated ore exclusively contains fluid inclusions were lost consequently due to
Fe-richsphaleritecontainingupto8.22wt%Fe. deformationandfracturingofsphalerite,allow-
Microscopic observations of ore texture ing fluid and gas phases to escape. Calcite,
showed a clear co-genetic relationship between carbonaceous matter and long chain n-alkanes,
sphaleriteandpyrrhotite. Incontrast,sphalerite identified via laser Raman micro-spectroscopy,
replacementofpyriteiscommon. Pressurecal- arecontainedinnegativeinclusions.
culations based on the sphalerite geobarome- Zinc isotope signatures of sphalerite suggest
terwereunsuccessfulandgavemeaninglessre- anaveragecontinentalcrustaslikelysourcefor
sults, confirming that sphalerite and pyrrhotite Znandthereforeotherbasemetals. Thesource
are not in equilibrium with pyrite. Pressure reservoirwaslikelyeffectivelyleachedbyahy-
calculations based on the sphalerite geobarom- drothermal fluid that subsequently inherited its
134 |
ADE | 4.6 Summary of findings
crustal isotopic signature. The inferred repeti- ThecoincidentalenrichmentofBiandAgor
tive depth-dependant Zn isotope fractionations Ag and Sb is caused by galena-matildite solid
in the upper and lower parts of the minerali- solutionsindeepestpartsoftheorebody,andby
sation are caused by syngenetic Rayleigh frac- galena-miargyrite solid solutions in the upper-
tionation. Vertical temperature gradients likely mostzonesofthemineralisation,respectively.
aided the fractionation. The fluid evolved from
Chalcopyrite is, apart from tetrahedrite, the
aninitialtypicalcrustalsignatureinmajorfluid
only visually identified copper-bearing mineral
influx zones to heavier signatures as it evolved
phase at Elura. It may occur as inclusions and
during its ascent. Additional analyses on a
exsolutions in sphalerite, intergrown with other
larger sample set are required in order to prove
basemetalsulphidesand, inplaces, isenriched
thistheoryasthistrialstudyisbasedonasmall
in zones enriched in iron carbonate and quartz.
numberofsamples.
Low concentrations of Co, Zn, Ag, Bi, As and
Galenafeaturesthemostpronouncedgrainsize SnweredetectedinchalcopyriteviaEMPA.
variability of all sulphide phases, occurring as
It has been shown that galena hosts significant
smallasafewmicronsinsizefillinginterstitial
Ag in some areas of the orebody. Apart from
space between pyrite or as several millimetres
these zones, argentian-tetrahedrite and minor
largepatches.
native silver were identified as the only sil-
The detailed mineral chemical study of
ver hosting mineral phases. Occurrences of
galena was undertaken in order to explain high
native silver and coarser grained tetrahedrite
levels of Bi in lead concentrate. No discrete
(15-200µm) are limited to the uppermost ar-
Bi-mineralphasesweremicroscopicallyidenti-
eas of the orebody. Here, tetrahedrite is com-
fied in any of the studied ore samples. EMPA
monly associated and intergrown with galena.
showed significant levels of Bi in galena up to
Throughout most of the mineralisation, tetra-
0.85wt%,evenlydistributedwithingrains. Bis-
hedriteissignificantlyfinergrainedandmaybe
muth concentrations increase with depth below
characterisedbymineralassociationswithother
surface. Galena contained in pyrrhotitic ore
phasesthangalena. Inplaces,tetrahedriteisal-
features slightly higher Bi concentrations com-
teredtonativesilverandchalcopyrite.
paredtootheroretypes.
The detected Ag concentration of all unal-
TraceelementdeterminationviaLA-ICP-MS
tered tetrahedrite grains via EMPA ranges be-
confirmed high levels of Bi in galena (max.
tween ~31 and ~19wt%. Some analyses with
5,645ppm) and its depth trend, but showed
Ag concentrations in excess of 31wt% feature
that EMPA is consistently overestimating Bi
calculatedstoichiometriessimilartofreibergite.
concentrations. The analytical error is most
Argentian-tetrahedrite is, in places, altered to
likely caused by interference of characteristic
native silver and Ag-rich chalcopyrite. Sto-
X-ray Mα peak positions of Bi and Pb. Sil-
ichiometries calculated on the basis of sev-
ver and Sb are both incorporated up to signif-
eralanalysesshowedsignificantdeviationfrom
icant concentration levels at maxima of 2,339
ideal tetrahedrite composition and reflect inter-
and 2,385ppm, respectively. The former ele-
mediatestagesofalteration.
ment is enriched in the uppermost as well as
lowermost parts of the orebody (mainly in the Four pyrite generations were identified in the
mainlodeorezone),whereasthelatterelement course of this study. Colloform, framboidal
features a converse enrichment trend compared and cloudy pyrite A formed prior to the base
to Bi. Otherelements detectedin galenaat low metal mineralisation. It is the most abundant
concentration levels and incorporated via sto- and represents the oldest pyrite variety in the
ichiometric lattice substitution are Se, Te, Cd mineralising system. Most of the second cata-
and Sn. Low levels of Tl are likely caused by clastic, sub- to euhedral pyrite B formed prior
nano-impuritiesofTl-bearingsulfosaltspecies. to the base metal sulphides during a period
135 |
ADE | 4. ORE MINERALOGY AND PETROGRAPHY
whenthefluidtemperatureincreasedandwhen notcontainanyelementinsignificantquantities
base metal concentrations began to increase. otherthanFeandS.
Both pyrite varieties were affected by tectonic
Arsenopyrite occurs in all different ore types
strain causing intense fracturing and were sub-
and depths but is enriched in the upper and
sequently intruded and partially replaced by
pyritic peripheral parts of the deposit. It is
base metal sulphides. Commonly coarse grains
observed in variable grain sizes, shapes and
of pyrite B feature typical and characteristic
textures. Partial replacements by base metal
cataclastic textures. The base metal minerali-
sulphides may occur. Cobalt and Sb are the
sation stage is dominated by pyrrhotite as the
only elements consistently detected in EMPA
main iron sulphide phase with only minor an-
of arsenopyrite. Some grains feature a well-
hedral pyrite C formation. In places, syn-base
defined growth zonation and a blurred, over-
metal pyrite C is intimately intergrown with
printing, deformational-induced zoning. The
magnetite. In peripheral pyritic dominated ore
formerzonationiscausedbyelevatedCoorSb
zone where pyrrhotite is lacking, minor pyrite
concentrations, the latter by changes in As/S
C is overgrowing earlier pyrite A and B. In
ratios. Elevated Ag and Au is observed in
these ore zones, however, pyrite precipitation
some analyses and caused by micro-inclusions
is subordinate. Replacement of pyrite by base
ofelectrum.
metal sulphides prevails. Late, euhedral and
finegrainedpyriteDisobservedinminorquan- Iron carbonates (ankerite and siderite) and
tities throughout the orebody and formed after quartz are the most important non-sulphide
thebasemetalmineralisationstage. gangue phases (NSG). The former preferably
Pyrite A and B compositions were inves- occursincorezonesoftheorebody,thelatterin
tigated via EMPA but showed no significant peripheral zones. Both phases occur interstitial
differences. Cobalt, Ni and As contents are to sulphides or form vein systems of complex
low at median concentrations of 0.07, 0.04 and timerelationshipsthroughouttheentireparage-
0.08wt%, respectively. Silver and Au were de- netic sequence. Minor barite is observed. Rare
tectedatamaximumconcentrationof0.2wt%, occurrencesofmagnetite,intimatelyintergrown
likely caused by micro-inclusions of unidenti- withpyritetypeC,areobserved.
fied phases. Trace element determination via Strongly altered wall rock clasts are em-
LA-ICP-MS suggests that Co, Ni, As, Sb, Au bedded in sulphide matrix in the semi-massive
and Tl are contained in pyrite at low concen- peripheral ore zones. Muscovite and chlo-
tration level. The impure nature of the pyrite rite are common alteration phases in silt- and
types makes is hard to distinguish between ele- mudstone rich, quartz in sandstone-rich frag-
mentsincorporatedinpyritevialatticesubstitu- ments. Minor goethite formed via alteration
tionfromthoselinkedtomicro-inclusions. of pyrite and siderite and preferably occurs in
pyritic ore types. Subordinate alkali-feldspar
Pyrrhotite dominates as iron sulphide in core andsodium-richplagioclaseisrecognisedasal-
zonesofthedepositwhereitrepresentsthesul- terationphase.
phide groundmass. It is clearly co-genetic with Compositionsofchloriteclearlyrelatedtothe
base metal sulphides, commonly occurring in- mineralisingeventweredeterminedviaEMPA.
timately intergrown in a myrmekitic-like tex- Most chlorite is ripidolite, some pseudothurin-
ture. Transformation of pyrrhotite to marcasite gite and daphnite. If oxidised, they would be
is observed in several samples throughout the classified as thuringite. Chlorite thermometry
entire depth extent of the orebody. In places, estimatedatemperaturerangebetween314and
transformation is complete, but texture typical 343°C, similar to temperature conditions pro-
for pyrrhotitic ore remains. Pyrrhotite com- posed for the mineralising fluid in other stud-
positions were determined via EMPA and LA- ies. A significant temperature change over the
ICP-MS. Results showed that pyrrhotite does depthextentoftheorebodyisobservedandmay
136 |
ADE | 4.6 Summary of findings
reflect changing crystallisation temperatures or clasts,whicharestronglyfracturedandpartially
changes in fluid characteristics. The tempera- replaced, occur in varying quantities. In zones
turedependencyasafunctionofdepthalsoap- ofhighdeformation,sphaleriteoccursasbands,
pearstoberelatedtochangesinthehostlithol- elongated grains and boudinaged fragments.
ogy, thus may be caused by differences of its Oretextureisreminiscentofmylonitesinplaces
chemicalcomposition. Inthiscase,thetemper- where strain was focussed into a rather con-
aturetrendwouldbeunrelatedtothemineralis- fined zone. Here pyrrhotite and galena reacted
ingevent. strongly ductile and were remobilised into dis-
crete sulphide bands. In contrast, sphalerite re-
Ore textures vary between ore types as well
acted brittler, forming sub-rounded clasts. De-
asbetweenzonesofintenseandweakdeforma-
formation of galena is shown by its bending
tion. Base metal sulphides in massive pyritic
cleavage planes. The sulphide body was af-
ore are commonly characterised by a reticular,
fected by prolonged deformation commencing
vein-style texture, where they fill fractures, in-
during the early hydrothermal stages as shown
terstitial space in rather compact but fractured
by the strongly tectonised early pre-base metal
pyrite groundmass, partially replacing it. Tex-
pyrite generations (type A and B) and the de-
tureofpyrrhotitic-dominatedoreissignificantly
formationofbasemetalsulphides. Remobilisa-
different, commonly featuring myrmekitic-like
tionandre-crystallisationatvaryingintensities,
intergrowths of sphalerite, galena, pyrrhotite
caused by the deformation, are responsible for
and chalcopyrite. Pyrrhotite, syn-genetic to
the significant grain size variability and hetero-
base metal sulphides, commonly features 120-
geneityoforefromtheEluraorebody.
degree grain boundaries. Pyrite B porphyro-
137 |
ADE | Chapter 5
Geochemical ore characterisation
THE ELURA DEPOSIT contains an estimated
pre-mining resource in excess of 50.7Mt
at 8.8wt% Zn, 5.6wt% Pb, 107g/t Ag and
∗
0.2wt% Cu . Samples from resource diamond
drilling are commonly assayed for Zn, Pb, Ag,
Cu and Fe and their spatial distribution mod-
elled in the resource block model. Apart from
thoseelements,othertraceelementspotentially
important for mineral processing (e.g. penalty
or potential beneficial elements) are not deter-
mined. No comprehensive trace element study
hasbeenperformedontheElura orebodysince
itsdiscovery.
A set of 124 samples, taken from locations
throughout the deposit and all occurring ore
types, were used for major and trace element
whole rock geochemistry as shown in figure
5.1 and summarised in table 5.1. Bulk sam- Figure 5.1: Samplelocationsformajorandtrace
elementwholerockgeochemistry. Numberofsam-
pleswerecrushedandgroundinachromesteel
ples is 124. Longitudinal view towards WSW
swing mill to analytical fineness at the Depart- (˜245°). Yellow shape is the stringer type ore
(VEIN)resourcedomain.
ment of Geology and Geophysics, School of
Earth and Environmental Sciences at the Uni-
versityofAdelaide. Samplealiquotswereanal- Mineralisation zone
Ore type ML z1 z2 z3 z4 z5 z6 W-Min
ysedformajorandtraceelementsbyALSLab-
Po 20 4 6 4 4 1 1 -
oratory Group, Australia. Geochemical pro- Py 20 1 2 3 3 3 - -
SiPo 9 2 - - 2 - 1 1
cedures used for different elements are sum- SiPy 11 6 - 1 3 - - 1
VEIN 5 1 - - - - - 1
marised in appendix B on page 480. Smaller MinA/ALT 2 - - - - - - -
6
sample sub sets were chosen for rare earth el- CSA (fresh unaltered, taken at greater distance to the orebody)
ement (REE, n=34, figure 5.2) and platinum Table 5.1: Number of samples for different ore
groupelement(PGE,n=8,figure5.4)determi- typesandorezones.
nation,includingRe-Osdating. Theseelements
were analysed at the Department for General
and Analytical Chemistry at the University of
∗.Calculated by the author based on the resource
Leoben,Austria. block model January 2008 (abbrev. Res08_jan) and a
cut-offgradeof10wt%Pb-Zncombinedgrade,supplied
byCBHResourcesLtd.
139 |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
5.1 Background and
ability and their spatial distribution throughout
methodology theentireorebody. Elementsidentifiedasbeing
highlyimmobileduringhydrothermalalteration
and sulphide precipitation, were used to calcu-
5.1.1 Major and trace element
whole rock geochemistry∗ laterelativemasschangesandsubsequentlyvol-
ume gains/losses during ore formation. Trace
Major element analyses of ore samples elementsenrichedincertainsulphidephasesup
(n=118) were used for to estimate their modal to significant concentration levels (e.g. Ag and
mineralogy and subsequently to calculate bulk Biingalena;seesection4.3onpage108)were
sample density. Density values are reconciled reconciledtowholerockgeochemistryinorder
to gravimetric determined ore density data to: (a)identifythepotentialoccurrenceofother
supplied by CBH Resources Ltd. If density mineral phases, which may incorporate partic-
estimates are sufficiently accurate, they will ular elements, and (b) estimate the relative per-
be used to calculate relative volume loss/gains centage of metals locked up in certain mineral
during ore formation. Major element compo- phases.
sition of unaltered fresh host rock will be used For most trace element determinations, ap-
for source rock provenance discrimination and proximately 0.250g of sample material was di-
toclassifyfortectonicsettings. gested via a multi-acid method using perchlo-
Astandardlithiummetaborate-tetraboratefu- ric (HClO 4), nitric (HNO 3) and hydrofluoric
sion (ME-XRF12 ALS lab-code) was used for (HF) acids. The solution was evaporated to
sampledigestion. Approximately0.66gsample neardrynessbeforehydrochloricacidwasused
was fused in a lithium metaborate-tetraborate forfurtherdigestion/leach(ME-MS61ALSlab-
flux at a ratio of 12:22. Lithium nitrate was code). Final sample solution was analysed
addedasoxidisingagent. Meltwaspouredinto via inductively coupled plasma-atomic emis-
platinummouldinordertoproduceglassdisks, sion spectroscopy (ICP-AES) and -mass spec-
which were analysed via XRF spectrometry on troscopy (ICP-MS). A Varian Vista Radial in-
a PANalytical AXIOS PW4400 system. Total strument and a Perkin Elmer ELAN 9000 sys-
sulphurandcarbonweredeterminedviaLECO tem were used for ICP-AES and ICP-MS, re-
CS carbon-sulphur analyser using 0.01 to 0.1g spectively. Elements incorporated in refractory
sample material (S-IR08 and C-IR07 ALS lab- mineral phases (e.g. Y, Zr, Sn, W, etc) were
codes). Determinationofloss-on-ignition(LOI) analysed using an alternative digestion method
ismeaninglessonsamplesthatcontainhighsul- in samples enriched in silicates in order to en-
phide quantities. That is because mass changes sure complete sample decomposition. Those
are predominantly caused by oxidation of sul- sampleswerefusedinalithiummetaborateflux
phides to either oxides (mass loss) or sulphates followed by dissolution in a mixture of HNO 3
(massgain)uponroasting, andthereforewould and HCl acids followed by analysis via ICP-
not accurately reflect contents of volatiles (e.g. MS.Samplescontaininghighconcentrationsof
water,organics,etc.). Samplesthatcontainsig- zinc, lead or silver were digested according to
nificant quantities of chlorite and sericite and ME-MS61 but more diluted (ME-OG62 ALS
thus water incorporated in crystal lattice (i.e. lab-code). Mercurywasdeterminedviaaquare-
siliceous semi-massive sulphide samples with giadigestionfollowedbyeitherICP-AES(ME-
high wall rock component) will result in totals ICP41 ALS lab-code) or ICP-MS (ME-MS42
below100%. ALS lab-code) according to concentration lev-
els. Gold analysis was by fire assay fusion fol-
A suite of 40 trace elements contained in
lowed by ICP-AES (Au-ICP21 ALS lab-code).
massiveandsemi-massivesulphideoreandwall
rock samples (n=118 and 6, respectively) was
∗.Summarised after geochemical laboratory proce-
determined in order to investigate affinities be-
dures ALS Chemex Laboratory group Australia (pers.
tween trace elements, their concentration vari- comm. StevenFinlayson,2010)
140 |
ADE | 5.1 Background and methodology
Approximately 30g sample material was fused causedbythehydrothermalmineralisingevent.
inamixtureofseveralreagents(e.g. leadoxide, REE commonly occur at low concentrations
sodium carbonate, borax, silica, etc.). The re- in geological material. Refractory and chemi-
sultingmetalbeadwasdigestedinamixtureof callyhighlyresistantmineralphase,e.g. zircon,
dilutedHNO 3 andconcentratedHClbyheating tourmaline,areimportanthostsforREEs. Com-
in a microwave oven. Diluted sample solution pletesampledissolutionisvitalforaccuratede-
wassubsequentlyanalysedviaICP-AES. termination of those elements. Sodium perox-
Several geological reference materials were ide sintering was used in order to deliver com-
used as laboratory standards for quality as- plete sample digestion, following the method
surance quality control (i.e. G2000 ALS described in Meisel et al. (2002). All acids,
Laboratory Group; GBM303-1, GBM305- sampledilutionsinadditiontorinsingsolutions
11, GBM306-12, GBM398-4C, GBM904-11, werepreparedwithpurifiedreverseosmosiswa-
GBM999-5 Geostats Pty Ltd; SARM-3 South ter (mQH O), delivered via a MilliQ system
2
African Bureau of Standards; ST321, ST381 (Millipore).
Gannet Holdings Pty Ltd). Analytical results Samples were ground to analytical fine-
were considered as valid based on standard, ness. Approximately 100mg of sample mate-
blankandduplicateanalyses. rial was mixed and homogenised with approx-
imately 0.6g sodium peroxide with a purity of
>95%m/minglassycarboncrucible. Themix-
5.1.2 Rare earth element analysis
tures were subsequently sintered at 480°C for
Rareearthelement(REE)concentrationsinore 30min. The crucibles were allowed to cool to
and unaltered fresh wall rock samples were in- room temperature, before mQH 2O was added
vestigated. Their concentrations and distribu- drop-wise, in order to control the intensity of
tion pattern were used to classify the host rock the dissolution reaction. Hot plates with mag-
(CSASiltstone)andtoinvestigateREEmobility neticmomentsmaintainedconstantsolutionag-
itationandatemperatureof90°C.Uponcessa-
100
10
Figure 5.3: Rare earth element standard analy-
ses (Penrhyn Slate OU-6). Chondrite normalised
concentrationdataofthisstudydisplayedasblack
Figure5.2: Sampleselectedforrareearthelement linessuperimposedtocertifiedvaluesofOU-6(blue
study. Numberofsamplesis34. Longitudinalview shape defined by uncertainties) after Kane (2004).
towardsWSW(˜245°). Yellowshapeisthestringer ChondritedatafromPalmeandO’Neill(2004)was
typeore(VEIN)resourcedomain. usedfornormalisation.
141
etirdnohc/elpmaS
aL eC rP dN mP mS uE dG bT yD oH rE mT bY uL |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
tion of the reaction, solutions were centrifuged 5.1.3 Platinum group element de-
for5minat4000rpminordertoseparateundis- termination (incl. rhenium)
solved hydroxides. The residue was dissolved
with 3ml 3M HCl. Graphite crucibles were Economic platinum group element (PGE)
rinsed with 2ml concentrated HCl. All sample mineralisation is commonly linked to
solutionswerecombinedina100mlvolumetric mafic/ultramafic orthomagmatic deposits,
flask(resultingdilutionis1:1000). e.g. layered mafic-ultramafic intrusions,
Oneml sample aliquots were diluted with Alaskan-type alkaline/calc-alkaline intrusions,
1%m/m HNO up to a volume of 5ml. As etc. Smaller occurrences are found in py-
3
internal standard, 100µl of a 0.1µgml−1 in- roxenite dykes or ophiolites associated with
dium/rhenium solutionwas added to each sam- “Schlieren”-type chromitites. Several studies
plesolution(finaldilutionis1:5100). showed that PGEs may be significantly mobile
All samples were digested and analysed as under aqueous hydrothermal conditions (e.g.
duplicates on a HP4500 Agilent Technologies Wood, 2002). Thus, significant PGE con-
ICP-MS. One blank and one standard solution centrations may occur in hydrothermal base
were analysed after every 5 samples. Penrhyn metal mineralisation hosted by a large variety
Slate (OU-6) was used as standard reference of different rock types, e.g. sedimentary,
material, certified by the International Associ- magmatic or metamorphic (Maier, 2005). For
ationofGeoanalysts(Kane,2004). Mostblank instance, significant PGE concentrations in
solutions yielded elemental concentrations be- excess of 1,000ppm Pd and Pt are reported
low 10ppb. All standard analyses (n=7) fea- for the Kupferschiefer in Poland (Kucha and
tured high reproducibilities with a maximum Przylowicz,1999;Bechteletal.,2001).
RSD of 2.9% with most analyses below 1.7%. The Elura deposit has previously been dated
Comparison of element concentrations of stan- based via Pb-Pb (411Ma David, 2005) and
dard analyses with certified standard reference 40Ar-39Ar (376-379Ma Sun, 2000; Sun et al.,
materialconcentrationdatashowedasignificant 2000) techniques, resulting in a significant age
and consistent overestimation of Ho caused by discrepancyforthegenesisofthesulphidemin-
laboratory wide contamination (pers. comm. eralisation (see section 3.2 on page 50). Thus,
Thomas C. Meisel 2009) and subsequently a further dating technique was sought in order
were rejected. Figure 5.3 shows chondrite- to better constrain the age of the ore-forming
normalised standard analyses (black lines, ex- event.
cluding Ho) superimposed to certified values Sphaleritehasbeensuccessfullydatedviathe
of OU-6 (blue shape defined by uncertainties). Rb-Srisotopesystem(e.g.Brannonetal.,1992;
Chondrite data from Palme and O’Neill (2004) Nakai et al., 1993; Christensen et al., 1995a,b;
wasusedfornormalisationsinthecourseofthis Schneider et al., 2003) and was initially con-
study. All REEs fall within certified values ex- sidered as a good option for age determina-
cept Ce, which is slightly lower. Median Ce tion. However, microscopic observations re-
concentrationdefinedbysevenanalysesisnev- vealed the pronounced heterogeneity and im-
ertheless very close to the certified value with purity of sphalerite contained within ore from
72.8ppm compared to 77.1ppm±2.7. Based Elura. Probably the most important and abun-
on blank and standard analyses, all analytical dant impurity in sphalerite are inclusions of
resultswereaccepted. iron-carbonatemineralphasesandsericite. The
formerphasesreadilyincorporateSrinsolidso-
lution and the latter Rb. Such inclusions would
certainlyhavedisturbedthehighlysensitiveRb-
Srisotopesystem. Subsequently,itsapplication
hasbeendismissed.
TheRe-Osisotopicsystemiscommonlyused
142 |
ADE | 5.1 Background and methodology
for dating mafic/ultramafic rock suits and asso- the orebody. If concentrations are sufficiently
ciated mineral deposits. Despite low PGE con- high, and PGEs detectable, their distributions,
centration levels, several authors were able to fractionationsandOsisotopecompositionsmay
usethedatingtechniqueforbasemetaldeposits help to determine fluid characteristics and con-
(e.g. Morelli et al., 2004; Keays et al., 2006; finepotentialmetalsources. Re-Osgeochronol-
McInnesetal.,2008). Traceelementgeochem- ogy will be used for age determination. If suc-
istryofthisstudyshowedaweakbutnoticeable cessful, the date would represent the first time
correlation between Re and Mo (figure 5.22), constraintestablishedbydirectlydatingthesul-
suggestingmolybdeniteasamajormineralhost phidemineralisation.
for Re. The whole rock Re concentrations are
The whole rock massive sulphide samples
variable and would cause the spread of Re and
were digested and analysed following the
Os isotope ratios needed for accurate isochron-
isotope-dilution,on-linecationexchangematrix
agedetermination.
separation procedure described in Meisel et al.
(2001,2003). Allacids,sampledilutionsinad-
dition to rinsing solutions were prepared with
purified reverse osmosis water (mQH O), de-
2
livered via a MilliQ system (Millipore). Sam-
plesweregroundtoanalyticalfinesseandthor-
oughly homogenised before an aliquot of 0.5g
wasextracted. Thesamplematerialwasweight
intoquartzglasstubesandamulti-elementPGE
spike solution containing 99Ru, 108Pd, 185Re,
190Os, 191Ir and 198Pt was added. The quartz
glass vessels were sealed with Teflon® tape
and glass lid immediately after adding 2ml
conc. sub-boiledHCland5mlconc. sub-boiled
HNO . In order to avoid excessive reaction of
3
thesamplepowderwithacids,potentiallycaus-
inglossesofvolatilephases(i.e. OsO ,RuO ),
4 4
quartz glass vessels were cooled in iced wa-
ter. The glass tubes were subsequently pres-
surised in an autoclave system (high pressure
asher HPA-S; Anton Paar – PerkinElmer In-
struments, Graz, Austria) under nitrogen at-
Figure 5.4: Sample selected for the determina-
mosphere at a pressure of 100bar. The sam-
tionofplatinumgroupelementcontent;numberof
samples is 8. Results of circled samples was re- ples were dissolved/leached for three hours at
jected due to significant analytical error. Longitu- aconstanttemperatureof300°Candallowedto
dinalviewtowardsWSW(˜245°). Yellowshapeis
cool for an additional hour. The sample solu-
thestringertypeore(VEIN)resourcedomain.
tions were decanted into Teflon® beaker. The
As a trial, a small set of eight whole rock volatile OsO phase was directly sparked into
4
massive sulphide samples was used for the de- thetorchofthequadrupoleICP-MSinstrument
termination of PGE and Re concentrations in (HP4500AgilentTechnologies)andOsconcen-
addition to the Os isotope ratio (187Os/188Os), trations determined via isotope dilution. The
never investigated at the Elura orebody before. samplesolutionsandtheundigestedsamplema-
The locations of these massive sulphide sam- terial was diluted to 50ml and carefully sepa-
ples, comprising 4 pyrrhotite- and 4 pyrite- rated via centrifuge. The supernatant was de-
dominated samples are shown in figure 5.4 and canted into Teflon® beaker, put on hot plates,
were taken from different locations throughout and allowed to evaporate until almost complete
143 |
ADE | 5.2 Major elements, modal mineralogy and density estimation
5.2 Major elements, modal Silicates Carbonates Py & Po Apy
Cpy Gn Sph
mineralogy and density
100%
estimation
90%
80%
MAJOR element compositions, calculated 70%
modal mineralogy and density of all 60%
analysed ore samples are shown in appendix 50%
section B.2 on page 480 and summarised for 4400%%
different ore types in table 5.2. Most iron is 30%
incorporated in sulphides, thus calculated and 20%
presentedaselementalconcentration. 10%
It needs to be stressed, that the calculated 0%
Po Py SiPo SiPy VEIN
modal mineralogy based on chemical data is at
MinA
best a rough estimation and has to be under- Ore type ALT
stoodasinformativeonly. TotalZn,Pb,As,Cu Figure 5.5: Calculated modal mineral composi-
tion based on major element whole rock geochem-
andBaconcentrationsweredirectlyusedtocal-
istryfordifferentoretypes. VEIN/MinA/ALTcat-
culate modal percentages of sphalerite, galena, egoryisdominatedbyVEINoretype(n=7outof
arsenopyrite, chalcopyrite and barite. Average atotalof9).
sphaleritecomposition,i.e. (Zn Fe )Sasde-
0.9 0.1
termined via EMPA (see section 4.2.1 on page
79) was used. Total Ca concentration was used spectively. Median Cu concentrations, with the
to calculate the amount of calcite, the remain- exceptionofVEIN/MinA/ALToretypes,range
ingcarbondioxidewasassumedtorepresentthe between 0.20 and 0.24wt% for Po, SiPo and
amount of siderite. The overall low manganese SiPy, and are lowest in massive pyritic ore at
andmagnesiumconcentrationswereignoredfor 0.16wt%. A weak tendency of higher Cu con-
carbonate calculation. Sodium and K concen- centrationsisindicatedforpyrrhotiticoretypes
trationswereusedtocalculatetheamountofal- as expressed by higher chalcopyrite content in
bite and orthoclase, respectively. A significant figure5.7.
proportion of K is associated with muscovite, The calculated modal compositions of dif-
butitwasnotpossibletocalculatebothphases, ferent ore types are summarised in figure
i.e. orthoclase and muscovite. The remaining 5.5. Total silicate content progressively in-
Al was used to calculate the amount of chlorite creases from approximately 2wt% (median)
usingitsaveragecompositionasdeterminedvia for massive, to 13wt% (median) for semi-
EMPA(seesection4.2.3onpage101). There- massive siliceous and up to 63wt% (median)
mainingSiwascalculatedasquartz. forVEIN/MinA/ALToretypes(figure5.7). In-
Median Zn and Pb concentrations are rather verse quantities are observed for total sulphide
uniform for massive (Po and Py) and semi- and total iron sulphide (Py + Po) content as
massive (SiPo and SiPy) ore types at approxi- themostabundantsulphidespecies(figure5.7).
mately 9.6 to 10.2wt% and 5.1 to 5.9wt% (ta- Massive ore contains around 85wt% (median),
ble 5.2). The highest Zn concentration is ob- siliceous ore 75wt% (SiPo) and 69wt% (SiPy)
served in massive pyrrhotitic ore at 16.4wt%. total sulphides. Iron sulphide content (pyrite
In contrast, maximum Pb is observed in mas- + pyrrhotite) are lowest in SiPy ore (median
sive pyritic ore at 30.30wt%. Stringer type 46wt%) and range between 54 to 50wt% (me-
ore (VEIN) including one minor mineralised dian) for Po, Py and SiPo, not considering
(MinA) and one altered wall rock (ALT) sam- VEIN ore. Total carbonate content is highly
ple feature variable Zn and Pb concentrations variable, ranging between 0.39 and 33.65wt%
between 0.2 to 7.7wt% and 0.1 to 8.1wt%, re- (overall range) with no significant differences
145
]]%%ttww[[
nnooiittiissooppmmoocc
llaaddooMM |
ADE | 5.3 Trace element signature - trends and element affinities
5.3 Trace element signature 7550 (a)
- trends and element 7450
7350
affinities Original data
7250
Projected data
TRACE element concentrations of individual 7150 Rotated final data
samples are presented in appendix B on 7050
page 486 and are summarised for different ore 6950
Best-fit plane
types in table 5.3. Despite a large sample set
6850 y = -2.1315x + 16389
(n=118), it is not feasible to accurately deter- R² = 0.9097
6750
mine lateral trace element variability of indi-
vidual ore zones (i.e. ML, z1-z6) due to their 6650
pipe-like geometry. Thus, the sampling strat-
egy focussed on testing compositional changes Mine East [m]
versusdepthoftheentirelongitudinalextentof
the Elura orebody. If significant concentration 10200 (b)
trendsversusdepthwereidentified,aprediction
map could be interpolated for better visualisa- 10000
∗
tion . Prior to the interpolation, east and north
coordinates of the 3D data set were projected
9800
onto a best-fit 2D plane defined by all sample
coordinates, oriented along the longitudinal di-
rectionoftheorebody. Subsequently,thosepro- 9600
jections were rotated into the N-S orthogonal
plane(figure5.8). 9400
9200
5.3.1 Inter-element affinities
†
Cluster analysis was initially used to investi- 9000
gate inter-element affinities (figure 5.9). Most 6650 6850 7050 7250 7450
highfieldstrengthelements(abbrev. HFSE;Sc,
Mine (cid:6)orth [m]
Ti, Y, Zr, Nb, Hf, Ta, Th), rare earth elements
Figure 5.8: Coordinate transformation for geo-
(abbrev. REE; La, Ce), lithophile elements, in-
statisticalmodelling. (a)Planviewofsampleloca-
cluding some large ion lithophile elements (ab- tions; (b) view towards W. Transformation is per-
formed in order to reduce a 3D data set to 2D.
brev. LILE; Li, Be, Rb, Cs), and elements
Transformation consists of projection onto a best
thatmaybehostedbydetritalrefractorymineral fit longitudinal plane followed by rotation into the
phasescontainedinthewallrockarecloselyas- N-Sorthogonalaxis.
sociated with each other (i.e. V, Ni, Cr, Sn).
Those elements are summarised as group A el-
ements and are clearly related to the influence
ofwallrockcomponents. TheassociationofGe ∗.Interpolation was performed via standard inverse
distance weighted (IDW) method implemented in the
to this group of elements suggests wall rock as ESRI® ArcGISsoftwarepackage;neighbourstoinclude
its source. The close relation between Zn, Cd was set to 20 with a minimum of 10; due to the pro-
nouncedpipe-likegeometry,ananisotropyfactorof2at
and Ga is caused by sphalerite, as the most im-
an angle of 0° was applied, with the search ellipsoid de-
portant host of those elements as suggested by finedat100mverticaland50mlateral;powerofweight-
ingwasoptimisedautomatically.
mineral chemical investigations (chapter 4 on
†.Spearman’s correlation coefficients are used as at-
page 79 and page 108) and affirmed by the lin-
tribute distance measure between elements. Cluster
earcorrelationtrendbetweenCdandZn(figure analysisisperformedaftertheapproachofWard.
149
]m[
htro(cid:6)
eniM
]m[
LR
eniM
0014 0024 0034 0044 0054 0064 |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
5.22). AssociationsofZnandPbwith groupA to the sulphide mineralisation and are hosted
elementsareartificial. Bothelementsarehighly by various sulphide phases. Cluster analysis of
enrichedinalloretypesincludingsemi-massive thoseelementsisshowninfigure5.10. Hafnium
siliceous ore. Wall rock components contained andZrwereincludedtoinvestigateproximities
in siliceous ore types are enriched in group A to wall rock. Elements with close associations
elements and are causing their proximities. Di- are grouped together: As, Sb, Ag, Au, Hg, Tl
lution due to the addition of sulphide is ampli- and Sn as group B; Co, Cu, Cd, In, Bi, Se,
fyingtheinter-elementassociations. and Te as group C. Although some Ga is in-
The remaining elements are clearly related corporated in sphalerite, its relative proximity
to Hf and Zr suggests wall rock as important
source. Figure 5.10 confirms that Ni and Ge
aregeneticallyunrelatedtothebasemetalmin-
eralisation. Their respective concentrations are
low and range between 0.4-80.7ppm and 0.07-
0.39ppm. Strontium and Ba are proximal to
group B (figure 5.9), most probably caused by
barite enriched in Sr. Barite is a common but
minornon-sulphideganguephase(chapter4on
Figure 5.10: Grouping of trace elements related
tobasemetalmineralisation. Dendrogramisbased
on cluster analysis (Spearman’s correlation coeffi-
cients are used as attribute distance measure be-
Figure 5.9: Dendrogram showing inter-element tween elements; cluster analysis is performed af-
affinitiesbasedonclusteranalysis(Spearman’scor- ter the approach of Ward). Element proximities
relation coefficients are used as attribute distance are predominantly caused by their spatial distri-
measure between elements; cluster analysis is per- bution. Elements grouped in orange are enriched
formed after the approach of Ward). Elements la- towards increasing depth below surface, whereas
beledinbluearesummarisedasgroupAelements those labeled in purple feature inverse characteris-
andareclearlyrelatedtotheinfluenceofwallrock tics. MolybdenumandRearecloselyrelated,indi-
components. ProximityofZn,CdandGaiscaused catingmolybdeniteastheircommonhost. Gallium
bysphaleriteastheircommonhost(green). Other appears closer to wall rock (defined by immobile
elementsarelinkedtothesulphidemineralisation. elements)thantosulphides.
150 |
ADE | 5.3 Trace element signature - trends and element affinities
page69). alised altered CSA wall rock (MinA) and up
Molybdenum, Re and U are closely corre- to 1.1wt% in massive pyritic ore. Median Ag
lated. Molybdenite is the most important ore concentrationsofmassiveandsemi-massiveore
for the production of Re. It is well known that types (Po, Py, SiPo and SiPy) are in the or-
Re is readily incorporated in molybdenite via der of 44 to 86ppm (table 5.3). Pyritic ore
lattice substitution. Association between Mo types are relatively enriched in Ag compared
and Re is, therefore, likely caused by molyb- to pyrrhotitic dominated ore (figure 5.11). It is
denite. All samples except one feature a weak wellknownfromtheresourceblockmodelthat
but noticeable correlation between Re and Mo silver is not only enriched towards the surface
(figure 5.22). Rhenium concentrations are very but also towards peripheral zones of individ-
low, commonly close to the detection limit of ualnear-concentricsulphidepipes(figure5.14).
2ppb, causing the pronounced scatter of data This enrichment explains the elevated concen-
points. Nevertheless, the trend affirms molyb- trations within deeper parts of the orebody as
denite as host for Re as suggested by the clus- indicated in figure 5.18. Due to close affinity
ter analysis. The proximity of U with Mo as of Ag to many of the group B elements, it is
observedinclusteranalysis(figure5.9)andthe likely that those elements feature similar char-
poorly defined element correlation between U acteristics. Theconcentrationincreaseofsilver
andMo(figure5.22)maybecausedbymineral towardsthesurfaceobservedinthisstudyisnot
phases containing both elements, e.g. sedovite an effect of supergene enrichment processes as
(U(MoO ) ). onlyfreshunalteredsampleswereanalysed.
4 2
A well-defined enrichment trend towards the
surface is exhibited by antimony (figure 5.18),
5.3.2 Trace element concentra-
partially linked to Ag and caused by the oc-
tions and their spatial distri-
currence of argentian-tetrahedrite (chapter 4 on
bution
page63). TheremainingSbmayoccurasanti-
The investigation of spatial element distribu- monite. Concentrationsvarybetween9.53ppm
tions revealed the reason for the observed clus- and 0.21wt% (table 5.3) and are commonly
teringofgroupBandCelements. GroupBele- higher in pyritic dominated ore types (figure
ments tend to be preferably enriched in the up- 5.11).
per parts of the orebody. Elements of group C Gold concentrations are highly variable,
commonly feature converse trends with enrich- rangingbetween2ppbinminormineralisedore
ment towards increasing depth below surface. and as high as 8.7ppm in siliceous pyritic ore.
Gallium, Mo and Re are not clearly related to Nocleartrendversusdepthbelowsurfaceisin-
eithergroup. dicated. Figure 5.11 shows Au variability for
Arsenic concentrations range between differentoretypes. Pyriticoretypesarethepre-
12.6ppm and 1.50wt% (table 5.3). An overall ferredhostforAuatmedianconcentrationsbe-
trend of increasing concentrations towards tween approximately 0.2 to 0.5ppm, whereas
the surface is observed despite some samples pyrrhotitic ore commonly contains ~0.09ppm
withinthelowermineralisationfeatureelevated (median).
As contents (figure 5.18). Pyritic ore types Mercuryandthallium,bothimportantsmelter
(Py and SiPy) are significantly enriched in As penalty elements, are most enriched in the up-
(median of ~0.1-0.5wt%, table 5.3) compared per main lode zone but are also elevated in up-
tootheroretypes(figure5.11). per parts of the lower mineralisation, in partic-
Silver shows enrichments in two zones: (a) ular in z2 and z3 zones (figure 5.17). Samples
intheuppermostareasofmineralisedzonesand from deeper parts of the orebody that are char-
(b)inthelowermostmainlode(ML)zone(fig- acterised by high Hg and/or Tl concentrations
ure 5.18). The overall detected Ag concentra- are from peripheral ore zones (SiPo and SiPy
tions range between 1.2ppm in minor miner- ore types), featuring similar enrichment char-
151 |
ADE | 5.3 Trace element signature - trends and element affinities
ticularoretypesorspatialtrendswereobserved. at median concentrations ranging between 0.16
The element cross-plot of Re vs Mo features and3.47ppmforalloretypes. However,signifi-
a good correlation trend when considering the cantconcentrationlevelsashighas714ppmcan
factthatdetectedReconcentrationsarecloseto be detected. A pronounced enrichment trend
the detection limit of 2ppb (figure 5.22). One withincreasingdepthbelowsurfaceisobserved
sample characterised by the highest observed (figure 5.20). The highest concentrations are
Reconcentrationmiss-fitsthistrend. Thistrend, limited to the lower parts of the Elura orebody,
in addition to the close association of both ele- inparticularinthelowermostMLzone. Noel-
ments in cluster analysis (figure 5.9 and figure ement affinity to any particular ore type is ob-
5.10),suggestsmolybdeniteasthemostimpor- served. Selenium and tellurium strongly cor-
tanthostmineralforRe. relate with each other in addition to Bi (figure
5.22)andfeaturealmostidenticalspatialdistri-
Trace element investigations of sulphides
butionsasshowninfigure5.20. Concentrations
have shown that sphalerite is the most im-
ofbothelementsareratherlowwithmaximaat
portant sulphide phase incorporating gallium.
45and1.88ppmforSeandTe,respectively.
Observed concentrations via whole rock geo-
chemistry are low and range between 1.71 and Cobalt and indium concentrations vary
24.00ppm. The spatial distribution of Ga in- strongly, but feature significant enrichment to-
dicates enrichment towards peripheral zones wards depth (figure 5.19) in a similar manner
of the mineralisation. Median concentrations as to Bi, Se and Te. Although some samples
of massive ore types are slightly lower (5.8- in upper and central ore zones contain elevated
6.4ppm) compared to semi-massive (6.6-7.0) Co. Cobalt concentrations range between 0.2
and are highest in stringer ore (10.7). This ob- and 288ppm with a clear affinity to pyrrhotite
servation, in addition to the proximity of Ga to dominated ore types at higher concentrations
HfandZrinclusteranalysis(figure5.10), sug- (figure 5.12). The maximum Co content is ob-
geststhatelevatedGaconcentrationsarelinked served in stringer type ore from the lowermost
toincreasedwallrockcomponent. ML zone. Indium concentrations are very low
between 5ppb and 3.2ppm. Highest enriched
The remaining investigated trace elements
samplesclusteratthebaseoftheMLzone,sim-
feature clear enrichment trends with increas-
ilartoBi,SeandTe.
ing depth below surface. The most important
amongst them is bismuth as a smelter penalty A very weak enrichment trend with increas-
element for lead concentrate. Concentrations ing depth is present for cadmium (figure 5.21).
are rather low throughout most of the orebody Itsinversedistanceweighteddistributionmodel
Figure 5.12: Box and Whisker plots for trace elements relatively enriched in massive and semi-massive
pyrrhotiticoretypes(SiPo,Po).
153 |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
Figure 5.17: Trace element variabilities vs. depth for Hg and Tl. Yellow shape is the stringer type ore
resource domain. Interpolated element distribution maps are based on standard inverse distance weighted
(IDW)methodimplementedintheESRI®ArcGISsoftwarepackage;neighbourstoincludewassetto20with
a minimum of 10; due to the pronounced pipe-like geometry, an anisotropy factor of 2 at an angle of 0° was
applied,withthesearchellipsoiddefinedat100mverticaland50mlateral;powerofweightingwasoptimised
automatically.
appears rather erratic. Nevertheless, the model ure 5.19). The overall measured Cu concentra-
indicates a pronounced Cd enrichment in the tionrangeisdefinedbyaminimumof155ppm
lower ML zone, which is defined by a cluster in minor mineralised rock, and a maximum
of several samples high in Cd. Concentrations of 0.78wt% detected in massive pyrrhotitic
rangebetween17and446ppmandareslightly ore. Median Cu concentrations range between
elevated in pyrrhotitic compared to pyritic ore 0.16wt% and 0.24wt% in massive and semi-
(figure5.12). massive ore types. The resource block model
(RBM) estimates a pre-mining Cu grade of
Many of the investigated elements feature
0.2wt% for the entire resource, thus falls well
spatialenrichmenttrends. Oneelementexhibits
withintherangeofobservedmedianconcentra-
a repetitive enrichment trend with increasing
tions of this study. A clear shift to higher con-
depth. Copper is relative enriched in most
centrationsisobservedinpyrrhotiticore(figure
samples from the lower mineralisation com-
5.12). The lowermost part of the ML zone is
pared to those from the upper ML zone (fig-
156 |
ADE | 5.3 Trace element signature - trends and element affinities
characterised by the highest Cu accumulation. figure 5.16). This study highlights the fact that
A second zone of enrichment, although only no large sample set, large budget or much ef-
defined by one sample, is observed at approx- fort is needed to gain at least a rough overview
imately RL9700m. At this depth, the positive of trace element distributions, in particular for
Cuanomalycoincideswithlithologicalchanges smelter penalty elements, even for such a large
ofthehostlithology(seechapter3onpage35). orebodyasElura.
The spatial distribution of Cu is well known
from the RBM. Both Cu enrichment zones de-
5.3.3 Element cross-plots
scribedabovearealsovisibleandconfirmedby
the RBM. The set of a total of 117 ore sam- Severallinearcorrelationtrendswereidentified
ples is certainly rather low if representing the in element-element cross-plots (figure 5.22).
basis for geostatistical modelling. Neverthe- Mostcorrelationsarecausedbymutualelement
less, rather good agreements are met for the enrichment in particular areas within the ore-
estimated inverse distance weighed distribution body, e.g. enrichment of Hg and Tl in the up-
modelsofAg(figure5.18)andCu(figure5.19) permostpartoftheMLzone(figure5.17). Geo-
whencomparedwiththeRBM(figure5.14and chemical characteristics clearly reflect miner-
alogicalandmineralchemicalchanges.
Figure5.22: Importantelementcross-plotsindicativeformajormineralogicalchangesand/ormutualelement
enrichmentinparticularpartsoftheorebody.
161 |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
Silver correlates with Sb in most samples, 5.3.4 Reconciliation between
caused by galena-miargyrite solid solution (see geochemistry and mineral
chapter 4 on page 108) and by argentian- chemistry for selected ele-
tetrahedrite (see chapter 4 on page 93) as com- ments
mon silver mineral. One sample features el-
evated Sb but low Ag concentration, indicat- Mineral chemical compositions were deter-
ing the occurrence of other Sb-bearing mineral minedforselectedsamplesaspartofthisstudy.
phases (figure 5.22). All samples elevated in Trace element concentrations of sphalerite and
Bi are characterised by well-correlated Ag and galenawereusedtoreconcilemineralchemistry
Bi concentrations (figure 5.22), reflecting solid to whole rock geochemistry based on modal
solution of matildite in galena as described in mineral composition, calculated from whole
chapter4onpage108. Silver-richsamplesthat rockgeochemicaldata.
lack significant Bi content contain argentian- Sphalerite is the most important carrier of
tetrahedrite, native silver, or reflect solid solu- Cd and In, although In occurs in very minor
tion of miargyrite in galena (see above). Thal- quantities of up to ~5ppm (chapter 4 on page
lium and Hg correlation is due to their strong 108). The reconciliation plot for Cd (figure
enrichment in the uppermost parts of the ML
zone, their host mineral phase is uncertain. In
500 Cd
contrast, the correlation between Bi, Se and
[ppm]
Te reflects their mutual enrichment in the low-
ermost parts of the ML zone, predominantly
400
hostedbygalena(figure5.22).
300
y = 0.9433x + 49.929
R² = 0.6855
200
200 300 400 500
10200
2 Conc. based on I n
10000 mineral chemistry [ppm]
1.5
9800
1
9600
0.5
99440000 y = 2.1617x + 0.0851
R² = 0.833
0
9200
0 0.5 1 1.5 2
Conc. based on
9000
mineral chemistry
0.01 1 100
Bi [ppm] Figure 5.24: Reconciliation of mineral chemistry
to geochemical data (Cd, In). Both elements are
Figure 5.23: Reconciliation of mineral chemistry predominantly incorporated in sphalerite at the
to geochemical data for Bi. Bismuth concentra- Elura orebody. Their concentration were deter-
tions incorporated in galena were determined via minedviaLA-ICP-MS.Blackdashedlinesareideal
EMPA.Redcirclesshowwholerockconcentrations 1:1ratio;solidblacklinesareestimatedlineartrend
calulated based on mineral chemistry; blue circles lines;R2 isthecoefficientofdetermination.
are concentrations determined via whole rock geo-
chemistry. Concentrations are plotted vs. depth
belowsurfaceasreducedlevels(RL).
162
]m[
LR
eniM
no
desab
.cnoC
no
desab
.cnoC
yrtsimehcoeg
kcor
elohw
yrtsimehcoeg
kcor
elohw |
ADE | 5.4 Host rock classification
elements, and Th are stronger enriched at fac- La-Th-Sc discrimination scheme (figure 5.29).
torsofapproximately6,3and2.5,respectively. The samples are more scattered in the Th-Sc-
Therelativemobileelementstrontiumissignif- Zr/10 diagram and crossover from the afore-
icantlydepletedbythefactor0.3. mentioned into the active continental margin
field. Theseresultsareinagreementwiththose
WhitbreadandMoore(2004)investigatedthe
basedonthediscriminationafterRoserandKo-
cryptic alteration halo within wall rock around
rsch(1986).
the deposit. They observed enrichment trends
Figure5.30showschondritenormalisedREE
ofCsandRbtowardsthemineralisation,caused
distribution pattern for unaltered and altered
by their incorporation in muscovite. Strontium
wall rock samples compared to turbidite and
is significantly depleted proximal to the ore-
marine shale reservoir data (REE pattern based
body, caused by its replacement in carbonates
onCSAsampledatapresentedintable5.6). The
byiron. Macroscopically,thesesamplesarenot
observed distribution patterns are very similar
affectedbyanysignificantalteration. However,
to marine shale. The better fit with shale than
selective trace element enrichment and deple-
withturbiditedatasuggeststhatREEcontentin
tion observed in the course of this study are
the wall rock sequence is predominantly con-
inagreementwithearlierinvestigationsandare
trolledbyREEscontainedinsilt-andmudstone,
caused by subtle mineralogical changes caused
sandstonecomponentappearssubordinate. This
by alteration. The results highlight the signifi-
is in agreement with studies by Cullers et al.
cant extent of cryptic alteration to distances in
(1987) that showed that REE patterns in sedi-
excessof200mtotheorebody.
mentaryrocksuitsarepredominantlycontrolled
Bhatia and Crook (1986) proposed tectonic by the clay-sized mineral fraction. The good
discriminationdiagramsforgreywackes,which correlation with reference data confirms that
represent a significant component within the REE analyses are valid and that the minor de-
host lithology as sandstone beds. All wall rock viation in Ce in standard analyses (figure 5.3)
samples, including the altered sample, closely hadnosignificanteffectonaccuracy.
cluster in the continental island-arc field in the
10
Wall rock samples
Altered wall rock
1
Marine Shale Reference
Upper continental crust
LLoowweerr ccoonnttiinneennttaall ccrruusstt
0.1
Cs Rb Ba Th U K Ta Nb La Ce Sr P Hf Zr Sc Ti Y
Figure5.31: Multi-trace-elementdiagramcomparingwallrocksamplestoreferencedata. Dataisnormalised
toaveragecontinentalcrust(Wedepohl,1995). UpperandlowercontinentalcrustreferencedataafterWedepohl
(1995),marineshaledataafterYuan-Hui(1991).
169
ttssuurrcc
llaattnneenniittnnoocc
..ggvvAA//eellppmmaaSS |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
5.5 Mass/volume and
element gain-loss
estimation
FLUIDS, introduced into and migrating
throughlithologies, willinalmostanycir-
cumstance result in its modification, as long
as fluid and bulk rock chemical compositions
El El El El El El El
are in disequilibrium. Modifications may oc- imm-A X-1 X-2 imm-B X-3 X-4 imm-C
cur via metasomatism, mineral replacements Scaled element concentrations of less-or
unaltered reference sample
and the formation of new mineral species.
The subsequent change of bulk rock chemi- Figure 5.32: Principles of the isocon method.
Scaled element concentrations are plotted against
cal composition is accompanied by mass and
each other. Immobile elements (open black cir-
volume changes. Gresens (1967) described cle, Elimm−A to Elimm−C) plot on a straight
linepassingthroughtheoriginanddefine"isocon"
composition-volume relationships caused by
lines. isoconslopesbelow1reflectmassgains,vice
metasomatismbycomparingmodalmineralogy versaforslopesgreaterthan1. Relativeconcentra-
aswellaswholerockcompositionsofunaltered tionchangesofmobileelements(filledbluecircles,
and altered samples. The author’s fundamental
ElX−1 to ElX−4) compared to reference sample
is defined by their deviation from the isocon lines
argument is that some elements are immobile andreflectrelativelyenrichmentiflocatedaboveor
during alteration, e.g. due to their incorpora-
depletionifbelow. ElX−1 featuresrelativeenrich-
ment for both example isocons compared to refer-
tion in highly refractory mineral phases not af- ence sample. ElX−2 is enriched for the mass gain
isocon whereas relative immobile for the mass loss
fected by metasomatic processes. When com-
isocon. ElX−3 is depleted for both, ElX−4 de-
paring concentration changes of identified im- pleted and enriched for mass gain and mass loss
mobileelements,massandvolumechanges,ac- isocon,respectively.
companyingalteration,canbeestimated. Grant
(1986)revisitedGresens(1967)propositionand
formulated a simplified method to describe
mass/volume changes during metasomatic pro-
sample(CRef )aredefinedaccordingtoeq. 5.6.
i
cesses and called it the isocon diagam, where
Byinspectingthedistributionofscaledconcen-
isoconsdescribestraightlinesofequalelement
trationsofpotentialimmobileelements(Al,Ti,
concentrations.
Y,Zr,Nb,HfandTainthisstudy),abest-fitiso-
conthroughtheoriginmaybeconstructed. Ac-
cording to Grant’s equation (eq. 5.7), which is
5.5.1 The Isocon-method
arearrangedversionofGresens’,thereciprocal
Theprincipleoftheisocon-methodisvisualised slope(k iso)oftheisoconisequaltotherelative
in figure 5.32. Scaled concentrations are used mass change (∆M) during alteration, because
foralteredandless-orunalteredreferencesam- theconcentrationchange(∆C i)ofimmobileel-
ple. Scaling is performed as described in Hus- ements is 0 (eq. 5.8). An isocon with k iso <1
ton (1993) in order to improve the presentation reflects mass gain, k iso >1 mass loss. If den-
ofisocondiagrams. Elementswereorderedac- sity values of altered (ρA) and reference sam-
cording to their expected mass changes begin- ple (ρRef) is known, relative volume changes
ning with those characterised by largest mass (∆V; eq. 5.9; Huston, 1993), and assuming
gains. Immobile elements were evenly dis- isotropic volume change, relative dimensional
tributed. Subsequently, an integer (n ) was as- change (∆L; eq. 5.10; Huston, 1993) can be
i
signed to each of the elements (i) and a scaling calculated. Relative concentration changes for
factor (F ) calculated (eq. 5.5). Scaled concen- other elements are calculated according to eq.
i
trations of altered samples (CA ) and reference 5.7.
i
170
snoitartnecnoc
tnemele
delacS
elpmas
deretla
fo |
ADE | 5.6 Rare earth element signatures of different ore types
wall rock fragments, preferably incorporate the of wall rock component. Massive ore samples
LREEs (Cullers et al., 1975; Laul and Lepel, yield maximum REE depletions in the order of
1987). 0.01, suggesting only a minor wall rock con-
tent. The consistently observed pronounced Eu
Considering mineralogy, remnant wall rock
anomaliesinsemi-massiveandmassiveoreim-
fragments, if contained in sulphide ore in suffi-
ply a rather reducing depositional environment
cient quantities, are clearly the main contribu-
and/or elevated fluid temperatures in excess of
tors to the observed REE distribution patterns.
250°C(Sverjensky,1984). AbsentorminorEu
If wall rock content within ore is low, these in-
anomalies in breccia-stringer type ore may be
herited patterns may, however, be masked by
caused by insufficient hydrothermal carbonate
REE-bearing mineral phases that precipitated
precipitation or deficient Eu2+ transported in
during ore formation, i.e. non-sulphide gangue
the hydrothermal fluid. The latter case would
(abbrev. NSG;ironcarbonate,quartz). Thehy-
beindicativeforadropintemperatureorchang-
drothermal fluid may either cause further mod-
ingredoxstateofthemineralisingfluidtowards
ifications via alteration of wall rock fragments,
higheroxygenfugacities.
accompaniedbytheformationofsericiteand/or
chlorite, or, if fluid rock ratios are extremely
ThreegroupsofREEdistributionpatterndif-
high,byremobilisationofREEs(Bau,1991).
ferenttowallrockwereidentified. Representa-
The pronounced decrease of total REE con- tive examples of these patterns are summarised
tent observed in different ore types (table 5.5), in figure 5.41. The example patterns are based
inadditiontotheLREE/MREE/HREEfraction- on average REE concentrations, which were
ation trends (figure 5.38) indicate a decrease normalised to average wall rock composition
1000 1000 1000
LREE depletion Similar to wall rock LREE enrichment
100 100 100
10 10 10
1 1 1
0.1 0.1 0.1
10 10 10
1 1 1
0.1 0.1 0.1
0.01 0.01 0.01
0.001 0.001 0.001
Figure 5.40: RepresentativeexamplesofdistinctREEdistributionpatternssignificantlydifferentto
unalteredCSAwallrock. Leftdiagram: significantdepletionofLREEandtosomeextentofMREE.
Middlediagram: REEdistributionsarerathersimilartowallrock,althoughslightlyenrichedinMREE.
Rightdiagram: patternsarecharacterisedbyLREE/MREEenrichment. Reddashedlinesareaverage
CSAwallrockREEdistributions. Greenlinesareaveragecompositionforindividualpatterns,shown
in figure 5.41. Data has been normalised to chondrite (upper diagrams; chondrite data after Palme
and O’Neill (2004)) and to post-Archian Average Australian Shale (lower diagrams; abbrev. PAAS;
normalisationdataafterNanceandTaylor(1976)).
179
etirdnohC/elpmaS
SAAP/elpmaS
aL eC rP dN mS uE dG bT yD )oH( rE mT bY uL
etirdnohC/elpmaS
SAAP/elpmaS
aL eC rP dN mS uE dG bT yD )oH( rE mT bY uL
etirdnohC/elpmaS
SAAP/elpmaS
aL eC rP dN mS uE dG bT yD )oH( rE mT bY uL |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
10
10
1
1
0.1
0.1 0.01
Figure 5.41: Three representative examples of Figure 5.42: REE content in carbonate refer-
distinctREEdistributionpatternssignificantlydif- ence data. Red solid line: ankerite after Davies
ferent to unaltered CSA wall rock. Pattern A et al. (1998); red dashed line: siderite after Bau
(red) features significant depletion of LREE and andMöller(1992);bluelinesareREEdistribution
to some extent of MREE. REE distributions are patternsofcalcite: solidafterDavieset al.(1998),
rather similar to wall rock in pattern B (blue), dashed after Bau and Möller (1992). Data is nor-
although slightly enriched in MREE. Pattern C malised to post-Archian Average Australian Shale
(green)ischaracterisedbysignificantLREEenrich- (PAAS)afterNanceandTaylor(1976).
ment. Data has been normalised to average CSA
wall rock composition and to Lu (yellow symbol).
Datausedinthisdiagramareaveragecompositions
showninfigure5.40asgreenlines.
to pattern B, indicative for samples dominated
in addition to Lu in order to visualise the rel-
by ankerite and/or calcite. Pattern C is exclu-
ative deviation from the host lithology. The
sively observed in pyritic ore types (SiPy and
differences are predominantly defined by vari-
Py). Both ore types commonly contain sericite
able intensities of depletion and/or enrichment
andchlorite,mineralphasesenrichedinLREE.
of the LREE and/or MREE. HREE are similar
Pronounced sericitisation and chloritisation is
towallrock. Onegroupofsamplesfeaturesig-
thus the most probable explanation for pattern
nificant depletion of LREE and to some extent
C.
of MREE (Pattern A). REE distributions rather
The well-defined and strong Eu anomaly in
similar to wall rock, although slightly enriched
additiontotheenrichmentofotherMREEsand
in MREE, are observed for a second group of
LREE in carbonates, sericite and chlorite im-
samples(PatternB).LREEaresignificantlyen-
plies that the hydrothermal fluid carried sig-
riched in patterns of in the third sample group
nificant REEs in solution. As the hydrother-
(PatternC).
malfluidascendedfromdeeperpartswithinthe
WhencomparingpatternAandB(figure5.41 Cobar Basin, REEs were most probable mo-
with REE distributions within siderite, ankerite bilised from the basinal stratigraphy via alter-
and calcite reference data (figure 5.42), it be- ationofrelativelyunstablemineralphases,such
comesclearthattheREEdistributionsaresolely asfeldsparandcarbonatephases.
controlledbycarbonates. SideritefeaturesREE ThisstudyshowedthatmodificationsofREE
pattern significantly diluted in LREE relative distribution pattern associated with hydrother-
to MREE and HREEs. Pattern A can there- mal fluid events at the Elura orebody are ex-
forebeexplainedbysamplesrelativelyenriched clusively caused by changes in mineralogy. No
in siderite. On the other hand, REE distribu- local REE remobilisation caused by extreme
tions of calcite and ankerite are relatively flat, fluid-rock interaction is needed in order to ex-
butmaybemoderatelyenrichedinMREE.Pat- plainpronouncedLREEdepletion.
terns of those two mineral phases are similar
180
kcor
llaw
morf
noitaived
.gvA
uL
ot
desilamron
aL eC rP dN mS uE dG bT yD )oH( rE mT bY uL
SAAP/setanobraC
aL eC rP dN mS uE dG bT yD )oH( rE mT bY uL |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
5.7 Platinum group
10
elements and the time of
ore formation
1
5.7.1 PGE concentrations and 0.1
their fractionations
0.01
The elements of the platinum group are with
increasing order of natural abundances Rh, Ir, 0.001
Os, Pd, Ru and Pt (based on data of Mc- Os Ir Ru Rh Pt Pd Re
Donough and Sun, 1995). The PGEs are
100
part of the D-block elements that mostly con-
sists of siderophile elements featuring simi-
lar atomic radii. The PGEs and Re have a
10
highlysiderophilecharacterandpreferablypar-
titioned into Fe-rich melts during the early ge-
netic stages of the primeval earth, as its differ- 1
entiation commenced. Subsequently, these el-
ements were incorporated into the earths’ core
0.1
during its formation, responsible for the very
Os Ir Ru Rh Pt Pd Re
low PGE abundances in mantle and crustal
Figure 5.43: Normalised PGE distribution di-
lithologies(FaureandMensing,2005;McInnes
agrams for massive sulphide ore samples. Red
et al., 2008). New chondritic material was patternarepyritedominatedoretypes,thosein
added during the late heavy bombardment at blue are pyrrhotitic ore samples. Dashed-black
lineistheaverageofall6samples. Datawasnor-
approximately 3.9 Ga, causing a re-enrichment
malisedtoprimitiveuppermantle(PUM)andto
of PGEs in the upper mantle. The PGEs are average continental crust using values after Mc-
Donough andSun (1995) andWedepohl (1995),
grouped according to their different geochemi-
respectively.
cal behaviour. The Pd-subgroup (PPGE), con-
sisting of Pd, Pt and Rh, features a stronger
affinity to sulphide phases than the elements IPGEsareratherlowandrangebetween29and
of the Ir-subgroup (i.e. Ir, Os, Ru; abbrev. 136ppt. Onaverage,theconcentrationsslightly
IPGE). The less chalcophile IPGEs preferably increase from Os (60ppt) to Ir (69ppt), to Ru
form alloys and are commonly associated with (74ppt)andRh(97ppt). Platinumisthehighest
chromites, whereas PPGEs preferably occur in enriched PGE with concentrations ranging be-
sulphide phases, or as variable separate PPGE tween 2.5 and 42.1 ppb (avg. 12.5ppb). Palla-
phases. PPGEaregenerallypartitionedintothe dium concentrations are significantly low com-
silicate melt, during partial melting of mantle paredtoPtatanaverageof0.5ppbandamaxi-
material, thus fractionating from the IPGE that mumof1.5ppb.
remainpreferablyinthemantleresidue(Barnes A clear fractionation between Pt and Pd is
etal.,1985). showninfigure5.43. Theprimitiveupperman-
The analyses of two samples (NP886-2 and tle (PUM) normalised PGE distribution pattern
560-MLR-S; figure 5.4) feature high analyti- (upper diagram in figure 5.43) features a rather
cal errors. Those results were rejected and flat pattern for the IPGs, which are depleted by
are not presented in this study. The results a factor of ~0.02 (average PGE composition of
of the remaining 6 samples are shown in ta- n=6 samples). Rhodium and Pd are with a
ble 5.7. The total PGE content ranges between factor of ~0.1 less depleted than IPGs. Plat-
3.9 and 49.2ppb with most being contributed inumrepresentstheexceptionofallPGEs,char-
byPtatapproximately90%. Concentrationsof acterised by almost PUM composition. PGE
182
MUP/elpmaS
.gva/elpmaS
tsurc
latnenitnoc |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
5.7.3 The isochron age and poten-
tial osmium sources
The analytical results of 6 massive sulphide
samples are presented in table 5.9. The ob-
servedtotalReandOsconcentrationsare0.802
to 7.580ppb and 0.039 to 0.103ppb, respec-
tively. Relative standard deviations (RSD) are
below1.5%. Thepresentday187Os/188Osratios
are commonly high and reach maxima in ex-
cessof3.5. Thehighlyradiogenicsamplesyield
present day γ values in the order of ~550 to
Os
2800 compared to present CHUR 187Os/188Os
ratios. RSDs of present day 187Os/188Os ratios
arehigher,particularlyforonesample(NP243)
Figure 5.45: Re-Os isochron diagram based on
with12.6%(table5.9).
whole rock geochemical analyses of 6 massive sul-
All 6 samples define a model 1 (Ludwig, phide ore samples. Isochron age and associated
data was calculated by use of the Geodate soft-
2003) isochron age of 377.94±14.98Ma (2σ)
warepackage(EglingtonandHarmer,1999). Data
with a MSWD of 0.237 and a probability of defines a model 1 isochron (Ludwig, 2003) age of
377.94±14.90Ma (2σ). Data-point error ellipses
0.964 (figure 5.45). The spread of data points
are 1σ; dashed lines define the 2σ isochron uncer-
is sufficient, although 3 samples cluster at a tinitycorridor.
187Re/188Os value of just over 500. There-
fore,theisochronmaybeviewedasbeingcon-
strainedbyonly4points,insteadof6. Aninitial suggestedbySelbyetal.(2009),mostlikelybe-
187Os/188Os value of 0.338±0.092 (2σ) is de- came significantly enriched in Re, bound to or-
finedbytheisochron, yieldinganinitialγ of ganicmatter,relativetoOs. Rheniumisreadily
Os
168.46, calculated based on time of formation incorporated in molybdenite during ore forma-
(378Ma) and the 187Os/188Os of the CHUR at tion that in turn decayed to radiogenic Os, sub-
thistime(eq. 5.13andeq. 5.14). Theinitialγ sequently causing the pronounced radiogenic
Os
valuesforindividualsamplesrangebetween68 signatureoftheEluraOrebody.
and 186. The percentage of 187Os from total Initial γ values are commonly used to
Os
Os ranges between 10 and 33%. The observed identify the extent of crustal components in-
present day γ Os values are highly radiogenic. volved during the genesis of ultramafic/mafic
Approximately 97% of all 187Os has been ra- magmatic ore deposits, i.e. genetically clearly
diogenically formed (187Osr; table 5.9) since linkedtoprimitivemantlesources(e.g. Stillwa-
thetimetheRe-Osisotopesystemclosed,asde- ter, Bushveld, Pechenga, Sudbury, etc.). Some
finedbytheisochronage. of these deposits feature significant radiogenic
Organic matter is a common minor con- initial 187Os/188Os composition. The Kam-
stituent within the host rock of the Elura mikivi Sill in the Pechenga Complex, for in-
(Sun et al., 2001) and probably occurs finely stance, has a γ of +251, some parts of the
Os
dispersed throughout most of the basin fill Sudbury Igneous Complex feature γ values
Os
metasediments. Hydrocarbons have been iden- as high as +814 (Shirey and Walker, 1998).
tified as an important component (fluid inclu- When comparing the initial γ of the Elura
Os
sion investigation of this study and e.g. Sec- orebody with reservoir data, it falls within the
combe, 1990; Lawrie et al., 1999) of the min- verylowerendofcontinentalcrustcomposition
eralising fluid. Interaction of the hydrothermal field(figure5.46). However,γ appearsrather
Os
fluid with organic-rich rock sequences would lowwhenconsideringtheratherradiogenicγ
Os
have created/mobilised hydrocarbons and, as values of the aforementioned ultramafic/mafic
186 |
ADE | 5.7 Platinum group elements and the time of ore formation
arc environment, characterised by pronounced
crustal extension and rifting (see chapter 2 on
page 17 and page 20). Magmatism accompa-
niedthebasinformation,reflectedbytheoccur-
rence of volcano-clastics, in particular within
theMountHopeandRastTroughsasthesouth-
ern extensions of the Cobar Basin, and the in-
trusion of post-orogenic Late Silurian granites.
Major crustal scale faults may have developed
in this geodynamically active environment and
mayhaveevenreachedthelowercrust-subcon-
tinental lithospheric mantle (SCLM) boundary,
allowing fluids/melts carrying relatively primi-
tive non-radiogenic Os to escape and to ascent
to shallower crustal levels. Such a primitive
Os contribution may explain the “relative non-
radiogenic” nature of Elura’s massive sulphide
ore.
Lithologies occurring in the Lachlan Fold
Figure 5.46: Diagram showing Os isotope vari-
BeltareofCambriantoEarlyDevonianage(see
ation (γOs) in various lithologies/reservoirs com-
paredtoElurasamples. γOsdataiscalculatedwith chapter 2 on page 10), relatively young com-
a187Os/188Osratioof0.127forpresentdaymantle
pared to mineralisation that formed at 380Ma
andsourcedfromShireyandWalker(1998);Pear-
son(1999). InitialγOs dataforEluraasshownin (figure 5.45). The basin is predominantly
table5.9. filled with silici-clastic rock suits, which were
sourcedfromthehinterland. Lithologieswithin
the basin and/or the underlying basement have
ore deposits. McInnes et al. (2008) interpreted beensuggestedasthemostlikelymetalsources
γ Os valuesbelow500generallyasrelativenon- for base metal mineralisation in the Cobar re-
radiogenic and derived from a non-radiogenic gion (see chapter 3 on page 50). Osmium and
source. Keays et al. (2006) observed an ini- Re are to some extent mobilised by hydrother-
tial187Os/188Osof0.16insulphideoresamples
malfluids(e.g.Marcantonioetal.,1994;Xiong
of the Century Pb-Zn-Ag deposit. The Os iso- andWood,1999). Themineralisingfluidwould
toperatiocorrespondstoγ Osof+35(calculated have inherited the 187Os/188Os of the rock se-
basedonCHURafterChenetal.(1998)andhas quences it interacted with as it had passed
beendescribedbytheauthorsasnon-radiogenic through during its ascent. Therefore, the initial
and mantle-like. The relative low γ Os value of 187Os/188Os or the corresponding γ Os value of
this study is clearly not as “mantle-like” as ob- theorebodyshouldreflecttheOsisotopesigna-
servedattheCenturydeposit,butmaystillrep- tureofthesourcerocks.
resent mixing of Os derived from a primitive
The calculation (eq. 5.13 and eq. 5.14) of a
and a crustal source. Another explanation for
theoreticalγ foranaverageuppercontinental
Os
the observed relatively non-radiogenic Os sig-
crust (data after Saal et al., 1998, Re=0.4ppb,
naturemaybearelativelyyoungcrustalsource. Os=0.05ppb,187Os/188Os=1.9256)relativeto
Juvenile crust, despite enriched in Re, features
CHUR (Chen et al., 1998) at the time of ore
relative low 187Os/188Os rations due to short
formation 380Ma ago (based on isochron fig-
residencetimeofRe.
ure 5.45) resulted in a highly positive value of
The Cobar Basin, part of the Lachlan +1190,byfarhigherthantheinitialγ of+169
Os
Fold Belt, developed at the northern end of of ore from the Elura. The value clearly does
the Wagga-Omeo-Metamorphic-Belt in a back- notcorrespondtotherelativenon-radiogenicOs
187 |
ADE | 5.7 Platinum group elements and the time of ore formation
youngest estimate may represent the oldest studies. This date of 378±15Ma represents
crustal lithologies in the Lachlan Fold Belt. the first age constraint established via directly
However, the overall age range suggests that dating the sulphide mineralisation. Model age
Lachlan lithologies are not old enough in order calculations showed that juvenile continental
torepresentafeasiblesinglecrustalsource. crust of the Lachlan Fold Belt has the potential
to cause the observed relatively non-radiogenic
Sun et al. (2003) suggested that Re con-
initialγ value,aslongasReconcentrationin
centrations of the upper continental crust are Os
thecrustisclosetoconcentrationssuggestedby
significantly higher than previously proposed.
Sun et al. (2003). The results, however, do not
Based on undegassed arc rocks, the authors es-
precludeacontributionofprimitiveOssourced
timated the average concentration at approxi-
fromthemantle.
mately2ppb. Rheniumconcentrationsthathigh
The investigation of PGEs and the applica-
would lead to a much quicker accretion of ra-
tionoftheRe-Osisotopegeochronometerinthe
diogenic187Os,thusyoungeragesofformation.
courseofthisstudywerebasedonaverysmall
ThecalculationofmodelagesusingRe(2ppb)
samplesetandutilisedaquick,inexpensivebut
and Os (0.05ppb) concentrations suggested by
relatively imprecise analytical method. Never-
Sunetal.(2003)andSaaletal.(1998),respec-
theless, the results are conclusive and highlight
tively, resulted in an age range between ~470
that age determination of whole rock sulphide
and ~416Ma. This data is well within the es-
samples, containing ultra-trace level PGEs, is
timated age range of crustal rocks occurring
possible without the need of highly accurate
withintheLachlanFoldBelt.
negative thermal ionisation mass spectroscopy
The Re-Os isochron age of 378±15Ma, de- (NTIMS). If a more precise age determination
spite featuring a relative high uncertainty, is in is desired, a larger sample set should be used,
agreementwithandvalidatestheformationage andNTIMStechniqueapplied.
determined via 40Ar-39Ar (376-379Ma; Sun,
2000; Sun et al., 2000) proposed in previous
189 |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
5.8 Summary of
highestvaluesinVEINorewhichiscommonly
geochemistry and in excess of 50wt%. By definition VEIN type
orecontainsbetween3and10wt%Zn-Pbcom-
integration of ore
bined grade. If ore of this type is close to the
petrology
upperlimitoftheaforementionedconcentration
rangeitmaylikelybeeconomictomineifmetal
Geochemical ore characterisation showed var- prices are sufficiently high and if exclusively
ious differences between massive (Py and Po) contained metal (i.e. Zn, Pb, Ag, Cu) is con-
and semi-massive (SiPy and SiPo) ore types. sidered. However,itisimportanttoaccountfor
Breccia-stringer type ore (VEIN) features sig- the negative effect caused by the significantly
nificantly different characteristics due to its increasedorehardness. Itsimpactonmineeco-
much higher wall rock component. Based nomicsneedstobecarefullyassessedinorderto
on median concentrations, massive and semi- accuratelydetermineresourcecut-offgradesfor
massive ore contains in the order of 9.6 to VEIN ore containing high content of silicates
10.2wt%Znand5.1to5.9wt%Pb. Thehighest andlowtotalsulphide.
Znconcentration(16.4wt%)wasdeterminedin Bulkoredensitiescalculatedbasedonmodal
massivepyrrhotiticore,whereasmassivepyritic mineral composition are close to those gravi-
ore contained maximum Pb (30.3wt%). The metrically determined. Calculated densities
large overall concentration range of Zn (4.6- are with 4.7gcm−3 highest for massive ore
16.4wt%) and particularly Pb (1.7-30.3wt%) and slightly lower for semi-massive ore at
concentrations in massive ore reflects the pro- 4.4gcm−3. Breccia-stringer style ore yields
nouncedheterogenictexturalandcompositional calculated median densities in the order of
nature of the Elura ore, also observed in petro- 3.4gcm−3.
graphicinvestigations.
Silver and Cu feature inverse affinities to Apartfromelementsthatareobviouslyenriched
pyritic (Py and SiPy) and pyrrhotitic (Po and intheorebody,e.g. Zn,Pb,AgandCu,analysis
SiPo) dominated ore. The former element is ofinter-elementaffinitiesidentifiedCo,As,Se,
commonly enriched in Py and SiPy at a me- Mo,Cd,In,Sn,Sb,Te,Re,Au,Hg,TlandBias
dian concentration range of 86 to 88ppm and clearly linked to the base metal mineralisation.
amaximumofinexcessof1000ppminPyore. Nickel and Ge, despite being elements com-
Coppercontentisslightlyelevatedinpyrrhotitic monly incorporated in sulphides, are predomi-
ore (0.21-0.24wt%; median values) compared nantly associated with and caused by wall rock
topyriticore(0.16-0.20wt%). Arsenicconcen- components. Depite sphalerite being a com-
trations are generally variable but may reach mon host mineral for Ga, the element features
significant concentrations in pyritic ore up to stronger affinities to wall rock than to the sul-
1.5wt%. phide mineralisation. Most high field strength
Total sulphide, silicate and carbonate con- and large ion lithophile elements are unrelated
tent are of great importance for mineral pro- tothebasemetalmineralisation. Theirelevated
cessingduetotheirsignificantlydifferenthard- concentrationsinsulphideorearecausedbyin-
nesses. High content of hard components, i.e corporatedwallrockfragments.
quartz, increases the milling time, the power Whole rock trace element geochemistry re-
consumption and lowers the mass throughput vealed several clear concentration trends as
in the mineral processing facility. Ore elevated a function of depth within the mineralising
in sulphides improves mineral processing per- system as well as element cross-correlations.
formance as they are commonly soft mineral Changesinmineralcompositionasdescribedin
phases. Median silicate content of massive ore the previous chapter were used to link and rec-
is low and yields ~2wt%, increases to a mod- oncilegeochemistrytomineralogy.
erate~13wt%insemi-massiveoreandreaches The identified element cross-correlations are
190 |
ADE | 5.8 Summary of geochemistry and integration of ore petrology
either caused by their common host mineral for S. These observations are confirmed by
phase and/or by their common spatial occur- reconciliation between geochemistry with min-
rence within the Elura orebody. Cadmium was eral chemistry of galena determined via LA-
consistently detected in sphalerite during min- ICP-MS.Samplesthatcontainminorargentian-
eralchemicalanalyses(EMPandLA-ICP-MS). tetrahedrite feature a good correlation for Ag.
Good correlation between Cd and Zn concen- Similarly, Bi and Te concentrations reconcile
trations in whole rock geochemistry suggests well. Minor Te may occur in phases other than
sphalerite as the most important host mineral galena as indicated by a weak excess in geo-
phase. Reconciliation with sphalerite mineral chemical data. Selenium does not correlate as
chemistry based on calculated modal mineral- itmostprobablysubstitutesforSinseveralsul-
ogy showed a good agreement but some whole phidephases.
rock samples feature an excess of Cd. A poor InadditiontoAg,CuandAs,whichhaveal-
correlation was observed in the reconciliation ready been mentioned earlier, further elements
of In contained in sphalerite. Trace levels of are preferably enriched in either pyrite (Py and
Cd and In were observed in galena but are too SiPy) or pyrrhotite (Po and SiPo) dominated
lowtoaccountforthedeficiency. Minoroccur- ore types. Elements that feature affinities to
rences of unidentified mineral phases contain- pyritic ore are commonly enriched in periph-
ingeitherofthoseelementsarethemostproba- eral zones of the orebody and particularly in
bleexplanationforthemismatch. theuppermostorezones,thusfeatureclearen-
Molybdenite readily incorporates Re into its richmenttrendstowardsthesurface. Thoseele-
crystal lattice. Correlation between Mo and Re mentsareAgandSbassociatedwithargentian-
indicates its presence in small quantities and tetrahedrite, Hg and Tl caused by pyrite en-
most likely small grain sizes as it was not en- richedintheseelements,Asduetothepreferred
counteredduringmicroscopicinvestigations. occurrence of arsenopyrite in upper ore zones,
A clear correlation is observed between Tl as well as Mo, Sn and Au. Gold was com-
and Hg for the upper observed concentration monly detected at low concentrations with an
range,whichisexhibitedbysamplestakenfrom overall median range for all different ore types
theuppermostpartsoftheorebody. Theircom- between 0.09 and 0.48g/t. In places, signifi-
monenrichmentiscausedbypyritethatfeatures cant concentrations were encountered in semi-
elevated Tl and Hg content in those ore zones. massive pyritic ore up to 8.7g/t (overall maxi-
AlowerHgbackgroundconcentrationinwhole mum). Copper,CoandCdarecharacterisedby
rocksupto~20ppmisduetosphalerite. a moderate enrichment in pyrrhotitic ore com-
Argentian-tetrahedriteisthemainsilverbear- paredtopyritedominatedorevarieties.
ing mineral phase in most areas of the ore- Inversetrends,i.e. enrichmenttowardsdepth,
body. Galena-matildite and galena-miargyrite are featured by the three aforementioned ele-
solidsolutionsareresponsibleforstrongenrich- ments in addition to In, Se, Te and Bi, which
ments of Bi/Ag and Sb/Ag in galena contained arehighestenrichedinthelowermostmainlode
in samples from the lowermost and uppermost orezone.
main lode ore zones, respectively. Good cross- Distinctly different geochemical characteris-
correlationsbetweentheseelementpairsareob- ticswhereidentifiedfor(a)theuppermainlode
served in geochemical data and confirm min- zone and (b) the lower sheet of mineralisation,
eral chemical data, as well as galena as impor- consisting of the lower main lode zone and the
tant host mineral for silver, and most probably northern ore pod z1 to z5. Indium, Se, Te and
theonlymineralphasecontainingBi. Sulphide Bi are relatively enriched in the latter, almost
ore samples enriched in Bi commonly feature undetectable in former parts of the orebody.
elevated Se and Te concentrations, caused by Copperfeaturestwomajorareasofenrichment.
the occurrence of unidentified sulfosalt species Firstly, it tends to be enriched throughout most
or by stoichiometric lattic substitution of Se of the deepest parts of the orebody and, in
191 |
ADE | 5. GEOCHEMICAL ORE CHARACTERISATION
places, in pyrrhotitic zones representing the Rare earth elements (REE) contained in the
central ore zones of the individual sulphide sulphide ore feature significant modifications
pipes. A second enrichment is observed at to their distribution pattern when compared to
the transition between the aforementioned unaltered host rock. It was shown that these
upper and lower parts of the mineralisation (a changes are exclusively controlled by modal
and b), coinciding with lithological changes mineralogy and are predominantly caused
in the host lithology, i.e. the occurrence of by the formation of muscovite, chlorite and
the upper laminated unit. This enrichment is carbonate gangue phases. Sulphide phases
reflected by results of this study and confirmed represent strong dilutents as they incorporate
by the resource block model. Antimony is negligible amounts of REEs. No local REE
the only element characterised by converse remobilisation caused by extreme fluid-rock
spatial distribution. High concentrations were interaction is needed in order to explain the
consistently encountered throughout most of observed characteristics. Pronounced positive
the upper main lode zone, whereas only the Eu anomalies indicate the existence of divalent
uppermost areas of the lower mineralisation Eu that was readily incorporated in carbonates
showedweakSbenrichment. as a major non-sulphide gangue phase. The
prevalenceofEu2+overEu3+isindicativefora
Theisochonmethodwasusedinordertoassess reducingfluidenvironmentand/ortemperatures
relative element mobility in different ore types inexcessof250°C.
and altered wall rock. Based on relative mass
changes, volume gain/losses were determined. Allplatinumgroupelements(PGE)withtheex-
Elements commonly incorporated in sulphides ception of Pt were determined at very low con-
(e.g. Co,Ni,Se,Mo,In,Hg,Tl,Bi)arestrongly centrations close to average continental crust
enriched as a function of increasing S content. reservoir data. Platinum is enriched by factors
EnrichmentsofMn,MgandCarecausedbythe ranging between 6 and 100 (average of 30) rel-
formation of carbonates. Other elements sug- ative to average continental crust. The strong
gestedtobeintroducedbythemineralisingfluid fractionation of Pt from Rh and Pd due to the
are Cr, Sr, Ba, W, U and Ga. Alteration proxi- selective enrichment of Pt is explained by hy-
mal to the base metal mineralisation is charac- drogenous and diagenetic ferromanganese rich
terisedbyamassgainofS,Fe,Tl,Mg,Rb,Cs, sediment sequences as the PGE source. Fer-
Ba and W and a relative loss of Ca, C, Na, K romanganese crusts, concretions or nodules are
andSr. enriched in Pt relative to other PGEs. The hy-
The assessment of mass/volume changes of drothermalmetal-bearingmineralisingfluidmi-
different ore types clearly showed that replace- grated through, and leached Pt from ferroman-
ment of the host lithology is a minor mecha- ganeserichsequencesduringitsascent. Mobil-
nism during ore formation. Some replacement isation of Pt likely occurred as either bisulfide
may have occurred in zones of elevated fluid- orhydroxidecomplexes.
rock ratios, i.e. in breccia-stringer type ore
Radiogenic age determination via the Re-
and, in places, in semi-massive ore zones. Re-
Os isotope system resulted in an isochron
sults for massive ore showed that most of the
age of 378±15Ma and represents the first
sulphide mass exclusively precipitated in sites
age constraint established via directly dating
of fracture-induced dilation, accompanied by
the sulphide mineralisation. The observed
negligible interaction with the wall rock. The
relatively non-radiogenic initial γ value is
Os
only substantial replacement that occurred dur-
either caused by juvenile continental crust
ingbasemetalformationisthatofpyritebybase
of the Lachlan Fold Belt as metal source or
metal sulphides in pyritic dominated ore vari-
a contribution of primitive Os from mantle
eties.
sources.
192 |
ADE | Chapter 6
Geo-metallurgical ore
characterisation study via
®
QEMSCAN
6.1 Introduction
(b) small grain sizes, in particular of galena (c)
overallpoorPbandZnrecoveryperformancein
AT THE ENDEAVOR MINE, three different addition to (d) potentially significant arsenopy-
metal concentrates are produced in its rite quantities. Leevers (2001) undertook ex-
mineral processing facility: zinc concentrate as tensive metallurgy test work on several zones
themainproduct,leadconcentrateassecondary of the Elura Orebody and tried to correlate ob-
product and copper concentrate in minor quan- servedplantperformancestodifferentoretypes.
tities as by-product. In the past, the metal re- He concluded that a prediction of plant perfor-
coveries during the mineral processing stage at mance based on geological data is not reliable
the Endeavor mine were and still are fluctuat- due to strong small scale variation of ore char-
ingasafunctionoforefromdifferentlocations acteristics including textures and grain sizes of
throughouttheorebodybeingminedandmixed differentoretypes.
beforedifferentialfrothflotation. The present study was undertaken in order
Theflotationrecoveryrateandthequalityof to better understand the ore characteristics of
the mineral concentrates produced are strongly the 5/5 stope and, furthermore, to test the pre-
affected by changes in whole rock geochem- dictability of how the ore of different 5/5 stope
istry, ore mineralogy and mineral chemistry, in areas will perform during mineral processing.
addition to textural characteristics of the mas- The investigation is based on a bigger sample
sivesulphideore. Apartfrommetallurgicaltest setthanpreviouslyused. Bulkaswellasinsitu
work on material from the flotation feed, pro- oresampleswerepreparedandpresentedaspol-
cess stream and tailings, only a few studies on ished block particle separates or as uncovered
ore characteristics have been undertaken since polishedthin-section. Initially,wholerocktrace
the commencement of mining activities at the elementgeochemistry,reflectedandtransmitted
Elura deposit. Butcher et al. (1998) and Barn- light ore microscopy are used to investigate the
field (1999) investigated ore characteristics of ore characteristics, followed by the application
proposedstopingzones,includingthe5/5stope ofQEMSCAN® technique.
area, however, the study is based on only one In order to investigate the nature of Ag and
sample from the 5/5 stoping zone. Their con- Bi occurrence, and furthermore, to attempt the
clusion was that ore from this zone will be dif- characterisation of different ore types through-
ficult to treat due to: (a) significant quartz con- out the orebody, additional samples were taken
tentandsubsequentlylowmillthroughputrates andpreparedaspolishedthin-sections.
195 |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
The following petrographic and textural ore A better understanding of the occurrence of
characteristicswereproposedofbeinganalysed pyrrhotite in different stope parts can be used
viaQEMSCAN®: as a tool to improve the mining schedule (e.g.
smaller portions of ore being fired and bogged
• Grain size distribution of ore and gangue faster in order to avoid excessive oxidation of
phases(e.g. haveasignificantinfluenceon pyrrhotite enriched ore, in addition to blend-
the milling time) and their modal percent- ing of high and low pyrrhotite ore). By that,
ages, the pH decrease caused by pyrrhotite oxidation
could be constrained, certainly of great benefit
• Theoretical grade-recovery graphs and
for the mineral flotation process. Furthermore,
concentratecomposition,
the Ag recovery, which is floated as a bonus
metal within the Pb flotation circuit, never ex-
• Particleliberationandfreeparticlesurface
ceeded recoveries above approximately 50%.
areaofimportantsulphidemineralphases,
The 3-dimensional silver grade distribution is
• Mineral associations of ore phases (none, very well known and is modelled within the
binary, ternary, etc.; which phases are in- blockmodel. However,theAgoccurrence(e.g.
tergrownwitheachother), asdiscreteAgmineralphases,assolid-solution
or as impurities in other sulphide phases), is
• Pyrrhotite content (significantly influenc-
onlypoorlyunderstood. Consequently,thelack
ingthemineralrecoveryrateduetooxida-
of this knowledge seems likely to represent the
tion and production of sulphuric acid sub-
explanation for the observed low Ag recovery
sequently changing the pH in the flotation
rates.
process),
Ore mineralogy directly influences ore to
concentrate metal recovery rates and thus the
• Identification of Ag-bearing mineral
mine economics. Furthermore, mineralogy,
phasesandpossiblyBiminerals,
mineralchemistryandwholerockgeochemistry
• Classification and distribution of low and are responsible for the quality of the mineral
highironsphalerite concentrate. Thisisbecause(a)worthlessmin-
eralspeciesareseparatedtogetherwithoremin-
• Ore textures (e.g. elongated grains po-
erals,loweringthemetalcontentintheconcen-
tentially intergrown with different miner-
trate and (b) penalty elements (e.g. Bi in the
als negatively affecting the flotation prop-
lead concentrate) reduce the quality and hence
erties),and
thevalueoftheproduct.
Despite the obvious importance of all those
• Exsolution and inclusion (identification of
ore characteristics mentioned above, very little
typeandquantity)
is known about them in general and basically
The improved understanding of ore charac- nothingisknownabouttheir3-dimensionalspa-
teristics and their spatial variability at stope tial variability at stope scale. By filling this
scale will improve mine planning and the en- knowledge gap, a significant contribution to a
tire mineral processing stage, from milling via better understanding of the ore forming pro-
screening to differential froth flotation. A de- cesses will be made, and will in particular be
tailed knowledge of grain size, shape, inter- of great benefit to the ore processing at the En-
growth as well as inclusion and exsolution are deavorMine.
very important for the correct determination of
millingtimesinordertogenerateliberatedmin-
eral grains for the subsequent flotation process.
196 |
ADE | ®
6.2 Fundamentals of the QEMSCAN technique
6.2 Fundamentals of the
overlappingBSEintensitiesforsomematerials.
QEMSCAN® technique Thus,theEDXspectrum,whichisadirectqual-
itativemeasureforthechemicalcompositionof
anunknownmaterial,isused.
The QEMSCAN® analysis system is a fully
Four different modes of measurement can
automated, non-destructive quantitative evalua-
be performed via QEMSCAN® (Intellec-
tionofmaterialstechniquethatusesascanning
tionPtyLtd,2008):
electron microscope (QEM*SEM), developed
and initially distributed by Intellection Pty Ltd.
1. Theparticlemineralanalysis(PMA)isa2-
The QEM*SEM technique has been developed
dimensional, pixel-based data acquisition
in the late 1970s by Dr Alan F. Reid, CSIRO,
technique. Each measured particle is rep-
Australia (Creelman et al., 1989). Other at-
resented as a compound of pixels and is
tempts in the development of image analysis
visualised as a particle image. The pixel
systemssimilartoQEM*SEMwereundertaken
size depends on the detail and resolution
betweenthelate1970sandearly1980sinEng-
which should be achieved during analysis.
land and Canada (Creelman et al., 1989). In
After an area of the sample has been de-
2003, Intellection Pty Ltd. was founded by
fined for measurement, the area is subdi-
CSIRO in order to market their product as
vided into frames. A BSE image scan is
QEMSCAN®. All QEMSCAN® assets, in-
conductedonaframebyframebasisprior
cludingsoftwareandintellectualproperty,were
to the acquisition of the chemical compo-
purchased by FEI Company in early 2009.
sition via EDX spectrometry. Epoxy resin
QEMSCAN® iswidelyusedintheminingsec-
is characterised by a rather low BSE in-
tor in order to investigate various mineralogi-
tensity,resultinginasignificantcontrastat
calandmetallurgicalorecharacteristicsapplied
the margin of particles embedded in resin.
toexploration,productionandmineralprocess-
Thisenablesafastidentificationofthepar-
ing. Gottliebetal.(2000)givesabriefoverview
ticlesandtheirexactcoordinativeposition
about fundamental of the QEMSCAN® tech-
withinthesample. Subsequently,everyin-
nique.
dividual particle is subdivided on a pre-
The QEMSCAN® system of AMDEL Lab- definedgridandeachpixelisanalysedfor
oratories in Adelaide has been utilised in the its individual EDX spectra. This informa-
course of this study. The system is based on tion is assigned to the particular pixel in
a Carl Zeiss EVO 50 scanning electron micro- additiontoitsspatialcoordinativeposition
scopeequippedwithfourBrukerenergydisper- andBSEintensityvalue. Thegridspacing
siveX-raydetectors. is chosen according to the desired analyt-
Thesamplematerialcaneitherbepreparedas ical resolution. Based on the acquired in-
polishedblock,polisheduncoveredthin-section formation, each pixel can be identified as
oraspolishedparticleseparatesmountedorem- a mineral phase and furthermore, the spa-
beddedinepoxyresin. Thelattertwowereused tial information enables a realignment of
in this study. The detailed sample preparation the analysed pixels in order to produce an
procedures are described in the following sec- actual image of the analysed particle sam-
tion6.3onpage199. ple area. This image information can be
usedinordertoinvestigatetheactualmin-
QEMSCAN® uses back scatter intensity
eral intergrowths and textures and can be
(BSE) and energy dispersive X-ray spectra
utilised for the calculation of parameters
(EDX) for mineral identification. The BSE in-
based on mineral areas such as liberation
tensityincreaseswiththeatomicnumberofele-
andlockinginformation.
ments and can therefore be used for fast identi-
ficationofchemicaldifferences. However,they
are not uniquely material specific resulting in
197 |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
2. The bulk mineral analysis (BMA) is a one In order to assign a mineral species to each
dimensional analysis mode. The particles pixel analysed, a database is required, which
inapolishedsampleblockarescannedon contains the entity of all mineral phases poten-
a line in X direction. The Y direction de- tially occurring in the studied samples. These
fines the line spacing and is set to a value mineral phases are defined by their BSE inten-
in order to each particle being intersected sityandEDXelementalspectralcharacteristics.
approximately once, hence representing a This database or list is the so called species
function of the analysed particle size frac- identification protocol (SIP). Several SIPs have
tion. This method produces a very large been developed in the past for various differ-
datasetfromrandomlyorientatedparticles ent types of ore deposits or materials investi-
and therefore is statistically highly robust. gated. A standard SIP for base metal deposits
The lack of the actual particle image was utilised and modified by the author in or-
is a disadvantage, compared to PMA. der to meet the observed mineralogy and min-
However, limited area based information eral chemistry of the deposit. An additional
can be arithmetically extrapolated from databaseorlistisrequiredinordertoassignthe
the data obtained via BMA mode (e.g. chemicalcompositionstotheidentifiedmineral
phasespecificsurfacearea). species. Thislist,thesocalledprimarylist,has
beenmodifiedaccordingtotheobservedchem-
ical compositions of the main sulphide min-
3. If certain mineral phases or those occur-
eral phases during electron microprobe investi-
ring in very low quantities are targeted
gations. Forviewing,investigationandcalcula-
for investigation, the specific minerals or
tion purposes, the primary list is still too com-
tracemineralssearchmode(SMSorTMS
plex and detailed. In order to simplify the list,
respectively) are used. Both modes are
mineralphasesaregroupedtogetherandstored
based on PMA. However, only particles
asso-calledsecondarylists.
containing the target mineral phase, as
Depending on the measurement mode used,
defined via a BSE intensity threshold, are
thefollowingoreormaterialcharacteristicscan
analysed.
be calculated: modal mineralogy, particle size
distribution, grain size distribution, elemen-
tal sample composition, phase specific surface
4. The field scan mode measures the entire
area, theoretical grade-recovery graphs, area
area of a sample and is used for thin-
based particle liberation and free particle sur-
sectionorpolishedblocksamples. There-
face,mineralassociations,irondeportment(i.e.
sultingimageisutilisedtostudytheinsitu
Fe content within concentrate resolved for dif-
texturalcharacteristicsofanysolidsample
ferentmineralphases),orclassificationofmin-
material.
eral species characterised by changes in their
chemicalcompositionsuchassphalerite.
198 |
ADE | 6.3 Sampling
6.3 Sampling
6.3.1 The 5/5 Stope
The stope is located in the northern part of
the southern upper main lode zone, at approx-
imately 6785 to 6820 Mine North and 9690
to 9785 Mine RL (see figure 6.1; RL is re-
duced level depth; coordinates are local mine
grid). Based on the stope reserve shape and
theresourceblockmodelfromDecember2009,
the overall tonnage and grade is estimated at
about 220,000t at 8.8wt%Zn, 5.7wt%Pb and
101g/tAg (data supplied by CBH Resources
Ltd). The mining design sub-divides the stope
into a total of eleven portions, which can be
combinedintothreemajorstopeparts,i.e. East-
ern, Central and Western part (figure 6.2). The
resource model is well constrained in this par-
ticular part of the ore body with a total of 35
DDHsinthevicinityorintersectingthereserve
stope shape. Table 6.3 summarises the tonnage
and grade as well as the percentage split of the
differentoretypesforthethreestopeparts. The
Figure 6.1: Thin-sectionsamplelocationsforore
eastern part contributes 66% of the total stope characterisation study; number of samples is 15;
sevensamplesarefromwithinthe5/5stopingarea
resource and is characterised by a significant
(grey stope shape), remaining samples were taken
quantity of massive pyrrhotitic ore as well as from upper and lower areas of the Elura orebody,
selected primarily focussing on elevated silver or
siliceous pyritic and pyrrhotitic ore varieties.
bismuth concentrations observed via EMPA/LA-
Thesmallesttonnageiscontributedbythecen- ICP-MSorwholerockgeochemistry. Longitudinal
tralpartwith10%ofthetotalresource. Massive view towards WSW (˜245°). Yellow shape is the
stringertypeore(VEIN)resourcedomain.
pyritic ore dominates, in addition to rather low
quantities of siliceous ore. The remaining 24%
arecontainedinthewesternstopeportion,com-
shows the sampled diamond drill holes. Sam-
prising 69% massive pyritic and 31% siliceous
pledataissummarisedintable6.1includingthe
ore.
sampleddiamondrillholedepthintervalandthe
relativequantitiesoforetypeswithineachsam-
6.3.2 Sample preparation
ple.
Ten bulk samples were taken with the aim to ThesamplesfororecharacterisationbyPMA
achieve good sample coverage for each of the via QEMSCAN® were taken as quarter core,
three stope parts. Thirty-five diamond drill groundinaLabtechEssa450mmhammermill
holes have been identified as being viable for to 100% passing 2.36mm, split by Essa ro-
sampling, however, only 19 were located at the tary sample divider followed by manual riffle
mine site’s core storage facility. Furthermore, split down to the required sample size of ap-
severalofthosediamonddrillholeswereeither proximately 1.5 to 3kg. Upon sample recep-
stronglyweatheredorpartiallymissing. Despite tion by AMDEL Laboratories in Adelaide (ab-
these difficulties, a rather good coverage has brev. AMDEL), the samples were weighted,
been achieved with the exception of the lower stagecrushedto100%passing600µmandsplit
portion of the eastern stope part. Figure 6.3 viarotarymicroriffler. Theappliedstagecrush-
199 |
ADE | 6.3 Sampling
Sampled
Diamond Sampled depth interval [m] Lithology split [%]
drill hole ID stope part Start End Po Py SiPo SiPy
DE020-1 Central 456.0 494.0 100 0 0 0
DE020-2 Lower West 512.5 550.0 0 65 0 35
DE307 Lower West 0.0 39.9 0 80 0 20
DE363 Upper West 0.0 42.3 0 100 0 0
DE367 Central 8.0 30.0 100 0 0 0
DE369 Central 0.0 33.0 21 62 0 17
DE377 Upper East 0.0 50.0 100 0 0 0
DE378 Upper East 0.0 25.8 66 0 11 23
DE381# Upper East 0.0 23.0 95 0 5 0
DE383# Upper East 16.5 31.0 0 0 0 100
DE381/3 Upper East na. na. 58 0 3 39
DE395 Upper West 0.0 50.0 0 61 0 39
Overall eastern stope part 64 0 4 32
Overall central stope part 74 21 0 6
Overall western stope part 0 77 0 24
Table6.1: Bulksamplesfor5/5stopeorechar-
acterisationstudy. # Sampleswerecombinedto
DE381/3.
Orebody Local mine grid [m] DDH Ore
Sample ID Zone East North RL sample depth type
475-z12-2-A z1 4477 6915 9482 grab sample Po
560-MLR-S ML 4464 6794 9567 grab sample Po
CAF-1LS-1-1 ML 4450 6841 10135 82 Py
CAF-6z3-1-A z3 4358 7109 9709 517 SiPy
DE008-2 ML 4510 6718 9993 264 Po
DE306 ML 4440 6792 9713 8 Py
DE367-1-A ML 4460 6787 9783 10 SiPy
DE367-3 ML 4458 6801 9725 69 Py
DE377 ML 4486 6801 9756 40 Po
DE381-1 ML 4508 6790 9784 11 Po
DE381-2 ML 4508 6799 9772 21 SiPo
DE398 ML 4439 6796 9775 23 Py
NP245-1 z4 4298 7250 9613 82 Po
NP549-1-B ML 4511 6850 9323 102 SiPo
NP680-1 z4 4331 7175 9458 47 Py
Table 6.2: Thin-section samples for ore
characterisationstudy.
5/5 Stope Tonnage Zn Pb Ag Cu Fe Po Py SiPy/SiPo VEIN MinA TOTAL
part [mt] [wt%] [wt%] [g/t] [wt%] [wt%] [mt] [%] [mt] [%] [mt] [%] [mt] [%] [mt] [%] [mt] [%]
East 145381 8.8 5.8 53 0.3 30.7 71509 49 14212 10 59274 41 354 0 32 0 145381 66
Centre 22188 9.4 5.4 166 0.2 29.6 6210 28 14192 64 1785 8 0 0 0 0 22188 10
West 52569 8.6 5.5 207 0.2 22.6 7 0 36235 69 15949 30 376 1 1 0 52569 24
TOTAL 220137 8.8 5.7 101 0.2 28.6 77726 35 64639 29 77009 35 730 0 33 0 220137 100
Table 6.3: Summary of tonnage and grade as well as the percentage split of the different ore types
forthethree5/5stopeparts. DatasuppliedbyCBHResourcesLtd.
201 |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
quentpreparationstepsandQEMSCAN® anal- are observed. Chalcopyrite is quite abundant,
yseswereundertakenbyAMDEL.Therawdata mostly associated with pyrite as interstitial
was subsequently processed to reportable data phase or as fracture infill at rather small
by the author without significant assistance by grain sizes below ~20µm. Sub- to anhedral
AMDELoranyotherinstitutionorperson. pyrrhotite, characterised by a granular texture,
represents the major iron sulphide species,
6.4 Investigations prior to forming the sulphide groundmass. Two pyrite
generations are present in several samples.
®
QEMSCAN
Pyrite type-B is mostly euhedral to subhedral
in shape, partially replaced by other sulphides
For quality assurance-quality control (QAQC) and frequently hosting inclusions of all other
purpose and in order to maximise the accuracy majorsulphidephases. Inthissample,however,
of the mineral identification process, it is im- pyrite-B grains contain only minor inclusions.
portanttogainaprofoundunderstandingofore Pyrrhotite and chalcopyrite frequently intrude
mineralogy and mineral chemistry prior to the this pyrite variety along fractures. The an- to
application of an automated analysis technique subhedral pyrite type-A may occur as small
suchasQEMSCAN®.
grains (10-30µm) but commonly forms large
A detailed study of ore mineralogy and min- dendritic compounds (>~100µm). Abundant
eralchemistryofthemajorsulphidephaseshas inclusions of predominantly gangue phases
beenundertakeninthecourseofthisthesis(see are observed. Arsenopyrite is found in minor
chapter4onpage63andchapter4onpage79, quantities. Only siderite is present as gangue
respectively). Transmitted and reflected light mineralphase.
microscopy of insitu ore samples indicate sub-
stantialvariationsofmineralassociations,grain
sizes, textures, and the mineral chemical com-
position of sphalerite. The microscopic obser-
vations made on a representative selection of
fivethin-sectionsfromsamplestakenwithinthe
5/5 stoping are qualitatively summarised in the
following. Quantitative mineralogical charac-
teristics(e.g. grainsizedistribution,modalmin-
eralogy, etc.) arediscussed inthe analytical re-
sults(section6.6onpage213).
6.4.1 Ore petrology
SampleDE377(figure6.4)
Sphalerite may occur as coarse patches or
as small grains in the order of 5µm intersti-
tial to all other major sulphide phases. It is
mostly anhedral in shape, fairly dark brown
in colour and contains abundant exsolutions
and inclusions of chalcopyrite and pyrrhotite.
Galena is commonly present as patches larger
than ~100-200µm and in minor quantities Figure6.4: Fieldscanimage(BSEandsulphides)
forsampleDE377;blueissphalerite,redisgalena,
as tiny interstitial grains in a similar way to
orangeischalcopyrite,yellowispyrite,mediumyel-
sphalerite. In places, intimate intergrowths low is pyrrhotite, olive green is arsenopyrite and
of chalcopyrite and inclusions of pyrrhotite non-sulphidegangueisshownindarkgreen.
202 |
ADE | ®
6.4 Investigations prior to QEMSCAN
SampleDE306(figure6.5) SampleDE381-1(figure6.6)
In this sample, sphalerite either occurs small This sample is characterised by a strong sul-
grained (below ~20-30µm) intimately inter- phide banding or layering. Sphalerite occurs in
grown with pyrite, or as coarse patches an irregular layered texture, characterised by a
(>~100µm) that contain significant quantities pronounced association with pyrrhotite as inti-
of siderite, quartz, sericite and chlorite. In mate intergrowths or as interstitial phase. Tiny
contrast to the previous sample, sphalerite is chalcopyrite exsolutions and slightly coarser
light brown coloured, indicating a decrease in exsolutions and/or inclusions of pyrrhotite are
Fe content. Only minor inclusions and exsolu- observed. Sphaleriteisrelativelydarkincolour
tionsofchalcopyritearepresent. Grainsizesof suggestingelevatedFecontent. Galenaappear-
galena strongly vary, at times occurring tightly ance and grain size strongly varies, occurring
intergrownwithchalcopyriteandsphaleriteand as coarse homogeneous patches (>~100µm),
frequently interstitial to pyrite. Chalcopyrite is or as rather small grains below 10 to 20µm,
quite abundant and mostly associated and in- intergrown with sphalerite and/or chalcopyrite,
tergrown with galena. Both are anhedral in or as interstitial phase to pyrrhotite. Fairly
shape. Pyrite is the dominant sulphide phase, abundant chalcopyrite is characterised by
frequently fractured forming a cataclastic tex- a noticeable affinity to pyrite, occurring in
ture, and is at times replaced by base metal the vicinity or interstitial to this particular
sulphides. Differentiation between both pyrite mineral phase. Only pyrite of type-B has been
types is difficult. However, it appears pyrite observed. It hosts inclusions of gangue and all
type-Bovergrowstype-Asubsequentlyforming other occurring sulphide phases, and appears
a massive, almost homogeneous groundmass. partially replaced at times. In some places,
Pyrrhotite is absent in this sample. The gangue pyrite shows cataclastic textures. Only siderite
mineralogy consists of quartz and carbonate,
chlorite,sericiteandbaritearepresentasminor
constituents.
Figure 6.6: Field scan image (BSE and sul-
Figure6.5: Fieldscanimage(BSEandsulphides) phides)forsampleDE381-1;blueissphalerite,red
forsampleDE306;blueissphalerite,redisgalena, is galena, orange is chalcopyrite, yellow is pyrite,
orangeischalcopyrite,yellowispyrite,mediumyel- medium yellow is pyrrhotite, olive green is ar-
low is pyrrhotite, olive green is arsenopyrite and senopyrite and non-sulphide gangue is shown in
non-sulphidegangueisshownindarkgreen. darkgreen.
203 |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
ispresentasgangue. arerathersmall(below~20µm),butcommonly
forms larger aggregates. Minor arsenopyrite is
SampleDE381-2(figure6.7) present. Theganguemineralogypredominantly
consists of quartz with subordinate siderite,
Sphalerite occurs as medium to coarse grains
accessory sericite and chlorite. A very weak
between ~50 and >100µm, tightly intergrown
sulphidelayeringorbandingisobserved.
with pyrrhotite in myrmekitic-like textures.
Minor exsolutions of chalcopyrite and in-
SampleDE398(figure6.8)
tergrows of pyrrhotite are present. Overall,
A moderate sulphide layering/banding, man-
relativesmallquantitiesofgalenaareobserved.
ifested mainly due to irregular zones signif-
When found, it occurs as coarse patches or
icantly enriched in sphalerite and pyrite, is
interstitial to sphalerite and pyrrhotite and
present. Sphalerite is coarser grained in these
at times associated as well as intergrown
irregularbandscomparedtograinsfoundinter-
with chalcopyrite. Chalcopyrite is present as
stitially to pyrite. Overall grain sizes range be-
accessory mineral phase together with pyrite
tween ~20 and >100µm. Relative low quan-
and/or gangue or rarely as tiny inclusions in
tities of galena are present, featuring similar
pyrite type-B. Subhedral pyrrhotite represents
grain sizes as sphalerite. Accessory chalcopy-
themajorsulphideconstituent,characterisedby
rite occurs mostly as fracture filling within or
agranulartexture. Bothpyritetypesarepresent
as replacement of pyrite and, furthermore, it
in this sample. The euhedrally shaped pyrite
is commonly intergrown with galena. Most
type-B contains abundant inclusions of galena,
pyriteiscolloformtospherulitictype-A.Minor
pyrrhotite, sphalerite, chalcopyrite and gangue
pyritetype-Bisfracturedtoacataclastictexture
phases. Pyrite type-A occurs rarely in some
and commonly intergrown with pyrite type-A.
areas, entirely absent in others. Single grains
Figure 6.7: Field scan image (BSE and sul-
phides)forsampleDE381-2;blueissphalerite,red Figure6.8: Fieldscanimage(BSEandsulphides)
is galena, orange is chalcopyrite, yellow is pyrite, forsampleDE398;blueissphalerite,redisgalena,
medium yellow is pyrrhotite, olive green is ar- orangeischalcopyrite,yellowispyrite,mediumyel-
senopyrite and non-sulphide gangue is shown in low is pyrrhotite, olive green is arsenopyrite and
darkgreen. non-sulphidegangueisshownindarkgreen.
204 |
ADE | 6.5 QAQC and data processing
6.5 QAQC and data
Sphalerite occurring within the Elura de-
processing positischaracterisedbyverylowMncon-
centrations(belowelectronmicroprobede-
tection limit) and Fe content ranging be-
Data validation is essential in any analytical
tween 2.4 and 8.2wt%. Therefore, three
procedure. Assaycorrelationbetweenchemical
sphalerite categories were defined to ac-
compositionscalculatedbasedonQEMSCAN®
count for the variation of Fe content.
resultsiscomparedtoassaydataobtainedfrom
Manganese has been ignored as very low
standard geochemical analysis. As already de-
concentrations were detected during EMP
scribed earlier, to each pixel of a particle, a
analysesofsphalerite. Inthecourseofthis
chemical composition is assigned according to
study it is aimed to identify, high and low
itsallocatedmineralspecies. Ifaveragedacross
Fe sphalerite as well as to quantify con-
the entity of particles, a bulk chemical compo-
tained impurities, such as inclusions and
sition of the sample can be calculated. A de-
alterations.
tailed mineralogical and mineral chemical un-
derstanding is necessary in order to achieve a
2. Two new categories were defined for Ag
goodassaycorrelation,subsequentlyindicating
and Bi. These categories have been de-
a correct mineralogical classification of sam-
signed in order to identify any mineral
ples analysed via QEMSCAN®. To ensure to-
phasescontainingtheseelementsabovethe
tal sample dissolution for major element deter-
EDSdetectionlimit.
mination, a lithium metaborate fusion followed
by analysis via ICP-OS was used. A modified 3. In the initial SIP, siderite as one of
aqua-regia digestion and analysis by ICP-OS the most abundant non-sulphide gangue
and ICP-MS is applied for trace elements. The (NSG)mineralphase,waslackingandhad
concentration of total S was determined by the to be defined. The subordinate occurrence
LECO technique. If sample material was un- of goethite, a mineral phase chemically
known,XRDanalyseswereusedtogiveanap- similar to siderite in respect to its EDS
proximation of mineral abundances, especially spectra, complicated this exercise. How-
clay minerals. The geochemical analyses used ever, by adjusting the BSE intensity range
forQAQCwereundertakenbyAMDEL. forbothmineralphases,acceptableresults
wereachieved.
6.5.1 Adjustment of mineral list
Attheinterphaseofdifferentmineralphases,
and pre-processing of raw
a composite EDS spectrum and BSE intensity
data
isgenerated. Similarly,BSEintensitydecreases
The SIP and the primary lists contain 336 at the margin of particles embedded in epoxy
and79mineralspeciesdefinitions,respectively. resin. The analysed particles, presented as par-
Both lists were slightly modified by the author ticleimages, needtobevisuallyinvestigatedin
according to the observed mineral occurrences order to detect these excessive boundary-phase
and the determined mineral chemistry of sul- artefacts or ambiguous identifications during
phidephasesanalysedviaEMP.Thelistsaretoo the measurement. If detected, the removal is
extensive to be incorporated within this thesis. performed via the boundary phase processor
The most important modifications performed whichenablesthemodificationofthemeasured
are: dataaccordingtothecontextofthepixels.
Althoughitistriedtoavoidparticleagglom-
1. Sphaleritedefinitionsweremodifiedbased erationduringsamplepreparation,occasionally
on the observed mineral chemistry. In they occur. These particles are identified and
the initial SIP, sphalerite species were far separated by a further pre-processor, the so-
too detailed for the purpose of this study. called“touchingparticleprocessor”.
209 |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
6.5.2 Assay correlation based on QEMSCAN®. Those samples were
lacking any common features, they were taken
A good agreement between chemical data ob- from different areas within the 5/5 stope (east-
tainedviaQEMSCAN®andconventionalmeth- ern and western stope part) and are of differ-
ods was met (figure 6.9). Analytical data is ent ore types. Therefore, an analytical error
presented in section C.1.2 on page 504. The seems the probable explanation. Copper con-
qualityofchemicaldataiscommonlydescribed centrations are in good agreement, with the ex-
via the double standard deviation corridor 2σ, ception of two samples with deviations greater
which corresponds to a maximum deviation of than10%. Acorrectestimationofsilicacontent
5.6%. For the QAQC assay correlation of inmassivesulphidesamplesappearstobeprob-
QEMSCAN®data,adeviationoflessthan10% lematic. Assay correlations of only four sam-
is considered as good analytical results by the ples resulted in deviations less than 10%; five
industry (pers. comm. Kellie Jones 2010). For samples deviate by 10 to 20% and one sample
the elements Fe, S, Pb and Zn, a deviation of byover20%. Thedeviationsarescatteredalong
less than or equal to 10% has been achieved the correlation line not indicating any system-
throughouttheentiresampleset. Sulphurcorre- aticanalyticalerror.
lated exceptional well, indicating accurate sul-
phide identification. A minor systematic over-
estimationofFeispresent,whichispotentially It needs to be stressed, that the elements af-
explainedbycompositionalvariationofcarbon- fectedbythedeviationsoccuratratherlowcon-
ates,notsufficientlyrecognisedduringanalysis. centrations, i.e. also low quantities of particles
Good As assay correlation has been achieved containing the corresponding mineral species
for the first set of seven samples. However, a were analysed and subsequently, an increase
sudden drop in analytical quality is observed of statistical uncertainty is inevitable. Overall,
fortheremainingthreesampleswithdeviations mostoftheobservedelementconcentrationsare
greater than 20% and characterised by a sys- well within analytical errors and therefore the
tematic underestimation of the concentrations resultsareconsideredvalid.
50 (a) 12 (b)
45
10
40
35 8
Zn
30
Pb
25 6
Cu
20
As
4
15 Fe
Linearcorrelation
10 2σcorridor 94.4% accuracy S
2
5 Data trend; R²=0.9986 Si
0 0
0 5 10 15 20 25 30 35 40 45 50 0 2 4 6 8 10 12
Chemical Assay Chemical Assay
[wt%] [wt%]
Figure 6.9: QEMSCAN® – chemical assay correlation; all elemental concentrations shown in diagram (a),
diagram(b)presentselementcorrelationsoflowerconcentrationlevels.
210
yassA
(cid:7)ACSMEQ ]%tw[ |
ADE | 6.5 QAQC and data processing
6.5.3 Important QAQC particle 100 Cent
features 90 East
West
80
Initially, a visual inspection of the resulting 45µm
70
particle images is carried out in order to de- 75µm
60
termine a successful and sufficient data pre-
50
processing, and to investigate potential particle
40
duplications during analysis. Based on statis-
tics, about 15,000 to 20,000 analysed parti- 30
cles are considered to be a representative sam- 20
ple. The particle count for the samples of 10
this study showed, that approximately 70,000- 0
260,000particleswereanalysedpersample(ap- 1 10 100
pendix section C.1.2, page 504). Only 0.1% of Particle size [µm]
allparticlesanalysedfallwithinthesizefraction
Figure 6.10: Particle size distribution for PS+
+80/-600µm. The low quantity is considered 3/−80µm. Black dashed lines show actual long-
termparticlesizedatafromtheEndeavorprocess-
notrepresentativeandtherefore,particleslarger
ing plant with a P78 of 45µm and ˜4wt% parti-
than 80µm are only used for textural investi- cles above 75µm(pers. comm. Andrew McCallum,
gations and filtered out in other calculations. 2010).
Overall, less than 10wt% of particles is con-
tained in the particle size fraction of +0/-3µm.
in appendix section C.1.2, page 504. Silver
A pixel size of 2x2µm has been used during
bearing mineral phases were found in a very
analysis. Smallparticlesintheorderofthepixel
low particle quantity as low as 51 and up to
size are likely to be affected by misidentifica-
346particles. Compared to the total number
tion due to boundary effects, and furthermore
of analysed particles this count is negligible,
would not lead to the gain of any mineral as-
subsequently, any calculations based on this
sociation data. In order to improve data man-
particle group need to be scrutinised and used
agementandsoftwareperformance,theseparti-
or interpreted in a cautious manner. No Bi-
cleswerefilteredoutwithoutsignificantlymod-
bearing mineral phases were observed during
ifyingthemodalmineralogyorothercalculated
PMA analysis of samples from the 5/5 stope.
ore characterisation data. However, a basic ore
For detailed and statistically more robust in-
characterisation and classification is performed
vestigations of Ag mineral phases, the earlier
onthefineparticlefraction(seesection6.6.1on
described specific minerals or trace minerals
page 213). The remaining particle size fraction
searchmodewouldneedtobecarriedout.
+3/-80µm was characterised by particle counts
inexcessof34,000particleswhichismorethan
6.5.4 Limitations and analytical
sufficient in statistical terms. At the mineral
problems
processing facility, the actual particle size dis-
tribution of the floatation feed is characterised QEMSCAN® delivers either one (BMA: line
byaP80(actuallyP78)ofapproximately45µm
scan) or two dimensional (PMA: particle scan,
with4wt%particlesabove75µm(pers. comm.
field scan) results. In most cases the particles,
Andrew McCallum, 2010). When comparing
which are randomly orientated, embedded in
the calculated particle size distributions of the
epoxy resin and polished on a planar surface,
selected size fraction +3/-80µm with the ac-
are characterised by a section exposure smaller
tualflotationfeed,agoodagreementisobserved
than its actual particle cross section. Subse-
(figure6.10).
quently, area and size calculations are always
Particle count as a function of different min- characterisedbyanegativebiasasareotherpar-
eral phases being contained in them is showen ticlefeaturessuchasvolume,surfaceexposure,
211
]%tw[
elcitraP |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
etc. However, an automatic stereological cor- during the raw data being transferred from the
rection is applied during these calculations in QEMSCAN® instrument to the storage media.
order to account for this deviation. Rim and It was recognised by the author, that the dupli-
boundary effects are identified and aimed to be cated particles are not identical, showing mi-
reducedviapre-processingoftherawdata. The nor differences in the composition and shape.
magnitudeofhowtheseeffectsareprocessedis Particle duplication due to software problems
primarilyafunctionoftheachievableassaycor- was therefore an illogical explanation. Analyt-
relationqualityandtosomedegreevariableand ical errors seem highly probable and initially,
operatororusercontrolled. the defect sample stage was thought to have
The initial field scans of thin-sections gave caused the duplication. However, the reanaly-
corrupt results due to a defective sample stage. sisof4samplesafterthesamplestagehadbeen
During the field scan measurement, the entire exchangedshowedreoccurringparticleduplica-
area selected for analysis, was subdivided into tions in 2 samples. Finally, AMDEL Labora-
frames and scanned individually. Upon com- tories identified the real cause of the problem
pletion of the measurement, the frames needed as being how the measurement setup is cho-
tobestitchedbacktogetherinordertoproduce sen. Ifthemeasurementissetto“fromthecen-
an actual image of the entire sample area anal- tre of the sample block” compared to “starting
ysed. The sample stage is a highly accurate from the edge”, particle duplication occurred.
X-Y positioning device, which allows the sam- Despite finding the root cause of this analyti-
ple to be moved and positioned under the elec- cal problem, no explanation for the actual rea-
tron beam during the scanning process. Due to son has been identified or communicated by
the defective sample stage, some areas of the AMDEL Laboratories. The samples were re-
frames were not analysed, and some analysed analysedwithmeasurementsetup,now“believ-
more than once. The subsequent stitching pro- ing”nottocauseparticleduplication.
cess gave unsatisfactory results. Therefore, all It needs to be stressed that visual inspection
samples were reanalysed upon exchange of the of particle images in addition to assay correla-
defective sample stage in a second field scan tion are the only QAQC instruments applica-
procedureinordertoobtainreliableresults. ble to the QEMSCAN® technique. Therefore,
a profound understanding of ore petrology and
Particle duplication has been observed dur-
mineralchemistrypriortotheutilisationofthis
ingvisualQAQCinspectionofthePMAresults
techniqueisabsolutelyessentialforhighquality
in six out of a total of ten samples. AMDEL
orecharacterisationresults.
Laboratories was convinced that the duplica-
tion is due to corrupt software and occurred
212 |
ADE | ®
6.6 QEMSCAN results from PMA and field scans
®
6.6 QEMSCAN results PMA – Particle size fraction +0/-3µm
from PMA and field
Analytical results are shown in appendix sec-
scans tion C.1.4, page 510. An east-west trend is ob-
served for modal mineral composition of ore
from the 5/5 stope (abbrev. 5/5-ST). Quanti-
All results were calculated utilising the iDis-
ties of sphalerite and galena are highest in the
cover software package version 4.3 supplied
eastern (abbrev. 5/5-E), pyrrhotite dominated
by FEI to the author. A detailed summary
stope part with approx. 22 and 18wt%, re-
of measurement setup and raw data of the
spectively. The lowest concentrations are con-
entire data set is given in appendix C, page
tained in ore from the western stope part (ab-
494 for particle mineral analysis (PMA) and in
brev. 5/5-W), which almost exclusively con-
appendix C, page 651 for field scans. Results
tainspyriticore. Thecentralstopesegment(ab-
from PMA and field scans are collectively
brev. 5/5-C) consists of pyritic and pyrrhotitic
described with reference to individual 5/5
ore and features intermediate Pb and Zn con-
stope parts and/or ore types. Massive sulphide
centrations. Chalcopyrite is recognised in all
samples (abbrev. Py for pyritic and Po for
samples but significantly elevated in 5/5-C and
pyrrhotiticdominatedsamples)werepreferably
5/5-E.Themostimportantmineralphaseshost-
selectedforfieldscans,thereforethenumberof
ingFeareobviouslypyriteandpyrrhotite. Iron-
samples of semi-massive pyritic (SiPy; n=2)
rich sphalerite and non-sulphide gangue min-
and pyrrhotitic (SiPo; n=2) ore is low. Field
eral phases (NSG) are responsible for an Fe
scan results of SiPy and SiPo are combined to
deportment of approximately 10wt% Fe each
one sample category. Due to the voluminous
character of the reported QEMSCAN® data, (calculated percentage of total Fe). Particles
falling within the PS +0/-3µm are highly lib-
only selected diagrams are included in this
erated with ~90wt% Zn, ~81wt% Pb, ~93wt%
section. All calculated data is comprehensively
Cu and 81wt% Ag contained in particles at a
compiled in appendix C, page 492. Numerous
∗
liberation grade above 90% . Particles featur-
references to the appendix and its sections are
ingliberationgradesinexcessof90%arecom-
notavoidableandarecontainedinthefollowing
monlyconsideredasfullyliberated.
description.
PMA – Particle size fraction +3/-
80µm
6.6.1 General ore characterisation
Results are shown in appendix section C.1.4,
The modal mineralogy, elemental composition, page 510. Variation of modal ore composition
iron deportment, calculated ESD (equivalent throughout the 5/5-ST is similar for this PS as
spherical diameter) sizes and mineral associa- observed for the fine PS +0/-30µm but not as
tions are presented in this section. These fea- pronounced. Highpyrrhotitecontentinthe5/5-
tures will give an overview of compositional Eand5/5-Ccorrespondstoelevatedconcentra-
variability throughout the sample set and the tions of sphalerite (max. Zn 10.7wt%), galena
threedifferentstopeparts. (max. Pb 9.6wt%) and chalcopyrite (max. Cu
In excess of 90wt% Zn, Pb and Cu is con- 1.0wt%). Arsenopyriteoccursinallstopeparts
tained in the particle size fraction (abbrev. PS) at rather constant concentrations ranging be-
+3/-80µm. Silver phases are commonly finer tween1.1and1.3wt%.
grained. Approximately 30wt% Ag is con- Siderite and other carbonate mineral phases
tainedinthefinePSbelow3µm. Silverphases
are not included in the calculation of modal ∗.Particle liberation was calculated on the basis of
areapercent,i.e. aparticlehasaliberationofmorethan
mineralogyandFedeportmentbecauseoftheir
90% in respect to Pb if more than 90% of the particle’s
lowabundance. areaconsistsofgalena.
213 |
ADE | ®
6.6 QEMSCAN results from PMA and field scans
sections contain on average between ~18wt% massiveoreandcontributes~69%oftotalNSG.
(Py) and ~16wt% (Po) sphalerite, as well as Siderite and other carbonate phases were ob-
~6wt% (Py ore type) and ~11wt% (Po ore served at ~23% but their contributions are sig-
type) galena. Chalcopyrite content is highest nificantly higher in massive ore ranging be-
at0.6wt%insemi-massiveore(SiPyandSiPo) tween ~70% in Py and ~97% in Po. Concen-
compared to 0.3wt% in Po and 0.4wt% in Py. trationsofotherNSGphasesarevariable.
The pyrite to pyrrhotite ratio is approximately Average ESD grain sizes of quartz range
1:1 in massive pyrrhotitic ore. Only 2.2wt% between ~60 and 110µm, coarsest in semi-
pyrrhotite was detected in massive pyritic ore. massive ore. Carbonate grain sizes are rather
Detection of Ag-bearing mineral phases was consistent and range between 70 and 80µm.
limited to massive pyritic ore. Arsenopyrite Other NSG phases, i.e. sericite, chlorite, iron-
concentration of 0.7wt% was determined for hydroxides, etc., feature ESDs close to the ear-
both massive ore types, whereas only 0.3wt% lier described limit of accurate determination
was observed in semi-massive ore. Quantities (~15µm).
of NSG phases range between ~7 and 11wt%
in massive ore but are significantly higher in
Theoretical grade/recovery diagrams
semi-massiveoreatanaverageconcentrationof
20wt%.
Theoretical grade/recovery diagrams were cal-
Sphalerite and galena are commonly coars- culatedforZn,PbandCubasedontheassump-
est in Po ore. Average grain sizes (ESD) range tionthatmineralliberation(calculatedbasedon
between 58 (Py) and 68µm (Po) for sphalerite areapercent)isdirectlylinkedtomineralrecov-
and 28 (Py) to 50µm (Po) for galena. Grain ery without consideration of mineral losses. If
sizes of chalcopyrite are consistent and range only particles characterised by high liberation
between 28 and 34µm. Pyrite and pyrrhotite arerecovered,ahighconcentratequalitywould
is significantly coarser at ESDs in excess of be achieved but at a low metal recovery rate.
~90µm. Pyrrhotite represents a minor con- The liberation value necessary for mineral re-
stituentinmassivepyriticoreandfeaturessmall covery is step-wise decreased and the compo-
grainsizeswithanaverageof~25µm. Accord- sition of the recovered particles calculated, re-
ing to field scan data, grain sizes of Ag-phases flecting the composition of the theoretical con-
areestimatedat18(Po)to33µm(Py). ESDsof centrate. Recoveriesandcorrespondingelement
small-grained mineral phases are significantly concentrationsdefinethegrade-recoverygraphs
overestimated based on field scan results be- (abbrev. GR). The GR diagrams are calculated
cause of the lower analytical resolution, which for different particle size fractions (PS) in or-
was set to 10x10µm. If a small grain of for der to investigate at which particle sizes libera-
instance 5µm happens to be identified during tion decreases, subsequently causing a drop in
field scan, the entire 10x10µm pixel would be concentratequality. Allresultsareshowninap-
assigned to this particular mineral phase. Such pendixsectionC.1,page520.
apixelwouldyieldacalculatedESDofapprox- Figure 6.13 shows the GR diagrams for
imately15µm,significantlylargerthanthetrue zinc, suggesting a significant decrease in
grain size. Only relative grain size changes be- grade/recovery performance for ore sourced
tweendifferentoretypesshouldbeusedforore from the 5/5-W. The long-term average plant
characterisation, in particular for small mineral performance for Zn showed recoveries ranging
phases. Theresultssuggestthatpyriticorecon- between ~77 and ~82% at concentrate quali-
tains larger Ag-bearing mineral phases when ties between ~48 and ~52wt% Zn. Recover-
comparedtopyrrhotiticore. Thisobservationis ies and grades of the 5/5-W, although perform-
inagreementwithmicroscopicinvestigations. ing poorer than the other stope parts, are close
Quartz represents the most important non- to the achieved plant performance. GR graphs
sulphide gangue (NSG) mineral phase in semi- for different PS populations indicate that par-
215 |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
70 100
90
60
80
50 70
Achieved Grade/Recovery
60 Achieved Grade/Recovery
40
50
30
40
20 Central Stope part 30 Central Stope part
Eastern Stope part 20 Eastern Stope part
10 Western Stope part Western Stope part
10
Combined Combined
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Zn Recovery [wt%] Pb Recovery [wt%]
Figure 6.13: Theoretical grade-recovery diagram Figure 6.14: Theoretical grade-recovery diagram
for Zn compared to achieved long-term plant per- for Pb compared to achieved long-term plant per-
formance data (range shown in orange; data from formance data (range shown in orange; data from
pers. comm. AndrewMcCallum,2010). pers. comm. AndrewMcCallum,2010).
ticles above 38µm are progressively less lib- aresignificant.
erated. For 5/5-W, a drop of liberation is al- At the Elura’s mineral processing facility,
ready indicated for PS +20/-38µm. The theo- chalcopyrite is initially recovered along with
reticalGRsuggeststhatorefrom5/5-Cand5/5- galena in a composite mineral concentrate.
Eshouldresultineithersignificantlybettercon- Only in a later processing stage both products,
centrate quality or higher metal recovery than i.e. lead and copper concentrate, are separated
thelong-termaverageplantperformance. How- by campaigned flotation. It needs to be inves-
ever,theseresultsreflectidealplantconditions, tigated whether dilution with chalcopyrite and
not considering unavoidable losses or dilutions its intergrown mineral phases is a possible ex-
causedbye.g. entrainment,overgrinding,slim- planation for the observed deviations between
ing, etc. Nevertheless, the results clearly high- theoreticalandrealplantperformance. Inorder
light that 5/5-W will most probably feature a to calculate a realistic best-case grade recovery
significantly lower recoverability by approxi- scenario, thefollowingamendmentstothedata
mately10%comparedtotheotherstopeparts. setaremade:
An even more significant deviation from the (a) Coarse particles above 53µm may be
long-term average plant performance is ob- highly liberated because of the near perfect
served for lead (figure 6.14). The deviation cleavage of galena. Such large particles may
translatesinadifferenceof~30wt%Pbconcen- only be recovered if the particle residence time
trategradeor~20%Pbrecoverywhencompar- within the flotation cells is long enough what
ingachievedplantperformancetothecombined theyinrealityrarelyare. Inordertoavoidanar-
5/5-ST theoretical GR diagram. The 5/5-W tificial increase in calculated theoretical recov-
will perform poorer than other stope parts. GR eries,particleslargerthan53µmareignoredin
graphs for individual PS suggest that particles thecalculations.
smaller than 38µm feature relatively good lib- (b) All Pb and Cu contained in the +0/-3µm
erationfor5/5-Eand5/5-C,whereasfor5/5-W PS is considered as a 100% loss to account for
liberation is lower and responsible of the over- sliming and entrainment mechanisms observed
all poorer performance for this particular stope inreality.
part. Even when considering that these results (c)Thetargetparticlesizefractionforgalena
are theoretical and do not reflect real plant per- flotationisbetween+3/-53µmandusedforcal-
formance as discussed earlier, the differences culationsastherecoverableparticlepopulation.
216
]%tw[
edarG
nZ
]%tw[
edarG
bP |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
rangebetween55and61%andcorrespondtoa dertakeninordertoassessthisissue.
theoretical concentrate quality between 33 and
34wt% Cu, close to the maximum possible of 6.6.2 Mineral specific characteri-
34.6wt%. sation
Zincrecoveriesandconcentratequalitybased
Sphalerite composition
onorecharacterisationshowedsimilarresultsto
actual plant data when minor metal losses and Mineralchemicalinvestigationsshowedsignifi-
dilution are taken into account. The same ap- cantvariationsintheFecontentincorporatedin
plies to Cu. In respect to the 5/5 stope, calcu- sphalerite (see section 4.2.1). These variations
latedtheoreticaldataindicateasignificantdrop may have an effect on sphalerite recoverability
in recoverability of ore sourced from the west- andthequalityofthezincconcentrate. Compo-
ernstopepart. sitionalcharacterisationofsphaleritewasthere-
Observed ore characteristics suggest that a fore undertaken as part of the geometallurgical
significant better grade/recovery performance study. The results of PMA and field scans are
for galena/lead should be achieved when com- presented in appendix section C.2.2, page 531
pared to the long-term mineral processing per- and appendix section C.7.5, page 670, respec-
formance. If Pb losses or significant dilu- tively.
tion by other mineral phase are the cause, they Three Fe categories were defined for spha-
are likely far too extensive as to be accept- lerite in addition to a category accounting for
able by industry standards. Another possi- smallimpuritiesandinclusionsofothermineral
ble explanation for a potential overestimation phases, which cause a mixed energy dispersive
of grade and recovery via QEMSCAN® may spectra. These mineral impurities/inclusions
be the sample preparation technique. It ap- commonlycompriseironcarbonates(calculated
pears feasible that stage crushing may lead to as siderite), chlorite and sericite. Another cat-
a significantly different mineral/particle liber- egory was defined in order to calculate the
ation compared to the grinding methodology quantity of sphalerite affected by boundary ef-
usedintheactualmineralprocessingplant(ball fects (i.e. analytical uncertainty caused by in-
mill). Galenaisasoftmineralwiththreeperfect tergrowth of sphalerite with different mineral
cleavages. Stagecrushingmayachievehighlib- phases or caused by the interphase between
erationwithouttheexcessivegenerationoffine sphaleriteandepoxyresin).
materialsmallerthan5µmwhereasgrindingvia Reconciliation between Fe content of spha-
ballmillmayproducesignificantquantities. In- lerite determined via EMP and QEMSCAN®
creasedamountoffinematerialmayleadtosig- showedaverygoodcorrelation,indicatingcor-
nificantmetallosstotailingsduetoe.g. sliming rectidentificationofdifferentsphaleritecompo-
and/orentrainment. Geometallurgicaltestwork sitions(figure6.17).
ontheactualflotationfeedwouldneedtobeun- Most of the sphalerite contained in 5/5-E is
Fe-rich. A decrease in Fe content is observed
fromtheeasternviacentraltothewesternstope
10
part. Mineral chemistry of sphalerite indi-
8
cated the preferred occurrence of Fe-rich spha-
6
SiPy & SiPo leriteinpyrrhotiticoretypes(seesection4.2.1).
4
Py QEMSCAN® PMA results confirm this obser-
2 Po
vationasthe5/5-Econtainsthehighestpercent-
0
0 2 4 6 8 10 age of pyrrhotitic ore. Less than 1wt% low
XFeSph -EMP [wt%]
Fesphaleriteiscontainedinmassivepyrrhotitic
Figure6.17: Reconciliationofsphaleritecomposi-
oreaccordingtofieldscanresults.
tiondeterminedviaQEMSCAN®-fieldscanvsmin-
eralchemicalcompositionbasedonEMPanalyses; The quantities of impure and altered spha-
solidblacklineisthe1:1ratio. lerite are in the order of 10-14wt% based on
218
]%tw[
(cid:14)ACSMEQ-
hpSeFX |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
of Ag-Tet, all phases featuring an EDS Sb a total number of 26 Bi grains. An analyt-
signature were assigned to Sb-phase category, ical resolution of 10x10µm was used during
composed of 100% Sb. Nevertheless, an even field scans. Considering that an approximate
greater underestimation compared to Ag was area of 10x20mm is scanned, this is a very
observed. Both,AgandSbmineralphasescom- detailed resolution. Nevertheless, small grains
monlyoccuratrathersmallgrainsizes. Despite below approximately 5-10µm are likely to be
using a high analytical resolution of 2x2µm misidentified or not identified at all. A signif-
during QEMSCAN® PMA, a significant quan- icantly higher resolution is used during PMA
tity of particles, containing Ag-Sb-phases, is (2X2µm). Notasinglegrainorevenpixelfea-
likely to have been either misidentified or sim- turinganEDSBisignaturewasidentifiedinany
ply not identified at all. Furthermore, particle of the analysed 1.5 million particles from 12
abundances containing Ag- and Sb-phases are samplesanalysedviaPMA.
overall rather low, consequently, analytical un- QEMSCAN® results do not preclude the oc-
certaintyishigh. currenceoftraceamountsofdiscreteBi-mineral
(c) Significant concentration levels of Ag phases. However, observations clearly suggest
and Sb were observed in galena (see section that their quantities are negligible and galena
4.3). Galena is a major constituent in sulphide is the only major Bi-carrying mineral phase
ore at Elura. Trace element compositions as suggested by observation made based on
were not accounted for during QEMSCAN® mineral chemical compositions determined via
analysis, thus may substantially contribute to EMP and LA-ICP-MS (see section 4.2.1 and
the systematic underestimation of Ag and Sb section4.3).
concentrations.
Iron sulphide characterisation
Microscopic and mineral chemical observa-
tions, part of this study, were confirmed by Four major pyrite generations were identified
the QEMSCAN® results. Ag-Tet the major during microscopic studies as part of this work
Ag-hosting mineral phase in addition to trace (see section 4.1.1). The early, pre-base metal
amountsofnativesilver. NoAs-Agphases(e.g. pyrite type A is commonly anhedral in shape
tennantite) were observed. Abundance of Ag and represents the most abundant pyrite type.
particles is low and mineral phases commonly It occurs as colloform, framboidal and cloudy
feature small grain sizes. Thus, QEMSCAN® varietiesandgenerallyformslargercompounds
resultsmaynotbesufficientlyaccurate. of individual grains. Pyrite type B also formed
prior to base metals but is exclusively euhe-
dral to subhedral in shape, although in places
Search for bismuth-bearing mineral
strongly corroded and replaced by base metal
phases
sulphides. Pyrite type C is contemporaneous
QEMSCAN® field scansof all15 thin-sections with base metal sulphides but occurs in minor
were used to search for potential discrete Bi- quantities. The post-base metal sulphide type
mineral phases that had not been identified mi- D is the fourth identified pyrite variety and is
croscopically. The identification process of Bi- observed in small quantities and small grain
phases was prioritised, i.e. any mineral phase sizes. The pyrite generations A and B in ad-
featuring an EDS Bi signature was assigned to ditiontopyrrhotiteaccountformostoftheiron
thebismuthmineralphasecategorypriortothe sulphides,thus,areimportantinrespecttomet-
assignment of any other mineral phases. An- allurgical problems. Grain size distributions,
alytical results are shown in appendix section pyrite/pyrrhotite ratios and the occurrence of
C.7.7,page675. pyrite varieties within different ore types were
NegligiblequantitiesofBi-phaseswereiden- investigated as part of this study. Results of
tified. Only 7 out of 15 thin-sections contain iron-sulphide characterisation are presented in
220 |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
100
SiPy & SiPo
6.6.3 Grain size distributions
Py
80 Po
Grain size data and distributions are based
60 on equivalent spherical diameter sizes (ESD)
calculated from 2-dimensional particle cross-
40
R² = 0.9375 sections based on PMA or field scan data. It
20 needs to be kept in mind that grain sizes cal-
culated from PMA results do not reflect real
0
insitu grain sizes, because results are based on
100 0 20 40 60 80 100
SiPy & SiPo
X of total Fe -sulphides broken sample material. These grain sizes are
Po Py
80 [wt%] Po nevertheless important for metallurgical prob-
lems. In order to gain at least some informa-
60
tion about insitu grain size distributions, they
40 arecomparedtothosecalculatedfromfieldscan
R² = 0.1718 results. However,theseresultsdonotnecessar-
20
ily reflect insitu grain sizes either, but reflect
0 an estimated combined size distribution of in-
0 20 40 60 80 100 dividual mineral grains and larger compounds.
X of total Fe -sulphides
Po [wt%] Small grains below ~15µm are not determined
correctly because of the lower analytical reso-
Figure 6.20: Associations of pyrite type A and
Bwithpyrrhotitecalculatedasfractionrelativeto lution (10x10µm pixel size) used during field
t Po yt Bal +iro Pn o)s ,u Xlp Ph yid Bes a; nw de Xre PX oP cay lAcul= ateP dy aA n/ a( lP ogy uA e.+ scans.
Grain size distribution data and diagrams for
(a)
100 sphalerite are presented in section C.2.1, page
90 529 (PMA) and section C.7.8, page 676 (field
80
scan). PMA analyses suggest that ore from
70
the eastern and central part of 5/5-ST con-
60
50 tains slightly coarser sphalerite at a P80 of
40 ~38µm (P80 defined as 80wt% of all grains
30 Cent
of a particular mineral phase are smaller than
20 East
10 West the given grain size) compared to 5/5-W with
0 a P80 of ~30µm. The average grain size, de-
1 10 100 fined as P50, yields ~20µm for 5/5-E and 5/5-
(b) Grain size [µm]
100 C but is ~12-13µm for 5/5-W. Individual sam-
90 ples from eastern and western stope parts fea-
80
ture rather consistent grain size distributions,
70
whereas those from 5/5-C are slightly differ-
60
50 ent and reflect the transition from pyrrhotite-
40 to pyrite-dominated ore zones. Field scan data
30
SiPy & SiPo showedthatsphaleriteinmassivepyrrhotiticore
20
Py
10 is slightly coarser compared to massive pyritic
Po
0 and semi-massive ore. Overall, grain sizes
10 100 1000
rangebetween~35(P10)and~500µm(P90).
Grain size [µm]
Galena grain size distributions based on
Figure 6.21: Grain size distributions of galena
PMA showed similar trends as sphalerite (sec-
based on (a) particle mineral analysis (PMA) and
(b)fieldscandata. tion C.3.1, page 563 for PMA and section
C.7.8, page 676 for field scan). P80 values
rangebetween~30and40µmforallstopeparts
222
sedihplus-
eF
latot
fo
X
sedihplus-
eF
latot
fo
X
]%tw[
AyP
]%tw[
ByP
]%tw[
nG
]%tw[
nG |
ADE | ®
6.6 QEMSCAN results from PMA and field scans
100 SiPy & SiPo nificantly lower for 5/5-W. Average grain sizes
90
Py (P50) are ~15µm for eastern and central stope
80
Po
70 parts and ~11µm in the west. According to
60 field scan results, the different ore types fea-
50
ture rather similar chalcopyrite grain size dis-
40
30 tributions. However, the quantity of fine min-
20 eral grains (<20µm) is with ~44wt% (Py)
10
and ~38wt% (Po) significantly higher in mas-
0
10 100 1000 sive compared to semi-massive ore with only
Grain size [µm] ~27wt%. The sizes of the largest observed
100
SiPy & SiPo grainsrangebetween60(Py)and100µm(SiPy
90
Py
80 andSiPo).
Po
70
Grain sizes of pyrite are rather uniform
60
throughout the 5/5-ST with a P80 of ~50µm
50
40 and a P50 ranging between ~23 and~28µm
30 (section C.5.1, page 631) according to PMA
20
analysis. Pyritecontainedinsamplesfrom5/5-
10
0 W is slightly smaller grained compared to the
10 100 1000 other stope parts. Field scan results (section
Grain size [µm]
C.7.8, page 676) showed that pyrite contained
Figure 6.22: Grain size distributions of pyrite inmassivepyriticoreismuchcoarsercompared
type A and B for different ore types. Data based
to pyrrhotite-dominated ore. The large grains
onfieldscanresults.
or aggregates predominantly consist of pyrite
type A and B as described earlier. Pyrite grain
with larger grain sizes observed in 5/5-E (fig- sizes range between ~60 (P10) and ~500µm
ure 6.21). Average grain sizes (P50) are rather (P90) in massive pyrrhotitic ore, and between
consistent and range between 12 to 14µm. In- ~80 (P10) and in excess of 1500µm (P90) in
dividualsamplesofthethreestopepartsfeature massive pyritic ore. Field scan grain size dis-
grainsizedistributionswhicharenotasuniform tribution data for pyrite type A and B are pre-
as those of sphalerite, indicating a pronounced sentedinfigure6.22. Grainsizesofpyritetype
heterogeneityofgalenagrainsizes. Particleim- Aarevariable. Itcommonlyformslargeaggre-
ages of galena (appendix section C.3.1, page gates(seesection4.1.1)asindicatedbyamini-
563) show the cubic breaking characteristics of mumgrainsize(P10)between150and300µm.
galena caused by its perfect cleavages. ESD Pyrite type-B tends to by smaller in pyritic ore
grainsizescalculatedbasedonfieldscanresults compared to pyrrhotitic ore with grain sizes
show that galena in massive Py ore is signifi- ranging between 45 (P10) to 200µm (P90) and
cantlysmaller(P50of35µm)comparedtomas- 75(P10)to250µm(P90),respectively.
sive Po (P50 of 60µm) and semi-massive ore Only minor pyrrhotite occurs in pyrite-
types (P50 of 80µm). Approximately 25wt% dominated ore and features small grain sizes
galena is smaller than 20µm. The upper grain at a P80 below ~9µm (PMA, section C.5.2,
sizes(P90)rangebetween~80(Py)and800µm page 633) or below ~35µm (field scan, sec-
(Po). tion C.7.8, page 676). The eastern and cen-
Similar characteristics were observed for tral stope parts contain significant quantities of
chalcopyrite grain size distributions based on pyrrhotitic ore, thus pyrrhotite, which is rather
PMA results (section C.4.1, page 599 for PMA coarse at a P80 of ~50µm. The overall grain
and section C.7.8, page 676 for field scan). sizerangeofpyrrhotiteinPooreisdefinedbya
The P80s feature grain sizes just below 38µm P10 of ~30µm and a P90 above 1000µm. The
for 5/5-E and 5/5-C but is with ~20µm sig- large values are likely caused by compounds
223
]%tw[
A
epyt-
yP
]%tw[
B
epyt-
yP |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
of pyrrhotite and do not represent individual 100
90
pyrrhotitegrains.
80
Silver phases preferably occur in peripheral 70
zones of the orebody (chapter 4 and chapter 60
50 5). Those zones are predominantly composed
40
of pyritic massive and semi-massive ore. The 30 Cent
number of particles containing Ag phases is 20 East
lower in the pyrrhotite dominated 5/5-E zone 10 West
0
(appendix section C.1.2, page 504). Grain size
1 10 100
distributions show that Ag phases are signifi-
Grain size [µm]
cantly coarser in those peripheral zones (5/5-
C and 5/5-W) at P80s ranging between ~7 and Figure6.24: Grainsizedistributionsofargentian
tetrahedrite for the three 5/5 Stope parts. Data
~14µm(figure6.24). MostAgphasesobserved
basedonPMA.
in samples from the 5/5-E, i.e. in excess of
90wt%,aresmallerthan1µm. Long-termplant
performance data showed that Ag recoveries ESDgrainsizesbasedonfieldscansaresignifi-
drop significantly for particles small than 5µm cantlyoverestimatedandshouldnotberegarded
(figure 6.23). According to grain size distribu- asrealisticvalues(seesection6.6.2).
tions,onlyoresourcedfromwesternandpoten- Threeoutofthefourvaluablemineralphases
tially central stope parts will feature good Ag at Elura feature rather similar grain size dis-
recoveries. tribution trends in respect to the investigated
Grain size distributions calculated based on stope parts. Results show that ore sourced
field scan data (section C.7.8, page 676) indi- from 5/5-E and 5/5-C contains coarser spha-
cateasignificantshifttocoarsergrainsinpyritic lerite, galena and chalcopyrite than the west-
ore (~50wt% Ag-Tet smaller than 20µm) rela- ern stope part. Ore sourced from the latter
tive to pyrrhotitic ore (~94wt% Ag-Tet smaller location will most probably feature relatively
than 20µm). It needs to be kept in mind that worse grade/recovery performance during min-
100
90
80
70
60
50
40
30
20
10
0
1 3µm 5µm 10 20µm 38µm 53µm
Particle size [µm]
May/Jun 2009 Jul/Aug 2009 Nov/Dec 2009
Jan/Feb 2010 Mar/Apr 2010 May/Jun 2010
3µm 5µm 20µm
Figure 6.23: Long-term Ag recoveries for different particle size fractions
achieved at the mineral processing plant. Data supplied by CBH Resources
Ltd and pers. comm. Andrew McCallum, 2010. Dashed red lines are median
recoveries.
224
]%tw[
yrevoceR
gA
]%tw[
teT-gA |
ADE | ®
6.6 QEMSCAN results from PMA and field scans
100 ated particles in 5/5-C and 5/5-E, whereas only
50wt% is fully liberated in 5/5-W. Compari-
75
son of achieved long-term plant recovery rates
PS +3/-5 µm withliberationdatasuggeststhatonlyfullylib-
50
PS +5/-20 µm
erated particles are recovered (i.e. liberation
PS +20/-38 µm
PS +38/-53 µm 25 between 92 and 97%). Calculated theoretical
copper recovery rates based on liberation data
0
hasalreadybeendiscussedalongwithgalenain
100 75 50 25 0
Liberation of Gn [%] thepreviousparagraph. Grainsizereductionto
20µmviasecondarygrindingmayliberated20
Figure 6.28: Lead recovery and corresponding
particle liberation of galena for different particle to 30wt% of the initially unliberated particles
size fractions. Long-term Pb recoveries are shown
containingchalcopyrite.
infigure6.26.
Argentian tetrahedrite
larger particles (figure 6.26). Theoretical re-
covery rates and Pb concentrate qualities are Liberation data for Ag-Tet is presented in sec-
calculated based on achieved real plan perfor- tion C.6.3, page 644. The pyrite-dominated
mance and calculated liberation data (figure western part of the 5/5 stope contains most Ag
6.28). Chalcopyrite is included in the calcu- (66%). Particles smaller than 38µm carry on
lation as it is initially recovered along with average ~83wt%. Only approximately 30wt%
galena. Resultssuggestanoverallrecoveryrate of all particles that contain Ag-Tet are fully
forgalenabetween74and77%,consideringthe liberated (>90%). Despite the significantly
fineparticlesize+0/-3µmas100%lossoratthe coarser grain sizes of Ag-Tet from 5/5-W com-
long-term recovery rate of 36%, respectively. paredtotheotherstopeparts,liberationispoor.
Calculated copper recoveries range between 57 It is not feasible to increase particle liberation
and 60% and are within the observed long- for Ag-Tet via grain size reduction because of
term achieved copper recoveries range of 55 to itsfine-grainednature. Silverisrecoveredalong
61% (pers. comm. Andrew McCallum, 2010). withPb,representingabeneficialelementinthe
The theoretical Pb concentrate quality, includ- concentrate. Apart from fully liberated parti-
ing particles recovered for copper, yields con- cles, Ag-Tet may only be recovered if associ-
centrations of 73wt% Pb and 2wt% Cu. The ated with galena. Mineral associations are in-
resultsconfirmthatsignificantlyhigherconcen- vestigated as part of this study and described
trate qualities should be achieved at the given laterinthetext.
recovery rates as suggested by the theoretical
grade/recovery diagram. Sample preparation
6.6.5 Free particle surface
may, however, cause unrealistically good parti-
cleliberationforgalenaasalreadydiscussedin
Calculatedparticleliberationbasedonareaper-
section6.6.1.
centage is a common parameter used for ge-
ometallugical ore characterisation. For some
mineral separation techniques, in particular for
Chalcopyrite
differential froth flotation, the free particle sur-
QEMSCAN® results are shown in section face of minerals is important as more bubble-
C.4.3,page604. Approximately75wt%oftotal mineral interfaces may be formed if free sur-
Cu is contained in particles below 38µm. The face is higher. Free surface data is estimated
relative Cu contribution between stope parts from the 2-dimensional particle cross sections.
slightly deceases from 5/5-E (42%) to 5/5-C If area specific liberation of a certain mineral
(32%) and 5/5-W (26%). Between 60 and phase is high, its free particle surface will be
65wt% chalcopyrite is contained in fully liber- high as well. However, if liberation is low,
227
]%tw[
yrevoceR
bP |
ADE | ®
6. GEO-METALLURGICAL STUDY VIA QEMSCAN
cate how extensive the Pb concentrate may be narymineralassociationif95area%ofthepar-
dilutedwithothermineralphasescausedbyre- ticlecross-sectionconsistsoftwominerals(e.g.
covered chalcopyrite-containing particles. The Bn: sphalerite-pyrite). Ternary mineral associ-
case scenarios yield Cu recovery rates between ation is defined analogue but allowing the in-
66and82%. Maximumachievedlong-termav- tergrowth of three phases (e.g. Tn: sphalerite-
erage Cu recoveries are in the order of 61% pyrite-pyrrhotite). Remaining unclassified par-
(pers. comm. Andrew McCallum 2010). Even ticles,i.e. notassignedtobinaryorternarycat-
though slightly higher, real plant performance egories, are classified as complex mineral as-
arebestmetbycaseI.AratherhighCuconcen- sociations,sub-dividedintotwocategories,one
tratequalityisdefinedbytheparticlesubpopu- containing in excess the other less than 10wt%
lationofthisparticularrecoverycaseconsisting of the investigated mineral phase (e.g. spha-
in excess of 98wt% chalcopyrite. The results lerite). Pyrite and pyrrhotite are combined as
suggestthatdilutionofthePbconcentratecon- iron sulphide and treated as one mineral phase
sequentlytochalcopyriterecoveryisnegligible. internarymineralassociations.
GalenaandchalcopyriteareseparatedviaPb Only particles characterised by liberation
rougher and scavenger flotation cells prior to grades of less than 90% (<90area% of a par-
Zn flotation. Sphalerite associated with parti- ticularmineralphase)areinvestigated. Itneeds
clesrecoveredduringthisprocessingstagemay to be kept in mind that the results will only es-
lead to losses of Zn. Lead and Cu concentrate timate the quantity of metal contained in par-
compositionsofthethreerecoverycasescenar- ticles of certain mineral association categories.
iosforgalenaandchalcopyritewereusedtoes- It, however, does not quantify the liberation
timate such potential metal losses. For cases grade of these particles. For instance, two hy-
I to III, Pb concentrate contains between 0.1 potheticalparticlesmayfeatureabinaryassoci-
to 0.6wt% Zn and Cu concentrate between 0.1 ationbetweensphaleriteandpyrite(bothphases
and 1.1wt% Zn. The Zn concentration in Pb >95area%,witharea%ofsphalerite<90). One
concentrateislowanddoesnotrepresentasig- ofthoseparticlesmayconsistof80area%spha-
nificant loss. Zinc recovered along with Cu is lerite, equivalent to a liberation grade of 80%,
higher,however,consideringthelowabundance and17area%pyrite. Suchaparticlewouldfea-
ofchalcopyrite(averageCuconcentrationinore ture a relatively high liberation and if recov-
is0.2wt%),Znlossisnegligible. Itneedstobe ered would cause minor dilution of the min-
stressed that this is a rough qualitative estima- eral concentrate. A second particle may con-
tion and does not represent actual mass flows tain the same phases but with the inverse spha-
within the processing stream. Real Zn losses lerite and pyrite area%. This particle, if recov-
maythereforebehigher. ered, would in contrast cause significant dilu-
tionduetothehighamountofpyriteintergrown
withsphalerite. Inotherwords,if20wt%spha-
6.6.6 Mineral associations and
lerite is contained in binary association with
locking
pyrite, 15wt% may be contained in highly lib-
Particle mineral associations were investigated erated particles. Therefore, important mineral
in order to assess mineral intergrowths and associations that potentially lock-up significant
locking characteristics of the valuable sulphide metalshouldbeassessedwiththeconsideration
phases. Results based on PMA are presented of particle liberation of this particular mineral
in section C.2.5, page 555 for sphalerite, sec- associationcategory.
tion C.3.4, page 588 for galena, section C.4.5, Fieldscandatacannotbeusedtoassessmin-
page 623 for chalcopyrite and section C.6.4, eral associations via the same approach be-
page 649 for argentian tetrahedrite. Associa- causeitisnotbasedonbrokensamplematerial.
tionswerecalculatedbasedonmineralareaper- However, by the determination of relative pixel
centage (abbrev. area%). Particles feature a bi- adjacencies, mineral associations can be esti-
230 |
ADE | ®
6.6 QEMSCAN results from PMA and field scans
mated. Transitions between neighbouring min- in complex association in particles containing
eralphasesarecalculatedviascanningthesam- >10wt% galena. Approximately 80% of this
pleimagehorizontally. Countsaresubsequently metal is linked to associations with iron sul-
normalised to 100% as relative transitions and phides.
reflectrelativemineralassociations. Particle liberation data of the binary galena-
iron sulphides show that most of the Pb is con-
tainedinrelativelyliberatedparticles(liberation
Sphalerite
>60%) with ~17 out of a total of 28wt%. Par-
Based on PMA, binaries with pyrite and ticles in complex locking feature lower galena
pyrrhotite are the most important associations liberationwithonly~5outofatotalof15wt%
with sphalerite and contain ~13 and ~7wt% Pb contained in relatively liberated particles
Zn in the respective particle subpopulation. (liberation >60%). Figure 6.31 shows the four
Ternaryassociationswithironsulphides(pyrite most important mineral associations of galena
and pyrrhotite) and galena hosts approximately in particle images. Calculations from field
4wt% Zn. Only 2.9wt% Zn is locked up scan data showed similar results but suggests
in complex association (particle with >10wt% that mineral associations with sphalerite may
sphalerite). In excess of 40% of those 2.9wt% behigherinpyrrhotite-dominatedorerelativeto
Zn contained in the complex association cate- otheroretypes.
gory is caused by intergrowths with iron sul- The results suggest that, despite significant
phides. Nosignificantvariationisobservedbe- galenabeinginassociationwithironsulphides,
tweenstopeparts. a good liberation is commonly achieved during
Iron sulphides in binary association with grinding. The theoretical grade recovery dia-
sphalerite contain ~20wt% Zn. Approximately gramforPbsuggeststhathighconcentratequal-
half of the Zn metal is contained in relatively ity (>50wt% Pb in concentrate) should easily
liberated particles (>60% liberation), causing be achieved even at high recovery rates above
only limited dilution. Nevertheless, if lower 85%. Mineralassociationandliberationdatais
mineral concentrate grade is observed, it is supportingthisobservation.
likelyaconsequenceofelevatedcontentofiron
sulphidesandtoalesserextentgalena.
The estimation of mineral associations of
Chalcopyrite
sphalerite based on field scan data is in agree-
mentwithPMAresultswithironsulphidesand Comparison of long-term plan performance
galenabeingmostimportant. Theresults,how- with QEMSCAN® data shows that only
ever, suggest a significantly higher association highly liberated particles containing more than
with non-sulphide gangue phases (NSG) for 90area% chalcopyrite may be recovered. The
semi-massiveaswellasmassivepyrrhotiticore. investigation of mineral associations is thus
The difference is explained by good separation needless for geometallurgical characterisation
ofNSGandsulphidesduringgrinding. as only near-pure chalcopyrite particles are re-
covered. Microscopic observations indicate an
affinityofchalcopyritetoNSGphases. Neither
Galena
PMAnorfieldscanconfirmedthisobservation.
Binaries with iron sulphides (pyrite and The most important mineral associations with
pyrrhotite) and with sphalerite, as well as chalcopyrite are binaries with pyrite (~12wt%
ternary associations with iron sulphides and Cu)andsphalerite(~5wt%Cu),internaryasso-
NSGwereidentifiedasthemostimportantmin- ciation with iron sulphides and NSG (~5wt%),
eralassociations. Thesecategoriescontain~28, andincomplexlocking(~9wt%Cu). Associa-
~7 and ~4wt% Pb, respectively. A signif- tion with pyrite and sphalerite are in turn most
icant amount of Pb (~13wt%) is locked up importantinthecomplexlockingcategory.
231 |
ADE | Chapter 7
Interpretation and conclusions
SINCE its discovery, several scientists have combination with alternative methods. Never-
studied the Elura orebody and other de- theless, chlorite thermometry estimated a tem-
posits inthe Cobar region. Their findings, con- perature range between 314 and 343°C,similar
clusionsandproposedgeneticmodelsaresum- totemperatureconditionsproposedforthemin-
marised in chapter 3 on page 50. The aim of eralising fluid by other studies. Under ele-
thisstudywastoinvestigategeochemical, min- vated temperature conditions >300°C, galena
eralogical, mineral chemical and textural ore andpyrrhotiteareductileandmobile;chalcopy-
characteristics on a deposit-wide scale based rite is relative mobile. Sphalerite and pyrite
on a sample set covering the entire extent of are brittle and are not remobilised (e.g. Mar-
the orebody and all different ore types. Ge- shall and Gilligan, 1993). Ore textures devel-
ometallurgical ore characterisation was under- oped consequently to these different mineral
taken on a smaller scale, i.e. proposed stop- characteristics. Sulphide banding is commonly
ing area, in order to test whether a prediction defined by zones enriched in sphalerite and/or
in respect to mineral processing characteristics pyrite and those enriched in galena, pyrrhotite
of ore sourced from different stope parts can and chalcopyrite. Rounded clasts of sphalerite
be made. Aspects important for ore genesis as may occur in mylonitic zones. Galena and
well as applied to mineral processing are sum- chalcopyrite feature significant grain size vari-
marisedinthefollowing. ations. Large grains formed due to remobilisa-
tion into fracture zones, commonly hosted by
pyriteand/orwallrockfragments.
7.1 Aspects to the genesis of
Primary mineral parageneses are preserved
the Elura orebody
in undeformed massive sulphide ore. Inti-
mate, myrmekitic-like intergrowths of spha-
Elura’s sulphide ore features pronounced het- lerite, galena and pyrrhotite in pyrrhotite-
erogeneity in respect to grain sizes, texture and dominated ore zones in central parts of indi-
chemical composition. In many areas, the sul- vidual sulphide pipes show their cogenetic na-
phide body was affected by intense deforma- ture. Pyrite-dominated ore exclusively occurs
tion, preserved as sulphide banding, grain size in peripheral areas of the orebody and is best
reduction and, in places, as mylonitic textures. developed in the upper apophysis of the main
Sulphidephasesbehavedifferentlyunderdefor- lode zone, forming an outer shell enclosing
mational conditions. Chlorite thermometry is pyrrhotitic ore. In these pyritic zones, base
weakasitiswellknownthatchloritecomposi- metalsulphidesoccurinterstitialtoandfillfrac-
tioncanbesensitivetomanyfactors,notsolely tures or vughs of pyrite, partially replacing it.
to temperature. Temperature values should al- Replacement of pyrrhotite by base metal sul-
ways be taken with caution and only used in phidesissubordinate.
233 |
ADE | 7. INTERPRETATION AND CONCLUSIONS
Fourpyritegenerationswereidentifiedinthis cipitation at initially higher fluid temperatures.
study. Finegrainedsub-toanhedral,colloform, Elements enriched in the upper and periph-
framboidal and cloudy pyrite A formed in an eral zones of the mineralisation (e.g. Ag, Sb,
early,mostlikelylowertemperaturestageofan Tl, Hg) remained in solution as the fluid tem-
evolvinghydrothermalsystem. Increasingfluid peratures decreased. Subsequently, these ele-
temperatures lead to the formation of coarse ments became enriched in areas distant to the
sub-toeuhedralpyriteB,tosomeextentcaused influx zones of the metal-bearing fluid. Pro-
bythere-crystallisationofpyriteA.Bothpyrite longed and repeated fluid pulses most proba-
varieties formed prior to the base metal miner- blycausedinternalzonerefinementdissolution-
alisation, which is dominated by pyrrhotite, in- reprecipitation processes, similar to that ob-
dicativeforaratherreducinghydrothermalfluid served for VHMS deposits and amplified the
and sulphur activities that favour the precipi- verticalelementzonation.
tation of mono-sulphur sulphides. Only minor Zinc isotope composition of sphalerite
syn-basemetalsulphidepyriteCformedandis, (δ66Zn 0.220-0.450 ) suggests average
JMC
inplaces,intimatelyintergrownwithco-genetic continental crust as meta(cid:104)l source and that the
magnetite. Pyrite D is unrelated to and formed source reservoir had been effectively leached
afterthebasemetalmineralisingstage. Mostof by the hydrothermal fluid. Vertical isotope
thetotalpyritemasscontainedintheEluraore- fractionation trends within the orebody are
body formed prior to base metal sulphides and likely caused by syngenetic Rayleigh frac-
pyrrhotite. Sphaleritegeobarometrywasunsuc- tionation, which was most probably aided by
cessful and confirmed disequilibrium between vertical temperature gradients. Initial light
pyriteandsphalerite/pyrrhotite. crustal isotope signatures define two major
Iron-rich sphalerite (overall max. 8.22wt%) zones of fluid influx. Both zones, one at the
is common in lower and pyrrhotite-dominated base of the entire orebody and a second at the
partsoftheorebody. AshifttolowerFecontent transition between the lower and upper main
(overall min. 2.41wt%) is present in sphalerite lode zones, feature coinciding enrichments in
fromtheuppermainlodezoneand,inplaces,in Cu,inagreementwithearlierdescribedinternal
peripheral pyritic zones. The lack of pyrrhotite refinement mechanism and fluid temperature
inthosezones,thereplacementofpyriteandthe gradients. The fluid became heavier in isotopic
lowerFe-contentinsphaleritereflectthechang- compositionasitevolvedandcooled during its
ing fluid chemistry characterised by a decrease ascenttohigherlevelsinthemineralisation.
ofFeSactivityaspyrrhotiteprecipitatedincen- The occurrence of the second fluid influx
tralorezones. zone is explained by temporal differences be-
None of the investigated samples contained tween the formation of the lower and the up-
intactfluidinclusionsinsphalerite. Thisobser- per parts of the mineralisation. The geometry
vation is not surprising considering that defor- of the orebody itself suggests some differences
mationoftheorebodyhasoccurred. Sphalerite, intheformationofthelargesouthernmostmain
abrittlesulphidephase,wasmicro-fracturedal- lode zone and the smaller northern ore zones.
lowingfluidandgasphasestoescape. Thepres- Only the former most prominent ore shoot in-
enceoforganicmatterandlong-chainn-alkanes tersects the upper laminated unit, which con-
suggestthatthemineralisingfluidcontainedhy- finesallnorthernorezones. TheEluraorebody
drocarbons. mayhavedevelopedintwostages. Asthebasin
The Elura orebody features a pronounced inversion commenced, dilation was initially fo-
vertical geochemical zonation. Copper and to cussed along a transpressional fault corridor.
a lesser extent Cd and Co are preferably con- Zones of pronounced fracture-induced perme-
tained in pyrrhotite-dominated ore and gener- ability occurred in sandstone-rich sediment se-
ally enriched towards increasing depth. Their quencesandwererestrainedbyhorizonssignif-
enrichment is caused by early sulphide pre- icantly enriched in silt- and mudstone (i.e. the
234 |
ADE | 7.1 Aspects to the genesis of the Elura orebody
upperandlowerlaminatedunits). Thenorthern genic than (b) an oxidising basement-derived
ore zones and the preliminary lower main lode fluid. The latter fluid led to the formation of
formed. A change in the regional stress regime Cu- and Au-dominated ore systems. It has
during progressive basin inversion caused the been shown by this study that the reducing hy-
development of major dilation and jogs at the drothermalfluidshiftedtohigheroxygenfugac-
southern end of the orebody. Subsequently, the ities during the latest stages of the mineralis-
lower main lode was further enlarged and the ing event. The changing fluid characteristics
uppermainlodeformed. suggest an increased proportion of basement-
Bismuth and Bi-bearing mineral phases are derived, less radiogenic fluid. The less radio-
common minor constitutes in ore from south- genic Pb signatures are exclusively observed in
ern deposits in the Cobar region, e.g. the the upper main lode zone. The changing fluid
New Cobar and the New Occidental deposits. compositiontherebyconfirmstheproposedtwo
A sulphur-poor, basement-derived hydrother- stagegeneticmodelwithaninitialformationof
mal fluid has been proposed as metal-carrying the lower mineralisation and later development
media in previous studies. Significantly ele- oftheuppermainlodezone.
vated concentrations of Bi, Se and Te were Changes in rare earth element signatures of
present only in the lowermost main lode ore differentoretypesareexclusivelycontrolledby
zone and reflect a change in fluid chemistry modal mineralogy, i.e. content of muscovite,
and temperature (Maslennikov et al., 2009). chlorite and carbonate gangue phases. No lo-
These element enrichments are most probably calrareearthelementremobilisationcausedby
caused by an evolving hydrothermal fluid, fea- extreme fluid-rock interaction is needed in or-
turing a decrease in sulphur activity, an in- dertoexplainobservedcharacteristics. Thepro-
creaseinoxygenfugacityandtemperatureover nounced positive Eu anomaly is indicative of a
time. The observed close genetic relationship reducingfluidenvironmentand/ortemperatures
between pyrrhotite, pyrite and magnetite, de- inexcessof250°C.
spiterare,confirmsthisproposition. Theevolv- Assessment of mass/volume changes and el-
ingfluidintroducedelementssuchasBi,Seand ement mobility via the isocon method showed
Te during the latest stages of hydrothermal ac- that replacement of the host lithology was a
tivity. This fluid features compositional sim- negligible mechanism during ore formation.
ilarities to the hydrothermal fluid proposed to Sulphides, including pre-base metal sulphide
haveformedCu-AuandAu-dominateddeposits pyrite, exclusively formed in sites of increased
south of Cobar (e.g. the New Cobar and the dilationandfracture-inducedpermeability. The
NewOccidental). only substantial replacement that occurred dur-
David (2005, 2008) compiled and presented ingbasemetalformationisthatofpyrite.
Pbisotopedataofsamplestakenfromtheupper Concentrations of most platinum group el-
(aboveapprox. 9600RL)andthelowerpartsof ements are close to average continental crust
the Elura orebody (figure 106 on page 152 in reservoir data. Platinum is relatively enriched,
David,2005). Awell-definedclusterofconsis- most probable due to the interaction of the
tent radiogenic Pb is present within the lower ascending hydrothermal fluid with ferroman-
mineralisation. Other samples, most of them ganeserichsedimentsequencesthatcontainel-
sourcedfromtheuppermainlodezone,feature evatedPtconcentrationscomparedtootherplat-
a trail of Pb isotope compositions originating inumgroupelements.
fromthedataclustertowardslessradiogenicPb. Radiogenic age determination via the Re-
Previous studies proposed two contrasting hy- Os isotope system resulted in an isochron age
drothermal fluids responsible for the genesis of of 378±15Ma, representing the first age con-
theCobardeposits: (a)areducingbasinalfluid, straintestablishedviadirectlydatingsulphides.
responsibleforPb-Zn-dominateddeposits. This The observed relatively non-radiogenic initial
fluid was proposed to be slightly more radio- γ of ~170 is either caused by juvenile con-
Os
235 |
ADE | 7. INTERPRETATION AND CONCLUSIONS
tinental crust of the Lachlan Fold Belt as metal ing the mineralising event. Semi-massive ore
source or by a contribution of primitive mantle formed during early fracturing events with sul-
Os. phidesprecipitatingandfillinginterstitialspace
betweenwallrockfragments. Asfracturingand
dilationcontinued,permeabilitywascreatedin-
7.1.1 The modified genetic model
ternally of the existing sulphide body. Subse-
for the Elura orebody∗
quently,massivesulphideorewithoutwallrock
The Elura deposit is epigenetic and syn- componentsformedincentralorezones.
deformational in nature, as it formed in Changing fluid temperatures, internal refine-
zones of pronounced fracture-induced perme- ment via dissolution-reprecipitation of earlier
ability in lithified turbiditic sedimentary strata. formed mineral phases and a changing fluid
The ore formation took place at 378±15Ma composition over time caused the pronounced
based on radiogenic Re-Os age determination. vertical mineralogical and geochemical zona-
Sandstone-rich sequences reacted more brit- tion.
tle during deformation compared to mud- and TheEluraorebodydevelopedintwoseparate
siltstone-rich zones and were the preferable mineralisingevents. Initially,lowerpartsofthe
site for sulphide precipitation. The deposit is mineralisationformedinzonesofdilationalong
therefore at least to some extent lithologically atranspressionalfaultcorridor,followedbythe
controlled. The rather compact and compe- formation of the upper and the upgrade of the
tentlimestoneunderneaththedepositmayhave lowermainlodezone.
aidedthedevelopmentofdilationalzones. Upon cessation of the hydrothermal activity,
Prior to the introduction of base metal sul- compression continued during basin inversion.
phides, a massive pyrite body formed in zones Therelativeductilecharacterofsulphidescom-
of positive dilation during an early, low tem- pared to lithified turbiditic sediments caused
perature stage as the hydrothermal system strain to be focussed within the orebody. The
ramped up. Prolonged fault activity and sub- sulphide body was subsequently deformed and
sequentfracturingallowedtheascendingmetal- vertically elongated. Deformational-induced
bearing fluid to influx the earlier formed pyrite sulphide remobilisation likely upgraded and
body. Decreasing fluid pressure and accom- amplifiedverticalorezonationinrespecttoore
panying adiabatic cooling caused the break- mineralogyandgeochemistry.
down of metalliferous complexes and initiated
the precipitation of base metal sulphides and 7.1.2 Differences to previous ge-
pyrrhotite. No evidence of fluid mixing as im- netic models and important
portant mechanism for ore formation was iden- new aspects
tifiedinthisstudy.
Repetitive fracturing events caused the accu- The genetic model presented in this study is
mulation of massive pyrrhotitic ore in central similar to epigenetic syn-deformational models
core zones, representing major fluid conduits. proposedfortheEluraandotherCobardeposits
The brittle character of pyrite caused the for- inpreviousstudies(Schmidt,1980,1990;Glen,
mation of reticular micro-fractures in the mas- 1987;Brill,1988;DeRoo,1989b;Hinmanand
sive pyrite shell. The mineralising fluid mi- Scott, 1990; Scott and Phillips, 1990; Perkins
grated and infiltrated those zones, precipitated etal.,1994). Differencesandimportantnewas-
basemetalsulphidesandpartiallyreplacedear- pectsarebrieflysummarisedinthefollowing.
lierpyrite,causingtheformationofpyritedom- The genesis of the deposit is to some ex-
inatedbasemetalore. tent lithologically controlled. Mineralisation
Microscopic and geochemical investigations
∗.Themodifiedgeneticmodelisbasedonresultofthis
in this study showed that replacement of host
studyandwiththeconsiderationofproposedmodelsof
sediments was a negligible mechanism dur- otherauthorspresentedinchapter3onpage50
236 |
ADE | 7.1 Aspects to the genesis of the Elura orebody
occurred in two separate mineralising events. malmetamorphism. TheRe-Osisochronageof
Zones of fracture-induced permeability were 378±15Ma established in this study estimates
initially focussed in sediment sequences en- the ore formation event and was most proba-
richedinsandstone. bly not affected by deformation. This age is in
DeRoo(1989b,a);Schmidt(1990)suggested agreementwithpreviousstudiesbySun(2000);
the orebody formed predominantly via metaso- Sun et al. (2000) that suggested an ore forma-
matism and selective replacement of host sedi- tion age of 376-379Ma based on 40Ar-39Ar of
ments. Infillofsulphideinfracture-induceddi- sericite contained in massive sulphide samples.
lation was a subordinate mechanism during ore The Elura orebody is strongly structural con-
genesis. This study showed that replacement trolled and ore textures clearly show that sul-
of the host lithology was negligible during ore phides were fractured and/or remobilised dur-
formation. Sulphides, including pre-base metal ing deformation. Whole rock dating via K-Ar
sulphide pyrite, formed exclusively in sites of and 40Ar-39Ar suggested that basin inversion
increased dilation and fracture-induced perme- and main cleavage development in the Cobar
ability. Prolonged periods of repetitive fractur- Group sediments occurred at 395 to 400 Ma
ing, fluid pulses and internal accretion of sul- (Glen et al., 1992), around 20Ma prior to ore
phides led to the concentric development and formation. Theageofthemaincleavagedevel-
orezonationoftheEluraorebody. opment reflects the climax of the basin inver-
Themostrecentgeneticmodelwasproposed sion. However,itsoveralldurationisunknown.
by David (2008) suggesting an early minerali- Sun (2000); Sun et al. (2000) identified a sec-
sation in semi-lithified sediments during basin ond major deformational event proximal to the
formationfollowedbydeformationandmodifi- Elura orebody at 385-389Ma via 40Ar-39Ar on
cation during basin inversion. Mixing of base- cleavage-parallel sericite, suggesting that basin
ment and basin derived fluids are an important inversion and accompanying compression and
mechanism in this model. The author argued fault activities continued over a prolonged pe-
that an epigenetic origin is unlikely as lithi- riodoftime. Agesoforegenesisandhostlithol-
fied sediments feature very low permeabilties, ogy clearly preclude a syn- but are in favour of
thus the mineralising fluid would have exclu- anepigenetic,syn-deformationalorigin.
sively migrated along reactivated faults. Seis- Ore textures show that the orebody had been
micpumpingwouldhavecreatedlargezonesof tosomeextentremobilised,elongatedandmod-
hydrothermal brecciation, which, according to ified by deformation. Peripheral parts of the
David (2008), do not exist at Elura. However, orebody are dominated by pyritic ore. Several
thisthesisshowsthatbrecciastringer-typemin- studies including this one (e.g. De Roo, 1989b;
eralisation surrounding the orebody is in fact Lawrie and Hinman, 1998; Sun, 2000) suggest
rathernarrowinmostupperorezones,whereas that a pyritic body formed prior to the intro-
at depth, it is large and forms a zone of pro- duction of base metal sulphides. The concen-
nouncedbrecciation,vein-stylesulphideminer- tric nature of the Elura orebody and its struc-
alisationandsignificantpervasivesilicification. tural framework doesn’t support a syngenetic
This zone was interpreted in this study as ma- originfollowedbydeformationalemplacement.
jor fluid influx zone. The mineralising fluid as- Certain sulphide phases are easy to remobilise
cended along reactivated faults and was chan- whereas others aren’t. The brittle nature of the
nelled and focussed into zones of pronounced outer pyritic ore shell makes it rather resistive
dilation. Such hydrothermal brecciation is ob- to deformational remobilisation. If the Elura
viously not possible in semi-lithified sediment orebodyformedinsemi-lithifiedsedimentsand
sequences. was subsequently remobilised and emplaced,
Steinetal.(2001)showedthattheRe-Ossys- the coherence between concentric pyrite and
temisremarkablyrobustandnotdisturbedeven pyrrhotitedominatedorezoneswouldhavecer-
duringintensedeformationandhigh-gradether- tainly been lost and a vertical mineralogical
237 |
ADE | 7. INTERPRETATION AND CONCLUSIONS
zonation likely developed. Mobile phases (i.e. man, 1998; Lawrie et al., 1999; Jiang et al.,
galena, pyrrhotite) would have travelled fur- 2000). The resulting reducing metalliferous
thest, whereas pyrite and sphalerite as rather pore fluids were expelled from those beds dur-
immobile phases would have became enriched ing diagenesis and metamorphosis and would
in zone proximal to the initial mineralisation. have accumulated in sandstone rich sequences,
SuchazonationisnotpresentatElura. which are characterised by higher permeabil-
According to Lawrie and Hinman (1998), ity. Other parts of the strata within the Co-
the Cobar deposits formed as a consequence barBasin,e.g. theMourambaGroupsediments
of mixing of two contrasting fluids: a base- at its base, consist of outwash fans, shallow
ment derived oxidising and a basinal reducing water clastics and locally of minor felsic vol-
fluid. This study confirmed a change in the hy- canic. Such lithologies most probably contain
drothermal fluid’s oxygen fugacity, but no evi- less organics, feature increased permeability
dence has been found that favours fluid mixing andmaycontainsignificantquantitiesofoxidis-
as important mechanism for ore genesis. Sun ingsulphatephasesinevaporatesequences. Hy-
(2000); Sun and Seccombe (2000); Jiang et al. drothermal fluids that migrated through and in-
(2000)reachedsimilarconclusionsbasedonhy- teractwithsuchlithologieswouldhavebecome
drogen/oxygenisotopeandfluidinclusionstud- saline, leach metals, release sulphur due to sul-
ies. phatereductionand,indoingso,thefluid’soxy-
Observations made during this investigation genfugacitywouldhaveincreased.
suggest that the reducing hydrothermal fluid It appears feasible that mineralising fluids
evolved temporal, characterised by an increase which formed within the basin but at differ-
in its temperature and oxygen fugacity. The ent stratigraphic positions also featured differ-
latter had a very minor impact on the mineral- encesintheirphysico-chemicalconditions(e.g.
isation at Elura and represents the latest fluid temperature, composition, pH, oxygen fugac-
pulses before the mineralising event ceased. ity, etc.). In the case of Elura, as the de-
Whether the changing fluid composition is, posit in highest stratigraphic position of all
as suggested by Lawrie and Hinman (1998), known deposits in the Cobar region, the hy-
caused by fluids sourced from different reser- drothermal mineralising fluid is predominantly
voirs (i.e. basement and basin) remains spec- sourcedfromaquiferswithintheturbiditicsed-
ulative. The reducing nature of the basinal imentary sequence that contain reducing met-
fluidiscausedbyabundantorganicmattercon- alliferous brines. Only during the latest ge-
tained within turbidites, particularly in mud- netic stages, minor basinal fluid characterised
andsiltstonebedsthatalsocontainsyn-anddi- byelevatedoxygenfugacityandincreasedtem-
agenetic pyrite as sulphur source, and poten- peratures ascended from deeper aquifers to the
tially as source for Zn, Pb, Cu and Au (Large site of ore formation. Deposits that formed
et al., 2009). Organics were oxidised, ther- in deeper stratigraphic position (e.g. deposits
mally matured and converted to hydrocarbons of the Cobar Goldfield) are more proximal to
during metamorphism. The presence of hydro- aquifers at depth and thus dominated by hy-
carbons has been verified by this and previous drothermal fluids that feature elevated oxygen
studies (e.g. Seccombe, 1990; Lawrie and Hin- fugacities.
238 |
ADE | 7.2 Important aspects for mineral processing
7.2 Important aspects for
compositionalnatureofElura’sore.
mineral processing Pronounced grain size variability of spha-
lerite and especially galena and the subsequent
decreased particle liberation were identified as
Zincand,inparticular,Pbrecoveryratessignif-
the main reasons for fluctuating Zn and Pb
icantlyfluctuatedsinceproductionbeganinthe
metal recovery rates. Base metal sulphides
early 1980s. Silver recoveries were commonly
in massive pyritic ore are commonly charac-
poor and rarely exceeded 50%. Recently, high
terised by a reticular texture, where they fill
concentration levels of Bi were encountered in
fractures and interstitial space in compact but
leadconcentrate. Partofthisstudywastoinves-
fracturedpyrite. Sphaleriteandgalenaaregen-
tigate general and geometallurgical ore charac-
erally much finer grained in this ore type com-
teristicsandtodeterminethecauseforthemin-
pared to massive pyrrhotite ore and rather vari-
eralprocessingproblems.
able in semi-massive ore types. Galena, how-
The integration of ore petrography and geo- ever, may occur as coarse patches in excess of
chemistry highlighted the pronounced hetero- 100µmthroughoutthedepositandinalldiffer-
geneity in respect to grain sizes, texture, min- entoretypesasitwas,inplaces,stronglyremo-
eral compositions and trace element geochem- bilised by deformational force. Grain sizes of
istry of Elura’s sulphide ore. Sulphide par- chalcopyriteareasvariableasthoseofgalena.
agenesis is simple, comprising major pyrite, The reasons for poor Ag recoveries are be-
pyrrhotite, sphalerite and galena, minor marc- cause grain sizes of the identified Ag-phases
asite, chalcopyrite and arsenopyrite, and trace (i.e. argentian-tetrahedrite and silver) are very
tetrahedrite (±freibergite), native silver and small in central pyrrhotite-dominated zones of
magnetite. Tennantite was described as Ag- the Elura orebody with approximately 70 to
hosting mineral phase in previous publications 98wt% smaller than 5µm. Particle liberation
but was not identified in this study. Silver con- will be poor for such small grains and entrain-
centrationsintetrahedriterangebetween~19to ment during flotation significant. In peripheral
~31wt%. Some analyses with Ag concentra- semi-massiveandmassivepyrite-dominatedar-
tions in excess of 31wt% feature stoichiome- eas, grain sizes are larger and reach maxima in
triessimilartofreibergite. Chalcopyriteis,apart theuppermostareaofthemainlodezoneupto
from tetrahedrite, the only other Cu-bearing 200µminsize.
phase. Significant quantities of wall rock as
As Ag-phases report to the lead concentrate
fragments, in places strongly silicified and in-
and upgrade its value, their associations with
corporated in sulphide groundmass, occurs in
other mineral phases are of great importance.
peripheral semi-massive (SiPy and SiPo) and
Giventhatgrainsizesaresmallandparticlelib-
brecciastringer-style(VEIN)oretypes. Siderite
eration is poor, Ag-phases may still be recov-
and to a lesser extent quartz represent the most
ered if intergrown with galena, but rejected to
important non-sulphide gangue phases in mas-
tailingsif,forinstance,associatedwithpyriteor
siveoretypes. Quartzisslightlymoreabundant
non-sulphideganguephases. Intergrowthswith
inpyrite-dominatedmassiveore.
pyrite in pyrite-dominated ore and with non-
Based on median concentrations, massive sulphidegangue, predominantlyFe-carbonates,
and semi-massive ore contains in the order of inpyrrhotite-dominatedorerepresentbyfarthe
9.6 to 10.2wt% Zn and 5.1 to 5.9wt% Pb. The most important mineral associations for Ag-
highest Zn concentration (16.4wt%) was de- phases. Abundant associations with galena are
termined in massive pyrrhotitic ore, whereas limited to the uppermost parts of the main lode
massive pyritic ore contained maximum Pb zone and, in places, in the uppermost areas of
(30.3wt%). The large concentration range of thenorthernorezones.
Zn (4.6-16.4wt%) and in particular of Pb (1.7- High Ag concentration levels up to
30.3wt%) reflect the heterogenic textural and ~2,300ppm were present in galena from
239 |
ADE | 7. INTERPRETATION AND CONCLUSIONS
the upper- and lowermost areas of the main concentratequalities.
lode zone, caused by galena-miargyrite and Complete transformation of pyrrhotite to
galena-matildite solid solutions, respectively. marcasite is limited to the uppermost areas of
Galena recovered from those ore zones would the sulphide mineralisation and caused by de-
benefit Ag recoveries, although having a neg- scending oxidising surficial water that became
ative effect on the concentrate quality in the acidic upon interaction with sulphides. Minor
case of galena-matildite due to the significant incomplete transformation is observed even in
Bicontent. deepest sections of the orebody. High marca-
Alteration of argentian-tetrahedrite to native sitecontentwillhaveanegativeimpactonmin-
silver and chalcopyrite is widespread. Inter- eral concentrate quality as it features a higher
mediate stoichiometric compositions, deviating floatability than pyrite under certain pH condi-
from ideal tetrahedrite, are common. Little tions(Bulatovic,2007). Furthermore,marcasite
is known about flotation characteristics of Ag- oxidises more quickly than pyrite, causing the
mineralphasesingeneral(Bulatovic,2007)and production of sulphuric acid, subsequently af-
presumably even less for those of alteration fectingthepHwithinthemineralprocessingcir-
products. Achangeofflotation/activationchar- cuit.
acteristicsofalteredtetrahedritemaybeanother
The geometallurgical ore characterisation
causeforlowsilverrecoveries.
study at stope scale (5/5-Stope) revealed sig-
Galena was identified as the only important nificant differences in composition, grain sizes,
mineral phase containing Bi with a maximum liberation,andthus,flotationcharacteristicsbe-
detectedconcentrationof5,645ppmandcaused tween the three investigated stope parts. Com-
bymatildite-galenasolidsolution. HighBicon- pared to the western stope part, sphalerite and
centrations are limited to lowermost main lode galena are coarser grained, feature higher par-
ore zone. Ore sourced from this area will in- ticle liberation and free particle surface in the
evitably cause significant Bi concentrations in easternstopepart. Calculatedtheoreticalgrade-
lead concentrate. The only feasible strategy to recovery diagrams suggest that ore from the
controlconcentratequalityisoreblending. western stope part will have by approximately
Manganese concentrations in sphalerite are 10% lower Zn and Pb recoveries. Whilst
low (<300ppm) throughout the deposit. Mer- calculated theoretical grade-recovery data of
cury is another important penalty element for Zn is close to long-term average plant perfor-
zinc concentrate and was determined at con- mance, a significant deviation was present for
centrations up to 100ppm but is generally con- Pb. The deviation translates to a lower con-
tained at concentrations in the order of 25ppm centrate grade (~30wt% Pb) or lower recov-
(overallmedian). ery(~20%Pb). Recoveryratesandconcentrate
Arsenicisapenaltyelementforconcentrates. qualities calculated on the basis of particle lib-
Elevatedarsenopyritecontentinoremayleadto eration and free particle surface confirmed this
poorer concentrate qualities if not sufficiently observationandsuggestthatasignificantlybet-
depressed during froth flotation. Arsenopyrite ter grade/recovery performance for galena/lead
features a clear affinity to pyrite-dominated ore should be achieved. Excessive Pb losses or
with a maximum determined As concentration concentrate dilution with sulphide and non-
of1.5wt%. ConsistentlyhighAsconcentration sulphide gangue phases may be the cause for
were present throughout the upper main lode the observed poor performance. However, the
ore zone. Ore from this part of the minerali- samplepreparationtechniqueusedinthecourse
sationmaycausepoorconcentratequalities. ofthisstudymayhavecausedsubstantiallydif-
High concentrations of Tl and Hg were ferent liberation compared to real plant grind-
present in pyrite from the uppermost main lode ing via semi-autogenous and ball mill. Con-
orezone. Insufficientdepressionofpyritewhen sequently, concentrate grade and plant recover-
orefromthosezonesistreatedwillleadtopoor iesmaybeoverestimated. Geometallurgicaltest
240 |
ADE | 7.2 Important aspects for mineral processing
workontheactualflotationfeedwouldneedto residuetime.
beundertakeninordertoassessthisissue. Most Improvement of Pb recoveries seems prob-
Agiscontainedinthewesternpyrite-dominated lematic for fine grained ore, as a further parti-
ore. cle size reduction would be needed in order to
liberate fine grained galena. This would cause
Results showed that, despite individual ore
over-grinding, sliming and significant entrain-
types that feature similar ore characteristics in
ment, and subsequently even lower lead recov-
different locations throughout the deposit, a
ery rates. The cause for the significant devi-
generalised characterisation is not feasible as
ation between theoretical and long-term plant
Elura’s ore is far too heterogeneous in respect
recovery-grade performance need to be identi-
to texture and composition. However, charac-
fied in order to assess whether significant im-
terisation on smaller scale, i.e. stoping areas,
provements in the lead flotation circuit are fea-
has the potential to identify significant varia-
sible. Optimisation of the secondary grinding
tions in flotation characteristics for ore sourced
stageofsphaleritemayhavethepotentialtoim-
from different stope parts. This knowledge
proveZnrecoveriesforfinergrainedore. Good
will enable the prediction of (a) relative re-
Ag-recoveriescanonlybeexpectedfororefrom
covery rate changes, (b) potential changes in
the uppermost areas of individual ore zones,
concentrate qualities and (c) changes in mill-
wheretetrahedriteiscoarseandintergrownwith
throughputasafunctionofsilicaandwallrock
galena. Considering the fine grain sizes, the
content. Consequently,predictionsofmineeco-
complexmineralassociationswithphasesother
nomics will be improved and strategies may be
than galena and the uncertain flotation charac-
setup and implemented in order to maximise
teristics of altered tetrahedrite, a recovery rate
the performance of the mineral processing fa-
inexcessof50%doesn’tseemveryprobablefor
cility. Amongstthosestrategiesareblendingof
ore from most areas within the Elura deposit.
ore sourced from zones of different ore char-
Elevated Bi in lead concentrate is easily man-
acteristics, screening, optimisation of type and
ageablebyblendingorefromdifferentsource.
quantity of chemicals used in flotation (frother,
collector or depressant) and mill and flotation
241 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Abstract
This thesis work has analysed environmental impacts and health aspects in the mining industry of
copper, uranium and gold with the aim of determining the relative performance, in a given set of
parameters, of the uranium mining industry. A selection of fifteen active mining operations in
Australia, Canada, Namibia, South Africa, and the United States of America constitute the subject of
this study. The project includes detailed background information about mineral extraction methods,
the investigated minerals and the mining operations together with descriptions of the general main
health hazards and environmental impacts connected to mining. The mineral operations are
investigated in a cradle to gate analysis for the year of activity of 2007 using the economic value of
the product at the gate as functional unit. Primary data has been collected from environmental
reports, company web pages, national databases and through personal contact with company
representatives. The subsequent analysis examines the collected data from a resource consumption,
human health and ecological consequences point of view. Using the Life Cycle Impact Assessment
methodology of characterisation, primary data of environmental loads have been converted to a
synoptic set of environmental impacts. For radiation and tailings issues, a more general approach is
used to address the problem. Based on the collected data and the investigated parameters, the
results indicate a presumptive relative disadvantageous result for the uranium mining industry in
terms of health aspects but an apparent favourable relative result in terms of environmental impacts.
Given the prerequisites of this study, it is not feasible to draw any unambiguous conclusions.
Inabilities to do this are mainly related to inadequate data availability from mine sites (especially in
areas concerning tailings management), and difficulties concerned with the relative valuation of
specific performance parameters, in particular radiation issues. Further studies are recommended
within tailings management issues, preferably performed at site, and for studies with a broader
sustainability approach.
iii |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Sammanfattning på svenska
Under de senaste åren har växthuseffekten och den globala uppvärmningen diskuterats flitigt i flera
medier. Detta har lett till ökad efterfrågan på mindre koldioxidintensiva energikällor. I denna debatt
har kärnkraften framförts som ett klimatvänligt alternativ. Frågetecken har dock uppkommit
angående brytningen av den råvara som används som bränsle i kärnreaktorer, uran, och hur stor
miljöpåverkan denna verksamhet egentligen har. Målet med detta examensarbete har varit att
utröna huruvida uranbrytning ur ett miljö- och hälsoperspektiv generellt sett är mer eller mindre
allvarlig än annan mineralbrytning.
Gruvbrytning i allmänhet är förknippat med stora konsekvenser för miljö och natur. Förutom tydliga
sår i jordskorpan märks också utsläpp av många miljö- och hälsofarliga ämnen. Stora problem är
framförallt förknippade med gruvavfallet som för lång framtid kan orsaka försurning och utsläpp av
tungmetaller som finns bundna i malmrester. Olika mineral förekommer i olika former av malmer
vilka i sin tur är förknippade med miljökonsekvenser av varierande allvarlighetsgrad.
För att genomföra studien har femton gruvor valts ut i olika delar av världen vilka har studerats i
detalj för att med hjälp av denna information försöka dra generella slutsatser. Studien har begränsats
genom att enbart undersöka gruvor som bryter någon utav mineralen koppar, guld eller uran.
Resultaten tyder på att urangruvor kan medföra allvarligare konsekvenser ur ett hälsoperspektiv men
att övriga gruvor påvisar skadligare konsekvenser ur ett miljöperspektiv. Studien har dock inte haft
möjligheten att täcka in samtliga områden som kan vara av intresse för att med säkerhet kunna ställa
ett entydigt svar på frågan om uranbrytning generellt är mer eller mindre allvarlig än annan
mineralbrytning ur ett miljö- och hälsoperspektiv. Vidare studier anses nödvändiga för att studera
viss problematik mer djupgående för att på så vis kunna dra mer definitiva slutsatser.
v |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Acknowledgements
To explore, mine and extract the truth about the mining industry is a vast and challenging task of
which we have managed to scratch but the surface in some areas and in other areas, we have been
able to dig a little deeper. A number of people have contributed to this work in one way or another
and deserve a special acknowledgement:
Lasse Kyläkorpi (Vattenfall AB Generation Nordic), our main supervisor of this thesis. Thank you for
all the support, for all discussions, for trusting us with this task and for making us feel welcome to
PQ.
Karin Andersson (Environmental Systems Analysis, Chalmers University of Technology), our
supervisor and examiner at Chalmers. Thank you for accepting this subject, for support and interest
in this project and for guiding us in the right direction.
Caroline Setterwall (Vattenfall AB Generation Nordic) for standing in when needed, for travel
company and interesting discussions.
Martin Luthander and Marthin Westman (Vattenfall AB Generation Nordic) for great enthusiasm,
interesting discussions and for doing that little extra to make us feel at home at Vattenfall AB
Generation Nordic.
Claes Lundström and Ali Etemad (Vattenfall AB Generation Nordic) for establishing contacts with
investigated mines and for interesting discussions about the uranium market respectively.
Synnöve Sundell-Bergman (Vattenfall AB Services Nordic) for valuable support concerning
radiological issues.
Lars Aili (LKAB) for showing us the Kiruna mine with great passion.
Olle Baltzari and Iiris Takala (Boliden Mineral AB) for showing us the Aitik mine and for taking time to
discuss environmental issues in the mining industry respectively.
Bengt Lilljha (Ranstad Mineral AB) for showing us the former Swedish uranium mine Ranstadsverket.
Bert Allard (Örebro University) for valuable discussions about the tailings issue.
The following people working at the investigated mines have answered questions and provided us
with invaluable data, this work could not have been performed without them and they deserve our
sincere gratitude (in alphabetical order): Frances Anderson (Rössing Uranium Ltd), Brent Berg
(Cameco Corporation), Claire Burke (Energy Resources of Australia), Louise Crogan (KCGM), Lachlan
Crow (BHP Billiton), Tony Da Cruz (AngloGold Ashanti Limited), Frank Harris (BHP Billiton), Per Jander
(Cameco Inc.), Tom Kung (BHP Billiton), Dean Roddis (Rio Tinto), Sue Sara (Xstrata Plc), Patty Simpson
(AREVA Resources Canada Inc.), Philipa Varris (Energy Resources of Australia), Peter Wollenberg
(AREVA Resources Canada Inc.) and Peter Woods (Heathgate Resources Pty Ltd).
Thank you also to people at the division of Environmental Systems Analysis at Chalmers University of
Technology and at the PQ and PN divisions at Vattenfall AB Generation Nordic that have welcomed
us to their workplaces.
Financial support from Vattenfall AB Generation Nordic is also acknowledged.
A final thank you from Johan to Katarina for being there when I am not.
vi |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
1 Introduction
In an era of increased attention to global warming, less carbon intensive energy sources are
requested to a large extent. In the light of this, the nuclear industry has witnessed a minor
renaissance. Consequently, increased demand of uranium has lead to increased prices of this
commodity and resources that were previously not profitable to mine have attracted new interest
(WNA: The Nuclear Renaissance 2007).
However, the extraction of uranium constitutes the part of the nuclear fuel cycle with the largest
environmental impact (Vattenfall AB Generation Nordic 2007). For obvious reasons, the mining of
uranium has therefore attracted considerable advertency in terms of environmental concerns. The
use of mineral resources, and thus also mining, is per definition unsustainable since it is based on the
production of non-renewable resources from finite deposits. Furthermore, mining potentially causes
severe environmental consequences and some of the worst industrial disasters are mining related
(Aswathanarayana 2003).
Having said this, without challenging the environmental inconveniences connected to mining, and
assuming that there is a presumptive future demand of uranium that has to be replenished by the
extraction from geological resources; does uranium mining differ in terms of environmental impacts
compared to the mining of other mineral commodities?
This is a question of particular interest in regions with rich holdings of mineral resources in general
and of uranium resources in particular. General opinions about uranium mining are commonly
strongly negative (e.g. Larsson 2008) while mining of other commodities often is met with great
enthusiasm (e.g. Lövgren 2008). Is this a rational behaviour, when considering only the extraction of
the different minerals respectively?
This project will try to shed some light to this issue by investigating environmental impacts and
health aspects in the mining industry in a comparative study of the mining and extraction of uranium,
copper and gold.
1 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
2 Project Outline
The structure of this report pursues the following outline. Initially, the problem and how it is
addressed is defined in general, and in time and space in particular in order to present the
prerequisites and limitations of the study. This is accounted for in the “Goal and Scope”.
Thereafter the reader is introduced to some background information about the mining industry and
the different processes it comprises. Furthermore, the general problems of the industry in terms of
the environment and human health are described. Special attention is given to the areas of radiation
and tailings. This is described in the parts: “Mineral Extraction” and “Health and Environment in the
Mining Industry”.
The subsequent parts of the report, “Investigated Minerals”, “Investigated Mines” and “Companies
Operating the Mines”, describe the investigated minerals: uranium, copper and gold, the specific
mines that are included in the study and the companies responsible for their operations.
After that, the methodology is explained and the results from the data inventory are presented and
analysed in terms of environmental impacts and health aspects. This is performed in “Methodology”
“Results and Analysis”.
Finally, conclusions from the study are presented and explained together with discussions dealing
with the validity of the findings and suggestions to further research.
The end of the report contains a complete list of referred literature and data sources, and detailed
raw data from the data collection, available in appendix.
2 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
3 Goal and Scope
The following part will describe the background of the project and define the problem and its system
boundaries, limitations and how it will be addressed.
3.1 Background
The authors’ interest in this particular issue originated from a broadcast in November 2007 called
“Uppdrag granskning” (“Mission to Investigate” freely translated) by the public Swedish television
dealing with environmental issues of the Australian uranium mining industry. In accordance with the
supervisor at Vattenfall AB, Lasse Kyläkorpi, and the examiner at Chalmers University of Technology,
Karin Andersson, the current problem definition was developed.
3.2 Problem Definition
The general question that is to be addressed in this study is whether or not uranium mining in
average has a better or worse environmental performance or health impact than the mining of other
commodities based on a given set of parameters.
3.3 Aim and Goal
The aim of the study is to compare health and environmental impacts from uranium mining with
mining of other minerals. The goal is to determine the relative environmental performance of the
mining and processing of the different minerals.
3.4 Method
The study has been performed in the form of an empirical study, which is both descriptive and
inductive, based on primary data from specific mining operations in order to come up with a general
conclusion.
In order to limit the investigation to a synoptic set of data, the mining of uranium will be compared
with the mining of two other mineral commodities.
The minerals chosen for comparison are gold and copper, due to their similarities as well as
differences, in order to obtain a wider picture. The market price for both uranium and gold is
relatively high and hence the minerals can be recovered at low mineral grades. Copper, on the other
hand, is priced significantly lower and is generally broken at ore grades higher than uranium and
gold. In addition to this, copper is a well-known metal of high utility in the society.
The investigated mining sites were chosen based on their geographical location. Initially three
different regions with active uranium mines were chosen. Based on this, copper and gold mines were
selected in the same geographical region in order to exclude contingent regional variations in both
geological prerequisites, and political and legal circumstances. Furthermore, the aim, which however
proved to be difficult to fulfil to a large extent, was to find mines operated by the same companies in
order to exclude variations based on company management and culture.
This study is mainly based on primary data provided by the corporations responsible for the
operation of the mines or data available in public databases managed by national environmental
authorities. Background information is compiled based on information provided in general literature
in the area of mining and environment and also on information from interest groups and associations
3 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
within the mining industry. Information is completed and compared with findings made in peer
reviewed scientific articles and publications from scientific associations.
The data collected has been compiled and classified into specific impact categories to allow for
comparison.
3.5 Impact Categories
Collected data has been classified into the following impact categories; resource consumption,
human health and ecological consequences and analysed based on the following parameters;
resource consumption, global warming, human toxicity, ecotoxicity, photochemical ozone creation
potential, acidification and eutrophication as suggested by Baumann and Tillman (2004). To allow for
comparison, when applicable, data will be analysed with the aid of characterisation indicators to
reduce the number of parameters.
3.6 System Boundaries
The nature of this study is a cradle to gate analysis investigating the impacts arising in all process
steps from the extraction until the final product that leaves the site. The geographical boundaries are
limited to what occurs at the sites of the mineral leases, which are performed by the companies
conducting the extraction. However, in terms of environmental impacts and health aspects, a global
perspective is used.
Electricity generated off site is included without accounting for the amount of primary energy used
for the electricity production and related conversion losses.
The temporal system boundaries are set to include activities of active mines occurring within the
year of 2007.
However, in terms of wastes generated at site and traces of the product contained in these wastes,
such as waste rock and tailings, these are followed to the grave and analysed over a longer period of
time.
Environmental impacts arising from the production of capital goods such as machinery and buildings
are excluded from this study.
3.7 Functional Unit
The collected data has been expressed and analysed based on three parameters: namely the amount
of ore mined, the amount of product produced, both expressed in tonnes, and the value of the
product produced expressed in million US dollars. Due to the inability to compare the utility of the
different commodities, and also because of differences in the nature of the final product at the
different sites, the most appropriate functional unit was determined to be the value of the product
at the gate.
In most cases, the final product at the gate is calcinated uranium octaoxide (U O ), also known as
3 8
yellow cake, in the case of uranium mining, and gold bullion and copper cathode in the case of gold
and copper mining respectively.
4 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
3.8 Data Quality
Since the data used in this study mostly is provided by the companies operating the mines, there is
little possibility to determine the validity of the data. However most of the information is found in
environmental reports and sustainability reports, which often follow the guidelines set up by the
Global Reporting Initiative (GRI).
The main problem with the data used is assumed to be differences in the way certain values are
calculated and presented and what is included in the data. This uncertainty is further increased by
the large variation of the sites in terms of age, mining methods, geology, climate and other local
prerequisites.
3.9 Limitations
Although this report deals with uranium and parts of the nuclear fuel cycle, it has nothing to do with
nuclear power generation and environmental consequences from that and it should not be used for
such purposes.
The aim is merely to compare the activities conducted at the mining sites and it is important to
emphasize the completely different utility of the different commodities. Impacts from the entire life
cycle of the products may generate a completely different result.
Furthermore, activities of exploration apart from the continuous exploration conducted in active
mines are not included. The majority of the data is based on one defined time period of the life of a
well developed and active mine. Similarly, although the study briefly investigates the environmental
prerequisites of the mine after closure by looking at waste and tailings generated during one year, it
is important to stress the fact that unless the mine is rehabilitated in an appropriate way, it can
continue to be an environmental issue for centuries.
5 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
4 Mineral Extraction
The U.S. Bureau of Mines defined in 1989 a mineral resource as “a concentration of naturally
occurring solid, liquid or gaseous material in or on the earth’s crust in such a form and amount that
economic extraction of a commodity from the concentration is currently or potentially feasible”. The
term “economic extraction” in the definition emphasizes the dynamics of what is a mineral resource.
Ore is defined as “mineral or rock that can be recovered at profit” (Aswathanaryana 2003).
Mining can generally be divided into four phases: exploration, development, extraction and
processing, and decommissioning (Ripley, Redman & Crowder 1996). Exploration is the defining of
the extent and value of the ore (Hartman & Mutmansky 2002). During the development stage, the
deposit is opened for production (exploitation or extraction), i.e. access is gained to the deposit. This
is done by either stripping the overburden to expose the ore near the surface or by excavating
openings as preparation for underground mining for deeper deposits (Hartman & Mutmansky 2002).
This study will look further into the consecutive phase, the extraction and processing of the ore.
The following text will provide an overview of different mining methods and processes involved in
the extraction of the mineral.
4.1 Mining
The choice of extraction method is based on the characteristics of the mineral, safety and
environmental concerns, and technology and economics (Hartman & Mutmansky 2002). The most
traditional methods are surface mining and underground mining. Surface mining can be further
divided in to mechanical excavation methods such as open-pit and open-cast mining and aqueous
methods, of which leaching is the most common. Underground mining methods are usually divided
into three classes: unsupported, supported and caving (Hartman & Mutmansky 2002).
4.1.1 Surface Mining
Surface mining is, as the name implies, mining methods of ore, coal, or stone that are carried out at
the surface with basically no underground exposure of miners (Hartman & Mutmansky 2002). This
group of mining methods is the dominating category worldwide (Hartman & Mutmansky 2002). Of
the global mineral production, 80% is performed by surface mining methods (Younger, Banwart &
Hedin 2002). Surface mining can be divided into mechanical excavation and aqueous excavation. The
mechanical excavation class consists of open pit mining, quarrying and open-cast mining (Hartman &
Mutmansky 2002). Aqueous extraction consists of all methods using water or a liquid solvent to
recover minerals (Hartman & Mutmansky 2002) and can be further divided into the subclasses placer
mining and solution mining methods (Hartman & Mutmansky 2002). In placer mining methods, water
is used to excavate, transport and/or concentrate heavy minerals from alluvial or placer deposits. In
solution mining, minerals that are soluble, fusible or easily recovered in slurry form are extracted,
normally by using water or liquid solvents (Hartman & Mutmansky 2002).
Below, the most common methods for the extraction of copper, gold and uranium are treated more
thoroughly.
4.1.1.1 Open-Pit Mining
Open-pit mining is used when the minerals occur near the surface (Aswathanarayana 2003). A
surface pit is excavated, using one or several horizontal benches. For mining thick deposits, several
6 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
benches are excavated and form the pit walls like an inverted cone. The reason for using benches is
that it enables control of the blast holes as well as the slope of the pit walls (Hartman & Mutmansky
2002).
For small open-pit mines, capital investments and running costs are low compared to underground
mines of equal size. However, large open-pit mines require much more preproduction investments
than comparable underground mines. Nevertheless, the running costs are still low for large open-pit
mines and the mining rate can be extremely high. Furthermore, ores can be mined at a grade that is
not economically feasible using other mining methods (Carr and Herz 1989).
The first step in open-pit mining is to expose the ore-body. This is achieved by stripping away the
overburden from benches and ramps. The overburden is transported to dumps near the pit. Apart
from handling equipment, economical characteristics and depth of the ore body as well as
characteristics of the ore and overburden are examples of factors that decide the design of the pit
(Carr and Herz 1989).
When removing the ore from the pit, holes are drilled into the bench and loaded with ANFO
(ammonium nitrate-fuel oil mixture) explosive. Nitroglycerine-type explosives are also used but in
smaller holes or for the blasting of incompletely broken material. After the blasting, the ore is
transported on ramps or spiral roadways or, in some cases, by beltways or tunnels leading to the
plant or dump site (Carr and Herz 1989).
4.1.1.2 Placer Mining
Placer mining is used for the mining of gravel and sand containing gold, tin, titanium and rare-earth
minerals. Placer mining can be done by dry-land methods or by dredging. Dry-land placer mining is
very similar to shallow open-cut mining and strip-mining, except that gravel banks can be removed
by pressurised water stream (Carr and Herz 1989), undercutting and caving it (Hartman &
Mutmansky 2002).
4.1.1.3 Leaching
Leaching is the chemical extraction of metals or minerals in the deposit or from material already
mined. Bacteriological extraction also exists but is not as common. There are two variations of
leaching, percolation leaching and flooded leaching. If the extraction is done within the confines of a
deposit, it is called in situ leaching, which will be treated separately. Heap leaching is the method
performed on already mined dumps, tailings or slag piles (Hartman & Mutmansky 2002).
4.1.2 Underground Mining
When ore veins are steep or deposits bedded, the costs of removing waste rock makes it impossible
to use surface mining methods. Instead underground mining can be used. Commonly, surface mining
methods are used to a certain depth until it is only economically feasible to continue excavations
with underground mining (Carr and Herz 1989). Underground methods differ by the wall and roof
support, the opening configuration and the direction of the extraction process. The alternative
methods of underground mining are unsupported, supported and caving operations (Hartman and
Mutmansky 2002).
The ore can be accessed by either a horizontal tunnel (adit), a vertical shaft sunk in the ore body or in
solid rock near the ore body. It can also be accessed by a decline, by an inclined shaft or a gently
7 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
inclined access spiral. Rubber-tired trucks can then be used to haul material to the surface (Carr and
Herz, 1989).
Underground mining operations generally follow a procedure of drilling, blasting, mucking (i.e.
removal of broken rock) and the installations of ground support (timber or roof bolts) (Carr and Herz,
1989). Waste rock and broken ore is collected and transferred to different haulage units by air- or
electric operated mechanical loaders, cable guided scraper systems or by mobile conveyors. Electric-,
diesel- or compressed-air-powered locomotives with trains of ore cars are normally used for haulage
to transfer points and the mine portal (Carr and Herz, 1989).
4.1.3 In Situ Leaching and Borehole Mining
In Situ leaching is a method using hydrometallurgy to recover copper, uranium and gold from mineral
ore that is permeable in its natural state or is made permeable by e.g. explosive shattering, block
caving or hydraulic fracturing (Carr and Herz 1989). The mine is composed of well fields that are
established over the ore-body (Hore-Lacy 2004). Instead of removing the ore-body from the ground,
a liquid is pumped through the ore, using the wells, dissolving the minerals (WNA: In Situ Leaching
2008). Initially, submersible pumps extract some of the native groundwater from the host aquifer
and then a leach solution is added and injected into the well field. The pregnant solutions from the
wells are pumped to the surface where the mineral is recovered in the same manner as for other
mining methods (Hore-Lacy 2004).
The ISL process is more or less a reversed mineral ore genesis and it can be done by an alkali or acid
solvent. What to use depends of the geology and the groundwater. If significant amounts of calcium
(as limestone or gypsum) are contained in the ore-body, alkaline (carbonate) leaching must be used.
This is the case for host aquifers in the US (WNA: In Situ Leaching 2008). Otherwise, acid leaching is
generally better and is the method used in Canada and Australia. Therefore, the oxidant and
complexity agents that are used differ in different locations and sites (Hore-Lacy 2004).
Positive aspects with ISL are that there is little surface disturbance and less waste generation but it is
crucial that the ore-body is isolated so that the leaching solutions do not contaminate groundwater
away from the ore-body. However, considering today’s technology, ISL is regarded as a controllable
and environmentally safe extraction method (Hore-Lacy 2004).
4.2 Beneficiation
The process of removing unwanted ore constituents in order to prepare it for subsequent processing
stages is the beneficiation. Beneficiation is carried out in a mill, which is usually located near the
mine site in order to reduce costs of transportation. Beneficiation consists of three stages:
preparation, in which the ore is comminuted by crushing and grinding; concentration, in which the
desired ore mineral is separated from gangue; and finally dewatering of the concentrate.
Concentration can be performed in various ways: gravity separation, magnetic separation and
flotation. Flotation is based on principles of surface chemistry and it uses a wide range of reagents. It
is the most common method used for concentrating base metal sulphide ores (Ripley, Redman &
Crowder 1996).
Crushing is performed in several stages until the ore is reduced to 5-25 mm size range. The following
procedure is the grinding in which the grain size is further reduced. In order to optimise the
8 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
comminution process in terms of resource use and efficiency it is important to keep track of ore
characteristics and adjust the process to suit the ore (Ripley, Redman & Crowder 1996).
The beneficiation stage results in atmospheric emissions mainly stemming from transportation,
crushing and grinding of the ore and from dry separation methods. These emissions can largely be
reduced by using closed process lines. In flotation however, the main emissions are hydrospheric in
nature. Some reagents may be passed on to metallurgical stages as components of the concentrate
while others may be transferred to the tailings in the bulk unless they are recycled (Ripley, Redman &
Crowder 1996).
4.3 Metallurgical Processing and Refining
Metallurgical processing and refining embodies all the treatment of the ore after extraction and
beneficiation. Contrary to earlier process steps, most of this treatment involves changes to the
chemical nature of the mined minerals. The purpose of these processes, known as extractive
metallurgy, is to isolate the metal from its mineral. Extractive metallurgy may be divided into three
groups: pyrometallurgy, using high temperatures to initiate and sustain extractive reactions;
hydrometallurgy, using liquid solvents to leach the desired metal from the ore; and electrometallurgy
in which electrical energy is used to dissociate the metal in aqueous solution. Some metals are
extracted by only one of these methods while others require combinations of two or several
methods (Ripley, Redman & Crowder 1996).
Pyrometallurgy is the oldest form of extractive metallurgy and its purpose is to convert the metal
from its sulphide or oxide form to a form closer to the pure element. It can be divided into four
stages: preparatory treatment, smelting, converting and fire refining. Smelting is the most widely
used pyrometallurgical process. Roasting, one preparatory treatment, burns off 20-50% of the
sulphur content of the concentrate. The subsequent smelting produces a so called matte, which is
approximately 50% pure. Both roasting and smelting release sulphur dioxide as flue gas. During the
conversion, the remainder of the sulphur is oxidised with an excess of air or oxygen, raising the metal
purity to about 98%. Normally, both slag, dust and fumes are recycled into the process feed in order
to reduce loss of valuable materials and to minimise emissions. However, emissions of sulphur
dioxide remain high. The converted metal may undergo fire refining, which further oxidises
impurities reducing them to about a half of previous levels (Ripley, Redman & Crowder 1996, p. 35).
Most of the ores processed by pyrometallurgy are sulphides containing both sulphur and iron, which
are oxidised during the smelting process. Sulphur is emitted to air as sulphur dioxide and iron ends
up in the slag in the form of iron oxide. Other emissions include particulates, which can be formed by
metallic and non-metallic compounds (Ripley, Redman & Crowder 1996).
Contrary to pyrometallurgical processes, hydrometallurgy does not produce sulphur dioxide, which is
an important environmental advantage. Water and aqueous solutions of acids and bases are the
most common hydrometallurgical solvents. The process involves four steps: preparation phase,
solution phase (leaching), solid liquid separation and treatment; and finally concentration and
deposition of the metal bearing solution. Hydrometallurgy consumes large amounts of water and
care must be taken to avoid contamination of effluent water (Ripley, Redman & Crowder 1996).
The use of electricity in the extraction of metals from their ores can be performed either by
electrolysis or by electrothermal methods. Electrothermal methods are a form of pyrometallurgy,
which compared with other pyrometallurgical methods have some advantages such as high
9 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
5 Health and Environment in the Mining Industry
The following text will discuss environmental impacts and health aspects in the mining industry in
general. Special attention will be given to the issues of tailings and radiation.
5.1 Environmental Impacts
Three main types of changes are distinguished as a result of mining: change in the natural
topography which results in restrictions in the possibilities of using the land for other purposes,
changes in the hydrogeological conditions with consequences for both groundwater and surface
water and finally changes in the geotechnical conditions of the rock (Aswathanaryana 2003). The
impact varies with local conditions of the specific site of mining. These changes caused by mining can
give rise to various impacts on the geoenvironment, described below.
5.1.1 Impacts on the Lithosphere
Depending on the type of mining conducted and the site of mining there are several types of impacts
on the lithosphere. The results range from formation of ridges, depressions, pits and subsidence on
the surface as well as underground cavities affecting the stability of the ground. Furthermore, both
the area for mining and the area used for waste dumps, occupy and degrade land that could be used
for e.g. farming and agriculture (Aswathanaryana 2003).
5.1.2 Impacts on the Hydrosphere
Impacts on the hydrosphere resulting from mining include lowering of the groundwater table, mine
water discharge into rivers, seas and lakes, leakage from settling tanks and evaporators that have a
negative effect on the groundwater quality and pumping of water into the ground for the extraction
of a mineral (Aswathanaryana 2003).
Significantly lowered groundwater levels can result in huge surface depressions and drained rivers
and lakes with serious impacts on surrounding agriculture for example. Furthermore, depending on
the chemical composition of the rock, the drained water usually becomes highly acidic with the
resulting capability of taking into solution a variety of toxic and heavy metals (Aswathanaryana
2003).
5.1.3 Impacts on the Atmosphere
Atmospheric emissions during mining occur not only from internal combustion engines in mining
machinery but dust and gases are also released from blasts and rocks and mineral masses. One tonne
of explosives produces about 40-50 m3 nitrogen oxides and huge amounts of dust (Aswathanaryana
2003).
Smelters are commonly used for mineral purification and emissions from these processes include
particulate matter and gases such as sulphur dioxide, carbon monoxide and carbon dioxide. Although
some installations use different kinds of flue gas purifications, these are never completely effective
(Carr & Herz eds. 1989).
11 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
5.1.4 Impacts on the Biosphere
The biosphere is adversely affected by mining mainly by pollution and by degradation of land and
vegetation resulting in loss in biodiversity. Mining can also have impact on local microclimate
(Aswathanaryana 2003).
5.2 Health Aspects
Mining is one of the most hazardous industrial occupations and during the period 1980-89, mining
was the industry with the highest annual number of traumatic fatalities. Health impacts from mining
can be divided into two categories: immediate impacts such as accidents; and accumulative and
progressive impacts such as stress, radiation and pulmonary diseases (Aswathanaryana 2003). In
terms of health hazards, four different types can be distinguished: physical, chemical, biological and
mental hazards:
5.2.1 Physical Hazards
Physical hazards include noise, heat, vibrations, falls and explosions, flooding and various forms of
dust, aerosols and fine particles with resulting fibrogenetic and carcinogenic effects (Aswathanaryana
2003). Ionizing radiation is included in the category of physical hazards.
5.2.2 Chemical Hazards
Chemical hazards arise from chemical pollutants in water, solid wastes and air with the most
common substances being carbon monoxide and dioxide, oxides of sulphur, nitrogen oxides and
fluorine compounds (Aswathanaryana 2003).
5.2.3 Biological Hazards
Biological hazards caused by living organisms such as fungus, bacteria and parasites are more
common among mine workers in developing countries with poor standards of hygiene and
sanitation.
5.2.4 Mental Hazards
Mental hazards involved with mining include claustrophobia, anxiety, tension or irritability involved
with the awareness of the dangerous working site. Fatigue and other disorders linked to shift work
are other potential problems among mine workers (Aswathanaryana 2003).
5.3 Tailings and Waste Rock
One of the most serious problems for the mining industry is the production of mine tailings, which
annually amounts to 18 billion m3. That is more solid waste than generated by any other industry
(Aswathanaryana 2003) and still, this figure is expected to double in the next 20-30 years as ores with
progressively lower grades are being worked (Aswathanaryana 2003).
It is estimated that of all material excavated by mining operations, more than 70% is waste. Surface
mining is the method that generates the most waste. At the beginning of the 21st Century, surface
mining contributed to 80% of the global mineral production but to as much as 99% of all generated
mine waste (Younger, Banwart & Hedin 2002).
Mine waste can be classified as waste rock or spoil and as tailings, also known as finings. Waste rock
can arise during both the extraction and the processing of ore, while tailings are only generated
12 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
through processing. The fundamental difference between the two is that of the grain size. Waste
rock is coarse-grained (1mm-50mm) and tailings can be characterizes as fine grained (< 1mm)
(Younger, Banwart & Hedin 2002). Furthermore, waste rock is normally tipped dry, while tailings are
deposited from flowing water (Younger, Banwart & Hedin 2002).
When waste rock is not used as back-fill or used as bulk-fill in construction projects, it is stored in
waste rock piles, generally formed by loose tipping from wagons or conveyor belts. Revegetation is
currently common practice but until just recently, little has been done to the drainage to minimized
leachate generation (Younger, Banwart & Hedin 2002).
Tailings used to be dumped in the nearest watercourse, until the early 20th Century, when it was
discovered that this was not a sustainable behaviour. The original reason for abandoning this tailings
management system was that the water routes required for shipping the products were obstructed.
Today, tailings dams and tailings dykes constructed from waste rock and tailings material are
generally used as storage facilities (Younger, Banwart & Hedin 2002).
The characteristics of the mine waste can be structured in different regions, where different
processes are dominating. These regions are the source term, the near field and the far field. The
source term is the actual waste dump, the near field deals with the treatment and management of
the waste and the far field concerns different interactions between drainage water and the soil and
water outside the constructed deposit (Höglund & Herbert eds. 2004). The terms are illustrated in
Fig. 1.
Fig. 1: Illustration of the dispersion of acidity and trace elements from the source term to the near field and far
field (Adopted from Höglund & Herbert eds. (2004))
As suggested by Bert Allard1, Professor in Chemistry and Project Manager for the Man-Technology-
Environment (MTM) research centre at Örebro University, six analytical parameters related to the
mine waste and important processes connected to the waste are described further:
1 Bert Allard, Professor in Chemistry, Örebro University, meeting at Örebro University November 3, 2008
13 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
• The ore: source term and water chemistry
• Age of the operations
• Deposits management
• Metals and substances related to the different minerals and ores
• Process chemicals and complexes
• Geography: precipitation and hydrology
The ore is used as a take-off point and the remaining parameters are presented and further on
discussed and compared in relation to the different types of ores of interest.
5.3.1 Source Term – Deposited Waste
The source term refers to the waste dump and characteristics that can be included are the total
amounts of waste and the amounts as well as concentration of contaminants. Factors such as grain
size, porosity and water retention properties, which are influencing the flow of oxygen and water
through the deposits, are also of interest (Höglund & Herbert eds. 2004). In this report, however, the
source term is limited to the amounts of waste produced and a basic composition of the tailings,
excluding the concentrations of different contaminants.
5.3.1.1 Mineral Composition of Tailings
Copper is most commonly contained in copper-iron-sulphide (e.g. chalcopyrite (CuFeS2) and bornite
(Cu5FeS4)) and copper sulphide minerals (e.g. chalcocite (Cu2S)). Also gold is often found associated
with sulphides (Mindat.org: Gold 2008).
Oxides as pitchblende (Mindat.org: Pitchblende 2008) and secondary ores formed from pitchblende
by weathering are the main uranium minerals (Metzler 2004), even if it can occur in sulphide
minerals as well (Ripley, Redman & Crowder 1996).
The mineral composition of tailings is directly linked to the water quality of the tailings discharge.
Reactions between minerals are dependent on tailings compositions and chemical properties of the
water (pH, oxygen dissolved solutes etc) (Höglund & Herbert eds. 2004). When sulphides get in
contact with water, the sulphide minerals are oxidized and sulphuric acid is produced, lowering pH of
the water. This phenomenon, termed Acid Mine Drainage, AMD (Akcil & Koldas 2006), is further
explained in the following section. Oxides are pH neutral1 while carbonates have a buffering capacity
(Höglund & Herbert eds. 2004) and constitute the most efficient minerals when it comes to
neutralising acid from weathering sulphides (Envipro Miljöteknik AB 2006). Silicates are weathered
when pH is lowered. Buffering silicates consume hydrogen ions when pH is low and can therefore
counteract a further pH decrease (Höglund & Herbert eds. 2004).
5.3.1.2 Acid Mine Drainage
Generally, minerals that constitute economically valued ores are mostly stable under the reigning
geological conditions in which they are found. When they are excavated and exposed to the
atmosphere, they become less chemically stable. Sulphide minerals spontaneously dissolve in
contact with water. The release of contaminants from ore and mine waste into hydrological cycles
through chemical weathering makes potentially toxic solutes mobile and thus bio available. Metal
ions and acidity are released by weathering of different sorts of sulphide ores (Younger, Banwart &
Hedin 2002).
14 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Pyrite (FeS ) is one of the most common sulphide minerals and is therefore used as an example of
2
oxidation of sulphide minerals (Akcil & Koldas 2006):
FeS + 7/ O + H O → Fe2+ + 2SO 2- + 2H+
2 2 2 2 4
Oxygen oxidizes the sulphide mineral into dissolved iron, sulphate and hydrogen (Akcil & Koldas
2006). If the resulting sulphuric acid does not react with acid neutralising minerals such as calcite
(Allard & Herbert 2006), it will induce a decrease in pH while total dissolved solids increase (Akcil &
Koldas 2006). The ferrous ion Fe(II+) can be oxidised to ferric ion Fe(III+), which is itself a strong
oxidant. The ferric ion can oxidise the sulphur and sulphide S(II-) in the sulphide minerals to sulphate
SO 2- and will then be reduced again to ferrous ion (Allard & Herbert 2006, p. 22). For these reactions
4
to occur, the pH must be low, between 2.3-3.5 (Akcil & Koldas 2006), otherwise the ferric ion will not
be in solution but instead be precipitated as ferric oxyhydroxide (Allard & Herbert 2006).
The following chain of effects caused by AMD is mentioned by Young, Mulligan and Fukue (2007):
• Severe health threats to aquatic species, native habitat and plant life.
• Groundwater and drinking water pollution.
• Decline of soil quality.
• Release of heavy metals that are otherwise contained by soil.
Although the sulphide minerals only constitute a small fraction of the broken ore, it can create major
environmental degradation lasting for decades or even centuries after mine closure (Younger,
Banwart & Hedin 2002).
Several factors determine the extent of the acidic leachate generation at a mine site. These factors
can be: the type and concentration of sulphide mineral in the host ore; spent ore and leach piles;
type of host rock; the availability of oxygen; hydrogeology at the site; pH of the water in the system
and finally; the presence (or absence) of bacteria (Young, Mulligan & Fukue 2007).
5.3.1.3 Age of Tailings
A prerequisite for the reaction described in the last section, where ferric ions acts as oxidants,
causing a continued acidification, is the presence of iron in the tailings deposition. Today, the ore is
processed to the extent that it is virtually free from iron. The same does not go for older process
waste, unfortunately. Even though old tailings might have been covered to shield them from contact
with oxygen, significant amounts of oxidised iron are most likely present in the waste dumps2. If the
pH drops, the ferric iron can dissolve and oxidise more of the sulphide minerals, hence decrease the
pH. However, if pH is kept high, the ferric ion will not cause any problem but instead absorb many
toxic metals in solution and decelerate their movement (Allard & Herbert 2006).
5.3.1.4 Metals and Substances of Importance
Consequences of iron in sulphide tailings have already been mentioned briefly. Other metals and
substances that that are important to consider when dealing with these issues are heavy metals and
uranium and its radon progenies.
2 Bert Allard, Professor in Chemistry, Örebro University, meeting at Örebro University November 3, 2008
15 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Heavy Metals
According to Young, Mulligan and Fukue (2007), common heavy metals that can be found in the
geoenvironment are results of human activities (e.g. mining). The more notable and interesting
metals for this study are those mostly connected to sulphide ores, and the acid mine drainage
dilemma, namely arsenic (As), cadmium (Cd), copper (Cu), lead (Pd) and zinc (Zn) (Young, Mulligan &
Fukue 2007). Copper needs no further presentation but the rest will be briefly introduced.
Arsenic is mostly found in iron ores and in sulphide form, the most common ores being, among
others, arsenopyrite (FeAsS) and realgar (AsS). Arsenic is a toxic element and most countries have
adopted a regulatory limit of 50 μg/l in groundwater for drinking. Ingestion of arsenic can lead to e.g.
death from hypertensive heart disease or skin diseases, such as keratosis and hyperpigmentation
(Young, Mulligan & Fukue 2007).
Cadmium can be found as greenocktite (cadmium sulphate CdS) and otavite (cadmium carbonate,
CdCO3) and is often associated with zinc, lead and copper in sulphide form (Young, Mulligan & Fukue
2007). Accumulation of orally ingested cadmium in the liver and kidney can cause organ distress and
the threshold value for drinking water, given by the US Environmental Protection Agency is 5 ppb
(Young, Mulligan & Fukue 2007).
Lead is in the nature found in sulphide minerals (galena, PbS and anglesite, PbSO ), oxide minerals
4
(minium, PbS) and carbonate minerals (cerrusite, PbCO ) (Young, Mulligan & Fukue 2007). The
3
central nervous system, kidneys and the reproduction systems can be affected by inhalation and
ingestion of lead.
Zinc is mostly found as zinc compounds, in combination with oxides, sulphides and carbonates, of
which the sulphide form is probably the most common one. Zinc is often associated with cadmium,
even if zinc is not as toxic (Young, Mulligan & Fukue 2007).
Uranium and Radon Progenies
Mining and milling of uranium can potentially result in residuals like waste rock and tailings, liquid
pollutants and particulate and gaseous emissions to the atmosphere, just as in the case for other
mining operations. However, uranium mining is significantly different in terms of health and
environmental prerequisites since the mined rock and the resulting mine waste can contain
radioactive materials. Radionuclides from the uranium decay chain such as radium, thorium and their
decay products; 210Pb, 210Po and 230Th are some of the most potentially harmful substances
associated with uranium mining. Furthermore, radon or rather the presence of its radon progenies
which are radioactive solids that can be ingested or inhaled are other sources of concern in the
uranium mining industry. These kinds of issues are mainly occupational hazard for people working at
the uranium mine and mill and for people, animals and plants living in the vicinities of mines, mills,
waste rock or tailings (Ripley, Redman & Crowder 1996).
5.3.2 Near Field – Prevention and Control Methods
The near field includes all the aspects of the wastes deposits and the engineered barriers and control
measures (Höglund & Herbert eds. 2004). Many aspects have to be considered for successful
management and remediation of tailings.
16 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Waste deposits management has primarily been focused at lasting re-vegetation of spoil and the
prevention of the release of polluting leachates from the spoil materials (Younger, Banwart & Hedin
2002). The type of waste in the deposit determines what issues to address and how they should be
dealt with.
5.3.2.1 Managing Acid Mine Drainage
The main process driving chemical changes in waste deposits from sulphide ore processing is the
intrusion of oxygen (Moreno & Neretnieks 2006). By decreasing the oxidation caused by oxygen, the
potential environmental loads are decreased as well and the soil and ground have a better chance of
buffering and neutralising the effects of the recipients. This can be done by either covering the
tailings with soil or water, two techniques that are considered the best available by EU. Which one of
the two that is the most appropriate method depends on site-specific characteristics (Höglund &
Herbert eds. 2004).
Soil covers constrain the oxygen flux in the tailings and can therefore prevent weathering of
sulphides. However, for succeeding with the method, appropriate soil qualities for the cover are
needed, enabling a maintained water balance for a long period of time. The soil cover should be able
to sustain high water saturation. Beside particle sizes in the soil, the degree of compaction is an
important factor, not only for the hydraulic conductivity of the sealing layer but additionally for
preventing roots from penetrating the cover, creating macro pores that can enhance oxygen
diffusion.
Finally, the soil cover also has to be protected from erosion by heavy rainfalls and snowmelt
(Höglund & Herbert eds. 2004) by runoffs, leading the large amounts of water from the top of the
impoundment. The transportation and control of groundwater is another critical issue (Höglund &
Herbert eds. 2004).
The oxygen diffusion rate in saturated tailings water is about 10000 times slower than in air. Hence,
the second technique, water cover, can be useful against oxygen intrusion. For this method to work,
the water must be stagnant and not allowed to be mixed. As well, freshwater containing oxygen
needs to be shielded from the tailings (Höglund & Herbert eds. 2004), including groundwater.
5.3.2.2 Uranium Tailings Management
About 85% of the radioactive material in the broken ore is discharged with the mill-tailings slurry and
the major risk of losses of solid materials from these wastes is associated with wind and water
erosion. For these reasons, a proper containment of tailings is the most important aspect in terms of
long term reduction of radiological and chemical pollution from uranium mining and some claim that
continuous improvement in these issues has resulted in uranium tailings being better managed than
those of other mining sectors. Open pit mines generally produce more particulate emissions than
underground operations due to their exposure to wind and rain (Ripley, Redman & Crowder 1996).
To deal with radon and radioactivity, uranium tailings are often covered by water. During
decommissioning, a couple of metres of clay and top soil are normally used to cover the waste to
reduce the radiation levels (WNA: Environmental Aspects of Uranium Mining 2008). Re-vegetation is
often recommended, in order to prevent erosion. However, potential transfers of radioactive
material might be caused by the vegetation and have to be considered carefully. Radioactive material
can be brought to the ground via the roots of vegetation and made available for feeding animals.
17 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Plants can also accumulate airborne radioactive materials that consequently become available for
herbivores. Finally, radon that is sufficient soluble to move up the plant with the transpiration
stream, can be released from the leaves of the plants. Plants with small leaf surfaces and shallow
root systems that are not palatable to animals, should be used when re-vegetating uranium waste
deposits (Ripley, Redman & Crowder 1996).
5.3.2.3 Process Chemicals and Transporters of Heavy Metals
There are a number of chemicals and processes that contribute to the transportation of toxic metals
and substances connected to tailings and waste management. Sulphuric acid is a commonly used
leaching chemical, used in e.g. uranium mining (Hore-Lacy 2004) but also when processing copper in
oxide ores (Davenport et al. 2002). Acidic leaching chemicals can be neutralized and is therefore not
discussed further (Hartman & Mutmansky, 2002). When extracting copper from the grinded and
milled ore, this is often done by flotation and thus by using a number of flotation chemicals. It is
reasonable to assume that the flotation chemicals affect the effluents in some way. The chemicals
used differ and what is used in the specific operations is unknown and hence, so is also the extent to
which they affect the effluents and surrounding.
Cyanide
Besides acid mine drainage of gold tailings connected to sulphide ores, the most toxic remainder
from lode gold mining is cyanide (Ripley, Redman & Crowder 1996).
The most problematic form of cyanide is free cyanide that includes the cyanide anion and hydrogen
cyanide, HCN (ICMI: Environmental and Health Effects of Cyanide 2006). Cyanide is very reactive and
forms simple salts with alkali cations and ionic complexes with metal cations. The strength of the
ionic complexes can vary. Weak-acid-dissociables (WAD) are weak or moderately stable complexes of
e.g. cadmium, copper and zinc, that are less toxic than free cyanide (ICMI: Cyanide Chemistry 2006).
Nevertheless, under varying environmental conditions such as ingestion and absorption by wildlife,
the dissolutions can release free cyanide as well as potentially toxic metal-ions. Cyanide can also
form very stable complexes with gold, lead, cobalt and mercury (Donato et al 2006).
The different types of cyanide commonly present in the tailings dams are free cyanide, WAD cyanide
and total cyanides. Free cyanide is not persistent in the tailings and will decay through different
processes to less toxic chemicals. The WAD cyanide on the other hand is persistent and can release
cyanide when for example being ingested as mentioned above (Donato et al 2006).
Cyanide salts and complexes stability is pH dependent and given that the bioavailable toxicity of
cyanide depends on the types of complexes present, the environmental impacts of cyanide
contamination can vary (ICMI: Cyanide Chemistry 2006). Still, cyano-compounds that liberate free
cyanide ions are highly toxic to almost all forms of flora and fauna (Souren 2000 see Donato 2007).
Cyanide is a fast acting poison. Poisoning of biota may occur by inhalation of dust, ingestion and
absorption through mucous membranes or through direct contact with skin. Cyanide binds to
enzymes and proteins containing iron, copper or sulphur, that are needed for the transportation of
oxygen to cells. This leads to cell death. In higher animals, the most affected organ is the brain,
resulting in coma and collapse of the respiratory and cardiovascular system (Donato et al 2007). In a
tailings environment, the species with the highest risks of contamination are those that drink, feed or
roost on habitats around cyanide bearing tailings systems (Donato et al 2007). This aspect is
18 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
especially serious in dry climates where animals are attracted to the water in tailings impoundments
(Balkau 1998).
Cyanide is a non-selective solvent and can therefore bring several hazardous substances in solution
(Ripley, Redman and Crowder 1996). When forming complexes with other metals from the ore, e.g.
copper and iron, cyanide is not available for dissolving the gold (ICMI: Cyanide Chemistry 2006). Gold
recovery from sulphide minerals requires additional amounts of cyanide to compensate for
decreased gold recovery rates due to the fact that the cyanide leach the sulphide minerals before the
gold (ICMI: Use of Cyanide in the Gold Industry 2006). In an environmental context the copper
content is of great interest as well, since copper-cyanide complexes are toxic to birds and bats
(Donato et al 2006).
Organic Material and Carbonates
Natural organic matter (NOM) is an important part of surface water biogeochemistry as well as of
reactions in the soil. Not only can NOM buffer pH in the range of weak acids, but it is also able to
form complexes with metal ions in solution, decreasing their transportation ability. Humic substances
are important with respect to these characteristics of NOM (Höglund & Herbert eds. 2004). However,
the organic content of tailings is in general very low (Höglund & Herbert eds. 2004).
Carbonates are the most efficient minerals when it comes to neutralising acid from weathering
sulphides (Envipro Miljöteknik AB), and carbonate ions are also able to slow down the transportation
of heavy metals from tailings water by the precipitation of carbonates. If the concentrations of metal
ions and carbonate ions are sufficiently high, these will react and form carbonates. However, to
maintain surplus concentrations of carbonate ions, pH levels cannot be too low (Berggren et al 2006).
5.3.3 Far Field
This perspective includes the area receiving the drainage water from the tailings deposit and is of
great importance to understand the consequences and environmental impacts caused by the tailings
(Höglund & Herbert eds. 2004).
5.3.3.1 Geography
Water and oxygen is of great importance for acid mine drainage. Flow direction, residence time and
groundwater levels and properties of surface waters are all of interest for the estimations of the
transport and spreading of contaminants (Allard &Herbert 2006).Therefore, the climate, landscape
scene and hydrological properties of the soil are examples of crucial aspects. A dry area will not have
the same problems with acid mine drainage as an area with much precipitation due to lower rates of
weathering of sulphide minerals. Similarly, water access differs between an operation situated on the
top of a hill and a site in a valley.
5.4 Rehabilitation of Mined Land
Mining operations affect the land in several different respects. It deforms the landscape and can for
example cause landslides, pollution of water and soil, lowering of groundwater levels etc. Mining
companies are today required to include rehabilitation plans and funds in the decommissioning
procedures of a mine. Restoration aims at restoring the productivity of the affected land area,
harmonize the landscape and reduce the risk of further land degradation. It involves landscaping and
revegetating of spoil heaps, pits, disused industrial areas etc. What rehabilitation procedure to use
19 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
depends on factors such as mining methods, climate, soil and hydrology as well as the intended
reason for the restoration (Aswathanarayana 2003). Despite all the various aspects influencing the
restoration activities, Aswathanarayana (2003) lists a number of elements that are common for
rehabilitation procedures:
• Removal and retention from topsoil that can be spread in the area of
rehabilitation.
• Reshaping the degraded areas and waste dumps for them to be stable, well
drained, and suitably landscaped for the desired long-term use.
• Minimizing the potentiality of wind and water erosion.
• Deep ripping of the compacted surface.
• Revegetation in order to control erosion and facilitate the development of a stable
ecosystem compatible with the projected long-term use.
In general, open pits are harmonized with its surroundings or filled up for new land use options. In
the past, pits have often been used for waste storage, but this is not common practice anymore due
to problems with water leachates and succeeding risk of groundwater contamination. Instead, pits
can be filled with water, preventing the mineral from being exposed to oxygen and by that avoiding
chemical weathering (see chapter 5.3.2.1 Managing acid mine drainage). The resulting artificial lake
can additionally be an asset for the near area. Another alternative is to seal the pit against seepage
and in that way reduce the acid mine drainage problems (Fukue et al 2007).
Apart from groundwater considerations, the rehabilitation of ISL mines is less dramatic. The wells are
sealed and capped, process facilities moved and evaporation ponds, if existing, are revegetated
(Hore-Lacy 2004).
The combination of economics (rehabilitation expenditures versus income from new vegetation) and
social priorities together with government regulations often decide the ambitions of restorations
(Aswathanarayana 2003).
5.5 Ionizing Radiation
Mining of uranium is unique compared to the mining of other minerals since the mineral extracted
can cause short-term and long-term damage to biological tissues due to its ability to emit ionizing
radiation (Ripley, Redman & Crowder 1996). This one but significant difference in the prerequisites of
the analysed objects deserves to be developed further.
5.5.1 Nature and Types of Ionizing Radiation
Radiation is the process of emitting radiant energy in the form of waves or particles. Ionizing
radiation refers to radiation that has the ability to remove electrons from the material absorbing the
radiant energy, thus producing ions. Ionizing radiation is produced by the nuclear disintegration of a
radioactive material but it can also be formed from other sources. The activity of a radioactive
material is measured by the number of nuclear disintegrations per unit of time expressed in the unit
Becquerel (Bq). One Becquerel is one disintegration per second (Ripley, Redman & Crowder 1996).
The decay of different unstable isotopes gives rise to different types of radiation with varying effects
on biological tissues.
Alpha radiation consists of helium nuclei and because of its relatively large size it can be screened by
thin material such as the skin or a piece of paper. Alpha radiation emitting isotopes are generally
20 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
considered to be a hazard only if ingested or inhaled (Ripley, Redman & Crowder 1996). Radon,
which is a gas produced from the radioactive decay of uranium, has decay products or so called
radon daughters that are alpha-emitters. Due to the gaseous nature of radon, it, or radon daughters
attached to dust particles, can be inhaled and radon daughters can be formed in the body, where
they constitute a significant hazard. Beta radiation consists of fast moving electrons and it can
penetrate a little way into tissues but it is easily shielded by a few millimetres of wood for example. If
kept in appropriate sealed containers, beta-radioactive substances generally pose no harm to
humans. Gamma rays are electromagnetic radiation of high energy requiring more significant
shielding (WNA: Radiation and Nuclear Energy 2007). Shielding by several centimetres of lead,
decimetres of concrete or metres of water is required to block gamma radiation (SSM 2008).
Other types of ionizing radiation include X-rays, cosmic radiation from outer space consisting mostly
of protons, and neutrons, which mainly originates from nuclear fission (WNA: Radiation and Nuclear
Energy 2007).
5.5.2 Doses of Ionizing Radiation
The amount of radiation energy absorbed by biological tissues is measured in Gray (Gy), where 1
Gray represents the deposition of one joule of energy per kilogram of tissue. Since different types of
radiation have different effects in different types of biological tissues, the unit Sievert (Sv) is used to
quantify the total dose received by an individual. This unit takes into account both the amount and
type of radiation and the relative sensitivity to radiation of the tissue receiving the dose. For practical
purposes, doses to humans are mostly measured in millisieverts (mSv) (WNA: Radiation and Nuclear
Energy 2007).
According to recommendations of the International Commission on Radiological Protection, ICRP, the
above background radiation dose limits for application in occupational exposure are set to a
maximum of 100 mSv over a period of five years on the further provision that the average dose
received during one year does not exceed 50 mSv (ICRP 1991).
Worldwide average background radiation dose is estimated to be 2.4 mSv per year and more than
half of this radiation originates from inhalation exposure from radon (The National Academies 2006).
In certain areas in the city of Ramsar in northern Iran, annual background radiation doses amount to
260 mSv (Ghiassi-Nejad et al 2002). Single doses of about 5000 mSv are considered to be fatal for
half of the exposed population and single doses of 10000 mSv are lethal within weeks (WNA:
Radiation and Nuclear Energy 2007).
Occupational radiation doses reported by the investigated objects range from averages below 1 mSv
per year to maximum doses on single individuals of just below 13 mSv per year. Doses below 100
mSv are defined as low-level doses (The National Academies 2006). The total effective dose normally
reported in the mines is the sum of the contributions including gamma, radon progeny and long lived
radioactive dust (AREVA Resources Canada Inc. 2008).
When comparing total average occupational radiation exposures however, workers in the mining
industry with an average effective dose of 2.7 mSv per year receive lower doses than for example air
travel crew (3.0 mSv/year) and above ground workplaces (4.8 mSv/year) (UNSCEAR 2006).
Anxiety about health aspects connected to the mine operations from people residing in the vicinity
of the mines is not uncommon. The mines often conduct monitoring of certain parameters, such as
radionuclide concentrations and radon levels, around the sites. Based on the obtained data, doses to
21 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
the public are calculated. At the investigated sites, doses to the public contributed by the operations
range from 0.003 to 0.1 mSv per year. This means that some levels above background radiation are
below detection limits.
5.5.3 Exposure Pathways of Ionizing Radiation
Doses of ionizing radiation are calculated values based on assessments using models of often very
high complexity. The first step in the approach of assessing doses of ionizing radiation is to estimate
the nature and magnitude of discharges of radioactive material into the environment. The emissions
of radioactive substances result in elevated concentrations of these substances in air, water and soil
and consequently also in terrestrial and aquatic foods (IAEA 2001). Concentrations can be obtained
either by modelling or by measurements (ICRP 2006). Based on this information, the absorbed dose
is calculated for a specified individual or group of individuals. The procedure is outlined in Fig. 2.
Fig. 2: Dose assessment process (adopted from ICRP (2006)).
The estimated doses are calculated for what is called a representative person, which is a hypothetical
construction of the more highly exposed individuals in the population. When considering doses to
the representative person, the following factors are taken into account: relevant pathways of
exposure, spatial distributions of radionuclides in the region assessed, habit data for the specific
population and different dose coefficients for specific age categories (ICRP 2006).
5.5.4 Health Impacts from Ionizing Radiation
The mechanisms behind the occurrence of adverse health effects in individuals exposed to ionizing
radiation are not fully understood. Risk models assessing risks associated with exposure to ionizing
radiation are based on epidemiologic studies of groups of populations that have received doses of
radiation. Survivors of the Hiroshima and Nagasaki atomic bombings are the most thoroughly studied
individuals for the evaluation of health effects of ionizing radiation. Other investigated groups
22 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
include persons that have been exposed because of medical reasons and nuclear workers (The
National Academies 2006).
Ionizing radiation has the capability to change the structure of molecules and DNA within the cells of
living tissues. If these changes cannot be repaired correctly, the affected cells can develop into
potential cancer cells. Scientific evidence suggests that there is a linear relationship between the
absorbed dose and the development of cancers in humans and that there is no threshold dose level
below which the risk of developing cancers is not affected. Consequently, even very low doses of
ionizing radiation result in slightly increased risks of developing cancer. This is the so-called linear-no-
threshold (LNT) model. The BEIR VII report by the National Academies (2006) predicts that there is a
one percent chance of developing cancer from a lifetime dose of 100 mSv and a 42 percent chance of
developing cancer from other causes. When dealing with lower doses, it is expected that one in 1000
individuals would develop cancer from a dose of 10 mSv (The National Academies 2006). Exposure to
an acute dose of 1 Sv is estimated to increase the lifetime risk of death from any form of cancer by
4.3 to 7.2 percent (UNSCEAR 2006). Increased risks of being affected by adverse health effects other
than cancer from radiation exposure have been observed. Particularly cardiovascular diseases have
been observed in individuals exposed to high or medium level doses. Increased risks of
cardiovascular diseases in low-level radiation groups have not been observed (The National
Academies 2006). Furthermore, factors such as gender, age at exposure and time since exposure
influences the effects of ionizing radiation (UNSCEAR 2006).
Concern has been given to the contingent health hazards connected with living in the vicinity of
uranium mines. Bollhöfer et al. (2006) have investigated concentrations of lead isotopes around the
Ranger mine and their results show that the relative contribution from the mine to the airborne lead
levels in the region was 13%. This was estimated to contribute to 40% of the total long-lived alpha
activity in the region and that the dose received by people from this source was approximately 0.002
mSv per year. This is well below the public dose limit of 1 mSv per year (Bollhöfer at al. 2006).
A similar study by Tripathi et al. (2008) around the uranium mining, processing and tailings
management facility at Jaduguda, India also concludes that radiation levels and radionuclide
concentrations within a distance of five kilometres away from the site are marginally higher than
elsewhere in the region. This includes radiation levels in surface water, ground water, soil and air.
23 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
6 Investigated Minerals
The investigated minerals in this study are all metallic minerals from three different categories. Gold
is a precious metal, copper is a non ferrous metal within the sub group base metals and uranium
belongs to the category of radioactive elements (Aswathanaryana 2003).
6.1 Uranium
Uranium is a naturally occurring element that can be found in approximately 50 minerals (Metzler
2004) and traces of uranium can be found in almost all natural materials (Carr and Herz, 1989).
Ore grades are commonly expressed in terms of uranium content or by their equivalent content of
triuranium octoxide, U O (Ripley, Redman & Crowder 1996) and the average crustal abundance is
3 8
between 2 to 4 parts per million (Metzler 2005). Economic deposits have concentrations of about
0.1% (Aswathanaryana 2003) and can only be found in limited areas (Carr and Herz, 1989). Among
the main uranium ores are uranite, also called pitchblende, carnotite, a uranium and vanadium
mineral, and secondary minerals formed by weathering (Metzler 2004).
Uranium consists of three semi stable radioactive isotopes, namely Uranium-238, Uranium-235 and
Uranium-234 (Carr and Herz, 1989). More than 99% of the naturally occurring uranium is 238U and
0.71% of it is 235U (Metzler, 2004). 235U is the only fissionable of these uranium isotopes and for the
uranium to be useful in a commercial nuclear power plant the share of 235U has to be increased. The
uranium is therefore enriched to normally 3-5% 235Uranium (McFarlane 2004). Hence significant
amounts of ore have to be mined and processed to cover the current usage of uranium of
approximately 65000 tU/year (WNA: Supply of Uranium 2008).
6.1.1 World Uranium Mining
More than 50 percent of the world’s uranium comes from Canada (23%), Australia (21%) and
Kazakhstan (16%). In 2007, underground mining made up half of the total uranium production. The
second largest mining method was in situ leaching, comprising 29%. The largest companies are listed
in Tab. 1 below (WNA: World Uranium Mining 2008).
Company tonnes U %
Cameco 7 770 19
Rio Tinto 7 172 17
Areva 6 046 15
KazAtomProm 4 795 12
ARMZ 3 413 8
BHP Billiton 3 388 8
Navoi 2 320 6
Uranium One 784 2
GA/ Heathgate 673 2
Other 4 919 12
Total 41 279 100%
Tab. 1: The seven companies marketing the most of the uranium production (85%) in 2007 (adopted from WNA
(World Uranium Mining 2008)).
24 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
6.1.2 Uranium Mining and Extraction
Generally, uranium mining does not differ from the mining of other minerals, unless the ore is of a
very high uranium grade. Ore bodies close to the surface are mined by open-cut mining and when
the deposit is deeper down, underground-mining methods are used instead (Hore-Lacy 2004).
Deposits lying in groundwater in porous material are generally mined using in situ leaching (Hore-
Lacy 2004).
To separate the uranium from the conventional hard rock mined mineral, the ore is crushed and
grinded, normally under water to avoid dust (SKI 2006) and at some sites, it is performed
underground (Hore-Lacy 2004). The result is a slurry of fine ore particles, mainly UO and UO ,
3 2
floating in water (Hore-Lacy 2004). Just like the process of ISL mining, the slurry is leached with
sulphuric acid or carbonate leaching, with the help of an oxidant like hydroxide peroxide or ferric iron
(Hore-Lacy 2004). In some cases, physical separation methods using the mass differences of the
compounds (e.g. gravity concentration or flotation) can be used in combination with the leaching
process (SKI 2006).
The liquid from the leaching containing the uranium is filtered and the uranium is separated from the
solvent and minerals still present (Hore-Lacy 2004, p. 320). If the leaching of the minerals has been
made with sodium carbonate and the solution is virtually free from other minerals, it is possible to
separate the uranium by precipitating the carbonate solution using sodium hydroxide (SKI 2006). If
sulphuric acid leaching has been used, the separation is instead performed by the use of ion
exchange. The two methods used for ion exchange of uranium solutions are solid ion exchange with
resin and liquid ion exchange using amines in kerosene. The latter, also known as solvent extraction,
is today the predominate method (Hore-Lacy 2004). In solvent extraction, organic molecules in
organic solutions bond to the specific targeted metal. Given that the organic solution and acidic
solution cannot mix, the metal can be separated from the other metals still in the acidic solution (SKI
2006). The solvents are then stripped in a counter current process, where an ammonia sulphate
solution is being used in the same time as gaseous ammonia is added, raising the pH, and yellow
ammonium diuranate is precipitated. The diuranate is dewatered in a thickener, followed by a filter
or centrifuge. Finally the diuranate is roasted in a furnace, producing the uranium oxide concentrate,
also called yellow cake due to its colour (Hore-Lacy 2004).
6.2 Copper
One of the earliest metals used by man was copper. The reasons for this are numerous; it is found in
mineral deposits in many areas around the world, both as a native metal and in minerals easily
smelted to obtain the metal. Furthermore, it is easily worked, it has an attractive colour and a high
resistance to corrosion. Apart from the above mentioned, its high thermal and electrical conductivity
are characteristics that make the metal popular in the industrial era and about 50% of the production
enters the electrical industry (Carr & Herz eds. 1989).
The continental crustal abundance of copper is around 47 ppm (Aswathanaryana 2003) and copper
ores typically contain from 0.5% copper in open pit mines to 1 or 2% copper in underground mines
(Davenport et al. 2002). Copper is mainly found in the crust in the form of copper-iron-sulphide (e.g.
chalcopyrite (CuFeS ) and bornite (Cu FeS )) and copper sulphide minerals (e.g. chalcocite (Cu S)) and
2 5 4 2
also to a lesser extent as oxidised minerals such as carbonates, hydroxyl-silicates and sulphates.
25 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Another source of increasing importance is recycled copper, which currently makes up between 10 to
15% of mine production (Davenport et al. 2002).
In the case of large near surface deposits, copper is extracted from open pit mines. When the
overburden of the ore is impossible to strip, the ore is broken in underground mines (Carr & Herz
eds. 1989). The total world production of copper in 1999 was 13200 ktonnes (Davenport et al. 2002)
and Chile is by far the world’s largest producer. It is estimated that nearly half of the world’s mined
copper comes from the Andes of South America (Davenport et al. 2002).
Depending on the type of broken ore, copper is refined in various ways. The two main extractive
methods for refining copper ores are pyrometallurgical and hydrometallurgical methods. Copper-
iron-sulphide ores make up about 80% of the world production. The ore has low solubility in aqueous
solutions why pyrometallurgical methods have to be used to refine the ore (Davenport et al. 2002).
This purification method involves the following steps:
Comminution (crushing and grinding) of the ore into finer particles (Davenport et al. 2002).
Concentration of the mineral particles by froth flotation. This is a method in which copper minerals
are selectively attached to air bubbles rising through a water suspension containing the finely ground
ore. Selectivity is achieved by adding reagents making the copper minerals hydrophobic. The
resulting floated copper-mineral particles have a Cu concentration of about 30% (Davenport et al.
2002).
The flotation concentrate is consequently smelted in furnaces with oxygen-enriched air to oxidise
sulphur and iron from the concentrate. The product is a copper-enriched molten sulphide phase
called matte with a Cu concentration of 45 to 75%. This is a process generating large amounts of SO
2
in the off gas (Davenport et al. 2002).
The matte is further converted by oxidation in oxygen-enriched air. The heat generated by the
oxidation of iron and sulphur is enough to make the process auto thermal. The product of the
converting process is crude molten “blister” copper with a Cu concentration of 99% (Davenport et al.
2002). In some cases, smelting and converting are performed as one step. A major problem with this
method however, is the high content (12-24%) of rejected oxidised copper in the slag (Davenport et
al. 2002).
Finally, the “blister” copper is electrorefined to produce copper cathodes with less than 20 ppm
undesirable impurities. To allow for electro refining, the blister copper is fire refined to produce
impure copper anodes that are dissolved in an electrolyte containing copper sulphate and sulphuric
acid (Davenport et al. 2002).
For oxide copper minerals and chalcocite ores, copper is refined with hydrometallurgical extraction.
In this process, copper is leached from broken or crushed ore with sulphuric acid. Pure electrolyte is
produced from the leaching solution by solvent extraction and the pure electrolyte is used to
electroplate pure copper cathodes (Davenport et al. 2002).
6.3 Gold
Gold has, since being used as currency from around 1000 BC, played an important role as a valuable
standard used to compare various currencies. Furthermore, apart from its attractive appearance
making it suitable for decorative purposes and jewellery, it has found an increasing use in
26 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
technological applications due to its high malleability, conductivity and resistance to corrosion (Carr
& Herz eds. 1989).
The average crustal abundance of gold is 0.005 g per tonne, which is much lower than for most other
metals, for example silver (0.07 g/t) and copper (50 g/t). Gold occurs mainly in two ionic forms, Au+
and Au3+, which can form a series of complex ions that are important for transporting gold in natural
environments (Carr & Herz eds. 1989). Native gold contains a few percent silver and alloys with
silver, copper and platinoid metals often occur (Carr & Herz eds. 1989). In order to obtain
commercial concentrations in ore however, the average crustal abundance must be upgraded by a
factor of 3000 to 4000. This can occur by several natural processes involving for example gravity,
leaching with natural fluids followed by redeposition in a more concentrated form (Marsden 2006).
Gold can be mined in various mines: primary gold mines, placer mining operations and in base metal
mines (Ripley, Redman & Crowder 1996). About 10% of gold production originates from base metal
mining such as copper (Ripley, Redman & Crowder 1996).
Lode ore deposits are normally mined with underground methods, which are more specific than
open pit operations (Ripley, Redman & Crowder 1996).
Purification of gold from ores involves several stages and methods differ depending on the type of
ore mined (Ripley, Redman & Crowder 1996). There are several different factors influencing the
process selection in gold extraction. The factors can be categorised into six main areas: geological,
mineralogical, metallurgical, environmental, geographical, economic and political. The process in
which the gold ore is purified into pure gold involves the following unit processes: comminution,
classification, solid-liquid separation, ore concentration, oxidative pre-treatment, leaching, solution
purification and concentration, recovery, refining and effluent treatment (Marsden 2006).
Comminution of ore into smaller particles is performed in order to make the ore amenable to
consecutive steps of gold extraction (Marsden 2006).
Classification, allows for selection of desirable particle size in various process steps. This is normally
achieved with the aid of cyclones or screens within grinding circuits (Marsden 2006).
Solid-liquid separation allows for separation between different processes phases in various ways.
Thickeners and filters are examples of the type of equipment used in solid-liquid separation and
especially thickeners also provide valuable retention time for chemical reactions (Marsden 2006).
Ore concentration is performed to upgrade the ore in order to reduce costs and process volumes in
subsequent process steps. The cost savings achieved by treating a smaller volume of material after
ore concentration must be outweighed by the loss of valuable material in the rejected fraction. Ore
sorting, gravity concentration, flotation, amalgamation, coal-gold agglomeration, electrostatic
separation and magnetic separation are different methods of ore concentration (Marsden 2006).
Gravity concentration and other ore concentration methods are used to remove larger gold particles
from the crushed and grinded ore. The product is sent directly to refining whilst the remaining ore,
after being recircled to crushing and grinding and gravity concentrator, is thickened and sent to the
cyanidation process (Ripley, Redman & Crowder 1996). Among the concentration methods,
amalgamation has been widely discussed due to the concerns over the health hazards associated
with the use of mercury (Marsden 2006). Mercury forms an amalgam with any gold and silver
present in the ore. In order to purify the gold from the amalgam, the mercury must be volatilized, a
process in which about 10% of the mercury used is lost to the atmosphere (Ripley, Redman &
27 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Crowder 1996). This has lead to reduced use of amalgamation in industry but it is still used in some
applications because of few suitable alternatives, especially in emerging and/or lesser-developed
countries. Placer ore concentrates are generally best treated with mercury (Marsden 2006).
However, the most important atmospheric emissions from lode gold mining occur when roasting
some types of ores in order to remove sulphur compounds (Ripley, Redman & Crowder 1996).
Certain types of ores may require oxidative pre-treatment before undergoing conventional leaching.
This step oxidizes the minerals in the ore in order to make it amenable to cyanide leaching. This pre-
treatment can either be hydrometallurgical or pyrometallurgical. Worth mentioning among the
hydrometallurgical methods are, pressure oxidation, which is both efficient and beneficial from an
environmental point of view although both capital and operating costs are rather high, and biological
oxidation with naturally occurring bacteria which can be a suitable but also rather slow process for
certain types of ores. Roasting, a pyrometallurgical method has been used successfully on a wide
range of ores for over 100 years. However, large quantities of gaseous effluents containing pollutants
such as sulphur dioxide and arsenic trioxide are emitted in this process, which probably will lead to a
decline in the application of roasting gold ores (Marsden 2006).
The following step in the purification of gold ores is leaching, a process in which the resulting product
is a gold bearing solution. Gold ore is currently exclusively leached with dilute alkaline cyanide
solutions, despite concerns over the toxicity of cyanide (Marsden 2006). The cyanide dissolves the
gold and a number of potentially hazardous substances from the ore, and cyanide is usually the most
important toxic remainder from lode gold mining (Ripley, Redman & Crowder 1996).
The leaching solution is subsequently purified and sometimes also concentrated before the gold is
recovered in reduction processes, either chemically, by zinc precipitation or cementation, or
electrolytically by electrowinning (Marsden 2006).
Refining is the step in which the final product, doré bullion with a precious metal content ranging
from 90 to 99% is obtained by smelting what is obtained in the recovery process (Marsden 2006).
Bullion can be refined further depending on the final use.
Finally, the waste products generated in all phases must be treated in an economic and
environmentally acceptable manner. Treatment of waste products may be conducted both to
detoxify a particular reagent but also to recover valuable elements from the waste stream. Generally,
recovery of valuable constituents in the waste flow is conducted by either recycling all or parts of the
waste flow back into previous processes or by treating the effluent in a dedicated process (Marsden
2006).
28 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
7 Investigated Mines
The following text will present the mines that constitute the core of the study. Their locations are
indicated in Fig. 3 and the operations are summarised in Tab. 2.
Fig. 3: The world map indicating the location of the specific mining operations of this study. 1: McClean Lake,
2: Rabbit Lake, 3: McArthur River/Key Lake, 4: Goldstrike, 5: Rössing, 6: Kennecott Utah Copper, 7:
Navachab, 8: Vaal River, 9: Palabora, 10: KCGM, 12: Olympic Dam, 13: Beverley, 14: Northparkes, 15: Mount
Isa, 16: Ranger.
Mineral Mine Company Country Method
Beverley Mine Heathgate Australia ISL
McArthur River/Key Cameco Canada Underground
Lake
McClean Lake Areva Canada Open pit
Olympic Dam BHP Billiton Australia Underground
Rabbit Lake Cameco Canada Underground
Ranger Mine Rio Tinto - ERA Australia Open pit
Rössing Mine Rio Tinto Namibia Open pit
Goldstrike property Barrick USA Open pit and
Underground
Kalgoorlie Consolidated Newmont Australia Australia Open pit and
Gold Mines Limited, Barrick Gold Underground
Corporation
Navachab AngloGold Ashanti Namibia Open pit
Vaal River AngloGold Ashanti/Rio South Africa Underground
Tinto
Kennecott Utah Copper Rio Tinto USA Open pit
Mount Isa Mine Xstrata Australia Underground
Northparkes Rio Tinto Australia Open pit and
Underground
Palabora Rio Tinto South Africa Open pit
Tab. 2: Table summarising the investigated operations.
29
muinarU
dloG
reppoC |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
7.1 Uranium Mines
The selected mines are all among the top twelve of the most productive uranium mines in the world
(WNA: World Uranium Mining 2008).
7.1.1 The Beverley Mine
The Beverley uranium mine in South Australia was discovered in 1969 (Mindat.org: Beverly
Mine…Australia 2008) and production began in 2001 (Heathgate Resources: Timing 2005). The mine
uses the in situ leaching (ISL) method to extract the uranium, which is contained in three zones at
depths of about 110 and 140 metres. It is expected that lifetime of the mine is 15-30 years
(Heathgate Resources: The Mine 2005). The ore was originally planned to be mined with open pit
operations (Heathgate Resources: Heathgate Resources 2005).
The Beverley Mine is entirely owned by Heathgate Resources Pty. Ltd. According to Heathgate, the
Beverley Mine is the most advanced ISL-mine in the world (Heathgate Resources: Sustainability
2006).
634 tonnes of uranium was produced from the mine in 2007. Looking at world production statistics,
there are two other ISL uranium mines, located in Kazakhstan and Uzbekistan, with higher annual
production (WNA: World Uranium Mining 2008).
The ore is leached from the aquifer, where it is located, with oxygen and a weak acid, which is
pumped through the ore body (Heathgate Resources: Uranium and the Beverly Deposit 2005).
7.1.2 McArthur River and Key Lake
McArthur River is the world’s largest high-grade uranium deposit (Infomine: McArthur River 2006)
and with a contribution of approximately 17% (7199 tonnes of uranium in 2007) of the total global
uranium mining production it is the world’s most productive uranium mine (WNA: World Uranium
Mining 2008). It is located in northern Saskatchewan in Canada (Infomine: McArthur River 2006). The
property is owned to about 70% by Cameco Corporation and to 30% by Areva (Cameco: McArthur
River – Summary 2008).
The resources were discovered in 1988 and mining began in 1999. The ore, which occurs in deposits
at a depth of between 500 and 600m (Infomine: McArthur River 2006) has an average ore grade of
20.7% U O (Cameco: McArthur River – Summary 2008) and is broken by non-entry underground
3 8
mining methods using remote controlled equipment in order to minimize employees’ exposure to
radiation. In some areas, the ore grade exceeds 40% U O . Freezing techniques are used for
3 8
groundwater control (Infomine: McArthur River 2006).
The broken ore is comminuted and processed into slurry in an underground processing circuit. The
slurry is subsequently pumped to the surface where it is thickened and transported 80km by truck in
specially designed shipping containers to Key Lake mill and processing plant (Infomine: McArthur
River 2006). McArthur River is blended down to a feed grade of about 4% U O , primarily for
3 8
radiation protection purposes, with remnant ore which has been stockpiled from the mined-out pits
at Key Lake3. The mined-out pits are today being used as tailings facilities. Key Lake is the largest
uranium mill in the world (Infomine: McArthur River 2006).
3 Brent Berg, Manager, Environmental Leadership, Cameco Corporation, personal communication (e-mail)
2008-09-10
30 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
The climate of the region is cold, sunny and dry, best described as sub-arctic, with temperatures
ranging from +30°C in summer to -30°C in winter (Infomine: McArthur River 2006).
Both McArthur River mine and Key Lake mill are certified according to ISO 14001 since 2003
(Cameco: McArthur River – History 2007).
7.1.3 McClean Lake
McClean Lake is an open pit uranium mine located in northern Saskatchewan, Canada. The resources
were discovered in 1979 and production began in 1999 (Infomine: McCLean Lake 2006).
Areva Resources Canada Inc. owns the McClean Lake mine to 70%. 22.5% is owned by Denison Mines
Inc. and the rest is owned by OURD Canada Co. Ltd (Areva: McClean Lake – quality and productivity
2008).
The property includes several deposits at a depth rarely exceeding 175 metres. Two of them have
been mined out and one is now used as a tailings management facility (Infomine: McCLean Lake
2006).
The production in 2007 amounted to 734 tonnes of uranium (WNA: World Uranium Mining 2008)
and the grade of the reserves, including stockpiles, is 1.7% U O (Areva: McClean Lake – quality and
3 8
productivity 2008).
In 2000, McClean Lake was certified according to ISO 14001 as the first uranium mine with this
certification in North America (Areva: McClean Lake – quality and productivity 2008).
7.1.4 Olympic Dam
Olympic dam is a multi mineral underground mine in South Australia owned by the global mining
company BHP Billiton. The deposit was discovered in 1975 and extraction began in 1988 (Infomine:
Olympic Dam 2007). It is the world’s largest uranium deposit, the world’s fourth largest copper
deposit, the fifth largest gold deposit and it also contains significant amounts of silver (BHP Billiton:
Olympic Dam 2008). Olympic dam, with an annual output in 2007 of 3388 tonnes of uranium,
accounts for 8% of total global uranium production, which makes it the world’s third largest uranium
producer (WNA: World Uranium Mining 2008). A further expansion plan of the mine, which would
double its production capacity, is currently being planned and is to be presented to environmental
authorities (BHP Billiton: About Olympic Dam 2008).
The operations are carried out in Australia’s largest underground mine (Infomine: Olympic Dam
2007). If expansion plans are completed, it would make Olympic Dam the largest mine in the world
and operations will eventually be converted from current underground mining, to an open pit
(Mining-technology.com: Olympic Dam Copper-Uranium Mine 2008).
The ore, which mainly consists of sulphide minerals and pitchblende, has a copper grade of 2.1%,
0.7kg/t U O , 0.8g/t gold and 4.5g/t silver (Infomine: Olympic Dam 2007).
3 8
The copper ore is recovered by copper sulphide flotation from slurry before the concentrate is
smelted and electrorefined to high purity copper. The uranium is recovered after flotation, together
with remaining copper, by leaching with sulphuric acid. The copper and uranium streams are
subsequently separated with solvent extraction methods (Infomine: Olympic Dam 2007).
31 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
The mine is located in an arid environment with no reliable sources of surface water, so water is
piped from artesian wells and daily consumption exceeds 30 million litres. Average daily temperature
range is 20-36°C in January and, 18-24°C in July (Infomine: Olympic Dam 2007).
In 2005, Olympic Dam received a partial environmental certification according to ISO 14001 criteria,
while the operations showed some non-conformances (Infomine: Olympic Dam 2007).
7.1.5 Rabbit Lake
The ore body in Rabbit Lake, located in Saskatchewan in Canada, was discovered in 1968 and
production began in 1975 (Cameco: Rabbit Lake – History 2008). The property consists of one mined
out open pit, today used as a tailings management facility, and several mined out underground ore
bodies as well as one active underground mine (Cameco: Rabbit Lake – Mining and Milling 2008).
Production in 2007 was 1544 tonnes of uranium (WNA: World Uranium Mining 2008). Ore reserves
have an average grade of 1.14% U O (Cameco: Rabbit Lake – Reserves 2007).
3 8
Rabbit Lake is entirely owned and operated by Cameco (Cameco: Rabbit Lake – Summary 2008).
7.1.6 Ranger Mine
The Ranger open pit uranium mine, located in Australia’s Northern Territory, is the world’s second
largest uranium producer accounting for 11%, and 4589 tonnes of uranium, of total world production
(Infomine: Ranger 2006 & WNA: World Uranium Mining 2008). Ranger mine is owned and operated
by Energy Resources of Australia Ltd (ERA), which is owned by Rio Tinto to 68.4% (Infomine: Ranger
2006). The ore bodies were discovered in 1969 and production began in 1981 (ERA: History – Ranger
2006).
The operations are located in the Kakadu National Park World Heritage area, which is subject to
monsoonal rainfall resulting in tropical vegetation. Water management is an important issue due to
the sensitive environments and the large amounts of water received (Infomine: Ranger 2006).
The operations include, apart from mills, crushers and uranium process plants, an electric power
plant and a sulphuric acid plant (Infomine: Ranger 2006), which was decommissioned in January 2008
(ERA: Acid Plant 2006). The Ranger power plant supplies power to the mine and the nearby town of
Jabiru, which was originally built by ERA to house workers at the mine.
ERA achieved environmental certification under ISO 14001 in 2003 (ERA: Environmental Certification
2006).
7.1.7 Rössing
The Rössing Mine is an open pit uranium mine situated in the Namib Desert of Namibia, Africa. The
low grade ore deposit at Rössing is the largest of its kind and the overall grade is 0.034% U O . The
3 8
mine was discovered in 1928 but production did not begin until 1976 (Infomine: Rössing 2007). With
a production of 2583 tonnes of uranium in 2007, representing 6% of global uranium production it is
the fifth largest uranium mine in the world (WNA: World Uranium Mining 2008). The life of the mine
has been extended from planned closure in 2009 to continue until 2021 (Rio Tinto: Rössing’s business
at a glance 2008).
The largest owner of the mine is Rio Tinto with a share of 69%. The rest is owned by the government
of Iran, 15%, the Industrial Development Corporation (IDC) of South Africa, 10%, the Namibian
Government, 3%, and local individual shareholders own the remaining 3%. However, the Namibian
32 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Government has the majority, 51%, of the voting rights (Rio Tinto: Rössing’s business at a glance
2008).
Haul trucks in the mine drive on trolley-assist on the ramps (Infomine: Rössing 2007). Trolley assist is
a system in which AC-current is supplied to the trucks, which are equipped with a pantograph, by
overhead power lines, which potentially can reduce diesel consumption (Ford 2006).
The region of the Namib Desert is very hot during the summer period between November and
February with temperatures exceeding 40°C. Daytime temperatures during the coldest months are
usually above 20°C but with potentially sub-zero temperatures at night. Rainfall and water is scarce
and Rössing uses 27% of total water used in the central Namib area (Infomine: Rössing 2007).
ISO 14001 environmental certification was received in 2001 and it was renewed in 2005 (Infomine:
Rössing 2007).
7.2 Copper mines
The following investigated copper mines are among the worlds largest, located within regions close
to the uranium mines selected.
7.2.1 Kennecott Utah Copper
The open pit in Kennecott Utah Copper, also known as the Bingham Canyon Mine, has been operated
since 1904. It is the worlds largest copper mine (Rio Tinto: Kennecott Utah Copper 2008). It is the
third largest gold producer in the United States and it also extracts significant amounts of silver and
molybdenum as bi-products.
The operations are located approximately 20 miles southwest of Salt Lake City in Utah, USA. The
temperature in the area is –7 to 2°C in January and ranges between 16 and 34 in July (Infomine:
Bingham Canyon 2006).
The orebody in Bingham Canyon is a sulphide mineralization (Infomine: Bingham Canyon 2006). The
ore is mined from the 4 km long and 1,2 km deep pit, known as the largest man made excavation on
earth (Rio Tinto: Kennecott Utah Copper 2008)
Kennecott is certified according to ISO 14001 Environmental Management System (Kennecott Utah
Copper 2008).
7.2.2 Mount Isa Copper Mine
The Mount Isa mine is one of the largest underground mine complexes in the world and includes two
separate operations, that is copper recovery and zinc-lead process streams (Xstrata: Mount Isa Mine
Sustainability Report 2007 2008). The area, Mount Isa in Queensland, Australia, is situated 2200km
northwest of Brisbane. A large area of Queensland is in the tropics and it is usually hot and sunny in
the state (Infomine: Mount Isa 2006).
The copper operations at Mount Isa consist of the two underground mines, a concentrator and a
smelter (Infomine: Mount Isa 2006). The Mount Isa operation started in the mid 1920s, when zinc
and lead were mined. Except brief operations during the Second World War, it was not until 1953
that copper was produced in parallel with the zinc-lead (Xstrata: Mount Isa Mine 2008). The ore in
Mount Isa is expected to cover operations for at least eleven more years, counting from late 2006
(Mining-technology.com: Mount Isa Copper Mine 2008).
33 |
Chalmers University of Technology | Environmental Impacts and Health Aspects in the Mining Industry
Sulphide minerals make up the copper ore body at Mount Isa. The ore, with a copper mineral grade
of 3.4% (Xstrata: Production Data – Xstrata Alloys 2008) is mined in the two underground mines,
crushed, hauled to the surface and separated from waste material through copper concentration.
From the smelter, copper anodes containing 99.7% pure copper are transported to Xstrata
Townsville Copper refinery for further refinement to copper cathodes (Xstrata: Mount Isa Mine
Sustainability Report 2007 2008).
The mill is certified according to the environmental management system ISO 14001 (Xstrata: Mount
Isa Mine Sustainability Report 2007 2008).
7.2.3 Northparkes
Northparkes Mines, producing copper and gold, is a joint venture between Rio Tinto (80%) and the
Sumitomo Group (20%). It is located at Goonumbla, 27 km north/northwest of the town of Parkes in
Central West NSW, Australia. The temperature in the region is normally hot in summertime to
regular frosts in the winter and the average rainfall is of 588mm per year (Infomine: Northparkes
2006).
The approval for mining at the site was granted 1992, even though the ground had been explored
since the early 1970’s. The copper and gold are predominately contained in sulphide mineral. The
copper content is 0.5% to 1.5%, while gold occurs in a grade ranging between 0.3 to 1 grams per
tonne ore (Infomine: Northparkes 2006).
Northparkes consists of two open pits and one underground mine. The mined ore is crushed, milled
and concentrated in on-site facilities. The resulting concentrate contains copper but also gold and
silver which increases the value of the product (Rio Tinto: Northparkes Mines Sustainability Report
2007 2008).
Northparkes owns 6000 hectares of land around the mine. Only 1 630 of these are covered by the
mining lease. The rest is farmed using best practice conservation farming by either a manager
employed by the company or leased to local landowners (Rio Tinto: Northparkes Mines Sustainability
Report 2007 2008).
The Northparkes Mines’ Environmental, Safety and Health Management System (ESHMS) is certified
according to ISO 14001 (Rio Tinto: Northparkes Mines Sustainability Report 2007 2008).
7.2.4 Palabora
Palabora is situated 360 km northeast of Pretoria, in the Northern Province of South Africa (Infomine:
Palabora 2005). The largest shareholders of the company are Rio Tinto (57.7%) and Anglo-America
(28.9%) (Mining-technology.com: Palabora Copper Mine 2008). Palabora is the leading copper
producer in the country but it also mines valuable bi-products such as zirconium chemicals, uranium
oxide, magnetite, nickel sulphate and small quantities of gold, silver and platinum. The copper ore
grade is 0.7% (Infomine: Palabora 2005).
Palabora was an open pit copper mine from 1964 to 2002, when it became an underground mine.
Palabora produces about 80 000 tonnes refined copper per year. Beside copper cathodes for export,
the refinery produces cast rod for the domestic market (Infomine: Palabora 2005).
The mine is located near Kruger National Park, South Africa’s biggest eco-tourism attraction. The
Palabora operations are therefore closely monitored by different shareholders, e.g. the government
34 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.