Anda di halaman 1dari 302

Sulphide deposits-their origin

and processing
Sulphide deposits-their origin
and processing

Editorial Committee: P.M.] . Gray, G.]. Bowyer,

J.E Castle, D.J . Vaughan and N.A. Warner

~) I
n~~ The Institution of
~U~UU Mining and Metallurgy
Published at the office of
~ The Institution of Mining and Metallurgy
44 Portland Place London W1 England
ISBN-13: 978-94-010-6851-2 e-ISBN-13: 978-94-009-0809-3
DOl: 10.1007/978-94-009-0809-3
The Institution of Mining and Metallurgy 1990

Front cover is a false colour backscatter electron image, obta ined from Cameca Camebax electron microprobe, showing colloform copper
sulphides from Cattle Grid orebody, Ml. Gunson, South Australia. Photograph kindly supplied by A.R. Ramsden and D .H. French, CSIRO
Division of Exploration Geoscience, Australia
The sulphide deposits pose a challenge to every technology that the mineral exploitation industry uses. Their structure,
composition and multi-element content could hardly be further from those of the end-products of highly refined single metals
which must be won from them . The location of the deposits is rarely on the surface, close to end-product markets, and the
disposal of the by-products of processing in environmentally acceptable ways is an unavoidable part of processing. The mining
of sulphide depos its is probably more expensive per tonne of product than for any other of the tonnage metals since low unit
cost mining methods cannot be used in deep veins or lodes or, if the sulphides are in disseminated form , the deposits are of
low grade.

Research engineers and scientists are striving to relate, in usable scientific terms, the properties of these natural deposits to
the physical and mechanical means of transforming a deposit into an engineering material.

The quantitative characterization of mineral textures has advanced considerably in recent years with the advent of more
powerful equipment. It is now possible to ident ify all minerals present in any but the very finest of microstructures, measure the
proportions of each present and obtain an accurate and full analysis of all the elements contained in them.

Even with these measurements we are still some way from being able to forecast the reaction of a natural combination of
sulphide and gangue minerals to, for example, the application of mechanical forces such as are used in ore grinding , the rates
of reaction in contact w ith an aqueous solution of well-defined and homogeneous characteristics or the behaviour of discrete
particles of that texture when subjected to gravity, magnetic or surface forces in mineral separation.

This gap in rational understanding is not too comfortable for the process engineer to live with and, even if he has grown
accustomed to it in dealing with mineral feed, it is, nevertheless, possibly still at its widest in dealing with the sulphides.
Production engineers are accustomed to organizing their data, and thus their manufacturing technology, into rigid programs
which can take command of day-to-day operation almost totally. It is a salutary experience for a mineral processing engineer to
observe a modern motor-car or electronic component production line and realize how far he still is from understanding the
basic parameters that still give him some unwelcome surprises - not infrequently at five o'clock on a Sunday morning!

This volume records some of the progress that has been achieved recently in understanding the processes by which sulphide
minerals have been assembled by nature and the way in which the properties of those assemblages influence the metallurgical
processes that have been derived by trial and error on a bulk scale to treat sulphides. There is still far to go.

Great advances in the technologies of medicine, electron ics, telecommunications and data processing have followed from
research into the fundamentals of the most basic molecular, atomic and electron units. Perhaps basic research into the
fundamental forces that have formed the Earth's crust will, in time, provide metallurgists with the right tools for transforming
minerals to metals on a controlled production line basis.

Meanwhile, the 'try-it-and-see' methods on the production or pilot plant and in the laboratory used by metallurgists for the
sulphides are becoming more sophisticated and efficient.

The future for the major non-ferrous metals and all their many associated elements from the sulphide deposits is bright. The
challenge to match the product quality and price demanded by the user is no less than it ever was.

Philip Gray
Technical Editor

July, 1990
Page Page
Foreword v Methods of recovering platinum-group
metals from Stillwater Complex ore
Geology, petrology and mineralogy E.G. Baglin 155

China's sulphide deposits - their

Compositional and textural variations of the occurrence and treatment
major iron and base-metal sulphide minerals Yu Xingyuan, Li Fenglou and Huang Kaiguo 165
James R. Craig and David J. Vaughan

Is flotation the unavoidable way for

Rio Tinto deposits - geology and geological beneficiating metal sulphide ores?
models for their exploration and ore-reserve J. De Cuyper and Ch. Lucion 177
F. Garcia Palomero 17 Concentrate processing and tailings disposal

The massive sulphide deposit of Aznalc6l1ar, Improved model for design of industrial
Spain, Iberian Pyrite Belt: review of geology column flotation circuits in sulphide
and mineralogy applications
loseflna Sierra 37 R.A. Alford 189

Comparison of methods of gold and silver

Precious- and base-metal mineralogy of the extraction from Hellyer pyrite and lead-zinc
Hellyer volcanogenic massive sulphide flotation middlings
deposit, northwest Tasmania - a study by D .W. Bilston et al. 207
electron microprobe
A.R. Ramsden et al. 49 Variables in the shear flocculation of galena
T.V. Subrahmanyam et al. 223
Mineralogy and petrology of the lead-zinc-
copper sulphide ores of the Viburnum Role of chloride hydrometallurgy in
Trend, southeast Missouri, U.S.A., with processing of complex (massive) sulphide
special emphasis on the mineralogy and ores
extraction problems connected with cobalt
D .N . Collins and D .S. Flett 233
and nickel
Richard D. Hagn i 73 Evaluation of the CANMET Ferric Chloride
lead (FCl) process for treatment of complex
base-metal sulphide ores
Ore processing and mineralogy w.j .S. Craigen et al. 255

Principles and practice of sulphide mineral lead production from high-grade galena
flotation concentrates by ferric chloride leaching and
R. Herrera-Urbina et al. 87 molten-salt electrolysis
l.E. Murphy and M.M. Wong 271
Chelating reagents for flotation of sulphide
minerals Mercury production from sulphide
A. Marabini and M . Barbaro 103 concentrates by cupric chloride leaching
and aqueous electrolysis
l.E. Murphy, H.G. Henry and l.A . Eisele 283
Mineralogy of and potential beneficiation
process for the Molai complex sulphide Arsenic fixation and tailings disposal in
orebody, Greece METBA's gold project
M . Grossou-Valta et al. 119 M . Stefanakis and A. Kontopoulos 289
Studies of mineral liberation performance in Acid mine drainage from sulphide ore
sulphide comminution circuits deposits
D.M. Weedon, L]. Napier-Munn and CL, Evans 135 Fiona M . Doyle 301
Geology, petrology and mineralogy
Compositional and textural variations of the major iron and
base-metal sulphide minerals
James R. Craig
D~p~,!ment of Geological Sciences, Virginia Polytechnic Institute and State University, Blacksburg,
Virginia, U.S.A.
David J. Vaughan
Department of Geology, The University of Manchester, Manchester, England

ABSTRACf ore deposits. These minerals range from sulphides in which

the ~rincipal metal extracted is a necessary, and usually
The crystal structures, stoichiometries, electrical and dO~lna.nt, c~nstituent (e.g ., galena or sphalerite) to sul-
magnetic properties, stabilities and mineral textures found in phides In which the most valued metal is a minor to trace
the metal sulphides are briefly reviewed. Eight of the major ~onstituent (e.g., gold in pyrite or arsenopyrite). Sphalerite
iron and base metal sulphide minerals, chosen because of IS an excellent example of a sulphide that serves in both of
their widespread occurrence (pyrite, pyrrhotite), role as the the ways noted above; it is mined for its zinc content but it
major ore mineral of a particular metal (chalcopyrite, spha- is today virtua~ly the only source of cadmium, an eiement
lerite, galena, pentlandite), or importance as a carrier of rare present as a minor to trace component. It also serves as a
or precious metals (arsenopyrite, tetrahedrite) are discussed source of gallium, germanium and indium. It is important to
in greater detail. The crystal structures and physical proper- remember that the ~ulphide minerals, especially pyrite, have
ties of these minerals are discussed, along with phase rela- also ~erved as maJ~r sources of sulphur (and more rarely
tions in the relevant sulphide systems. Particular emphasis selenium and tell unum) as well as metals. A general listing
is placed on the presentation of data on major and minor ?f the major types of sulphide-rich ore deposits, as presented
element compositional variations in these minerals and textu- in Table 2, documents the abundance of these minerals in the
ral features commonly observed in ores containing them, ore~. In addition, .these major sulphide minerals, especially
both of which are of crucial importance in their metallurgical pynte and pyrrhotite, occur as common accessory phases in
processing. a large variety of rock types .

These same minerals, so valuable as metal sources or as

INTRODUCflON hosts to the minerals that contain the metals, are today also
The naturally occurring metal-sulphur compounds, col- recognized as the potential sources of major environmental
lectively referred to as the sulphide ore minerals, serve both hazards, such as acid mine drainage and acid rain. The min-
as actual metal sources and as the hosts for many of the eral of greatest concern in this regard is pyrite also known
world's precious, base and strategic metals. Distinct, as "fool'," becaus~ of its very superficiisimilarity to
named, sulphide species now number in the hundreds and ~old and ItS a~undant WIdespread occurrence. Indeed, pyrite
have been variously classified on the basis of chemistry and IS overwhelmingly the most abundant metal sulphide in the
crystal structure. Despite this large group, most completely crust of the earth. Its content in sulphide ores may range
listed in Fleischer (1987), the majority of sulphide-bearing from only a few pe~cent of th~ sulphid~ mass in Sudbury-
deposits are dominantly composed of one or more of the type and some stratiform massive sulphide deposits, to vir-
small group of major sulphide minerals listed in Table 1. tually 100% of the sulphide minerals in coal beds.
This listing of only eight minerals is admittedly arbitrary, but
these minerals constitute more than 95% (and in many cases . The compositional variations of the major sulphide
99%) of the sulphide mineral volume in most sulphide-type minerals are reasonably well characterized, both as a result
of numerous analyses of natural samples from a wide variety
Table 1. The major iron and base-metal sulfide minerals.

Ideal Principal Elements Derived

Name Formula (* by-product)
pyrite (marcasite) FeS2 Co, Au, S
pyrrhotite Fel _xS Ni, S*
chalcopyrite CuFeS2 Cu
sphalerite ZnS Zn, Cd*, Ge* , Ga* , In*
galena PbS Pb, Ag*, Bi*, Sb*
arsenopyrite FeAsS As,Au
tetrahedrite CU12Sb4S13 Cu, Sb, Ag, As
pentlandite (Fe,Ni)9S g Ni, Co, Pd*

Table 2. Abbreviated listing of the major types of sulfide ore deposits. This
classification is modified and much simplified from that of Cox and
Singer (1987) .

Type Major Minerals" Metals Extracted Examples

Ores related to mafic and ultramafic inttllsions

Sudbury nickel-copper po, pn, py, cpy, viol Ni, Cu, Co, PGM Sudbury, Ontario
Merensky reef platinum po, pn, cpy Ni, Cu, PGM Merensky Reef, S. Af.
1M Reef, Montana

Ores related to felsic inttllsive rocks

Tin and tunsten skarns py, cass, sph, cpy, Sn, W Pine Creek, California
Zinc-lead skarns py, sph, go Zn,Pb Ban Ban , Australia
Copper skarns py,cpy Cu,Au Carr Fork, Utah
Porphyry copperI py, cpy, bn, mbd CU,Mo,Au Bingham Canyon, Utah
molybdenum Climax , Colorado
Polymetallic veins py,cpy,go,sph,ttd Camsell River, NWT

Ores related to marine mafic eXttllsive rocks

Cyprus-type massive py, cpy Cu Cyprus
Besshi-type massive py, cpy, sph, go CU,Pb,Zn Japan

Ores related to subaerial felsic to mafic extrusive rocks

~e-type epithermal py, sph, gn, cpy, ttd, Cu, Pb, Zn, Ag, Au Creede, Colorado
vems asp
Almaden mercury type py, cinn Hg Almaden, Spain

Ores related to marine felsic to mafic extrusive rocks

Kuroko type py, cpy, gn, sph, asp, Cu, Pb, Zn, Ag, Au Japan

Ores in clastic sedimentary rocks

Quartz pebble py, uran, Au Au, U Witwatersrand, S. Af.
conglomerate gold-
Sandstone-hosted lead- py, sph, go Zn, Pb, Cd Laisvall, Sweden
Sedimentary exhalative py , sph, gn, cpy, asp, Cu, Pb, Zn, Au, Ag Sullivan, BC
lead-zinc (Sedex) ttd, po Tynagh, Ireland

Ores in Carbonate rocks

Mississippi Valley type Py, go, sph Zn, Pb, Cd, Ga, Ge SE Missouri

Abbreviations used as follows: po = pyrrhotite, pn = pentlandite, py = pyrite. cpy = chalcopyrite. viol =

vio\arite, cass = cassiterite. sph = sphalerite. wolf = wolframite, go = galena. bn = bornite, mbd = molybdenite,
ttd = tetrahedrite, asp = arsenopyrite, cinn = cinnabar, uran = uraninite.
of deposits, and as a result of systematic laboratory investi- ion probe and the proton probe (PIXE), have yielded new
gations of the phase equilibria. We can here only present a data that appear to give much more accurate measurements of
few relevant phase diagrams; for additional information the the sulphide mineral compositions, especially for minor and
reader is referred to Barton and Skinner (1979) and Vaughan trace elements. Table 3 presents a listing of the maximum
and Craig (1978,1990). Table 3 contains a tabulation of the contents of many elements in the major sulphides considered
maximum concentrations of numerous elements in the com- in this paper. Sources used were limited to those employing
mon sulphide minerals. Analytical data for the common sul- modem analytical techniques that should have largely
phide minerals is abundant but widely scattered and largely avoided contamination by mineral inclusions.
redundant in displaying minor amounts of a variety of ele-
ments. The large number of analyses results from these
minerals being abundant and from the desire of investigators STRUcruRES AND PROPERTIES OF TIlE MAJOR
to ascertain the distribution of valued elements so that they SULPHIDE MINERALS
can be effectively extracted. There have been relatively few
extensive compilations of sulphide mineral compositional Cxystal structures
ranges. The largest (Fleischer, 1955) is now 35 years old
and contains data derived primarily using analytical methods Several of the common sulphide minerals were among
that indiscriminately included elements from mineral inclu- the first materials to be studied by X-ray crystallography,
sions as well as from the mineral being studied. The devel- and since that time the structures of nearly all mineralogically
opment of the electron microprobe, which allows analysis of significant sulphides have been determined. It is possible to
areas as small as a few micrometers, and more recently the categorize the mineral sulphides into a series of groups based

on major structure types, or having key structural features in are also, commonly, other minerals that have structures
cOrJ?11lon, as shown in Table 4 (modified after Vaughan and based on these "parent" structures and that can be thought of
Craig, 1978). In many cases, these are the classic structures as being "derived" from them. The relationship between a
of crystalline solids such as the rocksalt structure of the derivative structure and the parent structure may involve:
galena group (Fig. lA), the sphalerite and wurtzite forms of
ZnS (Fig. IB,C), or the nickel arsenide structure (Fig. lC). (I) Ordered substitution, e.g., the structure of chal-
The disulphides are characterized by the presence of dianion copyrite (CuFeS2) is derived from sphalerite (ZnS)
(S-S, S-As, As-As, etc.) units in the structure; as well as the by alternate replacement of Zn atoms by Cu and Fe
pyrite structure in which FeS6 octahedral units share comers resulting in an enlarged (tetragonal) unit cell (see
along the c-axis direction, there is the marcasite form of Fig. 2A). As also shown in Figure 2A, stannite
FeS2 in which octahedra share edges to form chains of (Cu2FeSnS4) results from further ordered substitu-
linked octahedra along the c-axis (Fig. 10). The structures tion of half of the Fe atoms in CuFeS2 by Sn.
of FeAs2 (loellingite) and FeAsS (arsenopyrite) are variants
of the marcasite structure that have, respectively, shorter or (2) A stuffed derivative, e.g., talnakhite (Cu9Fe8S 16) is
alternate long and short metal-metal distances across the derived from chalcopyrite by the occupation of
shared octahedral edge (see Fig. ID). A few sulphides such additional, normalIy empty cavities in the structure
as molybdenite or covellite (Fig. IF) have layer structures, (Fig.2B).
and a small number exhibit structures best characterized as
containing rings or chains of linked atoms (e.g ., realgar, (3) Ordered omission, e.g., monoclinic pyrrhotite
AsS). A diverse group of sulphides, referred to by Vaughan (Fe7S8) is derived from the NiAs structured FeS by
and Craig (1978) as the metal-excess group, is composed of removal of Fe atoms leaving holes (vacancies) that
an unusual and diverse set of structures well illustrated by are ordered (Fig. 2C).
the important example of the mineral pentlandite
Ni,Fe)9S8, see Fig. 10). (4) Distortion, e.g., the troilite form of FeS is simply a
distortion of the parent NiAs structure form (Fig.
As can be seen from Table 4, in many of these groups a 2C).
number of minerals share the actual structure type, but there
Table 3. Maximum concentratio-ns (hi ppm unless otherwise indicated) of numerous elements in the eight major sulfide minerals
discussed in the text All data are from studies employing techniques such as electron microprobe or PIXE that are both
sensitive and capable of avoiding contamination by mineral inclusions. References for the data are given in parentheses
after the data : Full references are given at the end of text.

Element pyrite pyrrhotite chalcopyrite sphalerite galena arsenopyrite pentlandite tetrahedrite SS

V 32 (11)
Cr 11 (11)
Mn 11.69% (17) 158 (3) 5.7% (I)
Fe essential essential essential 27.6% (17) essential essential 13.6% (29)
Co major ss 415 (3) 9.9% (33) 52 .6% (20) 4.2% (8)
Ni major ss 719 (3) 0.21% (24) 4.3% (33) essential 3.5% (8)
Cu 40 (11) essential 1.3% (26) 2000 (26) essential
Zn 3334 (5) 2570 (4) essential 1.19% (26) 12.7% (8)
Ga 0.16% (14)
Ge 0.14% (14) 1.3% (28)
As 8% (9) essential 30 .1% (8)
Se 644 (3) 180 (3) 4383 (5) 396 (3) 3681 (3) 682 (4) 41.1 % (13)
'h 37(11) -
Mo 7 (10) 0.8% (28)
Ru 80 (4)
Rh 5200 (4) 86 (4)
Pd - 1.42% (4)
Ag 0.12% (5) 1685 (3) 1.62% (19) 308 (5) 3.1% (10) 14.77% (27) 55 .0% (22)
Cd 2.84% (7) 899 (3) 11.9% (23)
In 1085 (5) 10.4% (31) -
Sn 0.41% (5) 2.34% (15) 286 (3) 7 (10) 14% (8)
Sb 900 (24) 7900 (24) 37.1 % (8)
Te 200 (10) 26.4% (18)
Au 110 (30) 1.8 (30) 7.7 (30) 3.4 (30) 1.6% (25) 2.26% (32)
Hg 17(11) 24% (8)
rt 400 (10) 2.6% (28)
Pb 51 (11) 0.38% (26) essential 6.3% (24)
Bi 0.58% (24) 6.2% (10) 19.7% (2)
(1) Basu(1984) (2) Boldryeva (1973) (3) Brill (1989) (4) Cabri (1984) (5) Cabri (1985)
(6) Cabri (1989) (7) Craig (1983) (8) DoeIter (1926) (9) Fleet (1989) (10) Foord (1989)
(11) Fralick (1989) (12) Harris (1984) (13) Johan (1982) (14) JOOan (1988) (15) Kase(1987)
(16) Kieft (1990) (17) Kissin (1986) (18) Kovalenker (1986) (19) Loucks (1988) (20) Misra (1973)
(21) Nikitin (1929) (22) Paar (1978) (23) Pattrlck (1985) (24) Pearson (1988) (25) Picot (1987)
(26) Scheubel (1988) (27) Scott (1973) (28) Spiridonov (1988) (29) Godorikov (1973) (30) Cook (1990)
(31) Burke (1980) (32) Kovalenker (1980) (33) Klemm (1965)

Table 4. Sulfide structural groups.


Pyrite Structure Marcasite Structure Arsenopyrite Structure Loellin2ite Structure
FeS2 pyrite \FeS 2 marcasite FeAsS arsenopyrite FeAs2 loellingite
COS2 cattierite! FeSbS gudmundite CoAS2 safflorite
NiAs2 rammelsbergite
derived by As/S ordered substitution
(Co,Fe)AsS cobaltite
(Ni,Co,Fe)AsS gersdorffite (I)


PbS galena
a-MnS alabandite
Sphalerite Structure .. derived by ordered substitution - stuffed derivatives
B-ZnS sphalerite CuFeS2 chalcopyrite Cu9FegS16 talnakhite
CdS hawleyite CU2FeSnS4 stannite CU9Fe9S16 mooihoekite
Hg(S,Se) metacinnabar CU2ZnSnS4 kesterite CU4FeSSg haycockite
Wurtzite Strucmre .. composite structure derivatives - ?further derivatives
a-ZnS wurtzite CUFe2S3 cubanite CU2Fe2SnS6 hexastannite
CdS greenockite ?AgFe2S3 argentopyrite
derived by ordered substitution
CU3AsS4 enargite


NiAs Structure .. distorted derivatives .. ordered ommission derivatives
NiAs niccolite FeS troilite Fe7Sg monoclinic pyrrhotite
NiSb breithauptite CoAs modderite Fe9SlO, Fell S12 hexagonal
pyrrhotite, etc.?


C03S4 linnaeite
FeNi2S4 violarite
CuC02S4 carrollite
Molybdenite Structure Tetragonal PbO Structure Covellite Structure
MoS2 molybdenite (Fe ,Co,Ni,Cr,Cu)l+xS CUS covellite
WS2 tungstenite mackinawite -Cu3FeS4 idaite
Pentlandite Structure Argentite Structure Chalcocite Structure
(Ni,Fe)9Sg pentlandite Ag2S argentite CU2S chalcocite
C09Sgcobalt pentlandite "" ?derivative
CU1.96S djurleite
Digenite Structure - - -...derived by ordered substitution
CU9SS digenite CU7S4 anilite Nickel Sulfide Structures
NiS millerite
Stibnite Structure Realgar Structure Cinnabar Structure
Sb2S3 stibnite AS4S4 realgar HgS cinnabar
Bi2S3 bismuthinite
In some cases, the relationships involved are more com- amples of ordered omission mentioned above) to regard the
plex, as, for example, in certain of the sulphosalt minerals" mineral sulphide structures as a relatively "rigid" sulphur
where the resulting structure is composite and made up of lattice framework from which metal atoms can be removed,
slabs or units of a parent structure (or structures) arranged in or to which metal atoms may be added.
some ordered fashion (an important example of a sulphosalt
mineral, tetrahedrite, is further discussed below). It is also Stoichiometty
useful in certain cases (such as the stuffed derivatives or ex-
Many metal sulphides show evidence that the elements
that comprise them are not combined in a simple whole
number ratios, i.e., they exhibit non-stoichiometty.
Defined as minerals with a general formula AmTnXp in
which common elements are A:Ag.CuPb; T:As,Sb,Bi; X:S. They In certain cases, the extent of deviation from a simple
contain pyramidal TS3 groups in the structure. ratio is considerable. For example, the pyrrhotites are


- l.7:J
./~ , I? ~ b
""-- ----
r-, .:

(Bl~m and Kroger, 1956). Galena is apparently stable over
a widerange of values of aSz, and at high aSz it has lead
vacancies, whereas at low aSz there are sulphur vacancies.
It has also been suggested that certain "polymorphic pairs"

J r~ L of minerals exhibit this type of non-stoichiometry, and these

reports conflict with the rigid definition of polymorphism.
For e~ample, Scott and Barnes (1972) have suggested that
.......,... -;j ~--.....
v v v b <, ~ ......
wurtzrte, formerly regarded as the high-temperature
a c (> 1020C) polymorph of ZnS, is actually sulphur-deficient
relative to sphalerite. Electrical measurements indicate that
zinc vacancies occur in sphalerite and sulphur vacancies in
wurtzite with a total range in non-stoichiometry of about 1
atomic %. Sphalerite-wurtzite equilibria would, therefore,

be a function of aSZ as well as temperature and pressure.
However, another pair of minerals to which such reasoning
might be applied are pyrite and marcasite, and here Tossell et
al. (1981) have offered an interpretation that is based on the
reaction mechanism by which marcasite is formed
e (commonly in acid solution) as a metastable species relative
., ,., t., '''1 to pyrite. An important consequence of non-stoichiometry is
d the effect that it has on physical (electrical, optical, hardness)
and chemical properties, although the latter have been less
well studied. Preliminary studies (Vaughan et al., 1987)
suggest that rates of surface alteration, that would in turn
have a marked influence on such extraction methods as
flotation, are strongly influenced by the stoichiometry of

Electrical and magnetic properties

As well as exhibiting a richness and diversity in struc-

tural chemistry, the metal sulphides also show a tremendous
range of electrical and magnetic behavior. As Table 5 indi-
cates, whereas such non-transition metal sulphides as spha-
9 lerite and galena are diamagnetic, diamagnetism is also ex-
Figure 1. Crystal structures of the major sulphides: (A) hibited by pyrite and also marcasite (the other FeSz poly-
galena (PbS) structure; (B) sphalerite (ZnS) structure; (C) morph). Substitution of iron for zinc in sphalerite leads to
wurtzite (ZnS) structure; (D) pyrite structure and the linkage paramagnetic behavior, and many transition metal sulphides
of metal-sulphur octahedra along the c-axis direction in (i) show various forms of magnetic ordering at lower tempera-
pyrite FeSz, (ii) marcasite FeSz, (iii) loellingite FeAsz, and tures including antiferromagnetism (e.g., chalcopyrite),
(iv) arsenopyrite FeAsS; (E) niccolite (NiAs) structure; (F) ferromagnetism (e.g., cattierite) and ferrimagnetism (e.g.,
covellite (CuS) structure; (G) cube cluster of tetrahedrally- monoclinic pyrrhotite, although "hexagonal" pyrrhotites are
coordinated metals in the pentlandite, (Ni,Fe>9Sg, structure . antiferromagnetic). Other transition metal sulphides , which
In each case , metals are shown as the smaller or shaded are metallic conductors, exhibit the weak temperature-inde-
spheres. pendent paramagnetism that is known as Pauli paramag-
netism. As well as the metallic conductivity occurring in
many sulphides of diverse magnetic character, numerous
sulphides are semiconductors, and a smaller number (e.g.,
pure sphalerite) are insulators. Among semiconducting sul-
sometimes given the general formula Fel_xS where 0 < x < phides, both intrinsic and impurity (or extrinsic) conduction
0.125 and the varying compositions correspond to a varying mechanisms occur, and conduction via electrons ~ or
concentration of vacancies "left" in sites that would other- via holes ~ is common. Such variations are often a
wise be occupied by Fe atoms . However, in systems like consequence of very minor impurities being present , or of
these, ordering of the vacancies occurs at low temperatures, slight non-stoichiometry (e.g., galena, PbS , exhibiting a
and the result may be a series of stoichiometric phases of total variation in composition of 0.1 atomic %, shows p-type
slightly different compositions. Although Fe7Sg has a conductivity in lead-deficient samples and n-type conductiv-
(monoclinic) superstructure (based on the NiAs cell) result- ity in sulphur-deficient samples).
ing from vacancy ordering (Tokonarni et al., 1972), the
situation in the "intermediate" or "hexagonal" pyrrhotites is Sulphide mineral stability
more complex. Some of these pyrrhotites may represent
ordered phases with clearly defined compositions (F~S 10, Much work has been done in the last 50 years to estab-
Fej j S lZ, etc.), but more complex and partial ordering in lish the stability relations of sulphide minerals in terms of
these systems may occur. One problem is certainly that free variables that include temperature, pressure, composition
energy differences between a series of phases resulting from and the activities of various components. Much of the data,
vacancy ordering would be very small, making a successful whether derived from synthesis experiments or thermochem-
investigation of relationships between synthetic products ical measu:ements, are reviewed by Vaughan and Craig
very difficult or impossible. (1978), Mills (1974), Kostov and Mincheeva-Stefanova
(1~8q and sources therein . Although it is not the primary
Other examples of non-stoichiometry involve deviations objective of the present paper to review these data indeed
from the simple ratio that are so small as to have been that ~ould not be possible in a relatively short contribution
ignored until fairly recently. For example, galena (PbS) certain aspects are worthy of discussion. '
exhibits a range of non-stoichiometry of 0.1 atomic %

Table 5. Magnetic and electric properties of some major sulfide minerals
(after Vaughan and Craig, 1978).

Mineral Species Magnetic Properties Electrical Properties

Sphalerite (ZnS) diamagnetic insulator (Eg-3.7 eV)

"Iron sphalerite" (Zn,Fe)S paramagnetic semiconductor (Eg-O.5 eV for -12 at % Fe)
Galena (PbS) diamagnetic semiconductor (n- and p-type, Eg-OAI eV)
Pyrite (FeS2) diamagnetic semiconductor (n- and p-type, Eg-O.9 eV)
Cattierite (COS2> ferromagnetic (Tc = 110 K) metallic
Chalcopyrite (CuFeS2) antiferromagnetic semiconductor (n-type, Eg-O.5 eV)
(TN = 823 K)
Covellite (CUS) diamagnetic(?) metallic
(Monoclinic) Pyrrhotite ferrimagnetic (T c = 573 K) metallic
Carrollite (CUC02S4) Pauli paramagnetic metallic
Pentlandite (Ni,Fe>9Ss Pauli paramagnetic metallic

Eg = band or "energy" gap, Tc = Curie temperature; TN = N6:1temperature.

... s

o c


() s.
B .

~~ e si



textures produced by exsolution or unmixing, as exemplified

by the Ni-Fe-S system (Fig. 4). Such changes occur be-
cause at elevated temperatures in many sulphide systems
b C",""",n.. there are extensive fields of solid solution that shrink with
falling temperature. In the case of the Ni-Fe-S system, it is
Figure 2. Parent and derivative crystal structures in the the breakdown of the so-called monosulphide solid solution
sulphide minerals: (a) the sphalerite (ZnS) structure with the and segregation of pentlandite according to the reaction
chalcopyrite (CuFeS2) and stannite CU2FeSnS4 structures;
(b) the sphalerite and chalcopyrite unit cells with an octahe- (Fe,Ni)S = Fel-xS + (Fe,Ni)9Sg
dron of metals outlined within which may be an additional monosulphide pyrrhotite pentlandite
metal ion in the minerals talnakhite (Cu9FeSS16), mooi- solid solution
hoekite (Cu9Fe9S 16) and haycockite (CU4FeSSS), the
arrangeinent of additional occupied metal sites being as that is central to understanding the assemblages and textures
shown (also shown are the dimensions of the parent spha- in the sulphide nickel ores. Such processes may lead to
lerite cell in A); (c) the niccolite unit cell of high temperature completely separate grains of the two phases being formed
FeS that has vacancies in place of Fe atoms in monoclinic on exsolution, or the two phases may be intergrown as laths,
pyrrhotite (Fe7Ss) that are ordered as shown in the diagram blebs, etc., often with a clearly-defined crystallographic
that has vacancies represented by squares (and only Fe atom relationship. The crystal structure of the two phases will
layers shown); also shown in a projection onto the basal partly dictate the kind of texture that forms, but other crucial
plane are the distortions that occur in the troilite modification factors are related to the monosulphide solid solution bulk
of FeS. "starting" composition and the cooling history. The impor-
tance of kinetic factors is clear from such studies as have
In terms both of the interpretation of assemblages and been undertaken (e.g., Kelly and Vaughan, 1983), although
textures in the context of understanding the genesis of the relatively little work has been done in this field. The major-
ores and from the point of view of mineral processing, ity of binary, ternary and many quaternary sulphide systems
temperature-composition relations in the subsolidus region have been studied as regards temperature-composition rela-
are of the most value. This is because many sulphide tions, yielding evidence of the maximum stability tempera-
minerals undergo changes in the solid state down to rela- tures of phases, solid solution limits, and coexistence of
tively low temperatures compared with silicates and oxides phases under the equilibrium conditions pertaining in such
(Fig. 3). These changes give rise to a wealth of intergrowth experiments.

Silicates and oxides


E .... pynte,
~ 'arsenopyrite

~ ~~~~PYrrhotite

1:~ eu-Fe sulfides

10- 10- 2 10- 1 1 10 102 uP 10 lOS 106
1.8 1.6 1:4 1.2 1.0
Time to equilibrate (yr) IOJ/T(K I

Figure 3. Relative rates of equilibration of various common Figure 5. Sulphidation curves for several sulphides as a
sulphide minerals as a function of temperature. (After function of temperature.
Barton, 1970.)
The stable phase relations may also be defined in terms
s of the activities of components and, in regard to the sulphide
minerals, it is the activity of sulphur (aSV that is of the most
importance. Thus, a "petrogenetic grid" commonly used to
present sulphide mineral relationships is that in which aS2 is
plotted against temperature (actually log aS2 versus Iff
because sulphidation boundaries then plot as straight lines in
many cases) as shown in Figure 5. Such diagrams show the
progressive development of assemblages through sulphida-
tion from metal to monosulphide to (where appropriate)
Another important type of phase diagram is that in which
the activity of sulphur is plotted against that of oxygen (at
constant temperature). Such aS2-a02 diagrams may be pre-
sented for one or several cations. In Figure 6 a plot of this
type is shown for the iron sulphides and oxides at 2S"C. In
this case, of course, the value of the diagram lies in the in-
terpretation of relationships and textures that may arise in the
B oxidative alteration of sulphide minerals during weathering
or some analogous process.

$ 10 \$
W(',"HT I't:ItCf;NT NICIl('-

Figure 4. (A) The condensed phase relations in the Fe-Ni-S

system at 650C. Note the absence of pentlandite and the
monosulphide s?lid solution (m.s:s) that spans the system
from Fel_xS to NI1_xS. (B) A pornon of the Fe-Ni-S system Figure 6. A log aS2 versus log a02 plot of iron SUlphides
showing the compositional limits of the monosulphide solid and oxides at 25C contoured in terms of log aS02.
solution (mss) in (A) at 600, 500, 400, and 300"C. SULPHIDE ORE TEXlURES: PRIMARY AND
Pentlandite flames, as those shown in Fig . 17, exsolve from SECONDARY DEVELOPMENT
!he mss as the sulphur-poor boundary retreats during cool-
mg. The major sulphide ore minerals occur not only in a wide
variety of deposits (fable 2) but also in an amazing array of
textures. The textures, in effect the shapes and spatial distri-
butions of mineral grains, are the products of the deposi-

tional and post-depositional histories of the minerals and the
deposits in which they occur. The terminology customarily
used to describe textures is sununarized below:

Primary - textures reflecting the form and distribu-

tion of the minerals as they were origi-
nally deposited

Secondary - textures that have formed as the result of

any post-depositional process

Hypogene - referring to the original minerals formed

in a deposit

Supergene - referring to secondary minerals and

textures that have formed as a result of
meteoric weathering .

It is not possible to herein exhaustively present or dis-

cuss the many varieties of textures exhibited by the major
sulphide minerals. We shall attempt instead to review some
of the most common textures and refer the reader to Craig
and Vaughan (1981) and Craig (1990) for additional discus-
sion and illustrations.

Primary textures of the major sulphides range widely',

depending upon the individual mineral phase or phases
present (e.g., pyrite that typically forms crystals, Fig. 7A,
versus pyrrhotite that is generally anhedral or interstitial to
other phases), the environment of deposition (e.g., open
space filling, Fig. 7B , versus replacement of preexisting
phases, Fig . 7C), and chemical reactions (of multiple
depositing solutions and between solutions and preexisting

Secondary textures of these minerals modify or replace

primary textures by processes such as post-depositional
deformation (e.g., brecciation, Fig . 8A, or flow), recrystal-
lization (e.g., to form crystals or polycrystalline aggregates,
Fig . 8B), exsolution (Fig. 9A), or supergene alteration
(e.g., to form alteration rims, Fig . 9B, or to differentially
corrode coexisting phases).

~nalysis of the textures of the major sulphide minerals

provlde~ not only much insight into the origin and history of
a de~slt or occurrence, ~ut also into the exploitability and
P:Otenn~ problems of mineral processing. Thus, the very
flOe-gr~un~ nature of an ore (e.g., McArthur River,
Australia, FIg. lOA) may result in its being very difficult to
econ?mically ~eparate into clean concentrates by standard Figure 7. (A) Euhedral pyrite crystals in a matrix of inter-
flotation techmq.ues. In con~~t, ores that are coarse grained stitial anhedral pyrrhotite (field of view = 1.1 mm). (B)
~s a !esult .of pnmary deposition or metamorphic recrystal-
Euhedral 7 mm sphalerite crystal formed atop a mass of
l~zanon (FIg. lOB) are relatively readily separated by flota-
saddle dolomite crystals in an open vug in the central
non . The.tex~s are ~so critical in determining which ores Tennessee ores. (C) Fractured pyrite crystal undergoing
m~y readily yield precious metals during cyanide leaching
replacement by chalcopyrite as a result of reaction of the
(FIg. I ~A) a~d thos~ ~at ar~ refra~tory (e.g., contain gold pyrite with copper-rich fluids.
as fine inclusions within pynte grams where it remains un-
reacted with cyanide solutions, Fig. liB).

The differences in sulphide mineral structures (both remains a rigid phase retaining primary structure and com-
crys~l and elecl!0nic structures) result in a broad range in position, whe~s p~ases such as pyrrhotite, chalcopyrite,
physical properties such as density, thermal stability, hard- an~ galena ~y YIeldand deform plastically. This differ-
ness, reflectance, electrical conductivity, etc. These varia- ential behavior under stress has been examined for several of
nons ~s~lt i~ signifi~ant diff~rences in the rates and degrees the major sulphide minerals by Kelly and Clark (1975).
of ~u.lhbration dw:ng coohng from original depositional
conditl~ns, and dunng post-.depositional metamorphism or
w~thenng. !bus, among rmxtures of the conunon sulphide
minerals, pynte, arsenopyrite and sphalerite are the most re-
fracto~ a~~ the sulph~~es most likely to retain the evidence In the more detailed discussion that follows the most
of their original de~SltlOn conditions, whereas chalcopyrite, impo~t of the sulphides are considered in turn ~d aspects
gal~na, and ~yrrhotlte most readily adjust their compositions of their s~ctures, chemistries and textural relationships in
o~s considered. A total of eight sulphides are considered in
dunng c~l~ng (Barton and Skinner, 1979). Similarly,
under conditions of moderate dynamic metamorphism, pyrite this way, chosen for their widespread occurrence (pyrite,

Figure 9. (A) Exsolution lamellae of chalcopyrite from
Figure 8. (A) Cataclastic texture of pyrite resulting from bornite (field of view = 1.1 mm) . (B) Veins of covellite
post-depositional deformation of sulphides (field of view = formed on chalcopyrite as a result of supergene alteration
0.6 mm). (B) Recrystallized pyrite exhibiting typical poly- (field of view = 1.1 mm).
granular texture with 120 interstitial angles (field of view =
0.6 mm),

pyrrhotite), role as the major ore mineral of a particular metal

(chalcopyrite, sphalerite, galena, pentlandite), or importance
as a carrier of rare or precious metals (arsenopyrite, tetra-

Pyrite (ideal formula FeS2) is the most abundant of all

sulphide minerals, occurring as a major phase in many sul-
phide ore deposits and as an accessory mine~ in ~y o.ther
ores and rocks. Pyrite has rarely been nuned for Its non
content but has served as an ore of nickel and cobalt, both of
which may substitute for iron in the structure. !he pyrite of
the Central African Copperbelt serves as a major source of
the world's cobalt supply. Before the development of the
Frasch technique for sulphur extraction, pyrite se~ed as. the
major source of sulphur; today, pynte from lbenan mme.s
remains an important sulphur source for Europe. Because it
is nearly ubiquitous and serves as a host for many other
minerals containing valued metals, pyrite has been much
analyzed . The maximum concentrations of many elements
held within pyrite are given in Table 3.

The crystal structure of pyrite is illustrated in Figure 10

and, as already noted, is distinguished by the presence of S-
S dian ion units. Pyrite is a diamagnetic semiconductor
(Table 4) that may exhibit p- or n-type conduction mecha-
nisms. Marcasite (ideal formula also FeS2) has generally
been regarded as the dimorph of pyrite in which ~-S dianion Figure 10. (A) Fine-grain~ , unmetam~rphosed pyrite-
units again occur, and Fe is octahedrally coordinated ~o ~. sphalerite ore from McArthur River, A.uStra!Ia. (B) Coarsely
However, the mode of linkage of FeS6 octahedral umts IS recrystallized, metamorphosed sphalerite-pyrrhotite ore from
different to that found in pyrite (Fig. 10). lorna, Norway. Both samples have been photographed at
the same scale (field of view = 0.6 mm).

Figure 12. Diagrammatic presentation of the structures
of Fe7Sg, FegSIO and FellSl2. These are the 4C, 5C and
6C pyrrhotites.
pyrrhotite or marcasite, leads to the ~eneration of ~trongly
acid waters that have been responsible for considerable
damage to stream and vegetation. The solid phase resulting
from pyrite oxidation commonly leads to the ~evelopment of
goethite (FeO-OH) pseudomorphous after pynte cubes.


Pyrrhotite (often simplistically referred to by the formula

B Fel-xS) is a major component in many important base metal
deposits. It has been mined for its iron content. The sn:uc-
Figure II. (A) Gold grain along pyrite boundary where ture of "pyrrhotite" is based upon t~e hexagona! NIAs
it would likely be dissolved by cyanide leach solutions. (B) (niccolite) structure (Fig. IE) , b~t this st:ru~ture I~ only
A gold grain completely enclosed in pyrite where it would stable at higher temperatures (5'150 C). StOlc~lOmetrlC ~~S
likely be protected from reaction from leaching solutions undergoes distortion at lower temperatures to give ~e troilite
(field of view = 0.6 mm). superstructure modification (see Fig: ~C). A v~~ unportant
feature of the pyrrhotites is the ability to omit tron. ~toms
from the structure, leaving vacancies up to a composlnon of
Pyrite, isostructural with cattierite (COS2) and vaesite -Fe7Sg. At higher temperatures (-100-300 C), these
(NiS2) exhibits temperature-dependent extensive solid solu- vacancies are disordered, but at lower temperatures they
tions toward both of these phases. Generally, however, the undergo ordering so as to result in a series of superstructures
very -low geochemical abundances of cobalt (25 ppm) and based on the parent NiAs-type structure. The b.est known
nickel (75 ppm) relative to that of iron (56,000 ppm) results and characterized of these superstructures IS that of
in the formation of pyrites that contain no more than a few "monoclinic pyrrhotite" (ideal composition Fe7Sg) illustrated
parts per million of cobalt and nickel. Pyrite is an important in Figure 2C. Here, the vacancies are ordered onto alternate
sink for cobalt and nickel and, occasionally, individual iron atom layers parallel to the basal plane and alt~mate ro~s
growth zones have concentrated the cobalt and nickel to form in these layers. The resulting structure has a urut cell w!th
bravoites. Because pyrite is often the dominant sulphide in four times the c-axis dimension of the parent structure With
many ores and must be processed to extract other intimately some collapse around the vacancies causing a monoclinic
intergrown ore minerals, much effort has been directed to distortion. The situation in the "intermediate pyrrhotites" at
determine what valuable elements the pyrite itself may low temperatures is more uncertain, but i~ seems likely. that
contain (see Table 3). Many pyrites have been thought to various other vacancy-ordered superlattices are possible,
contain gold within their lattices, but the highest documented centered on such compositions as FegS 10, Fe lOS 11. and
value is 110 ppm (Cook and Chryssoulis, 1990). Fej j S 12, and with 5C, llC and 6C superstructures as Illus-
trated in Figure 12 and further discussed by ~aughan and
Pyrite occurs in a very broad range of textures depending Craig (1990). An important.aspect of the ch~ffi1~try of these
upon mode of origin, abundance relative to other minerals intermediate pyrrhotites anses from the ~nencs o~ such
and post-depositional history. Although it can form in poly- vacancy ordering processes. Because the differences .m free
crystalline porous colloform bands at low temperatures, energies between various vacancy OI~ered structures ISvery
pyrite is best noted for its very common appearance as c~bes small, a variety of metastable alternative .structur~s and fine-
(Fig. 7A) or pyritohedra. Its great tendency to form into scale intergrowths can be formed on c<?Oling. !!tIS. has led to
euhedral crystals and its very great hardness (greater than all considerable confusion in understanding equ ilibrium phase
other sulphides and many silicates and oxides) results in it relations, although much progress has been made using
typically appearing as cubes, even in matrices that have Transmission Electron Microscopy.
undergone extensive deformation. Under conditions of
thermal metamorphism, polycrystalline masses of pyrite The vacancy-ordered superstructures found in the
recrystallize to typical annealed textures displaying 120' pyrrhotites at low temperatures also have important conse-
interstitial grain boundaries. Despite its hardness and refrac- quences for their ma~netic pro~rties. Whe~eas FeS and the
tory nature, pyrite, like all sulphides, is unstable in the oxi- intermediate pyrrhotites are antiferromagnetically ordered at
dizing conditions of the earth's surface and most surface room temperature, the ordering of vacll!lcies ?n to on~ sub-
waters. The oxidation of pyrite, and sometimes coexisting lattice in monoclinic pyrrhotite results m ferrimagnetism at

room te~perature. This property is of value in geophysical Atomic percent
prosJ>e<:ung for pyrrhotite-containing ore bodies, and in the
separation of monoclinic pyrrhotite by magnetic methods A
during mineral processing.
Although natural pyrrhotites rarely contain more than
trace quantities of other transition metals, the high tempera-
ture hexagonal NiAs form of pyrrhotite exhibits complete
solid solution with NiS (Fig . 4A) and COS, and has been
reported to accept up to -6.8 wt % MnS in solid solution at
9OOC (Skinner and Luce, 1971) Similarly, high tempera-
ture pyrrhotite can hold up to 4 wt % Cu in solid solution
(Yund and Kullerud, 1966), but natural pyrrhotites contain
20 30 4o- -"----ri,....-........---=-=---'
only trace quantities of copper. Natural samples of the Fe- I 50
C'C :" dg l ron - - -
Ni-S system monosulphide solid solution (mss) have been
reported from several localities, especially marine basalts,
where rapid cooling has preserved intermediate mss com- ATOMIC "I. 2~C

positions. Slow cooling apparently always results in ex- B

solution of the nickel - and copper-bearing phases to leave a
rather pure pyrrhotite.

Pyrrhotite occurs only very rarely in crystals; in ore

deposits, it invariably is present as anhedral, commonly
interstitial, masses. Although once considered a "high tem-
perature" mineral, it is clear that pyrrhotite can form in
deposits from magmatic temperatures above lOOOC to sea
floor exhalative vents at 300C and probably to much lower
temperatures. It is now apparent that pyrrhotites in ores tend
to continuously reequilibrate as temperatures drop. As noted
in Figure 3, the pyrrhotites equilibrate much more readily 30 0 so 60
than pyrite or sphalerite; hence, natural ores commonly
actually preserve a variety of different temperatures when Figure 13. (A) Phase relations in the central portion of
they are examined. ' the CU~F~-S system; (a) schematic relations at 300C.
Abbreviations: c~, chalcopyrite; dj, djurleite; di, digenite; al,
Pyrrhotites also reequilibrate rapidly in the weathering anilite; cv, ~ovelhte; bn, bornite ; id, idaite; cp, chalcopyrite;
realm by suffering ready removal of iron from their lattices . ~' talnakhite; mh, mooihoekite; he, haycockite; cb, cuban -
As a consequence, meteoric weathering of pyrrhotite results ite; mp?, monocli~ic pyrrhotite; hpo, hexagonal pyrrhotite;
in a progressive compositional change toward more sulphur- tr, trOllu7; py, pynte; ISS, Intermediate solid solution. (B)
rich varieties (e.g., monoclinic pyrrhotite). Continued oxi- Schematic low-temperature phase relations in the Cu-Fe-S
dation during weathering results in sufficient iron removal to ~ys!em (after Craig and Scott, 1974). The shaded arrow
form pyrite and/or marcasite. In~cates the genera;I trend of copper-iron-sulphide minerals
dunng the weathenng process. The removal of iron and
ChalcoPyrite subsequently copper, results in the formation of comPosi-
!I_ons_ncher In sulphur.
Chalcopyrite (ideal formula, CuFeSi) is the major source Sphalerite
of copper in almost all of the ore deposits worked for that
metal. The structure of chalcopyrite (Fig. 2A) is derived Sphalerite (ideal formula, ZnS) is the only ore mineral of
from the sphalerite structure by ordered substitution of Cu zinc, although it commonly contains other metals replacing
and Fe for Zn . However, it is also possible to add or Zn (most notably Fe, but also Cd, Co, Ni, Cu, Ga, Ge).
remove metal atoms from the structure , particularly at higher The sphalerite or "zinc blende" structure is one of the fun-
temperatures, such that extensive solid solutions exist in the damental structure-types found in inorganic solids (Fig. IB).
Cu-Fe-S system at higher temperatures. Some (in particular, However, the other major structure type found in ZnS, the
metal-rich) compositions near to that of chalcopyrite (such as hexagonal structure of wurtzite, is also of fundamental im-
talnakhite, CU9FesSI6; mooihoekite, CU9Fe9S16; and hay- portance (Fig . l C). The structural relationship between
cockite, CU4FesSs) appear to be preserved in ores, although sphalerite and wurtzite is such that, although both contain Zn
they may be metastable phases. Optically, they are almost and S in tetrahedral coordination, linking of the tetrahedra
impossible to distinguish from chalcopyrite and have very (via comers) in sphalerite results in a cubic unit cell, whereas
similar X-ray powder diffraction patterns. The structures of in wurtzite the cell is hexagonal . It is also a relationship that
these stuffed derivatives of chalcopyrite are shown in Figure can be viewed in terms of close-packing of the sulphur
2B and consist of ordered occupation of additional (normally anions in layers parallel to the basal plane and the stacking of
empty) sites. these layers to give cubic or hexagonal symmetry.
Chalcopyrite, like pyrrhotite, tends to occur as anhedral Variations in stacking sequence can give rise to different
masses interstitial to other sulphides, although it has often polytypes and a submicroscopic interlayering of sphalerite
been observed as euhedral tetrahedra lining vugs in ores and and wurtzite structure types. As already noted, some
in fluid inclusions. It is known to accept significant quanti- authors have proposed that a compositional difference exists
ties of other elements, especially zinc and tin (Table 3). The between sphalerite and wurtzite, although others would
composition of chalcopyrite lies centrally within the Cu-Fe-S regard wurtzite formed at low temperatures as a metastable
system; hence, it coexists with many of the common iron phase resulting from the mechanism of nucleation of ZnS in
and copper sulphides (Fig. 13A). The weathering of chal- solution (Vaughan and Craig, 1990).
copyrite, as well as that of bornite and the copper sulphides,
proceeds in a manner similar to that of pyrrhotite. The ready The sphalerite lattice and the similarity to the size of the
removal of iron first, and then copper, results in the driving Zn2+ ion permits the ready substitution of a number of diva-
of the copper-iron sulphide assemblages across the Cu-Fe-S lent transition metal ions for the zinc. Indeed, naturally-
system as shown by the arrow on Figure 13B. occurring sphalerites always contain metals (dominantly

iron) in addition to zinc. Iron, virtually always present,
ranges from trace quantities to greater than 27 wt % and is
the primary cause of color in sphalerite. Sphalerites from
Mississippi Valley-type deposits nearly always contain less 700
!han 2% iron (commonly <1%) and are usually pale yellow
10 color, whereas those from vein and massive sulphide
deposits may contain variable amounts of iron (5-10% are
common) and are reddish to nearly black . The most iron-
rich sphalerites occur in meteorites where values in the range
of 18-27.6 wt % Fe are reported. The iron content of the 600
sphalerite (generally quoted in terms of FeS) is controlled by
the activity of sulphur. The definitive study of the Fe-Zn-S
system (Barton and Toulmin, 1966) demonstrated that the u
FeS content of the sphalerite solid solution is buffered by ,;
coexisting iron sulphides. Using those data, Craig et aI. 3
(1984) demonstrated how knowledge of the dominant iron ..
e 500

sulphides (pyrite and/or pyrrhotite) in massive ores can be
used to predict the composition of the sphalerite and hence t-
the purity of the zinc concentrate that is produced.
Barton and Toulmin (1966) defined, and Scott (1973)
later calibrated, the sphalerite geobarometer. Application of 400
the geobarometer is based upon the observation that 10 kbars
increasing pressure reduces the FeS content of sphalerite
equilibrated with pyrite and hexagonal pyrrhotite in a manner
that is temperature-independent (Fig. 14A). Sphalerite is
sufficiently refractory to preserve the original composition 7.5 kba "
unless it remains in contact with pyrrhotite or chalcopyrite; 300
low-temperature reequilibration of sphalerite generally 5 kba"
reduces the original FeS content and results in pressure esti- o kba,
mations that are too high. A
22 10
The sphalerite structure easily accommodates a variety of
elements substituting for zinc; hence , sphalerites serve as
important sources of Cd, Ga, Ge, and In. Few data are
available for the Ga, Ge, and In contents of commercially
mined ores , but Table 3 presents values for the richest re-
ported sphalerites.

Sphalerites from many types of ores contain dissemi-

nated to rod-like inclusions of chalcopyrite < pyrrhotite) in a
texture (Fig. 14B) known as "chalcopyrite disease" (Barton
and Berthke, 1987). The quantity of copper present in the
chalcopyrite commonly far exceeds the solubility of copper
in the sphalerite structure, and the disease is believed to
result primarily from selective reaction of copper-bearing
fluids along crystallographically-preferred directions within
the sphalerite lattice.

Galena Figure 14. (A) Plot of the FeS content of sphalerite

coexisting with pyrite and hexagonal pyrrhotite at 0, 2.5, 5,
Galena (PbS) is the only significant ore mineral of lead. 7.5, and 10 kbar at temperatures from 300 to 700C. (After
The crystal structure of galena is the familiar "rock salt" Scott, 1976.) (B) "Chalcopyrite disease" consisting of blebs
structure, in which both metal and sulphur are in octahedral, (actually rods) of chalcopyrite dispersed with sphalerite. See
six-fold coordination. This structure may be one of the text; field of view = 0.6 mm.
building blocks found in some of the more complex
sulphosalts. As already noted , galena has been shown to lamellae of matildite in the host galena. Silver, commonly
exhibit a total variation in Pb:S ratio of -0.1 atomic % with sought in galena-rich ores, is not taken into the galena
both lead-deficiency and sulphur-deficiency being possible structure except as a coupled substitution where the silver
and hence both p-type and n-type semiconducting behavior. and bismuth atom are coupled to substitute for two lead
atoms .
Galena has been reported to contain small amounts of
several elements (Table 3). Experimental studies have ArsenopYrite
demonstrated that complete solid solution exists at high tem-
peratures between galena (PbS) and c1austhalite (PbSe) and Arsenopyrite (ideal formula, FeAsS) is worthy of inclu-
between galena and matildite (AgBiS2). Whereas the PbS- sion in this review because of its importance as a carrier of
PbSe solid solution remains stable at low temperature, the gold, either as very fine particles or as so-called "invisible"
PbS-AgBiS2 solid solution decomposes on cooling below gold. Arsenopyrite may also be of value to the ore mineral-
about 200C. The breakdown of this solid solution com- ogist as an indicator of the temperature of formation of an
monly results in the development of very fine exsolution assemblage (i.e., in geothermornetry, Fig. 15), whereas to
the mineral technologist it is a mineral more likely to cause
problems during tailings disposal or smelting due to its
arsenic content.

po. to. l.
I8a ' ) S CIV)



Co (Ill I''' ;'

,,+po+L S.

l soo


np+PY+L B .w

Figure 16.
(Ref 21).
(A) A half-unit cell of tetrahedrite
CU12Sb4S13. (B) Coordination polyhedra in tetrahedrite

The crystal structure of CUl2Sb4S13 tetrahedrite is
shown in Figure 16. In each half-unit cell, ten M1 and two
MOl atoms occupy six 4-fold and six 3-fold coordination
sites, the Mill atoms occupy the equivalent of a tetrahedral
site in sphalerite but are bonded to only three sulphur atoms
Figure 15. Pseudobinary temperature-composition plot resulting in a void in the structure; 12 S atoms are 4-coordi-
showing arsenopyrite composition as a function of tempera-
ture and equilibrium mineral assemblage. The assemblages nate and the other single S atom is 6-coordinate. Numerous
numbered 1-8 in the Fe-As-S phase diagrams on the right studies have been undertaken of the nature and extent of the
correspond to the labeled curves in the left-hand diagram. various substitutions that occur in the tetrahedrites. For
Abbreviations: asp, arsenopyrite; py, pyrite; po, pyrrhotite; example, using a variety of spectroscopic techniques,
10, lollingite; L, liquid. (After Kretsehman and Scott, 1976) Charnock et al. (1989) were able to show that silver goes
into trigonal rather than tetrahedral sites, cadmium into tetra-
The crystal structure of arsenopyrite is closely related to hedral sites and iron mainly into tetrahedral sites, although in
that of marcasite. It is, of course, a "disulfide" in which Fe natural tetrahedrltes of low silver content, it may enter the
atoms are octahedrally coordinated to S-As dianion units . trigonal site.
As in marcasite, the MX6 octahedra share edges normal to
the c-axis direction, but whereas in marcasite the Fe-Fe dis- The tetrahedrite -tennantite series exhibits a broad range
tances are of uniform length along the c-axis, in arsenopyrite of solid solutions (Table 3; Johnson et al., 1986). Most
there are alternate long and short Fe-Fe distances (Fig. ID). tetrahedrites are mined for their silver contents, and numer-
ous studies have demonstrated that the silver contents corre-
Arsenopyrite, even more than pyrite, exhibits a tendency late with antimony contents. It is clear that silver and arsenic
to occur in well-developed subhedral to euhedral crystals. have little tolerance for each other in the tetrahedrite
The correlation of arsenopyrite with gold has long been rec- structure . However, as antimony content rises, the capacity
ognized but remains poorly understood. Most gold with for silver substitution also rises.
arsenopyrite occurs as discrete electrum grains but Johan
(1982) reported concentrations of up to 1.6 wt % gold in Pentlandite
arsenopyrite, and Cabri et al. (1989) and Cook and
Chryssoulis (1990) report 0.44% and 1.3%, respectively. Pentlandite Ni,Fe)9Ss) is the major ore mineral of
nickel. A complete solid solution occurs between pentlandite
Kretschmar and Scott (1976) have demonstrated that the and cobalt pentlandite (CQ9SS). The crystal structure of
As content of arsenopyrite when equilibrated with pyrite and pentlandite is unique and contains "cube clusters" of tetrahe-
pyrrhotite varies as a function of temperature and hence can drally coordinated metals with very short metal -metal
be used as a geothermometer (Fig. 15B). This has found distances along the cube edges; the additional (ninth) metal
considerable application but appears to often suffer from atom in pentlandite occurs in octahedral coordination. The
low-temperature reequilibration. crystal chemistry and mineral chemistry of pentlandite
exhibits a number of interesting features. One of these is
Tetrahedrite that it undergoes substantial volume contraction on cooling,
evidenced in natural samples by the presence of numerous
Tetrahedrite is a complex sulphosalt mineral with an ideal fractures. It is also important to note that pentlandite almost
composition of Cu12Sb4S13 and exhibiting a complete solid always forms as a subsolidus phase in the Fe-Ni-S system
solution in natural samples to tennantite (Cu12AS4S13). (see Fig. 4).
However, almost all natural tetrahedrites contain substantial
Pentlandite formation occurs primarily as a result of ex-
amounts of other metals substituting for copper in the
solution in the form of so-called "flames" in nickeliferous
structure, notably Zn, Fe, Cd, Hg and Ag. \l}eHfral pyrrhotite. Exsolution occurs such that pentlandite (Ill),
f0rIl}ula for the tetrahedrites can be given as M~oM2 M4 S 13 (110) and (112) planes are parallel to the (001), (110) and
( M = Cu, Ag; MIl = Fe, Zn, Cd, Mn, Cu; (100) planes ofthe pyrrhotite (Francis et al., 1976). The ex-
Mill = Sb, As, Bi). As a consequence, tetrahedrite may be solution occurs as a consequence of the shrinkage of the
an important ore mineral for a number of metals and is, in (Fe,Nih-xS monosulphide solid solution (mss) field during
fact, the dominant ore mineral for extraction of silver. Thus. cooling. The mss phase that spans the Fe-Ni-S system at
although volumetrically minor, the "tetrahedrltes" are of high temperatures (Fig. 4A) beings to thin (Fig. 4B) as tem-
considerable technological significance. perature falls . Consequently, pentlandite exsolves as the


The authors acknowledge support from NERC Grant

GR3/6823 for collaborative research on iron sulphides and
the help of a large group of mineralogists and ore petrolo-
gists who have stimulated their curiosity toward, and
educated them about, sulphides. JRC also acknowledges
USBM allotment grant G1194151 to Virginia Tech that helps
support his sulphide work, and DJV acknowledges SERC
Grant GR/F/62841 that supports work on copper and
copper-iron sulphides.


Barton, P.B. and Berthke, P.M. (1987) Chalcopyrite

disease in sphalerite: pathology and epidemiology.
American Mineralogist, 72, 451-467.
Barton, P.B. and Skinner, B.J. (1979) Sulfide mineral sta-
bilities. In Barnes, H.L. (ed.), Geochemistry of
Hydrothermal Ore Deposits, 2nd edition. Wiley-
Interscience, New York, p. 278-403.
Barton, P.B. and Toulmin, P. (1966) Phase relations
involving sphalerite in the Fe-Zn-S system. Economic
Geology, 61, 815-849.
Basu, K., Bortnikov, N.S., Mookherjee, A., Mozgova,
N.N. , Svitsov, A.V., Tsepin, A.I. and Vrubeskaja,
Z.V. (1984) Rare minerals from Rajpura-Dariba,
Rajasthan, India. V: The first recorded occurrence of a
manganoan fahlore. Neues Jahrbuch Mineralogie,
Figure 17. (A) Pentlandite "flames" fonned as a resu1~ of Abhandlungen, 149, 105-112.
crystallographically-oriented exsolution from msson cooling Bloem, J. and Kroger, EA. (1956) The P-T-X phase dia-
(field of view = 0.6 mm). (B) Chain-like masses of pent- gram of the lead-sulfur system. Z. Phys. Chern. NF, 7,
landite between grains of pyrrhotite. The characteristic 1-14.
fractures in the pentlandite are believed to result from the Boldryeva, M.M. and Borodayev, Y.S. (1973) Zinc-
high coefficient of thermal expansion of the pentlandite bismuth tetrahedrite, a new variety of grey ore.
during cooling. Transactions USSR Academy of Science, Earth Science
Sections, 212, 180-181.
sulphur-poor boundary of the mss retreats towards more Brill, B.A. (1989) Trace element contents and partitioning of
sulphur-rich compositions. elements in ore minerals from the CSA Cu-Pb-Zn
deposit, Australia. Canadian Mineralogist, 27, 263-274.
Kelly and Vaughan (1983) have experimentally traced Buerger, N.W. (1934) The unmixing of chalcopyrite from
the development of the pentlandite textures that result from sphalerite. American Mineralogist, 19,526-530.
exsolution. Initial exsolution at the highest temperatures re- Burke, E.A.J. and Kieft, C. (1980) Roquesite and Cu-In-
sults in the formation of discrete blebs of pentlandite along bearing sphalerite from Langban, Bergslagen, Sweden.
the grain boundaries of the mss. During contin.uedexs?lu- Canadian Mineralogist, 18,361-363.
tion the blebs grow and coalesce to form connnuous nms Cabri, L.J., Blank, H., EI Goresy, A., Laflamme, J.H.G.,
between the mss grains. At lower temperatures, where rates Nobiling, R., Sizgoric, M.B. and Traxel, K. (1984)
of diffusion and exsolution are far slower, rims no longer Quantitative trace-element analyses of sulfides from
form but rather there are developed the fine blades and Sudbury and Stillwater by proton microprobe. Canadian
flam~s shown in Figure 17A. No doubt, some of the earliest Mineralogist, 22, 521-542.
formed lamellae coalesce into the rims, but because the Cabri, L.J., Campbell, J.L., Laflamme, J.H.G ., Leigh,
pentlandite exsolution continues during cooling, new flames R.G., Maxwell, J.A. and Scott, J.D. (1985) Proton-
continue to be formed as prior ones disappear. The pe~t microprobe analysis of trace elements in sulfides from
landite chains are characteristically highly fractured (Fig. some massive-sulfide deposits. Canadian Mineralogist,
17B), apparently because pentlandite has a coefficient of 23, 133-148.
thermal expansion two to 10 times grea.ter th!'" tha! of. ~e Cabri, L.J ., Chryssoulis, S.L., de Villiers, J.P.R. ,
sulphides (e.g., pyrrhotite and chalcopynte) With which It IS LaFlamme, J.H.G. and Buseck, P.R. (1989) The nature
commonly associated (Rajamani and ~ewitt, 1975). The of "invisible" gold in arsenopyrite. Canadian
greater shrinkage suffered by pentlandite results 10 greater Mineralogist, 27,353-362.
stress and subsequent cracking. Charnock, J.M ., Garner, C.D., Pattrick, R.A.D. and
Vaughan, D.J. (1989) Coordination sites of metals in
Pentlandite has traditionally been considered an ore tetrahedrite minerals determined by EXAFS. Journal of
mineral of nickel that yields by-product cobalt. In the past Solid State Chemistry, 82, 279-289.
20 years, however, pentlandites have been found that ~S? Chryssoulis, S.L., Cabri, L.J. and Lennard, W. (1989)
carry significant quantities of silver (Scott and Gasparrini, Calibration of the ion microprobe for quantitative trace
1973) and ruthenium and rhodium (Cabri et al., 1984) as precious metal analyses of ore minerals. Economic
shown in Table 3. Geology, 84, 1684-1689.
Cook, N.J. and Chryssoulis, S.L. (1990) Concentrations of
"invisible gold" in the common sulfides. Canadian
Mineralogist, 28, 1-16,
Cox, D.P. and Singer, D.A. (1987) (eds.) Mineral Deposit Johan, Z., Picot, P. and Ruhlmann, F. (1982) Paragenetic
Models. U.S. Geological Survey Professional Paper evolution of the uranium mineralization rich in selenium
1693. at Chameane (Puy-de-DOme), France: chameanite, gef-
Craig, J.R. (1990) Textures of the ore minerals. In Jambor, froyite and giraudite, three new selenides of Cu, Fe, Ag
J.L. and Vaughan, D.J. (eds.), Advanced Microscopic and As. Tscher. Mineral. Petrogr. Mitt, 29,151-167.
Studies of Ore Minerals. Mineralogical Association of Johnson, N.E., Craig, J.R . and Rimstidt, J.D. (1986)
Canada Short Course Series, voL 17, p. 213-262. Compositional trends in tetrahedrite. Canadian
Craig, J.R. and Vaughan, D.J. (1981) Ore Microscopy and Mineralogist, 24, 385-397.
Ore Petrography. Wiley Interseience, New York. Kase, K. (1987) Tin-bearing chalcopyrite from the Izurno
Craig, J.R. and Vaughan, D.J. (1983) Growth characteris- vein, Toyoha Mine, Hokkaido, Japan. Canadian
tics of sphalerite in Appalachian zinc deposits. In G. Mineralogist, 25, 9-13.
Kisvarsanyi et al. (eds.), ~oc. In!;1 C0!1f. ?n Kelly, D.P. and Vaughan, D.J. (1983) Pyrrhotite-pent-
Mississippi Valley Type Lead-Zinc Deposits, University landite ore textures: a mechanistic approach.
of Missouri-Rolla, p. 317-327. Mineralogical Magazine, 47, 453-463.
Craig, J.R. and Vaughan, OJ. (1986) Paragenetic studies of Kelly, W.C. and Clark, B.R. (1975) Sulfide deformation
growth-banded sphalerites in Mississippi Valley -type studies, ill. Experimental deformation of chalcopyrite
zinc deposits of the Appalachians. In J.R. Craig et al. up to 2000 bars and 500C. Economic Geology, 70,
(eds.), Mineral Paragenesis. Theophrastus Publishers, 431-453.
S.A Athens, Greece, p. 153-158. Kieft, K. and Damman, A.H. (1990) Indium-bearing chal-
Craig, J.R., Ljokjell, P. and Vokes, F.M . (1984) Sphalerite copyrite and sphalerite from the GAsborn area, West
compositional variations in sulfide ores of the Bergslagen, Central Sweden. Mineralogical Magazine,
Norwegian Caledonides. Economic Geology, 79,1727- 54, 109-112.
1735. Klemm, D.O. (1965) Synthesen und Analysen in den
Doelter, C. (1926) Handbook of Mineral Chemistry. Dreiecks-diagrammen FeAsS-CoAsS-NiAsS und FeSz-
Eldridge, C.S ., Boucier, W.L., Ohmoto, H. and Barnes, CoS r N is 2. Neues Jahrbuch Mineralogie,
H.L. (1988) Hydrothermal inoculation and incubation of Abhandlungen, 103, 205-255.
the chalcopyrite disease in sphalerite. Economic Kostov, I. and Mincheeva-Stefanova, J. (1981) Sulphide
Geology, 83,978-989. Minerals: Crystal Chemistry, Paragenesis and
Fleet, M.E., MacLean, P.J. and Barbier, J. (1989) Systematics. Bulgarian Academy of Sciences.
Oscillatory-zoned As-bearing pyrite from stratabound Kovalenker, V.A. and Rusinov, V.L. (1986) Goldfieldites:
and stratiform gold deposits: an indicator of ore-fluid peculiarities of the chemical composition, paragenesis
evolution. Economic Geology Monograph 6, 356-362. and conditions of formation. Mineralogicheskiy
Fleischer, M. (1955) Minor elements in some silfide Zhumal, 8, 57-70.
minerals. In Bateman, AM. (ed.) , Economic Geology Kovalenker, V.A. and Troneva, N.V. (1980) On the gold-
Fiftieth Anniversary Volume, p. 970-1024. bearing fahlore, Sulphosalts, platinum minerals and ore
Fleischer, M. (1987) Glossary of Mineral Species, 5th microscopy. Proceedings XI General Meeting of IMA
edition. The Mineral. Record, Inc., Tucson, 234 p. Moscow, p. 75-83.
Foord, E.E. and Shawe, D.R. (1989) The Pb-Bi-Ag-Cu- Kretschmar, U. and Scott, S.D. (1976) Phase relations
(Hg) chemistry of galena and some associated sulfosalts: involving arsenopyrite in the system Fe-As-S and their
a review and some new data from Colorado, California, application. Canadian Mineralogist, 14,364-386.
and Pennsylvania. Canadian Mineralogist, 27, 363-382. Loucks, R.R . and Petersen, U. (1988) Polymetallic
Fralick, P.W., Barrett, T.J., Jarvis, K.E., Jarvis, I., epithermal fissure vein mineralization, Topia, Durango,
Schnieders, B.R. and Vandekemp, R. (1989) Sulfide- Mexico: Part Il, Silver mineral chemistry and high reso-
facies iron formation at the Archean Morley occurrence, lution patterns of chemical zoning in veins. Economic
Northwestern Ontario: contrasts with oceanic hydro- Geology, 83, 1529-1559.
thermal deposits. Canadian Mineralogist, 27, 601-616. Marignac, C. (1989) Sphalerite stars in chalcopyrite: are
Francis, C.A., Fleet, M.E., Misra, K.C. and Craig, J.R. they always the result of an unmixing process?
(1976) Orientation of exsolved pentlandite in natural and Mineralium Deposita, 24, 176-182.
synthetic nickeliferous pyrrhotite. American McLimans, R.K. , Barnes, H.L. and Ohmoto, H. (1980)
Mineralogist, 61,913-920. Sphalerite stratigraphy of the Upper Mississippi Valley
Goble, R.J. (1980) Copper sulfides from Alberta: zinc-lead district, southwest Wisconsin. Economic
yanowite, CU9SS, and spionkopite, Cu3QS28. Canadian Geology, 75, 351-361.
Mineralogist, 18,511-518.
Goble, R.J. and Robinson, G. (1980) Geerite, CU1.60S, a Mills, K.C. (1974) Thermodynamic Data for Inorganic
Sulphides, Selenides and Tellurides. Butterworth,
new copper sulfide from Dekalb Township, New York. London.
Canadian Mineralogist, 18,519-523. Nikitin, W.W. (1929) Parallel intergrowths of fablore and
Godorikov, A.A. and Il'yasheva, N.A. (1973) Chemical their chemical constitution. Zeitschrift fiir
characteristics of fablores. Int'l Geology Review, 15, Kristallographie, 69, 482-502.
1413-1422. Paar, W.H., Chen, T.T. and Gunther, W. (1978) Extremely
Harris, D.C., Cabri , L.J. and Nobiling, R. (1984) Silver- silver-rich freibergite in the Pb-Zn-Cu ores of the
bearing chalcopyrite, a principal source of silver in the Bergbaues "Knappenstube," Hoctor, Salzberg.
lzok Lake massive-sulfide deposit: confirmation by Carinthia 11, 168,35-42.
electron- and proton-microprobe analyses. Canadian Pattrick, R.AD. (1985) Pb-Zn and minor U mineralization
Mineralogist, 22, 493-498. at Tyndrum, Scotland. Mineralogical Magazine, 49,
Hutchison, M.N. and Scott, S.D. (1981) Sphalerite geo- 671-681.
barometry in the system Cu-Fe-Zn-S. Economic Pearson, M.F., Clark, K.F. and Porter, E.W. (1988)
Geology, 76, 143-153. Mineralogy, fluid characteristics, and silver distribution
Johan, Z. (1988) Indium and germanium in the. s~cture .of at Real de Angeles, zacatecas, Mexico. Economic
sphalerite: an example of coupled substitution With Geology, 83, 1737-1759.
copper. Mineralogy and Petrology, 39, 211-229. Picot, P. and Marcoux, E. (1987) Nouvelles donnees sur la
metallogenie de l'or. Academy of Science [Paris]
Comptes Rendus, 304, 221-226 .

Rajamani, V. and Prewitt, C.T. (1975) Thermal expansion
of the pentlandite structure. American Mineralogist, 60,
Scheubel, F.R., Clark, K.F. and Porter, E.W . (1988)
Geology, tectonic environment, and structural controls in
the San Martin de Bolanos District, Jalisco, Mexico.
Economic Geology, 83, 1703-1720.
Scott, S.D. and Barnes, H.L. (1972) Sphalerite-wurtzite
equilibria and stoichiometry. Geochimica et
Cosmochimica Acta, 36, 1275-1295.
Scott, S.D. and Bames, H.L. (1973) Experimental calibra-
tion of the sphalerite geobarometer. Economic Geology,
Scott, S.D. and Gasparrini, E. (1973) Argentian pentlandite,
(Fe,Ni)sAgS s, from Bird River, Manitoba. Canadian
Mineralogist, 12, 165-168.
Skinner, B.J. and Luce, F.D. (1971) Solid solutions of the
type (Ca,Mg,Mn,Fe)S and their use as geothermometers
for the enstatite chondrites. American Mineralogist, 56,
Spiridonov, E.M., Kachalovskaya, V.M. and Chvileva,
T.N. (1988) Thallium-bearing hakite, a new fahlore va-
riety. Trans. USSR Academy of Science, Earth Science
Sections, 290, 206-208.
Tokonami, M., Nishiguchi, K. and Morimoto, N. (1972)
Crystal structure of a monoclinic pyrrhotite (Fe7SS).
American Mineralogist, 57,1066-1080.
Tossell, J.A., Vaughan, D.J. and Burdett, J.K. (1981)
Disulfide minerals: crystal chemical and structural prin-
ciples. Physics and Chemistry of Minerals, 7, 177-184.
Vaughan, D.J. and Craig, J.R. (1978) Mineral Chemistry of
Metal Sulfides. Cambridge University Press.
Vaughan, D.J. and Craig, J.R. (1990) Sulfide ore mineral
stabilities. In Bames, H.L. and Ohmoto, H. (eds.),
Hydrothermal Processes: Applications to Ore Genesis.
Vaughan, D.J., Tossell, J.A. and Stanley, C.J. (1987) The
surface properties of bornite. Mineralogical Magazine,
51, 285-293.
Wiggins, L.B . and Craig, J.R. (1980) Reconnaissance of
the Cu-Fe-Zn-S system: sphalerite phase relations.
Economic Geology, 75, 742-751.
Wu, X. and Delbove, F. (1989) Hydrothermal synthesis of
gold-bearing arsenopyrite. Economic Geology, 84,
Yund, R.A. and Kullerud, G. (1966) Thermal stability of
assemblages in the Cu-Fe-S system. Journal of
Petrology, 7, 454-488.

Rio Tinto deposits - geology and geological models for
their exploration and ore-reserve evaluation
F. Garda Palomero
Rio Tinto Minera, S.A. Minas de Riotinto, Hue/va, Spain

This naper describes the geology of the ,
Rio Tinto area, and the morphological anc
reserves, and also as a control on
subsequent reserve evaluation using
geostatistical methods.
genetic relationships between different
mineralisation types (stockwork, massive INTRODUCTION
sulphide and gossan) in each of the
The Riotinto sulphide deposits are
three deposits (San Dionisio, San
located in the Huelva province of
Antonio and Cerro Colorado). The
southern Spain, 90 km northwest of
geological relationships and spatial
Seville and ?O km northeast of Huelva
distribution of the ore mineralisation,
(Figure 1).
and associated alteration minerals,
~uning activity in the area dates
define a mineralisation model related
from pre-Roman times, with a peak of
to shallow submarine volcanic activity.
activity between the third and first
~lineralisation begins with precipi-
centuries BC (Rothenberg et aI, 198?)1,
tation of sulphides in fissures in the
as illustrated by the large amount of
volcanic hostrocks at estimated depths
slags (6 million tonnes) derived from
of 400 metres below the seafloor, and
treatment of gold and. silver ore.
temperatures of 4002C. Sulphide
After a long interval during which
precipitation from ascending hydro-
only sporadic mining occurred,
thermal fluids increases towards the
production of copper from the zones
seafloor, and leads to the development
of secondary sulphide enrichment, with
of mineralised stockworks as the
grades of 3-5% Cu, began in middle
temperatures decrease to 1002C at the
of the last century. Towards the
top of the volcanic pile. The remaining
end of the last century, as this material
sulphur and metals are deposited on the
was worked out, activity switched to
seafloor in massive sulphide lenses.
the exploitation of the primary massive
The distribution of minerals in
sulphides for sulphur production. With
haloes to both types of mineralisation
the drop in pyrite priCes, at the end
are related to cooling of the orefluids
of the 1960's, new mining operations
as they move outwards from the source
were started for the production of
area. copper from the low grade sulphide
After the cessation of volcanic
stockworks (enriched by supergene
activity in the area flysch sedimen-
activity near the surface), and
tation continued followed by tectonism
production of gold and silver from t he
associated ,dth the Hercynian orogeny.
surfical oxide cap (Gossan) formed
Subsequent erosion and strong
by the weathering of underlying
weathering of the deposits, during the
late Tertiary period, led to the
formation of extensive gossan deposits.
The geological model of metal
distribution in the sulphide and gossan
On a regional scale, the Riotinto
deposits has been successfully used
deposits lie in the Iberian Pyrite
as a guide in exploration for new
Belt, an upper Palaeozoic unit

con~aining volcanics (andesites and Riotinto anticline wh e r e it displays
rhyolites) at the base of the Carbonife- intense chloritic alteration associated
rous, with whd ch the abundant sulphide with the mineralisation process. It
deposits are associeted both spatially ahowa a fairly uniform thickness of
and genetically (Figure 1). 400-500 metres in the Riotinto area.


~ Post-Pa leozo ic s e d i rnen ts ~
~ Volcani c 12v 21 ~ LOW2r Pal~oZ\)ic
c.:==.::J Devon ian sla tes ~ Gran ite

Figure 2 shows the geology of the Acid Complex

Riotinto area, wh e r-e sulphide minerali-
sation is associated with the volcanic This unit consists of very homogeneous
rocks of .t h e Pdotinto anticline (Garcia acid volcanics (dacitic to rhyolitic
Palomero, 1974)2, being overlain by in composition), with predominant
barren rocks of Cul~ facies. The main pyroclastics, and locally abundant
lithostratigra?hic units are described breccias, autobreccias and lavas.
briefly b e Lo w, These rocks have a porphyritic texture
with phenocrysts of quartz and feldspar
Devonian in a microcrystalline matrix of the
Thi s unit outcrops to the south of the same composition, which shows a clear
schistosity and frequent alteration
Riotinto area, in the core of a mainly
vo lcanic anticlinal structure. It (chlorite, haematite, and sericite).
consists mainly of grey argillaceous The thickness of the acid volcanic
slates with some quartzitic levels near horizon is generally fairly uniform and
the top. The first evidence for Devono- similar to that of the Basic Complex
Carboniferous volcanic activity, dated (400-500m), although in the area of
as Tournaisian ( Schermerhorn, 1971)3, the mineralisation it diminishes
occurs near the top of this unit, considerably thinning to only 30-50
metres in the centre of Corta Atalaya
Basic Complex
(Figure 2).
This unit conformably overlies the
Devonian, and consists of alternations Transition Series
of basic pyroclastics, vesicular an desi-
tes, slates, fine grained acid pyro- This is typically a heterogeneous
clastics and a variety of breccias. pyroclastic level with abundant lateral
It outcrops to the south of the area, and vertical changes in both grain
as well as in the core of the mai n size and composition of the constituent

:"," -'__ -: _"1 .'

- . . - . ,===,,,,,,'000 m I ~

~ Devcni c n slc te s r:::-:.I Massiv2 s ul p hi d e s

I- +'++1 Basic co rn p l e x c=J Culm

E=,:=:,:'-J S2dim2ntary -Pyroclast ic hor izon ~ 'In situ' Gossan

~ A c id c ornp te x ~ Ir cns por te d Gossan

b,- ,..~- \I Tr ans it i 0 n SUi2S COlcio Pa lom et a F. "7 '

lithologies (polygenic breccias, litho- elevations (400 metres, 350 metres, etc)
crystalline tuffs, finely crystalline coinciding with possible terraces
tuffs, chloritised-haematised-serici- associated with the evolution of the
tised tuffs, massive sulphides, etc.). river systems.
These represent the lithologies formed All the units described, except
at the end of a phase of volcanic for the Tertiary, are affected by the
activity with the transition to the Hercynian orogeny. This resulted in
flysch sedimentation of the Culm. The various phases of folding which formed
thickness of this unit is very variable the main ea at--we at; folds, and an
(5- 50 metres). associated schistosity (dipping north
The most notable feature of this at 75-85 degrees), as we .LL as faultin G"
series is that mineralised massive and thrusting. A weak metamorphism of
sulphide lenses, of probable syn-sedi- greenschist facies is also developed.
mentary origin, occur within it. Geological studies of the units
described above have identified some
Culm features whi c h indicate the palaeogeo-
graphic conditions prevailing at the
This unit consists of a monotonous time of mineral isation. Some of these
series of turbiditic sediments, although phenomena are specific to the Rio Tinto
within it different units can sometimes area, an~ may be rel ated to the p r e s en ce
be distinguished on the basis of of mi n e r a l i s a t i on , including the
fossiliferous horizons, and differences follo wing:
in the relative proportions of gre~~acke 1. The predominance of pyr o c l a s t i c
and shale. rock s wi t h respect to l avas, in the
Basic Complex of the Pd o Tinto area.
Tertiary The associate d l ava s contain a b unda n t
vesicles of ~uartz an d sulphides.
This is represented by subhorizontal These features disappear a way from
deposits, 1-10 metres in thickness, the Pdo Tinto area.
of a limonite cemented mixture of 2. The decrease in thickness of
fragments of volcanic rocks, slates, the Acid Volcanic unit close to the
and iron oxides. These are laterally mineralisation (from 500 to 100 metres,
fairly restricted, and occur at specific '.d t h a minimum of 30 metres). A parallel

decr ea so in the t hickn es s of t ho Ma s siv e 3ulphi des:
.Jyroclastic e.Lemerrt s in t he transition Amount 500-600 Mt
series i s a l s o noted. Grades S 50% cu 1%,
3. At the e a stern and "{estern limit .:: Zn 2 ~~, Pb 1%,
of the area, coincident ' : i th strong Ag 30ppm, Au 0.3ppm
accwnulc.:tions of pyroclastic :i..n
t he acid volcanic level, abundant stock'-lork sulphi des:
fragments of oxidised lapilli tuff Amount 1900-2000 Ht
a nd coarser pyroclastics are seen, Gr a de s S 6 %, Cu 0.155'; ,
suggesting very aha.Ll.ow "later volc ani sm. Zn 0.15%, Ph 0.06%,
J!.. At the base of the Culm frequent Ag 7ppm, Au 0.07Ppm
slump features are seen, and the b Qsal
unit of the Culm is absent close to Of the total sulphide mineralisation
the mineralised zone. formed, at least 400 mi l l i on tonnes of
These features seem to indicate a massive sulphides and 400 million tonnes
very shallow marine e n v i r o n me n t of stockwork mineralisation have been
coinciding wi t h the present-day affected by weathering and erosion.
anticline, with marginal shallo\'! Part of this remains in the form of
oxide caps (Gossan) totalling in excess
volcanic centres. This difference in
of 100l-It ,dthin whi c h the concentration
relief wa s eliminated during deposition
of less soluble metals (Au, Ag, Pb, Da,
of the Culm.
30, etc.) has been enriched by the
"'e a t h e r i n g process.
The three different mineralise d zones
are described separately be.l.ow,
l-lin e r a l i s a t i on in the Riotinto area
occurs in three separate zones, situated
Sa n Dioni s io
in different parts of the Riotinto
anticline. Although locally inter- This zone is located on t he south fl a nk
connected, each zone belongs to a of the Riotinto anticl ine, associated
separate g e n e t i c unit, sharing a comnlon wi t h a secondary synclinal structure
g e n e t i c model. which plunges east a t 30-3 5 degrees
Th e three mineralised zones are: (Figure 2).
Sa n Dionisio, PlaneS/ San Antonio, Figure 3 shows a transverse N- S
Cerro Colorado/Filon Norte/Filon Sur. section through the centre of this
Each of these zones is made up of deposit, illustrating the interrela-
three different types of mineralisation, tionship bet~een the different types of
namely; mineralisation, and lithologies present.
SUlphide stockNorks in the volcanic The different types of mineralisation
r-o c k s , are described be Low,
St ra t i f o r m mass~ve sulphide lenses
si tuated on top of the stoek"lork Stockwork
mineralisation, ' a n d intercalated
tith other rocks of t h e transition Stockwork mineralisation in San Dionisio
consists of irregular veins containing
pyrite, chalcopyrite, galena, sphalerite,
Gossan (iron oxide cap), formed
magnetite, quartz, chlorite, calcite a nd
l:l y ";e a t h e r i ng of outcropping massive
b arite, cutting the volcanic host rock s.
sulphide and stock":ork mineralisation,
The top of the stockwork marks the
a nd r estric'c e cl to "lithin 70 metres
b a se of the ma s s i v e sulphides, an d it
of the surface.
extends for up to 400 metres stratigr a-
Th e total a mount of sulphide minera-
phically downwar-ds fro m thi s contact,
lisa tion in the ~otinto area sug gests down to the Basic Volcanic complex.
t h at t h e area wa s the locus of one of The minera i.isation takes the form of
t he most intense mi n e r a l isi ng episodes a n originally s emi v e r t i cal chimney with
do cum e n t e d to date. a s emi c i r c ul a r cross section of about
Th e i mportance of the mineralising 60 0- 70 0 metres in diameter. Folding
event i s indicated by the following s u b s e qu e n t to f o r mat i o n of the deposit
stuti stic s: h as t ilted this chimney through 90
d e gr e e s~ so that it now lies in a
J.ii n e ral i s e d area 4 km2 s u b - h or i z o n t a l p o s i t i on , elongated
Vertical ext e n t 4 00- 500 m. n or th- south.

N 100
s N

S~co n da ry
Su I p h j d2
U.;.~-r':":";~-t-----8 a si c I)b c k s
W~ l l def i n e d contact I Py PYR ITE
Stockwork- Massi v2 O r e


O UA R T Z .

S lat~s
.-l-:; C hi C H L OR IT E

Stockw!>rk . i~ Mog MAGNET ITE


--. G


Se r i c it i c Cu
R: - _
Ac id Ro ck s Pta- Zn
4 7. 20


Fi g . - 3


Se r i c i t i c Rocks --\
Chi - O-"t-CPy_'Sp h _ G - _---"\-_ Stockwork. a nd .M ass iv~ S ulph id ~

- -==
o lO Omis .

Ga.r cia Poio mHO F. 19 &3

Variations in vein density, thickness, The mineral paragenesis present is

texture, orientation, mineralogical indicated in Figures 3 and 4.
composition, wall rock alteration, and Sericite Envelope, consisting of
relationship with the massive sulphides, weak mineralisation in the outer
allow the definition of a spatial zona- fringes of the stoc~work zone. It
tion within the stockwork mineral~sation. is restricted to the acid volcanic
This reflects the physio-chemical rocks within which strong serici-
conditions affecting the hydrothermal tisation occurs contiguous with the
solutions as they ascended towards the top of the three dimensional stockwork.
seafloor. The zonation is " made up of
the following three units: Hassi ve Sulphides
- Pyrite Chimney (stockwork nucleus),
consisting of very intense pyrite !'lassive sulphide mineralisation in San
mineralisation, with local chalcopy- Dionisio forms a large lens of sulphides
rite, in which original wa l l r o c k sitting directly on top of the stockwork.
lithologies are almost indistin- This has a more irregular shape and
guishable. No clear contact with greater lateral extent than the
the massive sulphides can be deter- underlying stockwork.
mined (Figure 3). Figure 3 shows the folded nature
Three Dimen sional St o c k wo r k (typical of this lens, and its stratigraphic
at.ockwor-k ) , which is widely developed l o c a t i on between the acid volcanic
spatially, and surrounds the pyrite rocks below, and the Culm above. Its
chimney. It is made up of a three spatial relationship with respect to
dimensional network of sulphide the stockwork is also ahown ,
veins which do not cut the overlying Within the ma ssi ve sulphi de bo dy,
massive sulphides, with a clear the accumulation of sulphides is
contact between the two units. fairly monotonous with very few other
'. Vithin this type of stockwork a minerals, apart from some thin
varied mineral paragenesis is seen, pyroclastic horizons near the footwall
along with intense alteration (quartz, and hangd.ngwa.lL contacts, and a weak
chlorite, and sericite) of the hostrocks. interstitial siliceous matrix which

Pyroc las t ic s

A cid Se r i c i t e
Volcan ics Enve lope

. o
Pyro cl a s I i c
te ve t
Basic o lJ. ..
cornp t e x

lJ. COLL () IOA ~

Gn 'cia Pa to rnu o F "eo

never exceeds 2p by weight. The sulphide sphalerite-galena above the three

minerals are totally recrystallised and dimensional stock",'ork. Beyond the
ahow some relict textures of a primary limits of the stockwork mineral isation
sedimentary character (colloform, pyrite with sphalerite-galena occurs.
framboidal, etc.) (Read, 1967)4, b) Upper unit. This is made up of
particularly ldthin the pyrite. Less very fine grained pyrite, wi t h a low
stable sulphide minerals, such .as base metal content, and occasional
chalco~yrite, galena, tetrahedrite, disseminated silica especially tiovrards
stannite, and bornite, are recrystalli- the top. Local irregular concentrations
sed in intergranular spaces bet"'een the of galena and chalcopyrite are related
pyrite and sphalerite crystals, or in to remobilisation associated with
microfractures ,dthill them. ilespite Hercynian folding, which also caused
the monotono~s character of the sulphiG3 strong remobilisation of quartz.
body, t wo well defined units can be The massive sulphides are overlain
c i s t i ngui s h e d on the basis of degree by shales of the Culm series, a s
of recrystallation, and mineral parage- indicated in Figure 3. At the lateral
nesis. These are as follo ws: limits of the massive sulphide minera-
a) Ga sul unit. This shows occasional lisation interfingering ,-: itil the
sul~hicle banding (pyrite-sphalerite-
Transition Series occurs. This series
;;; al e n a ) , is nred Lum-cfLne grained, and
thins over the centre of the sulphide
contains the bulk of the base metals in
lens where it is represented by a 1-3
the massive sulphides. ':li t h i n this
metre thick pyroclastic level.
wlit, metal distribution appears to be
related to structure and/or the under-
lying stock~ork types. Gossan
Variations in paragenesis are
illustrated in Figure 3, wi t h the As indicated in Figure 3, the near
ureferential occurrence of pyrite- surface part of the stockwork and
~halcopyrite Lmme dd.a t.e.Ly above the massive sulphide mineralisation ahows
pyrite chimney, and pyrite-chalcopyrite- intense we a t.her-Lng , Oxidation of the

sulphide minerals and leaching of the As in San Dionisio, this stockwork is
volcanic rocks has led to the formation a gain made up of three different units,
of an Oxide Cap (Gossan), which extends namely:
down to a max:imm depth of 70 metres below - Pyrite Chimney, formed by semimassive
the surface in the most permeable parts sulphide mineralisation, generally
of the orebody (within the massive without base me t a l s , and lacking a clear
sulphide lens). contact with the overlying massive
sulphides. In this transition zone into
stratiform massive sulphides, at the
top of the pyrite chimney, a strong
This deposit is located at the eastern concentration of copper occurs.
end of the Riotinto anticline, where - Three Dimensional stockwork is
the anticlinal axis closes and plunges poorly developed here, and although
eastwards at 30-35 degrees beneath the lead-zinc sulphides are dominant over
Culm (Figure 2). copper, overal~ base metal contents
The same types of mineralisation, are low. The volcanic host rocks
as in San Dionisio, occur although some show less alteration than in San
differences in overall size and mineral Dionisio.
paragenesis patterns are seen. - Sericite Envelope, is fairly well
developed with several small minera-
Stockwork lised zones containing sphalerite and
The Planes stockwork, down to the ~s in San Dionisio, both the three
limits of evaluation by drilling dimensiona l stockwork, and the sericite
(Figure 5), is of reduced size and envelope, show a clear contact with
somewhat different characteristics to the overlying massive sulphides with
that of San Dionisio. Its structure no evidence for s t.ockv-or-k veins cutting
in terms of mineral paragenesis, a n d the ma s sd,ve sulphides.
relationship with the massive sulphides,
is shown in Figure 5.

w E


&~'-"':..~ Culm c==J Mass ive Sulphi de 1::::::::1 Ba sic Comp lczx ~ Stockwork
1 ~ / ~ : ." 1 Hor izo n Tr an si t i on ~ A ci d Complczx ~ P y r it ic - pipe

G arC i a PQt,),... crc r 19 83

ha ssi ve 3u1i:>hi de a suggest that the sulphides originate d
from th e pyrite chimney Loc a t.ecl , to tihe
The massive sulphiu~s in this deposit we st., upslope from the 3an An t o n i o l e ns
are planar in form, wi t h average (Garci a Palomero F, 19 80)5.
thicknesses of 10-15 metres. They
contain enhanceu b r ade s of co~per, as Cerro Colorado/Filon No r t e / Fi l o n Su r
wel l as several mi n o r metals such as ~ ~ ,
Au , As , and 3n . They appear to form This deposit, locate d a t the centre
t ~.O separate len ses 1d th some LsoLa't.e d of the Riotinto anticline, to the east
l i n k s between the b ,o (Fi bure 5). The of the Zduardo fault (Fi gure 2), is the
l e n s directl y overl ying- the stock work, most extensive mineralised zone in the
known as the Planes d eposit, has been area. It consists of s mall remnants
almost tota.Li.y wo r-k e d out. The other of ma s s i v e sulphide (Filon Norte and
lens, known as the .:Ja n An t o n i o deposit, Filon Su r ) connected by a large oxide
occurs dov.n dd.p , further a wa y from the cap and extensive at.ockvor-k mineralisa-
at.o ckwor-k , It VIa s di s c o v e r e d in the tion in the underlying volcanics. Fi gu-
1960 I S and remains une xpLo Lt.e d , re 6 ahows a NE- 3,-' s e c t i o n through the
The Planes lens, occurs directly deposit, indicating the relationship
above the p yrite chimney and shows no between mineralisation t ypes an d their
clear contact ,d t h it. It consists hostrocks.
mainly of p yrit e ,:i t h a bu n da n t chalco- The ma i n g e ol ogi c al f e atur es of thi s
.:> yrite. Apa r-t, from the mas s i v e sulphides zone are describ ed be.l ow, alt h o u g h it
the transition series is onl y represented s h o ul d be noted that t h e "!ide s p re a d
b y only 2-5 me t re s of tuffs. effects of w e a t her i n~ , a nd min i n g
The 3 a n An t o n i o len s occurs outside a c t i v i ty . deve loped in t his a rea mak e
the limits of the s t o ckwo rk in the it di fficul t to s tu dy s ome o f th e mo s t
underl ying roc ks, ,dt h a g r a d u a l develop- i mport ~ n t Ge o l ogi c a l asy c c t s .
me n t of a thick Tr ansition 3eries
bet.we en it and t h e Culm e a s c war-ds , An
increase in t y picaL s lwuping textures,
new mineral par a geneses, and additional :.s ah o wn in Figure 6, the stocl.:1: or' :
lithologies wi t h i n the pyroclastic .,lineralisation is very extensi v e , e:d~ 3 n
rocks, are a l s o seen going eastwards. ding over lOOOm from north to s o u th,
Al t h o ug h the avera;;;;e lens thickness an d 2000m from east to ~' e s t . It i s
i s Lo wer- than usual at 10-15 me t.r-e s , de v e l o p e d over a vertical thickn e s s of
b .o parallel e a a t.we sf zones show 400-500m, affecting the t09 of the Ba s i c
sulphide thicknesses of up to 5U metres. volcanics and all th e Ac i. d Complex.
' ;i t hi n these zone s a g r e a t variety of .~ g e o l o g i c a l r econ struction of the
sep a r a t e sulphide l e n se s , as well as unero ded remnant s define s a s imil a r
intercalate d p yroclastic horizons, are s t r u c t u r e to the s t o ck"o rks a Lr-e a d y
r eco gnised. These include compact described, ~ :i t h the p re s e n c e of a p y r i te
pol ymeta l lic massi v e sulphides, banded- chimney, t hree climen s ion a l stoc!,,:or ~( in
br ecci ated len s es, p yritic lenses, and chloriti sed volc anic s, a n d a s e r i c i t i c
rh yolite-sulphi d e breccia horizons. In envelop e. ;\ b e n e r a 1. de s c r i p t i o n of the
the s e t 1,'0 zones the l e n s e s and mineral s'cock wor-k r-a Ld a a t. Lo n i s not rep e atec:,
p a r-a g e n e s e s s now different l a t e r al and but some charact eri stics p articul ar to
vertical distriqutions, suggesting that this zone a r e l ist e d be.Lov i
the y occup y t.wo troughs, or parallel wi ele spatial e xtent,
depressions, predating the sulphide l ol'l intensity of mi neral i s a t i o n in
formation. Th e s e depressions we r-e the three dimen sional at.oclcv-or-k (4- 6 ~~),
fi l led wi t h sylphide an d p yroclastic l o c al i s a t i o n a t t h e ed:;;e of a pos sihl e
material, d i s pla ced from hi gher rhyolite d ome , inferr ed f'r-o m the
elevations, during movemerrt s of the thinnins of the a c i d volc anics from
sea f loor c aused by seismic or volcanic north to south,
,Ji despread disseminated copper
act ivity.
The tlistribution of the parageneses mineralisation,
1'!i d e s p r e a d remobili sed copper miner a-
ahown in Figure 5, together with the
lisation due to its structural position
changes in the nature of the sulphides
at an anticlinal crest,
and p yroclastics g o i ng e astwards,

S.W. N.E .

~oo Jollie

Ch lor it ic

Cu lm
Tra n si ti o n

;~:,t~ ~F
asi~-=,l ~

_. '" .""'-.,
. . T i " _

-+"I'--.=;-+_--,~_ I

........ ,
Trans it ion u -'"

ho r i zo n Z a::
. ")
~ Stockwork boundary'-..
VI I -- .


Show ing structure ot the St ockwo rk and Ma s s ive Su l p hide

- - - ---- Eroded a nd min e d

mater ia ls
Garcia Polom c ro F 1983

we a t h e r i n g wh.i. c h postdates the and lead in the massive iron oxide cap,
deposition of the overlying massive the four sulphide minerals (pyrite,
sulphides. chalcopyrite, sphalerite and galena),
The absence of visible magnetite at probably occurred in the massive
depth, and the abundance of rhyolite sulphide overlying the three dimensional
breccias and autobreccias confirms the at.ockwor-k , which is now totally oxidised
proximity to an active volcanic centre. or eroded away,.
Filon Sur shows a lens of
i-iassi ve Sulphi des lead/zinc-rich copper-poor massive
sulphide near the limits of stockwork
As mentioned above, almost all (7U-9U~) mineralisation, overlying the sericite
of the massive sulphides in this zone envelope, similar to the other massive
have either been eroded a way , or altered sulphide lenses in the area.
by weathering to form a very large iron Despite the lack of information in
oxide cap (Figure 6). Remnant some areas, the g e o l og i c a l characteris-
unweat.ner-e d massive sulphides are tics of Cerro Colorado, PlaneS/San
preserved on the south flank of the Antonio, and San Dionisio, show a
main anticline (Filon Sur), and in the distribution of sulphide mineralisation,
axes of subsidiary synclines on the and alteration patterns, etc. wh i c h
north flank (Filon Norte). indicate similar genetic processes
Only a small amount of information and conditions of formation. These
is available from the remaining genetic factors can be combined to
massive sulphides since they have form a single genetic model for all
almost entirely been removed by mining. the Riotinto SUlphide deposits, as
Available information shows several described in the next section.
parallels with the other massive
sulphide deposits, for example massive Gossan (Iron Uxide Cap)
copper-rich lead/zinc-poor massive
pyrite occuring on top of a pyrite Surface oxidation of both the massive
chimnej in the old Filon Norte pit. sulphides and stockwork sulphide
Dased on the abundance of b ar-Lum lIii n e r a l i s a t i o n , in all three mineralisec

zones, has led to the forl.ICltion of an Lower- zone of paler we anher-Lng rocks
iron oxide ca.:> up to 70 metres thick. in Kh i c h the more soluble metals are
This iron oxide c ap material is normal. l,l' enriched (silver and copper)
r-ef er-r-e d to a .s the Go s a an or-ebo dy , .3a n
;)ionisio and Planes /3an An cond,o ar-e not G 3N3TI C hOD3L
affected to a g r e a t extent becau se of
their location on the f'Lank s of the The historic development of ideas on
main anticline. By contrast, in the the genesis of sulphide deposits
,: ides")read near surface sul~)hice associated \'lith acid volcanic rocks is
mineralisation in Cerro Coloraclo, an not discussed in detail here. 11illiarr.s
estimated 70-9G~ of the ori~inal (1934)6, one of the earliest workers to
stratiform massive sulphides have been publish ideas on the g e n e s i s of the
oxidised. .~ total of bet.we en HAl and deposits, proposed an epibenetic origin
150' million tonnes of oxidised ma t er-Ls.l , for the deposits vrhdch wa s subsequently
has been formed from the weathering of revised in later publications (Williams
400-500 million tonnes of sul~hide et aI, 1975)7. According to this
bearing material. genetic theory the at.ockwor-k ".'as formed
The pr-o por-t.Lon of at.ockwor-k su.Lphd de s by the crystallisation of sulphides in
affected by oxidation is much less, and the charme.Lway s of a fumarolic system.
the ~rocesses involved in their Given the close spatial and g e n e t i c
,,;e a t h e r i n g are less we Ll, know. The relationship b ebv-een volcanism and
presence within the stockwork zone of sulphide mineralisation in the area, a
an extensive layer of secondary detailed genetic model can be defined.
sulphide enrichment, thought to be This genetic model relates the different
deri ved from the we a 't.h e r-Ln g of the nd.ner-a LLaat.Lon types in the three
overlying massive sulphides, indicate deposits described above, and takes
the importance of these weathering into account the following factors
processes in enhancing the economic (Garcia Palomero F, 1980 and 1982)5,8;
value of the original mineralisation. the g e ol o g i c a l characteristics of
Figure 2 indicates the wide extent the three mineralised deposits,
of the Go.:>san derived from the weathe- the relationship bet.ween the
ring of the massive sulphides. The mineralisation types and their
affects of F e a t h e r i ng on the Cerro host rocks,
Colorado deposit, are shown in the the relationship b et.ween different
transverse section presented in Figure ~ mineralisation types in each deposit,
30me of the main affects of the distribution of the mineral para-
weathering on the sulphide deposits geneses,
are listed below: the relationships b e t.we en the
oxidation of sulphide minerals to mineralisation and alteration in
fONu iron oxides and hydroxides the volcanic hostrocks
- total loss of sulphur, zinc, arsenic, Several aspects of this g e n e t i c
etc. model have been confirmed by g e o c h e mi c a l
loss of copper-, p art of whd c h is
studies. The genetic processes which
precipitated in a zone of secondary
formed the Riotinto deposits are thought
enrichment, the rest being lost in
to have evolved through the follo \,'ing
surface meteoric wa t er-
partial loss of silver (80%), some of
1) Submarine volcanic activity leading
whd ch is deposited in the secondary
to the development of rhyolitic volcanic
enrichment zone, the rest being lost
domes, and associated submarine topogra-
in surface meteoric wat-er-
phic variations. The abundance of
residual enrichment of the least
oxidation products sug gests that t his
soluble elements, including gold,
bar-Lurn, lead, and tin, some of which
volcanic activity occurred at sha.l Lo v
are enriched up to 5 times their depth. Formation of troughs parallel
levels in the massive sulphides to the Hercynian fold axes we r-e probably
T:le end. pr-o duc t, of sulphide rel ated to early tectonic events
,,: eathering is a Gossan deposit consis- ( 3trauss, 1970)9.
ting of an upper zone of massive iron 2) ~ :an i ng of volcanic activity wa s
oxic1es and hydroxides containing accompanied by strong fumarolic activity
enriched l e v e l s of the least soluble in areas on the edges of the volanic
metals (in particular gold), and a centres, 1'l h e r e the acid volcanic rocks

S .W. N.E.
. AREA'-:----'----- '


Acid Volcon ic Rocks ~ Gosson Seccn dcr y Sutphide ' kE<::>~J WoStll Dumps
(high in Cu, Ag, Au) 30m. thick
. ! '
1~~~~4 Basic Volc~nic Rocks . ~ Go sscn from Massivll SUlphidlll~1 Stockwork Minllroliza tion

V./ ;:.:;:/:,:.:::j Massivll Sulphidll ' ~Slatlls Silver rich we c ther ed zo nes

I I W~athllring Boundary 1:;.:::.:1 Anc ie nt Slogs

Gorcia Pa lamt'fo F e t 01 "86

are thinnest. This activity led to dropped causing some of the metals to
concentration of sulphur and metallic be precipitated in fractures in the
elements which precipitated in favoura- volcanic rocks giving rise to the
ble conditions giving rise to the stockwork mineralisation. This occur-r-ed
sulphide mineralisation. Conditions under the following conditions:
necessary for precipitation were created Precipitation started 40G-SOum depth
by a hydrothermal convention cell driven bel 0',' the seafloor l-:ithin the volcanic
by heat generated in magmatic bodies pile. The amount and intensity of
underlying the volcanic centres. The mineralisation increases upv-ar-d s from
fluids involved wer-e mainly phreatic this point., l! ith the possible addition The three different types of
of SOllie fluids of magmatic origin. The suoctcwor-k mi n e r a l i s a t i o n (pyrite
large amount of sulphides accumulated chimney, three ddrnen sional s t.oclo-or-lc
in the area suggest that sea,rater Ha s and sericitic envelope) appear ~o
the main fluid involved in the evolve out.war-ds from cylindrical cores.
convection cells, given that volcaaic These units, defined by chang es in
activity at this point wa s very we a k intensi ty of mineralisation, nri n e r-a L
and shortlived. Similar convection paragenesis, and ~al l r o ck a lteration,
cells have been suggested as the are t h o u g h t to be relatetl to the
formation mechanism for some Canadian coo.LLng of fluids as they mo v e d upvar- d s
de posits (Ridler 1973)10. a n d outwards from the hydr-o tihe r-maL
The sulphur and metals are thought centres. The pyrite chimneys are
to have a magmatic origin, since the thought to represent centres of max i.mum
large quantity of metals present are fluid movement.
unlikely to have been derived from the As well as the overall zoning of
thin volcanic sequence. This is the mineral paragenesis, a sequence of
supported by studies of the isotopic formation of different minerals can
composition of the sulphur. also be recognised (chlorite, ma Gnetit e,
3) As the metal bearing fluids ascended pyrite-chalcopyrite, :Jyri te-chalco~)yrita-
t-o war-ds the seafloor, temperatures sphalerite-galena-sericite, barite).

These features s Ut:>b e s t that at the start values of around -15;;., both in the
of h ydrothermal acti vit y the stock,'!ork at.ookwor-k and at the extremities of
va s an open system, ,,'ith sul;>hur an d me- the stratiform sulphi des, as possibly
tals deposited on the seafloor. In the relating to biogenic sulphur.
lat e r stages parts of the s.fstem The values of 34 5 found in the
continued to be open to ascending f luids stratiform sulphides contrast strongly
(the p yrite c ey s }, the remainder with those found in the stockqork,
consisting of ti bht fr actures wh i c h we r e wh e r e typical values in the range +6 %0
fi lled wi t h late stage mi n e r a l i s a t i on to +125~owere found by Ea s t o e et al l l ,
(sphal erite-galena-chalcopyrite-barite and +4.9 %oto +12.4 %uby I'iisuno et al 13
in the three dimensional atiockwor-k }, The These results sug gest derivation from a
s t.cckwor-k seen in Planes/San Antonio ",'as mixing of cold sea wa t e r and ascending
mainl y an open system, ,d t h most of the hydrothermal fluids, resulting in an
metals reaching the seafloor to form isotopically heavier product.
sedimentary massive sulphides, rather During the deposition of the
than being precipitated in at.ockwor-k stratiform sulphides, some movement of
veins. the seafloor occurred, associated v:i th
From the distribution of mineral volcanic activity. This caused slumping
paragenesis ahown in Figure 4, formation of sulphides and pyroclastic sediments
temperatures of 350-40U oC at the roots downslope, g i v ing 'r i s e to mixed
of the at.ookwor-k (quartz-chlorite-magne- sedimentary breccias, a s seen in some
tite), and around 1000C near the parts of the San An t on i o deposit.
seafloor (sphalerite-galena-barite- The spatial distribution of the
sericite-etc). ar e indicated (Garcia stockwork and the massive sulphides
Pulomero F, 1980)~ This is supported in the area a h ow s an asymmetry in
by g e o c h e mi c a l studies wh i c h indicate each deposit, sug gesting the presence
a similar range of formation temperatu- of a we .Ll, defined topography, wi t h
res, deDendent on the location of the deposi tion of massive do wn s .Lope from
samples used (Bastoe CJ, et a l., 1986) areas close to the volcanic centres.
Ox y g e n isotope studies indicate forma- 6) Deposition of the Culm sediments,
tion temperatures of 210-230 0C in the follo wed by the HerJcnian tectonism.
Cerro Colorado stocJ(\~ork (l-mnha H, et 7) Uplift and erosion during Tertiary
a l 1986 )12. times, leading to the exposure of the
4) Su l p h i d e de;>osition occurred on the volcanic rocks. Subsequent we at.her-Lng
seafloor in depressions close to the of the sulphide mineralisation led to
stock\wrk zones, ' d th sulphur derived the formation of the extensive gossan
from sulphates in the seawater, tempe- mineralisation.
rature g r a di e n t s , and thermal zonation, During weat.her-Lng oxidation of
e xi sting in these depressions at the sulphides occurred in both the massive
start of sulphide precipitation, led
sulphides and stockwork mineralisation.
to mineral and wetal zonations as shown
In both cases the more soluble elements
in Fi gures 3 and 5. Taking the pyrite
such as copper, zinc and silver we r e
chimneys as the centre of maximum
leached out and redeposited at depth
temperature, the paragenesis changes
in a zone of secondary enrichment
outwards in the follo wing sequence:
(70-100 metres below the surface), or
p y r i t e - c h al c opyr i t e , pyrite-chalcopyrite-
removed in surface water runoff. Less
sphalerite-galena, pyrite-chalcopyrite-
soluble element s such as tin, lead,
shpalerite-galena-barite, pyrite- .
barium, and g ol d remained in the oxide
cap wh e r e residual enrichment occurred.
calcium carbonate, jasper. At a later
In the case of g o l d the resulting
stage an accumulation of pyrite wi t h
enrichment reaches economically
Lo w contents of other metals was
exploitable g r a de s .
deposited, formin g a cap on top of
the other zones.
An average isotopic composition of BCONOHI C GEOLOGY AND lUNING
the stratiform sulphides of between
-6 %0 and +8 %034S was determined by Alt h ou g h the economic geology of mineral
Sa s t o e et al (1986)11, and -14.1i~to deposits is closely linked to mining and
+8.0560by Ilisuno et al (1986)13. Sastoe mineral processing activity, because
et al explain several isolated negative mining activity at Riotinto commenced

before there wa s much geological l a s t 110 yearsl
knowledge of the deposits, these two In 1970 the gossan derived from the
aspects are discussed separately below. massive au.Lph.i de s wa s again wo r-k ed,
this time for the extraction of g ol d
l'iining and silver. The initial production rate
of 1 million tonnes p er- annwu h as
The earliest known mining activity on g r a du a l l y increased, due to increases
these deposits occurred in pre-Roman in precious metal prices, and the
times, with a peak of activity in the definition of extensive a d di t i on al
period 300-100 DC (Rothenberg et a I, reserves. At present ~bout 6 million
1986)1. At this stage mining activity tonnes p e r year is treated, avera~ing
concentrated on zo n e s of secondary about 19/t gold and 56g/t silver.
mineralisation, at the base of the l,ost recent mining acti vi ty has
Gossan cap, whd ch contained high grades been by open ~J i t mining methods. The
of silver, copper and gold. The exten- increase in reserves, along wi t h
sive nature of these workings is improved mine de sing a n d production
evidenced by the large quantity of control methods, required to support
slags produced by treatment of the ores these large operations is due largely
(6-8 million tonnes in total). Known to the development of geological
vorkings from this period are concen- kno wledge of the deposits, as we .lL as
trated in the Filon 3ur and Filon Norte to n ew exploration techniques and
deposits. evaluation methods.
The next main phase of mining
activity commenced during the last Economic Geology
century, with the exploitation of the
main stratiform sulphide bodies Since the start of mining activity
containing high copper grades. Copper in the area, geologica~ criteria used
was produced either by directly leaching in the exploration for new reserves
the sulphides, or after pretreatment by concentrated on the search for
calc ination continuations of the massive sulphides.
\.t the turn of the century, the drop This focussed attention on the horizon
in copper g r a de s of the massive marked by the contact bebween the
sulphides, and their hi gh sulphur content slaty rocks of the Culm a n d underlying
(49-51 %), led to the initi ation of their porphyritic textured rocks (volcanics),
eXDloi~tim for sulphuric acid production without being a war-e of their volcano-
Ka ~ initiat ed, ~.-ith subsequent extraction sedimentary origin.
of copper from roaster slag s outside the In the 1960's, wi t h the acceptance
mine area. At the same time working of of a vo.Lcano-c s origin for
copper-rich s t.o ckwor-k zones was possible these sulphides (Rambeud F, 19 69) 1 '"', "
due to the development of sulphide and an increased a v areness of their
fl ot a t i on methods of ore treatment, ,dth regional stratigraphic setting, the
copp er being produced from the concentra- first r e gional geological/geophysical
te s a t a smelter in the mine area. exploration IJr o b r a liimes we r-e commenced
'I'he se t.wo typeS of production, namely in the a r e a . Such exploration v o r-k
p yrite for sul~Jhric acid production, and concentrated on the kno w host horizon
cop p er fro m sulphide stock,:ork zones, of the s u I phi des \': hich had been l~.efined
continue to the present-day, despite by earlier rebion al beologica l stu di e s.
the ups and do wn s of the market for the Durine; the 1970's and 1980' s, as
t,,'o e n d pro ducts. Overall the production a result of the detailed stu~es of
of copper from s t.ockwor-k zones has the paragenesis of the sulphides an d
incr ea se d ,.,,' i th time, as new reserves the we a t.h er-Lng processes wh d ch had
h ave been discovered, a n d ore with affected them, explor ation li a s
Lo wer- head- 6r <l (~ ~s has been worked. The concentrated on the ful l potencial of
amount of ore tre ated rose from 350 000
a ll the economic metals present (Cu,
tonnes per annum in the 1960's to Pb, Zn, Au, Ag), resulting in a large
S OCO 000 tonnes per year in 1985. The increase of know mineral reserves. .d:.
a mount treated has dropped a gain in the saine t inl ,; n e v metho ds of r eserve
more recent times, due to changinc;;
e v a l u a ti on ...e r- e implemented, based on
economic circwnstances. Production of
a n inte.:;rated system of computerised
pyrite for sulphuric acid production
r eserve estimation and mine plannin,;
has been at a more r-e g u.l ar- rate of
:Jr O;;r aIilS, a1.1 0' ,i n ,;; t he selection of
about 1 000 0 00 tonn es p er year for t h ~

the most appropriate mine design and (Pryor RN, Rhoden N, and Villalon l'~,
scheduling creteria for maximising the 1972) 15.
economic benefit from exploiting these In the mid seventies, as a result of
reserves. the synthesis of the available geological
The computerised reserve eva.Lua t .Lon studies' into an overall interpretation of
techniques used are based on a kno~ledge the form and origin of the mineralisation,
of the paragenesis and grade distribution a new hypothesis was established. The
of each element studied, in order to copper concentrations within the
define the geological/statistical suockwor-k were thus thought to be
divisions used as a basis for subsequent preferentially located in zones semipa-
geostatistical analysis. These techni- rallel ,d th the volcanic stratigraphy,
ques allow all the parameters or assay ,'lith l e s s e r amounts in vertical zones
grades available for each sample, in v.h d ch link the others, as ahown in
this case mostly derived from diamond Figure 8 (Garda PalomeroF, 1983 )8.
drillhole core (sampled on standard 2 3ubsequent exploration wor-k on the San
metre lengths) to be integrated into Dionisio and Cerro Colorado stockwork
the final reserve models. mineralisation wa s based on this
The capacity of computer processing hypothesis, and resulted in a spectacu-
methods for handling large quantities lar increase in kno~n reserves betveen
of data, and considering different 1977 and 1980. Reserve evaluation work
evaluation alternatives, is a fundamen- was based on similar geological thinking,
tal improvement on previously used which was incorporated into the geoesta-
techniques. Totally manual approaches tistical estimation methods implemented
wou.l d not all 0" ; detailed study of all on a Perkin-Elmer 3210 computer on site
the available geological and geochemical in 1983.
information derived from a total of The geoestatistically bases reserve
about 160 000 metres of drilling carried estimation work carried out on Cerro
out at Rio Tinto, from v-h d.c h about 80 000 Co Lor-a do and 3an Dionisio follo" 'ed a
samples have been taken, and analysed for similar approach, as summarised bel 0" :
six different elements in most cases (Garcia Palonero F et aI, 1987, Sides,
(Cu, 3 , Pb, Zn, Au, Ag , etc.). in pr ep , ) 16,17.
Geological studies carried out on 1. Database preparation.
the different deposits for reserve Drillhole data files are prepared for
evaluation purposes are described belo",: available orillholes, containing the
spatial ~ocation of each sample, along
with their geological characteristics
and available assay data for S, Cu, Pb,
~ }i t h i n the three stocln:ork deposits at Zn, etc.
~otinto, attention has concentrated 2. Geological interpretation.
on the search for zones of copper The spatial distribution of the main
enrichment ,; ithin the overall stock',ork lithological units, alteration types,
mineralisation. Only in the cases of and mineralised zones are defined on
3an Dionisio and Cerro Colorado have sections and plans.
areas ,d th economically viable copper As part of the g e ol ogi c a l mo~ellin G,
g r a d e s been encountered. different populations of sample grades
Initial explor ation for such are defined for each element studied,
mineralisation wa s carried out in the on the basis of statistical studies.
a b s e n c e of an overall geological In the case of the Cerro Colorado,
f'r-ame wor-k , being based on kIW\'I' mineral three different copper grade zones
occurrences intersected by development were defined as indicated in Figure 9.
carried out for mine infrastructure. In the case of the San Dionisio
During the mid sixties systematic stock~ork a similar approach is u s e d ,

studies we r e started, both in the but the basis for defining the zones is
exploration for further reserves, and slightly different.
also mine planning studies. This wo r-k 3. Experimental variograms.
wa s based on the hypothesis that the Samples are classified according to
copper mineralisation , localised in the zonal divisions defined spatially
the form of chimneys, or vertical by geological interpretations on
structures, ~hich cut across the sections and plans. Experimental
geological structure, and which alterna- variograms are then calculated for the
ted with other barren structures three principal co-ordinate axes,


~ I till SlatlZS" k :c;\;:;:, q Massi VIZ Sutph i de DO SlZricit ic AltlZration

~ G'o ssa n ( i n situ) I.::.1 MinlZd MassivlZ Sul p hid e ~ Ch lor it ic Altuat ion

~ (ero de d IminlZ) CD Acid Volcanic Rocks ~ Pr imary Cop per Zo ne s

CQ t.e c che d Ac id Volcan ics CD Bas ic Volca nic Roc ks ~ Se con dary Cop per ZonlZs
G c r c:i Pclom ,t o F tl ol 198 8

S. N.

"."' ..... _ 2

10 0 zoo Mis.


GradlZ Zo n e s
< 0.10 , . Cu
O.lO ~ 0.38 '. Cu

0 .38 '. Cu

~ SlZcondary CO pplZ r ZOnlZ

c==J P r i mar y Go rcie PQI~m e to F. ct e t 19e..8

seluratelj" for each zone, in order to reserve estimation to be r-e vd s e d and
study the autocorrelation of the metal improved. This is illustrate C: in the
being studied (co~per Gr a d e ) in each case of Cerro Colorado wh e r e block
axis direction, ~i th i n each zone. g rade s estimated from ori ginal explora-
4. illock Iilodel. Par ameter s of the tion holes on a g r i d of 50m by 50m a r e
reserve block models, or matrices, are used for long range mine planning
selected based on the nature of the (Primary Reserves). For short to m e d i.uru
deposit being- studied, and the Jnj.ning range planning the reserve model s ar e
methods to be investigated. In the case updated using information obtained f'r-orn
of Cerro Colorado b lock sizes of 25m in-pit percussion drilling carried out
ea sb-we s't , 20m north- south, and 10-12m on a closer spaced g r i d of ZOm by 25m,
vertical ly, are used. The co-ordinates in zones of int erest (Investigation
of each block are defined from its R e aer-ve s }, The estimation models used
position wi t h i n an array of blocks, and in this instance are selected after
detai l study of the relationship
each block is a s sd.grie d codes de f'Lnd.n.,
b e t. we en the e s t.Ln.a t e d reserves and
the main materia l ty~)es of interest
actual production in adjacent areas.
(rock type, alteration, grade zone,
Af t e r 16 year s of oper ation f'r-on, 197G
et.c , }, by digitis:Lng in the geological
interpretations made on section or plan, to 1986, the rem aining re serve " for t. he
Cerro Colorado op en-pit operation ar e
so as to build up a complete three
1 80 mil lion tonne s aver a ging 0 . 4 3 ~
dimensional model of the deposit.
cO.-Jper. The res erves of s t.oc k wor-k
5. Variogram models. Variogran models
tuinerulisation in 3a n Dionisio a re
are fitted to the e x perimental vario-
currently bein o r e- evaluated.
grams, defined by parameters such as
the ranGe, nugget variance, and sill
Gossan (Iron ~xi de Cap)
variance. Separate models are used
for each different g r a d e zone, and
The U S3 of &; e ol o c,;,i cal criteria in t h e
rao de.Ls are tested statistically by
ex )~ o r a t i o n and evaluation of reserve s
lcr i g i n g the g r a d e s of known sample
of pr ecious metals assoclated wi t h t h i s
:Joints usinG surroWlding samples (cross-
t ype of mineralisation, are describe ~
validation point :, r igi n.:;; ) , and carrying i n ~ 3 t ai l elsewhere (Garcia Palomero
out statistical cOffi?arison between the et a L , 1986) 1 8. . _ b r i e f description i..a t.ed and known g r a d e s , as of the evaluation of the different
illustrated in Fig. 10. types of gold-silv er b e aring gossanised
6 . ;~ i g i n g of block g r a d e s . The rocks i s b i ven b e L o v,
selected variogralli models are used to .\ recent exploration campaig-n for
e s t i ma t e grades for e a c h block in the Go s s a n reserves, based on the geologic al
matrix by kribint,;. UnJ..,)' from reconstruction of the sulphide minerali-
the same are u sed in estimating s a t i o n before ,:eathering commenced,
the grade of a bJ..ock ~ i t h i n it, the proved very successful. This wa s based
number of samples used being restricted on comparing the reconstructed pre-
t o tho se fal.Linb inside a search prism e r o s i o n g eo Lo g y h'i t h the situation
c entre d on the b lock. ....h e n mini gn activity commenced in th e
7. Pl anning. l' h ~ b .iock models are a (Figure 7) an d the present- c a y
cont ainin t,; a Ll the v ari ables estimated, 5i tuation wh e r- e mo s t of the zon es of
or assi gned to .t h e lil, f o r m the basic for v-e a t.her-e d miner a li sation had b een
studi e 3 of th e d.e s i !bn of future mining a f f e c t e d by e x ten s i ve mining a c t i v i ty .
uni ts, a s \:eli. a s production contro.L Th e g ol d - s il v e r cont e nt of the
\.: i thin exd s'tLn g o per ations. we a t h e r e d zone wa s estimate d b y
T h e pro cedure of reserve evaluation e x t r a p ol a t i o n from the g ol d - s i l v e r
used a L so a 110\: s the r eserve mo deLs to c o n ten t s of the kno ;:n sulphides, t.a kd ng
be updated u sinb r e sults obtained into account the a mo u n t of or e alrea cl.r
during exploitation of the deposits. ext r a c t e d . On this b asic, an overal l
Re c o n c i l i a t i o n studies are carried unex~loited pot ential of 100-20G tonnes

out to determine the predictive of c,;,old anc. 10uuu - 1 5000 tonnes of

a c c u r a c y of the methods used, both in silver wa s estim ated in 1984. Thi s
terms of compar-Lson of predicted and potencial form ed the basis for the
a c t ua l production figures (grades and e x p Lo r-a t.Lon c amp a Lgn c on.meric e d at
tonna ges), as \,'e.1 1 as the prediction that date, v-h.i ch concentr ated on
of ore/waste boundaries, a110\-:1ng the looking for both in-situ areas of

- , 2- 5 - &- 25 - > 25 POM S


II _

' .n . .'" ,. .. . .11
.~ l: " t .'" II . I. CIlII
srATrs':'tcs ~

c:- I.n c .. ... . u

/10. f' U U. D .. 3' 1.1 IC'. (l' T OUU 3""
IAIO lol a ... 110o'I 1.7 '9
.,~ ' .9 11'2 ,.o.uucz 1.0 712
naJ.lO~ r~. \ .38 P..c...<.'Dlrt"'..Jo ':"'"...:PI- '. 0 3
-...a: '" l'.It.O.O 0 .01 .9.96 A.uc;(f' T...,;,,;..cs: O.0 5 - 6 . 94
CG'ftCl"~ .' .S4 ~:I" ~ - 1. 20
CJC"I;IIC OJ' IU"Tl:IIIUII; 6 .10 CR": 0' ~ 1.,8 2
ou ~ 1'7'J .' .1 1074 , 0 DC ~ ITu.J &0 7.1

~9'l.9 15.



M~:'N D IFi' e;; <: N C ~ IN VALUES - -0 .010
AXIS 1 2 3
t~UGGET VAR. 0 .6 0 .6 0.6
SILL VAR. 1. 3 1.3 1. 3
Mi: AN ERROR OF ESTIl'A TIO~~ 7 0 . 0::
I RANE~ SO 25 so
HEA:: KRrGrNG VARIA I1Ce -K - 0 . e 30
5E,!.,;:;CH R.!.O II 90 25 50 EXF\.OR'CIo.~ MTNE"
RATrO ElK 0 . e9
Mill . 4 lU X. 14 VALUoS usa FOR KIHG,:'G

Gossan that had not yet been mined, and geological ideas on the behaviour of
also for potential reserves in waste these metals during weat.her-Lng j gold,
dwnps from earlier mining activity. due to its lower mobility, being
The evaluation campaign "as based concentrated in-situ being more affected
on geological mapping, drilling and by local geological controls; silver,
pitting aimed at determining the types being more mobile, is deposited by
of gossanised material, their in-situ cementation from groundwaters both
location in relation to sulphide in rocks whd ch previously contained
mineralisation, dWlp areas with gossan sulphides and also in barren waLLr-o ck s ,
potential, gold-silver grades, material The computerised reserve models
densities, metallur gical recovery, etc generated consisted of a block model,
This wo r k resulted in a database covering all the areas affected by
similar to that used for the Cerro g o s s a n i s a t i on and wi t h g o s s a n - bearing
Colorado copper evaluation, as descri- dump s , Evaluation p were
bed above. stored for each block (material type,
The reserve evaluation wa s car-r-Led grade zone classification, g ol d and
out using similar procedures to those silver g r a de s , density, etc.) Using
follo~ed for Cerro Colorado, except
the same sy st.em as for Cerro Colorado,
that in this case b ,'o economic metals, this a l l owe d the plotting of plans
g o l d and silver, wer-e studied. The an d sections, t.he se par ameter s,
result of the statistical and g eoesta- for use in mine planning and further
tistical analysis ahowed that wh.iLsf
geological studies.
gold behaved differently in each in-
The initial r esults of this ex~lor a
si tu zone, silver showed a similar
tion campaign we r e to outline identifi ed
pattern in all t ypes of in-situ gossan
resources (proven, probable and
( gossan derived f'r-om massive sulphides
possible reserves), at the start of 198 4,
and from stockwork, .; o s s a n i s e d shales
of 100 million tonnes averaging about
and g o s s a n i s e d pyroclastics). These
19/t Au, and 56 g/t Ag. The contained
g e o e s t a t i s t i c a l models of the g ol d and
precious metal content of these reser-
sil ver distribution support the
ves, l OOt of c o l d , and 5,600t of s il ver

cornp ar-e cl favour abl y wi t h the initial o f t h e s e me tal concentrations is
expl ora t i on t arget (l OOt Au , and s upp or ted by the experimental vario gr a ms
100 0 0-15 UO Ot Ag ) . The c lose ma t c h of and g e ost a t i s t i c a l mo del s use d. Vario-
the g old reflects its les ser mobility, Gram s ahow simil ar ran ge s in the
Fh e r e a s the defici t in silver reflect s v ertical an d ea st-" Jest di r e c t i on s ,
its g r e a t e r mobd.l. Lcy and pr-oba bLe losses corresponding ,~i th the pl ane of s t r a -
in a subgr-a de dispersion h alo us we .I l, tification, but have much shorter
as through r e moval in surface wat. e r- r a n ge s in the north- south direction,
runoff. whd ch correspon ds t o the stratigraphic-
Af t e r s e veral y ear s of mi n i ng using normal sense.
the n e w g e o s t a t i s t i c a l l y b ased reserve Th e in- situ g e ol ogi cal r e s erve s of
mo d el s as a basis for pl a nni ng , a very the Sa n Di on isi o ma s s i ve s u lphi de
g o od mat c h ha s been foun d bet"Jeen
d epo si t ",'e re evaluated u sing a block
estimat e d an d actua l production fi gures.
mode l. Thi s res erve mod e l is beinb
The c lo se match in p r e di c t e d g ol d an d
used a s a basi s for continuin~ studi e s
silver content, both in tonn a ges and
g ra d e s , as we.I L a s accurate s pa t i al n ecessary to define appropri ate mining
prediction, reflects i n p art the an d mi nara l beneficiation met h o d s for
homogeni sation of the me tal di stribu- e x p loitinG these mas sive sul phides.
tion s as a result of ,.'e a the r i ng The ~roject is currentl y in a p h a se
p r o c e sses . of det ailed me tall u r g i cal study v-hd c h
lfill determ i ne \!h e t h e r or n ot t h e s e
l'~a ssi v e Su lph ides sulphi de s can be econond c a L'l y e xploit e d
for their b a s e metal cont e n t s.
Hi s t or i cally , the ma ssive sulphide Th e tota l g e ol oci c a l re serves of
deposits exploited for sulphur, require the 3an Di on i s i o dep osit i n 19 8 6 wer-e e
l i t t l e or no g e o l ogi c al control during !;-51'it a v e r agi ng :
mining, once their overall l i nd ts, and 0 . 837~ c, 5 0 ;~ 3 ,
the nature of their contacts ~!i t h 0 .6 5% Pb , 2.1 4 ~ Zn
h o s t r o ck s , have been defined. However,
these d epo s i ts do al SO contain si g nifi- The current g e ol o g i cal r ese r v es o f the
cant contents o f base me tal s , 1 ~~ Cu, 1 5~ San Ant on i o ce ~ osi t , c a l c ul ated u sin G
? b , 2 ~ Zn in San Di on i s i o , a nd 1.6~ Cu, c l a s s ical g e ol oe:;i c a l me t.ho cls , a re :
1 ~~ P b an d 2% Zn in San Ant on i o . A 9.5ht averaging:
r econsi deration of these dep o s i t s in 1. 6% Cu, 1 ~ Pb, 2% 2n ,
t.he light of the recent studies of 60ppm Ag, 0 .6ppm Au
the p a r ag e n e t i c distribution of these
In order to i mprove the re s erve estima-
metals sug gested the possibility that
z o n e s of met al enriclunent, of potenci al t i on me t h o d s used in t h e s t rat i f o rm
economic interest for their base metal sulphi des, ",h i c h are gener ally f'o Lde c.
contents, might occur wi t hi n them. s u ch that the mai n co-cor- d dnat. e a xe a do
Dased on g e ol o g i c al i da a s of the not correspon d t o t h e s t r a t iGr a p h i c
ori~in and stratigr aphic control o f axes a t t h e t ir.le of depo sition, t h e
rue t.aL zonation p a t t e r n s , drilling ~ ossi b i l i t y of studyinG v 3r i o g r ams in

c a mp a .Lg n s "Je r e c arried out of San thepl a n 3 o f stratification an d

J i oni s i o and San :.nt on i o to test their 1, it;1in di f f e r e n t structural zo ne s i s
p oten t ial base me tal reserves. As a bei n ~ investi g ated. In mo r e compl ic a t ed
re sult of these campaign s z o nes , or foldee! situations lI unf ol c'.i n g ll proc e du-
len s e s, "Ji t h i n the massive sulphides, re s can be adopted, " :h :~ r e by the pr esent-
containing preferential concentrations day co-ordinat e s a r e tr an sformec into
of one or mor e b a se metal sulphides, s t r ati g r aphi c co-ordinates wh ar e
v-e r-e defi n e d (se e Figure 3 ) . use d du ri ng the g r a cb e stimat i o n p r o cess
Th a s e metal - r i c h zones h a ve b een ( 3i de s 3 , 198 4)19.
e Vul uated i n t he San Di o ni s i o d e p o s i t In g e n e r al the use o f computeri s e d
using t h e g e ostat i s t i c a l techni ques met h ods in the exploration a nd
al ready Lesc r i b e d a bove. In this case e v al uat i on of mi n e ral clep osi ts pr-ovd. d e s
ea~h o f the metal s o f p o t e n c ial econo- a n i n v a l u a ble tool. It i s i mport ant
I.dc inter est ( copper, lead, a n d zinc) to r e alise, however, that to obtain
,,'e re studied ".'Ii th separate gr a de zones a c c epta bl e r esults, ri gorous b e ol ogi cal
being define d for each. It is of creter i a must b e app.l.Led , A ma x d.mun,
int er e st t hut the s t r at i g r aph i c origin

amou nt of Geological detail, should concept a nev: tool for ex~lorat ion. The
also be used, since even i n app arently Northern Ianer, Ca n a cla ,
homogeneous geological zones, stratigra- 11. 2a s t o e CJ, .'3a l omo n 1'1, & Garci a
p h i c , p etrographic and structural cliscon- Palomero F (198 6). sulphur isotope
tinui ti es may clivide an area into study of massive and stock,;ork pyrit e
different spatial domains are d eposits a t ~io Ti n t o , Spai n . Trans .
best s t u cli e d separately. Inst. Ian. J.1et a l l. Sec f B. Vol 95,
12. l-amha J an d. 13arri ga FJAS (1986).
High 0 1 8 ore formin g flui ds in volc anic-
The assintance of all the geoloGical hosted b ase me t. a 1 n1u3::;i ve sulphide
tear:l at Rio Tinto hinera 3.1. is deposits. ~con. Geol. Vol 81.
a c kno wLedged, a Lorig ,'lith the support 13. IIIi tsuno Ch a nd cowor-ker- a (1986) .
received at <.\1 1 times from the company GeolOGical s t u d ie s of the Iberian
management. ':;pecial thanks are due to Pyrite De lt. Okayama Univ., Japan.
Prof. Antonio Ar r i b a s for his advice, 14. Rambau d F (1969). 31 sinclin al
a n d 3dmWld Sides for his teclmical and Carbonifero de Ilio Tinto y sus md.rre r-a-
lin b~istic assistance. l izaciones a s o c ia da s . He m. IGJ.i3 T.LXXI,
l-ia dr-Ld ,
11.3F:SR3NC Z3 15. Pryor nn, l1h o d cm !IN a n d Vill alon i.
i. RothenberG n, Garcia Palomoro F, (1972) .3a iiip l i n G of Cerro Color a do,
( 1986) The "Rio Tinto 3niCl;la - no more". Pdotinto, ':;~) ,::li n . Trans. En s c , Illin. i,ie tal .
IAllS No.8. London. 16. Garci a PaLome r-o F, l'lalave J,
i,ararifio h , SoboL F, 3i (le s S (1988)
2. Garcia Palomero F (1974) Caracteres
I\Jodelos Ge o1 6 Gi c o s par a La exp.Lor-ac I Sn
estratigra~icos del anticlinal de
Riotinto. StuUia Geologica VIII. y c~lculo de r cservas a l argo y corto
Salamanca. plazo, e n el ya c imien t o de Cobr e de
3. .schermerhorn LJ G (1971) An outline Cerro Colorado ( hin a s d e Iliotinto) .
stratigraphy of the Iberian Pyrite Dol. Geo1. y h i n . T. ::CI::-I, Ha dr ili .
Belt. Bolo Geol. y Hi n . LXXXII. Ha dr-Ld , 17. Si d e s ZJ (in pr ep). Computerise d
.;.. Read RA (1967) Deformation an d r eserve e s t ili;at i o n a t 3 antia Go an d
metamorphism of the Pyritic San Dionisio Rio Tinto, 3p a i n , a n d lJev e s - Co r v o , :'01'-
tugal. Ph D thesi s, Empc r-La L College,
orebody, Rio Tinto, Spain. PhD thesi s,
Royal 3chool of Ilines, Imperial College,
18. Garcia Pa F, Dedia Fe r-n an d e r
JL, Garcia h aga r i fio h , Side s 3J (19 86).
5. Garcia Palomero F (1980) Ca r a c t e r e s
Nuevas investi gaciones y trabajos de eY[~
estratigrAficos y relaciones lilorfologi-
luacion de reservas de gossan en llinas
cas y g e n e t i c a s de l as mineralizaciones
de Riotinto. Dol. Geol. y lun.
del anticlinal de Riotinto. ~xma. Dipu- T. XCVII-V, Ira dr-Ld ,
taci6n Provincial de lIuelva , 19. Sides 3 J (19 81;). Intern al['.l1:/
6. c'Jilliams D (1934) The g e o l o gy of report, l;.i O Tinto L in cre..
t he ru.o Tinto Ilines, Sp a i n . Trans. Inst.
L i n . a n d hetal. Vol. 71, London.
7. ' lilliams D, s t a n t on RL, Rambaud F
(1975) The Pl anes-San Antonio pyritic
dep osi t of Rio Tinto, Sp a i n : i t s
nature, environment and g e n e s i s . Inst.
of IUn. and l, e t al . Vol. 84, London.
8. Garcia Palomero F (1983) Geologia
d e las miner~lizaciones de Riotinto "j'
.:;u modelo Ce n e t i c o . Libro Homenaje a
Carlos rel gueroso, liacirid.
9. s t ra u s s G (1970) 30bre la geologI a
d e la provincia pirit!fera del 3uroes-
t o de la penInsula Iberica y d e sus
y a c i mi e n t o s , en especial sobr-e
d e pirita GO Lousal (Portugal). 1<"1301.
IG H3 -77, iie-cirid.
10. IlicUer RH (Nov. 1973) 3J:ild i t e

The massive sulphide deposit of Aznalc6l1ar, Spain, Iberian
Pyrite Belt: review of geology and mineralogy
losefina Sierra
Departamento de Crista/ograffa y Minera/ogfa, Facu/tad de cc. Geo/6gicas, Universidad
Comp/utense, Madrid, Spain

SYNOPSIS schistosity and partial obliteration of

Aznalcollar is a strabound deposit So planes.
located within the so-called Iberian Alteration is characterized by an
Pyrite Belt. The deposit has reserves early stage of silicification as a
of 44 x 10 6 t of pyrite ore grading consequence of submarine fumarolic
0.44 % Cu, 1.76 % Pb, 3.33 % Zn, activity. -This was followed in time by
6 the formation of new minerals (quartz,
67 g/t Ag, 1 g/t Au, and 34 x 10 t of
sulphides grading 0.58 % Cu, 0.40 % Zn plagioclase, chlorite, sericite) due to
and 10 g/t Ag, the latter belonging to hydrothermal alteration and
the so-called "cupriferous pyroclastic". devitrification. The massive
The pyrite-Cu-Zn-Pb orebodies strike mineralization is mostly formed by
E-W and dip 45 Q towards the north. The pyrite with minor sphalerite, galena,
pit is 1,500 m long, 700 m wide and has chalcopyrite, and lead sulfosalts. A 3D
a depth of 300 m. statistical analysis of 576 samples
The orebodies occur within a lower taken from the mine showed no zonality
Carboniferous volcano-sedimentary unit of metal distribution and a reverse
named "Complejo Volc'nico Sedimentario", behaviour of Cu versus the pair Pb-Zn.
in which the volcanic rocks account for The origin of Aznalc611ar is time-
most of the sequence. Three episodes of related to the initial stage of the
acid volcanism (V l' V V and one of first acid volcanic episode (V
Z' 3) l),
basic volcanism (Vb, which ended the during which exhalative activity
first acid volcanic episode) are occurred, thus leading to metal leaching
recognized within this unit. The from the volcanosedimentary complex,
massive sulphide mineralization is expulsion of brines from vents, and
genetically related to V V is eventually, to precipitation of the
l; z
thought to be responsible for some metal load onto the seafloor. Metal
minor jasperoid-Mn mineralizations, and transport was probably in the form of
V is "barren". The most important rock halide complexes in weakly acid brines,
types at Aznalcollar are tuffs, ashes, which shortly after expulsion reacted
rhyolites, volcanic breccias and with HZS to form colloidal sulphid e
shales. These rocks were affected by precipitates. Either gravitational
the Hercynian compressions, which instability or earthquake activity
resulted in folding, penetrative flow ultimately led to the generation of

turbidite flows with rh ythmic la yering Azn alcoll ar c ons t it u t e s a r a th er
o f bl a c k shal es and s u l phi d e l a y ers . la r g e E-W/45 Q min e r a l i z e d s eq u enc e
The acid v o lcan ic s a n d sulphi d e c o mpr is i ng th e old mi nin g work s o f
mi n e r al i z at i o n are her e t h ought t o be a Cuchillon , Si li l lo s an d Hi g u e r e t a s. As
c ons eq u e n c e o f th e emplacemen t o f a we r e most o f t h e d epo sits in t he
s ubvol cani c plug lo c a ll y known as th e Pyrit e b e l t , Azn al c oll a r wa s first
" f elsite". S i n ce th e o r e i s s patiall y mi ned by t h e Romans . I t s mod er n
r el a t ed to this subvolc a nic bod y, e x p l o i t a t i o n beg a n in 1876 b y th e
Azn alc oll ar is her e c o n s ide r e d to be a Se v i ll e Sulphur a nd Copper Co . Lt d.,
" pro ximal", although a non- root ed, a n d l as t ed until 1 94 2. Fu r th er
vo lca nose d i me n t a ry d ep osit. e x p l o r a t i o n and as s e s me n t o f r es e r v e s
wer e c a r r ie d out b y Pe narr o y a Esp ana i n
INT RODU CTION 19 56. In 1960, a holding h ead ed by th e
Th e or e d eposit o f Aznalcollar is Banc o Central (Spa in) took o v er a n d a
locat e d within th e so-call e d "Iberian new c omp a n y was f o r me d (And aluza d e
Pyr it e Belt" (Fig. 1), on e of th e Pir it a s S.A.). Finall y, in 1988, th e
world's most outstanding metallog enic Swedi s h comp a ny Boli d en a dq u i re d th e
prov i nce s of mass iv e sulph id e d epos i ts. min e properti es, a nd i s i t s cur r ent
This roughl y 30 Km wid e belt of own e r.
Di na n t i a n ag e s tre tch es from southern Or e re se rve s a mo unt t o (a ) 44 x 1 0 t
Po r t u gal to Sp ain. Th e b e l t tr ends NW of mass i v e sulphi d es g r ad i ng 0 .44 % Cu ,
i n Portug al and b ecomes p rogr essi vel y 1.76 % Pb, 3 . 3 3 % Zn an d 0 .67 g /t Ag;
E-W in Spain, a l o n g some 230 Km, and a nd (b) 34 x 10 t of th e so- c all e d
includ es important str atabound ma s s ive cupr i f erous p yrocl asti c ( "p i r o c l a s t o
sulph i d e d eposits, such a s those of cup rif ero"), which gr ad es 0.58 % Cu,
Nev es Co r vo, Ri o Tinto, Tharsis and 0.40 % Zn and 10 g /t Ag.
Aznalcollar. Its northern boundar y is
d ef in ed b y th e Variscan antiform of
Beja; southward, th e Pyrite Belt is
bound ed by Carboniferous formations and
Mioc en e marin e s ed imentar y s equences.

CJ ;.C ~ ..., Cfo'. l OIC D c ~ u,

l _. .... ... Fig. 2 View of th e Azn alc olla r o p e n

pit works (t ak en from t he Fels it a
hill) .
Fig. 1 Location map of the Iberian
Py ri t e Belt and Aznalcollar
Th e g eom etri c al c ha rac te r i s t i c s of
t h e mi nerali z e d bo d y, t h e o r e g rad e

distribution and geotechnical (pyroclastic circuit) 40,000 t/y of
characteristics of the host-rock made 21 % Cu, 200 g/t Ag copper concentrates,
it advisable for Andaluza de Piritas to and 7,000 t/y of 40% Zn concentrates.
exploit Aznalcollar as an open pit mine Sulphuric acid is produced as a
(production started i n 1979) (Fig . 2). byproduct at the current rate of
The pit is 1500 m long, 700 m wide 250,000-270,000 t/y.
(max imus) and 300 m deep. At the Current research on bio-leaching is
current extraction rate (3.5 Mt/y, with now being done at the University of
a 3.1:1 stripping ratio), ore reserves Madrid (Dept. of Cristalografia y Mine-
will allow continuous exploitation ralogia) in order to test whet en
till 1992. Boliden is currently metals could be selectively leached
carrying out a vigorous exploration from crude ores. Flask agitation,
programme aimed at finding new reserves. column tests and mineralogical studies
The mineral processing flowsheet are now in progress.
of Aznalcollar is shown in Fig. 3.
6000 tpd Three well differentiated units account
for most of the Pyrite Belt
unit 1)
I I A footwall sequence formed by probably
Devon ian slates and quartzites of some
- 150 um 500 m thick; unit 2) The Volcano-
800/0~20um Sedimentary Complex (CVS), which
I overl ies conformably the footwall
P~~g8"~ THICK EN
sequence. This Dinantian complex hosts
KAX + coso, the mineralized bodies, and has a
so, + KAX KAX _HZSO.


maximus thickness of 800 m. At the


pyrite Belt scale, the CVS consiste of

a sequence of acid and basic volcanics,
and interbedded shales/slates. Three
DISC F ILTER episodes of acid volcanism (VI' V2, V3)
THERMAL DRYER and one of basic volcanics (Vb; which
ended the first acid volcanic episod e)
17% Cu 48% Pb 47 % Zn are recogniz ed within .t h e Iberian
3% Z n 61o Zn 2 % Cu
2% Pb Pyrite Belt. The mass ive sulphides are
l oo0 g/1 Ag
3% Sb genetically related to the first
volcanic episode (VI). V is thought
Fig. 3 Mineral processing flowsheet to b e responsible for some minor
jasperoid-Mn mineral izations and V3 is
of Aznalcollar (Fox, 1990).
"barren". Unit 3) conformably overlying
the CVS, a Culm sequence completes the
i Paleozoic st ratigraphy. These Culm
According to Fox the plant has a
facies are formed by slates and
production capacity of: a) (pyrite
greywackes of variable thicknesses
circuit) 25,000 t/y of 21.5 % Cu,
(several thousands metres in the west
1,500 g/t Ag copper concentrates;
and 500 m in the east). The Culm is
38,000 t/y of 47.5 % Pb, 500 g/t Ag
eroded at Aznalcollar and the CVS is
lead concentrates; and 100,000 t/y of
unconformably covered by Miocene
47.5 % Zn concentrates; and b)

marin e s e d i me n t s. f ormi ng th e so- c all e d " p i r o c l asto
Th e Dinan tian vol canis m that fo r med c upr i fer o " , a zone loca lly en r i c hed
th e CVS i s to b e r egard ed a s bi moda l , i n c ha l c opy rite .
and can b e re lat ed to an important Th e ac id volcan ic s und e r went a
e x te n s i o n a l episod e of c on t i n ental serie s of c omp l e x processes of mineral
rifti ng in t h e s o u t h e rn r ealm o f the recry s ta l lization (Tab l e 1) . Up to fiv e
Iberian pen insu la . Th i s e p i s o d e gen eratio n s of quartz ar e r ecogni z ed
belongs to t h e i n i t i al stage s of th e ( QI- Q5 ): th e f o rm e r tw o were r e lated to
Va r i s c a n Cyc le . ma g ma t i c (Ql ) (Fig . 5 ) and ear l y
Th e loca l st ratigraphy of t he CV S at
Az na lco llar (f r om t he base t o t he t op )
inc ludes th e f ollow in g fac ie s (Fig. 4):


~ uOfl1ct\ lor.',c luf '

D o,tl . e-'o'OP"'Y'~ luP

I::~ : 1 khyOIII l! I~f '

D c,n~ ,.! ~

Q OOCI" C l uf f

D r t li ,t t

o S"o'~ . Im cO' bono:~s mal 't'

Is ~ ss4eM'"
' [ rT'I COU OSCo oo

I".u "" t t rous 0 '1'0(10,) '

F i g. 5 He x ag o nal c lasts of igneous

q uartz (Ql) . Cross ed po l ars x lO .

Fig . 4 Azna lcol l ar stra t igraphic

c o l umn (S i erra , 1984 a ) . fumarolic alt eration (Q2) pro ces s es ,
r espectivel y . Th e last thr e e stag es of
quart z cr ystalli z at ion wer e r elat ed to
1 ) a thick s equ enc e of rh yo lit es ; 2) later Var iscan t e ctonometamorphic
c a rbo n - r i c h sl at es and cherts; 3) proc ess es ( F ig . 6) . P la g i o c l a se was
da c it i c tuf fs ; 4) mas siv e su lphid es formed d u ring two e p i s o d e s : a magmatic
(th e so -c a ll ed " F i l o n Sur " bod y , host e d (tim e -r elated to Ql) a nd a metamorphic
by a s he s ) ; 5) an a lternat ing s eq uenc e (time-r elat ed to Q4) . Ch l o r i te is
of tuffs and slat es ; 6 ) a seco nd body most l y a p rod uct of d evitrification
of mass i ve pyrite (t h e so -called and alt e rat ion of magmatic b io t it e .
" F i l o n medio " ) ; 7) pyroc last ics ; 8) a Thr e e g e nerations of t his mine ral ar e
t hird body of ma s s i v e s u lp h id es (th e . e d: 4 th e form er lS
r e c o gnlz . 0
s o - c a l l ed " Filon Si li l los " ) ; a n d 9) un c e r t ain r el a t i v e ag e , b ut it mi gh t
q uar t z- chlo r i tic tuffs and as h es to r ela ted t o e a r l i e r a l te r a t i o n

processes (time-related to Q the
second and third episode of chlorite
formation are of metamorphic origin and
are time-related to Q and Q4'
respectively. Sericite formed in
relation to devitrification-type
processes during early metamorphic
processes. K-feldspar is the most
important mineral within the rhyolitic
tuffs, and formed during a late
magmatic stage. In the other rock
types, K-feldspar is of metamorphic
origin and time-related to Q4.
Minerals of minor importance are:
sphene, zircon, apatite, rutile,
carbonates, tourmaline, kaolin and
barite (the last three are rare) .

Similar to the other ore deposits of
Fig. 6 Quartz-chlorite devitrified lava.
, the Iberian Pyrite Belt, Aznalcollar
Pyrite euhedra surrounded by Q quartz.
3 is pre-orogenic mineralization, which
Crossed po lars x 16.

o s.

s-: - '"'"T __ - '~

D Gra phit iC Sha le s. ( T ournOISlo n- V ISE AN )

RH YOL I TE pl uq

. 100

o~--...,jl oo~

Fig . 7 Schematic cross-section of Aznalcollar (Sierra et al., 1985).

underwent a series of deformation of the mineralized zones, which
phases during Variscan time. The completely discard it as a possible
general structural trend of the Pyrite feeder zone
Belt follows a NW (portugal) to E-W Mineralogy
(Spain) arc in which folding was The ore mineralogy is mostly formed by
extensive and intensive. Four Variscan massive recrystallized pyrite. The
phases of deformation are recognized content of pyrite decreases towards
within the belt. A) Phase 1 (mid the orebody contacts with the barren
Westphalian ) : is responsible for the host-rocks. A wide variety of are
4, , ,10
d '~n pyr~te
major structures of Aznalcollar and textures are recogn~ze
generated tight isoclinal recumbent a) annealed, b) euhedral, c) "caries"
folds vergent to the South with E-W (Fig. 8), d) atolls in sphalerite,
axes (Fig. 7). This phase is
characterized by reverse faulting and
overthrusting leading to the complete
obliteration of the isoclinal folds.
The Sl schistosity developed during
this phase. B) Phase 2 (Stephanian
it generated N90 a-120aE coaxial
assymetric folds and the S2 foliation
(a non-penetrative fracture cleavage
locally observed in the metapelitic
rocks). C) Phase 3: this phase produced
E-W orientated kink-bands in relation
to a late episode of thrust tectonics
and D) Phase 4: this phase is of late
variscan age and is characterized by
the development of E-W and NE brittle

As mentioned earlier, four main
Fig. 8 Caries texture 'of pyrite cemented
mineralized zones are recognized at
by chalcopyrite. Plain polars x 200.
Aznalcollar: Filon Sur, Filon Medio
and Filon Silillos (massive sulphides
and the Firoclasto Cuprifero, a mostly
disseminated mineralization. Stockwork- e) zoned, f) brecciated, g) framboidal
type mineralization has been also (Fig. 9) and h) colloform. The rest of
suggested for Aznalcollar but a the minerals infilled spaces left by
series of features makes that very the uncomplete annealling of pyrite
improbable: 1) absence of zoned related during the metamorphism. Other textural
alteration processes 2) the minor features include the e n ve l o p i n g of
importance and small size of the pyrite by sphalerite, galena,
mineralized veinlets (as compared to chalcopyrite or bournonite.
"true" stockworks e.g. Rio Tinto); and Sphalerite is second in volumetric
3) the supposed "stockwork-like" importance within the or e overwhelmingly
structure occurs towards the top of one dominated by pyrite. This mineral occurs

Aznalcollar does not display any
f e a t u r e s reflecting ore zoning at
orebody scale. In fact, minerals such
as chalcopyrite, sphalerite and galena
follow a rathe r random distribution l O.
Alteratio n
As stated earl ier, the or ebodies occur
within a volcanogenic sequence which
has undergone a variety of alteration
related processes. These phenomena can
be regarded as either hydrothermal or
metamorphic. Common alteration types
throughout the deposits are:
silicification, chloritization and
sericitizac ion. Local albitization and
K-metasomatism have been also observed.
Silicification began during the time
that e xhalative activ ity took place,
and continued during later metamorphic
ph enomena. K-metasomatism is also time-

Fig. 9 Framboids and colloforms of r e l a t e d to the exhalative activity.

pyrite in sphalerite. Plain polars x 200 Chloritization and sericit ization are
related to the vitrification processes
that place during th e first
tectonometamorphic event.

as thin film coatings, colloforms, Geochemistry

mineralized fractures, and bands. A comprehensive geochemical study of

Chalcopyrite infills cavities and Aznalcollar was carried out in
fractures in pyrite. Lead occurs as e a r l y 80's. Th e aim of this study was
galena, bournonite and/or menenghinite. focused in establishing whether metal
All these minerals infill cavities in zoning existed or not at the mine
pyrite or occur in close association to scale. 576 samples from drill cores
sphal erite. Other mineral phases of very were analyzed for Cu, Pb and Zn. The
minor importance are: melnicovite, results showed no zoning, however, a
ars enopyrite, pyrrhotite, tennantite~
reverse behaviour of Cu versus the pair
tetrah edrite, stannite, gudmundite, Pb-Zn was inf erred. Log -probability
marcasite, nagyagite, silver and plots were used to stud y the
bismuth. The paragenetic studylO statistical behaviour of Cu, Pb and Zn
indicates that the different mineral (Fig. 10, 11, 12). Copper has a rather
phas es must have been precipitated simple distribution and only one
simultaneously (first stage of mineral population is observed, with a

formation). Mineral precipitation must background of 0.45 %. Lead and zinc

have been collo idal, and later, due to have distributions displaying two
diagenetic and metamorphic processes, populations each; their backgrounds
the minerals underwent remobilizations are of 1.7 % and 3.1 %, respectively.
and r ecrystallizations.
Different from other examples of
massive sulphid e min eralization,

.,. Probobllity
% Probabil ity



"., 98

00 90



., 60

'" j.L50

'0 4a




,.,, ro

0'> 3
,. ~. Cu
0, . ...
, - - - - ----~~-~-~.l,_-----


Fig. 10 Log-probability plot of copper. % Zn

4 5 6 7 8 9 la'

% Probo bliity
Fig. 12 Log-probability plot of zinc


Aznalcollar is primarily the result of
hydrothermal phenomena related to the
extrusion of the early acid volcanics

(V and can be regarded, to an extent

40 l),
as an integral part of the CVS
stratigraphic column. The V magmatism
.0 l
not only gave rise to the extrusion of
the acid lavas and pyroclastics, but
, triggered important exhalative activity

." that expelled into the seawater the
metals and H The origin of metals
0.' can be regarded as the consequence of
convective cells of seawater operating
." . .
"I. Pb

.s .6 7 .8 . to" 6 within the volcanic pile, from which
they were leaches, and la~er
Fig. 11 Log-probability plot of lead
precipitated by exhalative flows. It

can be envisaged that a system H pyrite-rich lobes of mineral were formed,
-rock accounted for the initial stage gravitational instability induced (Fig.
of mineral precipitation at Aznalcollar. 13 II) the generation of turbidite
First mineral phases in precipitating flows (Fig. 13 III). As shown by the
would have been of the type greigeite stratigraphic and sedimentalogical
and mackinawite, plus Pb, Cu and Zn record, this activity led to the
sulphides. formation of turbidite lobes of pyrite
At Aznalcollar, one or more vents enclosed by black shales (Fig. 14).
fed a continental platform with Similar to other Kuroko-type deposits,
sulphides (Fig. 13 I). Later, once the the early metallogenic history of

SEA L EVEL - -- - - - - - _

- - - - - - - - -- - - - - - - SEA L EV EL - - - - - - - - - -

-------~--------- S E A LEVEL - - - -- -- - - - - -

~ Turbid ites pyri t e

~Felsite ~ Block sholes c:::::::::s lob es
~ Acids volcanites ~ Massive sulphides ~ Source

Fig. 13 Genetic model for Aznalcoll ar (Si erra et a l . , 1985).

Fig. 14 pyrite turbidite lobes .

TI\BlE 1


--- -------
Q1 Q2 Q3 Q4 Q5

Plagioclase carbonates Ollor-ite (1,2) Ollorite (J) carbonate

~q.JeS K-spar sericite Plagioclase Fe-oxidcs

Apatite Plagioclase siderite K-spar .

ZiI'OCX1 Rutile ~ Siderite

Toorlnal1oo st:nene Rutile Muscovite



Hyd.rot.henMl MetarrotPlic Ii1enarena: recryst.alliza-

activity tion, devitrification

Aznalcollar seems to be strongly linked volcanosedimentary massive sulphides.

to a subvolcanic rhyolitic plug (in Moreover, this deposit can be
this case, the so-called "Felsita"), to genetically considered as a typical
which the orebodies are spatially Kuroko-type mineralization. Finally,
related. Since, as discussed earlier, the stratigraphic and mineralogical
no clear evidence supports the existence characteristics of Aznalcollar allow a
of a "true stockwor~" at Aznalcollar, classification of this deposit into the
the orebodies here described can be 12
type III of Colley (F ig. 15).
regarded as "proximal", non-rooted

Sed ime nts Volconic rocks Emergent
- dominat e Stratovolcano
dominate - - --
Lateral vent
Exhalat ive
Sea level

ry sequence
Porphyry -style

Fig. 15 The five types of Kuroko-like mineralizations (after Colley, 1976).

REFERENCES 7. Hofstteter, J.P. L'amas sulfure a

1. Fox, K. Aznalcollar. Mining Magazine, Cu-Pb-Zn d'Aznalcollar (Sevilla),
January 1990, p. 20-25. Espagne. Geologie, paleogeographie et
2. Strauss, G.K., Madel, J., Alonso, F. metalogenese de l'extremite sud-orie~
Exploration practice for strata-bound tal de la ceinture sud Iberique. These
volcanogenic sulphide deposits in the 3eme cycle, Universite de Pierre et Ma-
Spanish-Portuguese pyrite Belt:Geology, rie Curie, France, 1980.
geophysics and geochemistry. In:~ 8. Garcia Palomero, F. Caracteres geo-
Klemm and H.J. Schneider (eds.), Time logicos y relaciones morfologicas y g~
and Strata-Bound Ore Deposits, Springer neticas de las mineralizaciones del an
Verlag, Berlin, 1977). ticlinal de Rio Tinto. Acta Salmanti-
3. Sawkins, F.J. and Burke, K. censia, 1977.
Extensional tectonics and mid-Paleozoic 9. Sierra, J., Arribas, A. and Gumiel,
massive sulphide occurrence in Europe. P. Geologia, mineralogia y metalogenia
Geologisches Rundschau, vol. 69, 1980, del yacimiento de Aznalcollar: Metalo~

p. 349-360. genia. Bol. Geol. Min., XCVI, 1985,

4. Sierra, J. Geologia, mineralogia y p. 23-30.
metalogenia del yacimiento de Aznalco- 10. Sierra, J. Geologia, mineralogia y
llar_ Litoestratigrafia y tectonica. metalogenia del yacimiento de Aznalco-
Bol. Geol. Min., XCV-V, 1984, p. 440- lIar: Mineralogia y sucesion mineral.
455. Bol. Geol. Min., XCV-VI, 1984, p. 553-
5. Schermerhorn, L.J.G. An outline 568.
stratigraphy of the Iberian Pyrite Belt. 11. Sierra, J., Astudillo, J. and Lunar,
Bol. Geol. Min., LXXXII, 1971, p. 239- R. Estudio geoquimico del yacimiento de
268. Aznalcollar (Cinturon Piritico Hispano
6. Hernandez Enrile, J.L. Marco geologi Portugues). I Congreso Espanol de Geo-
co estructural de los yacimientos de logia, II, 1984, p. 685-705.
sulfuros de Aznalcollar (Region Orien- 12. Colley, H. A classification and
tal de la banda piritica Iberica). exploration guide for Kuroko-type
Reunion de Xeologla e Mineria do deposits based on occurrence in Fidji.
Noroeste peninsular, 1981. Trans. Instn. Min. Metall. (Sect. B:
appl~ earth Sci.), 1977, p. 190-199.

Precious- and base-metal mineralogy of the Hellyer
volcanogenic massive sulphide deposit, northwest
Tasmania - a study by electron microprobe
A.R. Ramsden
K.M. Kinealy
R.A. Creelman
R.A. Creelman and Associates, Epping, New South Wales, Australia
D .H. French
CSIRO Division of Exploration Geoscience, North Ryde, New South Wales, Australia

zone of relatively high-Fe sphalerite

(5 - 7 wt%) which is also distin-
guished texturally by the presence of
SYNOPSIS fragments of concentrically zoned
sphalerite crystals carrying abundant
sub-micron chalcopyrite. The crystals
The Hellyer volcanogenic massive probably formed in a vent, from which
sulphide deposit, discovered by they were expelled to accumulate as
Aberfoyle Ltd in 1983, is a poly- clasts on the sea floor. Expulsion
metallic orebody. Although Cambrian, occurred after formation of the lower
it is relatively undeformed and unit, possibly following closure of
contains abundant textural evidence the original vent system, and resulted
to indicate that the sulphides were in deposition of a 'blanket' of high-
extremely fine-grained, perhaps even Fe sphalerite over part of the pre-
amorphous, at the time of deposition. existing sulphide mound. The upper
There are clear affinities with the unit, which contains most of the
Tertiary Kuroko deposits of Japan, for precious metals, accumulated above
which a submarine volcanic- this.
hydrothermal origin is now widely
Gold shows three associations: 1) with
barite, 2) with arsenopyrite and 3)
As part of a broader metallurgical with pyrite. Gold associated with
investigation aimed at developing the barite is probably particulate but
optimum treatment strategy for the gold associated with arsenopyrite and
ore, detailed electron microprobe pyrite occurs as a trace element in
studies have been carried out to both minerals. The arsenopyrite asso-
determine the mineralogy of the ciation is the most important,
deposit and provide quantitative data electron microprobe analyses indi-
on the mineralogical residences of the cating Au-contents up to 400 ppm for
metals. Three drill holes have been this mineral. The pyrite association
analyzed to investigate both strati- is inferred, since the Au-content is
graphic and lateral variations along generally less than the detection
the axis of the orebody. The data limit for electron microprobe analysis
obtained are relevant not only to the (lOOppm). The arsenopyrite is fine-
metallurgy, but also lead to a better grained 20pm), well crystallized and
understanding of the origin of the shows cryptic oscillatory growth
deposit. zoning due to small fluctuations in
Sb-content. The complexity of the
zoning indicates crystallization from
The average Fe-content of sphalerite a turbulent fluid in which early
in the ore (based on 1009 analyses) is formed crystals were fragmented and
2.7 wt%, but trends within the overgrown by more arsenopyrite as
sphalerite divide the deposit into an crystallization proceeded. The inten-
upper and lower unit. The boundary, in sity of zoning and fragmentation
all three holes, is marked by a thin decreases away from the presumed vent

area consistent with crystallization LONGITUDINAL PROJE CTION 100m

in a quieter part of the plume some
distance from the vent. HL38A

Silver occurs as a major element in

tetrahedrite and as a trace element in 2
galena'. The tetrahedr i te and galena
are closely associated in the ore, and
therefore there is an excellent DRILL HOLE
correlation of silver with lead.
However, it is tetrahedrite (and HL58 PLAN OF OREBODY
freibergite) that are the major hosts I
rather than galena. Tetrahedrite at
some distance from the presumed vent
area is less arsenical and more
argent ian than that from close by.
Textural evidence suggests that Ag-

enrichment occurred during late stage
replacement of the tetrahedrite. Some D Bar ite

of the additional silver may have been ~ Mal lin

released from argentian galena at the
hanging wall, that is now largely
EJ Stri nger

HL38A I o
replaced by bournonite and boulang- 100

erite. Such argentian tetrahedrite is


itself replaced by minor pyrargyrite
and carries inclusions of acanthite-
.. "
Que Rive !
Rosel>er y

argentite. :~/

Fig.l Geographic location of the

INTRODUCTION Hellyer deposit and location of the
three dr ill-holes sampled for the
present study. These holes provide
The Hellyer volcanogenic massive stratigraphic cross-sections at
sulphide deposit, N.W. Tasmania, was interva;s along the axis of the
discovered by Aberfoyle Limited in orebody west of the Jack Fault.
August 1983, about 5 km northeast of
the Que River mine (Fig.l) where ' a
similar polymetallic deposit is being
mined by this company. Both deposits The deposit is highly pyritic,
are within the Cambrian Mt Read comprising on average 54 wt% pyrite,
Volcanics and consist of massive 20 wt% sphalerite, 8 wt% galena, 2 wt%
sUlphides with very little gangue. arsenopyrite and 1 wt% chalcopyrite
with minor tetrahedr i te and other
trace sulfosalts. The principal gangue
minerals are barite, quartz, ankerite,
The Hellyer orebody comprises a siderite, chlorite and sericite.
single, faulted, elongate massive
sulphide body overlying a well-
developed pyrite footwall alteration Mineragraphic studies (this work and
pipe (McArthur 2 l). It shows many Fander, pers.comm.) have shown that
features of the Tertiary Kuroko ores primary (depositional) banding is a
of Japan, including the presence of conspicuous feature of the are, and
barite overlying massive are, and an there is abundant textural evidence
underlying stockwork of stringer (especially with spongy, radiating-
mineralization. Unlike the Kuroko fibrous, concentric and colloform
deposits, however, copper is only a pyrite intergrowths) to indicate that
minor constituent (averaging <1.0 wt%) the sulphides were extremely fine-
and there is no gypsum. Precious metal grained, perhaps even amorphous, at
values are concentrated towards the the time of deposition. The are is
stratigraphic hanging wall, as with recrystallized in varying degrees,
other deposits of the gold-zinc resulting in coarsening and re-
association in the Mt Read Volcanics mobilization of galena, sphalerite and
(Large et al. 1 7). tetrahedrite.

The work reported in the present paper The drill holes were sampled at 1 to 2
is based primarily on the results of metre intervals, each sample being
detailed electron microprobe investi- mounted in epoxy and polished for
gations which were carried out as part microscopic and microprobe investiga-
of a much broader collaborative pro- tion using standard petrographic
ject between Aberfoyle Ltd and the procedures. In addition, 9 samples
Commonwealth Scientific and Industrial from HL38A were prepared as doubly-
Research Organization (CSIRO) aimed at polished thin sections, (after
developing the optimum treatment Barton 2), in order to carry out more
strategy for Hellyer ore. Overall, detailed examination of sphalerite
this project involved metallurgy, textures in transmitted light. The
minerals engineering, chemistry and samples were vacuum coated with
geology, with a strong bias towards approximately 20 nm of carbon to
mineralogy and examination of the ore provide a conducting path during
response to unit processes such as analysis in the electron microprobe.
crushing, grinding, flotation, hydro-
metallurgy and pyro-metallurgy.
A Cameca Camebax scanning electron
microprobe was used for both quanti-
As well as providing data for the tative and qualitative analysis of the
process mineralogist, the electron ore minerals. This instrument is
microprobe studies also contribute fitted with four wavelength dispersive
towards a better understanding of the X-ray spectrometers (WDS) analyzing
origins of the mineralization. This from boron to uranium and a Kevex 7000
aspect of the work, however, must be energy dispersive X-ray system (EDS)
seen in the context of broader analyzing from sodium to uranium.
geological studies being carried out
by Aberfoyle staff (McArthur, 21) and
others (Large, pe r s .comm.) , Quantitative analysis can be carried
out on regions down to about 2 pm in
area with semi-quantitative and
qualitative analyses possible at sub-
SAMPLES AND METHODS OF INVESTIGATION micron levels. The WDS system was used
for both major and trace element
analysis. The EDS system was used
Detailed mineralogical studies using primarily for rapid qualitative iden-
electron microprobe techniques were tification of minerals prior to
carried out on samples from three quantitative analysis, and to monitor
drill holes (Fig.l) that provide the progress of trace element analyses
stratigraphic intersections from where long counting times are involved
hanging wall to foot wall along the (i.e. to ensure that the electron beam
axis of the orebody west of the Jack remained on the mineral of interest).
Faul t , Hole HL38A intersects the ore-
body at its southern end where it is
overlain by barite ore. Hole HL70 is Qualitative analysis involved scanning
located some 200m NNE of HL38A and electron microscopy using back-
intersects minor barite-enrichment scattered electrons (BSE) to display
within the massive sulphide (possibly mineral relationships, and the
correlating with barite ore overlying imaging and profiling of X-ray signals
the orebody immediately to the east). to display element distributions.
Hole HL56 is located some 300m further Images obtained in BSE-mode contain
NNE where barite is negligible. The much compositional information, the
orebody east of the fault has not been contrast being determined mainly by
sampled in the present study. differences in average atomic number,
so that heavy minerals such as gold or
galena appear bright whereas light
Overall, the depos it opens out to the minerals such as quartz, carbonate and
north and plunges in this direction silicate gangue appear much darker.
(Fig.l). Thus, the hanging wall is Thus fine-grained minerals (eg micron-
encountered at increasingly greater sized gold and chalcopyrite) that can
depth from 152m in HL38A to 223m in be very difficult to distinguish
HL70 and 300m in HL56. unambiguously with optical microscopy
can be readily identified with the
electron microprobe, and their

compositions determined. small 20 11m) inclusions of the other
sulphides. Galena also forms coarse
patches and veins, but these are
comparatively free of other sulphides,
RESULTS occurs as complex intergrowths with
pyrite and sphalerite and as minute
disseminated grains. Minor chalco-
Oualitative mineralogy pyrite occurs generally as very small
S 11m) inclusions in sphalerite, and
less commonly as coarse (100 11m)
Detailed mineragraphic examination has patches.
not formed part of the present work,
but it is known from other studies
(Fander, p e r s , comm.) that pyrite, Minor euhedral arsenopyrite occurs as
sphalerite and galena are the major fine 20 11m) intergranular crystals
sulphides with minor arsenopyrite and intimately associated with pyrite
chalcopyr i te and t race amounts of (Fig.4a) and as inclusions in
tetrahedrite and other sulphosalts. sphalerite (Fig.4b). The mineral is
auriferous (see below) but its fine
grain size makes it difficult to
The complexity of the ore is liberate. It commonly displays cryptic
illustrated in Figs.2 to 6, using oscillatory zoning at the micron and
back-scattered electron images to sub-micron scale. Such zoning is
reveal the textures. Several of these clearly related to growth of the
are not readily visible with optical crystals (Fig.4c) and many of the
microscopy. larger grains appear to have formed
as a result of overgrowth on
fragmented cores (Fig.4d). The zoning
Broad-scale features of the ore, is most conspicuous in HL38A and least
arising from the volcaniclastic obvious in HLS6. Detailed X-ray
sedimentary origin of the deposit and profiles indicate that it is due to
remobilization and recrystallization variations in the minor Sb-content
of the more ductile minerals such as (Fig.S)
galena, tetrahedrite and sphalerite,
include possible sedimentary load-
structures (Fig.2a) in fine-grained Argentian tetrahedrite is a trace
sphalerite ore, bands of coarse (1 mm) mineral, concentrated in the upper
concretionary pyrite (Fig.2b), frag- part of the orebody (the Precious
mental and brecciated pyritic ore Metals Zone of Aberfoyle) where it
(Figs.2c and 2d) and complex fragmen- occurs as coarse (to 100 11m) grains of
ted intergrowths of colloform pyrite complex composition (Fig. 6a) mainly
and galena (Figs.2e and 2f). in association with remobil ized
galena. Micron-sized inclusions of
acanthite (A9 2S) have been observed in
At a much finer scale, primary such tetrahedrite (Figs.6b , 6c and
(depositional) textures are represen- 6d). The cubic morphology of the
ted mainly by framboidal pyrite and inclusions (Fig.6d) indicates they
colloform intergrowths of pyrite and were originally argentite, the high
sphalerite to 20 11m in diameter temperature ()177 oC) polymorph.
(Figs.3a and 3b). The similarity of Tetrahedrite also occurs as cross-
these textures to those found in cutting veinlets; less commonly as S
hydrothermal sulphide deposits to 10 11m disseminated grains and, in
currently forming at the sea floor is the barite-ore, as S 11m inclusions
striking (Koski et al.,lS; Paradis et within sphalerite. Very rarely, it
al.,22). The close affinities with occurs intergrown with colloform
Kuroko deposits are clearly evident pyrite.
when the microscopic features of
sphalerite-rich ore (Fig.3c) are
compared with those of black ore from Other trace minerals found at the
the Shakanai deposit of Japan hanging wall include boulanger i te
(Fig.3d) (SPbS.2Sb 2S 3) and bournonite
(2PbS.Cu2S.Sb2S3) replacing galena,
pyrargyrite (3A9 2S.Sb 2S 3) replacing
Sphalerite typically forms coarse (mm- tetrahedrite and gudmundite (FeSbS) -
size) patches, commonly full of very the Sb-analogue of arsenopyrite. These

Fig.2 Backscatter electron images showing broad-scale textural
features of the ore. Scale bar = 0.1 mm. a) Sedimentary load
structure in sphalerite ore. Sphalerite (light grey), pyrite
(dark grey), galena (white) and quartz (black). b) Coarse
concretionary pyrite (dark grey) with galena (white) and
sphalerite (light grey). c) Fragments of pyrite ore (dark grey)
containing abundant arsenopyrite (light grey), cemented by
remobilized sphalerite (grey), galena (white) and quartz (black).
d) Fragments of sphalerite ore (grey) and pyrite ore (dark grey)
cemented by quartz (black) and rimmed by remobilized galena
(white) and tetrahedrite (light grey). e) Complex colloform
intergrowths of galena (white), sphalerite (light grey) and
pyrite (dark grey). f) Fragments of colloform galena (white) and
pyrite (grey) in a matrix of calcite (black).

have no economic value but are of the gold associated with barite-rich
genetic interest, their presence being ore is amenable to direct cyanidation
interpreted as the result of late- and therefore probably free (Woodcock,
stage Sb-rich solutions concentrated p e r s v c o mm, ) This conclusion is
at the hanging wall. supported by the observation of a
single particle of gold in barite ore
(McArthur, pe r s.comm.) , No particulate
gold has been observed in the present
Ouantitatiye mineral chemistry study.

Mineralogical residence of gold The Au-content of pyrite is generally

below the electron microprobe
detection limit of 100 ppm. Electron
On the basis of the current mineral- microprobe trace analyses of the fine-
ogical results as well as extensive grained arsenopyrite in HL38A and
metallurgical testing, there is now HL56, however, indicate that this
widespread agreement among Hellyer mineral is indeed auriferous with Au-
geologists and metallurgists that at contents up to about 400 ppm. In HL56,
least 90 % of the gold occurs in auri- there is sufficiently good agreement
ferous arsenopyrite and/or auriferous between trends shown by the whole rock
pyrite. However, a small proportion of Au-assays (Aberfoyle data) and the Au-

Fig.3 Backscatter electron images showing fine-scale textures of

the ore. a) Pyrite framboids (grey) in a matrix of barite (light
grey). Minute inclusions of galena (white) are present in the
framboids and laths of interstitial chlorite (black) occur with
the barite. Scale bar = 10 ~m. b) Colloform grains composed of
pyrite (dark grey), sphalerite (mid grey), barite (light grey)
and galena (white). Flakes of chlorite (black) are visible (upper
right). Scale bar = 100 ~m. c) Detail of sphalerite-rich ore
showing abundant interstitial galena (White) and minor
disseminated pyrite (black) in a matrix of sphalerite (grey).
Scale bar = 100 ~m. d) Detail of Kuroko ore from the Shakanai
mine, Hokuroko district, Japan, showing interstitial galena
(white), tetrahedrite (light grey) and disseminated pyrite
(black) in a matrix of sphalerite (grey). Scale bar = 100 ~m.

Fig.4 Backscatter electron images of arsenopyrite. a) Fine-
grained arsenopyrite (light grey) and minor galena (white) along
the grain boundaries of pyrite (grey). b) Coarse-grained
arsenopyrite (light grey) and minor galena (white) associated
with sphalerite (mid grey) and pyrite (dark grey). c) Detail
(high contrast) of cryptic oscillatory growth zones in
arsenopyrite. The brighter zones have higher Sb-contents. d)
General view (high cont rast) of zoned arsenopyr i te show ing
overgrowth on complex fragmented cores. Scale bar = 10 pm.

Electron microprobe analyses of

content of the arsenopyrite (Fig.7) to tetrahedrite show a wide range of
conclude that th is is the dominant compositions, average values for which
host mi n e r a l in this section. Overall, are summarized in Table I. The number
however, there is no simple of analyses averaged per sample ranges
relationship between whole rock As- from 1 to 20 and gives a rough idea of
and Au-contents (see Geochemical the variation in relative abundance
Trends), possibly due to the presence (ease of location) of the mineral.
of arsenical pyrite . The Ag-content of individual grains
varies from 1.9 to 17.9 wt % in
HL38A, from 3.0 to 17.9 wt% in HL70
Mineralogical residence of silver and from 3.0 to 28.3 wt% (freibergite)
in HL56. Large variations also occur
within individual grains that appear
Tetrahedrite and galena are the homogeneous under a light microscope.
principal hosts for silver in the The complex internal structure
deposit. Tetrahedrite, although only a revealed in backscatter electron
trace constituent, accounts for about images of such grains has been noted
60 % of the Ag-content of the ore and earlier (Fig.6a).
is the dominant host in the upper part
of the orebody (the Precious Metals
Zone of Aberfoyle). Galena is dominant Although acanthite-argentite inclu-
i n the lower part of the deposit sions in tetrahedr ite and pyrargyrite
(where t elrahedrite rarel y occurs). replacing tetrahedrite have ver y h igh
Ag-contents by definition, these concentrates from the deposit that
minerals are rare in the Hellyer agree well with the bulk Ag-assays.
deposit and have no economic signif- The presence of antimony accounts for
icance. Boulangerite and bournonite the enhanced solubility (Amcoff l).
from the hanging wall in HL38A are not
argent ian.
Reconnaissance electron microprobe
analyses for trace silver in the
Electron microprobe analyses for trace pyrite of HLS6 indicate Ag-contents of
silver in the galena show a range from 2S0 to 3S0 ppm in the Precious Metals
less than 100 ppm to about 0.42 wt% Zone; values that are well above the
(Table II). The highest values are detection limit of 100 ppm. Sub-micron
found at the base of the orebody in inclusions of the associated galena
HL70, and at the base of the the cannot account for such levels because
Precious Metals Zone. Similar levels the Ag-content is not high enough
have long been known in galena from (although, of course, the inclusions
the Ja~anese Kuroko deposits may be atypical), and inclusions of
(Yamaoka 8. Shimazaki 2 7), and Foord tetrahedrite are extremely unlikely in
and Shawe 8 have recently reviewed the view of the scarcity of this mineral.
chemistry of galena from a variety of The existence of pyritic silver may
sources. There can be little doubt well explain the anomalously high Ag-
that silver in Hellyer galena is in assays reported by Aberfoyle for
solid solution despite the fact that pyrite concentrates from the hanging
the solubility of A9 2S in PbS is wall of the deposit. Analyses of about
insi~nificant below soooe (Van 200 ppm Ag are typical for such
Hook 3 ). This is confirmed by electron concentrates compared to levels of
microprobe analyses of galena in Pb- about SO ppm Ag from the Footwall.


P Y fi l T E PYfilTE


o 3 5 6 7 8 11 12
I( l c r o n s

Fig.S X-ray line-scan showing the distribution of As, Fe and

minor Sb across a grain of zoned arsenopyrite in contact with
pyrite. Length of scan = 13 ~m. Note: maximum for As corresponds
to about 42wt% As, maximum for Fe corresponds to about 46wt% Fe
and maximum for Sb corresponds to about 2wt% Sb.

Fig.6 a) Backscatter electron image showing a large grain of
complex argentian tetrahedrite (shades of grey) associated with
base metal sulphides (black) and galena (white). The different
shades of grey reflect variation in the Ag-content which ranges
from 11 wt% for the darkest zones to 28 wt% (freibergite) for
the lightest. Scale bar = 100 11m. b) Backscatter electron image
showing another complex argentian tetrahedrite (dark grey to
light grey) associated with remobilized galena (white) and minor
pyrite (black). In this case the tetrahedrite also contains
abundant inclusions of acanthite-argentite (also white). Scale
bar = 10 11m. c) Backscatter electron image at high magnification
showing the cubic morphology of the acanthite-argentite
inclusions in (b). Scale bar = 1 11m d) X-ray element map showing
the distribution of silver in (b).

below). Rare zoning has been observed

in some crystalline pyrite.
As-content of pyrite

Sb-content and Zn-content of arseno-

Reconnaissance data for the pyrite pyrite
show a wide range in As-content from
less than 500 ppm to about 2 wt%, and
the distribution of results indicates Electron microprobe major element data
that several populations (generations) for arsenopyrite from the arsenical
are present. No assessment has been zones of HL38A, HL70 and HL56 are
made of relative abundance, but the summarized in Table III. The only
majority of grains have low to minor elements of significance are
moderate As-contents on the order of antimony and zinc. The average Sb-
0.5 wt%. Pyrites of very high As- contents range from 0.5 to 1.2 wt% but
content are represented by the large show no systematic trends between the
(1 mm) colloform concretions, illus- holes. The effects of the cryptic
trated in Fig.2b, that have been zoning, however, are evident in the
observed only in a zone of high As- variability of the individual analyses
content in the upper part of HL56 (see which range from about 0.3 to 3.0 wt%

ppm ppm Sb. This variability is greatest in
2 4 6 81012 100 200 300 400 HL38A. The average Zn-contents are
300 generally less than 1 wt% and may, in
part, reflect contamination from
associated sphalerite.

Fe-content of sphalerite

320 Detailed electron microprobe analyses

of sphalerite in HL38A, HL56 and HL70
are summarized in Tables IV, V and VI,
respecti vely. The composi tion within
330 individual samples is generally very
uniform, and the average values well
represent the stratigraphic variation.
Au in Mean Au content
Depth (rn) Drill Core Assay of Arsenopyrite The mean Fe-contents for HL38A, HL70
and HL56 are, respectively, 3.0, 2.9
Fig.7 Plot showing the Au-content of and 2.2 wt%, the average for all
arsenopyrite from the Precious Metals sphalerite analyses (1009 in total),
Zone in HL56 and correlation with the including a few samples from other
assay for total gold. locations within the deposit, being

Table I Average composition (wt%) of Hellyer tetrahedrites

Depth Cu Pb Zn Fe Ag Sb As S Total N*

(m) BL38A
154.5 21.5 <.1 10.6 4.7 15.7 19.3 3.1 26.9 101.8 5
161.5 39.5 <.1 4.7 3.3 2.1 16.5 9.1 26.8 102.0 3
164.9 37.1 <.1 5.5 3.0 3.1 18.5 7.5 26.3 101.1 10
165.9 38.7 <.1 5.1 3.6 2.6 16.5 9.4 25.9 101.8 3
191.8 33.0 <.1 4.9 5.9 5.3 26.8 0.6 25.3 101.8 4
208.9 36 .4 <.1 5.9 2.5 3.3 25.5 2.7 25.3 101.8 10

301.9 27.3 <.1 9.1 3.0 11.0 25.1 0.4 24.5 100.5 3
303.0 22.5 <.1 3.6 3.3 20.3 26.4 0.3 22.7 99.1 10
304.0 24.0 <.1 2.8 3.8 18.2 27.1 0.2 23.3 99.5 20
305.0 32.1 <.1 4.0 3.1 7 .7 27.9 0.0 24.6 99.6 5
306.0 32.3 <.1 4.4 3.7 5.9 27.7 0.9 24.5 99.4 10
308.0 33.7 <.1 5.0 2.5 5.2 28.1 0.8 24.9 100.2 1
309.0 30.8 <.1 2.9 4 .3 9.0 27.3 1.3 24.3 99.9 2
310.0 30.8 <.1 4.0 3.5 7.5 27.7 0.5 24.2 98.2 10
311.0 26.8 <.1 4.2 3.9 12.6 27 .1 1.8 23 .8 100.2 5
312.0 32.0 <.1 3.8 6.1 4.2 26.6 1.0 25.5 99.3 5
313 .0 33.9 <.1 5.4 4.2 3.0 27.1 0.6 26.2 100.4 1
315.0 35.0 <.1 2.8 6.0 3.4 28.6 0.6 25.5 102.0 1

223.9 32.5 <.1 3.6 3.5 7.4 26.7 1.4 24.8 99.9 5
226.0 31.7 <.1 3.7 3.7 7.7 26.9 1.4 24.7 99.9 10
228.0 32.2 <.1 3.4 4.9 6.0 27.3 1.0 25.0 100.0 10
230.0 24.6 <.1 3.2 4.3 17 .5 27.5 0.5 23.2 100.8 1
232.0 32.2 <.1 4.5 3.1 6.6 27.6 1.1 24.7 99.8 11
234.0 31.3 <.1 5.0 2.7 8.2 27.6 1.1 24.9 100.7 11
238.0 27.4 <.1 3.0 5.2 13 .6 27.3 0.6 24.0 101.2 2
242.0 25.8 <.1 3.7 3.7 15.2 27.1 0.9 23.8 100.2 10
244.0 29.6 <.1 3.1 5.4 8.1 26.2 1.4 25.3 99.2 11
256.0 34.8 <.1 4.8 2.6 4.1 24.3 3.2 25.2 99.0 4

* Note: N = number of analyses.

2.7 wt% Fe. Whilst these average Table II Ag-content (ppm) of
figures are useful for assessing Zn- Hellyer galena
recovery from the ore, they conceal a
marked stratigraphic increase in Fe- BL38A BL56 BL70
content (to about 7 wt% in HL38A, 5 Depth Ag Depth Ag Depth Ag
wt% in HL56 and 6 wt% in HL70) that (m) (ppm) (m) (ppm) (m) (ppm)
is of considerable genetic interest
(see DISCUSSION). 161.5 230 301.9 180 223.9 320
164.9 <l00 303.0 170 228.0 290
165.9 150 304.0 290 232.0 150
Geochemical trends 166.9 590 305.0 620 236.0 230
167.9 190 306.0 170 240.0 1200
171.8 3360 307.0 890 244.0 940
Figure 8 summarizes geochemical trends 175.9 590 308.0 330 248.0 140
for Au, Ag, As, Ba, Cu and Pb within 191.8 1100 309.0 120 260.0 970
each hole (based on Aberfoyle assays 208.9 260 310.0 270 270.0 540
of the samples used in this study). 215.9 380 311.0 370 280.0 4220
Because the orebody is intersected at 312.0 910 290.0 3510
progressively greater depth from south 313.0 920 300.0 1180
to north, it is convenient to present 315.0 110
the data for each element in such a 323.0 <l00
way that the profiles for HL38A, HL70 329.0 200
and HL56 are projected onto the one 338.0 <l00
diagram. It should be emphasized,
however, that there is a spatial
separation along the axis of the is clearly seen in HL38A where there
deposit of 500m between HL38A (on the is a good correlation with As-content
left) and HL56 (on the right). (Fig.8b) for the lowermost Au-zone
(where auriferous arsenopyrite is
abundant) and with barite (Fig.8c) for
As is to be expected in a complex the uppermost Au-zone (at the hanging
volcaniclastic-sedimentary environment wall). The association with arseno-
where lateral continuity of facies is pyrite is also quite clear in HL56
limited, much of the detailed down- where the Au-content of this mineral
hole variability cannot read ily be closely parallels the bimodal
correlated along the axis of the distribution of whole rock gold
depos it. Neverthless, on the broad (Fig.7). In HL70, however, only gold
scale, trends are present that provide at the hanging wall can be correlated
a useful basis for discussing the with arsenopyrite; the lowermost Au-
mineralogical results. zone (which occurs about halfway down
the hole) is associated with barite.

Figure 8a shows the distribution of

gold. Assuming an economic cut-off of A third associat ion (with pyrite) may
4 ppm Au, there are three Au-rich be inferred for the central zone of
zones (shaded) in HL38A, two in HL70 Au-enrichment in HL38A. Although there
and two in HL56. The Au-concentrations is a correlation with minor arsenic in
are highest in HL56, where there is a this zone, this cannot explain the
well defined bimodal distribution in bulk Au-assay, for the arsenopyrite
the upper part of the deposit, and carries only 300 ppm Au. It is
lowest in HL70, where a bimodal dis- concluded, therefore, (though not
tribution is only broadly discernible. proved) that auriferous pyrite is

To assist visual comparison of trends, Figure 8d shows the Ag-distribution in

these zones have been plotted as the three holes. The Ag-content is
reference intervals on all the lowest in HL38A and highest in HL70
geochemical profiles. and HL56. Although there is a good
correlation with bimodal gold (shaded)
at the hanging wall in HL56 this is
Two Au-assoc iations are well defined, not the case further south in HL38A
one with arsenic (arsenopyrite) and and HL70. Indeed, there is no
the other with barium (barite). This correlation with gold in HL38A, other

Table III Average composition (wt%) of arsenopyrite

Depth Cu Pb Zn Fe Ag Sb As S Total N*

(m) BL38A
164.9 <.1 <.1 0.5 35.3 <.1 0.5 43.0 20.8 100.1 6
182.9 <.1 <.1 0.3 35.4 <.1 1.2 41.3 20.7 98.9 13
191.8 0.1 <.1 0.7 35.2 <.1 0.9 42.4 20.7 100.1 10

303.0 <.1 <.1 1.4 35.6 <.1 0.3 41.6 22.3 101.3 1
306.0 <.1 <.1 1.2 35.0 <.1 1.2 41.5 21.8 100.8 10
311.0 <.1 <.1 0.6 35.3 <.1 0.5 42.0 21.4 99.9 23

223.9 <.1 <.1 0.3 34.8 <.1 0.7 43.7 21.2 100.7 5
226.0 <.1 <.1 1.9 35.5 <.1 0.4 40.3 23.6 101.8 2
232.0 <.1 <.1 0.5 35.0 <.1 0.8 42.2 20.5 99.1 15
236.0 <.1 <.1 0.0 35.2 <.1 0.9 42.2 20.2 98.5 6

* Note: N number of analyses

Table IV Average compositions (wt%) of sphalerite in HL38A

Depth Cu Pb Zn Fe Ag Sb As S Total N*

154.5 0.1 <.1 65.7 0.6 <.1 <.1 <.1 33.0 99.4 3
160.9 0.2 <.1 65.8 1.0 <.1 <.1 <.1 33.5 100.5 8
161.5 <.1 <.1 65.6 1.2 <.1 <.1 <.1 33.6 100.4 11
162.9 0.1 <.1 64.9 1.3 <.1 <.1 <.1 33.2 99.5 10
164.9 0.1 <.1 66.2 1.4 <.1 <.1 <.1 33.6 101.3 9
165.9 0.1 <.1 65.6 1.5 <.1 <.1 0.1 33.4 100.7 10
166.9 0.1 <.1 62.5 3.1 <.1 <.1 <.1 33.9 99.6 10
167.9 <.1 <.1 66.0 2.2 <.1 <.1 <.1 33.4 101.6 11
170.9 0.3 <.1 61.1 5.0 <.1 <.1 <.1 33.5 99.9 10
171.8 <.1 <.1 60.7 6.1 <.1 <.1 <.1 32.8 99.6 10
172.9 <.1 <.1 60.6 6.7 <.1 <.1 <.1 34.0 101.3 9
174.9 <.1 <.1 65.4 2.6 <.1 <.1 <.1 33.2 101.2 10
175.0 <.1 <.1 63.5 2.8 <.1 <.1 <.1 33.0 99.3 9
176.9 0.3 <.1 63.5 3.5 <.1 <.1 <.1 33.4 100.7 10
178.9 0.1 <.1 62.2 3.6 <.1 <.1 <.1 33.9 99.8 10
182.9 <.1 <.1 63.9 3.4 <.1 <.1 <.1 33.6 100.9 10
184.9 0.2 <.1 62.9 3.7 <.1 <.1 <.1 33.7 100.4 10
186.9 <.1 <.1 62.5 3.9 <.1 <.1 <.1 33.4 99.8 10
188.9 <.1 <.1 62.6 4.2 <.1 <.1 <.1 33.9 100.7 10
191.8 <.1 <.1 63.1 3.5 <.1 <.1 <.1 33.7 100.3 10
192.8 <.1 <.1 64,0 3.1 <.1 <.1 <.1 33.6 100.7 11
194.9 0.1 <.1 63.8 2.3 <.1 <.1 <.1 32.8 99.0 10
196.0 <.1 <.1 63.3 3.3 <.1 <.1 <.1 33.6 100.2 13
200.8 <.1 <.1 64.4 2.0 <.1 <.1 <.1 32.6 98.0 10
202.5 <.1 <.1 64.5 2.1 <.1 <.1 <.1 33.4 100.0 10
204.5 <.1 <.1 63.3 2.2 <.1 <.1 <.1 33.1 98.6 10
207.9 <.1 <.1 63.5 2.9 <.1 <.1 <'1 33.7 100.1 10
208.9 0.1 <.1 62.9 3.7 <.1 <.1 <.1 33.5 100.2 10
209.9 <.1 <.1 63.4 3.7 <.1 <.1 <.1 33.9 101.0 10
211.0 <.1 <.1 62.4 3.7 <.1 <.1 <.1 33.5 99.6 10
213.4 0.2 <.1 62.6 3.3 <.1 <.1 <.1 33.9 100.0 10
217 .9 <.1 <.1 64.1 2.4 <.1 <.1 <.1 33.6 100.1 10

All 0.1 <.1 63.6 3.0 <.1 <.1 <.1 33.5 100.2 314

* Note: N number of analyses.

Table V Average compositions (wt%) of sphalerite in HL56

Depth Cu Pb Zn Fe Ag Sb As S Total N*

301.9 0.4 <.1 63.3 2.8 <.1 <.1 <.1 33.6 100.1 10
303.0 1.1 <.1 60.9 3.8 <.1 <.1 <.1 33.8 99.6 9
304.0 0.9 <.1 61.6 3.1 <.1 <.1 0.2 33.6 99.4 8
305.0 0.2 <.1 62.5 3.1 <.1 <.1 <.1 33.6 99.4 10
306.0 0.3 <.1 61.4 3.6 <.1 <.1 0.2 33.4 98.9 16
307.0 0.4 <.1 63.0 4.0 <.1 <.1 <.1 34.0 101.4 8
308.0 0.1 <.1 62.5 3.6 <.1 <.1 <.1 33.1 99.3 10
310.0 0.2 <.1 62.3 3.4 <.1 <.1 <.1 33.2 99.1 10
311.0 0.2 <.1 63.3 3.5 <.1 <.1 <.1 34.7 101.7 7
313 .0 0.5 <.1 62.7 3.2 <.1 <.1 <.1 32.5 98.9 18
315.0 0.6 <.1 62.5 3.1 <.1 <.1 <.1 33.7 99.9 10
317 .0 0.6 <.1 60.8 3.8 <.1 <.1 <.1 33.4 98.6 7
319.0 0.2 <.1 62.3 3.6 <.1 <.1 <.1 33.6 99.7 8
323.0 <.1 <.1 60.6 5.0 <.1 <.1 <.1 34.1 99.7 10
325.0 <.1 <.1 63.5 3.2 <.1 <.1 <.1 33.9 100.6 10
327.0 <.1 <.1 64.3 2.7 <.1 <.1 <.1 33.1 100.1 10
329.0 0.2 <.1 62.4 3.1 <.1 <.1 <.1 33.7 99.4 10
331.0 0.2 <.1 64.7 2.0 <.1 <.1 <.1 33.1 100.0 9
333.0 <.1 <.1 65.2 1.6 <.1 <.1 <.1 33.0 99.8 9
335.0 <.1 <.1 64.9 1.4 <.1 <.1 <.1 32.7 99.0 10
337.0 0.1 <.1 65.2 1.3 <'1 <.1 <.1 33.1 99.7 10
338.0 0.1 <.1 65.1 1.3 <.1 <.1 <.1 33.4 99.9 10
339.0 <.1 <.1 65.4 1.0 <.1 <.1 <.1 33.3 99.7 10

All 0.3 0.2 62.8 2.9 <.1 <.1 <.1 33.4 99.6 243

*Note: N = number of analyses.

than overall enrichment in the upper distribution (due to concentrations of

part of the deposit (the Precious chalcopyrite), and lowest in the north
Metals Zone). In HL70, the zone of (HL56). The Cu-content increases with
highest Ag-concentration also is not depth in HL70, reflecting an increase
associated with gold, although Ag- and in chalcopyrite, but is more or less
Au-enrichment coincide at the hanging uniform in HL56.

The Pb-distribution (Fig.8e) closely DISCUSSION

follows that of silver in all three
holes, but it is incorrect to conclude
that galena is the dominant Ag-bearing On the basis of its general geological
mineral. The correlation merely ref- setting, ore mineralogy, host rocks
lects the close physical association and associated alteration (McArthur
between galena and argent ian 21), there is general agreement
tetrahedrite. Only in the lower part amongst Hellyer geologists that the
of the deposit in HL38A is the Ag- orebody is of volcanic-hydrothermal
distribution dominated by the galena. origin, formed at or near the
discharge sites of submarine hydro-
thermal activity. The characteristic
Copper (Fig.8f) is a very minor features of such deposits have been
constituent of the ore, the dominant reviewed by Franklin et al.,9. The
host mineral being tetrahedrite rather results of the present study provide
than chalcopyrite. It is highest in additional information on the
the south (HL38A) where there is a processes that have operated to form
suggestion of layering in the the Hellyer deposit.

Table VI Average compositions (wt%) of sphalerite in HL70

Depth Cu Pb Zn Fe Ag Sb As S Total N*

223.9 0.1 <.1 62.2 3.1 <.1 <.1 <.1 33.3 98.7 10
226.0 0.1 <.1 63.5 2.6 <.1 <.1 <.1 33.4 99.6 9
228.0 0.1 <.1 62.9 3.0 <.1 <.1 <.1 33.5 99.5 10
230.0 0.1 <.1 62.4 3.4 <.1 <.1 0.1 33.3 99.3 10
232.0 0.1 <.1 62.6 3.3 <.1 <.1 0.2 33.4 99.6 10
234.0 0.2 <.1 63.3 3.2 <.1 <.1 <.1 33.0 99.7 10
236.0 0.1 <.1 63.2 3.4 <.1 <.1 <.1 33.0 99.7 10
238.0 <.1 <.1 63.4 3.1 <.1 <.1 <.1 33.4 99.9 10
240.0 0.1 <.1 62.2 4.2 <.1 <.1 0.1 33.4 100.0 10
242.0 <.1 <.1 62.9 3.7 <.1 <.1 <.1 33.5 100.1 10
244.0 0.5 <.1 60.3 5.6 <.1 <.1 <.1 33.5 99.9 10
246.0 0.1 <.1 64.1 2.3 <.1 <.1 <.1 33.1 99.6 10
248.0 <.1 <.1 64.8 1.3 <.1 <.1 0.1 33.2 99.4 10
250.0 <.1 <.1 64.9 1.6 <.1 <.1 <.1 33.1 99.6 10
252.0 0.1 <.1 64.9 1.7 <.1 <.1 <.1 33.4 100.1 10
253.0 0.1 <.1 64.6 1.6 <.1 <.1 <.1 33.1 99.4 10
254.0 0.1 <.1 64.3 2.1 <.1 <.1 <.1 32.9 99.4 10
256.0 0.1 <.1 64.6 1.3 <.1 <.1 <.1 33.2 99.2 9
258.0 <.1 <.1 64.6 1.8 <.1 <.1 <.1 33.2 99.6 11
260.0 0.1 <.1 64.2 1.9 <.1 <.1 <.1 32.9 99.1 18
264.0 0.1 <.1 63.6 2.5 <.1 <.1 <.1 33.0 99.2 20
268.0 0.1 <.1 64.5 1.7 <.1 <.1 <.1 32.2 98.5 11
270.0 0.3 <.1 64.6 1.2 <.1 <.1 <.1 34.2 100.3 11
272 .0 0.1 <.1 64.5 1.8 <.1 <.1 <.1 33.2 99.6 10
264.0 0.1 <.1 64.7 1.9 <.1 <.1 <.1 33.6 100.3 10
276.0 0.1 <.1 64.3 2.2 <.1 <.1 <.1 33.6 100.2 10
278.0 0.1 <.1 62.4 2.9 <.1 <'1 <.1 33.3 98.7 9
280.0 <.1 <.1 64.4 1.9 <.1 <.1 <.1 33.5 99.8 10
282.0 0.1 <.1 64.5 1.6 <.1 <.1 <.1 33.2 99.4 10
284.0 0.1 <.1 64.9 1.5 <.1 <.1 <.1 33.2 99.7 11
286.0 0.1 <.1 64.5 1.4 <.1 <.1 <.1 33.0 99.0 11
288.0 0.1 <.1 64.7 1.5 <.1 <.1 <.1 33.2 99.5 12
290.0 <.1 <.1 64.3 1.6 <.1 <.1 <.1 33.1 99.0 9
292.0 0.3 <.1 64.2 1.6 <.1 <.1 <.1 33.5 99.6 9
294.0 0.1 <.1 64.7 1.4 <.1 <.1 <.1 32.8 99.0 10
296.0 0.3 <.1 64.9 1.5 <.1 <.1 <.1 33.7 100.4 10
298.0 <.1 <.1 64.6 1.7 <.1 <.1 <.1 33.6 99.9 10
300.0 0.1 <.1 64.2 2.0 <.1 <.1 <.1 32.9 99.2 11
302.0 0.1 <.1 63.5 2.5 <.1 <.1 <.1 33.2 99.3 10
302.0 <.1 <.1 63.5 2.3 <.1 <.1 <.1 33.0 98.8 10
304.0 <.1 <.1 64.8 1.4 <.1 <.1 <.1 32.7 98.9 10

All 0.1 <.1 63.9 2.2 <.1 <.1 <.1 33.2 99.4 431

*Note: N number of analyses

Significance Qf ~ sphalerite progressively greater depth below the

compositions hanging wall in HL38A, HL70 and HL56.
In the south, it is clearly related to
the abundance of sphalerite in HL38A
The fact that a zone of comparatively (Fig. 9b) and to a lesser extent also
high-Fe sphalerite (Fig.9a) is in HL70, but there is no such
encountered at depth in all three correlation in HL56 in the north. In
holes suggests that this is a HL70 it lies immediately above a zone
widespread horizon. It transgresses of minor barite-enrichment (cf.Fig.
the stratigraphy, appearing at 8c) that, in this study, has been

1\1I ell ~g_

r.oo - - - - - - - - - - - - iiLSG--
II IIL70 m -=:J
- I
'0 IIL38A
rn 1m : -- 500


L IIL7 0 E 111- . ~1=

e) r. _ ==::I!iilI_ _ J O . 300
e) o

'Ih1\ ~"K
>00 .

J ~\I~I "Jl" I' 1,,1'0' rL LIlI


O . - ' - -- - - - - - - - ---------'

---- Sample Depth ---- Sa mple De p t h ---.

til I\s e) Pb
20 0 -
n 0 III ====:l
"" _=:J
r. 0
.,0 ;,~
~. 0
-tJ 10 . 0 -

~ II U 8A .~
0 rIl

" n~~~,
1 0

> 0

, 0

o 0 0 .0

Samp l e Dept h ---- ---- Sample De pt h

40 . 0
Ba r) ell
20 -
liClllJ:l!lII J
ilL _llm
II L38A _

]0 0 -

I J ?O 0 - .W \. 0 -
<: 3:

'0 0
1Il:1 -==:J

---- Sa mp le De p th ---- Sample Depth ---.

Fig.8 Geochemical profiles for HL38A, HL70 and HLS6 illustrating

both vertical and lateral trends for a) Au, b) As, c) Ba, d) Ag,
e) Pb and f) Cu. Note that there is a separation of SOOm along
the axis of the deposit between HL38A in the south and HLS6 in
the north. Shaded sections indicate intervals containing more
than 4 ppm Au.

interpreted as an extension of a fragments embedded in a matrix of
barite capping immediately to the east yellow-brown sphalerite. The zoned
(Fig. 1). crystals contain abundant sub-micron
inclusions of chalcopyrite, described
al Fe (In sone I er 1 t e) as 'chalcopyrite disease' by Barton 2
8 0 Similar crystals have been reported in
HL38A the Kuroko ores of Japan, although
7 0
HL70 without evidence of stratigraphic
60 HL56
cont rol, and Eldr idge et al., 6 have
MEl described them as facies 2 sphalerite
5 0 (recrystallized by 250 0 C fluids) in
~ terms of their maturation model for
.w 0
3: Kuroko deposits.
30 -

2 0

1. 0 Figure lOa shows an unbroken zoned

crystal viewed in transmitted light;
0 .0 opaque 'growth' zones are conspicuous.
Detailed electron microprobe analyses
---- Sampl e Depth ---. across this crystal, however, show no
difference in composition between
b) Zn (in assay)
'0 0 -r--riT'i"0A"- - - ' - . . : . - - - - - - - - -- --, translucent and opaque zones, and it
1lL3 A
is concluded that the opacity arises

from 'chalcopyrite di$ease' concen-
trated along growth planes below the
30 0
"""",,----"-,, O====~
H L7.=..

0 .0 - ' - - - - - - - -J

- -- Sample Depth - - -.-

Fig.9 Down hole profiles illustrating

both the vertical and lateral trends
in a) the average Fe-content of
sphalerite and b) the Zn-content of
the ore. Shaded sections indicate
intervals containing more than 4 ppm

Examination in transmitted light of

polished thin-sections from HL38A
shows that sphalerite in the ore is Fig.lO a) Photomicrograph in trans-
commonly homogeneous and varies in mitted plane polarized light showing a
colour from straw yellow to almost crystal of concentrically zoned high-
colourless as the Fe-content decreases Fe sphalerite in a matrix of high-Fe
from about 3.0wt% (typical of the sphalerite from the base of the upper
deposit) to about 0.5wt%Fe (in the unit. b) Backscatter electron image
barite zone at the hanging wall). of the same crystal showing minor
Sphalerite in the zone of high Fe- chalcopyrite (dark grey) concentrated
content, however, is markedly on the growth planes. Other minerals
different and consists of concen- visible are pyrite (black) and galena
trically zoned crystals and crystal (white). Scale bar = 100 11m.

surface of the section. These and a new vent (or vents) opened up
intersect the surface as very thin further south. A new mound with a
concentric zones as can be seen from barite capping formed above this. The
the backscatter electron image zone of high-Fe sphalerite contains
(Fig.lOb). The width of the opaque crystals expelled from this vent and
zones in transmitted light is thus deposited on the sea floor as clasts.
determined by the inclination of these These spread over the slopes of the
planes to the surface of the thin original mound, burying attenuated
section. sections of the barite capping in the
vicinity of HL70 and thinning out
further north beyond HL56.
This interpretation is consistent with
the conclusions of Barton 2 and Scott 26
that 'chalcopyrite disease' is brought A model that well fits this hypothesis
about by the reaction of Cu-bearing is deposition from a plume of buoyant
solutions with the FeS component of sulphide forming solutions, as
sphalerite. Perhaps copper in excess proposed by Solomon and Walshe 2 9
of the equilbrium solubility was Maturation of the deposit, as
concentrated in regions of high envisaged in the Kuroko model of
dislocation densities as suggested by Eldridge et al.,6 and Pisutha-Arnond
Tauson et al.,33. Such regions may be and Ohmoto 2 4), appears to have been
expected to exist at the crystal limited to the formation of facies 2
surface. The zonal pattern suggests assemblages ( recrystallization
episodic growth which could indicate and coarsening by fluids at about
that the crystals formed in a vent. 250 oC). Other models for the flow of
hot saline solutions from vents on the
sea floor have been reviewed by Turner
Although data from only three holes is and Gustaffson 35
not sufficient to prove lateral
continuity of the high-Fe sphalerite
zone, nevertheless it provides a At low temperatures and pressures,
strong indication that a marker phase relations in the Fe-Zn-S system
horizon is present which divides the (Scott 26) indicate that the Fe-content
orebody into lower and upper units. A of sphalerite should be highest at low
plausible interpretation can be sulphur activity (the onset of
developed as follows:- solfataric activity) and decrease as
the sulphur activity increases during
construction of the mound. The
1. The lower unit was derived from a relatively high Fe-content of
submarine hydrothermal vent (or vents) sphalerite at the base of the upper
somewhere between HL70 and HL56, where unit, therefore, is to be expected
extensive stringer mineralization from resurgence of hydrothermal
underlies the deposit (Fig.l). The activity at a new vent, and is
mound of ore that formed on the sea consistent with the generally held
floor above this had a bar i te capping belief (Moles 2 2) that Fe-rich
that thinned over the area now sphalerites of submarine hydrothermal
intersected by HL70 (Fig.l). That such origin have precipitated from the
mounds can form in this environment is metalliferous solutions during their
by now well established from study of expulsion from the hydrothermal vents
modern day hydrothermal activity and carried in suspension in the
(Goldfarb,lO,Graham et al.,ll; laterally spreading brines.
Hannington and Scott,13; Marchig et
al.,20; Thompson et al.,34) . They grow
and mature by way of a complex Decrease in Fe-content of sphalerite
succession of precipitation and with stratigraphic height in Kuroko
replacement processes in response to deposits has been reported by Urabe 36,
the waxing and waning of a thermally who ascribed the trend largely to a
intensifying system; more or less as decline in temperature, and by
envisaged in the Kuroko model of Bryndzia et al .,3. However, as noted
Eldr idge et al., 7. above, th is trend may also be related
to an increase i n total sulphur
concentration (and oxygen fugacity)
2. At some stage the activity ceased which also favours the formation of

barite (Finlow-Bates 7). Such a was likely transported as Au(HS) 2 and
systematic decrease in Fe-content that high gold grades are promoted by
through to the barite ore is clearly near neutral oxidised and low
seen within the upper unit of the temperature (150 - 27S oC) conditions.
Hellyer deposit. Huston and Large 1 4 have proposed
further that partial oxidation of
reduced sulphur caused by mixing with
~ precious metals seawater may remobilize this gold
along the seawater interface to distal
parts of the deposit in the form of
With the exception of minor gold at AUS~ (n>2). This model can well
the top of the lower unit (associated explain the higher gold grades in HLS6
with arsenopyrite in HL38A and with and would indicate that the auriferous
barite in HL70) most of the precious arsenopyrite crystallized within the
metals mineralization is found within upper parts of the mound close to the
the upper unit. sea floor rather than within the
buoyant plume itself.

Gold Although the average Sb-content of the

arsenopyrite appears not to vary
significantly from hole to hole (Table
Cathelineau et al.,S have shown, by III) there is a very clear diminution
means of ion microprobe analysis, that in both the frequency of the cryptic
the Au-content of hydrothermal zoning and its compositional range
arsenopyrites commonly ranges up to from south to north. This is
1000 ppm, and that in most cases the consistent with a change from
gold is heterogeneously distributed turbulent conditions of crystal-
within growth zones or overgrowths in lization near the hydrothermal vent
the crystals due to an antipathetic (HL38A) to quiet conditions in remote
relationship between Au and the Sb- parts of the plume (HLS6). Overall,
content. Concentrations as high as the comparatively uniform conditions
0.44 wt% Au in arsenopyrite have been of the distal environment would be
reported by Cabri et al.,4, but even expected to favour incorporation of
at these levels the gold is invisible gold into the arsenopyrite compared
to high resolution electron microscopy with the turbulent proximal environ-
indicating that it is probably in ment where co-precipitation of gold
solid solution. The average Au-content would be inhibited, by the
in arsenopyrite is thus highly antipathetic behaviour of gold and
variable within a deposit, as a antimony, during the frequent
function of the distribution of Au- excursions to higher Sb-content,
rich grains. Cathelineau et al.,S have
also shown that Au-rich arsenopyrites
crystallize for the most part at quite Silver
low temperatures (170 - 2S0 oC).

Natural tetrahedrite shows complex

The results of the present study find compositional variation in which As
ready interpretation in the light of and Sb display extensive mutual
these data. The Au-contents of the substitution, and elements such as Fe,
arsenopyrite determined with the Zn, Hg and Ag commonly substitute for
electron microprobe fall within the Cu. Because of this complexi ty, it is
range cited by Cathelineau et al.,S not possible to use the composition as
and the patchy distribution of gold a guide to the physico-chemical
noted by these authors can be well conditions of formation, despite
understood in terms of the complex extensive experimental studies
oscillatory variation in Sb-content .
( Sklnner et al., 26 , tuce et al., 18 ,
described in the present paper. Tatsuka and Morimoto 31,32, Makovicky
and Skinner 1 9), since few of the
equilibrium reactions can be directly
Thermodynamic modelling of the applied to natural mineral assemb-
hydrothermal fluid (Large 16, Eldrid%e lages. Nevertheless, systematic
et al.,6, Pisutha-Arnond and Ohmoto 2 ) compositional variations within the
indicates that gold in the upper part Hellyer tetrahedrites imply that Ag-
of zinc-rich massive sulphide lenses enrichment at the hanging wall is the

result of late stage Sb-metasomatism. a) Ag l in Te t r a ne dr tel
HL 56
On average, the most argentian HL 70
tetrahedrites are found at the top and 20 .0
bottom of the upper unit (Fig.lla).
They are also quite arsenical in
HL38A, whereas further north, in HL70
and HL56, the As-content is
negligible (Fig.llb). When the ' 0 .0

individual analyses are taken into

account, three populations can be
distinguished in HL38A: 1) tetra-
hedrite, 2) argentian tetrahedrite and
3) arsenical tetrahedrite (Fig.12) ; o.o -L- - - - - - - - - - - - - - - ---'

the argentian and arsenical varieties
occur only in the upper unit. These ---- Sample De p th
distinctions cannot be made in HL70
and HL56 where the tetrahedrite shows b) As (i n Te tr ah e dr i t el
a continuous range of compositions
through to freibergite (Fig.12). The HL38A
division between freibergite and '0 .0

argentian tetrahedrite is here taken

at 20 wt%Ag, following the class-
ification of Ril e y25. As in HL38A,
trace tetrahedrite in the lower unit
contains only minor silver in HL70. No
tetrahedrite was observed in the lower HL70
unit in HL56. HL56

0 .0 -1- '- -'

Figure 13 shows the tetrahedrite
analyses plotted in terms of the
principal 'end member' components. ---- Sa mple Depth -----
Such plots, adapted here from the
" Mo d i f i e d Johnson Spinel Prism" Fig.ll Down hole profiles showing
described by Stevens 30 and Haggerty12, both vertical and lateral variations
provide a useful summary of the in a) the average Ag-content and b)
complex compositions. In HL38A, for the average As-content of tetra-
example, the composition of the hedrite. Shaded sections indicate
arsenical tetrahedrite trends along intervals containing more than 4 ppm
the back plane of the projection and Au.
can be understood primarily in terms
of substitution of As for sb , This
population is sUbstantialy diminished
in HL70 and is absent in HL56. The
argentian tetrahedrites, which trend However, in HL70 and HL56, grains of
along the front triangular surface, highly argentian tetrahedrite and
have a more complex composition freibergite in the Ag-rich parts of
involving significant Fe as well as the deposit are coarse (lOO~m) and
Ag subsitution in HL38A but associated with recrystallized
comparatively simple compositions with galena. Such grains are inhomogeneous
negligible Fe subsitution in HL70 and and show large variations in Ag-
HL56. Similar trends are apparent content due to complex zoning
with regard to Zn substitution (not (Fig.6a). Similar variability has been
plotted) reported for tetrahedrites in the
Kuroko deposits of Japan (Shima-
zaki 2 7). The source of additional
In HL38A, the most argent ian silver in the Hellyer deposit is
tetrahedrite occurs as 5 ~m inclu- believed to have been argent ian galena
s ions in low-Fe sphaler i te in the that has been replaced by non-
barite-ore, implying that a primary argentian boulangerite and bournonite
trend towards Ag-enrichment existed at the hanging wall. The present
during accumulation of the deposit. bimodal distribution of silver and its



Cul25b4 51 3
v _ _v._-",-_-, A9
Cu _ -->L-------'' ' ' -v. \I

Agl 2AS451 3
Ag l25b 45 13

HL 70

Cu 12511 45 13

Agl2As4 51 3

A9125b 45 13
_ _v_ _v_ _V
\I V

Cu 125 11 451 3

. iH'gpnlian Agl2As4 S 13
.,~' ar snn i cn 1 Agl25b4S13

Fig.13 Tetrahedrite compositions

plotted in the 6-component prism
Cu Ag
cUl2Sb4Sl3 - Fe12sb4sl3 - Agl2 Sb4Sl3 -
Wt % CUl2As4Sl3 - Fel2As4sl3 - Ag l2 As4 Sl3
showing variation in trends from south
(HL38A) to north (HL56) in the
Fig.l2 Tetrahedrite compositions orebody.
plotted in the ternary diagram Cu-Ag-
Sb showing variation in trends from
south (HL38A) to north (HL56) in the

northwards enrichment at the hanging precipitation of gold in arsenopyrite
wall may be a consequence of of higher Sb-content being inhibited
metasomatic solutions moving along the by the antipathetic behaviour of gold
top and bottom contacts of the upper and antimony.

4) Tetrahedrite is the dominant host

The Ag-content of galena in HL56 for silver, but galena is also
parallels the bimodal distribution argent ian. Recrystallization assoc-
shown by the Ag-content of the iated with late stage Sb-metasomatism
tetrahedrite. The maximum Ag-content, resulted in formation of highly
however, is much lower 0.1 wt%) than argentian tetrahedrite and freibergite
in the presumed vent area to the south at the hanging wall. Replacement of
despite the higher Ag-content of the argent ian galena by non-argentian
associated tetrahedrite. This is bournonite and boulangerite provided
consistent with the hypothesis, the additional silver.
outlined above, that silver has been
released from the galena through
replacement by Sb-rich minerals.

CONCLUSIONS The authors wish to thank the

Directors of Aberfoyle Limited for
providing access to the Hellyer
It is concluded that: deposit at an early stage of
exploratory drilling and development,
1) The Hellyer volcanogenic massive and for financial assistance towards
sulphide deposit can be divided into conduct of the research. Drs W.L.
upper and lower units, representing Griffin, J.R. Wilmshurst and Mr G.J.
accumulation of sulphides from McArthur are thanked for reviewing the
different hydrothermal vents. A zone manuscript.
of high-Fe, in part clastic,
sphalerite at the base of the upper
unit over-rides the lower unit and
marks the onset of hydrothermal REFERENCES
activity from a new vent further
1. Amcoff O. The solubility of silver
and antimony in galena. ~ Jahrbuch
2) Most of the precious metals ~ Mineralogisch~ MQng~shefte, Heft
mineralization is associated with the 6,1976, p. 247-261.
upper unit.
2. Barton P.B.Jr. Some ore textures
involving sphalerite from the Furotobe
3) Arsenopyrite is the dominant host mine, Akita Prefecture, Japan. Mining
for the gold but an association with ~ology, vol. 28, 1978, p. 64-72.
pyr i te is infer red and independent
evidence suggests that minor free gold
may occur in the barite ore. The
arsenopyrite is believed to have 3. Bryndzia L.T., Scott S.D. and Farr
crystallized at the unconsolidated J.E. Mineralogy, geochemistry and
surface of the mound close to the sea mineral chemistry of siliceous ore and
water interface where gold may have altered footwall rock in the Uwamuki 2
been present as AU(HS)j, in the and 4 deposits, Kosaka mine, Hokuroko
proximal facies, or AUS n (n>2), in district, Japan. In: ~ ~QkQ ~nQ
the distal facies. Cryptic oscillatory related volcanogenic massive sulphide
zoning of the arsenopyrite reflects ~ ~ . Economic Geology Monograph
fluctuations in the Sb-content of the No.5, 1983, p. 507-522.
fluids. These fluctuations were
greatest in the proximal facies and 4. Cabri L.J., Chryssoulis S.L., De
least in the distal facies. This may, Villiers J.P.R., Laflamme J.H.G. and
in part, explain the lower gold grades Buseck P.R. The nature of "invisible"
of ore in the proximal facies, co- gold ina r sen 0 p y r it e ~ngQ.i~n

Mineralogist, vol. 27, 1989, p. 353- 13. Hannington M.D. and Scott S.D.
362. Mineralogy and geochemistry of a
hydrothermal silica-sulfide-sulfate
5. Cathelineau M., Boiron M.C., spire in the caldera of axial
Holliger Ph. and Marion Ph. Gold-rich seamount, Juan de Fuca ridge. Canadian
arsenopyrite: Crystal chemistry, gold Mineralogist, vol. 26, 1988, p. 603-
location and state, physical and 625.
chemical conditions of crystal-
lization. In: Bicentennial ~ aa, 14. Huston D.L. and Large R.R. A
Symposium sponsored by The Australian chemical model for the concentration
Institute of Mining and Metallurgy, of gold in volcanogenic massive
Melbourne, 1988, p. 235-240. s u I phi d e de po sit s Q.t:.. ~.QIQgy
Revieli~, vol. 4, 1989, p , 171-200.
6. Eldridge C.S., Barton P.B.Jr. and
Ohmoto H. Mineral textures and their 15. Koski R.A., Shanks W.C., Bohrson
bearing on the formation of the Kuroko W.A. and Oscarson R.L. The composition
orebodies. In: ~ Kuroko ~ related of massive sulfide deposits from the
volcanogenic massive sulphide depo- sediment-covered floor of Escanaba
~ . Economic Geology Monograph No. Trough, Gorda Ridge: Implications for
5, 1983, p, 241-281. d e po sit ion a l p roc e sse s ~n~in
Mineralogist, vol. 26, 1988, p. 655-
7. Finlow-Bates T. The chemical and 673.
physical controls on the genesis of
submarine exhalative orebodies and
their implications for formulating 16. Large R.R. Chemical evolution and
exploration concepts. A review. zonation of massive sulfide deposits
Geologische Jahrbuch, vol. D40, 1980, in volcan i c te r ra ins. ~~QnQmQ.i.~
p. 131-168 Geology, vol. 72, 1977, p , 549-572.

8. Foord E.E. and D.R. Shawe. The Pb- 17. Large R.R., Huston D.L.,
Bi-Ag-(Hg) chemistry of galena and McGoldrick P.J. and Buxton P.A. Gold
some associated sulfosalts: a review distribution and genesis in Australian
and some new data from Colorado, volcanogenic massive sulfide deposits
California and Pennsylvania. Canadian and their significance for gold
Mineralogist, vol. 27, 1989, p. 363- transport models. Econom~ Geology,
381. Monograph 6, 1990, p. 512-518.

18 . Luce F.D., Tuttle C.L. and Skinner

9. Franklin J.M., Lydon J.W. and B.J. Studies of sulfosalts of copper:
Sangster D.F. Volcanic-associated V. Phases and phase relations in the
massive sulphide deposits. Econom~ system Cu-Sb-As-S between 350 0 and
~.QIQgy, 75th Anniversary Volume, 500 0C. ~~QnQmi~ ~.QIQgy, vol. 72,
1981, p , 484-627. 1977, p. 271-289.

10. Goldfarb M.S., Converse D.R., 19. Makovicky E. and Skinner B.J.
Holland H.D. and Edmond J.M. The Studies of the sulfosalts of copper.
genesis of hot spring deposits on the VI. Low-temperature exsolution in
East Pacific Rise 21 0N. In: ~ Kuroko synthetic tetrahedrite solid solution
n~ ~~l~.~ m~~iY. CU12+xSb4+yS13. Canadian Mineralogist,
sulphide deposits. Economic Geology vol. 16,1978, p. 611-623.
Monograph No.5, 1983, p. 184-197.
20. Marchig V., Rosch H., Lalou C.,
11. Graham U.M., Bluth G.L. and Ohmoto Brichet E. and Oudin E. Mineralogical
H. Sulfide-sulfate chimneys on the zonation and radiochronological
East Pacific Rise, 11 0 and 13 0 relations in a large sulfide chimney
latitudes. Part 1: Mineralogy and from the East Pacific Rise at 18 025'S.
paragenesis. Canadian Mineralogist, Canadian Mineralogist, vol. 26, 1988,
vol. 26, 1988, p. 487-504. p. 541-554.

12. Haggerty S.E. Opaque mineral 21. McArthur G.J. Hellyer. In: Geology
oxides in terrestrial igneous rocks. 9.rul Mi n era 1 ~ sou r c e s Q!.
In: Oxide Minerals, Mineralogical Chapter 4, 1989, p. 144-148.
Society of America, Short Course
Notes, 1976, p. 101-300. 22. Moles N.B. Sphalerite composition
in relation to deposition and

metamorphism of the Foss stratiform 32. Tatsuka K. and Morimoto N.
Ba-Zn-Pb deposit, Aberfeldy, Scotland. Tetrahedrite stability relations in
Mineralogical Magazine, vol. 47, 1983, the CU-Fe-Sb-S system. AID~.I.i~,gn
p. 487-500. Mineralogist, vol. 62, 1977, p, 1101-
23. Paradis S., Jonasson I.R., Le
Cheminant G.M. and Watkinson D.H. Two 33. Tauson V.L., Makeyev A.B., Akimov
zinc-rich chimneys from the plume V.V. and Paradis L.F. Copper
site, souther Juan de Fuga ridge. distribution in zinc sulfide minerals.
Canadian Mineralogist, vol. 26, 1988, Ge 0 k h i IDll , No.4, 1 9 8 8, p. 4 9 2 - 5 0 5
p, 637-654.
34. Thompson G., Humphris S.,
Schroeder B., Sulanowska M. and Rona
24. Pisutha-Arnond V. and Ohmoto H. P.A. Active vents and massive sulfides
Thermal history, and chemical and at 26 0N (TAG) and 23 0N (SNAKEPIT) on
isotopic composition of ore-forming the mid-Atlantic ridge. ~~n~gi~n
fluids responsible for the Kuroko Mineralogist, vol. 26, 1988, p. 697-
massive sulphide deposits in the 711.
Hokuroko district in Japan. In: ~
KY~QAQ ~ng ~~~~~~g YQ~~~nQg~ni~ 35. Turner J.S. and Gustaffson L.B.
massive sulphide deposits. Economic The flow of hot saline solutions from
Geology Monog r aph , No.5, 1983, p. vents on the sea floor - some
523-558. implications for exhalative massive
sulfide and other ore deposits.
25. Riley J.F. The tetrahedrite- Econom~ Geology, vol. 73, 1978, p.
freibergite series with reference to 1082-100.
the Mt Isa Pb-Zn-Ag orebody.
Mineralium Deposita, vol. 9, 1974, p. 36. Urabe T. Iron content of sphal-
117-124. erite co-existing with pyrite from
some Kuroko deposits. In: Geology Qf
26. Scott S.D. Chemical behaviour of Kuroko Deposits. Mining Geology, Tokyo
sphalerite and arsenopyrite in Special Issue No.6, 1974, p. 377-84.
hydrothermal and metamorphic environ-
ments. Mineralogical Magazine, vol. 37. Van Hook H.J. The ternary system
47,1983, p , 427-435. A9 2S-Bi 2S 3 -PbS. Economic Geology, vol.
55,1960, p , 759-788.
27. Shimazaki Y. Ore minerals of the
Kuroko-type deposits. In: Geology Qf 38. Yamaoka K. Metallic minerals of
Kuroko Deposits. Mining Geology, Tokyo the Kuroko deposits in Northeast
Special Issue No.6, 1974, p , 311-322. Japan. In: Proceedings SYmRQiym Qn
':Min~~~~ ~Qn~~i~Y~n~~ Q.f. K!!.~QAQ
28. Skinner B.J., Luce F.D. and Deposits ~ Geochemistry Qf Sulfide
Mackovicky E. Studies of the Minerals ~ Hydrothermal Deposits",
sulfosalts of copper. III. Phase and Morioka, 1969, p. 1-38.
phase relations in the system Cu-Sb-S.
Econom~ Geology, vol. 67, 1972, p.

29. Solomon M. and Walshe J .L. The

formation of massive sulphide deposits
on the sea floor. Econom~ Geology,
vol. 74, 1979, p. 797-813.

30. Stevens R.E. Composition of some

chromites of the Western Hemisphere.
American Mineralogist, vol. 29, 1944,
p. 1.

31. Tatsuka K. and Morimoto N.

Tetrahedrite stability relations in
the Cu-Sb-S system. Economic Geology,
vol. 72,1977, p, 258-270.

Mineralogy and petrology of the lead-zinc-copper sulphide
ores of the Viburnum Trend, southeast Missouri, U.S.A.,
with special emphasis on the mineralogy and extraction
problems connected with cobalt and nickel
Richard D. Hagni
Department of Geology and Geophysics, University of Missouri-Rolla, Rolla, Missouri, U.S.A.

SYNOPSIS character of the ore textures include zinc in

The ore deposits of the Viburnum Trend in the the lead concentrates; magnesia, cobalt, and
Southeast Missouri Lead District form the nickel in the zinc concentrates; lead in the
world's largest lead producing district. The copper concentrates; lead, zinc, and copper
district is comprised of el even mines, ten of los ses to the ta ilings. Ore microscopic study
which are currently in production. The of the ores and beneficiation product s
principal minerals are few in number, simple in contribute to a better understanding of the
composition, and form simple superposition and mineralogical and textural characteristics that
replacement textures. The economically valuable cause these metallurgi cal problems.
minerals are the sulphides , galena, sphalerite, CODalt and nickel minerals are present in
and chalcopyr ite; silver is recovered from small amounts in the Viburnum Trend ores, and
galena and cadmium from sphalerite at the lead they comprise the second largest reserve of
smelter and zinc refinery. The most abundant cobalt in the United States. Cobalt occurs
gangue minerals are dolomite, pyrite, marcasite, primarily in siegenite; nickel is present mainly
calcite, and quartz . The Viburnum ores, in siegenite and bravoite. Although small
however, contain a diverse assemblage of amounts of cobalt and nickel are present in the
minerals uncommon in most Mississippi Valley- ores at all of the mines, their concentrations
type ore deposits . These include: siegen ite, vary throughout the district. The greatest
bravoite, vaesite, fletcherite, nickelean concentration occurs in the Fred~icktown
carroll ite, bornite, gersdorffite, tennantite, subdistrict where the ores average 0.23% cobalt.
enargite, luzonite, millerite, polydymite, Within the Viburnum Trend, cobalt and nickel are
chalcocite, digenite, anilite, djurleite, most abundant in its central portion at the
covel lite, blaubleibender covellite, pyrrhot ite, Magmont and Buick mines .
magnetite, anhydri te, and dick ite. In r ecent ly discovered ores at the
The valuable ore minerals exhibit a variety Magmont-West mi ne , nic kel i s locally
of open space and r epl acement textures. Most concentrated in ni ckel -ar seni c sulphide ore s
gra ins of galena, sphalerite, and chalcopyrite along portions of a linear breccia ore deposit.
are sufficiently coarse to provide an adequate This deposit is unique in that it occurs two
separation by flotation at a grind of abou t 60% miles west of the main Viburnum Trend and i n its
minus 200 mesh. Most mills in the district location close to sulfide deposits in the
produce three flotation concentrates: lead, underlying Precambrian basement. The Magmont-
zinc, and copper. Portions of the ores are West ores are predominantly collomorphic in
fine-grained or intricately intergrown and character and composed of millerite,
present a variety of recovery, concentrate gersdorffite, polydymite, and vaesite, partly
grade, and deleterious constituent problems. rep1aced and i ntergrown with chalcopyrite ,
Beneficiation problems that result from the sphalerite, galena, and quartz.
Although cobalt and nickel formerly were production currently is derived from the
recovered in the Fredericktown subdistrict, the district . A total of about 380,000 short tons
metals have not yet been recovered at mines in of recoverable lead metal is recovered annually
the Viburnum Trend nor elsewhere in the from the district. 1
Southeast Missouri Lead District. The recovery Important amounts of additional metals are
problems are related to the fine-grained recovered as by-products from the lead ores at
character, coating and intergrowth textures, and the mill and smelter. About 40,000 short tons
flotation properties of siegenite. Current of zinc metal recovered from the district places
research efforts to recover cobalt and nickel Missouri third in U.S. zinc production. Nearly
involve fine grinding, flotation, and separation 15,000 tons of copper metal places Missouri
by pyrometallurgical procedures. sixth among copper-producting states. The
recovery at the smelters of about 1.2 million
troy ounces of silver annually places Missouri
The Southeast Missouri Lead District has been seventh among silver-producing states . The
the largest producer of lead in the Un ited annual value of these four metals currently
States since 1907, except for 1962 when a strike amounts to about 400 million U. S. dollars.
disrupted production. It has been the world's The ore deposits are distributed in f ive
largest producer since 1970. More than 90%of subdistricts as shown in Figure 1 from
United States and about 15%of world lead 2
Kisvarsanyi. The most productive subdistricts

'_0..0-;)1 .;~
/ .J' \ ~

-, .:,' ~:,-.
"" , Bonne Tene :TWleS
~n'Wl8'$......" . " .: . \ ~ '"
/!~ ~.. . Flat ~ nw>e
~ .
Czar Knob I ' , No 29
~:.~~ :-,.., " :' {0lD LEAD BELTI
. __
No 27 : ~ , .' ::.'; '.' J ' -...

,..... ~ .. . ,/
.. ( mine
t . :1

o s
= kin

- Maplaull

Fig. 1. Map of Southeast Missouri Lead District showing subdistricts, mines,

mineralized areas, and Precambrian outcrops (modified from Kisvarsanyi 2).

are the old Lead Belt that was mined from 1865 currently is purchasing all of Homestake's
to 1972. and the Viburnum Trend or new Lead Belt holdings.
that was first mined in 1960 and from which all The stratigraphy. structure. forms of ore
of the current production i s derived. Smaller bodies, and ore controls recently have been
subdistricts include Indian Creek. Frederick- reviewed by Hagni 4 who included numerous
town. and Annapolis. Current exploration is references on specific aspects. The exploration
centered on the Winona area. southwest of the history in the district has been reviewed by
Viburnum Trend. Ohle and Gerdemann. 5 Recent descriptions of the
Total production from the older subdistr icts geology of individual mines i ncl ude: Viburnum 27
in the Southeast Missouri Lead District was 220 (Grundmann 6). Viburnum 28 (Pettus and Dunn
million metric tons of ore. 3 Production to date Casteel (Dunn and Grundmann 8). Magmont (Sweeney
from the Viburnum Trend totals about 140 million et al. 9 ), Magmont-West (Bradley and Krolak10),
metric tons of ore. In contrast to ores from Buick (Rogers and Davisl l; Mugel and Hagni I2).
the old Lead Belt. the principal district prior Brushy Creek (Evans 13 ) West Fork (Dlngess
to the discovery of the Viburnum Trend. that had Fletcher (Paarlberg and Evans 15). Sweetwater
an average grade of about 2.8% lead. 3 the ores . 16
(Mouat and Clenden ln ; Larsen et al. 17;
from the Viburnum Trend average 5.8 % lead. 0.8 % Walker I8). The sedimentary facies and their
zinc. 0.14 % copper. and about 1/4 ounce silver influence have been discussed by Larsen 19 and
per ton of ore. 4 Gregg and Gerdemann . 20
The Viburnum Trend extends for a distance of
about 72.5 km in a north-south direction. The MINERALOGY
town of Viburnum is about 8 km south of the The principal sulphide minerals in the ores of
northern end. Ten mines currently are operating the Viburnum Trend are galena, sphalerite.
in the Viburnum Trend . Their initial mining chalcopyrite, marcasite, and pyrite. All of the
dates are given below beginning at the north end sulphide minerals were repetitively deposited
of the Trend. Mining in the Trend began in 1960 and occur in a variety of forms. Galena occurs
at the Viburnum No. 27 mine. a deposit that as massive replacements of the hostrock
subsequently has been exhausted. The Viburnum dolomite. as replacements of earlier deposited
28 and 29 mines were opened in 1962 and 1964 . sulphide minerals. and as euhedral crystals
The Casteel or Viburnum 35 mine. where mining deposited in vugs. Sphalerite occurs dominantly
began in 1983, has mined unusually copper-rich as relatively fine-grained replacements of
ores . Although the Magmont mine began in 1968, hostrock dolomite, but small, subseGuently
mining to the west from the Magmont-West mine deposited. vug-1in ing crystal also are present.
began in 1979, and the latter has produced from Chalcopyrite occurs commonly as massive to
a zinc-rich deposit. The Buick mine opened in colloform replacement masses. but also as
1969. Brushy Creek was opened in 1973 and was subhedral crystals disseminated in host rock
temporarily shut down from 1986-1988. West dolomite and as euhedra1 crystals in vugs.
Fork. the newest mine in the Trend, was opened Marcas ite and pyrite form small dtss ennnateo
in 1985. Mining began at the Fletcher mine in crystals, large co110form masses, and euhedra1
1967. The Sweetwater mine. formerly called the crystals in vugs.
Ozark Lead mine and the Milliken mine and The main gangue minerals. in addition to
located at the south end of the Trend. was marcasite and pyrite. are dolom ite, quartz, and
opened in 1968. closed in 1983. and reopened in calcite. The host rock Cambrian Bonneterre
1987 . The West Fork and Sweetwater mines are 1imestone has been dolomitized nearly everywhere
operated by ASARCO. Magmont and Magmont-West that it contains ore deposits. Dolomite also
mines by Cominco American. and all other mines was deposited as euhedral crystals in vugs in
by Doe Run Company. Doe Run Company. created in association with the sulfides. Cathodoluminesc-
1986, has been owned 57.5 % by Fluor Corporation ence microscopic study of those dolomite
and 42.5 % by Homestake Mining. but Fluor crystals by Voss and Hagnl.21 and Voss et dl.
. 22

shows that they have been deposited during four marcasite, and dolomite crystals. The
periods that are closely associated with the octahedral galena may be partially or completely
periods of deposition of the ore sulphides. leached and leave behind the coatings with empty
Quartz is most abundant at the Magmont and Buick casts.
Mines where it occurs as jasperoid replacements
of dolomite and as euhedral crystals formed in ORE TEXTURES AND RELATED MINERAL BENEFICIATION
vugs. 23 Quartz also was deposited during the PROBLEMS
span of sulfide mineral deposition, and it The Viburnum ores exhibit a wide variety
contains fluid inclusions, studied by Roedder 24 textures that result from combinations of open
and Hagni,25 that indicate it was deposited from space filling and replacement processes. Open
ore fluids with elevated temperatures. Calcite space filling textures are characterized by the
forms euhedral crystals that were deposited in simple superposition of one mineral upon another
vugs after the sulphide mineral deposition had in vugs, fractures, and spaces between breccia
largely ceased . fragments. Replacement textures include those
Although the principal minerals are few in in which host rock dolomite crystals are
number and simple in composition, many minor replaced by sulphides and those in which earlier
minerals of more complex compositions also are sulphide minerals are replaced by later
locally present in the Viburnum ores. Many of sulfides. Typical open space and replacement
these minerals are ones that are not commonly textures have been illustrated previously by
associated with Mississippi Valley-type ore Hagni and Trancynger. 30 Partial replacements,
deposits. They include: siegenite, bravoite, fine sulphide grain sizes, and other intricate
vaesite, fletcherite, nickelean carrollite, intergrowths tend to develop locked particles
bornite, gersdorffite, tennantite, enargite, when these ores are ground prior to
luzonite, millerite, polydymite, chalcocite, beneficiation. The partitioning of such
digenite, anilite, djurleite, covellite, particles is difficult to control during the
blaubleibender covellite, pyrrhotite, magnetite, subsequent flotation steps. Beneficiation
anhydrite, and dickite . practice at the mills in the Viburnum Trend has
Zinc sulfide with wurtzite morphology has recently been reviewed by Watson,31 and the role
recently been found to occur in bands of mineralogy in beneficiation problems in the
interlayered with fine-grained normal Viburnum Trend has been discussed by Hagni.32
sphalerite. 26 Such schalenblende zinc sulfides Six examples of current beneficiation problems
are moderately abundant at the new West Fork in the Viburnum Trend, which are a function of
mine in the Viburnum Trend where they occur as ore mineral textures, are given in the following
one to three mm thick yellow bands that are paragraphs.
deposited on the surface of colloform masses of Galena, the most abundant ore mineral in the
sphalerite. They are composed of radiating Viburnum Trend, is deposited during two main
acicular crystals with much porosity between the periods, an earlier octahedral generation, and a
crystals. later cubic generation. 22,32 Both generations
Although barite is absent from the Southeast of galena were deposited after most sphalerite
Missouri Lead District, a major barite district, grains, and locally galena exhibits a preference
the Washington County Barite District, is to replace the earlier sphalerite grains.
located immediately to the north of the lead Partially replaced sphalerite grains form
district. Recent descriotions of the barite intricately locked sphalerite-galena grains that
deposits include Wharton 27 and Kaiser 28 result in increased zinc contents of the lead
All of the sulfide minerals and calcite are concentrates and losses of lead to the zinc
locally leached, but the most prominent leaching concentrates.
25 29 Most of the sphalerite grains were deposited
is that of galena.' Galena that belongs to
the octahedral generation has been leached after early in a disseminated texture that formed
the deposition of crusts of chalcopyrite, principally by replacement of the host rock

do10mite. 30 This feature, together with the concentrates from portions of some ore bodies in
fact that sphalerite forms smaller grains than the Viburnum Trend have had to be st ockpi l ed
those of galena, causes abundant locked because of cobalt and nickel contents.
~pha1erite-do10mite particles in the zinc Fine-grained sulphide replacements of host
concentrate. 33 Because the magnesium from the rock dolomite are locally present . Galena,
dolomite is a deleterious constituent at the sphalerite, and chalcopyrite may form grains as
electrolytic zinc refineries, the refineries small as 10 ~m that are disseminated through
limit the magnesium content to 0.6% in the zinc dolomite crystals. Although these intergrowths
concentrates. Until recently, all zinc are not common, where they are present they form
:oncentrates in the Viburnum Trend had to be locked particles upon grinding. During
treated by sulfuric acid digestion to diminish flotation these locked particles carry lead,
their magnesium contents . Some of the final zinc, and copper losses to the tailings. Other
zinc concentrates, such as those from the metallic losses to the tailings, however, are in
Magmont mine and mill, have been reground and the form of free galena, sphalerite, and
subjected to additional flotation in order to especially cha1copyrite. 33 Research by the U.S.
avoid costly acid treatment. Bureau of Mines to recover the metallic losses
Chalcopyrite was the earliest of the to the Viburnum tailings by ref1otation using a
economically valuable minerals to be deposited. variety of flotation reagents has produced con-
Chalcopyrite commonly was followed directly by centrates with acceptable lead contents but with
siegenite, the principal Co-Ni mineral. . . f ilcan t cont ent s 0f 'lmpurlty
51gnl . e1ements. 35
Siegenite formed crusts upon the surfaces of the Zinc-rich ores at some of the Viburnum mines
earlier chalcopyrite and usually partly replaced have recently been found to float poorly with
the underlying chalcopyrite gra ins or conventional treatment . It is suspected that
crysta1s. 30 Those partial replacements lead to the presence of abundant organic matter or clays
the formation of cha1copyrite-siegenite locked with the zinc ores result in excess ive
particles that cause cobalt and nickel to be consumption of flotation reagents and result in
concentrated in the copper concentrate. In very significant zinc losses to the tailings .
addition, very fine-grained inclusions of Mineralogical studies in progress are designed
siegenite may occur within chalcopyrite in some to i nvest i gat e the mineral phdses responsible
copper concentrates. 34 , 35 The fact, however, for that mill problem.
that the flotation properties of siegenite also
are similar to those of chalcopyrite causes part COBALT-NICKEL MINERALOGY
of the free siegenite particles to also The mineralogy of the cobalt and nickel phases
partition into the copper concentrates. in the Viburnum Trend is of special interest
In some ore bodies in the Viburnum Trend, because the Trend is the second largest reserve
such as Magmont East, siegenite tends to replace of cobalt in the United States, behind that of
spha1erite . 33 Where this is the case, partial the Blackbird district in Idaho. Also, cobalt
replacements lead to the formation of spha1erite- is a strategic metal important to the welfare of
siegenite locked particles during grinding and the country. Although the cobalt and nickel
flotation . Those locked particles constitute an occurs in the Viburnum Trend ores that are
important beneficiation problem because their currently being mi ned, the met al s are not being
presence raises the cobalt and nickel contents recovered. The mineralogy of cobalt dnd nickel
in the zinc concentrates above those allowable in the oistrict has been investiqated by several
at the zinc refinery. Cobalt and nickel are individuals: Hagni 23 , 25, Pigno1et-Brandom 37
deleterious constituents in zinc concentrates at Horra11 38 ,and Jessey. 39
the refineries, and it is necessary to hold The most important cobalt-nickel mineral is
cobalt below 0.13 %, and to restrict cobalt plus siegenite. Siegenite forms octahedral crystals
nickel below 0.25 %. At times, some of the zinc that typically form coatings on chalcopyrite
ores have not been mined or the zinc grains and crystals. The thickest coating of

siegenite occur at the Madison mine in the brown, less reflective zones carrying the
Fredericktown subdistrict, south of the Old largest nickel content, up to 12.0 %. Although
Leadbe1t, where the ores averaged 0.23 % bravoite is an iron-nickel sulphide, it may
coba1t. 40 The metals were commercially contain significant amounts of cobalt replacing
recovered from that subdistrict during World the nickel. Bravoite crystals from the lead
Wars I and II by producing an iron sulfide district have been found to contain as much as
flotation concentrate that contained much of the 9.2 %coba1t. 36,41 Bravoite was an abundant
cobalt and nickel. The concentrate was then constituent in the cobalt-nickel concentrates
subjected to a roast and the cobalt and nickel from the Fredericktown ores.
were recovered by hydrometallurgy. The Madison A nickel-rich mineral association is locally
mine was dewatered in the late 1970's, shortly present in some of the mines; it is known
after a disruption of cobalt supplies from Zaire especially at the Buick mine, but it is present
and Zambia had caused the price of cobalt to locally at most of the mines. The association
escalate to about $74 per pound, but operations consists of prismatic to acicular crystals of
subsequently ceased when the price returned to millerite that were deposited after octahedral
about $8 per pound. Although some cobalt and . 25 32
galena but before CUblC galena.' Where the
nickel production was made from concentrates crystals were enclosed by later cubic galena
formerly developed and stored on the property, deposition, they are unaltered. Most millerite
no new mine production was achieved before the (NiS) crystals were not coated by galena and
price declined. they exhibit all degrees of alteration to
Siegenite also occurs in smaller amounts as po1ydymite (Ni3S4). The volume change involved
coatings of chalcopyrite in all of the Viburnum in this pseudomorphic alteration results in the
mines. Larger amounts of cobalt and nickel and development of prominent shrinkage cracks
higher cobalt-nickel ratios occur at the Magmont aligned largely along the basal cleavage. The
and Buick mines in the central portion of the subsequent deposition of vaesite (NiS2) occurred
Viburnum Trend and appear to indicate that a on the surface of the polydymite crystals and
significant portion of the ore fluids entered within the shrinkage cracks within the
the Trend in that vicinity.23,39 Because larger polydymite. Nickel for the deposition of
amounts of cobalt and nickel and higher vaesite was supplied by the excess nickel from
cobalt:nicke1 ratios also occur at the the pseudomorphic alteration of millerite to
Sweetwater mine at the southern end of the polydymite and by the introduction of additional
district, another important source of fluid nickel from the ore fluids. 23,25,32
probably entered the Trend from the south. Chalcopyrite-bornite pOds and lenses that
Recent research on the recovery of cobalt and locally are present in most of the mines in the
nickel from the ores in the Viburnum Trend has Viburnum Trend also contain cobalt-nickel
centered on efforts to recover those metals from 32
minera1s. The pods are horizontal lenses that
. 34
the copper concentrate at t he Magmont mlne. are from about four inches to about four feet
Due to the fine intergrowths between thick and may be 15 feet or more in lateral
chalcopyrite and siegenite, it has been found extent. They are composed of massive sulfides
that an extremely fine grind to a particle size that have totally replaced the host rock
of 10)Urn is necessary to achieve separation of dolomite at an early time in the paragenetic
siegenite from chalcopyrite. sequence and commonly are located low in the
Bravoite is the second most important Viburnum Trend orebodies. The mineralogy and
cobalt-nickel mineral in the ores from the paragenetic sequence for these ores has been
Fredericktown and Viburnum Trend subdistricts. discussed by Hagni. The chalcopyrite and
Bravoite occurs as disseminated pyritohedral bornite exhibit a wide variety of textural
crystals that are closely associated with intergrowths that include exso1ution,
siegenite and other ore minerals. The bravoite replacement, and co110form relationships.
crystals typically are chemically zoned with the Gersdorffite (NiAsS) crystals were deposited

early and were successively altered to the The nicke1-arsenic-su1phide-rich ores were
copper-arsenic minerals, tennantite and deposited dominantly as millerite that formed
enargite, with the introduction of the during four generations of crystalline and
copper-rich ore fluids that deposited massive co110form nickel sulfide as shown in Figure 2.
bornite and cha1copyrit~. Compositionally The first generation of millerite (generation I)
zoned, coba1tian pyrite cubes were subsequently was deposited as coarse prismatic crystals.
deposited in these ores . 35 F1etcherite (Cu(Ni, Millerite I was followed shortly by two
CO)2S4) is locally present in the cha1copyrite- generations of col10form millerite (generations
bornite pods and it was also replaced by II and III) deposition . Millerite II was
tennantite and enargite. Vaesite crystals deposited in three pulses (Millerite IIA, lIB,
formed euhedral crystal replacements of and IIC) . Millerite lIB constitutes the main
chalcopyrite and bornite, commonly in the portion of millerite II deposition, and IIA and
vicinity of enargite crystals. Nicke1ean IIC were composed of thin layers of acicular
carro11ite occurs in some chalcopyrite-bornite millerite that were deposited before and after
pods. 36,38 Late vein1ets composed of millerite lIB and were especially suscept ible to
chalcocite, digenite, ani1ite, djur1eite, replacement by subsequently deposited ore
cove11ite, and b1aub1eibender cove11ite traverse minerals. Millerite II consisted of two pulses
the massive chalcopyrite-bornite pods. of deposition, an earlier main pulse, and a
later thin layer of acicular millerite crystals
MAGMONT-WEST ORES that commonly has been partially to completely
A drift, more than 2 km long and initiated in replaced by later ore minerals. The fourth
1979, provided access to the new Magmont-West generation of millerite (millerite IV) crystals
orebody. Since 1985, that deposit has provided consist of thin prismat ic crystals.
more than half of the production from the All four generations of millerite have been
Magmont mine. 1U The Magmont-West ore body intensely altered to and replaced by especially
differs in many respects from typical Viburnum the three minerals, polydymite, gersdorffite,
Trend deposits. It lies about two miles west of and vaesite, but also by the subsequently
the main ore bodies in the Trend. The position deposited sphalerite, chalcopyrite, and galena.
of the ore is controlled by a narrow linear The millerite crystals of generations I and IV
brecc ia that is associated with fracturing. In have been partially to completely altered to
contrast to the main ore bodies in the Viburnum polydymite. Shrinkage cracks developed by this
Trend that trend northly, the Magmont-West ore alteration are partially filled with vaesite.
body trends north-northwest and locally to the The mineralogy and textures of th is mineral
east-northeast. Its stratigraphic position i s assemblage are similar to that noted earlier for
rather high, and its greatest concentration of ores elsewhere in the distr ict, such as at the
mineralization is centered upon the False Davis Buick mine. The colloform millerite of
horizon within the Bonneterre Formation. The subsequent generations, however, has not been
ore body is unusually zinc-rich and averages found el sewhere in the district. This
about 3%zinc, approximately three times that of fine-grained millerite is especially susceptible
the average deposit in the Viburnum Trend. A to replacement and only small remnants of the
portion of those ores, especially along the millerite usually remain to attest to its
easterly side of the ore trends, contain original mineralogical character. Gersdorffite
nickel -arsenic-rich ores that are unlike is the most common mineral that replaces the
nickel-rich ores elsewhere in the district. millerite, but millerite IIA, IIC, and IIIB may
Because the mineralogy of the Magmont-West be almost completely replaced by vaesite that
nickel-arsenic-rich ores has not previously been forms prismatic pseudomorphs after the original
described, some emphasis is given here to the mi 11 erite.
mineralogy and paragenetic sequence of those After the deposition of millerite III, a
ores. coating of euhedra1 pyritohedra1 vaesite

Figure 2.

Paragenetlc Sequence for Collomorphlc Nickel-Arsenic Sulfide Ores,

Magmont West Orebody

~ ~
2 3 4



Gersdorfflte ~

Carrollite ~





crystals were deposited upon the surface of They are more intensely altered to polydymite
colloform millerite III, and these crystals were and vaesite . Small octahedral crystals of
followed by a coating of cubic and pyritohedral carrollite commonly were deposited on the
crystals of gersdorffite. Locally, those surfaces of the millerite IV prisms. Another
gersdorffite crystals form thin veinlets that generation of vaesite crystals, in turn, were
traverse the earlier generations of colloform deposited on the carrollite.
millerite II and III. The small crystals of The copper, lead, and zinc sulfides were
vaesite and gersdorffite were both deposited deposited after all of the above minerals in the
before the deposition of millerite IV. Magmont-West nickel-arsenic sulphide ores.
Millerite crystals of generation IV are not Sphalerite was the first to be deposited and it
as coarse as those belonging to generation I. filled small vugs between the millerite IV

crystals. It also locally completely replaced that has a mineralogy that is comparable to that
layers of millerite IIA, lIB, and IIIB. of the Olympic Dam deposit in Austra1ia . 42-44
Chalcopyrite also was deposited in the spaces The very close proximity of these two deposits
between millerite IV crystals, selectively raises the question of their genetic
replaced layers of millerite IIA, lIB, and IIIB, relationship . The Boss-Bixby is a major iron
and locally veined and replaced sphalerite. The deposit that has been thoroughly drilled but has
initial deposition of copper was in the form of not been opened to mining. If this deposit is
1uzonite that occurs at the margins of mined, copper and cobalt will be recovered as
chalcopyrite areas. Pyrite and marcasite by-products. One interpretation is that the
commonly are associated with the deposition of copper and cobalt sulphides in the Boss-Bixby
chalcopyrite. Galena formed in the spaces deposit were deposited after the deposition of
between millerite IV crystals and locally the Precambrian iron ore and that they were
replaced both earlier deposited sphalerite and deposited at the same time and from the same ore
chalcopyrite. Locally, galena exhibits fluids that deposited the Viburnum ores . This
prismatic forms where it pseudomorphica11y interpretation appears not to be the case
replaced vaesite that is itself pseudomorphic because Precambrian granite dike1ets appear to
after prismatic crystals of millerite IIA. 45
traverse the copper-cobalt sulphide ores.
Sphalerite, chalcopyrite, and galena Another interpretation for the unusually
preferentially replaced carro11ite crystals that nickel-rich character of the Magmont-West
were deposited before vaesite. That replacement deposit is that the copper, cobalt, and nickel
resulted in octahedral pseudomorphs of may have been leached out from the underlying
sphalerite, chalcopyrite, and galena that are Precambrian Boss-Bixby ore aeposit, by the
contained within vaesite. saline connate brines that deposited the
The paragenetic sequence for those ores was Viburnum Trend ores, and redeposited in the
completed with the deposition of quartz and overlying Bonneterre Formation. The fact that
significant leaching of the earlier deposited the Boss-Bixby ores are cobalt-rich but are
ore minerals. Because the major portion of nicke1- and arsenic-poor in contrast to the
these ores are unlike those found elsewhere, it Magmont-West ores that are nicke1- and
is difficult to fit their deposition into the arsenic-rich and cobalt poor, however, does not
paragenetic sequence previously constructed for appear to support this interpretation. It
Vlburnum ores found throughout the Trend. 22' 32 appears from the information known to this
Although the millerite crystals of generation IV point, that the genetic connection between the
are similar in many respects to those found at two deposits is a more remote one, if any at
the Buick mine and elsewhere in the district alL
where they were deposited late in the mineral References
sequence, other features suggest that the 1. Esparza L.E. United States Bureau of Nines,
nickel-arsenic sulfide ores in the Magmont-West Mineral Industry Surveys, Washington, D.C.,
orebody have been deposited at an earlier time. 1989, 3 pp.
The fact that all of the nickel-arsenic sulphide 2. Kisvarsanyi K. The Role of the Precambrian
minerals were deposited before sphalerite, Igneous Basement in the Formation of the
chalcopyrite, and galena supports this Stratabound Lead-Zinc-Copper Deposits in
interpretation. The fact that the most abundant Southeast Missouri. Economic Geology, vol. 72,
nickel-arsenic sulphide mineral is Qersdorffite. 1977. D. 435-442.
a mineral found to be deposited early in the 3. Wharton H.M. Letter to Paul A. Gerdemann on
paragenetic sequence elsewhere in the Trend also lead-zinc production totals for Southeast
supports this idea. 32 Missouri Lead District, Missouri Department of
The Magmont-West ore deposit occurs within Natural Resources, Division of Geology and Land
about 0.3 km of the Boss-Bixby deposit, a. Survey, 1981, 2 pp.
deposit in the underlying Precambrian rocks, 4. Hagni R.D . The Southeast Missouri Lead

District: A review. In: Mississippi Valley-Type mine, Viburnum Trend, Southeast Missouri, In :
Mineralization of the Viburnum Trend, Missouri, Applied Mineralogy: Proceedings of the Second
Society of Economic Geologists, Guidebook International Congress on Applied Mineralogy in
Series , vol . 5, R.D. Hagni, and R.M . Coveney the Minerals Industry, Park, W. C., Hausen, D.
Jr., eds., Fort Collins, Colorado, 1989, p. H. and Hagni R. D., eds., AIME, New York, N.Y.,
12-57. 1984, p. 965-986 .
5. Oh1e, E.L. and Gerdemann, P.E. Recent 13. Evans L.L. Geology of the Brushy Creek
exploration history in Southeast Missouri, In: mine, Viburnum Trend, Southeast Missouri.
Mississippi Valley-Type Mineralization of the Economic Geology, vol. 72, 1977, p. 381-390.
Viburnum Trend, Missouri, Society of Economic 14. Uingess P.R. Geology of the ASARCO West
Geologists, Guidebook Series, vol. 5, R.D . Fork mine, in Mississippi Valley-type
Hagni, and R.M . Coveney, Jr, eds, Fort Collins, mineralization of the Viburnum Trend, Missouri,
Colorado, 1989, p, 1-11. In: Mississippi Valley-Type Mineralization of
6. Grundmann, W.H. Jr. Geology of the Viburnum the Viburnum Trend, Missouri, Society of
No. 27 mine, Viburnum Trend, Southeast Missouri. Economic Geologists, Guidebook Series, vol. 5,
Economic Geology, vol . 72, 1977, p. 349-364. R.D. Hagni, and R.M. Coveney, Jr, eds, Fort
7. Pettus, J.R. Jr. and Dunn R.G. Jr. Geology Collins, Colorado, 1989, p. 96-110.
of the St. Joe Minerals Corporation Number 28 15. Paarlberg N. L. and Evans L. L. Geology of
mine, In: Sediment-Hosted Pb-In-Ba Deposits of the Fletcher mine, Viburnum Trend, Southeast
the Midcontinent, Hagni, R. D. et a1., eds. Missouri. Economic Geology, vol . 72, 1977, p.
Society of Economic Geologists Field Guidebook, 391-397.
1988, p, 63-76. 16. Mouat M. M. and Clendenin, C.W. Geology of
8. Dunn, R.G . Jr. and Grundmann, W.H. Jr . the Ozark Lead Company mine, Viburnum Trend,
Geology of the Casteel mine : Copper-rich ore in Southeast Missouri. Economic Geology, vol. 72,
a MVT setting, in Mississippi Valley-Type 1977, p. 398-407.
Mineralization of the Viburnum Trend, Missouri, 17. Larsen, K.G. Stratigraphic and facies
In : Mississippi Valley-Type Mineralization of nomenclature of the Viburnum Trend, Southeast
the Viburnum Trend, Missouri, Society of Missouri. In: N. Paar1berg, ed., The 26th Annual
Economic Geologists, Guidebook Series, vol. 5, Field Trip Guidebook, Association of Missouri
R.D. Hagni, and R.M . Coveney, Jr, eds, Fort Geologists, 1979, p. 15-19.
Collins, Colorado, 1989, p. 58-83. 18. Walker W.B . Geology of the Sweetwater mine,
9. Sweeney, P.H., Harrison, E.D. ana Bradley, Viburnum Trend, Southeast Missouri, In:
M.F. Geology of the Magmont Mine, Viburnum Mississippi Valley-Type Mineralization of the
Trend, Southeast Missouri. Economic Geology, Viburnum Trend, Missouri , Society of Economic
vol. 72, 1977, p. 365-371. Geologists, Guidebook Series, vol. 5, R.D.
10. . Bradley, M. F. and Krolak, T.E., 1989, The Hagni, and R.M . Coveney, Jr, eds, Fort Collins,
Magmont-West 1ead-zinc-copper mine (Cominco Colorado, 1989, p, 111-131.
American Incorporated - Dresser Minerals), 19. Larsen K.G . (1977) Sedimentology of the
Viburnum Trend, Southeast Missouri, In: Bonneterre Formation, Southeast Missouri.
Mississippi Valley-Type Mineralization of the Economic Geology, vol. 72, 1977, p. 408-419.
Viburnum Trend, Missouri, Society of Economic 20. Gregg J.M., and Gerdemann P.E., Sedimentary
Geologists, Guidebook Series, vol . 5, R.D . facies, diagenesis, and ore distribution in the
Hagni, and R.M. Coveney, Jr, eds, Fort Collins, Bonneterre Formation (Cambrian), Southeast
Colorado, 1989, p. 84-95. Missouri, In: Mississippi Valley-Type
11 . Rogers R.K. and Davis J.H. Geology of the Mineralization of the Viburnum Trend, Missouri,
Buick mine, Viburnum Trend, Southeast Missouri. Society of Economic Geologists, Guidebook
Economic Geology, vol. 72, 1977, p. 372-380. Series, vol . 5, R.D. Hagni, and R.M . Coveney,
12 . Muge1 U.N. and Hagni, R.D. Mineralogy and Jr, eds, Fort Collins, ColoradO, 1989, p.
geology of the blanket lead-zinc deposit, Buick 141-154.

21. Voss R.L. and Hagni R.D. The application of Shanks W.C. III Geologic and geochemical
cathodoluminescence microscopy to the study of controls of mineralization in the Southeast
sparry dolomite from the Viburnum Trend, Missouri Barite District. Economic Geology vol.
Southeast Missouri, Chapter 5 in D.M. Hausen and 82, 1987, p. 719-734.
O.C. Kopp, eds., Mineralogy--Applications to the 29. Clendenin C.W., 1977, Suggestions for
Minerals Industry, Proceedings of the Paul F. interpreting Vibu rnum Trend mineralization based
Kerr Memorial Symposium, AIME, New York, NY, on field studies at Ozark Lead Company,
1985, p, 51-68. Southeast Missouri . Economic Geology, vol. 77,
22. Voss R.L. Hagni, R.D. and Gregg J.M., 1977, p. 465-473.
Sequential deposition of zoned dolomite and its 30. Hagni, R.D. and Trancynger T.C. Sequence of
relationship to sulfide mineral paragenetic deposition of the ore minerals at the Magmont
sequence in the Viburnum Trend, Southeast mine. Economic Geology, vol . 72, 1977, p.
Missouri, Special issue on the "Relationships 451-464.
between dolomitization, basin development, and 31. Watson J.L. South East Missouri Lead Belt -
Mississippi Valley-type ore deposits, Carbonates A review. Minerals Engineering, vol. 1, no. 2,
and Evaporites, vol 4, no. 2, 1989, p. 195-209. 1988, p. 151-156.
23. Hagn i, R.D . Minor Elements in Mississippi 32. Hagni, R.D. Ore microscopy and paragenetic
Valley-type ore deposits: Chapter 7, In: Cameron sequence of the ores of the Southeast Missouri
Volume on Unconventional Mineral Deposits, Lead District, International Mineralogical
Shanks W.C. III, ed, Society of Economic Association, 14th General Meeting, Abstracts
Geologists - Society of Mining Engineers, AIME, with Program, Stanford University, Stanford,
New York, 1983, p. 71-88. California, 1986, p. 118.
24. Roedder, E. Fluid inclusion studies of ore 33. Hagni, R. D., 1983, Applications of the ore
deposits in the Viburnum Trend, Southeast microscope to ore dressing problems in the
Missouri . Economic Geology, vol. 72, 1977, p. Viburnum Trend, Southeast Missouri, U.S.A., ICAM
474-479. 81: Proceedings of the First International
25. Hagni, R.D. Ore microscopy, paragenetic Congress on Applied Mineralogy, de Villiers
sequence, trace element content, and fluid J.P.R. and Cawthorn, P.A., eds., Special
inclusion studies of the copper-lead-zinc Publication No.7, The Geological Society of
deposits of the Southeast Missouri Lead South Africa, Marshalltown, South Africa, 1983,
District, In : International Conference on p, 209-215.
Mississippi Valley Type Lead-Zinc Deposits, 34. Cornell W.L. and O'Connor, W.K., 1987, Ore
Proceedings Volume, Kisvarsanyi G., Grant S.K., microscopy in support of appl ied benefi ciation
Pratt W.P ., Koenig J. W., eds., University of research to recover cobalt from Mis sour i l ead
Missouri-Rolla, Rolla, Missouri, 1983, p. ores. In : Process Mineralogy VII: Applications
243-256. to Mineral Beneficiation Technology and Mi neral
26. Mavrogenes, J.A. Hagni, R.D., Dingess, Exploration, with Special Emphasi s on
P.R., Metal sulfide mineral zoning at ASARCO's Disseminated Carbonaceous Gold Ores, Vassiliou
West Fort mine, Viburnum Trend, Southeast A.H ., Hausen D.M., and Carson D.J.T., eds.,
Missouri (abs.), Geological Society of America, American Institute of Mining, Metallurgical and
North-Central Section annual meeting, vol. 22, Petroleum Engineers, Warrendale, PA, p. 603-615.
no. 5, Notre Dame University, South Bend, 35. Co rnell W.L., Wethi ngton A.M., Holtgre fe
Indiana, Apr. 1989, p. 14. D.C. and Sharp F.H., Continuou~ flotation
27. Wharton, H.M., Total mine production and testi ng to recover cobalt fro m Missouri l ead
value from the Viburnum Trend mines from the ores, U.S. Bureau of Mines Report of
State of Missouri in 1964, Missouri Department Investigations 9072, 1987, 11 p.
of Natural Resources, Division of Geology and 36. Frank Sharp, 1980, oral communication.
Land Survey, 1986, 1 p. 37. Pignolet-Brandom, S., ~li lieralogy,
28. Kaiser C.J., Kelly W.C., Wagner R.J., and paragenesis and el ect r on mi cr oprobe analysis of

the cobalt-nickel minerals in the Mississippi 45. Brandom R.T. and Hagni R.D., Ore Microscopy
Valley-type deposits of Southeast Missouri, of the Boss-Bixby, Missouri, Cobalt-Copper-Iron
Unpub. Ph.D. dissertation, University of Ores. Society of Mining Engineers of AIME
Missouri-Rolla, Rolla, Missouri, 1988,221 pp . Transactions, vol. 280, 1987, p. 2139-2142.
38. Horrall, K.B . Mineralogical , textural, and
paragenetic studies of selected are deposits of
the Southeast Missouri Lead-Zinc-Copper District
and their genetic i mpl i cat i ons , Unpub. Ph.D.
dissertation, University of Missouri-Rolla,
1982, 650 pp.
39. Jessey, D.R. Mineralogical and
Compositional Variations of the Nickel-Cobalt
Mineralization in the Southeast Missouri Mining
District . In: Process Mineralogy: Extractive
Metallurgy, Mineral Exploration, Energy
Resources, Hausen O.M. and Park W.C., eds, AIME,
New York, 1981, p. 159-177.
40. Brooke G.L., Mad ison cobalt mine, Madison
County, Missouri, American Institute of Mining,
Metallurgical and Petroleum Engineers, Preprint
82-324, Warrendale, PA, 1981, 3 p.
41. Moh, G.H . and Ottemann J. X-ray
fluorescence and electron-probe analysis of some
pyrite-type minerals, Year Book, 1963-64,
Carnegie Inst. of Washington, 1964, p. 214-216 .
42. Brandom R.T . and Hagni R.D ., A Comparison
of the mineralogy of the Boss-Bixby, Missouri
Copper-Iron deposit, and the Olympic Dam (Roxby
Downs) copper-uranium-gold deposit, South
Australia, Geological Society of America Annual
Meeting, Orlando, Florida, 1985, p. 530-531.
43. Hagni R.D. and Brandom R.T., Mineralogical
similarity of the Boss-Bixby, Missouri
copper-iran-cobalt deposit to the Olympic Dam,
South Australia are deposit : potential for
additional deposits in the Midcontinental United
States (abs.), 28th International Geological
Congress Program, Washington, D.C ., vol. 2,
1989, p, 7-8.
44. Hagni, R.D . and Brandom R.T., The
Mineralogy of the Boss -Bixby, Missouri
Copper-Iran-Cobalt Deposit and a Comparison to
the Olympic Dam Deposit at Roxby Downs, South
Australia. In: Olympic Dam-Type Deposits and
Geology of Middle Proterozoic Rocks in the St.
Francois Mountains Terrane, Missouri, Society of
Economic Geologists, Guidebook Series, vol . 4,
V.M. Brown, E.B. Kisvarsanyi, R.D. Hagni, eds,
Fort Collins, Colorado, 1989, p. 82-92.

Ore processing and mineralogy
Principles and practice of sulphide mineral
R. Herrera-Urbina
J. S.Hanson
G . H. Harris
D. W. Fuerstenau
Department of Materials Science and Mineral Engineering, University of California
at Berkeley, Berkeley, California, U.S.A.

SYNOPSIS compiled recently by the United States Bureau of Mines,

This paper presents both the fundamentals of sulphide the largest tonnage of ores treated by flotation
mineral flotation and the practical aspects of sulphide corresponds to sulphide ores, which also consume the
ore flotation. The fundamental aspects of the flotation largest amount of reagents.2 Today, froth flotation of
of sulphide minerals both in the absence and presence of sulphide ores constitutes the single most important
thiol collectors are discussed in relation to their crystal operation for the concentration of sulphide minerals.
structure, surface chemistry and electrochemical Current flotation practice often involves the separation
characteristics. Since the electrochemistry of the of three or four sulphides from the same ore.
sulphide mineral/aqueous solution interface and of Industrially important sulphide minerals include
various sulphide mineral/aqueous thio-compound galena, chalcopyrite, sphalerite, pyrite, chalcocite and
systems has been amply investigated over the past two molybdenite. Traditionally, these minerals have been
decades, the results of these modem electrochemical separated from each other using relatively simple
studies are briefly reviewed. Mechanisms of the action flotation schemes to process high-grade, coarse-gra ined
of various flotation reagents, including collectors, ores. Because of the depletion of these easy-to-process
activators and depressants, commonly used in sulphide ores, however, flotation scientists and engineers are
mineral flotation are discussed in terms of the chemical now confronted with the need to exploit existing low-
and electrochemical principles presented within this grade sulphide deposits with complex mineralogical
paper. Simplified flow sheets of representative flotation composition and finely disseminated values. As a result
plant operations are given to illustrate how these of this need, in recent years the flotation of complex
principles are applied for the commercial separation of sulphide ores has been the topic of a number of
sulphide minerals from their ores. Various novel symposia and workshops.3-5
techniques that take advantage of controlling the In contrast to the extensive amount of progress
electrochemical and physical behaviour of the flotation that has been gained ind ustrially, a clear understanding
system are also presented. of the basic principles of the flotation of sulphide
minerals has not yet been achieved though most of the
early fundamental flotation research involved
The practice of sulphide minerals flotation is as old as investigation of the flotation behaviour of this group of
the flotation process itself. In the United States, for minerals. The way by which those sulphides that do not
example, the first froth flotation plant began operation exhibit natural flotability are made to respond to
in 1911 at the Butte and Superior Copper mill in Basin, flotation has been perhaps the aspect investigated more
Montana and recovered sphalerite from slimes intensively and in the greatest detail. This aspect
(material 92% finer than 150 meshl.I Although in 1911 involves the interaction of the mineral with
sphalerite was the only sulphide mineral concentrated amphiphatic surfactants, which are called collectors in
commercially by flotation, over the years the industrial flotation term inology, that adsorb at the solid/solution
flotation of sulphides has grown impressively in size, interface and make the attachment of an air bubble to
complexity and importance. According to statistics the mineral surface thermodynamically favourable.

Since the discovery in 1925 by Keller6 that the alkali sulphide minerals flotation and the practical aspects of
xanthates act as collectors of sulphide minerals, these sulphide ores flotation . Since the crystal structure and
thio-eompounds have become the universal collectors surface chemistry of sulphides determine their response
for sulphide minerals other than molybdenite. In spite to flotation, the nature of the sulphide mineral surface
of over half a century of continued research on the will be examined first to delineate the effect of crystal
interaction of this thiol collector and sulphide minerals, structure on the inherent hydrophobicity of some of
this system is nowyielding both new results and these minerals. For sulphide minerals that do not
problems as new experimental techniques are exhibit natural flotability, approaches to render their
introduced in modem flotation research and ultimately surfaces hydrophobic and the phenomena involved are
into flotation plant practice. In addition to xanthates, discussed. How the surfaces of sulphides interact with
such thio-eompounds as dithiophosphates (patented by thiol collectors, depressants and activators used
Whitworth in 1926) and thiocarbamates (patented by commercially for the concentration of sulphide
Harris and Fischback in 1954) are also common sulphide minerals by flotation are analysed. Over the past two
minerals collectors. Although the interaction of thiol decades, the electrochemical aspects of these interactions
collectors and sulphide minerals is not completely have received considerable attention. Results of these
understood, both its chemical and electrochemical studies will be reviewed briefly and discussed to the
nature has been amply documented. Galena, bornite extent that various working hypotheses can be applied
and chalcocite, for example, interact chemically with to plant practice. To illustrate how these fundamental
thiol collectors producing a surface metal thiolate.7 On aspects of flotation chemistry are applied for the
the basis of modem electrochemical studies, on the separation of sulphide minerals from their ores,
other hand, it has been possible to propose a mixed- simplified flowsheets of representative commercial
potential model to rigorously explain the electro- operations will be given. In addition, various novel
chemical nature of this interaction. This model states techniques that take advantage of controlling the
that the hydrophilic/hydrophobic transformation of the electrochemical and physical aspects of the system are
sulphide mineral surface by the interaction with a thio- also presented.
compound involves simultaneous electrochemical
reactions. The cathodic step is usually the reduction of CRYSTAL STRUCTURE AND SURFACE CHEMISTRY
oxygen, while the anodic step involves the oxidation of OF SULPHIDES
the collector.8 Surface chemistry is the very basis of flotation.
Apart from the issue of how thiol collectors Therefore, a complete understanding of the physical-
interact with sulphide minerals making them amenable chemical characteristics of solid surfaces is essential to
to flotation, various aspects of these flotation systems better understand the complex phenomena taking place
remain unclear. For example, the role of oxygen has during flotation . These surface characteristics are
become the most controversial issue mainly because of determined by such crystal properties as the types of
lack of adequate oxygen control in many investigations. forces that hold the solid together, that is the chemical
Based on the most recent results on electrochemical bonding, and its atomic or molecular structure. Surfaces
research, however, it is now widely recognized that an created by the rupture of bonds other that van der Waals
electron transfer is associated with the adsorption of a bonds contain very reactive, ionic sites, while those
thiol collector to make most sulphide surfaces surfaces created by the rupture of van der Waals bonds
hydrophobic. Industrially, oxygen acts as the electron are capable of reacting with the environment only
acceptor but can be substituted by other oxidizing agents through van der Waals type interactions. Furthermore,
such as sodium hypochlorite,9 or by anodic oxidation . the interaction between the surfaces of solids and their
After many years of being denied as a real phenomenon, environment depends on the solid's cleavage
the flotation response of sulphide minerals in the characteristics, which are determined by the type of
absence of collectors and under certain conditions is bonds that are most susceptible to breakage during size
now well documented. lO Mildly oxidizing reduction. Using such techniques as X-ray diffraction
environments render some sulphide minerals and 8 -a solution diffractometry, Lindqvist and Stridh ll
amenable to flotation without collectors. have demonstrated that the flotation behaviour of
This paper presents both the fundamentals of galena, pyrite and sphalerite is directly related to a

Table I. Classification of sulphide minerals based on the environments, the surfaces of sulphide minerals
relative abundance of surface sulphide sites. undergo an oxidation process and, since the solubility of
these oxidation products is higher than that of their
Sulphur-rich surfaces Sulphur-deficient surfaces sulphide counterparts, the solubility of the mineral
Pyrrhotite Wurtzite increases. As a result, oxidation phenomena alter
Marcasite Sphalerite (Blende) significantly the surface and aqueous chemistries of
Arsenopyrite Niccolite
sulphide minerals and therefore directly affect their
Pyrite Bismu thinite
Cinnabar Stibnite flotation behaviour.
Molybdenite Tetrahedrite
Chalcocite Surface oxidation
Bornite Oxidizing environments, chemically or
Galena electrochemically induced, are known to greately
Realgar influence the flotation response of sulphide minerals,
both in the absence and the presence of collectors,
specific crystallographic cleavage plane . For galena and through oxidation of their surfaces. Surface oxidation
pyrite, this plane corresponds to the (tOO) plane, while phenomena are closely related to the interfacial
for sphalerite it corresponds to the (t 10) plane. electrochemical potential, which determines both the
The atomic or molecular arrangement of the solid, type and the rate at which a reaction may take place.
on the other hand, determines the position of the During the last decade, modem sulphide flotation
different chemical species and their distance from the research has been increasingly involved with
surface. In the case of sulphides, their surfaces can be delineating the effect of potential on the extent of
chemically classified as sulphur-deficient and sulphur- surface oxidation under different solution conditions.
rich depending on the relative abundance of sulphur Such techniques as linear potential sweep voltammetry,
atoms. Table I lists as class A those sulphide minerals intermittent galvanostatic polarization, and
whose surfaces contain exposed sulphur atoms, and as A-C impedence have been invaluable in achieving our
class B those sulphides with surfaces characterized by the present understanding of surface oxidation phenomena
absence of sulphur atoms, which are located inside the and their relation to flotability in sulphide mineral
elementary cell.12 Those sulphide minerals listed in flotation systems. Woods and Richardsonf have
class A have been found to respond well to flotation obtained linear potential sweep voltammograms for
with thiol collectors, while those in class B respond fresh surfaces of various sulphide minerals, and these
poorly to flotation or do not float at all, such as are reproduced in Figure 1, which shows the potential
sphalerite. According to this classification, it seems regions in which surface oxidation occurs.
reasonable to assume that in order for this zinc sulphide Characterization of oxidized sulphide surfaces has
to respond to flotation with short-ehain xanthates, the been undertaken to identify the chemical nature of the
superficial layer of zinc atoms must be removed to oxidation products. Electrochemical techniques,
expose the hidden internal sulphur atoms.1 2 however, do not allow identification of the chemical
Another aspect related to both the crystal structure composition and structure of sulphide mineral surfaces
and surface chemistry of sulphides that has been found subjected to oxidizing conditions. For this purpose,
to influence their flotation behaviour is the solubility of flotation scientists have used such spectroscopy
the mineral. Theoretically, the bonding between metal techniques as X-ray photoelectron spectroscopy (XPS).
atoms and sulphur is mostly covalent, and is Among these researchers, Buckley and Woods 13 applied
responsible for the low solubility of sulphide minerals XPS to identify the chemical species formed when the
in water. Under common working conditions, surfaces of galena, pyrrhotite, pyrite and chalcopyrite are
however, it is well known that most sulphides are not oxidized. Their results indicate that the initial surface
as chemically inert as predicted by classical product in these systems is a metal-deficient (or
thermodynamics, which does not take into sulphur-excess) sulphide. Further oxidation can result
consideration surface phenomena related to the in the formation of sulphide species of progressively
instability of sulphide surfaces in the presence of lower metal content and eventually elemental
oxygen-containing fluids such as water or air. In these sulphur.H

CuaS i
(0) (b)

Fe,_"S I

.' ,/
Fe ~
, ._..... /'
-,-"'.; '1- -... , ..._.., .'

o 00 0

-0.4 -0.2 0 0.2 0.4 0.6 (e) (d)

Potential (V vs SHE) Fig. 2: The crystal structure of Molybdenite.

(a) Environment of the molybdenum atom - the
Fig. 1. Linear potential sweep voltammograms for fresh trigonal prismatic coordination polyhedra.
surfaces of sulfide minerals : positive-going scans at (b) The 5-Mo-S sandwich layer.
20 mV/s in solutions of pH - - 4.6; - - - 6.8; (c) The hexagonal (2Hl) crystal molybdenite.
- - - 9.2; 11; -.-.-. 13. (After Ref. 8) (d) The environment of the sulfur atom.

Guy and Trahar15 have summarized the different NATURAL FLOTABILITY OF SULPHIDES
oxidation mechanisms that may take place when a Many sulphides were once thought to exhibit natural or
sulphide mineral is exposed to aqueous solutions under inherent flotability . It is now well known, however,
acidic or alkaline conditions. In both cases, elemental that this surface property arises from the inherent
sulphur is suggested to form on the surface; under acidic hydrophobicity displayed only by those sulphide
conditions metal ions pass into solution, while under minerals with special crystallographic structures, of
alkaline conditions a metal hydroxide is formed. It is which molybdenite is a typical example. Gaudin et a1.,16
also suggested that further oxidation may lead to the had long recognized that the natural flotability of a solid
formation of oxy-sulphur species according to the arises when a new surface is created without breakage of
following reaction: interatomic bonds other than residual van der Waals
x MzS + (zxn + y)H20 = zx M(OH)n + Sx~ bonds . Two sulphide minerals exhibit natural
(2y + zxn)H+ + (2y + zxn-Zle" flotability: molybdenite, MoS2, and stibnite, Sb2S3. As
The subsequent flotation behaviour of an oxidized can be seen in Figure 2, molybdenite exhibits a layered
sulphide mineral will depend upon the nature of the crystal structure consisting of 5-Mo-S sheets held
surface products formed . If sulphur or a sulphur-like together by van der Waals bonds. Rupture of these
species is formed at the surface, the mineral will bonds results in a hydrophobic surface, while rupture of
respond to flotation in the absence of collectors. In the the mostly covalent Mo-S bond produces hydrophilic
case of strong oxidizing conditions, the hydrophilic oxy- sites. This breakage pattern produces predominantly
sulphur species present at the interface will suppress hydrophobic faces and to a lesser extent edges which are
flotability. hydrophilic. As a consequence, it has been reported that

the flotation response of molybdenite is determined by collectors, a behaviour closely resembled by galena and
the surface made up of faces.1 7 In the case of stibnite, its pyrrhotite, while pyrite and sphalerite float poorly
crystalline structue consists of chains where the Sb and S without collectors. The trend on the collectorless
atoms are held together by weak bonds of unknown flotation of these common sulphide minerals and the
character.1 7 Another characteristic of naturally flotable greatest response of the sulphide of copper and iron has
sulphides is that, on oxidation, they generally form been confirmed by Leroux et a1. 20 The flotation of
anionic species that remain as individual sites. In oxidized chalcopyrite without collectors has been
contrast, non-naturally flotable sulphides form cations attributed to the formation of a hydrophobic surface
that may hydrolyze to form insoluble oxides (or species, which has not been identified with certainty,10
hydroxides) on the particle surface . These oxides can but is thought to be a sulphur-enriched surface
grow as patches even on the hydrophobic part of the species,14 or a metal polysulphide.21
surface rendering it hydrophilic. In addition, those Collectorless flotation has been observed to be a
sulphides that exhibit natural flotability float under very real property of several common sulphide
reducing conditions, while other sulphide minerals do minerals and to occur in typical sulphide flotation
not. By exploiting this phenomenon, it is possible to environments, that is, in air saturated solutions ranging
separate chalcopyrite from molybdenite. from moderately acidic (pH 3) to moderately alkaline
(pH 11) conditions.1 0 When a frother is added to the
COLLECfORLESS FLOTATION OF SULPHIDE flotation pulp, collectorless flotation is more
MINERALS pronounced. Recently, Hayes et al.,22 have reviewed
The response of sulphide minerals to flotation without several aspects of classical surface oxidation studies
conventional collectors was reported by Gaudin as early through to modem electrochemical, surface chemical
as 1932. 6 At that time, however, lack of a satisfactory and spectroscopic studies relevant to the collectorless
explanation led to the suggestion that contamination of flotation of sulphide minerals.
the mineral surface was responsible for this unusual In addition to electrochemical techniques, X-ray
flotation behaviour of sulphide minerals, and the photoelectron spectroscopy (XPS) has been used to
subject was not investigated further. It was not until the determine the nature of the sulphur enrichment at
mid 1970s that the flotation of sulphide minerals sulphide mineral surfaces and its role in flotation, a
without collectors was again reported.9,18,19 Flotation subject recently reviewed by Buckley and Walker,23 who
of sulphide minerals other than molybdenite without concluded that the collectorless flotation of sulphide
the addition of a conventional thiol collector is now minerals is controlled by the concentration of surface
termed "collectorless flotation," and is used to describe oxides as well as the degree of sulphur enrichment at
flotation systems where a mineral acquires some degree the surface, regardless of the nature of the surface
of hydrophobicity in the absence of a collector due to sulphur. Metallic compounds precipitated on the
chemical changes at the surface20 These chemical surface of the mineral have been found to be especially
changes, which are closely related to the interfacial effective for the supression of collectorless flotation. As
electrochemical potential of the system, strongly affect can be seen in Figure 3, which shows the collectorless
the nature of the sulphide mineral surface and therefore flotation of chalcopyrite in the absence and presence of
control its collectorless flotation . For example, the more copper salts reported by Guy and Trahar,15 the recovery
cathodic or anodic the rest potential and the less easily of chalcopyrite is significantly reduced as the
oxidized the sulphide surface, the more readily it floats concentration of added copperfll) species increases. The
without collector. This appears to be due to the depressant action of this compound is closely related to
formation of sulphur and sulphur-rich mineral patches the precipitation of copper hydroxide at the mineral
at mildly oxidizing potentials. surface.
Chalcopyrite9 and galena 15 have been shown to The industrial impetus towards collectorless
float in the absence of collectors under mildly oxidizing flotation is the major reduction in reagent
conditions. Chalcocite, on the other hand, does not float requirements, particularly reagents for pH control.
significantly under similar conditions. More recently
Shannon and Trahar10 have also shown that
chalcopyrite exhibits flotability in the absence of

100 ~~""""--'----"'-"""'-"""",-"""'---r--~-.., Table II. Thio-compounds commonly used as collectors
in sulphide mineral flotation.

eo Collector Structural Formula

(Trade Name)

Alkyl disulfide I
Q.alkyl dithiocarbonalCS or
unthales (C)'IIIIIIIidAET 300,
DOW Z-3 ID Z-12 Series)

~S S~
Diallcyl disulfide or dilwuhogens R-O-C~ '\:C-O-R

2 4 8 8
Xanthic CSlCrS R-O-C~
T1WE (min) 'S-R'
Fig. 3. Influence of added copper on the collectorless
flotation of chalcopyrite. Xanthogt71 formate R-O-C~S ,&0
(Minem;) -, C~
S- 'O-R'
/p~ e

Sulphide minerals other than molybdenite are

Diallcyl dithiocarblunalt:
inherently hydrophilic under common flotation
conditions. To effect the hydrophilic/hydrophobic
transformation of the surfaces of these minerals, and Isopropyl thionocarbamate
hence make them amenable to flotation, a collector (DowZ200)

must be added to the system. Since 1925, when alkali

xanthates were first used as collectors, these compounds Men:aplDbenmthiazole
(Cyanamid AERO400 Series)
have become the standard collectors in the flotation of
sulphide minerals. 6 Dithiophosphates, whose use as
collectors was patented a year later,6 have remained the
next most commonly used collectors in the flotation of 1,3-diphenyI2thiourea
or Thioc:artJujJidc
sulphide minerals. Over the years numerous chemical (C)'IIIIIIIid AERO 130 Series)
reagents have also been reported to have collecting
properties. Of these chemical compounds, those that
possess a polar group containing at least one sulphur
atom not bonded to oxygen are the most common Diphenyl Guanidine
sulphide mineral collectors, and they belong to the type
of thio-compounds, These thiol collectors are listed in
Table II, which also gives their structural formula and
trade name. account for oxygen's role in the adsorption process fall
The role played by oxygen in the adsorption of thiol basically into two categories. The first and oldest
collectors onto sulphide minerals has been perhaps the mechanism proposed is the metathetical, ion-exchange
most controversial issue in sulphide mineral flotation type reaction. Although this hypothesis of oxygen's role
investigations. Dewey24 recognized in 1933 that oxygen in flotation has been used for many years, it is often
must be present in order for xanthate to adsorb onto a found to be inconsistent with experimental observation.
sulfid e mineral surface. The theories proposed to The past thirty years have seen the development of

another approach to explain the adsorption of flotation 1100 1000 900 101 700
collectors on sulfide minerals, namely the use of
electrochemical principles to account for the role of (0)
oxygen. The electrochemical mechanism is still the 50

focus of numerous investigations and undergoes

continual refinement. z 0
0 50 lb)
Early physical-ehemical theories of sulphide flotation a:
(/) 0
The interaction of sulphide minerals and thiol collectors III
4( 50
was explained at the beginning of flotation research in ~
terms of an ordinary chemical reaction between these o
a: 0
chemical species,25 and by means of an adsorption w
D- 50
mechanism that led to the formation of non-Daltonian
compounds.26,27 Taggart's chemical theory25 proposes
that because of surface oxidation the sulfide mineral 50 (e)
surface is actually composed of metal oxides, carbonates
and sulfates, which upon addition of a thiol collector O......_ _--' L.-_ _L.-_ _L.-_ _l..-_---'

such as xanthate react in a metathetical, ion-exchange Fig. 4. Infrared spectra of: a) bulk lead ethyl xanthate
reaction. Metathesis between these surface oxidized solid as mujol mull, b) freshly evaporated PbS film after
compounds, MA(s), and xanthate ions in solution, X- atmospheric oxidation, c) freshly evaporated PbS film
treated with aqueous ethyl xanthate solution, d) freshly
(aq), results in the formation of a metal xanthate, MX(s), evaporated PbSfilm after prolonged washing in ether,
on the sulphide mineral and the transfer of the oxidized and e) freshly evaporated PbS film after washing in
pyridine. (After Ref. 28).
surface anions, A-, into solution:
MA(s) + X(aq) = MX(s) + A(aq) energetically equivalent to the metathetical, ion-
The chemistry of thiol collector/sulphide mineral exchange type reaction proposed by Taggart for the
interaction has been investigated more recently using formation of lead ethyl xanthate on galena immersed in
such techniques as infrared spectroscopy28 and xanthate solutions.
calorimetry.29 In the case of the galena/ethyl xanthate
system, the infrared spectra obtained by Leja et a1. 28 and Electrochemical aspects of sulphide flotation
reproduced as Figure 4 reveal that indeed a chemical The difficulty in accurately evaluating the amount of
compound, namely lead ethyl xanthate, forms at the dissolved oxygen in solution and assessing the degree of
surface of lead sulphide immersed in solutions of ethyl oxidation of the sulphide mineral surface make
xanthate. Spectrum (a) in this figure is for bulk lead electrochemical techniques ideal for investigating the
ethyl xanthate, while spectrum (b) is for a lead sulphide adsorption of thiol collectors on mineral surfaces.
film oxidized under atmospheric conditions. After Further complicating the adsorption process is the ease
exposing this film to a xanthate solution, spectrum (c) in which the sulfhydryl collectors oxidize. This
was obtained. As can be seen, this spectrum is virtually phenomenon often obscures the effect of oxygen in the
identical to spectrum (a) thus confirming that a new flotation system because the adsorption mechanism may
surface chemical compound is formed, most probably be altered if the mineral surface is not well characterized
lead ethyl xanthate. Finally, after this sample was or the collector itself has undergone some degradation.
washed with a strong solvent, namely pyridine, In most electrochemical studies the role of oxygen as
spectrum (e) was obtained. This spectrum shows that electron acceptor is replaced by an external electric
this washing procedure completely removes the lead circuit.
ethyl xanthate leaving behind only an oxidized lead Electrochemical techniques were used in sulphide
sulphide surface. The heats of reaction measured by flotation research as early as 1931.30 It was not until
Mellgren,29 on the other hand , clearly indicate that 1954, however, that the sulfide mineral/thiol collector
the uptake of ethyl xanthate by oxidized galena, which interaction was explained in electrochemical terms,
had been treated with 0.1 M H2S04 and washed, is and a sulfide flotation model , known today as the

mixed-potential model, was proposed.31 This model, in Table III. Rest potential and reaction product of
its most general form, considers that sulphide minerals, minerals in xanthate solutions (After Ref. 7).
being good electronic conductors, interact with thiol
Mineral Rest Potential, V Reaction Product
collectors through a mechanism that involves
Sphalerite -0.15 NPI
simultaneous electrochemical reactions at the mineral
Stibnite -0.13 NPI
surface. The important role played by electrochemical
Galena +0.06 MX
reactions during the interaction of thiol collectors with
Bornite +0.06 MX
sulphide minerals has been demonstrated by the results
Chalcocite +0.06 NPI (MX)
of numerous electrochemical investigations reported in
Chalcopyrite +0.14 X2
the last three decades. These recent results, which have
Molybdenite +0.16 X2
been thoroughly reviewed and summarized by
Pyrrhotite +0.21 X2
Woods 32,33 and Woods and Richardson,8 have helped
Pyrite +0.22 X2
in identifying the simultaneous electrochemical
Potential for X2+ 2e- -+ 2X-at 0.625 roM KEX = 0.13 volt
reactions that occur in these systems. The cathodic step
usually involves the reduction of oxygen,
potentials are above this value oxidize the thiol to its
0:2 + 2H20 + 4e- = 4OH-
disulphide, and this product is now accepted as the
and it is coupled with an anodic step involving
collecting species in these systems. Similar results have
oxidation of either the collector or the mineral. The
been reported for sulphide/diethyl dithiocarbamate
products of the anodic reaction depend on the mineral
systems. 34
and collector used, and the pretreatment of the mineral
A necessary condition for sulphide mineral
surface. They have been identified as chemisorbed
flotability is the thermodynamic stability of the
collector, dithiolates and metal/collector compounds.
hydrophobic entity formed at the surface. If this
The reactions for the anodic oxidation of the collector
compound forms at a rate too slow compared to the rate
can be written as follows:8
at which the sulphide mineral oxidizes, however, the
(a) Charge transfer chemisorption of a thiol ion (X-):
mineral will not float because surface oxidation may
X- = X(ads) + e-
result in the formation of oxide-type materials that may
(b) Oxidation of a thiol ion to its disulfide:
not only impede electron transfer from the collector to
2X- = X2 + 2e- the mineral but is also extremely hydrophilic. In
(c) Formation of a thiol compound with a metal
addition, if the hydrophobic species is decomposed by
component of the mineral:
oxidation, strong oxidizing conditions would be
MS + 2X- = MX2 -+ So + 2e-
detrimental to flotation. Knowledge of the oxidation
The mixed-potential model does not actually
potential (Eh) of the system is therefore vital in sulphide
contradict earlier theories of sulphide flotation. Wark's
mineral flotation. It is possible to predict the oxidizing
and Gaudin's adsorption model would be equivalent to
and reducing conditions necessary for a substance to be
reaction (a) while Taggart's chemical reaction theory
thermodynamically stable by constructing the
would be represented by the anodic process given by
appropriate Eh-pH diagram. These diagrams, first
reaction (c), which can be a two-step reaction involving
constructed for sulphides/thiol collector systems by
oxidation of the mineral and metathesis.
Toperi and Tolun 35, have been used for comparing
The product resulting from the interaction of
experimental results with thermodynamic
sulphide minerals and thiol collectors has been
equilibria.36,37 Figure 5 presents the Eh-pH diagram for
identified and correlated with the rest potential of the
the chalcocite/diethyl dithiophosphate/water system.
mineral. Allison et all identified the surface products
This diagram shows the stability doma ins of the
in sulphide/ethyl xanthate systems, and a selection of
different mineral and collector species. It can be noted
their results is reproduced in Table Ill. These results
that copper diethyl dithiophosphate, CuDTP, the species
clearly indicate that, with the exception of covellite,
responsible for chalcocite flotation, is unstable in
sulphide minerals that display a rest potential below the
alkaline solutions regardless of the oxidation potential.
corresponding reversible potential for the oxidation of
In acidic, neutral and mildly alkaline solutions, as the
the collector to its disulphide will react with thiols
potential becomes increasingly positive, chalcocite first
forming metal thiolates. Those sulphides whose rest


1.6r---.....--r---r--oy---r---r--...., o
I C ChoIeoc:it.
- I 80 6 Pyrite O'
HS04 1 _..0
I ~u

0.8 I
CuO ~...
I 20
:~"':"'_0= cz:t:-:::..-::
. . . -. 6

Eh,V - 0 .5 - 0 .3 -0.\ 0.1 0. 3

o POTENTlAl. V (SeE)

Fig. 6. Flotation recovery of chalcocite and pyrite at

pH 9.2 as a function of potential for single and mixed
-0.4 mineral beds. (After Ref. 42).

H25 1 dependent, flotation ceases. Identification of the

-0.8 .1 HS- products of the sulphide mineral!ethyl xanthate
KOTP COnc. 1 interaction was achieved using such techniques as linear
5 I sweep voltammetry and ultraviolet
-1.2 - - 10- ~ 1 [5] _10. 8
- - - 1O- !! :[CU++] _ 10- 8 spectrophotometry.41 The species responsible for
- - - - 10- !! I
imparting surface hydrophobicity appear to be metal
-f.6 .......- .........- ........----'----"--''---........- ...........- .....
o 2 4 6 8 10 12 14 xanthates or surface analogues of metal xanthates
formed by chemisorption on chalcocite and bornite,
Fig. 5. Eh-pH diagram for the chalcocite/diethyl dixanthogen on pyrite, and both metal xanthate and
dithiophosphate system. dixanthogen on chalcopyrite.
Gebhardt, et al.,42 working with mixed beds of equal
oxidizes to CuDTP, which then oxidizes to Cu 2+ (or weights of chalcocite and pyrite, have achieved flotation
CuO). Strongly oxidizing conditions leading to the separations of these sulphide minerals using xanthate as
formation of Cu 2+ or CuO have a deleterious effect on the collector by controlling the electrochemical potential
the flotation of chalcocite with diethyl dithiophosphate of the flotation cell. Their results are reproduced in
as the collector. Figure 6, which presents the flotation recovery of both
It is now widely accepted that the flotation of Single-component mineral beds and the components of
sulphide minerals with thiol collectors is controlled by mineral mixtures for three different methods of
the interfacial electrochemical potential. Using in-situ addition of the mineral mixture to the cell. This figure
electrodes in microflotation cells, Gardner and Woods 38 clearly shows that in addition to the general xanthate
and Chander and Fuerstenau39 were the first to flotation behaviour of sulphides, which are depressed at
independently and simultaneously demonstrate that the reducing potentials and floated at oxidizing potentials,
electrochemical potential can control the flotation of the flotation of chalcocite and pyrite from the mixture
sulphide minerals. More recently, Walker et al.,40 does not follow that of the single mineral. The presence
developed an electrochemical micro flotation cell that of chalcocite enhances pyrite flotation in the potential
uses a packed bed of conducting sulphide particles as the range of "{).4 to -0.1 V and promotes pyrite depression
working electrode to correlate interfacial electrochemical above 0.1 V. Pyrite flotation below the reversible
reactions with flotation response. Richardson and xanthate/dixanthogen potential, namely "{).13 v,7 was
Walker4 1 used this cell to investigate the xanthate attributed to activation by Cum) ions produced by the
flotation of chalcocite, bornite, chalcopyrite and pyrite as anodic dissolution of chalcocite.
a function of the electrochemical potential. The Through control of the electrochemical potential at
flotation of these sulphide minerals was found to be the solid/liquid interface, Chander and Fuerstenau43
strongly dependent on the conditioning potential and to have also been able to separate sulphides, namely
occur only under moderately oxidizing conditions. At a chalcocite from molybdenite, by flotation using diethyl
certain reducing potential, whose value is mineral dithiophosphate as the collector. Choosing the

appropriate flotation potential, chalcocite can be shown to activate sphalerite for xanthate flotation.
depressed and molybdentie floated, or vice versa . In this Through contact angle measurements and flotation
investigation, molybenite was found to be relatively experiments, Sutherland and Wark47 examined the
electrochemically inactive, and therefore its flotation- activation behavior of sphalerite with various metal
response was not suppressed by applying low potentials. salts. Their results indicate that successful activators are
Chalcocite flotation is affected by the rate of oxidation, salts of metals which themselves or their sulphides
which was controlled by changing the sweep rate to alter respond readily to flotation with thiol collectors . Salts of
the value of the conditioning potential to that of lead, cadmium, silver and mercury fall within this
flotation. category. Since the flotation behaviour of sphalerite
activated with metal salts is similar to that of the
ACTN AnON AND DEPRESSION IN SULPHIDE activator metal sulphide mineral, there appears to be a
MINERAL FLOTATION direct correlation between flotation response and the
With the exception of sphalerite and pyrrhotite, solubility product of the metal xanthate. The chemistry
sulphide minerals of the base metals respond well to of sphalerite activation has been thoroughly reviewed
flotation with short-chained xanthate collectors. Upon by Finkelstein and Allison. 48
the addition of copper salts, however, sphalerite Cufll) ions have also been found to activate other
responds readily to flotation with xanthates. This sulphide minerals for xanthate flotation. Using
means of enhancing the flotability of sphalerite has electrochemical techniques and a rotating disc
become the classic example of activation in flotation electrodes, Nicol 49 has recently investigated the
technology, and since 1913, when Leslie Bradford activation mechanism of galena and pyrrhotite with
discovered in Australia that copper sulphate enhances copper salts; the results show that CuS forms on the
the flotability of zinc sulphide,6 this inorganic salt has surface of these activated minerals, which in tum
proved to be indispensable for effectively concentrating exhibit flotation behaviour typical of copper sulphides.
zinc sulphide ores by flotation. The general consensus The activation of sulphide minerals with heavy metals
on the activation of sphalerite by Cu(II) ions has been has been recently reviewed by Wang et a1.,SO who point
that this process involves an ion exchange mechanism out that the ion-exchange mechanism postulated to
in which cupric ions are exchanged for zinc ions, explain this process is oversimplified because it takes in
according to the following reaction: to consideration neither surface oxidation phenomena
nor metal ion hydrolysis. An electrochemical approach
ZnS(s) + Cu 2+ = CuS<S) + Zn 2+
was taken by these researchers to propose that the
The stoichiometric 1:1 Cu/Zn exchange and the overall activation process can be explained in terms of a
coating of sphalerite by covellite were observed by mixed-potential mechanism.
Cooke44 in a long-term (50 days), high-temperature A number of different approaches have been found
(1000 C) study of the activation of sphalerite with copper to be effective for sulfide depression 33 which can be
chloride. The results reported by Gaudin et a1.,45 on the accomplished by preventing the interaction of thiol
uptake of Cu(ll) ions by sphalerite also indicate that an collectors with the sulfide mineral resulting in the
ion exchange reaction occurs in this process, and that formation of a hydrophobic surface:
there is a 1:1 Cu/Zn exchange. Using 64Cu as a 1. enhancement of sulphide surface oxidation,
radioactive tracer, and starting with 3 micromoles Cu(II), 2. inhibition of oxygen reduction,
these researchers reported that 2.8 micro moles Cu(II) 3. decrease in oxygen concentration,
were abstracted and 2.9 micromoles Zn(II) were 4. inhibition of collector oxidation,
displaced in this reaction. This chemical exchange 5. complexation of surface sites.
reaction was also reported by Mellgren et a1.,46 to be In practice, flotation engineers achieve selectivity by
responsible for the activation of sphalerite by copper changing the pH of the pulp and by adding to the system
ions, and through calorimetry they measured the heat of such depressant-acting reagents as sodium salts of
adsorption of Cum) on sphalerite. The enthalpy change cyanide, sulfide, sulfite, and potassium dichromate. The
for the exchange reaction was found to be 62.8 kJlmole addition of these chemicals alters the electrochemical
Cu at 250 C. potential of the system thus affecting mineral flotation.
Salts of metal ions other than copper have also been The mechanisms by which these depressants inhibit

Table IV. Mechanism of Depressant Action on Sulfide Mineral Flotation:


Hydroxyl Due to the formation Xanthate oxidation to Cuprous ethyl xanthate is so stable that
of lead hydroxy species dixanthogen does not hydroxide ions can function as a depressant
at about pH 11 occur above pH 11. only at very high pH values for both chalcocite
Also Fe(OHh is the and chalcopyrite.
stable species above
pl-l-l l .

Cyanide No effect because there Formation and In absence of Fe only Similar to pyrite
is no lead cyanide chemisorption of high concentrations
complex . Fe(CN)6-4at the of CN will depress,
surface due to the formation
of Cu(CN)2-

Sulfide PbS is more stable than Reduces the oxidation Due to the forma- Similar to pyrite
PbX2so collection re- potential so that the tion of CU2S at the
action is reversed. formation of X2is surface.
inhibited. Also there
can be chemisorption
of sulfide ion.

Chromate Due to the formation

of lead chromate at the

Sulfite Reduces oxidation

potential below that is
required to form

flotability are summarized in Table IV. In terms of the chalcopyrite, pyrite and marcasite.52 The gangue in this
oxidation potential of the system, the collector species ore is dolomite. In this operation, the first flotation step
responsible for flotation, and the species leading to aims at recovering all of the lead and copper minerals
mineral depression, Chander51 has classified sulphide while depressing iron and zinc sulphides, and dolomite.
minerals as reversible and passivated. This This is achieved by flotation at pH 8.5 to 9.0 using amyl
classification is presented as Figure 7, which shows the xanthate as the collector, lime for pH control, and very
Eh-pH conditions for the stability domain of sulphides, small quantities of cyanide (0.02-0.03 kg/tonne) for
metal!collector compounds for reversible sulphides and pyrite, sphalerite and marcasite depression. In addition,
disulphides for passivated sulphides. This figure also zinc sulphate (1.6 to 20 kg/tonne) is used to help
gives the depressant species for both groups of minerals. enhance sphalerite depression. Separation of galena
The group of reversible sulphides include galena and from chalcopyrite is possible using liquid sulphur
chalcocite, while the group of passivated sulphides may dioxide and starch as galena depressants while floating
include pyrite, chalcopyrite and bornite. The reactions chalcopyrite. At the Magmont Mill, this method was
of passivated sulphides are irreversible. Using this type chosen because the copper content of the bulk lead-
of information, the appropriate conditions for copper concentrate is usually low (2 to 7% Cu).52 The
selectively separating sulphide minerals by flotation tailings from the lead-copper flotation circuit are treated
with thiol collectors can be predicted. with copper sulphate to activate sphalerite for its
subsequent recovery. Lime is used to maintain a pH of
INDUSTRIAL PRACTICE IN THE FLOTATION OF 11.0 to 11.5 in the cleaners and 9.0 to 9.5 in the roughers
SULPHIDE ORES during sphalerite flotation with xanthate. Cyanide is fed
Flotation plant practice at the Magmont Mill, to the cleaners to depress the iron and unrecovered
Southeast Missouri, will be described to illustrate a copper minerals. The flowsheet of this operation is
typical separation of lead, zinc and copper sulphides shown in Figure 8.
from a sulphide deposit containing galena, sphalerite,

Research on the electrochemistry of sulphide
ReVERSIBLE SULFIDes mineral flotation with thiol collectors has been
flotation Species conducted at a rapid pace during the last thirty years on a
1. 0 , - - - - , - - - - - - ,
metal-collector compound(s) variety of minerals by numerous investigators.
MX, MX2,etc.
However, the only industrial use of electrochemical
Depressant Soecjes control was recently described by Heimala et a1.,54 who
(a) mineral oxidation product report that Outokumpu Oy has developed an industrial
layer MO, M(OH)2, etc.
(b) Hydrophilic MD layer cell incorporating electrochemical control to effect a
(c) Removal of surface MX,
better separation and concentrate grade in the flotation
.I.OO'---~---.l etc. by oxidation, reduction
7 14 or hydrolysis . of a copper/nickel ore. Their results show that by
pH simply monitoring and controlling the electrochemical
potential of the flotation cells, both the amount of
PASSIVATED SULFIDES collector and the amount of reagents used for pH
adjustment can be reduced significantly, while
flotatign Species
1.0, - - - - r - - - - - ,
oxidation product of
improving the grade and recovery of the concentrate.
collector, X2 Controlling the electrochemical potential of the
flotation system may also be accomplished on an
> Depressant Species
(a) mineral oxidation product industrial scale by controlling the oxygen concentration
layer, Mo. M(OH)n,etc. in the slurry. One technique to help keep the
(b) hydrophilic layer of a
metal-depressant electrochemical potential down is to use nitrogen as the
- LO ' -_ _~_ _- ' compound, MD.
o 7 14 flotation gas. This approach is being used industrially in
(c) removal 01 surface X2 by
pH reduction copper-molybdenum separations, using hydrosulphide
in closed flotation cells. Closed flotation cells are
Fig. 7. Conditions for flotation and depression of
reversible and passivated sulphide. (After Ref. 51). extremely effective in reducing reagent cost by excluding
oxygen. Dissolved oxygen not only.raises the potential
An extensive and detailed study of the Climax but when hydrosulphide is present it increases the
molybdenite ore (0.3-0.4% MoS2) flotation has been consumption of this reagent. At Anamax, the use of a
reported by Hoover and Malhotra .53 The Climax Mill closed flotation system reduced the molybdenite
flowsheet is presented in Figure 9. It consists of a single- depressant by 68%, with no nitrogen make up being
stage grind to 35-40% plus 100 mesh for rougher necessary.55
flotation followed by several stages of fine grinding
restricted to 3-6% of the total feed. Satisfactory recovery
at an extremely coarse grind is reportedly possible by the

effective flotation of a large amount of low grade quartz-

molybdenite middlings due to the high flotability of
molybdenite. Overall recovery at the Climax Mill
~ Flec:IMner Zn CardIonIng

historically has been 89-90%. Molybdenite ore flotation
at Climax uses a battery of reagents carefully designed to
Pb + Cu CondlIclnIng
ensure maximum recovery and grade. Molybdenite,
similar in flotation behaviour to unoxidized, higher- ~
Cu Rougher- - - . . ,
rank coal, is hydrophobic and is floated with oily
collectors like vapor oil (light kerosene). An
emulsifying agent, called Syntex, is used to enhance
Cu L FlMI Tellngi

molybdenite flotation by more effectively distributing

J Pbc:o.- Znc:o.-
the collector. In addition to vapor oil and Syntex,
molybdenite ore flotation requires other reagents to
Fig. 8. Flowsheet of the copper-lead-zinc sulphide
depress impurities: NaCN to depress pyrite, Nokes
separation by flotation at the Magmont Mill.
reagent to control galena and copper flotation, and
sodium silicate to depress the silicate gangue.

Mill Feed Reagent Kgltonne
ROUGHER PLANT 1. Fuerstenau D.W. In: Froth Flotation . D.W.
~ .._-------
Ball Mllls ....
Pine Oil 0.07
Fuerstenau, Ed. New York: American Institute of
Vapor Oil 1.00
Syntex 0.06 Mining, Metallurgical and Petroleum Engineers, Inc.,
Sodium Silcate 1.20
Rougher --...., Lime 0.80 1962,677pp.
Flotation Nokes 0.18
2. Martin T.W. and others. In: Minerals Yearbook 1985.
Concentrate SCavenger Reagent Kgltonne Washington: Bureau of Mines, vol. 1, 1987. p. 7-65.
Flotation 3. Flotation of Sulphide Minerals. KS.E. Forssberg, Ed.
Vapor Oil 0.10
Syntex 0.01 New York: Elsevier, 1985, 480pp.
4. Complex Sulphide Ores. M.J. Jones, Ed. London : The
Institution of Mining and Metallurgy, 1980, 278pp.
Vapor Oil 1.80
5. Complex Sulphides. A.D. Zunkel and others, Eds.
Sod ium Cyanide 0.50 Warrendale: The Metallurgical Society, Inc., 1985, 938pp.
tst Stage-SCavenger Nokes 1.80
Grind Float Flotation Dowfrolh 250 0.06 6. Gaudin A.M. Flotation. 2nd ed. New York: McGraw-
Nalco 1801 0.012
Hill, 1957,573pp.
1 TJng
All reagent additions are based
7. Allison S. A. and others. A determination of the
2nd Stage-SCavenger on cleaner plant feed tonnage. products of reaction between various sulphide minerals
G,ind Float Flotation
and aqueous xanthate solutions, and a correlation of the
3rd Stage ~I'
Tal 109
Vapor oil, Dowfroth 250 and
sodium cyanide are added to products with electrode rest potentials. Metallurgical
Grind Float all stages of flotation according
Transactions, vol. 3, 1972, p. 2613-2618.
to stage tonnages .

Conditioners Nokes Reagent is added to

8. Woods R. and Richardson P.E. The flotation of

4th Stage Float
all stages of grinding and
cond itioning according to
sulphide minerals-electrochemical aspects. In: Advances
in Mineral Processing. P. Somasundaran, Ed. Littleton:
stage tonnages.

Concentrate Nalco 1801 is added to all
Society of Mining Engineers, 1986. p. 154-170.
stages of thickening. 9. Heyes G.W. and Trahar W.J. The natural flotability of

Fig. 9. Flowsheet of the copper-molybdenum separation chalcopyrite. International Journal of Mineral

by flotation at the Climax Mill. Processing, vol. 4, 1977, p. 317-344.
10. Shannon L.K and Trahar W.J. The role of collector
SUMMARY in sulphide ore flotation. In: Advances in Mineral
The results of numerous investigations on both pure Processing. P. Somasundaran, Ed. Littleton: Society of
and impure systems have helped to further our Mining Engineers, 1986. p. 408-425.
understanding of the complex flotation behaviour of 11. Lindqvist O. and Stridh K Structural studies on
sulphide minerals with thiol collectors. It is now well adsorption of collectors on mineral surfaces. In: XVI
accepted that the flotation response of sulphides is International Mineral Processing Congress. E. Forssberg,
determined by three characteristics: 1) their low Ed. New York: Elsevier, 1988, p. 717-725.
solubility, 2) their surface oxidation behaviour and 3) 12. Scordamaglia R. and others. Correlation of collecting
their semiconducting properties. Modern power of function alized organic molecules and their
electrochemical studies have made possible the molecular structure by quantitative chemical and
establishment of a mixed-potential model to rigorously statistical methods of calculation. In: Reagents in the
explain the hydrophilic/hydrophobic transformation of Minerals Industry. M.J. Jones and R. Oblatt, Eds. London:
a sulphide mineral surface upon its interaction with a Institution of Mining and Metallurgy, 1984. p.189-192.
thiol collector. This model states that this surface 13. Buckley A.N. and Woods R. In: International
transformation involves simultaneous electrochemical Symposium on Electrochemistry in Mineral and Metal
reactions; the cathodic step usually being the reduction Processing. P.E. Richardson, S. Srinivasan, and R.
of oxygen, while the anodic step involves the oxidation Woods, Eds. Electrochemical Society, 1984. p. 286-302.
of the collector or the formation of a metal thiolate.
This information helps mineral processing engineers
develop effective schemes to selectively recover
sulphide minerals from complex ores.

14. Buckley AN. and others. Investigation of the surface 25. Taggart A.F. and others. Chemical reactions in
oxidation of sulphide minerals by Linear Potential flotation. Technical Publication of the American
Sweep Voltammetry and X-Ray Photoelectron Institute of Mining and Metallurgical Engineers, no . 312,
Spectroscopy. In: Flotation of Sulphide Minerals. KS,.E. 1930,p. 3-33.
Forssberg, Ed. New York: Elsevier, 1985. p. 41-59. 26. Gaudin A.M. Flotation mechanism, a discussion of
15. Guy P.J. and Trahar W.J. The effects of oxidation and the functions of flotation reagents. Technical Publication
mineral interaction on sulphide flotation. In: Flotation of the American Institute of Mining and Metallurgical
of Sulphide Minerals. KS.E . Forssberg, Ed. New York: Engineers, no. 4, 1927, 27pp .
Elsevier, 1985. p. 91-109. 27. Wark I.W. The chemical basis of flotation.
16. Gaudin A.M. and others. In: 2nd International Proceedings. Australasian Institute of Mining and
Congress of Surface Activity. London: Butterworths, Metallurgy, no. 90, 1933, p. 83-123.
1957. p. 202-210. 28. Leja J. and others. Transactions. Institution of
17. Chander S. and others. On native floatability and the Mining and Metallurgy, vol. 72, 1962-63, p. 414-423.
surface properties of naturally hydrophobic solids. In: 29. Mellgren O. Heat of adsorption and surface
Advances in Interfacial Phenomena of reactions of potassium ethyl xanthate on galena.
Particulate/Solution/Cas Systems. Applications to Transactions. American Institute of Mining and
Flotation Research. P. Somasundaran and RG. Grieves, Metallallrgical Engineers, vol. 235, 1966, p.46-60.
Eds. New York: American Institute of Chemical 30. Karnienski B. So-called flotation . Przem. Chern., vol.
Engineers, vol. 71, no. 150,1975. p. 183-188. 15, 1931, p.201-2. (Chemical Abstracts, vol. 26, 1932, p. 53).
18. Lepetic V.M. Flotation of chalcopyrite without 31. Salamy S.G. and Nixon J.e. Relation between a
collector after dry autogenous grinding. Canadian mercury surface and some flotation reagents: an
Institute of Metallurgy Bulletin, 1974, p. 71-77. electrochemical study. Australian Journal of Chemistry,
19. Finkelstein N.P. and others. Natural and induced vol. 7, 1954, p. 146-156.
hydrophobicity in sulphide minerals systems. In: 32. Woods R Electrochemistry of sulphide flotation. In:
Advances in Interfacial Phenomena of Flotation. A.M. Gaudin Memorial Volume. M.e.
Particulate/Solution/Gas Systems. Applications to Fuerstenau, Ed. Vol. 1,1976 . p. 298-333.
Flotation Research. P. Somasundaran and RG. Grieves, 33. Woods R Electrochemistry of sulphide flotation. In:
Eds. New York: American Institute of Chemical Principles of Mineral Flotation. The Wark Symposium.
Engineers, vol. 71, no 150, 1975. p. 165-175. M.H. Jones and J.T. Woodcock, Eds . Parkville: The
20. Leroux M. and others. Collectorless flotation in the Australasian Institute of Mining and Metallurgy,
processing of complex sulphide ores. In: Advances in symposia series no. 40, 1984. p. 91-115.
Coal and Mineral Processing Using Flotation. S. 34. Finkelstein N.P. and Goold L.A. The reaction of
Chander and RR Klimpel, Eds. Littleton: Society for sulphide minerals with thiol compounds. Nat. Institute
Mining, Metallurgy and Exploration, Inc., 1989: p. 65-70. of Metallurgy. South Africa, report no. 1439, 1972.
21. Luttrell G.H. and Yoon RH. The collectorless 35. Toperi D. and Tolun R Electrochemical study and
flotation of chalcopyrite ores using sodium sulphide. thermodynamic equilibria of the galena-oxygen-
International Journal of Mineral Processing, vol. 13, xanthate flotation system. Transactions. The Institution
1984,p. 271-283. of Mining and Metallurgy, vol. 78, 1969, p. C191-C197.
22. Hayes RA. and others. Collectorless flotation of 36. Chander S. and Fuerstenau D.W. Electrochemical
sulphide minerals. Mineral Processing and Extractive reaction control of contact angles on copper and
Metallurgy Review, vol. 2, 1987, p. 203-234. synthetic chalcocite in aqueous potassium diethyl
23. Buckley AN. and Walker G.W. Sulphur dithiophosphate solutions. International Journal of
enrichment at sulphide mineral surfaces. In: XVI Mineral Processing, vol. 2, 1975, p. 333-352.
International Mineral Processing Congress. KS.E. 37. Pritzker M.D. and Yoon RH. Thermodynamic
Forssberg, Ed. New York: Elsevier, 1988. p. 589-599. calculations on sulphide flotation systems: I. Galena-
24. Dewey F.D. Reactions of some sulphur-bearing ethyl xanthate system in the absence of metastable
collecting agents with certain copper minerals. Ph. D. species. International Journal of Mineral Processing, vol.
Dissertation, Montana School of Mines, 1933. 12,1984, p. 95-125.

38. Gardner J.R and Woods R An electrochemical SO. Wang X. and others . The aqueous and surface
investigation of contact angle and of flotation in the chemistry of activation in the flotation of sulphide
presence of alkylxanthates. 1. Platinum and gold minerals-a review. Part 1: an electrochemical model.
surfaces. Australian Journal of Chemistry, vol. 27,1974, Mineral Processing and Extractive Metallurgy Review,
p. 2139-2148. vol. 4, nos. 3-4, 1989, p. 135-165.
39. Chander S. and Fuerstenau D.W. Effect of potential 51. Chander S. Oxidation/reduction effects in
on the flotation and wetting behaviour of chalcocite and depression of sulphide minerals-a review. Minerals and
copper. Transactions. Society of Mining Engineers, vol. Metallurgical Processing, vol. 2, no. 1,1985, p. 26-35.
258,1975,p. 284-285. 52. Sharp, F.H. Lead-zinc-copper separation and current
40. Walker G.W. and others. Electrochemical flotation practice at the Magmont Mill. In: Flotation. A.M.
of sulphides: reactions of chalcocite in aqueous solution. Gaudin Memorial Volume . M.e. Fuerstenau, Ed. New
International Journal of Mineral Processing, vol. 12, York: American Institute of Mining, Metallurgical, and
1984,p. 55-72. Petroleum Engineers, Inc., vol. 2, 1976. p. 1215-1231.
41. Richardson P.E. and Walker G.W. The flotation of 53. Hoover RM. and Malhotra D. Emulsion flotation of
chalcocite, bornite, chalcopyrite, and pyrite in an molybdenite. In: Flotation. A.M. Gaudin Memorial
electrochemical-flotation cell. In: XVth International Volume. M.e. Fuerstenau, Ed. New York: American
Mineral Processing Congress, tome 11,1985. p. 198-210. Institute of Mining, Metallurgical, and Petroleum
42. Gebhardt J.E. and others. Electrochemical Engineers, Inc., vol. 1, 1976. p. 485-505.
conditioning of a mineral particle bed electrode for 54. Heimala S. and others. New potential controlled
flotation . United States Bureau of Mines. Report of flotation methods developed by Outokumpu Oy. In:
Investigation 8951, 1985, 10pp. XVth International Mineral Processing Congress, tome
43. Chander S. and Fuerstenau D.W. Electrochemical III, 1985. p. 88-98.
flotation separation of chalcocite from molybdenite. 55. Onstott K.Y. and Person P.L. By-product
International Journal of Mineral Processing, vol. 10, molybdenum flotation from copper sulphide
1983,p. 89-94. concentrate with nitrogen gas in enclosed Wemco
44. Cooke S.RB. Flotation. Advances in Colloid Science, nitrogen flotation machines. Preprint No. 84-65.
vol. 3, 19SO, p. 357- 374. Littleton: Society of Mining Engineers of AIME, 1984,
45. Gaudin A.M. and others. Activation and 8pp.
deactivation studies with copper on sphalerite.
Transactions. American Institute of Mining,
Metallurgical and Petroleum Engineers, vol. 214, 1959,
46. Mellgren O. and others. Thermochemical
measurements in flotation research. In: 10th
International Mineral Processing Congress. London: The
Institution of Mining and Metallurgy, 1974. p. 451-472.
47. Sutherland K.L. and Wark 1.W. Principles of
flotation. Melbourne: Australasian Institute of Mining
and Metallurgy, 1955.489pp.
48. Finkelstein N.P. and Allison S.A. The chemistry of
activation, deactivation and depression in the flotation
of zinc sulphide: a review . In: Flotation. A.M. Gaudin
Memorial Volume. M.e. Fuerstenau, Ed. New York:
American Institute of Mining, Metallurgical and
Petroleum Engineers, Inc., vol. 1, 1976. p. 414-457.
49. Nicol M.J. In: International Symposium on
Electrochemistry in Mineral and Metal Processing, The
Electrochemical Society, 1984. p. 152-168.

Chelating reagents for flotation of sulphide minerals
A. Marabini
M. Barbaro
Istituto per il Trattamento dei Minersli, C.N.R., Rome, Italy

ii fficult to recover one or more selective

Olelating reagents , because of their metal products of acceptable quality and economic
specificity, can function as effective mineral value with minimal losses and at reasonable
processing reagents.
cost . Complex sulphide ores are fine-grained,
A brief review of the work carried out in this
intimate associations of chalcopyrite
area is presented in particular , with regard (CuFeS ,) sphalerite (ZnS) and galena (PbS),
dissemtnated in dominant pyrite, and which
- Calculation method for the selection of contain valuable amounts of silver and, in
chelating functional groups
some cases, gold.
- Criteria for designing and synthetizing
As to the second problem it is known that the
chelate-type collectors with optimal
surface that sulphide minerals present to
structure for a given metallic mineral
collectors is generally not rather well
- Comparison between conventional collectors
defined; nor does it correspond closely to the
and new synthetic aliphatic-aromatic
bulk mineralogical species. In fact, it is
known that sulphides are often strongly
The application of chelating reagents in oxidized and this is considered to influence
sulphide mineral flotation is reviewed 1ll adsorption of collectors. Despite numerous
particular, with respect to influence of studies , 1-08 the adsorption mechan1sm ' 0
molecular structure of chelating reagents on
known collectors on sUlphi1~\lurfaces is not
their collecting power (effect of alkyl yet cOffipletely clarified, and it has not
substituent and chain branching) . yet been established to what extent the
adsorption 15-~B affected by surface
The difficulties in clarifying the
The beneficiation of sulphi1e minerals is an interaction mechanism of collectors ar e
extremely difficult problem first, because a demonstrated by the number of studies carried
sulphide ore contains many valuable minerals
out on ~~:~r adsorption. Spectroscopic
with similar surface properties and secondly,
techniques 4 and, in3~~56icular, infrared
because of the high number of parameters that absorption spectrometry and attenuated
play an important role in the flotation total refJLection infrared spectrometry
process. It is interest~g to note that in (ATR-IR) 37- 4 or X-ray photoelectron
1955 Sutherland and Wark listed 32 variables 43-44
spectroscopy (XPS) have been used ,
in sulphide flotation, of which only 22 can
5g~51her with. sol~tion ~quilWia studies
usually be controlled by the metallurgist. and calor1metr1c stud1 es.
A complete evaluation of the variables in Generally, the collectors employed are
sulphide mineral flotation is still a thiols, reported in (~ig. 1), and the most
challenge. commow are xanthates. In practice , however ,
As to the first problem , this is particularly xanthates are active towards the whole class
evident for ores containing several of sulphide minerals rather than towards an
commercially useful mineral components finely individual mineral.
intergrown with the gangue and with one Thus, in order to float a given mineral from a
another, or for ores characterized by a very mixture of minerals belonging to the same
fine size analysis. This is the case of the sulphide cIas s , modifiers always have to be
so-called complex sulphide ores which have used in order to render the action of the
been defined as those ores from which it is collector more specific .

Typic.!Typeof Campaunda Uald At Sullldl Min." CallKlan
I'lrlnl FlmllyMlmbln

... Cllbonla
...I Alkyl
DllhloCirbanll. R-S-C
, S-M
OH S-M (Xanlhlill

H ... 0 R' , .1 R'... S

, N-C"
... Clrbl,"la
N-C , S-M Dlilkyl N-C
, OR" Alkyl
H OH n'" DllhloClfbamal. H' lhlonocllblmlla

HO, .0 Photp/!orla
R'O , .1 Dlilkylind R'O,
,P, AcId ,P, Dlaryl , P, Thlopholphonyl
HO OH R"O S-M Dllhlophotphllal R"O CI Chlorldl

ROH Alcoholl R-S-H Alkyl Mllclptan R'- S.- R" I'olylullld..

In-2,3 ...1

H,N, , B, DJllkyl
, c-o Ur. H-N N-H Thlour.
H,N k ~" CThI0C8rbanllldl,

.S ~S .S
R-O-C , Xlnlhll. R'- O-C , ..0 XII11hoJIIR R'- O-C , a-On Xlnlhla
s-c , Formal.. Eilin

h .. N+
C - S-M MarCiploblnlOlhltlolt

Fig. 1 Typical type of Compounds used as sulfide mineral collectors 91

Many studie s have been perfor~~itn modulation CHELATING AGENTS AND CHELATE COMPOUNDS
of flotation with xanthates in order to
i mpr ove separation efficiency.
However, there are often considerable Chelating agents are those particular
difficulties involved in this procedure and complexing reagents which are formed of large
the de sired results are not always obtained, organic molecules capable of bonding to the
especially i n the case of minerals of a metal ion via two or more functional groups
c ompl ex composition-hence the importance of with the formation of one or more rings. The
s eeki ng collectors capable of linking compounds formed in this way are known as
selectively with given minerals. Selective "chel a t e s " or "chelate compounds" .
l inkage i s possible if the collector structure The bonds that exist between the chelating
incorporate s active groups with a specific agent and metal vary according to the nature
affinity for certain cations characteristic of of the metal and that of the functional groups
the mineral surface. of the chelating agent , and include
The decision to investigate the possibility of coordination bonds, covalent bonds, ionic
utilizing chelate-forming reagents for bonds and bonds that ar e partially ionic and
flotation was first taken by Gutzeit, 9 who partially covalent.
used chelate-forming compounds in the anionic in any case, every chelating agent is
flotat ion of ox idized i ron ores, their ro l e characterized by the presence of at least one
being to "complex" the heavy metal i ons so a s group capable of donating a pair of electrons
to prevent quartz from floating. to the metal to which it bonds, with the
formation of a coordination covalent bond. In
pract ice, this virtually restricts the choice
to groups cont a i ning N, 0 and S.
Regarding metal ions capable of forming
chelate compounds , the tendency for chelate

formation with donor molecules may be
interpreted as the tendency for met al ions to
fill up unoccupied orbitals and thereby
approach the electronic configuration of the
nearest inert ~as (Table I) .
Tab. I Periodic classifica tion of chela t e-forming metals

Common chelate forming metal s

Metals which f orm few chelates


Li Be B C N o F Ne
I Na Al S1 r S Cl A
I Dr Kv
K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se
I Rb Sr y Zr Cb Mo Tc 'Ru Rh Pd Ag Cd I n Sn Sb Te I Xe

I Cs Ba Lanthanides Hf Ta W Rh Os I r Pt Au JIg Ti Pb Bi Po At Rn
I Fr Ra Ac t i ni de s
1 .L- ---:-~ ~ - _ _ _ :~

basic feature of the ch el at e non-electrolyt es

It is generally true that the farthe r removed is al so their very low electrolytic
the configuration of the metal ions is from dissociation, their i nsol ubi lit y or low
that of the rare gases, the greater will be solubility in water and their s ol ubil i t y in
their tendency to accept pair of electons to non-polar solvents.
approach the structure of the nearest rare gas The factors governing the stability of a
to form chelates . chelate are , es sentially, th e structure of the
The atoms of t he alkali metals and the chelating agent and the natu re of the metal
alkaline-earth metals instead , which on the ion involved.
loss of one and two electrons, respectively , For a given metal i on the sta bi l i t y of a
have the exact configuration of the nearest chelate depends on t he number and size of the
rare gas , have little or no tendency to accept chelate rings, stability i nc r eas i ng with the
a pair of elec trons or to form chelate number of rings up to a maximum of five or
compounds . six. The basic strength and the degre e of
The heavy metals and the transition metals in dissociation of a li gand are a ls o dec i s i ve
particular , have come to be considered as factors for th e s ta bi l i t y of chelat es .
having the s t r onges t tendency to form Other f actors that have a marked inf l ue nc e on
chelat es . the degree of s t a bil ity of a chel ate are
On t he basis of the points made so far it is steric f~~tors and re sonan ce.
apparent that chelating compounds form a class Martell has summariz ed t he e ff ec t of the
of r eagent s that ca n be used for the flotat ion variation vf st r uc t ure of t he che l a t i ng agent
of metallic ores . Their inability to form on the chelating f orc e (Tabl e II ).
bonds wi t h silicon, aluminium, sodium and
potassium guarantees their complete
selec tivity as regards all siliceous and
silicate gan gues , whereas their very s light Table II Ligand s t r uc t ur e i nf l uenc e on
tendency t o link calcium al so en sures good chelat e s t ab il i ty
diff erential action for carbonate gangues.

Bas i city of the l i gand

Number of metal chelate per l i gand
Size of chelate ring
The main properti e s of ch elat e compounds ar e Steric effects
their high stability and the selectivity or Specific effects
specificity of their format ion reactions. A

The stability of the chelate naturally depends It would be as well to replace the expression
not only on the structure of the chelating II specific or selective reagent II with the more
agent but also on the nature of the metal ion appropriate term " specific or selective
to be held . The influence of the metal ion is, reaction", thus clearly demonstrating the
in turn, governed principally by two factors - importance of the reaction medium and reaction
the influence of the donor atoms and that of conditions in eliminating interference caused
the properties of the metals themselves. by other metal ions.
As regards the influence of the donor atoms,
the main donor atoms in chelate structures
are, in prac tice, limited to 0, Nand S, as
Hetals can be divided into three groups FOR METALLIC MINERALS.
according to their tendencies to combine with
oxygen (usually through a normal covalent
bond) or with nitrogen (usually through a At this poin!, from a purely theoretical
coordinate covalent bond). aspect, it may be thought that chelating
These three groups are: agents are compounds with characteristics
- Group 1 - Bonding to oxygen more strongly ideal for the flotation of given metal ores.
than to nitrogen Not only should such reagents react in a
Group 2 - Bonding to oxygen and nitrogen highly selective if not actually specific
comparable in strenght manner on a given mineral but they should also
- Group 3 - Bonding to oxygen less strongly form a highly insoluble coating on the surface
than to nitrogen (Table III) thereof by nullifying the ionic charge of the
surface ion thereby preventing all fo rm of
hydration of the mineral. However, although
Table III Metal tendencies for ~8mbination the resulting chelated particle is fairly
with oxygen or nitrogen hydrophobic, it is, nevertheless, not
sufficiently aerophilic to ensure flotation
because the chelating agents commercially
available are almost all without long
0) N Mg , Ca, Sr, Ba, Ga, In, TI, Ti, Zr , Th,
hydrocarbon chains.
V IV V V V VI The line followed by Marabini et al . in
Si, Ge, SI1, V ,V , Cb , Ta , Mo , U ,
IV I II CoIl their early investigations was that of
U , Fe , rendering particles hydrophobic by making
contemporaneously available for this purpose
e, r III , Fe II ,platinum metals long-chain organic groups (as fuel oil or oily
frothed capable of adhering to the surface
chelate formed by chemical reaction .
I I I II From 1970 systematic research was carried out
Cu , Ag Au, Cu , Cd, Hg,
I II, ll by Marabini et al. to evaluate effectivenes of
Co Ni
chelating agents a11 fuel oil f or metallic
mineral flotation . Most of these studies
h~wever 7~~75 performed on oxidized
The second influence consists in the ml.nerals.
properties of the metals themselves and The first application on sulphide minerals was
particularly, the ionic forces that are in 1973. A chelating agent, 8-hydroxy
related to the charge and to the radius of the quinoline (Fig. 2), with fuel oil was employed
metal ion, as already noted. For instance , in to float f'jxed oxide-sulphide minerals of Zn
the case of the transit ion metals and for most and Pb, Table IV reports the results of
metals in general, stability of the chelate this work. Tests in laboratory cellon an ore
increases as the ionic radius decreases and as containing 7,3% Zn with 1,4% as sphalerite,
the metal charge increases. and 0,9% Pb with 0,40% as galena, led to good
Contrary to what happens in the case of common recoveries that (Table !bV) ,
chemical compounds or non-chelated complex Also Mangalam et. al. proved efficiency of
salts, numerous and diverse factors influence 8-hydroxyquinoline in chalcopyrite flotation.
chelate stability. It is precisely the Another example of chelating reagent in
multiplicity and complexity of these factors , sulphide treatment is the flotwon of
and especially their interrelationships, that cobaltite with~ -nitroso- 6-naphtol .
determine the extent to which chelating agents It is known that dv -nitroso- ~-naphtol (1., N~N)
are selective and , in certain cases, specific is a chelating agent of cobalt . Fig.3 reports
in their ac tion. its formula and chelate structure. Flotation
tests in the Hallimond tube on pure cobaltite

HO .co
;:,... ~

Fig. 2 8-Hydroxyquinoline and 8-Hydroxyquino- Fig . 3 ~ -Ni troso-HNapthol (5).NnN) a nd

line - Zn chelate 73 d-NnN-Co che l a t e 77

Table IV Flotation results of Zn-Pb ore with 8-hydroxyquinoline (oxine) 73

Test Product Conditioning Oxine Gasoline % wt Grade of % Recovery

time, min. g ml float product
% Zn % Pb Zn Pb
1 Concentrates 5 3-5 210 33.29 19.84 2.03 91.33 77 .74
Tailings 66.71 0.94 0.29 .ssi 22.26
Feed 100.00 7.22 0.87 100.00 100 .00

2 Concentrates 10 3-5 210 34.65 19.92 2.09 94.29 80.48

Tailings 65.35 0.64 0 .27 --i.:1.!. 1952
Feed 100.00 7.32 0.90 100.00 100.00

3 Concentrates 20 3-5 210 29 .08 21.99 2 .33 90.20 77.97

Tailings 70.92 0 .98 0.27 ~ 22 .03
Feed 100.00 7.09 0.87 100.00 100.00

with ~-NnN and fuel oil as collector (Fig. 4) 100

gave a 100% recovery. Moreover , tests on
cobaltite/hematite and cobaltite/niccolite 80
gave good selectivity for cobaltite (Table V).
Rinelli and Marabini interpreted these results
on the basis of the greater affinity of d.,- NnN ....."
~ 60

for cobalt than nickel and copper. In fact ,
the graph of conditional constants of a.... -NnN
complexes with Co, Ni and Cu (Fig. 5) shows
this selectivity. 20
At this stage, the results had confirmed
the soundness of the basic theoretical
premise. However problems had emerged that 0 1 4 6 8 IV 12 14
made it difficult for the chelating agents
commercially available to be used on the plant
The main hindrages to the pra tical use of
commercial chelating reagents with aromatic
structure are : Fig . 4 Relationship between cobalti t e recovery
- lack of long-hydrocarbon chain in the and pH (fuel-oil con centration ~ 0.1
molecule gil) Tt
- costitutional insolubility in water of most _ _ _ _ _ with cSl..NnN 0 .025 gil
commercial chelating reagents wi thout &NnN
- hi~h con sumpt ion needed to ensure

Table V Results of cobaltite-9~ccolite and cobaltite-hematite flotation test s
(feed 50% cobaltite)

NBN Fuel-oil %
Mixture concentration concentration pH Products Weight RenW Cl'Y
gil gil

Cobaltite Niccolite

Cobaltite 0.05 0.1 3 Float 42.36 80 .26 12. 81

Niccolite Non-float 54.64 19.73 87.Il)
100.00 100.00 100. 00

Cobaltite 0.03 0 .05 3 Float 60.75 98.73 23.00

Hematite Non-float 1.27 77.00
- - -
100.00 100 .00

this type of reagent in comparison with

'cQ. 40t.- tog
~~ -Co
-_ classical collectors.
II To conclude, for all practical purposes che
chelating agent must have a long chain, be
30 water-soluble and have proper chelating groups
endewed with good but not too high an affinity
for mineral cations and capable of acting
selectively on a given mineral against an
At this point Marabini et a l. orientated
their research towards the design and
synthesis of new compounds incorporating into
o a single molecule long-chain hydrocarbons and
chelating groups satisfying the requirements
5 L-,r-"I'......._.....,...~.................................. _ for specific collectors.
1 2 3 4 5 8 7 B 9 10 11 12 13 14 Nevertheless, to envisage an ideal chelating
collector, it was necessary:
- to state a method capable of evaluating in
advance the efficiency of a ligand as
Fig . 5 Conditional constants of~NBN complexes selective collector towards a given mineral
wi th Co - Ni and Cu as a func tion of pH 'i7 - to fully understand the intimate mechanism
of the interaction between ligands and
satisfactory flotation results mineral surfaces, studying the associated
static and kinetic equilibria
The first factor generally makes the chelating
agent alone incapable of rendering the mineral
surface ufficiently aerophilic. To overcome
this difficulty the chelating agent was
provided artificially with a long-chain
organic portion by introducing a neutral oil
(fuel oil) . But the introduction of a new
liquid phase into the flotat ion pulp is
To solve the first problem, a method has been
damaging to the system as a whole and is not
developed, with the help of a computerized
available on the industrial scale .
program, for the selection of chelating
The insolubility in water and high consumption
reagents theoretically provided with a
are both related to the chemical
selective collecting power towards one given
characteristics of the reagents commercially
mineral rather than another , based the
available , and particularly to the nature of calculation of conditional constants 780n

their chelating groups.
Conditional constants are s o called because
Furthermore, too strong an affinity of
they depend on the experimental conditions
chelation groups for mineral ca t i ons can lead
such as metal-water and ligand-water reactions
to detachment of the chelate from the particle
and consequently they vary as a function of
surfaces and consequential high consumption of

pH. These constants can be calculated from rutile from iron oxide, and for the selection
published values of equilibrium constants. It of ligands capable of separ:t9~~ Zn and Pb
can be assumed that a given ligand L acts as a oxides from Fe-Mg-Ca oxides Two series
selective collector for a mineral that of reagents containing chelating functional
contains the cation M against another that groups for Pb and Zn (Fig. 6) and an alkyl or
contains the cation M if it forms with M a alcoxyl chain linked to the aromatic ring have
2 1
complex M L the conditional constant of which MERCAPTOBENZOTHIAZOLE AMI NOTHIOPHENOL
. 1
Log K'M L'>6
and if the difference between the conditional
constants of the complexes M and M is
1L 2L

Log K'M L - log K'M L~5

1 2 Where R alkyl chain ( 12 C atoms)
The first criterion characterizes the absolute alcoxyl chain
chelating power toward the cation M. The
.1 .
second defines the difference in the re 1 at1ve
chelating power of a ligand vis a vis two
cations M and M .
F or a gener1c. 2 cat10n
. M th e compI '
reaction is
M + L ~ ML
the equilibrium constant of which is defined
by the e~ression
ML = [M][L]
where K is the stability constant dependent
on ioni~L strength and temperature, and[M,][L] Fig. 6 Mercapto Benzothiazole and Amino Thi _
and [ML]are the concentrations of free met~l' phenol structure and chelating bond 1
free ligand and complex, mole 1, 81
been prepared and tested successfully on
oxidized Pb-Zn ores.
From consideration of the influence of all the Although obtained on oxidized ores, the
side reactions in which M and L are involved a results obtained with these two series of
new constant may be defined for the same reagents of provide interesting pointers on
equilibrium: the criteria to be followed in designing mixed
K' ML =
fi.lL] aliphatic-aromatic chelate-type collector
[M'][L'] molecules. Studi~s are now being done g their
where K' ML 1S the conditional constant and application in sulphide ore treatment

[MI] [L,l are concentrations of free and

hydroxylated m=ial and of free and protonated
ligand, mole 1 , r es pec t i vel y . CHELATING COLLECTORS FOR SULPHIDE MINERALS
From the dissociation coefficients
[MJ Examples of the use of chelating agents in
[L'1 sulphide flotation are known from studies on
[L] traditional thiol co Ll.ect or s , In fact, both
the conditioRal constant may be written dixanthogen (Fig. 7) and thigocarbamate (Fig .
ML = --r- ~ 8) act as chelating agents.
M L Many researchers have studied the use of
and chelating reagents in metallic mineral
Log K'ML = log K ML -log M -log L flotation. Recently , a review of ~ese studies
where M and L are calculated from has been published by Pradip. ' He showed
mathemat1cal relat10nships considering the that most of the studies concern ox i di.zed
equilibrium constants of the metal-water minerals there are very few applicat~gns to
reactions. The validity of these criteria has sulphides. However, hav ing observed that
been proved in many systems by flotation tests thionocarbamates8~ave chelating properties ,
in the Hallimond tube and in laboratory cells. Ackerman et al. began to investigate their
The method has been applied both to the action.
selection of collectors for separation of


"~I N - CZH
\ /
\ I
\ I
\. I

Fig. 8 Proposed attachment of thionocarbamate

to a mineral surface (modified after
Fig. 7 Proposed attachment of xanthogen to the Bogdanov et al., 1976) 86
surface of a c_opper mineral 86
They studied alkyl substituent effects on the
oJ" ~ I I I I I I

I ~ ~-=1 _8
performance of thionocarbamates as copper
I ' \
sulphide and pyrite collectors and found that 8i \ .\ \
a-alkyl or N-alkyl substituent better controls r \\ \
hydrophobici ty, electron density and oJ 60'- \ ' \
accessibility of the adsorbed collectors, with .
~ -~
~. ~

resulting influence on collector density. This
~~ in accordance with Somasundaran' s studies 0 1
2 20

on substituent effect on functional groups QJ

of a chelating agent. 90 C < (01

Moreover, Ackerman et al. proposed for OM " oH


the flotation of sulphide minerals a new

commercial chelating agent , I-hydroxy- Fig . 10-a) Chalcopyrite flotation with various
cthyl-2-heptadecenyl glyoxalidine (amine 220), non-sulphydryl collectors.(a) % reco-
the structure and mechanism of surface very vs. pH, (b) rate constant, K, vs.
attachment of which are reported in Fig . 9. pH. l=sodium laureateJ 2=sodium oleateJ
They also found that the collecting power of 3=sodium dodecyl sulfate: 4=dodecyl a-
glyoxalidine is superior to that of other mine acetate, 5=Amine 220 90
non'-sulphydryl collectors (Fig. 10).

CH 2

Fig. 10-b) Chalcocite flotation with various

non-sulphydryl collectors. (a) %reco-
very vs. pH. (b) rate constant, K, vs.
pH. 1= sodium laureate, 2=sodium oleate,
3=sodium dodecyl sulfateJ 4=dodecyl
amine acetate, 5= Amine 220
Fig. 9 Proposed attachment of l-hydroxyethyl-
2-heptadecenyl glyoxalidine (Amine 220)
to the mineral surface by forming a Two classes of chelating collectors for
copper chelate 90 sulphide mineral flotatior~.)l have been
synthesized by Klimpel et al ; the general
molecular fonnula of the two classes are
reported in Fig. 11. They also studied the

1 11 1 u
R-X-(CH) - R -S-R or R -C-y-R
2 n \S


Fig. 11 Chelating collectors for sulphide minerals 91

effect of substituents and chain length of Fig. 14 reports the selectivity of the three
these reagents on copper recovery (Figs. 12 selected reagents over pyrite and
and 13) and chose the examples reported in copper/molybdenum recovery, and shows that F2
(Fig. 11).
57 _ _

1.0421,,- )KAX/57

'KAX -
57 c:::::::. / ' L02I1

iI 701-
:> FZI.02I1 / 1 . 0211 -

0.0 L...L--IL.L....J_.L....l_.L....l_.L--l._J......!:---IL...-~
Coppor RICOW'f
1 1 1 10 12 I R..,...,.,
.. c;."a-I

eol-_ _--lI ~I ---~I ---:

Fig. 12 The effect of chain len~ on the
recovery of copper 91
o 10
" Pyrltl In C..-......
" za

104 r-...,...-r-.,........,~ .....--r--r--r-.-""T-r-.,..-r-..,

Fig. 14 Relationship of copper/moly~~enum
0.1 recovery to pyrite content.
collector dosages in kg/metric ton

I ..
indicated in () MIBC frother at 0.015
kg/metric ton

J F1, F2, and 57 are experimental collect

ors products of the Dow Chemical Co.
-;::..... =.-:.,.-----.---
] 0..

0.0 l-..L---J'-.L.--I_...L-I._...L-l._.l.......L._J....-'-_'--..J
and S7 collectors increase recovery above
1 10 12 I. commercial standards.
In another work Klimpel and Hansen
Fig. 13 The effect of chain length on the reco 92 analysed all influences of controlled
very of copper with various chemical changes in chemical structure of this proposed
collector structures 91 series of sulphide mineral chelating

They confirmed that several members of this
family have the potential of achieving usage
on a commercial scale .
Recently, new classes of collectors with
chelating properties for copper sulphide
minera~1 have been presented by Nagaraj et
al. They are alkoxycarbonyl alkyl
thionocarbamates and thioureas, dialkyl (or
diaryl) monothiophosphates and dialkyl
monothiophosphinates . Their structure is
reported in Fig. 15. FTIR ATR and
thermodyn!~ic studies of these structures have
shown that alkoxycarbonyl alkyl
thionocarbamates and thioureas adsorb on
sulphide minerals via the formation of a
six-membered chelate between metal cation and
groups C=S and c=o. There are, however,
differences between these two collectors in
terms of flotability of the various sulphide

Appearance p" range of

Stabi lity

o s 10.5 Clear brown oily liquid 3-11

1. AlkoxYcarbonyl Alkyl
Thionocarbamates R-O-C-N"-C-O-R (for R C2"S; R' i-C4"gl

o s
2. Alkoxycarbonyl Alkyl
R-O-C-N"-C-N"-R >12 low melting solid with 3-12
appreciable solubility in
(R C2"S; R' 1-C4"gl

3. Oialkyl or diaryl S 1.55

R-~. Clear pale yellow 1-12
Honothlophosphates p-o- liquid
(R 1-C4"g)

3. Oialkyl S
3.17 Clear pale yellow 1-12
Honothiophosphlnates ~p-o- liquid
(R i-t 4"g)

Fig. 15 New Classes of Collectors

As regards the phosphorus aci~~ (Fig. 13)

In fact , as can be seen from Figs. 16 and 17 Nagaraj et a.I , have shown that the
and 18, alkoxycarbonyl thionocarbamate floats monothiophosphates are acid-circuit col l ector s
copper-rich minerals such as bornite, of sulphides, the monothiophosphinates are
covell i te and chalcoci te faster than effective in neutral and midly alkaline
chalcopyri te, whereas the modified thioureas circuit and the dithiophosphates are effective
float chalcopyrite very effecti vely. Such at pH 9 (Fig. 19), and they proposed a
difference s are observed not only \Yith pure mechanism to explain the differences in such
minerals but also with natural ores containing collector activities between the phosphorus
the various mineral s . acids .

COLUCTO". _ I.,.
QAI! ,
-.-- IIeCTC. 0..12

- +-
llCTV . " 12'
II: ~

> ,:
o .......-+
c ;;
~ TO o
U g

II: COLUCTOA., VI f R. . . .)

+ .CTU 10 .1.:1.0

RO lSe ell GRAD!!. "


-+- ecre, 0 .1:"

- +- llCTV . 0.1:'
Fig. 16 The difference in col le c t or activity
between Ethoxycarbonyl Thionocarbamate
and Ethoxyca rbonyl Thiourea (Isobutyl
homologs) for a Western U.S. Copper ore
.~ , ......
--<~ . . .". '+

cont a i ni ng pr edomin antly chaleopyrite <3 +.-

(pH 9.3; Cu r eco very vs. grade ; Cumula
tive time s 1, 3 and 12 min. ) 93


Fi g . 18 The di ffe re nce i n Col l ec t or act 1v1 t y

CeLL_eTCR. 1II+laI\
O"I! C
be t ween Et hoxy carbonyl Thionocarbama te
IlCTU.o.on and Et hoxyc arbonyl Thiourea ( Is obut yl
homologs ) a s a f unction of pH f or two
South Ame ri can Cu ore s con taini ng pre -
dominan tly co pper - ri ch mine ra l s ( chal -
coci te , bo r n ~; e and covel l i t e ) . Flo t .
Time 7 min. 9 J


Fig. 17 The diff erence i n Col lector act1v1ty

between Ethoxycarbonyl Thionocarbamate 80 :
and Ethoxycarbonyl Thiourea as a fun-
ction of pH for a Western U.S. Copper 'f.
ore containing pr edominantly cha lc o-
a: 80-
+ !
pyrite. (6 min. Flot.) 93 >
- . - DIBMTP. 0.01
-*.- DIBDTP. 0.01

o ~f-,-.""""T"""""""""T"""""""""'T"""""""'''''''''''''''''''TT'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''.,...,.....t
2 3 7 10 11
There have been fewer studies on chelating
reagents as collectors for sulphide flotation
Fig . 19 The diffe r ence s in co l lecto r ac t iv1t1e s
than on oxidized minerals. In fact, reagents
betwe en monoth ioph osph at e , monothio -
for flotation of oxidized Zn and Pb in the
phosphinat e and dithioph oshat e (a l l
pilot plant have already been proposed. This
Isobutyl homologs ) as a f unc t i on of pH
opens up new prospectives in the application
for a West e rn U.S, Copper ore. Tot al
of these reagents.
Flotation Time 7 min . 93
Applications on sulphides are presently only
on laboratory scale and on a limited number of
reagents , but further results obtained show
that this s ubj ec t deserves attention. All
studies should consider not only the active

functional group but also steric and Sulphide, Galena and Metallic Lead
thermodynamic factors . Subs~~~es . , Trans. IMM, 72, 1963, 407.
On the one hand the mechanism of action of 10) Plante E.C. and Sutherland K.L., Effects
these reagents should be studied and, on the of Oxidation of Sulphide Minerals on Their
other criteria should be defined in order to Flotation Properties., Trans. AIME, 183,
project optimal chelating structures for a 1949, 160.
given mineral relative to its reticular 11) Herd H.H. and Ure W., Surface Chemistry in
structure and to its metallic active sites. the Flotation of Galena. , J. Phys . Chern.
Moreover, efforts should be made towards 45, 1941, 93 .
reducing the reagent cost and reagent 12) Taylor T.C. and Knoll A.F ., Action of
consumption in order to make these reagents Alkali Xanthates on Galena., Trans. AIME,
economically advantageous in mineral 112, 1934, 382.
processing application. However, as many 13) Sparrow G., Pomianowski A. and Leja J.,
problems in mineral and sulphide mineral Soluble Copper Xanthate Complexes.
processing not can be solved by conventional Separation Science 12, 1977, 87.
reagents, probably in the coming decades a 14) Gaudin A.M. and Schuhmann R., The action
number of reagents already tested in of Potassium n-Amyl Xanthate on
laboratory and at the pilot-plant scale will Chalcocite. , J. Phys. Chern. 40, 1936, 257.
be commercialized and the mineral industry 15) Heyes G.W. and Trahar W.J ., Oxidation-
will collaborate with the production of Reduction Effects in the Flotation of
tailor-made chelating collectors. Chalcocite and Cuprite., Int. J. of Min.
Proc., 6, 1979, 229.
16) Kowal A. and Pomianowski A. , Cyclic
Voltammetry of Ethyl Kanthate on a Natural
References Copper Sulphide Electrode., Electroanal.
Chern . and Interfac. Electrochem. 46, 1973,
1) Healy TW. , Trahar WJ Challenges in 17) Richardson et aI, Electrochemical
mineral sulfide flotation. Challenges in Flotation of Sulfides: Chalcocite-
mineral processing , Sastry K.V. ed., AlME, Ethylxanthate Interaction. Int. J. of Min.
Colorado 1989. Proc , , 12,1984,87 .
2) Sutherland K.L. and Wark I.W . Principles 18) Walker G.W., Stout J.V. and Richardson
of flotation. Australasian Inst. of Mining P.E., Electrochemical Flotation of
metallurgy, Melbourne, 1855. Sulfides : 1. Reactions of Chalcocite in
3) Rogers J . : Princ iples of sulfide mineral Aqueous solutions. Int. J. Min. Proc. 12 ,
flotation . Fuerstenau M.C . ed , Froth 1984, 55.
Flotation , Gaudin A.M. memorial volume , 19) Gaudin A.M. and Finkelstein N.P.,
AlME New York 1976. Interaction in the System PbS-KEX-0
4) Poling G.W. : Reactions between thiol Nature, 201, 1965, 389 .
reagent and sulfide minerals. Furestenau 20) Finkelstein N.P., Kinetic and
M.C. ed. , Flotation , Gaudin A.M. memorial Thermodynamic Aspects of the Interaction
vol., AIME, New York 1976 pp . 334-363 . Between Potassiwn Ethyl Xanthate and
5) Ackerman P.K., Harris G.H. , Klimpel R.R. Oxygen in Aqueous Systems., Trans IMM, 76
and Aplan F.F .: Evaluation of flotation 1967, C5l.
collectors for copper sulfides and pyrite 21) Finkelstein N.P. , Quantitive Aspects of
I. Common Sulphydryl collectors . Int . Min. the Role of Oxygen in the Interaction
Proc ., 21, 1987 pp. 105-127. Between Xanthate and Galena., Sepa r ation
6) Granville R. , Finkelstein N.P . and S.A . Science 5, 1970, 227.
Allison. Review of the reactions in the 22) Gardener J .R. and Woods R., The Use of A
flotation system Galena-Xanthate-Oxygen, Particulate Bed Electrode for the
Trans. IMM, B, 1972. Electrochemical Investigation of Metal
7) Hagihara H. Mono and Multilayer Adsorption Sulfide Flotation., Aust . J . Chern., 26,
of Aqueous Xanthate on Galena Surfaces. J . 1973, 1635.
Phys. Chern. , 56, 1952, 616. 23) Healy T. W. and Moignard M. S.: A review of
8) Greenler R.G., An Infrared Investigation electrokinetic study of metal sulfides.
of Xanthate Adsorption by Lead Sulfide, J . Fuerstenau M.C. ed., Flotation, Gaudin
Phys. Chern. , 66 , 1962, 879. A.M. memorial vol., AIME N.Y. 1976, p ,
9) Leja J., Little L.H. and Poling G.W., 275.
Xanthate Adsorption Using Infra-Red 24) Woods R.: Electrochemistry of sulphide
Spectroscopy. I-Oxidized and Sulphidized flotation. ed . Flotation, Gaudin A.M.
Copper Substrates . 2-Evaporateu lead memorial vol. , Fuerstenau M.C. AIME NY

25) Richardson P.E., Monst E.E. Surface 38) Mielezarski J. In situ ATR-IR
stoichiometry of Galena in aqueous spectroscopic study of xanthate sorption
electrolytes and its effect on xanthate an marcasite. , Coll. Surf. , 17, 1986,
interactions. Fuerstenau M.C. ed., 235-248.
Flotation, A.M. Gaudin memorial vol., AIME 39) Leppinen J.O., Basilio C.I. and Yoon R.H.
N.Y. 1976. FTIR study of thionocarbamate adsorption
26) Nowak P. et al. The applicability of Emf on sulfide minerals. , Coll. Surf. , 32,
measurements to evaluation of 1988, 113-125.
thermodynamic properties of the Cu-S 40) Leppinen J.O., Basilio C.I. and Yoon R.H.
system., J. Electroanal. Chern. 171, 1984, In situ FTIR study of ethylxanthate
pp. 335-358. adsorption on sulfide minerals under
27) Nowak P., Krauss E., Pomianowski A.. The conditions of controlled potential., Int.
electrochemical characteristics of the J. Min. Proc. , 26, 1989, 259-274. ----
galvanic corrosion of sulfide minerals in 41) Johansson L.S., Juhanoja J., Laajalehto
short-circui ted model galvanic cells. , K., Suoninen E. , Mielczarski J. XPS
Hydrometallurgy 12, 1984, pp. 95-110. studies of xanthate adsorption on metals
28) Barzyk W., Malysa K., and Pomianowski A.. and sulfides., Surf. Interf. Anal., 9, pp.
The influence of surface oxidation of 1986 , 501-505.
chalcocite on its floatability and 42) Termes S.C., Richardson P.E . Application
ethylxanthate sorption., Int. J. Min. of FTIR spectroscopy for on situ studies
Proc., 8, 1981, pp. 17-28. of sphalerite with aqueous solutions of
29) Chander S.: Oxidation reduction effects in KETX and with diethyldixanthogen. , Int. J.
depression of sulphide minerals. A Min. Proc., 18, 1986 , p. 167-178.
review., Min. Met. Processing, 22, 1985, 43) Mielczarski J . , XPS study of ethyl
pp. 26-35. xanthate adsorption on oxidized surface of
30) Luttrell G.M., Joon R.M . Surface studies cuprous sulphide., J. Colloid Interface
of the collectorless flotation of Sci., 120, 1987, 201-209.
chalcopyrite. , Coll. Surf, 12, 1984, pp , 44) Laajalehto K. , Johansson L.S., Mielczarski
239-254. J. , Anderson S. and Souninen E. . Electron
31) Giesekke E. W.. A review of spectroscopic spectroscopic studies of interaction of
techniques applied to the study of xanthate collectors with different
interactions between minerals and reagents substrates. In K.S.E. Forssberg ( Edi t or) ,
in flotation systems., Int. J. Miner. Proceedings of the XVI International
Process., 11, 1983 , 19-56. Mineral Processing Congress , Stockholm,
32) Kongolo M., Cases J.M., Burreau A. and 1988, pp. 691-702.
Predali J .J . Spectroscopic study of 45) Mielczarski J. and Minni E.. The
potassium amylxanthate adsorption on adsorption of diethyldithiophasphate on
finely ground galena. M. Jones and R. cuprous sulphide ., Surf. and Interf. Anal.
Oblatt (Editors), Reagents in the Minerals 6, 1984, 221-226.
Industry. IMM, Rome, 1984, pp. 79-87. 46) Mielczarski J. and Suoninen E.. XPS study
33) Walker G.W., Richardson P.E. and Buckley of ethylxanthate adsorbed onto cuprous
A.N. Workshop on the notation related sulphide., Surf and Interf. Anal., 6,
surface chemistry of sulfide minerals., 1984, 221-226.
Int. J.M. Proc. 25, 1989, 155-158. 47) Mielczarski J. and Suoninen E.. XPS study
34) Little L.M., Poling G.W., Leja J.: IR of the oxidation of cuprous sulphide in
spectra of xanthates c ompounds , , Can. J. aerated aqueous s ol ut i ons., Coll. Surf ,
Chern. , 39, 1989, 745-754. 33, 1988, p. 231-237.
35) Marabini A.M., Cozza C.. Determination of 48) Cecile J.S. , Bloise R., Barb ery G. Galena
lead ethylxanthate on mineral surface by depression with c hr oma t e i ons after
IR spectroscopy., Spectrochim. Acta, 388, flotation with xanthates. A kinetic and
1983, p. 215. spectrometric study i n Jones M. ed.
36) Mielczarski J. and Leppinen J.: Infrared Complex sulphide ores. IMM. , London, 1980,
reflection-absorption spectroscopy study pp. 159-170.
of adsorption of xanthates on copper., 49) Predali K.C. et al. XPS study of xanthate
Surf. Sci, 187, 1987 , pp. 526-538. adsaption on peprite mineral surfaces., J.
37) Mielczarski J., Nowak P., Strojek J.W. and Coll Interf. Sci , 103, 1985, n. 1, 1985.--
Pomianoxski A. Investigation of the 50) Buckley A.M., Woods R. and Wauterlood H.J.
products of ethylxanthate sorption on An XPS investigation of the surface of
sulfides by IR-ATR spectroscopy, in J. natural sphalerites under flotation
Laskowski ed , Proc of the XIII Int. Min. related conditions., Int. J. Min. Proc.
Proc. Congress, Warszawa Part. A 1981, pp. 26, 1989. 29-49.
40-131, 1981.

Steiningen J. . The depression of
51) Pillai K.C., Young U.Y. and Bankris a.M..
XPS studies of xanthate adsorption auto sphalerite and pyrite by basic complexes
of copper and sulphidryl flotation
galena surfaces ., Appl. Surf. Sci. , 16,
collectors., Trans. AIME, 1968 , p. 34 .
1983, pp. 322-344 .
Finkelstein N.P. Allison S.A.. The
52) Page P.W. and Hazell L.B. X ray
photoelectron spectroscopy (XPS) studies chemistry of activation, deactivation and
of potassium amylxanthate (KAX) adsorption depression in the flotation of zinc
on precipitated PbS related to galena sulphides: a review. Fuerstenau M.C. ed ,
Flotation, Gaudin A.M. memorial volume,
flotation., Int. J. Min. Proc , , 25, 1989,
AlME New York, 1976, pp. 414-457.
pp. 87-100.
66) Ball B., Richard R.S.. The chemistry of
53) Pillai K.C. , Young V.Y. and Bockris Y.O.M.
X-ray photoelectron spec t roscopy studies pyrite activation and depression.
Fuerstenau M.C . and Flotation, Gaudin M.A.
of xanthate adsorption on pyrite mineral
memorial volume, AlME, New York , 1976 .
surfaces ., J. ColI. Int. Sci, 103, 1985,
Ghiani M., Satta F., Barbaro M. and
pp , 145-153.
Passariello B Flotation of sphalerite
54) Predali J .J., Brion D., Harper J.,
from pyrite by use of copper xanthate and
Pelletier B.. Characterisation par
sodium cyanide. Reagents in minerals
spectroscopie ESCA des etats de surface de
industry , ed. Jones M. and Oblatt, IMM
mineraux sulfures fines de Pb-Zu-Cu-Fe
lors de la flotta t i.on , Proc . of the XIII 1984.
68) Chander S. Inorganic depressants for
Int. Min. Proc , Congress , Warsaw part A,
sulfide minerals" Reagents in Mineral
1981, p. 55.
Technology, P. Somasndaran and B. Moudgil
55) Mielczarski J. , Suoninen E. Johansson L.S.
ed s , , Marcel Dekker, New York, pp ,
and K. Laajalehto An XPS study of
adsorption of methyl and amyl-xanthates an 428-468.
copper., Int. J . Min. Proc , 26,1989, p , Gutzeit G. Chelate-forming organic
compounds as flotation reagents.,
Trans-AlME 1946, v. 169, p. 272.
56) Bertil I. Palsson and K.S.E. Fors sberg.
70) Martell A., Calvin M. Chemistry of the
Computer-assisted calculations of
thermodynamic Equilibria in Sphalerite- metal chelate compounds, Prentice Hall,
xanthate system., Int. J. Min. Proc., 1952.
71) Usoni L., Rinelli G., Marabini A.M ..
1989, 26, 223-258.
Chelating agents and fuel oil: A new way
57) Bertil I. Palsson and Forssberg RSfi.
Computer-assisted calculations of to flotation, AlME Centennial Annual
thermodynamic Equilibria in the galena - Meeting, New York , February 26-March 3,
ethylxantate system., Int . J. Min. Proc., 1971.
72) Marabini A.M . Flottazione e flocculazione
23, 1988, 93-121.
selettiva., Boll. Ass. Min. Subalpina XIX,
58) Partyka S., Arnaud M. , Lindheimer M..
n , 3-4, 1982.
Adsorption of Ethyl-xanthate onto Galena
at low s ur f ace coverages., ColI. Surf., 73) Rinelli G., Marabini A.M .. Flotation of
zinc and lead oxide-sulphide ores with
26, 1987, 141-153.
Maillot M. , Cecile J L. , Bloise R chelating agents , 10th Int. Min. Proc.
Stability of ethylxanthate ion in neutral Congress, IMM, London, 1973.
and weakly acidic media ., Int. J. Min. 74) Marabini A.M. New collector for
cassiterite., Trans IMM, Sect. C. vol. 84 ,
Proc . , 13, 1984, 193-210.
60) Shimoizaka J. et al . Depression of galena 1975, C 177.
flotation by sulfite or chromate ion , in 75) Rinelli G., Marabini A.M., Alesse V.
Fuerstenau M.C . cd ., Flotation , Gaudin Flotation of cassiterite with
salycilaldheide as a collector, Flotation,
A.M . memorial vol . , AIME , NY, 1976.
Gaudin A.M. Memorial volume Fuerstenau
61) Rinell i G., Marabini A., Alesse A..
Depressing action of permanganatc on M.C .ed., vol. 1, 1976, p. 549.
76) Mangalam V. et al . Zeta potential and
pyrite and galena flotation. Jones M.J.
flotation studies of chalcopyrite fines
ed. Complex sulphide ores. Rome IMM,
London, 1981, pp. 199. with 8-hydroxyquinoline., ColI. Surf., 7,
1983, pp. 209-220.
62) Popov 5 .5. et al.. Effect of the
77) Marabini A.M., Rinelli G. Flotation of
depre ssing agents FeSo and NaCN on the
cobalt minerals with chelating agents as
surface properties of Galena in the
collectors. , ATB Metallurgie , XXII 1,
flotation system. , Int. J. Min. Proc. 24,
1982, p. 17-22 .
19 88 , 111-12.1.
63) Elgillani D.A. and Fuerstenau M.C . 78) Marabini A.M., Barba ro M., Ciriachi M. A
Mechanisams involved in cyanide depres s ion calculation method for sel ect i on of
of pyrite., Trans . AIME, 241, 1968.

complexing collectors having selective based on chelation chemistry, SME/AIME
acion an a cation., Trans IMM, Sec C, 92, Annual Meeting Las Vegas, Nevada, Febr. 27
1983, czo-c26. - March 2, 1989.
79) Marabini A.M . Study od adsorption of 93) Nagaraj D.R Basilio C. and Yoon R.M
Salicylaldhiede on cassiterite., Trans. The chemistry and structure-activity
IMM, Sect. C., 87, 1978, C75. relationships for new sulfide collectors.
80) Marabini A.M., Rinelli G. Development of 118th SME/AIME annual Meeting, Las Vegas,
a specific reagent for rutile flotation., Nevada, Feb . 27 - March 2, 1989.
SME Trans. , 274, 1983, p. 1822.
81) Marabini A.M., Alesse V. , Barbaro M. New
synthetic collectors for selective
flota tion of zinc and lead oxidized
minerals., XVI Int. Min. Proc. Congress,
Forssberg ed. Elsevier, Amsterdam Vol.
1988, pp. 1197-1208.
82) Marabini A.M., Barbaro M., Passariello B.
Flotation of cerussite wi th a synthetic
chelating collector., Int. J. Min. Proc.,
25, 1989, p. 29.
83) Marabini A.M., Cases J. Barbaro M.,
Chelating reagents as collectors and their
adsorption mechanism, Challenges in
Mineral Processing , Sastry K.V., ed AIME,
Colorado, 1989.
84) Cozza C., Di Castro V., Polzonetti G.,
Marabini A.M.. An X-Ray Photoelectron
Spectroscopy Study of the Interaction of
Mercapto-Benzo-Thiazole with cerussite, in
press on Int. J. of Min. Proc.
85) Bornengo G., Marabini A.M. and Alesse V. ,
Italian Patent 22019 A/89, 1989.
86) Acherman P.K. , Harris G.H., Klimpel R.R.
and Aplan F.F., Evaluation of flotation
collectors for copper sulfides and pyrite.
III Effect of Xanthate chain length and . , Int. J. Min. Pr-oc . , 21, 1987
87) Pradip: Application of chelating agents in
mineral processing., Min. Met. Proc., 80;
88) Ackerman P.K., Harris G.H. , Klimpel R. and
Aplan F.F., Effect of alkyl substituents
perfomance on thionocarbamates as copper
sulphide and pyrite collectors. Reagents
in mineral industry, ed , Jones M. J. and
Oblatt, IMM , Lond, 1984.
89) Somasundaran P. and Nagaraj P.R.,
Chemistry and applications of chelating
agents in flotation and flocculation,
Reagents in mineral industry , ed. Jones
M.J. and Oblatt R., IMM, London, 1984.
90) Ackerman P.K. , Harris G.H. , Klimpel R.R.
and Aplan F. F., Evaluation of flotation
collectors for copper sulfides and pyrite
II. Non-sulphydryl collectors. , Int. J.
Min. Proc., 21, 1987 , pp 129-140.
91) Klimpel R.R. , Hansen R.D. and Fee B.S. ,
New collector chemistries for sulfide
mineral flotation ; SME/AIME, Annual
Meeting, Phoenix, Arizona , Jan 25-8, 1988.
92) Klimpel R.R. , Hansen. Recent work on
developin~ new sulfide mineral collectors

Mineralogy of and potential beneficiation process for the
Molai complex sulphide orebody, Greece
M . Grossou-Valta 0

Department of Technology, Mineral Processing and Metallurgy, Institute of Geology and Mmeral
Exploration, Athens, Greece
K. Adam
Department of Research and Technology, Aegean Metallurgical Industries (M ETBA SoA. J,
Athen s, Greece
DoC. Constantinides ,
Department of Economic Geology, Institute of Geology and Mineral Exploration, Athens, Greece
J.M. Prevosteau , 0 0 " ,

Department of Analysis, Bureau de Recherches Geologiques et MlnJeres, Orleans, France

E. Dimou l OA h
Department of Mineralogy and Petrology, Institute of Geology and Mineral Exp oretion, tens,


The Molai sulphide deposit situated in

S.E.Peloponnese, Greece, was located after a
systematic exploration programme which
commenced in 1980. Geostatistical evaluation of
the drilling campaign results indicated mineable ore INTRODUCTION
reserves of 2.3 x 106 tons with an average
estimated content of Zn: 9,0%, Pb: 1,8% and Ag: The Molai polymetallic sulphide deposit is
50 glt. situated in the Lakonia county, S.E. Peloponnese
The geological and mineralogical features (Fig. 1). It was located after a systematic
led to the conclusion that the Molai ore body exploration programme carried out by the Institute
belongs to the stratiform, syngenetic, volcanogenic of Geology and Mineral Exploration (lGME), which
massive sulphide deposits . Detailed mineralogical commenced in 1980 and included geological
analysis of drill core and representative bulk ore mapping in various scales (up to 1:1000),
samples indicated sphalerite as the main Zn bearing geochemical and geophysical surveys and an
mineral , while smithsonite and monheimite were extended drilling campaign. The mineralisation was
also observed. Galena and small quantities of explored to a maximum distance of 1300m along
anglesite and cerrusite were recorded as the strike and 200m across dip and was found to be
Pb-bearing minerals. Silver is encountered in the developed along rather well defined zones and
tetrahedrite-tennantite mineral group mainly sub-zones. Geostatistical evaluation of the drilling
associated with sphalerite and in the Ag-Sb bearing campaign results from the surface (+ 165m) down
minerals diaphorite, pyrargirite and stephanite to -5Om level indicated mineable reserves of 2.3 x
observed as inclusions within galena. 106 tons with an average estimated grade of Zn:
The fine dissemination and association of 9.0%, Pb: 1.8% and Ag: 50 g/t. (1) Above data will
base metal sulphides, combined with the presence be reevaluated after the completion of ongoing
of oxidised zinc and lead minerals adversely affect research activities in the Molai area. (2,3) It is
the beneficiation of the Molai sulphide deposit. The suggested that the sulphide ore body continues at
liberation size varies from 26J.l.m for sphalerite to depths below the levels already explored. Presently
12J.l.m for galena and 2J.l.ID for anglesite. Prolonged an exploratory gallery is being opened in the main
grinding and high reagent consumption is required ore body (Western zone) by Aegean Metallurgical
to improve zinc recoveries. Metal losses in slimes, Industries (METBA S.A), with the aim to
early zinc activation in the galena circuit and silver investigate the optimal mining method and
distribution between the lead and zinc concentrate estimate the corresponding capital and operating
are recorded, as major problems encountered in costs.(2, 3) During trial mining exploitation bulk ore
the differential flotation of the Molai sulphide ore. material will be produced for subsequent
In this article mineralogical analysis data are beneficiation tests in laboratory and pilot plant
employed to corroborate the prevailing theories on scale . (4 ,5) Technical and economical data thus
the Molai ore body genesis and formation as well obtained will be employed in the final feasibility
as to assist in developing an economic beneficiation study for the potential exploitation of the Molai
process for the treatment of this complex, finely sulphide deposit.
disseminated sulphide mineralisation.

GEOLOGICAL SETTING - SULPHIDE minor and occurs occasionally. The
MINERALISATION stratiform development of the mineralisation is
evident both in macroscopic and microscopic ore
The geological environment of the deposit textures. Extensive brecciation is another main
comprise an assemblage of clastic and volcanic characteristic of the ore.
rocks belonging to the Tyros beds. The base metal
sulphides are hosted in the volcanic rocks
consisting of tuffs, tuffltes and other pyroclastics,
as well as massive aphanitic and porphyritic lavas, ORE GENESIS
which have an andesitic to basalt-andesitic
composition and calc-alkaline character. The base metal mineralisation of the Molai deposit
Intense alteration has affected both the is closely associated with the permotriassic volcanic
volcanic rocks hosting the mineralisation and the activity. The age of the ore was determined by
adjoining rocks with an adverse impact on their seven lead isotopic analyses organised by IGME
mechanical properties. (2) The alteration is mainly and conducted by N.Gale at Oxford University.
represented by silicification, sericitisation and The mean value was plotted on Z06Pb/Z04Pb versus
chloritisation. The Tyros beds have undergone 207Pb/ 204Pb diagram shown in Fig.2 indicating an
regional metamorphism, which is placed on the average age of 240 m.y. in agreement with that
boundary of Eocene to Oligocene. The deduced from paleontological evidences on
metamorphic temcr.erature is estimated to have metasediment rocks .(10) These data support the
reached 320- 350 C with a pressure of about 2 theory for the syngenetic origin of the ore .
Kb.(6) Furthermore, it is suggested that during the Molai
Despite the existence of two fold systems in ore genesis, volcanosedimentary conditions of
the ductile formations, the deformation in the deposition prevailed as evidenced by the following
more massive volcanics is manifested by the observations:
development of faults resulting in the displacement i) The ore consists of bedded sulphides that
of the ore bodies and the Variability in their commonly display sedimentary features such as
thickness. Nevertheless, the latter as well as the graded bedding or rhythmic layering (Fig. 3(a) . (6)
ore grade variability is mainly attributed to the it) Soft sediment deformation of the
initial conditions prevailing during deposition . semi-consolidated material resulted in slump folds
The mineralisation is developed along and intraformational faults.
several zones and subzones. The altered and iii) The presence and form of imbricated
mineralised zones exhibit a N-S trend and an mineralised breccias within the semi- massive fine
eastern dip of 55 to 75, Fig. 1. The main ore grained ore, indicates that the above breccias have
body of economic significance lies within the been transported from their original depositional
Western subzone whose reserves were given sites, commonly lying within a short distance, Fig.
above. The average thickness of this subzone is so,
3.8m. Three other orebodies, namely: the Eastern iv) The existence of framboidal and colloidal
zone, Western B-subzone and Western C-subzone textures within pyrite and sphalerite minerals as
of total reserves averaging 1.5 x 106 tons are well as the presence of linear banded pyrite
considered as not mineable due to their limited inclusions within sphalerite.(4,5)
thickness of 1.o-1.4m. O) Although further study is required for the final
The presence of different mineralised zones classification of the Molai ore body into a certain
in the strata is attributed to the repetition of the group of deposits, the features already described
ore genetic process and not to tectonic events. indicate that many similarities exist between the
The overall characteristics of the above zones examined ore body and the deposits of the
including the stratiform development, the type of Kuroko type. (11)
alteration and the ore mineralogy are very
similar. (6)
The spatial distribution of Zn, Ag, Cd, Fe ORE MINERALOGY
and Ba Obtained from the geostatistical analysis of
drill-core samples assays (programme BECES(7) A detailed mineralogical stydy was conducted on
is irregular and expressed by rich local representative drill-core and bulk ore samples
concentrations. It has to be noted that the originating from the main part of the ore body
correlation coefficient between Zn and Ag (0.71) is (Western zone). Microscopic studies indicated that
higher than the corresponding value between Pb the major sulphide minerals are: sphalerite (6-25
and Ag (0.41). There is a tendency for increased wt%), pyrite (3-7 wt%) and galena (1-3 wt%).(4)
Ba concentrations towards the South where the ore The minor and rare metallic minerals observed, as
body is observed to transcate. (8) well as the supergene and gangue minerals
indentified are listed in Table 1. Microanalysis
Chemical analysis of drill core samples
indicated that across the ore body the oxidation results for the characteristic metallic minerals are
degree of the mineralisation varied significantly. presented in Table II.
Average zinc oxidation in the Eastern and Western Zinc is mainly associated with sphalerite,
zone was estimated to 30-40% and 5-15%, however an amount of approximately 5-15% is
respectively. distributed within the supergene minerals,
Two main types of mineralisation can be smithsonite-monheimite. Galena is reported as the
distingulshed: major lead-bearing mineral, while small quantities
1) the disseminated and semi-massive of anglesite and cerrusite were also observed in
flne-gralned sulphides with both fragmental (lithic the samples examined.
and sulphide fragments randomly distributed in the The ore minerals mainly occur as
matrix) and bedded textural varieties. polycrystallic aggregates and rarely as unmixed
2) the coarse-grained massive ore, which is concentrations or independent crystals, Fig. 4(a) &
recrystallised and remobilised. The latter type is 4(b).


brecc ia.

o_ _ _100 200m
I ! ! ! ! = ~ I


Fig. 1 Geological map of the Molai deposit area

l ,~ - - ~-----
c-> tOJ 0 _-
~ 7SJl 'j_ ~ ~-- " pptf' cru. s~
bT~ ~~
. ~: :
-:-- 1
~__ -
.34 ---= . ' .- I -

I ~'---- . I I : ______
: '_ !- -T ~ __-
1- - , I I I
32 ....-; : -
~ -- , I I . 1- " ~ _ _
~ : I ; ._ ____

~, I ~ . ... , - -
I I j."'-- I I ;

,a. --,I
1 I -t~
l--r- I
)0 .............: I ..- I I
.v-, !.J--

: i '-r: : .-
/G I , . ....--
..-r i !OIAI
10 I I _ J' "
, I I 1 __
I t f'\\\~ I I _ f"
~~> t
lSI I -f
20 _
180 190 100

Fig. 2 : The mean value of the ratios 206Pb/204Pb, 207Pb/204Pb plotted

on the Amov diagram(9)

(~ ~)
Fig. 3 Macroscopic observations of the Molal sulphide ore:
(a) Stratiform mineralisation of sphalerite (light grey) and tuff.
(b) Massive ore, sphalerite (light grey) with imbricated partly
mineralised breccias.
Drill -core sample from the Western subzone. Scale: lcm-1.5cm

Mine ra ls of Mol ai ore body

A. Hypogene sulf ides-s ul fosalts -oxides mi neral s Mai n Minor Rare

Sphaleri te lnS *
Py rite FeS 2 *
Gal ena PbS *
Tetrahedr i t e-Tennant i t e Cu 12(As -Sb)4S1 3 *
Magneti te Fe304 *
Hematite Fe 203 *
Geo kronite Pb sSbAsS a *
Freibergite (CuAgFe)12Sb4S1 3 *
Chalcopyrite CuFeS 2 *
Marcas i t e FeS 2 *
Pyrrhot ite FeS *
Born ite CuSFeS4 *
Ena rgi te Cu 3As S4 *
Famatini te Cu 3Sb \ *
Arsenopy r i t e FeAsS *
Bou langer i te PbsSb4S 11 *
Sel igma nnite ( ?) Cu PbAsS 3( +Sb) *
Diaphorite Ag/b 2Sb3Sa *
Stephani te Ag sSbS4 *
Py rarg i ri te Ag 3SbS3 *
Lazarevic ite Cu 3(As,V)S 4 *

13. Supergene minerals

Cerussi te , Anglesite PbC0 3 ' PbS0 4 *
Smi t hsonite (monhe imi te) lnC03 ,(Fe,ln,Mn,Ca)C0 3 *
Geoth ite a-Fe 203 H 2O(FeH02) *
Chalcosite Cu 2S *
Covell ite CuS *

c. Gangue minerals
Quartz Si0 2 *
Ca 1ci te CaC03 *
Dolomite Ca( Mg ,Fe)(C03 )2 *
Siderite FeC0 3 *
Kaol in ite A1 203 2Si022H2O *
Il l ite KAI2(OH )2A1Si3(0.OH)10 *
Chlor ite Mgs (Al ,Fe)( OH)a (Al ,S;)4014 *
Ba rite BaS0 4 *

Microanalysis of Molal metallic minerals
% Tennantite
ZnS PbS FeS2 CU12(A.-Sb)4S13 (CuAgFc)12Sb4S13 Ag3Pb2Sb3Sg

Ag 0.02 0.4 23.9 23.8

Sb 29.3 11.2 26.0 27.4
Zn 62.9 65.7 0.1 2.2 0.04 1.1 0.02
Cu 0.1 37.0 44.7 24.3 0.2
Fe 3.1 0.3 46.8 4.2 0.4 1.1

S 34.1 33.7 14.1 53.2 25.0 32.2 24.0 19.4

Pb 0.2 0.4 87.3 0.2 0.1 29.2
As 0.9 10.5 0.2 0.01
Cd 0.5 0.3

Total 100.9 100.1 101.5 100.0 99.5 99.1 100.6 100.1

Sphalerite occurs in two types. The first one The minor and rare minerals such as
is Fe free and consists of grains and aggregates geokronite, seligmanite, bornite, famatinite
ranging from 30 to 300ilm in size. The second enargite ~d lazarevicite do not form intergrowth
one, less frequently observed, contains up to 3% textures with the other metallic minerals but they
Fe (Table II) and consists of finer grains ranging occur as individual crystals.
from 20 to 200ilm in size. As shown in Fig.4(c) The supergene minerals such as smithsonite
sphalerite grains contain numerous and variou~ monheimite and anglesite are often observed i~
inclusions - exsolutions covering their surface such cl?se assosiation with the base metal sulphides.
as: pyrite crystals from a few urn to 100~ in Zinc carbonates also occur as fine inclusions
size, pyrrhotite, arsenopyrite, hematite, (4-SIlffi) within the quartz and silicates.
chalcopyrite, galena, anglesite tetrahedrite - . Sil~er was located within the following
mineralogical phases: in tetrahedrite-tennantite
freibergite - tennantite. In some cases these serie~, mainly in freibergite and in the three Ag-Sb
inclusions exhibit a linear-banded form into beanng. mmeral~, diaphorite, pyrargirite and
sphalerite, Fig. 4(d), implying that during the ore stephanite. Freibergite occurs in the form of
genesis sedimentary conditions of deposition inclusions or intergrowths mainly associated with
prevailed. Also, in some sphalerite crystals the sphalerite while th~ other three Ag-Sb bearing
rare phenomenon of Fe exsolution in the form of minerals are exclusively found as microinclusions
magnetite can be observed, as illustrated in Fig. within galena.
4(e). It is anticipated that inclusions of iron bearing The main characteristics of the Molai ore
minerals within sphalerite grains would adversely are the fine-grained texture, the intergrowths often
affect the kinetics of zinc separation by observed between sphalerite - galena, galena -
flotation. (12,13)
Furthermore, on the boundaries of pynte and the numerous inclusions present
particularly in sphalerite grains. The above
sphalerite grains galena was observed in a form mineralogical features are expected to adversely
that w~ difficult to liberate and would potentially affect the flotation response of lead and zinc
result ill the early activation of zinc in the lead minerals with conventional flotation reagents.
circuit. (14) It has to be noted that sphalerite
mineral grains contain approximately 0 5%
Cd. (4,5,14) ,
Galena crystals are smaller in size ranging LIBERATION STUDY
~rom S .to 1OOIlffi, Fig.4(f). They contain very fine
in cl us i o n s Sllm) of anglesite The grain size of the metallic minerals
tetrahedrite-tennantite, boulangerite, and silve~ significantly affect the selection of the subsequest
minerals. When galena is forming aggregates benefication scheme. It is known that for selective
pyrite is often included in a fine grained flotation as mineral grains become finer liberation
intergrowth texture. Microanalysis of galena grains and selectivity becomes less effective. '
indicated only traces of Ag. A detailed liberation study was conducted on
Pyrite occurs as grains and aggregates a composite bulk sample from the Molai mine with
ranging from 30 to 200ilm in size, and as the ~ to .identify the optimum grinding size for
th~ differential and bulk flotation of lead and zinc
inclusions of a few microns within sphalerite and
IDlne~als . The average chemical analysis of the
galena. In certain mineral specimens pyrite was examined sample IS shown in Table III. Although
also observed in the form of framboids.
Tetrahedrite, freibergite and tennantite are the base metal assays are lower than the
enclosed or intergrown mostly with sphalerite and previously examined drill-core samples the
to a lesser extend with galena. average mineralogical analysis of the bulk' sample
is reported to be the same.{4,5}




Fig. 4: Microscopic observations of mineral samples. Reflected light, / /

Nicols. (a) Polycrystallic aggregates: sphalerite(sph), galenatga), pyrite(py),
tetrahedriteuetl.Ib) Aggregates and independent crystals of sphalerite,
galena, pyrite into a smithsonite matrixtsrnt). (c) Pyrite, galena inclusions
within sphalerite crystal.




Fig. 4: Microscopic observations of mineral samples. Reflected light, / /

Nicols. (d) Linear banded pyrite inclusions within sphalerite crystal.
(e) Magnetite exsolution(mag) within sphalerite.
Oil immersion. (f) Fine grained galena-sphalerite intergrowths.

Chemical analyses of the Molal composite bulk sample

Zn 6.3
~ 1~
Fe ~7
S 5.9
S(S04) 0.6
Ag (ppm) 31.3
Cd (ppm) 332.0
Si0 2 52.0
~03 12.0
Kz0 4.3
N~O 0.7
cao 2.4
MgO 1.7
CO2 3.5

The study was implemented with an Image On the other hand for bulk flotation the
Analyser Cambridge Quantimet 900, coupled with a ~be~a~ion size of sphalerite estimated to 701!m, is
Scanning Electron Microscope (S.E.M), Cambridge significantly coarser than the corresponding values
Stereoscan 250. The procedure followed is for zinc carbonates, galena and anglesite.
presented in detail elsewhere.(S) The majority of zinc carbonates is liberated
It has to be noted that the size distribution at 201J.ffi, however at this mesh of grind a fraction
obtained does not itself represent the actual is still bound in the form of inclusions within
mineral grain size, but the smallest dimension to silicates. (S) Galena and anglesite were the most
which the ore must be ground in order to liberate difficult to liberate, the reason being their fine
the valuable minerals from the gangue material for dissemination not only within zinc bearing minerals
bulk flotation, or to extract monomineral grains but within pyrite, quartz and other silicates.
for differential flotation. For the purposes of this Similar conclusions were reached after heavy
study, galena (PbS) and anglesite (PbS0 4) were liquid tests on drill-core samples from the same
part of the ore body. (IS) Tests conducted on
defined as the lead bearing minerals, while particle size fractions ranging from -37+34I!m to
sphalerite (ZnS) and zinc carbonate series of -13+ 1OI!m indicated that lead losses in the light
smithsonite (ZnC0 3) monheimite fraction, varying from 5.0% to 21.0% were
(Fe,Zn,Mn,Ca)C0 3 were considered to be the zinc systematically higher than zinc ranging between
carrier minerals. Above phases are also presented 3.0% and 15.0%. For both zinc and lead significant
in the backscattered electron images seen in portion of the metal values reporting in the tails
Fig.5(a) and so, were in the oxidised form.OS)
From analytical results in Table IV it is Thus, if bulk flotation is applied to the
deduced that the Molai ore is classified in the fine treatment of the Molai ore a grinding size of
grained minerals group. (12) To liberate the majority 60l!m would be adequate, in order to liberate
of sphalerite and galena as monomineral grains, sphalerite and a considerable fraction of lead
the required grinding size is 261!m and 121!m minerals from pyrite and gangue material. In that

liberation size of major Pb-ZD minerals
of the Molal ore
Pb-minerals Sphalerite Oxidised Zn Galena and
liberation and sphalerite minerals Anglesite
Flotation 26* 12/2
Bulk Flotation 55 70** 20 7

* monomineral
** with metallic inclusions


Lead minerals



Sphalerite + Lead Minerals

Fig. 5: Backscattered electron images with S.E.M.

(a) Inclusions of lead bearing minerals within sphalerite matrix
(b) Zonation within zinc carbonate minerals: Smithsonite (light grey) on
monheimite grain boundaries (dark grey).

case, a percentage of valuable metals is not grinding time and size analysis, reagents
anticipated to report in the tails. combination and dosage, nH, conditioning and
Above liberation study has also shown that due to flotation time was evaluated~15
the fine intergrowths between lead and zinc Flowsheet derived from the above tests is seen in
minerals and the numerous inclusions observed Fig. 6. Fine grinding, 80% passing 40~m, was
within sphalerite finer grinding of approximately applied to liberate galena and sphalerite. Increased
10~m is required for the separation of lead reagent consumption was observed for the
minerals. The effect of liberation on the overall depression and subsequent activation of sphalerit~
performance of the subsequent beneficiation as compared to published data on similar ores . O
process is discussed in detail in the following Metallurgical results of batch tests shown in Table
section. V indicate that overall Zn and Pb recovery was of
the order of 80% and 78% respectively. Based on
laboratory test results marketable lead and zinc
BENEFICIATION PROCESS products can be recovered with subsequent cleaning
The beneficiation study of the Molal ore was The finely disseminated intergrowth of galena,
initially conducted on characteristic drill-core sphalerite and pyrite combined with the presence of
samples originating from different parts of the ore oxidised minerals were the main reasons for
body. (14 ,IS) Above data formed the basis for an reduced selectivity of separation and low metal
extended research programme, being currently in recoveries. As already described in the previous
progress, with the aim to develop an economic sections at this fine mesh of grind galena is not
process route for the treatment of the Molal fully liberated. More than ten per cent of lead
sulphide minerals. (4,S) Testwork results presented in reports in the sphalerite concentrate in the form
this section refer to composite drill-core samples of locked particles (middlings). Fine intergrowth
from the main ore body (Western zone). OS) The textures also resulted in valuable metal losses in
average chemical analysis of the examined sample the tails . High zinc assay in the galena concentrate
is given in Table V. was mainly attributed to early sphalerite activaction
Due to the mineralogical features already decribed, caused by the galena and anglesite slimes covering
separation by flotation was considered as the most its surface. Mineralogical observation of the galena
appropriate method for the beneficiation of the cleaner concentrate further support the above. (is)
Molal ore. Preconcentration for the upgrading of The presence of oxidised minerals such as
the flotation feed material was also examined. smithsonite and anglesite adversely affected the
Laboratory and Pilot Plant testwork have indicated overall flotation recoveries. Despite the strong
that separation with heavy media cyclone may be sulphidisation applied, more than 50% of the
applied in order to scalp 10-25% of run-of-mine contained oxidised lead and 70% of oxidised zinc
ore (r.o.m) with Pb and Zn losses in the tailing report in the tails, Table V.
ranging from 2-10% to 1-5% respectively. OS) Thus, Silver was equally distributed among the two
preconcentration has to be considered as a process rougher concentrates due to the association of Ag
option in view of the selected mining method and bearing minerals with both galena and sphalerite.
subsequent dilution of the r.o.m ore. To further illustrate the above, Ag distribution
versus the Pb and Zn recoveries in the galena and
Differential Flotation sphalerite cleaner concentrates is presented in
Fig.7. The silver recovered in the zinc concentrate
To investigate the sulphide mineral response under does not linearly correspond with the Pb contained
various flotation conditions, batch laboratory tests in this product but follows zinc distribution as
were initially conducted in order to produce evidenced by mineralogical analysis, indicating the
separate lead and zinc concentrates. During the presence of silver bearing minerals such as
testwork the effect of process variables such as freibergite within sphalerite grains.


Metallurgical balance of differential flotation tests

on drill-core samples from the Molal ore

Concentrate 41.9 1.2 19.7 0.5 515 60 11 4 2 27
PbS cleaner
Tails 8.1 1.1 26.2 0.8 188 18 16 8 4 15
Zns cleaner
Concentrate 1.0 0.2 59.3* 0.5 123 9 13 74 12 39
Zns cleaner
Tails 1.0 0.6 16.6 1.1 57 3 10 7 8 6
Tails 0.2 0.2 1.1 0.7 8 10 50 7 74 13

Head 1.7 0.3 11.6 0.7 46 100 100 100 100 100
*Cd: 0,35%

: r,SO : 2. 5k/t
4 C,,0 fH 11
WI : r. ~
CIJSO : 2 . S ~ / t
II.. SO : 0. 4" 4
2 3 KEX :2119/ t
II I: ~ X : se "
~ nIBC: 20 "
~ ~

UU -. -.

rl~~ation -t

UU1 -t --+

CI t
rlo~~tion - +



1~D ~D
r" ' ~~~;~'D
'(" "1 r'i :~~~~"D'T"'-1
I I L _J
, CI t ,

3D : ,
Zn cone .
L _J

rlJ cone .

Fig. 6 Flowsheet for differential flotation of Molal are.


~o r- .
PbS c 1 ane r cone

-- - - ~ _ ... -


GO 60
'"u>o ~ .~ I nS cl ea n r cone .
~ 40 --- /-- t 40 ---


.. 20

I 20
I nS cl ~ane r cone

60 80 100
20 40 60 80 100

I Ag recovery

Fig. 7: Silver versus lead and zinc distribution in the galena

and sphalerite cleaner concentrates.

Bulk Flotation To improve the grade of above concentrate
various process schemes might be employed
To mitigate the adverse effect of fine grinding on including regrinding followed by subsequent cleaning
the overall flotation performance, an alternative stages to separate pyrite and galena. (14) The
process scheme was also examined in order to optimum beneficiation scheme has to be selected in
evaluate the recoveries attained when a bulk Pb-Zn view of the subsequent metallurgical processing
concentrate is produced from the Molal ore. With options to obtain the maximum economic benefit
the flow-sheet applied, Fig. 8, significant grinding from the ore resource. As already stated an
energy and reagent consumption savings were extended research project is currently in progress
observed. for the development of the Molal ore process
From the corresponding metallurgical results shown route. (4,5) Technical and economical data thus
in Table VI, it is deduced that with bulk flotation obtained will be incorporated in the fmal feasibility
higher overall base metal recoveries were study for the potential exploitation of the Molal
achieved. Due to the high Zn grade of the bulk sulphide deposit .
concentrate, this product could be potentialy
considered as a marketable zinc concentrate. (17)

Metallurgical balance of bulk flotation tests
on drill-core samples from the Molal ore


Concentrate 7.1 0.5 48.0 0.1 180 76 32 75 6 70

Cleaner Tails 1.9 0.5 18.1 0.4 70 16 29 21 14 21
Tails 0.2 0.1 0.7 0.4 6 8 39 4 80 9
Head 1.7 0.3 11.6 0.7 46 100 100 100 100 100
*Cd: 0,3%

r- --- --i--~- -.-.--1


'CI t

L. I _ _ . JI

Bulk cone.

Fig. 8: Flowsheet for bulk flotation of Molai ore .

CONCLUSIONS and anglesite slimes. With bulk flotation lead, zinc
and silver recoveries of the order of 90% were
1. The Molal polymetallic sulphide deposit achieved with significant grinding energy and
is situated in S.E.Peloponnese, Greece. To date reagents savings. Due to the high zinc grade of
evaluation of drilling campaign results indicated the bulk concentrate this product could be
mineable ore reserves of 2.3 x 106 tons with an potentially considered as a marketable sphalerite
average estimated grade of Zn: 9.0%, Pb: 1.8% concentrate.
and Ag: 50 g/t. Presently an exploratory gallery is
excavated in the main ore body to investigate the 7. The optimum beneficiation scheme for
optimal mining method. the Molal SUlphide ore has to be selected in view
of the subsequent metallurgical processing options
2. The Molal deposit belongs to the with the aim to maximise the economic benefit
volcanic associated massive SUlphides. The host from the ore resource. An extended research
rocks are mainly volcanics, such as tuffs, tuffites, programme is being currently in progress for the
massive arphanitic and porphyritic lavas. The ore development of the above process route.
predominantly comprise stratiform accumulation of
base metal sulphide minerals which have been
formed in a submarine volcanosedimentary
environment by precipitation of hydrothermal ACKNOWLEDGMENTS
fluids. Although the final classification of this
deposit into a specific group is a matter of further The authors would like to acknowledge the
investigation, the Molal ore deposit presents financial support of the European Community,
many similarities to the Kuroko type deposits.(11) Contract No. MA1M-0051-C(AM), for part of the
work presented in this article. They would also like
3. The genetic conditions have affected the to thank Mrs. R. Vasmoulakls for typing this
mineralogical characteristics of the ore (fine manuscript.
grained textures, complex intergrowths between ore
minerals, numerous inclusions particularly in
sphalerite), the geometry of the ore bodies REFERENCES
(thickness and average grade variation) and the
rock mechanical properties . These features would
SUbsequently affect the mining and mineral 1. IGME: "Concise report on the Prefeasibility Study
processing methods to be employed for the of the Molal massive sulphide deposit in Laconia".
potential development of the Molal deposit. Institute of Geological and Mineralogical
Exploration, Athens, 1985.
4. Sphalerite is the main zinc bearing
mineral in the Molal ore, although depending on 2. METBA: "Optimisation of the exploitation of thin
the oxidation degree of the sulphide mineralisation, vein polymetallic sulphide deposits through
a fraction of contained zinc is distributed within the mathematical modelling and rock mechanics, an
supergene minerals, smithsonite - monheimite. The application to the Molal mine" CEC contract No.
occurrence of these two minerals would result in MA1M-0025-C(TT) 2nd Progress Report, Athens,
lower overall recoveries due to their poor December 1989.
response to conventional flotation reagents. Galena
is the main Pb bearing mineral. Anglesite and 3. METBA: "Molal Base metal massive sulphides"
cerrusite are also observed occasionally, on the CEC Project GRE 02109/002, 2nd Progress
boundaries of galena crystals. Silver is associated Report, Athens, April 1990.
with the minerals of the tetrahedrite-tennantlte
group, mainly with freibergite. Furthermore, part 4. IGME, BRGM, ITMI: A process design study for

of silver is also distributed within the Ag bearing a finely disseminated partially oxidized complex
minerals i.e. stephanite, diaphorite, pyrargyrite, sulphide ore from the Molal area Southern
exclusively reported as microinclusions in galena Greece" CEC Contract No. MA1M-002S-C(TT).
crystals. 2nd Progress Report, Athens, May 1989.
5. As a result of the close association S. IGME, BRGM, IMM: A process design study

between lead and zinc minerals and the numerous for a finely disseminated partially oxidized complex
microinclusions observed, the liberation sizes sulphide ore from the Molai area Southern
determined ranged from 281lm for sphalerite to Greece" CEC Contract No. MA1M-0025-C(TT).
12J.tffi for galena and 2J.tffi for anglesite. However 3rd Progress Report, Athens, November 1989.
valuable minerals would be liberated from gangue
material at coarser grinding sizes of approximately 6. Angelopoulos C. and Constantinides, D.C.: "The
60llm. Molai Zn-Pb-Ag
M, deposit, Lakonia,
S.E.Peloponnese Bull Geol.Soc. Athens, Greece,
6. For the preliminary beneficiation study Vol.XXI2, 1988, p.305-320.
of the Molai sulphide ore, laboratory differential
and bulk flotation tests were conducted on 7. Christakos, G.: A computer program for the

composite drill core samples. Testwork results analysis and estimation of spatial functlons", Intern.
indicated that with differential flotation, after fine Report, Civil Eng.Dpt, MIT< Boston, Mass. 1980.
grinding and increased reagents consumption,
marketable lead and zinc concentrates could be 8. Angelopoulos, C. & Constantides, D.C: MThe
produced at recoveries lower than 80%. High zinc Molai zinc-silver-lead deposit, Lakonia/Greece",
assay in the galena concentrate was mainly Intern. South Eur. Symp. on Expl. Geoch., Field
attributed to early sphalerite activation by galena trips guide book, Athens 1986, pp.61-67.

9. Amov, G.B : "Evolution of uranogenic and
thorogenic lead. A dynamic model of continuous
isotopic evolution". Earth and Planet Sci. Letters
~ Elsevier, Amsterdam, 1983, pp 61-74.

10. Brauer, R. : "Das Praneugen im Raum Molai

Talanta/S.D.-Lakonien (Peloponnes, Griechenland)"
Inang .- Dissert. zur Erlangung des Doctorgrades.
Frankurt am, 1982.
11. Ohmoto, H and Skinner, J.P .: "The Kuroko and
Related Volcanogenic Massive Sulphide depos its".
Econ. Geol. Monogr. 5, 1983.
12. Norrgran, D .A. and Armstrong, R.O. :
"Developing a selective flotation concentration
technique for polymetallic ores" in Complex
Sulph ides. eds. A.D .Zunkel, R.S .Boorman,
A.E.Morris, R.J.Wesely, TMS, 1985, pp . 37-54.
13. Finkelstein, N.P. Allison , S.A.: "The Chemistry
of Activation, Deactivation and Depression in the
Flotation of Zinc Sulphide. A review". In Flotation,
ed. M.G.Furestenau, Vol. 1, New York, 1976, pp.
14. Ingerttila, K. Kaukkanen J., Hintikka, V.: "On
Mineralogy and Flotation test of Molai ore",
Technical Research Center of Finland, 1986.
15. Grossou-Valta , M. et al.: "Discussion on the
possible treatment of Molal mixed sulphide ore
after experimental investigation of typical samples".
Metallurgical Research Reports No.37, Institute of
Geology and Mineral Exploration, Athens , 1984.
16. Bulatovic, S.M., Wyslouzil, D.M.: "Selection of
reagent scheme to treat massive sulphide ores" in
Complex Sulphides ed. D.Zunkel, R.S. Boorman,
A.E .Morries, R.J.Wesely, TMS. 1985, pp. 101-137.
17. Cases, J.M. : "Finely disseminated complex
sulphide ores" in Complex SUlphide Ores, ed.
M.J.Jones, IMM Conference Proceedings 1980, pp

Studies of mineral liberation performance in sulphide
comminution circuits
D.M. Weedon
T.J. Napier-Munn
julius Kruttschnitt Mineral Research Centre, University of Queensland, Indooroopilly, Queensland,
CL . Evans
Mount Isa Mines Ltd., Mount lse, Queensland, Australia (formerly, julius Kruttschnitt Mineral Research

This paper describes studies of two relatively coarse- The principal purpose of any grinding circuit in a
grained sulphide ore grinding circuits (one copper and one sulphide mineral processing plant is to break the ore to a
lead/zinc), carried out to quantify the behaviour of the size sufficiently small to release or liberate the value and
liberated minerals. The work formed part of a larger study gangue mineral grains, so as to permit the valuable
whose objective is the modelling of mineral liberation for component(s) to be recovered by downstream
comminution circuit simulation. concentration processes, usually flotation. Penalties exist
for undergrinding (inadequate liberation) and overgrinding
The circuits were sampled and mass balances (poor flotation response and excessive power costs) .
calculated. The mineralogy of the solids in each process There is therefore considerable incentive to design or tune
stream was quantitatively determined using QEM*SEM, the grinding circuit to provide adequate liberation without
from which mass flows of liberated mineral were overgrinding. Such optimisation is difficult in practice
calculated. Samples of circuit feed were also broken in a because:
laboratory pendulum apparatus, and the degree of
liberation measured using QEM*SEM. 1. The assessment of the extent of liberation in
different parts of the circuit is time-consuming
The plant results confirmed the strong effects of and expensive (to the extent that it is not a
classification on liberation performance, and the routine function in most concentrators).
differences in the behaviour of different classifier types.
The effects are usually negative, as the denser liberated 2. The modelling of liberation in comminution is
values tend to concentrate in the classifier underflow not yet sufficiently advanced to permit
product, resulting in large circulating loads of liberated optimisation by simulation, although some
mineral which are reground unnecessarily, whereas the progress has been made in recent years (1, 2).
liberated gangue is preferentially selected for the
downstream separation step (flotation in the case 3. The distribution of liberated species around the
considered here). The data also demonstrated that the net circuit, and the associated circulating loads, are
liberation in secondary grinding can be quite small, and strongly influenced by the characteristics of the
that the extent of liberation in products from the coarse comminution device itself, its associated
(primary) grinding stage can be substantial. classifier, and the properties of the different
mineral phases. This introduces complex
The evidence in one case suggested that the interactions and renders an analysis of liberation
proportion of liberation in a particular size interval is performancedifficult
independent of comminution history, and that the
propensity of a mineral to be liberated can be assessed The Julius Kruttschnitt Mineral Research Centre is
using single particle breakage tests and QEM*SEM currently investigating ways of incorporating liberation
analysis. information into established mathematical models of
comminution processes, to permit more comprehensive
These observations imply the need to consider aspects design and optimisation of grinding circuits. In order to
such as coarse flotation early in the grinding process, provide a base level of data for this programme, and to
positive sizing devices to close the grinding circuit (eg advance the understanding of the liberation process in
screens rather than hydraulic classifiers), and whether conventional sulphide concentrators, the grinding circuits
conventional multiple grinding stages are always justified
in practice.

in two well known such concentrators in Australia were summary of the mass balanced data is given for each
surveyed to determine their liberation performance. This circuit on the circuit flowsheets (Figures 1-4).
paper presents a preliminary analysis of the results of this
exercise, and draws conclusions relating to the incidence Detennination of Mineral Liberation by OEM*SEM
and behaviour of liberated minerals in the circuits.
Sub-samples of the solids in all process streams were
EXPERIMENTAL DETAILS sized and submitted to the CSIRO Division of Mineral
and Process Engineering for analysis by QEM*SEM, to
Circuits determine the size-by-size incidence of composite and
liberated particles in each stream.
Two plants were surveyed. The first was the ZC
Mine Concentrator of Pasminco Mining at Broken Hill, "QEM*SEM" is an acronym for "Quantitative
formerly the Zinc Corporation, in which the valuable Evaluation of Materials (Minerals) using Scanning
minerals are galena and sphalerite; it was surveyed in Electron Microscopy". It is an automated image analysis
September 1987. The ZC plant contained three parallel system developed by the CSIRO and capable of scanning
grinding circuits, all of which were sampled . Circuits 1 large numbers of particles to determine their mineralogical
and 2 each incorporated an open circuit rod mill followed composition. It analyses scanning electron microscope
by a closed circuit ball mill, lead rougher flotation, and a images of samples using back scattered electrons and X-
second closed circuit ball mill treating the lead rougher ray emissions, and is a useful tool in the estimation of
tailing. In Circuit 1 the ball mill was closed by a rake liberation in mineral processing circuits , its results being
classifier, whereas in Circuit 2 classification was by quantitative and reproducible (3).
hydrocyclones. Circuit 3 was configured similarly to
Circuit 2, but without the lead roughing and secondary QEM*SEM measures mineral samples in polished
ball mill stages; the feedrate to this circuit was also about sections . A size fraction from a circuit stream sample is
half that of the other two. Flowsheets of the three mounted in an epoxy resin briquette before cutting flat and
circuits are shown in Figures 1 - 3. polishing with diamond paste. Great care is taken to
present the particles in the resin in a random manner to
The second plant was the No.4 Copper Concentrator ensure no bias exists. The briquette is mounted in a
of Mount Isa Mines, in which the principal valuable motorised stage which is computer controlled in the x - y
mineral is chalcopyrite; it was surveyed in October 1986. plane, and the surface is then systematically scanned by an
This concentrator contained two parallel grinding lines, electron beam . The back scattered electrons and emitted
one of which was sampled. An open circuit rod mill was X-rays are recorded and interpreted to provide details of
followed by two stages of closed circuit ball milling, the particle numbers, mineral phases and associations with
classifiers being hydrocyclones in each case. The other minerals .
flowsheet is shown in Figure 4.
Two modes of analysis were used by QEM* SEM in
Surveys and Sample Analysis this work:

Each stream of each circuit was sampled 1. Bulk Mineralogical analysis.

incrementally over a period of about 1-2 hours while the 2. Particle Analysis
circuit was as far as possible in steady state, as determined
by the stability of plant instrument readings. A variety of The first technique involves linear scans of a large
plant instrumentation was available, but for the purposes number of particles in polished section. The X-ray
of this exercise only the mass flow of ore to the rod mill spectrum and back scattered electron intensity for each
in each circuit was determined instrumentally. Fresh measurement point are collected. Each mineral phase is
water additions were also measured by plant identified and the volume of mineral present is estimated
instrumentation. All other mass flows were calculated by statistically. The weight percent is calculated by
mass balancing. multiplying the volume fraction by the density of each
phase and normalising the results. Elemental
The samples were processed to determine solids size comparisons are calculated using the weight percent of
distribution (by sieving to 38 urn), solids concentration each phase and the known mineral stoichiometry, which
(by drying and weighing), and metal assay. Sub-samples is supplied to the system prior to analysis (4).
were taken for liberation analysis using QEM*SEM (see
below); these were pre-screened into individual size The second technique gives a quantitative measure of
fractions using conventional woven wire sieves and the degree of liberation and intergrowth of individual
microsieves. particles for up to six nominated phases or groupings of
phases. All particles measured are classified into
The mass flow, solids concentration, assay and percentile categories which represent the modal abundance
particle size distribution data were then mass balanced to of the reference material in each particle. Also shown are
produce self-consistent estimates of flows and size the number and percentage of particles in each of the ten
distributions in each stream, subjeet to the constraint that liberation classes, the area percent of the particles in that
input = output, both at each process node and over the class and the grain size of the phase of interest.
whole circuit. In nearly every case, the data adjustments
required to meet this criterion were small, confirming that There are limitations to the size of particle which can
the sampling and analysis procedures were satisfactory. A be measured by QEM*SEM. The minimum size is
10-20 urn because of difficulties in obtaining
138.9 64.9
68.1 257

8.9 tph

water OIF

30.9 tph
water 83.4
138.9 4.7
96.7 12006 464


2 .44m-dia

20 .4 tph 126.7 27.44
82.2 381
138.9 35.6
79.6 778

Pb Rougher

.---__~~ 132.6 59.9
6.34 5.0
68.9 210
56 75

2.44m -dia

132.6 59.9
68.9 300 1.:l86.8
319.4 106
75.1 259

+ = sample points

Sol ids Water
tph tph
0/0 P
Solids 80 um
Figure1. Flowsheet of ZC Mine grinding circuit N01 showing mass
balance data and sample points. 137
143.9 82.8
63.5 217

5.6 tph
water OIF

3.05m 165.6
143.9 3.2m
85.6 dia i:;~;!J,: : ~:.t:.:i:.:.i: : ;'i:: i 472

U/ F

2.44m-d i r F = = ;i;iJ - -r-

23 tph

72.6 2
Pb rougher

6.2 4.8 ....-._---1. 137.7 78
56.4 75 63.8 213

137.7 78 265 .1
63.8 266 81.5
286 402.8 138
74.5 277

1- sample points

Solids Water
tph tph

% P
Figure 2. Flowsheet of ZC Mine grinding circuit N02 showing mass Solids 80 um
balance data and sample points.
60.3 82.8
42.1 204

20 tph
water OIF

29.7 tph
2.74m r-
60.3 2.2
96.3 12063


30.9 tph 223 112
66.5 326

60.3 31.9
<, 283.3 174.8
65.4 413 61.8 325

So lids Water
tph tph
0/0 P
solids 80 urn


Figure 3. Flowsheet of ZC Mine grinding circuit No3 showing mass

balance data and sample points.

390 173.1
69.3 43.7


water 6.1m
I 5.0m-d ia .
, 5 .55m

I ..:.:.:.:.:.:.:.:.:.:.:.:.,.:.:.: .

81.1 117.6
99 .7tph


..--....-~ 390 399.4
49.4 67.1

6.1 m

5.0m-d ia

1445 720.4
water 226.8 66.7 36
tp h

-I- sample points

So lids Water
tph tph

Figure 4 Flowsheet of Mt Isa Mines No4 Copper concentrator grinding clcult showing mass
balance data and sample points.
unambiguous X-ray signals at finer resolutions. The 2. The substantial extent to which liberated mineral
maximum size is about 300 um due to time and cost is returned by the classifiers for regrinding, and
constraints in measuring the many particles necessary to the variation in the performance of different
obtain a statistically significant sample. For the present classifiers.
work it has been arbitrarily assumed that all the minerals
in the fine sizes (-28 um in this case) are liberated and all 3. The different behaviour of the individual mineral
particles above 300 um are unliberated. The former phases in the circuit.
assumption is reasonable (by extrapolation from coarser
sizes), but the latter is open to some doubt, and the mass 4. The very limited amount of additional apparent
flows of liberated mineral may therefore be underestimated liberation achieved in some of the second stage
in streams containing significant proportions of +300 um grinding circuits.
These aspects are discussed in more detail below.
Stereological bias is an unavoidable problem in the
current state of image analysis technology. It arises due Production of liberated mineral
to the two dimensional representation of three
dimensional particles. In simple terms, sectioning the The incidence of liberated mineral in the early stages
particle will result in some composites being seen by of grinding was relatively high in both sites. For
measurement systems such as QEM*SEM (incorrectly) as example at ZC the rod mill discharge contained a mass
monornineralic, whereas truly liberated particles will flow of liberated galena and sphalerite which was at least
always be seen (correctly) as monomineralic. As a result, 60-70% of the amount of liberated mineral reporting to
image analysis by measurement systems such as lead roughing (i.e, after one further stage of grinding). At
QEM*SEM will overestimate liberation by a proportion MIM, the rod mill produced nearly 50% of the tonnage of
inversely related to the apparent degree of liberation. liberated chalcopyrite reporting to the final circuit product
(i.e, after two further grinding stages). This incidence of
There is also the question of defining a liberated liberated mineral after primary grinding suggests that in
particle. Even though the proportion of truly liberated some cases it may be appropriate to consider a coarse
particles (those which are monomineralic) in a particular particle concentration stage early in the comminution
fraction might be quite small, many of the composites process. The benefits would be both in improved mineral
will be sufficiently liberated for their recovery kinetics not recovery by minimising overgrinding, and in a reduction
to be limited by the degree of liberation. In this work, a in downstream milling tonnages, by reduction in the
liberated particle was arbitrarily defined as one in which considerable circulating loads of liberated values occurring
the plane section surface area was composed of at least in most circuits (see below).
80% of the phase of interest. Barbery (5) has estimated
that the error in the estimate of degree of liberation caused The creation of additional liberated mineral in the
by stereological bias inherent in such a definition might final grinding stages was variable and mineral-dependant.
be up to 15% absolute. In the MIM case, it was interesting to note that the
secondary ball mill (drawing about 2.5 MW) increased the
Once the proportion of "liberated" particles in a tonnage of apparently liberated chalcopyrite by less than
particular stream/size interval had been determined, the 2%, which is negligible in the context of the prevailing
mass flow of liberated mineral in each size/stream was experimental error. This conclusion should however be
calculated by multiplying this proportion by the total viewed in the context of the stereological bias mentioned
solids mass flow in that size interval (estimated from the earlier. Espinosa-Gomez (12) has estimated (after
mass balancing step), and then re-balancing the mineral applying the 20/30 correction to the 90-100% liberation
flows . class) that the increase in liberation may actually approach
12%, and this was supported by laboratory flotation tests
DISCUSSION OF RESULTS which suggested that the flotation of primary cyclone
overflow operated on a lower grade recovery curve than
Distribution of liberated mineral in the circuits that of the secondary cyclone overflow. This illustrates
the difficulty of interpreting liberation data of this kind.
The production and behaviour of liberated mineral
becomes very apparent when the mass flows are tracked For the ZC Circuit I, the tonnage of liberated galena
through the circuit. Figures 5-8 map the size-by-size apparently declined slightly after regrinding of the lead
mass flows of four liberated phases through the ZC Nos . rougher tailing. This is an artefact of the data due to
1-3 Circuits, and Figure 9 presents similar data for MIM. uncertainties in the incidence of (unmeasured) + 212 um
The plots are constructed such that material flows through and -28 urn mineral and the vagaries of the definition of
the circuit approximately from left to right. The more "liberation" mentioned earlier. However the additional
obvious features of such plots include the following: liberation achieved, if any, was clearly small in this case
also (though it should be stated that the sphalerite
1. The relatively high degree of apparent liberation liberation was substantial). In some situations, therefore,
achieved early in the comminution process, as there is justification in considering whether the expense of
evidenced by the tonnages of liberated mineral in multiple grinding stages is warranted in view of the
rod mill discharge compared with final classifier limited additional liberation sometimes achieved.
The relative behaviour of the different sizes of
liberated mineral in grinding, as observed in Figures 5-9,
Figure 5a. GALENA Nol Circuit

35.00 .----~---_._--___,---__r_---;_-

30.00 -+----+----+-----1- ---+----I-----+-------i

25.00 +----+- - - ;-- - - ; - ---+-- - - 1 - - - - + ----;

20.00 +----+---;----;----+----1----+----;
15.00 +----+---;----;----+----1----+----;
10.00 +-----::l4~~~~-.....~--_+_--+--+_-_l

0.00 f!!~~~!

Figure 5b. GALENA No2 Circuit

35.00 r------r---r----,-----,----,---~--__.

30.00 +----{r:"..;





Figure Sc. GALENA No3 Circuit

35.00 r-----..,.-----r-------,

30.00 I----+----+-----l
25.00 I----+----+-----l o 0-28

20.00 t-----.~;;;;::;:;;;:k_ --_l Gl 28-38
Ell 38-53
15.00 + - - --fl
IlllI 53-106


Figure 5. The tph of liberated galena flowing through

circuits 1, 2 and 3 at ZC Mine.

PCLASS U/F =primary rake classifier underflow .
PBMD =primary ball mill discharge.
SCYCO/F =secondary cyclone overflow .
RMD =rod mill discharge.

Figure 6a. PYRITE Nol Circuit

14.00 ,-------,c__---,---....- - - , - - - - - - - , - - - - , - - - ,

12.00 +-------if----+----+----+-------i----+----t
10.00 +-------if----+----+----+-------i----+----t
8.00 +---~f-----+---~c__--+--____,,,,
6.00 +-----+----+---7"
4 .00

RMD PBMD PClass PC lass Pbtail SBMD

Figure 6b. PYRITE No2 Circuit

14.00 r---~---.__--___,r---_._---..---_._--_____.

12.00 +----t---+----II-- - +---+--"'7""" 'r-----t

10.00 +-----+---+- - -1----+----,1'

8.00 +---/


0 .00

Figure 6c. PYRITE No3 Circuit



10.00 0 0-28

8.00 0 28-38
6.00 Gl 3853
IlIlI 53-106
B 106-212wn


Figure 6. The tph of lib erated pyrite flowing through

circuits 1, 2 and 3 at ZC Mines.

PCLA SS UfF =prim ar y rake cla ss if ier under fl ow.
PBMD =pri ma ry ball mill di schar ge.
SCYCOfF =se cond ary cyc lone over fl ow.
RMD =rod mill disch arge .

Figure 7a. SPHALERITE Nol Circuit



15 .00


o.oo _ _
U/F Off

Figure 7b. SPHALERITE No2 Circuit

25.00 ~--~-----.------,r-

20.00 +----/P.~--t_--_t_--___1r_____:;~i


5.00 _ _ ~==.

Figure 7c. SPHALERITE No3 Circuit


0 0-28
0 28 -38
I]] 38-53
II1II 53-106
5.00 B 106-212wn


Figure 7. The tph of liberated sphalerite flowing through

circuits 1, 2 and 3 at ZC Mine.

PCLASS U/F =prim ar y rake class ifi er underflo w.
PBMD =pri mary ball mill discharge .
SCYCO/F =s econdar y cycl one overfl ow.
RMD =rod mill discharge.

Figure 8a. NON-SULPlllDE GANGUE Nol Circuit

140.00 . - - - - - - , - - - . - - - - - - , - - - , - - ---,- - - , -- - --,

120.00 t-- --+- - - +-- -- + -- -i -- -- + - - - i -- - - - i

100.00 t - - - - + - - -+-- - - + - -- i -- - - + - - -i - - ---i

80.00 i - - - - t - - - - i - - - - +- - - t - - --t-- - - t - - - - ;
60.00 i----t----i----7c==---t-

2O.00 _ _

Figure 8b . NON-SULPmDE GANGUE No2 Circui1

140.00 .-------.---- -r--- - --r- - --,-------,,..---.....- -----,

120.00 +----4--- - 4-- - -+---+------1---+ - - -1

100.00 +---~~--+----+---+--____::i"'

80.00 + - - -1 /

20.00 O . 0 0 ! i l l l l 1 ! 1 1 1


Figu re 8c. NON- SULPHIDE GANGUE N03 Circ uit.



0 <r28
80.00 0 28 -38
60.00 EJ 38-53
40.00 llIII 53-106
B 106-212um

R.\ID B!dD crCU/F crCO/F

F igure 8. The tph of liberated non-sulphide gangue

flowing throu gh circuits 1, 2 and 3 at ZC Mine.

PCLASS U/F =p rim ar y rake class if ie r unde rflow .
PBMD =p rimar y ball mill discharge.
SCYC O/F =secon dar y cyc lone overflow .
RMD =rod mill discharge.





tph 150




Figure 9b. QUARTZ




tph 150




Figure 9c. PYRITE




tph 150



o 0-28 ~ 28-38 [] 38-53 rn 53-75
EI 75-106 J.i] 106-150 III 150-212 13 212-300

Figure 9. The tph of liberated chalcopyrite, quartz and

pyrite flowing through the No4 Copper concentrator
circuit, Mt Isa Mines.
PMD Primary mill discharge.
PCYC U/F Primary cyclone underflow.
SCYCO/F Secondary cyclone overflow.
SMD Secondary mill discharge

is a function of two mechanisms which control the phenomenon well (Figure 9a). Figure 10 shows the size
appearance of liberated mineral into, and removal from, distribution of liberated chalcopyrite entering and leaving
each size interval: the size reduction of existing liberated the mill, the tonnages of which were apparently nearly
grains, and the creation of newly liberated grains from identical. It suggests that the coarse mineral remained
(coarser) composites. With the data presented in this unground, whereas the finer mineral (below about 150J.un)
paper it is not possible to decouple these effects, but was reduced in size.
attention will be given to this problem at a later stage of
the investigation, in the context of the modelling of the
liberation process. For the present it is sufficient to
observe that in most of the comminution stages
illustrated in Figures 5-9, the size reduction mechanism is ....

dominant. This is best seen in the behaviour of liberated Q) 80
galena in the ZC Circuit 2 primary ball mill (Figure 5b). :;::
Reading the size-by-size tonnage from the ball mill feed ~
0 60
(primary cyclone underflow) to ball mill discharge (i.e.
from right to left), the production of additional liberated
galena is seen to increase with decrease in particle size
o 40 Product
(i.e. the gradient of the line increases). The increases in
:~ ...
the size-by -size tonnages of liberated galena resulting 20
from the primary ball milling stage were: Feej

Size Interval (urn)

% increase in liberated galena "

106-212 -1 10 2
53-106 16 Size (micron)
38-53 11
28-38 28 Figure 10 Size distributions or liberated
0-28 327 chalcopyrite, MIM secondary ball mill

(The negative increase in the coarsest size implies that When designing a sulphide grinding circuit, therefore,
breakage out of that size interval exceeded breakage into it a distinction should if possible be ,made between the roles
from coarser composites). of size reduction per se, to render an appropriate size
distribution for flotation, and size reduction to achieve
A similar trend is seen in many of the other ball mill liberation. In practice, the appropriate conditions are
stages. The MIM secondary ball mill illustrates the usually assessed through batch grinding and flotation

Table 1- Ratio or underflow to overflow mass flows or liberated mineral,

for all classifiers and minerals
Mineral Galena PYrite Chalcopyrite Sphalerite NSG** Quartz

Nominal SG* 7.5 5.0 4.2 4.0 2.7 ? 2.65

Hardness (Moh) 2.5 6-6.5 3.5-4.0 3.5-4.0 6-7 7
Classifier ***
ZC 1 PRAKE 1.30 0.42 - 0.55 0.48 -
ze 1 seve 3.26 1.88 - 0.66 0.73 -
ze 2PCye 3.38 1.17 - 1.03 1.23 -
ZC 2 SCYC 0.93 2.77 - 1.65 1.23 -
ZC 3 PCYC 5.25 1.30 - 1.09 2.02 -

MIMPCYC - 1.64 1.72 - - 1.08

MIM SCYC - 2.12 3.03 - - 1.24

* Average SG of pure mineral (from literature). In the present work, significant deviations from this figure could
occur for "liberated" grains due to the definition of liberation and the stereological bias mentioned earlier.
** Non-sulphide gangue, mostly quartz; variable SG.
*** For identification of classifier codes see Appendix.
testing, but the present work has demonstrated that the low density quartz had the lowest circulating load (as
interaction between these roles is complex, presumably expected) but the chalcopyrite consistently exhibited a
due to the different liberation sizes and grinding rates of greater circulating load than the denser pyrite. Again, this
the different minerals. is attributable to the finer size of the pyrite (Table 2), and
the relatively small difference in densities.
Circulating Loads
The key conclusion apparent from Table 1 and the
In dealing with liberated mineral, the term mass flow maps of Figures 5-9 is that a large proportion
"circulating load" (referring to the flow of liberated of liberated mineral is returned by each classifier for
mineral returning to the mill) is not strictly valid in view further grinding, and that this proportion is greatest for
of the creation of new liberated mineral in the mill. the denser and thus the valuable minerals. The effect was
However it is a helpful concept, and will be used here. In particularly severe at ZC where the hydrocyc1ones returned
all the circuits studied, the circulating load of liberated about 75% of the liberated galena to their associated ball
mineral varied systematically with mineral density and mill, compared with only 53% of the NSG. Since the
size distribution, as would be expected from the NSG was presumably also harder than the galena, this
hydrodynamic principles of classification, and as has been ensures that the galena reporting to lead roughing was
observed by other workers (6, 7, 8). In Figures 5-9 the significantly finer than the NSG - an average of 45% -
effect is seen in terms of the difference in mass flow of 38Jlm, compared with 21% - 38 urn for the NSG. The
classifier underflow and overflow, and the systematic same applies to the MIM situation: the liberated
change in this difference with mineral density. The chalcopyrite reporting to flotation (secondary cyclone
difference in the underflow-overflow distribution for the overflow) was 66% - 38 urn, whereas the associated quartz
ZC rake classifiers and cyclones is particularly striking. was only 46% - 38 urn.
Table 1 shows the ratio of classifier underflow to
overflow mass flow for each mineral circuit, as a measure It is probable that some grinding capacity is taken up
of circulating load. with the breakage of mineral grains which are already
liberated, and although in some cases it may be necessary
In the ZC data, the circulating load (ratio) generally to achieve a size reduction to optimise flotation recovery,
follows mineral density, galena usually having the largest it seems likely that both overgrinding of liberated mineral
circulating load and the non-sulphide gangue the smallest. and unnecessary consumption of power and grinding
There are some exceptions. The galena ratio for the No.2 media do occur. There would therefore seem to be some
circuit secondary cyclone was anomalously low, and it is incentive in sulphide comminution circuits to optimise
not known why this should be so. In three cases (Circuit the liberation/size reduction conditions for each valuable
I primary classifier, Circuit 2 primary cyclone, and mineral. However, such benefits are often difficult to
Circuit 3 primary cyclone) the NSG ratio exceeded that of realise in practice. A project some years ago to reduce the
the pyrite. Anomalous pyrite behaviour can be attributed preferential recirculation of liberated galena at the North
to its relatively fine size (see Table 2 below) and the Broken Hill Concentrator, through modifications to the
uncertainty in the density of "liberated" minerals is such classifiers (including the installation of a sieve bend), was
that the true densities of sulphide (dense phases) and non- not successful (13).
sulphide (light phases) will be closer to each other than
indicated in Table 1; for example in extreme cases the The Influence of the Classifier
density of a "liberated" pyrite grain could be as low as 4.2
and that of a "liberated" NSG grain as high as 4.3. Some Classifiers behave according to hydrodynamic
inconsistencies in the trends seen in Table 1 can therefore principles, and their performance has a profound influence
be expected. on the associated grinding circuit. The present case is no
The same trends are apparent in the MIM data. The

Table 2 - Fineness of Selected Classifier Feeds

Mineral Pyrite Chalcopyrite NSG Quartz

% - 38 urn % - 38 urn % - 38 urn % - 38 urn
ZC 1 PRAKE 49 .0 - 18.0 -
ZC 2 PCYC 36.5 - 17.1 -
ZC 3 PCYC 37.7 - 12.1 -
MIMPCye 58.7 34.8 - 38.7
MIM seve 45.0 27.1 - 31.0

Table 3 - Classification parameters for all classifiers (total material)

Classifier ex d5(k; Rf

ZCIPRAKE 1.59 362 0.198
ZC 1 SCYC <0.5 334 0.421
ZC 2PCYC 0.81 233 0.493
ZC 2 SCYC <0.5 193 0.426
ZC 3 PCYC 0.47 122 0.507
MIMPCYC 2.48 565 0.582
MIM SCYC 0.94 122 0.411

In general, the performance of the hydrocyclones in 100 -r------,r------:::::=:~=!1

all circuits was relatively poor, with high recoveries of
water to underflow, high corrected cut-points and low
efficiency values. Table 3 summarises the classification == 80
parameters for all the hydrocyclones and for the rake o
classifier in the ZC Circuit 1. The parameters are those ~ 60
of the well-known Whiten efficiency curve (7): c: / Fitted curve
.9 ZC Circuit 2
(1) Primary cylones
20 1::1 ZC Circuit 1
where x = d/d5(k; Primary rake
d = particle size (urn)
d5(k; = corrected cut-point (J.UTl) o 200 400 600 800 1000
ex = efficiencyparameter
Particle Size (urn)
and Ec is the efficiency (or partition number) corrected for Figure lla - ZC Ccts.l&2 primary classifier
by-pass according to the expression uncorrected efficiency curves

--~ t2) 100
1 - Rr Hydrocyclones
0 80 dSOc=233Jlm
where Ea = actual efficiency (calculated from mass Q) ~
balanced size distributions). and Rf = proportion of water c: 60
reporting to underflow.

The parameters in Equation I were estimated by non-

40 Rake classifier
linear fitting procedures.

A comparison of the ZC primary rake classifier in a ~~--r-~--,r-"""---r-T--,---'r--"1

Circuit 1 and the corresponding primary cyclones in a 200 400 600 800 1 000
Circuit 2 is particularly instructive. Figure lla shows
the experimental points and fitted uncorrected efficiency Particle Size (urn)
curves for these two units. It is clear that the rake Figure lib - Corrected efficiency curves for
classifier cut coarser and recovered much less water than ZC rake and cyclones
the cyclones; its classification efficiency was also higher
(Table 3). As a consequence, the circulating load of
liberated mineral associated with the rake classifier was total product from the rake classifier circuit was coarser
much lower than that of the hydrocyclones (Figures 5a/b, than that of the hydrocyclones, even though the liberated
6a/b.7a/b; Table 1). Another consequence was that the galena size distribution was much the same. The lead
rougher flotation performance in both cases was similar,
most of the losses of liberated lead reporting to the species) cut finer than the least dense phase (liberated
secondary grinding circuit being in the coarse sizes quartz) but coarser than the denser phases. Again, this is
(Figures Sa and 6a). to be expected from the distribution of the phases in the
hydrocyclone feeds.
Figure 11b shows the corrected efficiency curves for
the two units, confirming that even after by -pass is All other things being equal, the ratio of the cut-
accounted for, the rake classifier cuts coarser and points of any two phases would be expected to be defined
somewhat more sharply than the hydrocyclones. by:

The data from all the classification units confirm that

the various liberated minerals separated according to their
densities, although the effect was often masked by the
relatively poor overall performance of the classifiers,
particularly the high water recoveries. The effect is best
seen in the MIM data. Table 4 gives the fitted
classification parameters for the MIM primary and where 0 = mineral density
secondary hydrocyclones, for liberated pyrite, p = water density
chalcopyrite, quartz and the bulk material, and Figure 12 n = hydrodynamic exponent (0.5 for
shows the fitted corrected efficiency curves for these laminar flow and 1.0 for turbulent flow) .
Lynch (1977) presented the results of Marlow's work
(9) to show that Equation 3 was valid, with n =I, for ore
treated by the ZC Mine Concentrator (formerly NBHC);
this was based on lead and zinc assays and the assumption
that these metals were present exclusively as liberated
mineral. The present data do not quantitatively fit
~ 80 Equation 3, even though the trends evident in Table 4 and
0 Figure 12 are as expected. This is almost certainly
Q) because of the large uncertainties in the estimates of a
c: 60 and dSOc due both to the spectrum of densities inherent in
the definition of liberation being used (as mentioned
.9 earlier), and to the relatively poor classification
~ 40
P - Pyrite efficiencies which themselves generate large standard
C - Chalcopyrite deviations in the parameter estimates.
B - Bulk
Q - Quartz
The relationship between the extent of liberation
0 achieved in a mill and the circulating load ratio of its
associated classifier is illustrated for all minerals and
0 100 200 300 400 500
circuits in Figure 13. The increase in liberation is defined
Particle Size (J.1m) as the ratio of the tph of liberated mineral leaving the mill
to the tph of liberated mineral entering the mill. The
Figure 12 - Corrected efficiency curves for circulating load ratio is defined as previously (Table 1).
MIM liberated minerals The inverse trend is clear and merely confirms that as the
circulating load (of liberated mineral) increases. the
increase in liberation tends to unity (i.e, no increase) until
in the limit all the mineral returned to the mill is already
Clearly the corrected cut-points increase liberated, and no additional liberation can by definition
monotonically with decrease in density, as expected and as occur. In such cases the mill acts only to reduce the
has been observed by other workers (6, 8, 9). The bulk mineral size distribution, an example of this being the
material (incorporating all liberated and composite MIM secondary mill (Figure 10). In simple terms, at low

Table 4 - Classification parameters for different minerals, MIM primary and secondary circuits

Mineral Primary Hydrocyclones Secondary Hydrocyclones

(notional SG) dSOc (urn) a dSOc (J.1m) a
Pyrite (SG 5.0) 166 4.50 42.5 1.79
Chalcopyrite (SG 4.2) 183 2.08 63.3 4.17
Quartz (SG 2.65) 1020 <0.5 145 0.60
Bulk 565 2.48 122 0.94

circulating loads the liberation mechanism is dominant, discharge (PBMD) and the secondary ball mill discharge
whereas at high circulating loads the size reduction (SBMD).
mechanism is dominant. Figure 13 also emphasises the
relatively high incidence of large circulating loads and
associated minimal liberation. 100 -r---------------,

Q) 90
r:J ZCGaJena :9
.Q a

ZC Sphalerite
ZC Pyrite
' 2.0 ZCNSG

MIM Pyrite
MIM Chalco.

r:J Rfv[)
.~ ~
A MIM Quartz 70

Q) 1.5 PBMD
...o SBMD

"- Predicted
60 +----r----r----,,....--.,---.---l
a o 100 200 300
Geometric Mean Size (urn)
0.5 +-..,...--,--,.-,...--r---.~__r-...__,.-,.-1
, I T I

o 123 4 5 6
Figure 14 Proportion liberated vs size for
Circulating Load Ratio
different mill products, MIM chalcopyrite
Figure 13 - Increase in liberation vs
circulating load, for all minerals In this case it does seem that the relationship is
linear, there being no end-point in liberation until the
In summary, it is clear from the data that the liberated theoretical limit of zero size is reached (the finest mean
minerals separated according to density, and that the size plotted is 32.6 urn, so the resolution of extrapolation
circulating loads apparent in Figures 5-9 (and Table 1) is limited). Indeed, the correlation can be represented
were a direct consequence of this. The ZC rake classifier adequately by the very simple relation
cut coarser and more efficiently than the corresponding
hydrocyclones. ThC? large water recoveries (40-60%) % liberated = 100 - 0.1 (size) (4)
experienced by the hydrocyclones in all circuits also
contributed to the circulating loads, particularly of the' fine (It should be remembered that the stereological bias is not
minerals. High circulating loads of liberated mineral accounted for in Figure 14; since the bias is an inverse
returned by the classifier will always be associated with function of liberation, it is unlikely that the relationship
low or zero liberation in the mill. for ~ liberation is as given in Equation 4).

Particle Breakage and Liberation Trends In addition, the data for all three mills fall naturally
on the same line with no mill-specific trend being
It is self-evident that in general terms the proportion evident. This implies that, for this ore, mineral and
of a particular mineral that is liberated will increase with circuit, a given size interval will always exhibit the same
decreasing size. Two interesting questions then arise, of degree of liberation, regardless of the number or type of
some process significance: grinding stages which those particles have experienced.

I. Does the proportion liberated increase linearly with A long term objective of this work is to model
decrease in size, or is the trend non-linear, liberation the liberation process so that it can be incorporated into
increasing rapidly below a certain critical "liberation comminution models to predict both the size distribution
size"? and composition of circuit products. To do this, the
liberation behaviour of each ore must be characterised. In
2. Does the proportion liberated in a given size interval the present study, a twin pendulum single particle
depend on breakage history, i.e. on the number and breakage apparatus was used to break particles of ZC rod
type of comminution machines that have acted on the mill feed to assess their liberation behaviour. The
particle? pendulum ap~tus, which has been described in detail
elsewhere (1 ), was used to break individual particles with
Figure 14 summarises the answers to these questions a ran~e of known initial sizes at a range of known specific
for the MIM data, relating to the liberation of energies (a procedure used to determine the appearance
chalcopyrite. The graph shows the proportion of function of an ore for use in comminution models).
chalcopyrite liberated in each of seven size intervals
(represented by their geometric mean sizes), for three mill The broken product was sized and analysed by
products: the rod mill discharge (RMD), primary ball mill QEM*SEM to establish the incidence of liberated mineral
in the various size fractions. Figures 15a and 15b
151 summarise the liberation behaviour of QM initial size of
particle (-5.6 + 4.75mm) at ~ input energy (1.3 kWh/t),
for the ZC ore, and they illustrate some interesting 2000
features of the liberation behaviour of this material.
0 Galena
Figure 15a displays the liberation information obtained
directly from the pendulum test. It records the
~ x Pyrite
Spha lerite
distribution of liberation of each mineral with particle
size. Here "distribution" for a particular size means the
proportion of the .mtal mineral present in the starting 1000

material which exists in a liberated form in 1ha1 size
interval, accumulated over all size intervals from fine to
coarse, for the particular conditions of the test (initial size
and energy level). Such plots therefore reflect three
- -
characteristicsof the ore:
The size distribution to which it breaks (under the 0 I I

specific test conditions). 0 100 200 300

Geometric Mean Size (urn)
The inherent size distribution of the mineral phases
in question.

The phase-specific texture of the ore, and the mineral Figure ISb Phase specific surface area for
hardnesses. ZC sulphides
Pendulum breakage: 4.75-5.6mm, 1.3 kWh/t
These plots increase from fine to coarse sizes,
reaching a maximum equal to the overall proportion of
the mineral liberated in the breakage event (not reached in
the present data). Over the range illustrated in Figures There could be several reasons for this. One likely
15a, the plots are approximately linear. reason is evident in Figure 15b, which is a plot of the
phase specific surface area (PSSA) against mean size, for
each mineral. The PSSA is a parameter calculated by
50 ""T"""------------71 QEM*SEM which measures the surface area of the phase
in question as a proportion of the total volume of that
.Q phase present (11). Such plots are approximately
Cti 40
..... horizontal at coarse sizes , the value of PSSA in this
.D region (corresponding to composite particles) giving an
:.:::i indication of the relative inherent grain size of the
'0 different minerals. At lower sizes, the value of PSSA
c:: increases rapidly as the mineral is liberated and then
iii broken further, since the breakage of liberated grains
0 20
increases surface area without increasing total volume.

10 Figure 15b shows clearly that the pyrite had a

significantly finer inherent grain size than the other
sulphides. The liberated pyrite would therefore tend to be
0 finer than the other liberated minerals, and this is clearly
0 100 2 00 3 00 seen in the mass flow maps of Figures 5-9 and in Table
Mean Particle Size (Ilm)
The inherent grain sizes of galena and sphalerite
appear very similar (almost indistinguishable in Figure
Figure ISa - Liberation distribution from 15b), even though their actual liberation behaviour
pendulum test, for ZC sulphides appears different in Figure 15a. This is probably because
the texture, mineral association or hardnesses are such that
the galena of a given size liberates more freely after
It is clear that the galena liberates most easily, and fracture than the sphalerite of the same size.
the pyrite least easily. The differences are quite
substantial - over 30% of the galena is liberated in the -
2121lm fraction, whereas only about 6% of the pyrite is
liberated. (The absolute liberation values should be
treated with caution, as they are a function of test
conditions, particularly the input energy. The full picture
is only apparent when considering the complete range of
initial particle sizes and input energies).

7. By utilising QEM*SEM to analyse the products of
SUMMARY AND CONCLUSIONS laboratory breakage tests, it is possible to draw
conclusions about the propensity of each mineral to
Clearly the incidence of liberated mineral in a liberate following a breakage event. This propensity
particular process stream in a comminution circuit is a is a function of the inherent mineral grain size, the
function of the properties of the ore and the characteristics size distribution of the product of the breakage event,
of the comminution and classification machinery. The the texture of the ore and the mineral hardness. Of
investigation which is the subject of this paper seeks to the sulphides in the ZC case, pyrite showed the
quantify and decouple these effects so that they can be lowest liberation, probably due to its finer grain size.
incorporated into models which can predict both the size Galena showed the highest liberation, possibly due to
distribution and composition of the product from a a greater tendency to break along grain boundaries
particular comminution circuit. The early conclusions of than the sphalerite, even though these two minerals
the investigation relating to the ZC and MIM surveys can exhibited similar grain size distributions.
be summarised as follows:
It is to be hoped that the trends reviewed in this paper
1. Significant amounts of valuable mineral are liberated will provide some helpful guidelines for the designers and
early in the comminution process. Where operators of sulphide mineral comminution circuits, and
practicable, this may justify the use of coarse particle expose the issues which must be addressed in maximising
concentration processes in cases where there are the metallurgical performance of such circuits. The
benefits to be gained thereby from reducing challenge for the future is to decouple and model the
circulating loads and overgrinding . liberation process quantitatively for use in circuit design
and optimisation by simulation. An essential component
2. The circulating load of liberated mineral (that of this work will be the refinement of the procedures by
reporting to classifier underflow) can be large. Its which QEM*SEM data are interpreted (including the
magnitude is a function of mineral size and density definition of liberation), the extension of QEM*SEM's
and the classifier characteristics. The circulating load capabilities to finer sizes, and the development of valid
is unnecessarily inflated when the classifier methods for the correction of stereological bias.
performance is poor, particularly in regard to a high
water recovery to underflow. There may therefore be Acknowledgements
benefits in maximising classifier performance
through correct selection of dimensions, feed solids The authors would like to thank the Management and
concentration and feedrate, and use of automatic Metallurgical Staff of the ZC Mine Concentrator of
control. Pasminco Mining (formerly ZC Mines), and the Mount
Isa Mines Limited Copper Concentrator, for permitting
3. The ZC rake classifier cut coarser and operated more access to their plants and for the extensive help which
efficiently than the corresponding hydrocyclones. they gave in sampling and sample analysis. Both
companies also contributed materially to the cost of
4. All the classifiers separated the liberated minerals as conducting the QEM*SEM analysis. The work was
expected from hydrodynamic principles; dense funded by a research grant from the Australian Mineral
minerals cut finer than light minerals, and thus Industries Research Association for a collaborative
generated larger circulating loads. The physical JKMRC project in mineral processing, of which both
process of flotation is also hydrodynamic in nature companies were sponsors.
and classification is therefore often an appropriate
method of preparing feed for flotation. However in We wish to thank Dr. David Sutherland, Paul
some cases there may be benefits in seeking to Gottlieb and Greg Wilkie of the CSIRO Division of
promote the sizing rather than the classification Mineral and Process Engineering QEM*SEM Project for
function in order to limit the circulating load of dense many useful discussions and for providing the
(valuable) mineral. The use of fine screening devices QEM*SEM analysis service. We also thank our
should be considered, where practicable. colleague Dr. Toni Kojovic for fitting the cyclone
parameters quoted in the paper.
5. In some grinding stages, minimal liberation was
apparently achieved, and even size reduction of the Finally, we wish to express our great sadness at the
liberated mineral was sometimes limited. It is recent death of Professor Gilles Barbery of Laval
important to establish the preferred role of each University, with whom we had some very helpful
grinding stage (liberation, or size reduction of discussions and who had agreed to participate in the
liberated grains, or both), and to maximise that role analysis of the data generated in this work. Professor
where possible. Barbery's wide-ranging talents were well known, but his
work on the complex problem of stereology in mineral
6. There is some evidence (for the MIM copper ore) that liberation will be particularly missed .
liberation is independent of grinding history, i.e. that
a given size interval will always contain the same References
proportion of liberated mineral, regardless of how
many breakage events it has experienced (at least over 1. Barbery G. and Leroux D., 1988. Prediction of
the range of sizes studied here). particle composition after fragmentation of
heterogeneous materials. Int. J. Mineral Proc.,
2. King RP., 1979. A model for quantitative
estimation of mineral liberation by grinding. Int. APPENDIX
J. Mineral Proc. 6,207-220.

3. Sutherland D.N. et al, 1988. The development

and applications of QEM*SEM. Chemeca 87. Codes for Classifiers. Tables 1 - 3
Melbourne, 2, 106.1 - 106.6.
ZC 1 PRAKE ZC Circuit 1,
4. Miller P.R, Reid A.F. and Zuiderwyck M.A., primaryrake classifier
1982. QEM*SEM image analysis in the ZC 1 SCYC ZC Circuit 1,
determination of model analysis, mineral secondary hydrocyclones
associations and mineral liberation. Proc XIV ZC2PCYC ZC Circuit 2,
Int. Mineral Proc. Cong. Toronto. CIM. primary hydrocyclones
ZC 2 SCYC ZC Circuit 2,
5. Barbery G., 1989. Private communication. secondary hydrocyclones
ZC 3 PCYC ZC Circuit 3,
6. Finch J.A. and Matwijenko 0 . 1977. Individual primary hydrocyclones
mineral behaviour in a closed grinding circuit. MIMPCYC MIM
CIM Bull., Nov. primary hydrocyclones
7. Lynch A.J., 1977. Mineral crushing and secondary hydrocyclones
grinding circuits. Elsevier.

8. Manlapig E.V., Drinkwater DJ., Munro P.D.,

Johnson N.W., and Watsford RM.S., 1985.
Optimisation of grinding circuits at the lead/zinc
concentrator, Mount lsa Mines Limited. Proc
IFAC Symp. "Automation for Mineral Resource
Development", Brisbane. Aus. Inst. Min.
Metall., 265-274.

9. Marlow D., 1973. A mathematical analysis of

hydrocyclone data. MSc thesis, University of
Queensland (JKMRC).

10. Narayanan S.S. and Whiten W.J ., 1988.

Determination of comminution characteristics of
ores from single particle breakage tests, and its
application to ball mill scale-up. Trans . Inst ,
Min. Metall., Section C. 97, 115-124.

11. Jackson B.R, Gottlieb P. and Sutherland D.N.,

1988. A method for measuring and comparing
the mineral sizes of ores from different origins.
Third Mill Operators Conf., Cobar, May, 61-65
Aus. Inst. Min. Metall.

12. Espinosa-Gomez R., 1989. Private Communic-

ation. (Mount lsa Mines).

13. Watters TJ., 1990. Private Communication.

(pasminco Mining, Broken Hill).

Methods of recovering platinum-group metals from
Stillwater Complex ore
E.G. Baglin
Reno Research Center, Bureau of Mines, U.S. Department of the Interior, Reno, Nevada, U.S.A.

SYNOPSIS currently Bureau research is focussing on

bioleaching of stillwater flotation
As part of its strategic and critical minerals concentrate with acidophilic chemolithotropic
program, the u .S. Bureau of Mines investigated bacteria, such as Thiobacillus ferrooxidans,
methods for beneficiating and recovering to destroy the sulfide minerals and liberate
platinum-group metals (PGM) from ores found in the PGM. A process that can oxidize sulfide
the stillwater complex, the only primary minerals without generating S02, and, at the
resource of platinum-group metals located in same time, liberate the precious metals so
the United States. Because the ore is that they are .ame nab l e to subsequent chemical
sulfidic in nature, it was concentrated by leaching, may provide an inexpensive and
froth flotation . A 45 kg/h process environmentally preferable alternative to
development unit (PDU) was constructed and 85 smelting for processing the sulfidic
to 95% of the platinum and palladium were concentrate.
recovered from the ore by using
mercaptobenzothiazole or sodium isobutyl SAMPLE DESCRIPTION
xanthate as collectors. Bench-scale research
showed that conventional matte smelting and Ore samples
leaching methods could be used to recover the
precious metals from the flotation The most significant PGM mineralization in
concentrate. More than 95% of the precious stillwater complex ores occurs in conjunction
metals reported to the matte phase during with iron, nickel, and copper sulfides, which
smelting . Several leaching schemes were are found in mineralized layers of the Banded
evaluated that selectively extracted the base Zone of the Complex. The depos it being mined
metals from the matte and allowed recovery of is known as the J-M reef, which is continuous
high-grade PGM residues that were suitable as for about 45 km in the rugged Beartooth
feed material for a precious metals refinery. Mountains, near Nye, MT. Early Bureau work
oxidative leaching of roasted flotation utilized ore from the west Fork exploration
concentrate to selectively extract precious adit, but most samples originated in the
metal values was also investigated as an Minneapolis adit, which is part of the current
alternative to conventional matte smelting- mine workings. Average grade of the
leaching strategies. stillwater ore is about 26 g combined platinum
and palladium per tonne with a 1:3.4 ratio of
ongoing research i s evaluating the effect of platinum to palladium . Geological,
bioleaching using Thiobacillus ferrooxidans on stratigraphical, petrographical, and
sulfide oxidation and liberation of precious mineralogical descriptions of the zones of PGM
metals present in stillwater complex flotation concentration are discussed in references 1-4.
Flotation Concentrates
Flotation concentrates were obtained from
Exploration work in the stillwater complex, several sources : (1) bench-scale flotation
located in south-central Montana began in tests conducted in Bureau laboratories S , (2)
1967. By 1976, four zones with high platinum- the 45 kg/h PDU operated by the Bureau 6 - ' , (3)
group metal values had been identified. At a pilot mill operated by Anaconda Minerals
that time, the Bureau of Mines initiated a Co., and (4) stillwater Mining Co.'s
research program to define methods for production mills.
upgrading and processing the ore to recover
the precious metal values. Because the most Microprobe examination showed that the
significant PGM mineralization in the concentrates were composed primarily of
stillwater complex occurs in conjunction with sulfide minerals in siliceous gangue. The
iron, nickel, and copper sulfides, principal accessory minerals were pentlandite
conventional froth flotation was selected as [(Fe,Ni)9sS)' pyrite (FeS 2), and chalcopyrite
the most logical means for concentrating the (CUFeS2), with minor to trace amounts of
ore. This paper summarizes development of the heazlewoodite (Ni 3S 2), galena (PbS),
flotation process from bench to PDu-scale and sphalerite (zns), and millerite (Nis). The
also describes several methods for extracting PGM were distributed in the form of platinoid
the precious metals from the flotation minerals (pt-Fe alloys, PGM sulfides) and in
concentrate. solid solution in the pentlandite, where

palladium partially replaced nickel in the Because of acid-consuming minerals in the ore,
crystal lattice. The other principal PGM the flotation scheme devised for the
minerals observed in the samples were anorthositic ore was not satisfactory for
vysotskite, (Pd,Ni,Pt)s, and braggite, theserpentinized ore 7 Bench-scale tests were
(Pt,Pd,Ni)s. The gangue minerals were AI, ca, conducted to screen collectors for the
Fe, and Mg silicates. serpentinized ore, and satisfactory results
were obtained with xanthates combined with
normal dodecyl mercaptan.
Three 32-h campaigns in the PDU were run to
Flotation studies determine the effect of potassium amyl
xanthate (AERO 350), sodium amyl xanthate
The Bureau began research on stillwater (AERO 353), and sodium isobutyl xanthate (AERO
complex ores in the mid 1970s with a bench- 317), each in combination with normal dodecyl
scale study employing froth flotation to mercaptan (Pennfloat 3) . During the three
recover platinum and palladium values in a campaigns, an average of 39 kg of ore per hour
sulfide concentrates . The f irst samples were was fed to the mill. Reagent dosages were 150
obtained from the Johns-Manville west Fork g xanthate, 100 g Pennfloat 3, and 7.5 g
exploration adit. The platinum-palladium Dowfroth 250 per tonne of ore. Flotation was
mineralization in the west Fork ore occurs in conducted at 16 wt% solids and a natural pH of
sulfide-bearing unaltered anorthositic rocks. 8.2. Best results were obtained with AERO
317, and results are outlined in Table II .
Recovery of 80 to 90% of the pt-pd values was Recoveries were acceptable, and the
achieved in laboratory flotation units from an concentrate averaged 220 g/tonne PGM .
acidified pulp with a mercaptobenzothiazole
(AERO 404)-polypropylene glycol methylether A 68-kg sample of the rougher concentrate was
(Dowfroth 250) reagent suite.* Typical cleaned in the bank of six pilot mill cells .
rougher concentrates contained 345 g PGM per Cleaning upgraded the concentrate
tonne . cleaner concentra~es that analyzed 860 approximately two-fold to an average grade of
to 1,030 g PGM per tonne were obtained by 463 g/tonne PGM.
refloating the rougher concentrate.

The west Fork adit was abandoned in 1976 Table II pilot mill flotat ion of
because water under high pressure was serpentinized ore with
encountered. The Bureau work continued with AERO 317-Pennfloat 3
ore samples received from Anaconda Minerals
Co. Anaconda 's major focus of exploration was Concen- Tail- compos-
the development of the Minneapolis adit, which trate ing ite
i s located in an area east of the west Fork
Adit. Ore from the Minneapolis adit was Solids kg Ih 2 .4 37 39
separated into two fractions: (1) fresh Analysis :
anorthositic gabbro containing lesser Pt g/tonne 58 0.17 3.8
quantities of serpentinized and sericitized Pd g/tonne 161 1.71 11. 3
rocks, the products of alteration, and (2) Au g/tonne 3 .4 <0.03 0 .34
material that was predominately serpentinized Cu wt% 0.31 0.009 0.03
and sericitized rocks with some anorthositic Ni wt% 0 .45 0.030 0 .06
gabbro. Distribution, %:
Pt 96 4 100
A PDU flotation mill was constructed to pd 86 14 100
concentrate the unaltered anorthositic ore 6 Au 95 5 100
A 5-day campaign processed an average of 48 cu . 69 31 100
kg/h of ore through the mill. The reagent Ni . 49 51 100
suite used for floating the ore was 200 g AERO
404, 15 kg H2S0 4 , and 5 g Dowfroth 250 per
tonne of ore. The solids content of the
flotation slurry was 36 wt% and the acidity Table I pilot mill flotation of
was maintained at pH 4. anorthositic ore with
AERO 404-Dowfroth 250
Results are summarized in Table I. Although
approximately 90% of the platinum and Concen- Tail- compos-
palladium were recovered in the concentrate, trate ing ite
the overall grade was only 38 g Pt and 120 g
pd per tonne . During the last few hours of solids kg Ih 4.1 44 48
the campaign, samples of the flotation froth Analysis:
were taken simultaneously from each of the six Pt g/tonne 38 0.34 3 .8
flotation cells. The analytical results pd g/tonne 120 1.71 12.0
indicated that the grade decreased from the Au g/tonne 2.1 0.07 0.24
first to the last cell . If the froth products cu wt% 0 .24 0.01 0 .03
i n the last four cells had been upgraded in a Ni wt% 0.35 0 .03 0.06
cleaner circuit , the overall grade of the Distribution, %:
final concentrate would have been considerably Pt 91 9 100
higher. Pd 87 13 100
Au 74 26 100
Cu . ' 69 31 100
*Reference to specific brand names is for Ni 52 48 100
identification only and does not imply
endorsement by the Bureau of Mines.

Table III Analysis of stillwater flotation concentrates
and smelting products

Weight, g/tonne wt%

Pt pd Au cu Ni Fe S

Flotation con-
centrate 500 127 278 21 1.4 2.1 10.6 7.2
Matte 93 521 1,423 93 7.5 10.5 53.0 31.5
slag 438 5.5 12 .7 . 1 . 1 1.2 1.0

Flux: CaO 37 g, CaF 2 18.3 g, si0 2 13.3 g.

To prepare the mattes, 500-g charges of

Table IV Analysis of products from stillwater flotation concentrate were smelted
first-stage leaching of in an inductively heated graphite crucible
stillwater matte with a flux cons isting of 37 g cao, 18.3 g
caF 2, and 18.3 g Si0 2 An argon blanket was
Products Analysis Distribution, maintained over the charge to partially
% exclude oxygen, to prolong crucible life, and
to minimize the oxidation of the sulfidic
Solution, gIl: charge. After melting, the charge was
cu <0.001 <0.01 maintained between 1,400 and 1,600C for
Ni 12.1 99.1 approximately 1/2 h. The molten charge was
Fe 56 99.8 poured into a graphite mold. After cooling,
pb <0.0015 <1.9 the matte and slag were separated and weighed .
Residue : More than 94% of the copper, nickel, and
Cu wt% 66 >99.9 precious metals reported to the matte.
Ni wt% 0.9 0 .9
Fe wt% 1.0 0.2 Analyses of a stillwater flotation concentrate
pb Wt.% 3.2 >98.1 and the matte and slag obtained from smelting
S wt% 18.9 NAp are presented in Table III. Volatiles and
Pt kg/tonne 5.59 >99.9 material remaining in the crucible account for
pd kg/tonne 13 .03 >99 .9 the 42-g weight loss during smelting. The
Au kg/tonne 0 .79 >99 .9 principal identifiable mineral phase in the
matte was troilite (FeS).
NAp Not applicable .
Conditions: 275 g minus 200-mesh several two-stage leaching schemes were
matte, 2.58 1 2.1M H2S0 4 (H2S0 4-to- developed that removed the base metals from
matte ratio = 1 .9), 4 h, 95c. the matte. All began with a sulfuric acid
leach, which selectively removed more than 99%
of the iron and nickel.

Matte Smelting and Leaching The following conditions gave the best results
in the first-stage H2S0 4 leaching :
Matte smelting followed by leaching of the
metal values is the conventional metallurgical o Particle size: minus 200 mesh
processing sequence for treating sulfide o H2S0 4 concentration: 2.1M
flotation concentrates containing copper, o H2S0 4-to-matte ratio : 1.9
nickel, iron , and PGM. The processing o Leaching temperature : 80-95c
sequence includes: (1) smelting to remove o Leaching time: 4 h
gangue material and to produce an iron-nickel-
copper-PGM green matte, (2) converting by Analyses of a typical pregnant liquor and
blowing air through the green matte to oxidize residue from f irst-stage leaching of
iron sulfide and to transfer the oxidized iron stillwater matte are summarized in Table IV.
to the slag as fayalite (2FeOsi0 2), (3) Distribution data demonstrate how completely
leaching the resultant copper-nickel-PGM white the iron and nickel were separated from the
matte to dissolve the copper and nickel and to copper, PGM, and gold. The small amount of
obtain a PGM residue, and (4) refining the lead in the matte reported to the residue.
residue for reco~ery of PGM. converting is The major mineral in the first-stage residue
accompanied by evolution of large quantities was djurleite (CU1 9 3S). The residue was
of S02 gas. leached in a second stage t o solubilize
copper .
Direct leaching of the green matte to extract
metal values without S02 evolution was Ferric chloride and ferric sulfate were
investigated as an alternative to converting selected as second-stage lixiviants because of
and leaching of white matte 9 Oxygen pressure their low cost and high capacity for
leaching of green mattes has been em~loyed by extracting copper f rom first-stage residues .
industry to avoid S02 generation 1 0 - 1 , but Ferric iron oxidized the copper and sulfide in
high-pressure autoclaves are required . In the the djurleite and solubilized copper, leaving
Bureau's i nve s t i gat i on , green mattes were the PGM in the residue . The reaction between
prepared and leached under ambient pressure ferric iron and the djurleite was
conditions. The objective of this research stoichiometric, according to the following
was to selectively extract the base metals equation:
from the mattes and to recover the PGM in a
high-grade residue.

Table V second-stage ferric chloride or sulfate leaching of
stillwater matte

Ferric chloride Ferric sulfate

Analysis Distribution, Analysis Distribution,
% %

Cu g/l 5.0 98.7 5.2 99.4
Ni g/l 0.21 98 .4 0.21 97.4
Fe g/l.. 8 .7 >99.9 9.2 >99.9
pb mg/l.. 202 99 .4 <7.5 <1.0
Pt mg/l.. <0.02 <0.06 <0.02 <0.05
pd mg/l.. <0.02 <0.02 <0.02 <0.02
Au mg/l <0.004 <0.07 <0.004 <0.06
Cu wt%.. 4.9 1.3 1.7 0.6
Ni . wt%. 0.26 1.6 0.32 2.6
Fe wt%. 0.15 0.02 0.06 0.01
pb . .. wt%. 0 .09 0.6 10.5 >99.0
s ........ wt% 79.8 NAp 75.8 NAp
Pt kg/tonne . 24.1 >99.9 21.2 >99 .9
pd kg/tonne . 63 .8 >99.9 56.2 >99.9
Au oz/ton 3.9 >99.9 3.5 >99.9

NAp Not applicable.

Conditions: 5 g first-stage residue, 600 cm3 0.15M Hcl or
H2SO C containing stoichiometric quantity of ferric chloride or
sulfate, respectively (pH 1.0), 6 h, 70c.

The second-stage residue contained less than 5 410 g/tonne in the flotation concentrate to
wt% copper and more than 75 kg/tonne PGM more than 200 kg/tonne in the residue. The
(Table V). final residue contained more than 99% of the
PGM in the matte.
The following conditions gave the best results
in the ferric chloride and ferric sulfate Roasting was investigated as an alternative to
leaching: oxidative second-stage leaching. Roasting
oxidized the djurleite to .the leachable copper
o Ferric salt - stoichiometric amount species , Cu2o, cuO, CUSO c ' and CUOCUSO c
according to equation 1
o pH 1.0 (adjusted with Hcl or H2SO C) Dilute sulfuric acid solubilized more than 98%
o Leaching temperature: 70C of the copper from first-stage residues when
o Leaching time: 1 h, Fe 2(SOC)3; 4 h, the roasting temperature was at least 360c.
Fec1 3 One mole of sulfuric acid was required per
mole of copper solubilized, and the PGM
After ferric sulfate leaching, the principal remained in the solids. Roasting between 360
species in the residue were elemental sulfur and 500c was the best range for separating
and lead sulfate . The residue was upgraded copper and PGM. Roasting at temperatures
two-fold by dissolving the elemental sulfur in below 360c resulted in decreased copper
carbon disulfide. Lead sulfate would be extraction, while PGM values were lost to the
removed during refining. After ferric leaching liquor when roasting was conducted at
chloride leaching, the second-stage residue 500c or above.
contained sulfur and a noncrystalline,
unidentifiable phase. The lead content was Analyses of the products of the roasting-
0.09 wt% compared with 10 .5 wt% in the sulfate leaching procedure are reported in Table VI.
leached residue. Ferric chloride solubilized copper extraction was complete in 1/2 h at
the lead, while the ferric sulfate 25C with 0.2M H2SO C The copper content of
precipitated it. the pregnant liquor was i nc r e a s e d to 48 gil,
which is satisfactory for copper
Advantages of the sulfate system over the electrowinning, by increasing the slurry
chloride system are (1) the second-stage solids in leaching from 23 to 100 gil. The
extraction t ime is shorter, 1 h at 70 c for final residue contained 300 kg precious metals
ferric sulfate compared with 4 h for ferric per tonne. A flow diagram for the roasting-
chloride leaching, (2) ferric sulfate and leaching procedure is shown in Figure 2.
sulfuric acid are less expensive than ferric
chloride and hydrochloric acid, (3) sulfate is Leaching of Roasted Concentrate
less corrosive and less volatile than
~hloride, (4) sulfate solution is more A procedure was investigated for selectively
suitable for electrolytic deposition of leaching PGM and gold directly from the
copper, and (5) lead sulfate in the residue flotation concentrate, which would eliminate
can be easily removed during refining, whereas the need for smelting and significantly
lead in the second-stage chloride liquor decrease processing time compared to
requires additional processing steps for conventional treatment of PGM sulfide
removal. concentrates 1 2

Figure 1 shows a flowsheet summar1z1ng the The concentrate used was a cleaner concentrate
two-stage procedure developed for leaching prepared from Minneapolis adit anorthositic
stillwater matte. The PGM content was ore in the Bureau's PDU using an acidified
upgraded almost 500-fold, from approximately pulp and AERO 404 as the collector. The

Stillwater concentrate
(~410 g/tonne PGMl

Flux: CoO, CaFz,

and S i02 Siao

Matte (~2 kg/tonne PGMl

L -_ _---. ........ Solution (Ni, Fe)

to metals recovery
First - stage residue
(~20 kg/tonne PGMl

Fe2(S04h + H2S04 Second - stage Solution (cc, Fe)

or FeCI3 + HCI leaching to metals recovery

Second-stage residue
(~90 kg/tonne PGMl

Final residue
(~200 kg/tonne PGMl
to refinery

Fig. 1 Two-stage matte leaching.

St illwater concentrate
(~410 g/tonne PGMl

Table VI Analyses of products

from leaching of Flux : CaO, CaF21
and Si02 Slag
roasted first-stage
residue from
stillwater matte Matte (~2 kg/tonne PGMl

Products Analysis

Solution, gIl:
Cu. 12.3
Hi 0.04
Fe . 0.04
Pt . <2xlO- 5 Solut ion (Ni, Fe)
pd .. <2xlO - 5 to metals recovery
cu . wt% 10.2 First - stage res idue
Hi . wt% 0.19 (~20 kg/tonne PGMl
Fe wt% 0.44
pb .. wt% 7.0
Pt kg/tonne . 67.2
pd kg/tonne 224 .6
Au kg/tonne . 7.0

Conditions : 4.0 g first-

stage res idue roasted at Solution (Cu)
350 to 400C, 175 cm] to metals recovery
0.2M H2SO. (H2S0.-to-cu
molar ratio = 1:1), 0.5 h,
Final residue
(~300 kg/tonne PGMl
to refinery

Fig. 2 Matte r o asting-leaching p roc e du r e .

Table VII Analysis of flotation and (2) high chloride ion concentrations (>6M)
concentrates promote the oxidation of the precious metals
by complexing and stabilizing dissolved ions .
Concentrate sample
Further investigation showed that leaching
A B e D temperature had a significant effect on the
selectivity of metal extraction. The data in
Analysis : Table VIII show that leaching at ambient
Pt 9 /tonne 301 300 105 342 temperature (approximately 25C)
pd 9 /tonne 556 1,139 296 1,175 preferentially extracted the precious metals
Au g/tonne 15 14 6 19 from the concentrate and left almost all of
Cu wt% 1.4 2.0 0.6 1.4 the base metals in the residue. When the
Ni wt% 0.8 2.5 0.7 3.2 leaching temperature was increased to 105C ,
Fe wt% 8.2 12.0 7 .4 11. 6 PGM extraction increased 8 to 20%, while
S . wt% 5.9 8.7 2.8 7.0 coextraction of base metals and gangue
A120 J wt% 7.1 4.3 4.9 5.2 minerals increased several-fold. utilizing
eaO . wt% 3.3 3.0 5.0 3.2 preferential extraction of the precious metals
MgO wt% 20.3 20 .8 24.1 18.5 at 25c simplifies processing of the roasted
Si0 2 wt% 46.7 40.7 48.1 43.0 concentrate by eliminating steps for base
metal recovery. Loss of the copper and nickel
in the tailings represents less than 5% of the
monetary value of the concentrate, and their
analysis of this concentrate is shown in Table recovery could be undertaken if the economics
VII (sample A). Experiments were conducted were favorable.
using 25 to 300-g samples.
Table IX shows the effect of roasting
preliminary studies indicated that roasting temperature on the extraction of precious
the concentrate at temperatures above 800C metals during ambient temperature leaching.
made the platinum, palladium, and gold Extraction increased with temperature up to
amenable to oxidative leaching in strong Hel 1,050 oe and decreased at higher temperatures
solutions. Effective oxddants included owing to fusing of the charge . Maximum
hydrogen peroxide, chlorine, sodium platinum, palladium, and gold extractions of
hypochlorite, nitric acid, and persulfate 95, 90, and 99%, respectively, occurred after
salts. Leaching was effective only when roasting at 1,050 oe. Hydrogen peroxide
solution potentials were above 1.1 V, Eh. additions totaling 0.005 to 0.03M were
Most of the experimental data were generated sufficient to maintain the oxidation potential
using hydrogen peroxide, which is easy to during ambient-temperature leaching of the
handle and yields water as the reduction roasted concentrates.
product. On the basis of cost and
selectivity, sodium hypochlorite, chlorine, X-ray diffraction examination of concentrates
and hydrogen peroxide are the most favorable roasted at temperatures greater than 800c
reagents. If sodium hypochlorite were used showed a ferrite-type major phase, MFe 20. (M
for leaching, control of sodium ion buildup in eu and/or Ni), possibly mixed with maghemite
the pregnant solution would be necessary. (a-Fe 20 J ) , which has a very similar crystal
Peroxide and chlorine add no contaminating structure. Microprobe studies identified a
ions to the strong Hel leaching solution. silicate phase and a number of nickel, iron,
and nickel-iron oxidic phases. Some PGM were
Ninety percent or more of the platinum, converted from sulfide minerals to their
palladium, and gold were extracted by hot elemental states by the roasting process.
H20 2-6M Hel solution when the roasting Some very small palladium-rich particles
temperature was 800C or higher. However, existed as relatively pure inclusions in the
appreciable quantities of base metals and nickel-iron oxide phase. Larger discrete
gangue minerals were also solubilized . grains of PtO-pdo particles, which ranged in
Hydrogen peroxide-strong Hel solutions size up to 15 pm, were also observed . These
effectively solubilize precious metals from particles exhibited varying ratios of
roasted concentrate because"(l) the high acid palladium to platinum. Thus, both locked and
concentration increases the half-reaction liberated grains of palladium occurred in the
potential for H20 2 acting as an oxidizing roasted concentrate, but platinum existed
agent, primarily in liberated form.
EhO 1.78 v, (2)

Table VIII Effect of leaching temperature on H20 2 -

6M Hel leaching of 800c roasted

Leaching Extraction, %
temp, e
Pt pd Au eu Ni Fe A120 3 eaO MgO si0 2
25 87 72 96 3.6 2.5 2.8 32 42 11 0 .6
70 : 98 82 >99 18 50 37 55 57 22 1.3
105 95 92 >99 47 91 84 81 69 52 .2

Conditions: 50 g roasted concentrate A, 840 cm3

0 .7M H20 2-6M Hcl leaching solution, 6 h.
1 22 h.

The lower solubility of palladium compared to Concentrate C was prepared by twice cleaning
platinum and gold can be attributed to locked rougher concentrates recovered from
palladium and to the fact that Pt0 2 , which can serpentinized ore in the Bureau's pilot mill.
form during roasting, decomposes to metal and An AERO 317 -Pennfloat 3 collector combination
oxygen above 650C, while pdo, the most stable was used at a natural pH of 8.2 . Concentrate
oxide of palladium, can exist at red heat and o is production mill concentrate obtained from
is insoluble in mineral acids. Higher stillwater Mining Co. The SMC flotation
roasting temperatures favor decomposition of circuit employs AERO 350 as the collector at
pdo to metal and result in increased palladium pH 7.8 to 8.2.
extraction during leaching.
None of the concentrates leached as well as
Roasting temperature also affected base metal the original sample and palladium was
and gangue mineral attack. coextraction of cons istently less soluble than platinum and
base metals after roasting at 1,050C was less gold. Extraction from concentrate 0 peaked at
than after roasting at lower temperatures, and a roasting temperature of 990C, with 94% of
magnesium extraction decreased to the platinum, 77% of the palladium, and 97% of
approximately 1%. since Mgo comprises about the gold being solubilized. SEM examination
20 wt% of the concentrate, any decrease in of the residue showed palladium-rich
solubilization is significant to the overall inclusions locked in the ferrite phase, as
process. The principal contaminants of the well as a small percentage locked in the
pregnant solution were A1 3 + and Ca 2 + ions. silicate phase. It is believed that by
judicious choice of grind size, roasting
cursory roasting-leaching experiments were temperature, and leaching conditions, the
undertaken with three other concentrates palladium extraction can be increased to above
prepared from Minneapolis adit ore. Analyses the 77% obtained in these tests.
are summarized in Table VII and leaching
results are given in Table X. Concentrate B Incorporating a H2S0 4 preleaching stage into
came from flotation of a mixture of the processing scheme significantly decreased
anorthositic and serpentinized ores using a contamination of the pregnant solution because
natural pH-xanthate circuit that included easily soluble gangue minerals were removed
three cleaning cycles. This concentrate was prior to oxidative leaching. Pre leaching was
received from a pilot mill operated by conducted by heating a pulp containing 10 wt%
Anaconda Minerals Co. in Tucson, AZ. solids in 1M H2S0 4 at 70 C for 4 h.

Table IX Effect of roasting temperature on H20 2-6M Hcl

leaching of stillwater concentrate at ambient

Roasting Eh, Extraction, %

temp , C V
Pt pd Au cu Ni Fe A120 3 CaO Mgo s i0 2

NR 1 1.12 1.2 8.9 40 12 14 21 42 30 7 .4 1.8

750 1.14 54 52 85 20 8.9 6.3 36 47 22 .8
800 1.18 90 73 99 3.3 <.1 1.8 28 42 11 1.2
880 . 1.18 92 75 99 3.0 3.5 2 .0 33 40 7 .6 1.6
960 1.18 92 81 98 1.3 < .1 .8 32 40 3.3 1.2
1,050 1.17 95 90 99 .9 < .1 1.1 38 40 1.0 1.0
1,110 . 1.20 83 74 78 2.3 1.8 3.7 20 21 .7 .5
1,130 1.20 86 78 82 6.9 6.9 7.0 3.2 3.3 2.0 .1

NR Not roasted.
Conditions: 25 g roasted concentrate A, 420 cm3 0.03M H20 2 -
6M Hc1 leaching solution, 25 c, 6 h.
10.2M H20 2-6M Hcl.

T~ble X Effect of roasting temperature on H20 2-6M Hcl leaching

on different stillwater concentrates

Roasting Extraction , %
Concentrate temp, C
Pt pd Au CU Ni Fe Al 20 3 CaO MgO si0 2

A . . . 800 90 73 99 3.3 <0.1 1.8 28 42 11 1.2

960 92 81 98 1.3 <.1 .8 32 40 3 .3 1.2
1,050 95 90 99 .9 <.1 1.1 38 40 1.0 1.0
B . 800 76 45 92 5 .9 4.0 4 .2 20 53 20 2.0
960 68 60 91 1.3 .8 .6 18 20 2.6 .9
1,040 89 75 96 .8 .6 .5 20 19 1.4 1.0
c 800 89 48 99 7.8 5.3 5.5 29 60 9.8 1.6
960 85 72 99 <.2 <.2 1.2 22 36 2.7 .7
1,050 90 82 98 . 1 1.0 .2 26 27 2 .5 1.7
D 950 64 72 84 11. 0 11.0 6.7 46 50 9 .6 4.4
990 94 77 97 8.4 9.0 5 .5 43 48 5.7 3.8
1,050 82 78 97 6.0 5.8 4.6 39 42 3.7 3 .7
1,070 74 77 92 9.7 8.8 7.4 46 51 5.4 3 .9

Conditions: 25 g roasted concentrate, 420 cm3 0.03M H20 2-6M

Hcl leaching solution, 2SC, 6 h.

Table XI shows the selectivity of two-stage Based on the experimental results described, a
leaching, and Table XII compares analyses of flow diagram for the roasting-leaching
the pregnant solutions and tailings obtained procedure was devised and is presented in
by oxidative leaching with and without H2S0 4 Figure 3. Flotation concentrate is roasted at
pretreatment. The precious metals content 1,050C and pre leached in H2S0 4 to dissolve
exceeded 60 mgll in both pregnant solutions, impurity elements. oxidative leaching in 6M
but contamination by base metals and gangue Hcl is conducted on the solids to extract the
minerals was decreased from a total of 3.2 gIl PGM and gold. The precious metals are
to 0.1 gIl when the concentrate was pre leached recovered from solution for treatment at a
prior to oxidative leaching. refinery .

Sulfuric acid consumption during pre leaching Bioleaching of stillwater Flotation

depends on the nature of the gangue minerals Concentrate
in the concentrate and the duration and
temperature of the pre leaching operation. New research has been initiated by the Bureau
Acid consumption for the samples tested ranged to biologically oxidize the sulfide minerals
from 0.04 to 0.22 g H2S0 4 per gram of in stillwater flotation concentrate with .
concentrate . Dilute H2S0 4 for pre leaching Thiobacillus ferrooxidans . Bioleaching avoids
could be produced from the S02 in the roaster the generation of S02 by oxidizing sulfide to
offgas. sulfuric acid and is considered to be a more
environmentally acceptable way of removing
Platinum, palladium, and gold were effectively sulfide than processes that include high-
recovered from the pregnant solution by three temperature smelting or roasting steps. The
techniques: (1) precipitation as sulfides by concept is to decompose the sulfide minerals
sparging the solution with hydrogen sulfide, and liberate the PGM so that they become
(2) cementation on nickel granules, and (3) amenable to subsequent chemical leaching.
collection on a column of activated coconut Results of this research will be published in
charcoal. the near future.

Table XI Two stage leaching of

1,050C roasted SUMMARY
Bureau of Mines research showed the
Extraction, % feasibility of upgrading stillwater Complex
ore by flotat ion concentration. In addition,
stage 1 stage 2 procedures were developed for recovering
(1M H2S0 4 (H 20 2-6M Hcl platinum-group metals and gold from the
concentrate by matte smelting and leaching and
Pt 0 .2 97 by direct leaching of preroasted concentrate.
pd .. . .3 92 Current Bureau research efforts are directed
Au .. <.5 99 towards liberation of the precious metals from
Cu ... 5.1 .06 stillwater complex flotation concentrate by
Ni 3.2 .13 bioleaching, to provide an alternative to
Fe .. 3.9 .62 high-temperature processes that include
49 17 smelting or roasting steps
A1 2 0 3
CaO .. 56 1.6
MgO .. 4.0 .20
Si0 2
Conditions: stage 1 - 300 g The author gratefully acknowledges the
roasted concentrate A, 2.8 1 1M assistance of Anaconda Minerals Co. (Atlantic
H2S0 4 , 70C, 4 h. stage 2 - Richfield corp.), Johns-Manville Corp., and
150 g preleached concentrate, stillwater Mining Co. for providing ore and
2.5 1 O.03M H20 2-6M Hel leaching concentrate samples used in this
solution, 27C, 6 h. investigation.

TABLE XII Analysis of products from single-and two-stage

leaching of 1 ,050C roasted concentrate

pregnant solution analysis, mgll Tailing analys is, wt%

single-stage Two-stage s ingle-stage Two stage

Pt 21.8 24.6 1
20 1 12