Prévia do material em texto

Published by:
The Australasian Institute of Mining and Metallurgy
Ground Floor, 204 Lygon Street, Carlton Victoria 3053, Australia
IRON ORE CONFERENCE 2011
11 - 13 JULY 2011
PERTH, WESTERN AUSTRALIA
The Australasian Institute of Mining and Metallurgy
Publication Series No 6/2011
© The Australasian Institute of Mining and Metallurgy 2011
All papers published in this volume were refereed prior to publication.
The Institute is not responsible as a body for the facts and opinions advanced
in any of its publications.
ISBN 978 1 921522 43 7
Desktop published by:
Kylie McShane, Anthony Salerno and Olivia Tet Fong
The Australasian Institute of Mining and Metallurgy
Compiled on CD ROM by:
Visual Image Processing Pty Ltd
PO Box 3180 
Doncaster East VIC 3109a
Keynote Addresses
The Roy Hill Project J M F Clout and B Fitzgerald 3
The SNIM Guelb II Project in Mauritania – 
Process Design Challenges
G Cooper, R Lapointe 
and J Routhier
11
The Benef ts of Computational Modelling in 
Improving Iron Ore Production and Processing
M P Schwarz and P Cleary 17
Dewatering Marra Mamba – Lessons Learned K Sommerville 
and B Douglas 
19
Recent Iron Making Topics in Japan N Takamatsu 21
Developing the West Pilbara Iron Ore Project K Watters, A Priddy 
and D Ducler des Rauches
27
Development Options for the Eyre Peninsula 
Magnetite – An Owner’s Perspective
J White and B Hammond 29
Environment
Overcoming the Challenges of Managing Tailings 
from Iron Ore Mining in Western Australia
D R Anstey and D A Reid 33
Iron Ore Holdings Ltd – Case Study of the 
Environmental Approval Challenges for an 
Emerging Junior
H Fletcher and J Moro 43
Using X-Ray Diffraction for Grade Control and 
Minimising Environmental Impact in Iron and 
Steel Industries
U König, L Gobbo 
and K Macchiarola 
49
Exploration
New Ore Types from the Cauê Banded Iron 
Formation, Quadrilátero Ferrífero, Minas Gerais, 
Brazil – Responses to the Growing Demand
L Q Amorim and F F Alkmim 59
From Banded Iron Formation to Iron Ore – 
Genetic Models and Their Application in Iron 
Ore Exploration in the Hamersley Province, 
Western Australia
H J Dalstra 73
CONTENTS
Exploration continued...
(U-Th)/He Geochronology of Channel Iron 
Deposits, Robe River, Hamersley Province, 
Australia – Implications for Ore Genesis
M Danišík, E R Ramanaidou, 
N J Evans, B J McDonald, 
C Mayers and B I A McInnes 
83
Contrasting Styles of High-Grade Iron 
Mineralisation at Weld Range, Western Australia
P Duuring and 
S G Hagemann 
87
Discovery and Geology of the McPhee Creek Iron 
Deposit, Northern Pilbara, Western Australia
J D Goldsworthy, R M Joyce, 
P Bonato and A D'Hulst
93
Realising the Potential of the Boolgeeda Iron 
Formation – Stratigraphy and Iron Mineralisation 
at McCamey’s North, Hamersley Province, 
Western Australia
P J Howard and P Darvall 103
Airborne Gravity Gradiometry and Magnetics in the 
Search for Economic Iron Ore Deposits
R Miller and M Dransf eld 109
Microplaty Haematite of the High-Grade Iron Ores 
– Its Nature and Genesis
R C Morris 117
Channel and Detrital Iron Deposits of the Flinders 
Mines Pilbara Iron Ore
A E Petts, G D McDonald 
and N J Corlis
125
The Tonkolili Iron Ore Deposits, Sierra Leone M S Reston, H T Baker, 
R D Elvish, C A Reardon and 
B J W Young 
133
Mapping Quartz, Carbonates and Riebeckite in 
Banded Iron Formation Iron Ore Drill Core Using 
CSIRO’s TIR-HyLogging System
M C Schodlok and 
E R Ramanaidou 
147
Mining
In-Pit Crushing and Conveying Bench Operations T Atchison and D Morrison 157
The Use of Integrated Web-Based Solutions in 
Productivity Improvement
P Cooper 165
Performance Enhancing Blends – Concentrator 
Feed Blending at Iron Magnet
M Darvall 173
Identif cation of Shale and Ore Boundaries Using 
Gaussian Processes
K L Silversides, 
A Melkumyan, D A Wyman 
and P J Hatherly
179
Research on the Shape and Particle Size 
Distribution of Caved Ore Mass in Sublevel Caving
G J Zhang 185
Ore Characterisation
New Innovations in Field Portable X-Ray Fluorescence 
for the Iron Ore Industry
S Bailey 191
Real-Time Image Analysis of Iron Ore Cores and 
Drill Chips to Complement Spectral Measures
T Belligoi, E R Ramanaidou 
and E Pirard 
207
Comparative Study of Iron Ore Characterisation by 
Optical Image Analysis and QEMSCANTM 
E Donskoi, J R Manuel, 
P Austin, A Poliakov, 
M J Peterson and 
S Hapugoda
213
Methodology for Selecting Borehole Samples of Iron 
Ore for Mineral Processing Characterisation
G J I Dos Santos and 
D T Ribeiro 
223
Technological Characterisation of West African Iron 
Ores in Order to Predict their Performance in the 
Benef ciation Process
L Dubron, E Pirard 
and A Pirson
229
Use of Metallurgical Test Data in Resource 
Evaluation for Magnetite Deposits
J N Farrell and A D Miller 241
Characterisation of Bedded and Channel Iron Ore 
Deposits Using CSIRO’s HyLoggingTM Systems
M Haest, T Cudahy, 
C Laukamp, 
E R Ramanaidou, S Gregory, 
J C Stark and D Podmore
249
Determination of Iron Ore and Gangue Mineral 
Associations Using Optical and Backscattered 
Electron Images with Electron Probe Microanalysis
S Hapugoda, J R Manuel, 
M J Peterson and E Donskoi 
257
Quantitative X-Ray Mineralogy of Iron Ore 
and Scales
K Knorr and N Yang 265
Geometallurgy and Ore Processing of the Hibbing 
Taconite Lake Superior-Type Magnetite Taconite 
Deposit, Mesabi Iron Range, Minnesota, USA
J D Lubben, 
P K Jongewaard 
and M E Young 
271
The Occurrence of Phosphorus and Other 
Impurities in Australian Iron Ores
C M MacRae, N C Wilson, 
M I Pownceby and 
P R Miller 
281
The Scratch Test – An Attractive Method to 
Measure the Strength of Iron Ore Material
L Mariano, T Richard 
and E R Ramanaidou 
291
Ore Characterisation continued...
Real-Time Online Analysis of Iron Ore, Validation 
of Material Stockpiles and Roll Out for Overall 
Elemental Balance as Observed in the Khumani 
Iron Ore Mine, South Africa
D Matthews and T du Toit 297
Geochemical Characterisation of Flinders Mines’ 
Pilbara Iron Mineralisation
G D McDonald 307
Nature and Distribution of Gibbsite in Some 
Western Australian Iron Ores
M Paine, E Ryan 
and P Mackenzie 
315
Measurement of Iron Ore Phases R Pax 321
Particle Size Analysis – Optimising Returns in 
Iron Ore
B Stump and B McPherson 323
Occurrence and Mineralogical Association of 
Phosphorus in Australian Bedded Iron Ore Deposits
M A Wells and 
E R Ramanaidou 
331
Ore Reserve Estimation
Resource Estimation Using Reverse Circulation and 
Blasthole Samples in a Bedded Iron Ore Deposit
C Boyle 339
Raman I Do – Raman Spectroscopy for the 
Mineralogical Characterisation of Banded Iron 
Formation and Iron Ore
E R Ramanaidou 
and M A Wells 
345
Application of Optimisation Technology to Resource 
Planning Within BHP Billiton Iron Ore
D Xu, R Pasyar and O Wang 351
Processing
Investigation on Collector Optimisation in the 
Reverse – Iron Ore Flotation
M S Cassola and K U Pedain 361
The Integrative Technology of SLon Magnetic 
Separator and Centrifugal Separator for Processing 
Oxidised Iron Ores
X Dahe 367
The Effect of Seasonal Variations on the 
Performance of Mineral Processing Plants – A Case 
Study of Deposit-10/11A Mine of NMDC Ltd
S Das, T Anantharaman, 
T N S Kumar, P K Satpathy 
and S Bose
373
Processing continued...
Use of an Online Elemental Analyser to Optimise 
the Sinter Process at ThyssenKrupp Steel Europe, 
Duisburg, Germany 
C Delwig, H Fettweis, 
T Schnitzler, S Wienströer, 
S Ferguson and G Noble
381
Utilisation Technologies for Australian Iron Ore 
in China
D Duan, H Han and S Wu 389
Bio-Benef ciation of Australian Iron Ores – 
Potential Applications of Indigenous Bacteria for 
Flotation, Flocculation and Phosphorous Removal
R Dwyer, S Rea, W Bruckard 
and R Holmes 
397
Removal of Phosphorous from Australian Iron Ores C I Edwards, 
M J Fisher-White, R R Lovel 
and G J Sparrow
403
Understanding Reactions in Iron Ore Pellets A R Firth, J DDouglas 
and D Roy
413
Thickening at Karara – The Role of Test Work and 
Innovative Design
G Hart 425
Calculating the Value of Iron Ores in Ironmaking 
and Steelmaking
T Honeyands and L Jelenich 
431
Innovative Benef ciation of Iron Ore Fines with 
allf ux® Two Stage Fluidised Bed Classif ers and 
gaustec® Wet High Intensity Magnetic Separators
A Horn and M Wellsted 437
Dephosphorisation of Limonitic Concentrate by 
Roasting, Acid Leaching and Magnetic Separation
K Ionkov, S Gaydardzhiev, 
A Correa de Araujo, 
H Kokal, A Pirson 
and D Bastin 
445
Conversion of a Pelletising Induration Furnace from 
a Haematite to Magnetite Feed
R Jones, M Bannear and 
R Martin
453
A Study into Bonding within Reduced 
Titanomagnetite-Coal Compacts
R J Longbottom, 
B J Monaghan, 
S A Nightingale 
and J G Mathieson 
459
Processing continued...
The Dispersion of Kaolinite X Ma 471
Investigation of Mixing and Rolling Drum 
Performance at Port Kembla’s Sinter Machine 
Through Full-Scale Sampling and Laboratory 
Scale Experiments
D Maldonado, R Davis, 
S Haehnel, P Drain 
and J Heslin 
475
Process Control Integration into a Brownf elds Iron 
Ore Plant as Applied to the Khumani Iron Ore Mine
D Matthews 485
Expansion of Iron Ore Processing Plant Using a 
Modular Design Basis as Applied to the Khumani 
Iron Ore Mine, South Africa
D Matthews and T du Toit 491
Alumina Reduction Challenges in the Benef ciation 
of Low-Grade Haematite Iron Ores
N Murthy and B Karadkal 499
Improving the Fluxed Pellets Performance by 
Hydrated Lime Instead of Bentonite as Binder
J Pan, D Q Zhu, M Emrich, 
T J Chun and H Chen 
509
Oversize Reduction Project at the Iron Ore 
Company of Canada
R Pinksen and J P R Proulx 513
Characterisation of Long-Term Scaling Effects of 
Ceramic Filter Media Used in the Dewatering of 
Iron Ore
R Salmimies, A Häkkinen, 
J Kallas, B Ekberg, 
J P Andreassen and R Beck 
521
Unlocking the Value in Waste and Reducing Tailngs 
– Magnetite Production at Ernest Henry Mining
J Siliézar, D Stoll 
and J Twomey
529
In Situ Diffraction Studies of Phase Formation 
During Iron Ore Sintering
N A S Webster, 
M I Pownceby, I C Madsen, 
N V Y Scarlett, L Lu 
and J R Manuel
537
Benef ciation of Low-Grade Haematite Ores D Wei, Z Guan, S Gao, 
W Liu, C Han and B Cui
545
Fine Sizing in Magnetite Concentration J Wheeler 557
The Product Quality System at Cliffs Natural 
Resources – Koolyanobbing Iron Ore Operations
A L Wills, K F Jupp and 
T J Howard 
563
Improving the Granulating and Sintering 
Performance by Pretreating Concentrates Using 
a Roller Press
D Q Zhu, J Pan, T J Chun 
and D Chen 
575
Project Development
The Central Eyre Iron Project – Process Selection 
and Design
D Connelly and L Ingle 581
Identify, Convert and Sustain – Resource 
Development in the Pilbara – The Rio Tinto Way
G Danckert and 
J Masterman 
589
A Collaborative Approach to Iron Ore Benef ciation 
at Assmang Khumani Processing Plant
S Döpp, T du Toit and 
D Ziaja 
597
Process Development for the Marampa Iron Ore 
Tailings Project, Sierra Leone
P Dunn, M Nevens, J Pease 
and R Nardi 
607
Razorback Iron Ore Project – A New Iron District 
to Meet the Growing Demand of an Iron 
Hungry World 
M Flis, G England 
and T Thomas
615
Extension Hill Direct Ship Iron Ore Project – 
Seven Years in the Making
R Forster 625
Development of the Hawsons Low-Grade Magnetite R L Koenig and 
K T Broekman 
633
Cape Lambert Upgrade Project S Russell 639
Solomon Iron Deposits – Continued Growth and 
Development
C Simpson, D Storey, 
D Kepert, R Boyd and 
N Nitschke 
649
Author Index 661
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 3
INTRODUCTION
The Roy Hill iron ore deposit is located in the Chichester 
Range on the northern side of the Fortescue River valley, 
in the Pilbara region of Western Australia. The mine area is 
located approximately 1300 km north of Perth and 100 km 
north of the regional centre, Newman. The export facility at 
Port Hedland is located 342 km by rail from the mine. 
The iron ore occurrences in the Chichester Range were 
initially identif ed by Broken Hill Propriety Limited (BHP) in 
the 1950s during a regional search for manganese. The area 
was mapped and sampled by the Geological Survey of Western 
Australia in 1959, and after the export embargo on iron ore 
in Western Australia was lifted in 1960, BHP was granted a 
Temporary Reserve 3358 in 1964. Following various tenement 
changes in the deposit area, two exploration licenses granted 
to BHP in 1991 over the area were relinquished and picked up 
by Hancock Prospecting Pty Ltd (HPPL) in 1993 and f nally 
the last tenement was dropped by Hamersley Iron Pty Ltd and 
picked up by HPPL in 2005. These three tenements comprise 
the Roy Hill Mining Area and have subsequently been replaced 
by Mining Leases M 46/518 and M 46/519, granted to Roy 
Hill on 1 November 2010. 
HPPL began an initial drilling program in 1993. The 
systematic evaluation of the deposit has been underway since 
2003, with approximately 98 per cent of drilling taking place 
in that period.
A preliminary feasibility study was concluded in August 
2008. The bankable feasibility study (BFS) was commissioned 
in January 2010 and completed in January 2011 by the project 
owner’s team in conjunction with a team of experienced 
consultants on behalf of Roy Hill Holdings Pty Ltd (RHH). 
The Project is 100 per cent owned by RHH, currently a wholly 
owned subsidiary of HPPL. The following paper outlines the 
deposit and project details.
MINE 
Geology
The Roy Hill deposit is located in the northern part of the 
Hamersley Basin. The Hamersley Basin is approximately 
400 km × 500 km in areal extent and contains up to a 10 km 
thickness of metasediments and metavolcanics that includes 
the banded iron formations (BIF) of the Hamersley Group in 
which the majority of the iron mineralisation has developed 
(Figure 1). The Hamersley Basin or Mount Bruce Supergroup 
is a late Archaean to Palaeoproterozoic platformal cover 
sequence of low grade metamorphosed sedimentary and 
volcanic rocks which unconformably overlies the mid Archaean 
granite – greenstone terrain of the Pilbara Craton (Trendall 
and Blockley, 1970 and Trendall, 1983). The Mount Bruce 
Supergoup comprises the regionally conformable Fortescue 
(ca 2770 to 2630 Ma), Hamersley and Turee Creek Groups. 
The Roy Hill deposit occurs at the base of the Hamersley 
Group in the Marra Mamba Iron Formation (MMIF), which 
conformably overlies the Jeerinah Formation at the top of 
the Fortescue Group (Harmsworth et al , 1990; Thorne and 
Trendall, 2001) (Figure 2). In the Roy Hill mine area the 
MMIF is represented by the lowermost Nammuldi Member 
which consists of pale chert and cherty BIF interbedded with 
thin shales. The trace shown on the right of the MMIF column 
in Figure 2 is the characteristic gamma log trace that re f ects 
the internal shale/tuff horizons within the BIF. The overall 
structure of the Roy Hill project area is relatively simple with 
the stratigraphy dipping very gently at 02° to 5º toward the 
southwest. The Hamersley Group is unconformably overlain 
by the Oakover Formation, a sequence of younger Tertiary 
lacustrine carbonate, silcrete and mudstone rocks that and 
has been deposited in the palaeodrainage of the Fortescue 
Valley. The Fortescue Valley is mantled by a thick (up to 50 m) 
blanket of Quaternary f ood plain alluvial sediments, derived 
from the erosion of the Hamersley and Chichester Ranges.
1. FAusIMM, John Clout and Associates, 41 Hardy Road, Nedlands WA 6009. Email: jmfclout@gmail.com
2. Executive General Manager, Carbon Steel Materials, Hancock Prospecting Pty Ltd, Level 4, 28-42 Ventnor Avenue, West Perth WA 6005. Email: barry.fi tzgerald@royhill.com.au
The Roy Hill Project
J M F Clout1 and B Fitzgerald2
ABSTRACT
The Roy Hill Project is an independent integrated port, rail and mine development in the Pilbara 
region of Western Australia. The bankablefeasibility study for the project was completed in 
January 2011 by Roy Hill Holdings Pty Ltd, a wholly owned subsidiary of Hancock Prospecting 
Pty Ltd. The Project will deliver 55 Mt/a of premium quality iron ore to a rapidly growing market 
with a mine life of over 27 years. This study is based on exploiting a 1.2 Bt bedded resource of 
microplaty haematite-rich Marra Mamba, from which 562 Mt of reserves has been delineated to 
date. A further 1.1 Bt resource of detritals and low grade bedded is proposed to be mined once the 
bedded high grade resource is depleted. Roy Hill lump and f ne ore products will compete strongly 
with other Pilbara products. As a low cost producer, Roy Hill is on track to capture four per cent of 
the global seaborne traded iron ore market. The project decision to proceed is targeted for Q2 2011 
to enable f rst ore on ship by Q2 2014.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
J M F CLOUT AND B FITZGERALD
4
FIG 1 - Location of the proposed Roy Hill mine, rail and port facility.
TUREE CREEK GROUP
FORTESCUE GROUP
Boolgeeda Iron Formation
H
A
M
ER
SL
EY
G
RO
U
P
Woongarra Volcanics
Weeli Wolli Formation
Marra Mamba Iron Formation
Wittenoom Formation
Mount Sylvia Formation
Mount McRae Shale
Joffre Member
Whaleback Shale Member
Dales Gorge Member
Yandicoogina Shale Member
Brockman Iron
Formation
M
et
re
s
0
200
400
600
800
1000
1200
1400
1600
1800
2000
2200
2400
2600
Banded Iron Formation (BIF), cherty BIF
Shale
140
Thickness
(m)
Hamersley
Range Natural
Gamma Log
W
it
te
no
om
Fo
rm
at
io
n
Mo
un
t N
ew
m
an
 
Me
m
be
r
Mt
 M
ac
Le
od
 
Me
m
be
r
Na
m
m
ul
di
 M
em
be
r
50
55
Je
er
in
ah
Fo
rm
at
io
n
Ro
y H
ill 
Sh
ale
Mt Tom Price,
Mt Whaleback
Fe ore deposits
M
ar
ra
M
am
ba
Ir
on
Fo
rm
at
io
n
W
es
t A
ng
ela
 
Me
m
be
r
Hope Downs, Mining Area C,
Marandoo, West Angelas
Iron ore deposits
hosted in Marra
Mamba Iron Formation
Chichester
Range
Roy Hill,
Christmas Creek,
Cloudbreak
0
40
Thickness
(m)
Ro
y H
ill 
Sh
ale
Na
m
m
ul
di
 M
em
be
r
W
it
te
no
om
Fo
rm
at
io
n
BIF
FIG 2 - Stratigraphy of the Hamersley Group (Harmsworth et al, 1990) and the Marra Mamba Iron Formation in the Hamersley Ranges (Lascelles, 2002) versus the 
Chichester Range stratigraphy adapted from Hannon, Kepert and Clark (2005) and the Roy Hill Iron deposit.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
THE ROY HILL PROJECT
5
Bedded iron mineralisation is restricted to the base of the 
Nammuldi Iron Formation and varies from 2 m to 44 m 
thickness, averaging 12 m and 1 km to 7 km wide by 23 km 
strike length (Figure 3). The mineralisation is f at lying and 
dips at between two to ten degrees to the south and south 
west (Figure 4). The central northwest portion of the deposit 
consists of a well developed, thick low phosphorous microplaty 
haematite (mplH) (Figure 5) mineralised core that has been 
overprinted by later supergene martite-brown goethite 
mineralisation. The northwest and southeast portions of the 
deposit consist of a dominant subhorizontal sheet of typical 
supergene martite-brown goethite replacement of BIF with 
minor amounts of ochreous goethite. Vertical and lateral 
ore zonation patterns are similar to those described by Clout 
(2005) and Clout and Simonson (2005) for other Chichester, 
Marra Mamba and Brockman iron ore deposits.
The Roy Hill deposit is believed to be the most well developed 
mplH mineralisation hosted by the MMIF outside of the large 
mplH deposits hosted by the Brockman Iron Formation. 
Mineral resources and ore reserves
More than 8700 vertical RC percussion holes have been drilled 
(over 460 000 metres) plus 269 HQ3 and PQ3 diamond holes 
(over 8500 metres), largely since 2006 (Figure 3). Typical 
drill hole spacing varies between 50 m × 50 m to 400 m × 
400 m with location of RC drill holes given in Figures 3 
and 4. Chemical analysis was by XRF and grade interpolation 
was by ordinary kriging. The mineral resource statement in 
Table 1 has been prepared according to the JORC Code (2004). 
The total resource of 2.3 Bt comprises:
  1.2 Bt of low phosphorus +55 per cent Fe cut-off, mainly 
Marra Mamba bedded iron ore, with some detritals; and
  1.1 Bt of 50 - 55 per cent Fe cut-off low-grade detrital and 
bedded ore resource.
As part of the bankable feasibility study, 562 Mt of reserves 
were de f ned by the independent mining engineers (Coffey 
Mining), suff cient to support the f rst ten years of mine life 
(Table 2). The mining study included examination of all key 
modifying factors including: 
  mining,
  plant recovery (80 per cent),
  economic,
  marketing,
  legal,
  environmental,
  social, and
  government approval considerations. 
The conversion factor from indicated and measured 
resources to reserves was 80 per cent. The resources given in 
Table 1 are inclusive of the reserves in Table 2.
FIG 3 - The Roy Hill deposit.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
J M F CLOUT AND B FITZGERALD
6
Mining 
The Roy Hill orebody has simple stratigraphy enabling a 
conventional, low risk drill, blast, load and haul mining 
strategy. The mineable resource is more than 991 Mt of 
bedded ore (55 per cent Fe cut off grade) with a strip ratio of 
4.56:1. Total material moved will peak at around 340 Mt/a, 
excluding rehandle, with 63.5 Mt/a of run-of-mine (ROM) fed 
to the wet process plant. Mining waste will be returned to the 
pit after the second year of operations, resulting in a smaller 
footprint for rehabilitation. 
Primary and secondary crushing will be carried out at three 
locations near the pit rim, with overland conveying to a 55 kt 
stockpile at the process plant to minimise haulage costs 
(Figure 6). Approximately 70 per cent of the mineable reserve is 
below the water table. A detailed dewatering strategy and water 
management plan has been developed to ensure dry mining and 
to provide the necessary process plant water. No geotechnical 
issues have been identif ed that will impact on pit design.
FIG 5 - Refl ected light photomicrograph of microplaty haematite, RHD 005, 
25.6 m.
Ore type
Resource 
category
Cut-off 
Tonnes 
(millions)
Fe % SiO
2
 % Al
2
O
3
 % P % Mn % LOI % S %
Density 
(g/cc)
Bedded
Measured 55% Fe 143 61.2 4.15 2.21 0.067 0.64 4.36 0.041 2.74
Indicated 55% Fe 505 59 6.01 2.92 0.054 0.88 4.69 0.034 2.73
Inferred 55% Fe 440 58.8 5.67 2.86 0.057 0.79 5.42 0.043 2.74
LG Bedded Inferred 55% Fe 15 55.8 9.92 4.82 0.041 0.46 3.83 0.028 2.67
Detrital 
Indicated 55% Fe 5 56 5.61 4.76 0.042 0.6 7.8 0.046 2.58
Inferred 55% Fe 120 56.2 4.01 4.07 0.053 0.73 9.62 0.06 2.59
Sub-total 55% Fe 1228 58.8 5.51 2.96 0.056 0.8 5.41 0.041 2.72
Bedded Various 50%-55% 160 52.9 11.71 4.48 0.047 1.08 5.61 0.041 2.58
Detrital Various 50%-55% 935 52.5 7.54 5.74 0.053 0.89 9.31 0.058 2.56
Grand total 2323 55.9 6.74 4.18 0.054 0.85 6.99 0.047 2.65
Note: The resource classifi cation for this estimate was prepared according to JORC (2004)
TABLE 1
July 2010 Geological Mineral Resource Estimate for Roy Hill.
 
 
FIG 4 - Cross-section through the Roy Hill Deposit, looking north.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
THE ROY HILL PROJECT
7
Mining will incorporate in-pit blending supported by 
detailed mine planning and grade control drilling. This will 
manage grade variability and enable on-spec product to be 
loaded on trains, using stockpile blending capability only by 
exception.
The Project includes the infrastructure and services to 
support construction and operations including accom- 
modation village, airstrip (Figure 6) and power station. 
Processing
Extensive metallurgical testing has been conducted in 
Australia and overseas to ensure the process f ow sheet has 
been optimised for both above and below water table ore. 
Importantly the f ow sheet is based on technology proven both 
in the Pilbara and internationally and is designed to producelump and f ne ore products. The wet process plant will have 
a nominal capacity of 63.5 Mt/a dry of ROM to produce 
55 Mt/a wet product with an overall dry mass yield of 
80 per cent from ROM.
Crushed ore from the live coarse ore stockpile is processed 
at the plant through a wet front end comprising, six 
wet scrubbers and wet screening into four size fractions 
(Figure 7). Tertiary crushing and screening produces lump 
and f nes product. Material less than 1 mm is wet processed 
and desanded using a f ow sheet similar to that described for 
the nearby Cloudbreak process plant by Clout and Rowley 
(2009) (Figure 8). The desanded f ne product is dewatered 
before transportation to the f nes blended stockpiles. Tailings 
will be thickened and the slurry pumped to a tailings storage 
facility. Up to 46 Ml/day of process water will initially be 
sourced from mine dewatering bores, and later from a remote 
bore f eld. 
A product stockyard of six 150 kt stockpiles will deliver 
blended and sampled lump and f ne products onto trains.
Rail 
The Roy Hill mine will be linked to Port Hedland by a world 
class 342 km, standard gauge, heavy haul railway with 
f ve passing loops, marshalling yard, loops both ends and 
14 bridges. The railway has been designed for 32 000 tonne 
trains of 232 ore cars, 40 tonne axle loads and speeds up to 
80 kph. The railway will incorporate a best practice moving 
block signalling and scheduling system, with in-cab signalling 
and automatic train protection, and electrically controlled 
pneumatic braking to manage in-train forces. The railway 
will be managed by a Train Control Centre located in the 
integrated Remote Operations Centre in Perth.
Classifi cation
Mineral Reserves
Tonnes Fe SiO
2
Al
2
O
3
P
Mt % % % %
Proved 127.3 61.0 4.24 2.26 0.066
Probable 434.3 58.8 6.10 3.03 0.052
Total 561.6 59.3 5.68 2.86 0.055
TABLE 2
Proved and probable ore reserves as of December 2010, prepared 
according to JORC (2004).
FIG 6 - Proposed Roy Hill mine and infrastructure layout.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
J M F CLOUT AND B FITZGERALD
8
Port 
The Roy Hill port facilities will include a world class stockyard 
facility at the Boodarie Industrial Estate in Port Hedland for 
receiving, stockpiling and exporting Roy Hill products using 
the 55 Mt/a port capacity allocated to Roy Hill (Figure 9). The 
port facility will include:
  A two-cell rotary car dumper with two positioners, capable 
of unloading more than f ve trains per day.
  A stockyard with 2.9 Mt live capacity (equivalent to four 
weeks of shipping) and a further 3.6 Mt of dead capacity.
  A lump rescreening plant to remove the less than 
6.3 mm undersize from the f nal lump product prior to 
ship loading.
  A 4 km overland conveyor from the stockyard to wharf 
and marine facilities at Stanley Point. Fines and lump 
buffer storage bins are installed to maintain reclaiming 
operations during hatch changes.
  Two berth wharf designed to cater for two 206 000 DWT 
ships.
  A 12 000 t/h shiploader capable of achieving an average 
total time in port of 107 hours.
The development of the port will require 7.5 Mm3 of dredging 
and is intended to commence in mid-2011. 
REMOTE OPERATIONS CENTRE
The Roy Hill Project will minimise on site labour by 
automating plant and equipment, and controlling operations 
from a Remote Operations Centre (ROC) located near Perth 
Airport. This will allow full operational integration of the 
supply chain.
Desanding Section of Roy Hill Processing Plant
-1mm Fines
Up-Current 
Classifier
Product Dewatering
Spirals
Tailings
Product
Cyclone 
Underflow
Cyclone 
Overflow
Middling
Concentrate
Overflow
Concentrate
Reject
Flowsheet options that give grade/yield flexibility
FIG 8 - Proposed wet desanding fl ow sheet.
FIG 9 - Proposed Roy Hill port stockyard, wharf and berths at Stanley Point, 
Port Hedland.
Fines 
Desanding
At Mine Lump
At Mine Fines
40mm
Reject
8mm
1mm
ROM Tertiary
Crush
Satellite 
Crushing 
Stations
Wet
screening
63.4M dmtpa
12.7M dmtpa
24.9M dmtpa
25.8M dmtpa
Scrubbing
40mm
8mm
FIG 7 - Proposed Roy Hill plant fl ow sheet.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
THE ROY HILL PROJECT
9
At the mine, all mobile equipment will be enabled for 
maximum remote operation and future automation and 
driverless operation. The design of the railway and control 
systems allows for early implementation of Automated Train 
Operation over the next decade. And the majority of port 
operations will also be automated and controlled remotely 
from the ROC in Perth.
PRODUCT QUALITY
Roy Hill lump and f nes are high quality haematite-dominant 
products, with very competitive iron grade, low contaminant 
levels and generally better physical characteristics than 
other Pilbara products (Table 3). The f nal product split on 
ship has been simulated from detailed drop test conditioning 
of PQ and 250 mm wide diameter diamond drill core to be 
60 per cent f nes and 40 per cent lump.
Unlike other Marra Mamba deposits in the Pilbara, Roy 
Hill f ne ore product is composed largely (~60 per cent) of 
the premium microplaty haematite ore type rather goethite 
(Figure 10). Some sections of Roy Hill may be >90 per cent 
mplH. Considering the large diversity in mineral composition 
of the Marra Mamba f ne ore products from goethite-dominate 
to martite-goethite and microplaty haematite rich, it is far 
easier to compare different Marra Mamba and Brockman 
products using the ternary composition diagram in Figure 10.
PROJECT EXECUTION AND NEXT STEPS
A detailed integrated planning schedule has been prepared 
and the key milestones are shown in Table 4:
Funding has been approved for precommitment expenditure 
to fast track the Project to June 2011, to achieve f rst ore on 
Production Dry weight %
Calcined 
weight 
%
Product target grades Mt/a (wet) Fe SiO
2
Al
2
O
3
P Mn S LOI total Fe
Lump product 24 61.0 4.5 1.8 0.060 0.70 0.040 5.0 64.2
Fines product 31 61.0 4.5 2.2 0.055 0.70 0.035 4.6 63.9
Weighted mean 
(Lump + Fine)
55 61.0 4.5 2.0 0.057 0.70 0.037 4.8 64.1
TABLE 3
Lump and fi ne ore product target grades.
Hematite
Hard Brown
Goethite
Ochreous 
Goethite
75
50
25
Roy Hill
Roy Hill Hematite
Brockman N
Brockman P
MM-W
MM-M
MM-F
CID-Y2
CID-J
Brazil-SS
Brazil PF
FIG 10 – Ternary mineralogy plot for various iron ore products from the Pilbara and Brazil. Note the normative mineralogy has been calculated 
from assay and normalised to 100 per cent, with goethite abundance calculated from loss on ignition at 375°C 
and goethite type from polished section microscopy.
Precommitment work including orders for long lead items Q1/2 2011
Commencement of dredging works Q2 2011
Project decision to proceed mid 2011*
Start of mining operations Q3 2013*
First ore shipment Q2 2014*
* Subject to timely approvals and weather.
TABLE 4
Key milestones of the integrated planning schedule.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
J M F CLOUT AND B FITZGERALD
10
ship Q2 2014. The Roy Hill Project was granted ‘Major Project 
Status’ by the Commonwealth Government. Project approvals 
are either in place or well advanced and the few remaining 
primary approvals are expected to be in place by Q2 2011.
SUMMARY
The Roy Hill project bankable feasibility study has been 
completed and early works has commenced in anticipation 
of a decision to proceed planned for the second half of 2011. 
The Roy Hill project will be a large exporter of quality low 
phosphorous lump and f ne ore products. The project will 
benef t from independent infrastructure from mine to port.
ACKNOWLEDGEMENTS
Hancock Prospecting Pty Ltd would like to acknowledge 
the tireless efforts of its current and former employees as 
well as strategic project partners POSCO and STX of South 
Korea. Numerous engineering contractors contributed to the 
bankable feasibility study including mine planning and ore 
reserves (Coffey Mining),mine process plant (Lycopodium 
Minerals), rail and port landside (WorleyParsons), mine 
infrastructure (GHD) and water management (MWH).
REFERENCES
Clout, J M F, 2005. Iron formation-hosted iron ores in the Hamersley 
Province of Western Australia, in Proceedings Iron Ore 2005 
Conference, pp 9-19 (The Australasian Institute of Mining and 
Metallurgy: Melbourne).
Clout, J M F and Rowley, W G, 2009. The FMG story – From 
exploration to the third largest iron ore producer in Australia, in 
Proceedings Iron Ore 2009 Conference, pp 3-10 (The Australasian 
Institute of Mining and Metallurgy: Melbourne).
Clout, J M F and Simonson, B M, 2005. Precambrian iron formations 
and iron formation-hosted iron ore deposits (eds: J W Hedenquist, 
J F H Thompson, R J Goldfarb and J P Richards), Economic 
Geology 100th Anniversary Volume 1905-2005 , pp 643-679 
(Society of Economic Geology).
Hannon, E, Kepert, D A and Clark, D, 2005. From target generation 
to two billion tonnes in 18 months – The reinvention of the 
Chichester Range, in Proceedings Iron Ore 2005 Conference , 
pp 73-78 (The Australasian Institute of Mining and Metallurgy: 
Melbourne).
Harmsworth, R A, Kneeshaw, M, Morris, R C, Robinson, C J and 
Shrivastava, P K, 1990. BIF – Derived iron ores of the Hamersley 
Province, in Geology of the Mineral Deposits of Australia 
and Papua New Guinea (ed: F E Hughes), pp 617-642 (The 
Australasian Institute of Mining and Metallurgy: Melbourne).
JORC, 2004. Australasian Code for Reporting of Exploration 
Results, Mineral Resources and Ore Reserves (The JORC Code) 
[online]. Available from: <http://www.jorc.org> (The Joint Ore 
Reserves Committee of The Australasian Institute of Mining and 
Metallurgy, Australian Institute of Geoscientists and Minerals 
Council of Australia).
Lascelles, D, 2002. A new look at old rocks – An alternative model 
for the origin of in situ iron ore deposits derived from banded 
iron formation, in Proceedings Iron Ore 2002 , pp 107-125 (The 
Australasian Institute of Mining and Metallurgy: Melbourne). 
Thorne, A M and Trendall, A F, 2001. Geology of the Fortescue 
Group, Hamersley Basin, Western Australia. GSWA, Bull 144.
Trendall, A F, 1983. The Hamersley Basin (eds: A F Trendall and R C 
Morris) Iron-Formation: Facts and Problems, Developments in 
Precambrian Geology, 6:69-129 (Elsevier).
Trendall, A F and Blockley, J G, 1970. The iron formations of the 
Precambrian Hamersley Group, Western Australia, with special 
reference to the associated crocidolite, GSWA, Bull 119.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 11
INTRODUCTION
The Société Nationale Industriel et Minière (SNIM) is 
expanding its ore and concentrate production capacity in 
Mauritania, West Africa. This government-controlled mining 
company has been in production since 1963 and currently 
exports some 12 Mt/a. The operations currently consist of 
mines producing direct shipping ore in three locations TO14, 
Mahoudat and Rouessa and an open pit magnetite mine and 
concentrator called Guelb el Rhein.
SNIM is presently doubling its magnetite concentrate 
production at the Guelb el Rhein mine. An existing processing 
line of 4 Mt/a will have a second parallel production line of the 
same capacity added. The mine is located some 600 km inland 
from the Atlantic coast in the heart of the Sahara desert. 
Electricity is generated at the site using diesel generators 
fuelled by heavy fuel oil. Water is provided by a borehole fi eld 
60 km away. The site is serviced by a heavy haul railway which 
takes the products to the port of Nouadhibou on the Atlantic 
coast (Figure 1).
The existing magnetite operation uses primary crushing, 
Aerofall type dry Semi-Autogenous Grinding (SAG) mills, 
screening at 1.6 mm and magnetic separation to produce 
a concentrate containing around fi ve per cent silica. The 
concentrate is transported to the port by a railway owned and 
operated by SNIM. 
The original installation has had some additions over the 
years. A gravity separation plant uses spiral concentrators 
to recover haematite from portions of the orebody. High 
pressure grinding rolls (HPGRs) have been added as pebble 
crushers in the SAG milling circuit. Recently, a wet magnetic 
separation plant has been added to recover concentrate from 
fi ne dust collected from the electrostatic precipitators in the 
Aerofall circuit.
The Guelb II project consists of a new processing line to 
recover magnetite, and will include new stockpiles, tailings 
disposal facilities and a train loading station. The electricity 
generation capacity is being increased, a new borehole fi eld is 
being added and other infrastructure is being added. 
SNC-Lavalin is managing the Guelb II project as part of an 
Engineering, Procurement and Construction Management 
(EPCM) contract. The open pit mine expansion is being 
managed by SNIM.
The ore is a banded iron formation consisting of magnetite, 
quartz and minor amounts of other minerals. There is some 
haematite present which is selectively mined for processing 
separately from the magnetite. The magnetite ore is classifi ed 
by the operators into two grades – TS1 with a liberation size of 
>400 microns, and TS2 with a liberation size of <400 microns. 
The Guelb II project will treat a blend of equal amounts of the 
two types with a liberation size of 400 microns and 96 per cent 
of the iron present as magnetite. The properties of the ore are 
fairly typical with a crushing work index of 10 kWh/t and a 
ball mill work index of 18 kWh/t.
1. Senior Process Specialist, Metallurgy, Mining and Metallurgy Division, SNC-Lavalin Inc, 455 René Lévesque Boulevard, West Montreal, Quebec H2Z 1Z3, Canada. Email: gordon.cooper@snclavalin.com
2. Process Engineer, Mining and Metallurgy Division, SNC-Lavalin Inc, 455 René Lévesque Boulevard, West Montreal, Quebec H2Z 1Z3, Canada. Email: remi.lapointe@snclavalin.com 
3. Project Area Manager, Mining and Metallurgy Division, SNC-Lavalin Inc, 455 René Lévesque Boulevard, West Montreal, Quebec H2Z 1Z3, Canada. Email: jean.routhier@snclavalin.com
The SNIM Guelb II Project in 
Mauritania – Process Design Challenges
G Cooper1, R Lapointe2 and J Routhier3
ABSTRACT
The Société Nationale Industriel et Minière (SNIM) is expanding its ore and concentrate production 
capacity in Mauritania, West Africa. This government-controlled mining company has been in 
production since 1963 and currently exports some 12 Mt/a.
SNIM is presently doubling its magnetite concentrate production at the Guelb el Rhein mine. An 
existing processing line will have a second parallel production line of 4 Mt/a added. The mine is 
located some 600 km inland from the Atlantic coast in the heart of the Sahara Desert.
The metallurgy of the ore is reasonably straightforward. It is a magnetite banded iron formation 
that is amenable to magnetic separation. While the existing plant uses semi-autogenous grinding 
and dry magnetic separation, the new processing line will use HPGRs, dry magnetic separation, 
and a wet milling and magnetic separation circuit for treating middlings.
The process choices refl ect the scarcity of water in the region, the high cost of locally generated 
power and the operating experience and preferences of SNIM. Dust control in the desert 
environment has been a particular challenge.
The metallurgical design aspects as well as the infl uence of the environmental and commercial 
aspects of the project are presented.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
G COOPER, R LAPOINTE AND J ROUTHIER
12
The plant extends over a distance of 3 km from the primary 
crusher to the train loading station.
The offi cial languages in Mauritania are Arabic and French. 
The project is being executed in French, and all documentation 
is in the French language.
The project CAPEX is estimated at US$630 M and the iron 
ore production is scheduled to start in fi rst quarter of 2013. 
The project offi ce is located in Montreal and a construction 
site office is located in Zouérate.
All process units were procured via an Engineering, 
Procurement and Services (EPS) contracting packaging 
strategy. This EPS strategy was imposed by SNIM to maximise 
the vendor responsibilities during the project engineering, the 
construction and start-up of the process units. 
The EPS strategy requires the vendors to validate the design 
concepts and take responsibility for the detail engineering, to 
supply the mechanical equipment, some of the bulk materials 
and instrumentation required for the complete operation of 
the plant. The EPS packages also include process performance 
guarantees such as throughput, particle size distribution, 
silica content, iron content, dust capture, etc for a 72 hour test 
executed after the mechanical completion phase. 
The EPS approach was part of the fi nancing strategy 
considering diffi culties in previous SNIM projects to meet the 
design performance.
The main EPS packages were divided as follows:
  crushing plant (to an equipment supplier), 
  crush ore stockpile,
  dry concentrator (to an engineering/integrator fi rm),
  wet concentrator (to an equipment supplier),
  concentrate stockpile,
  reject stockpile, and
  train loading.
A prequalifi cation period of three months was used to 
evaluate and identify the bidders for each package. Following 
the request for tender, nine months passed during the 
procurement cycle of request for bids, evaluation and issuing 
the purchase orders. All EPS packages include liquidated 
damages associated with the performance tests and equipment 
delivery schedules.
The EPCM mandate was centred around the pre-engineering 
required to prepare the EPS tender packages, evaluation, award 
and contract follow-up and construction management. The 
EPCM mandate included the civil, structural, electrical, and 
control systems engineering as per EPS supplier requirements.
The EPCM mandate includes the preparation, evaluation, 
award and management of a single site erection package to 
build the new Guelb II plant. The objective is to manage the 
general contractor with an integrated team (SNC-Lavalin and 
SNIM) under the leadership of SNC-Lavalin.
FIG 1 - Project location.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
THE SNIM GUELB II PROJECT IN MAURITANIA – PROCESS DESIGN CHALLENGES
13
An EPC contract was also awarded to extend the existing 
power plant with two new 10 MW diesel generators.
PROCESS SELECTION
The main factors affecting the process selection are:
  dry processing as much as possible to conserve water,
  SiO
2
 limits in concentrate,
  high electricity cost because of on-site generation,
  conservative approach by SNIM to new technology,
  selection of equipment and practices that is familiar to the 
plant operators, and
  water quality available from the boreholes.
The basic process was established in an early study. After 
small scale exploratory testing, pilot testing was done on 
samples of 10 tonnes taken from the mine. The resulting fl ow 
sheet and design parameters were carried forward to the 
preliminary design (Figure 2).
A primary gyratory crusher and closed-circuit secondary 
crushing circuit is used to reduce the ore size to less 
than 50 mm.
The HPGRs feed medium intensity magnetic separators 
(MIMS) to eliminate about 36 per cent of the feed material 
as tailings. The magnetic fraction is screened at 1.6 mm to 
recirculate the oversize material to the HPGRs. The minus 
1.6 mm material is fed to two stages of low intensity magnetic 
separators (LIMS) to produce concentrate (34 per cent of 
feed), tailings (13 per cent of feed), and middlings fraction (17 
per cent of feed).
The dry concentrator produces about 80 per cent of the 
concentrate and 80 per cent of the tailings.
The middlings fraction goes to a wet ball mill circuit with 
a 180 micron screen to close the circuit. This feeds 3-stage 
magnetic separators to produce more concentrate (55 per cent 
of feed), and tailings (45 per cent of feed).
The wet concentrate goes to a storage tank and is then 
fi ltered on a belt fi lter and joins the dry concentrate.
The tailings are thickened and fi ltered on a belt fi lter and 
joins the dry tailings.
The process packages were sent out for bids to qualifi ed 
vendors along with new samples of ore from the mine. These 
packages specifi ed the major process equipment types and 
the circuit confi gurations. Process guarantee specifi cations 
were included. Given the limited time for bidding and the 
vendors’ budgets for preparing bids, a very limited amount of 
metallurgical testing was performed during the tender period 
by the vendors to provide process guarantees. Consequently 
the vendor responses were naturally cautious with respect to 
performance guarantees and equipment sizing. All bids for 
the dry concentration process provided process guarantees 
that were subject to further testing.
In the case of the dry concentrator, further testing after the 
contract award produced some different results in the dry 
concentration circuit. It was found to be diffi cult to respect 
the performance guarantees and to replicate the results of the 
initial test work in the dry magnetic separation. Signifi cantly, 
the MIMS separation on the HPGR product could not reject as 
much tailings material as the original pilot testing had shown. 
The fi nishing LIMS stage of the magnetic separation also 
showed poor results. The original tests produced concentrate 
with less than 4.7 per cent of silica while the new test results 
indicate higher than 6.5 per cent of silica. The equipment 
supplier was unwilling to guarantee the performance of the 
original fl ow sheet.
Another issue is a technical risk in the use of MIMS as 
compared to the use of LIMS. LIMS equipment is common 
in the iron ore industry around the world and achieves good 
effi ciency. LIMS is also used in the existing SNIM plant at 
Guelb el Rhein. On the other hand, MIMS equipment is still 
relatively new in this application.
FIG 2 - Original fl ow sheet.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
G COOPER, R LAPOINTE AND J ROUTHIER
14
The concentrate quality results can be explained by the 
fact that the dry magnetic separation is not effi cient for the 
very fi ne material (less than 50 microns). During the lab 
tests, dust removal were done many times on the material 
tested apparently resulting in the removal of high SiO
2
 fi nes 
to the dust collector. In order to avoid contamination of 
the concentrate by the high level of silica, a re-design was 
performed. It was decided to add air separators to remove 
the SiO
2
 fi nes and to treat this fi ne material (representing 
two per cent of the feed) in the wet concentrate plant where 
wet magnetic separation is effi cient for the fi ne material. 
In the fi nal fl ow sheet dust will be removed from the material 
before production of the fi nal concentrate, and this dust will 
be treated in the wet plant and all magnetic separation will be 
made by LIMS and no MIMS will be used (Figure 3).
The changes to the dry concentration have eliminated the 
MIMS separation on the HPGR product. The circuit now uses 
a type of air classifi er on the <1.6 mm screen underfl ow and 
only uses LIMS in three stages. 
The HPGR product is screened into three fractions. The 
fraction over 10 mm is recycled to the HPGR feed bin. The 
1.6 - 10 mm fraction passes through a LIMS separator to reject 
a tailings stream and the magnetic fraction is returned to the 
HPGR. The minus 1.6 mm fraction goes to rougher, cleaner and 
scavenger stages of LIMS. The rougher stage rejects a tailings 
stream. The cleaner stage produces fi nished concentrate and an 
intermediate stream which is treated in the scavenger stage to 
produce a middlings stream and more tailings.
The fi rst stage magnetic separators are treating material with 
narrower size ranges than in the original concept. The MIMS 
in the original design was treating materialranging from 
fi nes up to nearly 50 mm. This has changed to size ranges of 
0 - 1.6 mm and 1.6 - 10 mm. In principle, this should provide 
better separation.
DUST CONTROL
Control of dust is a particular concern to SNIM. The existing 
operation has problems with dust emissions and dust in the 
workplace environment. There are concerns that have arisen 
due to silicosis among the employees as a result of the dusty 
environment in the plant.
The main constraints and principles for the design of the 
dust control system are:
  minimise the use of water;
  minimise the amount of regular inspection and 
maintenance required;
  avoid returning dust to the dry process system once it has 
been removed;
  robust, reliable, easy to maintain;
  recovery of iron where practical; and
  separation of parallel production lines.
The limit for ambient levels of respirable silica in the 
workplace as mandated by the World Bank’s International 
Finance Corporation is 0.025 mg/m3. During strong winds a 
value of 0.060 mg/m3 has been measured in the region. Higher 
levels have been measured within the existing plant. The 
international development banks involved in fi nancing of the 
project have mandated the adoption of the World Bank standard 
of 0.025 mg/m3. While the standard is very low compared to 
the ambient levels in the Sahara desert, the basic principle is 
that the plant must meet the best international practices.
The standard for stack emissions at 50 mg/m3 is not 
diffi cult to meet. Conventional baghouses can easily meet that 
performance standard. This emission limit dates back many 
years and it appears likely that it will be revised downward in 
the near future.
The approach to dust collection design has followed the 
guidelines for no adjustments and ease of maintenance. 
This led to the design where there are no control dampers 
in the system and no more than six pickup points for each 
baghouse. This is in contrast to the common approach where 
a large number of collection points are connected to one large 
baghouse per building. 
To provide good air fl ow without fl ow control dampers, non-
standard duct diameters are used to balance the fl ows between 
the branches of the system. Flow balancing is therefore 
achieved at the design stage and is not left to the discretion of 
the operators.
Wherever possible, parallel production lines or pieces 
of equipment each have their own baghouse. This avoids 
imbalance of fl ows when one line is out of service. This also 
permits a baghouse to be serviced when a production line is 
down for repairs.
Welded construction rather than bolted construction is used 
for baghouses and ductwork. This is to prevent in-leakage 
in the equipment when it ages and could otherwise loosen 
up over the long-term due to thermal cycling and vibration. 
The ductwork design avoids horizontal ducts. All ducts are 
either sloping up or down at 60o to prevent accumulation of 
dust along the bottom and the consequent loss in fl ow. This is 
considered to be the best approach in an environment where 
routine inspection and cleaning is diffi cult.
FIG 3 - Dry concentration – fi nal fl ow sheet.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
THE SNIM GUELB II PROJECT IN MAURITANIA – PROCESS DESIGN CHALLENGES
15
Conveyors are not covered outside of the process buildings. 
This is a choice by SNIM resulting from past experience. 
Instead, conveyor speeds are limited to 2.5 m/s. It is accepted 
that some material may blow off the conveyors during high 
winds. However, these high winds also carry sand off the 
desert which is a much larger source of silica in the air.
Conveyor transfer points are designed with extended belt 
skirts and enclosures to permit the material on the belts to 
settle properly before exiting from the ventilated area.
The result of taking this approach is that there are over 60 
individual baghouses. Each of these baghouses has a fan and a 
stack and is regulated as an emission point. Most of them are 
relatively small units in the range of 10 000 - 20 000 m3/h. 
The negative side of this selection is that the plant has a large 
number of fans to maintain. However, these are small units 
which are easy to replace and less likely to have problems with 
impellor balance. Variable speed drives are used instead of 
dampers on fans over 35 kW to regulate fl ow. These do not 
have moving parts and so will be more reliable.
Water sprays are not used at conveyor transfer points. The 
quantity of transfers and their widely separated locations has 
led to the conclusion that this would create a situation that 
would be diffi cult to maintain.
Dust generation on the mine haul roads and the in-plant 
roads is reduced by spraying with the waste brine from 
the plant’s reverse osmosis water purifi cation unit. This is 
supplemented by limited amounts of raw water when required. 
Dust control on the stockpiles and the tailings pile is the 
most diffi cult source to control. Water sprays for limited use 
are included in the current design. It remains to be seen how 
effective these will be. The concentrate and tailings are blends 
of about 80 per cent dry material and 20 per cent moist fi lter 
cake which will give four per cent of moisture in the mixed 
material. Experience in the existing plant has shown this to be 
very effective in suppressing dust generation at the conveyor 
transfer points and the discharge point. There is however 
considerable dust generation when material ‘avalanches’ 
down the slopes of the piles. This mainly occurs after some of 
the moisture has evaporated from the piles.
The stackers for crushed ore and concentrate are 
conventional design with automatically controlled luffi ng 
booms to limit dust generation.
The tailings stacking system is the leach pad type stacker 
using a long conveyor travelling sideways on crawlers with a 
tripper and discharge conveyor moving along its length. The 
discharge conveyor is reversible to stack ahead or behind 
the travelling conveyor. The stacking height is designed for 
15 m ahead of the travel (down-stacking) and 7 m behind (up-
stacking) and therefore minimises the avalanche effect. It also 
means that the dust is generated in a location that is relatively 
sheltered from the wind. Water sprays are not used (Figure 4).
DUST TREATMENT
Samples of dust taken from the existing plant showed that 
wet magnetic separation can easily produce a good quality 
concentrate. This equates to a small but signifi cant portion of 
the total production. The dust is sent to the wet concentration 
plant and combined with the ball mill product as feed to the 
magnetic separation.
Transporting the dust to the wet concentration plant 
presents some challenges. The collection locations are spread 
over a length of 1.7 km (excluding the product conveying and 
loading), the quantities are hard to predict and the fl ows are 
all intermittent.
A number of options were considered:
  pneumatic conveying,
  trucking in pneumatic discharge trailers,
  trucking in concrete mixing trucks, and
  slurrying and pumping.
After a thorough review, the option selected is the slurry 
pumping system. A circuit with a continuous fl ow of water 
in a loop around the plant is used. The collected dust is 
added at several points in mix tanks and pumped to the wet 
concentration plant. The slurry contains variable amounts 
of solids in a fi xed amount of water and is discharged into 
a thickener. The thickener overfl ow is returned to the 
circulation loop. The thickener underfl ow joins the feed to the 
wet magnetic separation (Figure 5).
PROJECT MANAGEMENT ISSUES
Health and safety
HSE plan and objectives of zero tolerance was a challenge. The 
HSE culture was developed during the site preparation phase 
with the client and the local contractor involved. A training 
program was prepared by type of construction activities and 
presented to all workers.
Logistics
The logistics for the project will be a significant challenge. The 
traffi c through the port of Nouadhibou will be triple its normal 
level during a period of six months from November 2011 to May 
2012. Additional storage facilities and coordination teams have 
been planned to achieve three months total delivery time to site. 
Although all equipment and construction tender packages 
were prepared in French there was a positive market response. 
The major equipment suppliers have all submitted a complete 
offer in French and accepted to execute the contract in French. 
However, all day to day communication with the vendors is 
conducted in English.
Long procurement schedule
The procurement schedule is longer than under a typical 
EPCM project execution for a number of reasons:
  A preliminary plant design in substantial detail is required 
as a basis for the bid packages so all the vendors provide 
costs on a comparable basis.
  An extended period is required for preparing detailed bid 
package specifi cations that match the capabilities of the 
vendors.
  Process plant bidders must perform in-house testing, 
process design, subcontract bidding and cost estimating 
before producing fi rm bids. This cannot be easily done in 
three months and is also very costly to the vendors. These 
costs are indirectly refl ected in the fi nal project overall 
cost.
  The bid evaluation time is longer than for bare equipment 
purchases because of the amount of detail required and 
the requirement for process guarantees.
  Detailed engineering for buildings, electrical work, etc 
cannot proceed until the major packages are awarded and 
vendor engineering is well advanced.
Management of Engineering, Procurement and 
Services packages
Management of the EPS packages has proven to be diffi cult 
to perform for some of the engineering deliverables such as 
P&IDs. In general, the vendors have underestimated the time 
required for design, review, Hazop review and coordination. 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
G COOPER, R LAPOINTE AND J ROUTHIER
16
Standardisation of equipment, documentation and design 
approach is diffi cult to achieve in all aspects.
Performance guarantees are diffi cult to negotiate with limited 
time available, limited samples and laboratory availability. 
The vendors are unwilling to guarantee performance based on 
third party test work results. 
The performance guarantees oblige the equipment vendors 
to increase their design factors and potentially increase the 
cost of the project. This is accepted as part of the price paid for 
the process guarantee. As an example, slurry piping could be 
overdesigned for high fl ow rates leading to blockage at normal 
fl ow rates. Suitable compromises must be agreed to make the 
plant operable and cost-effective.
CONCLUSIONS
The Guelb II project presents some unique challenges:
  limited water supply requiring dry processing as much as 
possible;
  dust emission control on dry ore with a minimum of water 
consumption;
  control of ambient levels of silica in the workplace, given 
the desert environment;
  contracting strategy using complete, process-guaranteed 
equipment packages; and
  construction in a remote site.
A fi nal overall conclusion concerns the use of process-
guaranteed equipment and engineering packages. This 
approach implies certain cost increases but also signifi cantly 
extends the project schedule, mainly on account of test work 
required for process guarantees.
FIG 5 - Dust collection and treatment.
FIG 4 - Tailings down-stacking, up-stacking and prevailing winds.
PREVAILING 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 17
The Benefi ts of Computational Modelling in 
Improving Iron Ore Production and Processing
M P Schwarz1 and P Cleary2
ABSTRACT
A major issue presently faced by companies producing and processing iron ore is the need to maximise 
throughput while minimising capital expenditure. While present high demand means this is of prime 
importance, the perennial issues of maintenance, operating costs, reliability, quality and environmental 
sustainability are also critical to profi tability. Attempts to address these issues generally require 
modifi cations to existing equipment, or trials of new equipment when designing plant for new projects. 
The risk involved can be substantially minimised by testing equipment improvements using advanced 
computational models prior to implementation. Examples of application of Computational Fluid Dynamics 
(CFD) and Discrete Element Method (DEM) modelling to sampling, cyclones, conveyor transfers, 
classifi ers, dust control, banana screens, comminution and smelt-reduction will be given to illustrate the 
benefi ts of such an approach.
1. MAusIMM, CSIRO Mathematics, Informatics and Statistics, CSIRO Minerals Down Under Flagship, Private Bag 312, Clayton South Vic 3169. Email: phil.schwarz@csiro.au
2. CSIRO Mathematics, Informatics and Statistics, Private Bag 33, Clayton South Vic 3168. Email: paul.cleary@csiro.au
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 19
Dewatering Marra Mamba – Lessons Learned 
K Sommerville1 and B Douglas2
ABSTRACT
BHP Billiton’s Area C iron ore operation has been in production since 2003. During 2010, below water table 
mining in the Marra Mamba formation commenced and dewatered ore has been successfully processed 
through the plant.
The ability to mine below water table material increases reserves. Pits can be dewatered using a 
combination of ex-pit and in-pit bores. Critically it is important to allow suffi cient lead time to dewater. For 
Area C, in-pit bores yield the most water which presents challenges in every day operations when mining 
around bores. Operationally there are steep learning curves in obtaining equipment, training personnel 
and setting up infrastructure. Core risks include lifting, refuelling and maintaining gensets. The pipe-
work, extraction and re-establishment work can be very manual. 
The challenges include blasting, mining, stockpiling, processing, water balance and disposal. Maximising 
up time on bores whilst minimising impacts to blast practices have evolved. Blast practices now align bores 
to the edge of patterns. Protecting bores have evolved from stemming bores with rock and re-drilling to 
using gas bags which require no drills. Based on moisture and material properties, mining of dewatered 
material can be categorised into three groups:
1. normal, 
2. sheeting, and
3. sheeting and top loading.
Ore can go straight to the plant ore to drying pads where it takes between one and three months to drain. 
Stock management of these piles is important to ensure effi ciency of the plant. Rules have been established 
for blending wet material to maintain feed capacity and some modifi cations to plants have been made.
Dewatering of pits poses a challenge for water use. Initially turkey’s nests were established to capture 
water and use for dust suppression. As water abstraction has increased methods of management include 
using exhausted pits for storage and evaporators. Managed Aquifer recharge and use in potable water are 
being considered. An over arching coordination of the capacity and use are required.
Important elements in dewatering include: 
  capacity models, 
  capital, 
  long to short planning, 
  day to day coordination, and 
  water level monitoring and balances. 
Success in dewatering requires a cross function team working towards a common goal. 
1. FAusIMM(CP), Mine Manager, Area C, 225 St Georges Terrace, Perth WA 6850. Email: kate.m.sommerville@bhpbilliton.com
2. Water Resources Manager, 225 St Georges Terrace, Perth WA 6850. Email: Blair.R.Douglas@bhpbilliton.com
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 21
INTRODUCTION
The Japanese steel industry has a long history of introducing 
innovative technologies in the fi eld of iron making. 
Technologies adopted during the past ten years include: 
  technologies to use cheaper and lower-grade raw materialsand fuels, 
  measures to prolong the service life of blast furnaces and 
coke ovens,
  promotion of energy saving, and 
  solutions to environmental problems. 
This report outlines the recent technical developments in 
iron-making and CO
2
 reduction technologies being developed. 
RECENT IRON PRODUCTION AND CHANGES 
IN NUMBER AND CAPACITY IN BLAST 
FURNACES 
The Japanese productions of crude steel and pig iron were 
about 100 M and 80 Mt, respectively, from the mid-1970s to 
end of the 1990s. From the start of the 2000s, the production 
gradually increased to meet the economic growth of the 
BRICS. However, last year's production fell widely due to the 
worldwide economic crisis (Figure 1).
The number of blast furnaces has been reduced from 65 in 
1971 to 28 in 2010, as a result of increasing the production 
capacity of each blast furnace by increase in its inner volume 
(Figures 2 to 4).
THE PRETREATMENT TECHNOLOGIES 
AGAINST THE DEGRADATION OF IRON ORE 
AND COAL RESOURCES
Changes in the iron ore resources 
During the 1990s, the Japanese steel mills increased imports 
of Australian iron ores to reduce costs. In the 2000s, they 
gradually began increasing the potion of Brazilian ores 
to reduce Al
2
O
3
 content in sinter to meet needs of high 
productivity operation of their blast furnaces (Figure 5).
Types of the Australian iron ores have changed from hard 
haematite (Brockman) to goethite (Pisolite) and goethite 
haematite (Marra Mamba). The average grain size of sinter feed 
has decreased as ratio of Marra Mamba increased (Table 1).
1. Director, Iron-making Technical Division, Nippon Steel Corporation, 6 - 1 Marunouchi 2-Chome, Chiyoda-ku, Tokyo 100-8071, Japan. Email: miwa.takashi@nsc.co.jp
Recent Iron Making Topics in Japan
N Takamatsu1
ABSTRACT
The environments surrounding the iron and steel industry have changed greatly in recent years. 
Reported in this paper are the technologies adopted by the Japanese iron and steel industry to fi t 
itself to change in global steel demands, degradation of iron ore and coal, prolonging service life 
of the facilities and requirements of energy saving. In addition, innovative technologies for CO
2
 
emission control in iron making processes are also discussed. 
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
19
51
19
55
19
59
19
63
19
67
19
71
19
75
19
79
19
83
19
87
19
91
19
95
19
99
20
03
20
07
20
08
4
6
20
09
4
20
09
8
20
09
12
20
10
4
Pr
od
uc
ti
on
(k
t/
)
Pig Iron
Crude Steel
P '08FY(q) '09 10FY(m)
S '08FY(q) '09 10FY(m)
FIG 1 - Trend of pig iron and crude steel production of Japan.
0
10
20
30
40
50
60
70
0
1000
2000
3000
4000
5000
6000
7000
19
45
19
53
19
57
19
61
19
65
19
69
19
73
19
77
19
81
19
85
19
89
19
93
19
97
20
01
20
05
20
09
Max. IV
Min. IV
Ave. IV
Number of BF
Callender year
B
la
st
Fu
rn
ac
e
IV
N
um
berofB
lastFurnaces
(n)
FIG 2 - Number and size distribution of blast furnace in Japan.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
N TAKAMATSU
22
Process of iron ore pretreatment technologies
Degradation in size of sinter feed results in lower productivity 
of sintering machines. To compensate the loss of productivity, 
several types of pretreatments for the sintering process have 
been applied.
JFE Steel Corp developed a new granulation process for the 
sinter feed treatment (Oyama et al, 2005). The technology 
controls melting behaviours of iron ore and limestone by 
coating the appropriate amount of coke breeze and limestone 
on the pseudo-agglomerates of iron ore. The process improves 
productivity of a sintering machine and the sintered ore's 
reducibility.
The process is applied into four sintering machines at the 
West Japan Works of JFE Steel Corp and achieves an annual 
capacity of 19 Mt (Figure 6). 
MEASURES FOR PROLONGING SERVICE LIFE 
OF FACILITIES
Prolonging of blast furnace life
Carbon bricks have been used for the blast furnace hearth 
refractory for 50 years. Nippon Steel Corp has improved 
carbon bricks for long years (Nitta, Nakamura and Ishii, 2008) 
(Figure 7). A brick of the blast furnace hearth needs high 
corrosion resistance and thermal conductivity to control the 
self-protective layer. The carbon brick recently developed by 
Nippon Steel Co is high in corrosion resistance. It is designed 
to form an adhesion layer at the boundary between the carbon 
brick and molten iron. 
Extension of coke oven life
A technology is being developed to extend service life of coke 
ovens. In Japan, many of the coke oven batteries were built 
‘92 ‘93 ‘94 ‘95 ‘96 ‘97 ‘98
Nagoya #1
Kimitsu #2
Muroran #2
Kimitsu #3
Oita #2
4,650m3 PW
4,250m3 Bell
4,822m3 PW
3,890
PW 4,403
PW 4,063
Nagoya #3
4,884m3 BellBell 4,158Oita #1
Tobata#1 #4
Kimitsu #4
‘90 ‘91 ‘99 ‘00 ‘01 ‘02 ‘03 ‘04 ‘05
3,273m3 PWBell 2,884
4,300m3 PWPW 3,424
‘07 ‘09‘08‘06 ‘10 ‘11 ‘12
2,902m3 PWPW 2,296
5,555m3 PWPW 5,151
5,775m3 BellBell 5,245
’92/5
’93/5
’94/11
’98/2
’00/4
’01/5
’01/11
’03/5
’04/5
’07/4
’09/8
’?/?
actual plan
Furnace inner volume 
enhancement
35,514 m3
42,095 m3
3,946 m3
4,677 m3/unit
The largest BF
in the world
5,775m3 Bell
5,443m3 PW
FIG 3 - Size enlargement of blast furnace (NSC).
FIG 5 - Changes in the iron-ore sources used by Japanese steel industry 
(1971 - 2007).
Inner Volume
The 2rd
campaign
The 3rd
campaign
Throat diameter
Belly diameter
Hearth diameter
Tuyeres
Tapholes
Maximum production
design
record
5,245m3
10.5m
16.6m
14.9m
40
4
12,550 t/d
13,368 t/d
5,775m3
11.1m
17.2m
15.6m
42
5
13,500 t/d
13,810 t/d
Working Volume 4,312m3 4,753m3
132 days 79 days
Throat Cast stave Cast stave
Relining days
Cast stave Copper+cast stave
hearth Cast stave
High-erosion-
resistance type 
carbon brick
Hearth brick Carbon brick
Cast stave
shaft 14900 15600
17200
10500 11100
I.V.5775m3
W.V.4753m3
O2R(2) O2R(3)
16600
I.V.5245m3
W.V.4312m3
H h
FIG 4 - Biggest blast furnace in Japan (NSC Oita #2).
A B C (2007) D E
Ore type Brockman Brockman
Marra 
Manba
Pisolite Pisolite
T.Fe % 62.40 61.80 61.78 56.40 58.30
AI
2
O
3
 % 2.47 2.24 2.26 2.79 1.52
-0.15 % 12.10 13.10 28.08 3.20 4.58
L0 I % 3.20 5.20 5.98 9.40 10.00
TABLE 1 
Alumina component, fi ne particle ratio and LOI of typical iron ores 
(2008 company data).
FIG 6 - Process fl ow of limestone and coke breeze coating granulation 
technology.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
RECENT IRON MAKING TOPICS IN JAPAN
23
around 1970, and their average age is 36 years. Coke ovens 
tend to be lower in productivity due to more frequency of high 
pushing loads and more through holes on the walls as they 
get older. 
To overcome this problem, Nippon Steel Corp has developed 
a technology for diagnosing and repairing damaged fi rebrick 
walls in the coking chambers quite precisely and effi ciently. 
The innovated machine, Doctor of Coke Oven (DOC), is of a 
technology of combination of mechanical engineering, optical 
sensing and data analysis and processing (Figure 8).
MORE EFFICIENT AND MORE STABLE 
OPERATION 
Development of hybrid bonded lumpy ore for 
blast furnaces
Kobe Steel, Ltd developed a technology of a new type of feed 
material for a blast furnace (New Energy and Industrial 
Technology Development Organization, 2007) (Figure 9). 
With this technology, the agglomerates, new blast furnace 
feedstock, are made from low quality ore with help of coking 
property of mixed coal. Carbon remained in the agglomerates 
functions as an auxiliary reducing agent in a blast furnace. Its 
aims are to use iron ore with various qualities and to reduce 
CO
2
 emission. Targets of the research were to make a lumpy 
material with strength enough to be used in a blast furnace, 
and to obtain an appropriate control of burden distribution to 
materialise its highly reactive effi ciency fully in a blast furnace. 
The target of energy conservation is 72 000 kiloliters/year ofcrude oil equivalent (Fiscal, 2010) by using 30 kg of the hybrid 
bonded lumpy ore per ton of pig iron.
CO
2
 reduction in the sintering process by 
blowing natural gas onto the sinter cake
JFE Steel Corp developed Secondary-fuel Injection Technology 
for Energy Reduction (Super-SINTER)(JFE Steel Co, 2009) 
(Figure 10). In this process, natural gas injected from nozzles 
is sucked into the sinter bed. This technology maintains the 
optimum sintering temperature in the sinter bed for longer 
time than the conventional technology does. It improves yield 
of lump sinter and energy effi ciency. Reduction in CO
2
 emission 
by this technology is estimated to be about 60 000 t/a. 
Start up of SCOPE21 – a new coke oven battery 
A new coke oven that adopted the SCOPE (Super Coke Oven 
for Productivity and Environmental enhancement toward 
the 21st century) technology was in operation on February 
2008 at Oita Works of Nippon Steel Corp SCOPE allows high 
temperature coal charging up to 250°C to improve coking 
property and coke oven’s productivity. The battery reached a 
productivity at as high as 1.85 charges per day, and also has 
maintained a stable operation since the beginning of operation 
(Doi et al, 2009) (Figure 11).
The operation is fully automated and stable operation was 
confi rmed within four months from the star-up. It took only 
less than a year to reach the full production capacity.
High-speed coal preheating and high bulk density of the 
charged coal contribute to a high ratio of semi-soft coking coal 
in total coal as high as 50 per cent or more. Moreover, the coke 
strength is 2.5 per cent higher than that of the conventional 
coke ovens. 
FIG 7 - Development of carbon brick.
FIG 8 - Schematic diagram of the quantitative damage information analysis 
system of Doctor of Coke Oven (Sakaida, Awa and Sugiura, 2006).
FIG 9 - The producing process of hybrid bonded lump ore.
FIG 10 - Schematic fi gure of Super-SINTER.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
N TAKAMATSU
24
Improvement of production effi ciency of blast 
furnace
New numerical simulation technology adopted in 
the iron-making processes 
The discrete element method (DEM) is being used for 
simulation of the blast furnace process. This is applied to 
analysis on behaviour of the burden particles when they pile 
up on top of the burden and when they descend in the furnace 
shaft. This also applied to analysis on a agglomerating process 
in drum mixers of raw material preparation processes. This 
model gives us useful information for operation control (Mio 
et al, 2009) (Figure 12). 
Researches and developments for effi cient blast 
furnace operation with an experimental blast 
furnace
It is important to examine changes in the blast furnace 
operation when the raw material qualities change. In addition, 
it is also important to obtain conditions of highly productive 
and energy efficient operation of blast furnaces. However, 
such examination with a commercial furnace is diffi cult, and 
that with a bench scale furnace is not accurate. Therefore, 
Sumitomo Metal Ind Ltd has conducted various researches 
using an experimental blast furnace with an inner volume of 
3.8 m (Nitta, Nakamura and Ishii, 2008; Ujisawa et al, 2006) 
(Figure 13). 
Recently, a series of experiments were conducted to quantify 
effects of hot briquette iron (HBI) on productivity and rate of 
the reducing agent in a blast furnace. In the experiments, HBI 
is found to be as a future source of iron in a blast furnace for 
high production and lower CO
2
 emission. 
Natural gas injection to blast furnace
JFE Steel Corp introduced natural gas (NG) injection 
to increase productivity of a blast furnace and to reduce 
CO
2
 emission (Yamamoto, Kashiwara and Tsukiji, 2008) 
(Figure 14). The trial of a commercial plant using No 2 Blast 
Furnace of JFE Steel’s East Japan Works (Keihin) proved that 
NG injection improved gas permeability in the furnace and, as 
a result, increased productivity of a blast furnace.
Moreover, because ratio of oxide reduction by hydrogen 
increased when NG is injected, CO
2
 emission was reduced. The 
blast furnace made the world’s highest productivity of blast 
furnaces bigger than 5000 m3 during the trial (2.56 t/d-m3). 
CONSERVATION OF ENERGY AND CO
2
 
REDUCTION TECHNOLOGY
Energy consumption in the iron and steel 
industry in Japan
The Ministry of Economy, Trade and Industry (METI) 
fi nalised ‘the Cool Earth Energy Innovative Technology Plan’ 
in March 2008 and have just released it. In May 2007, the 
prime minister at that time announced the ‘Cool Earth 50’ 
initiative. The long-term goal of this initiative was a 50 per 
cent reduction of the world's total greenhouse gas emissions 
by 2050. The Prime Minister of Japan stated at the United 
Nations Summit on Climate Change in September, 2009 
that Japan will set a middle term target of reduction in CO
2 
emission by 25 per cent in 2020 from that in 1990.
FIG 11 - The process fl ow and main specifi cations of Oita no 5 coke oven.
FIG 12 - An example of burden distribution in the blast furnace 
simulated by discrete element method.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Fluidized 
bed dryer 
Hot briquetting machine 
Pneumatic 
preheater 
High 
temperature 
coal bin 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
RECENT IRON MAKING TOPICS IN JAPAN
25
The Japanese steel industry exhausts CO
2
 about ten per 
cent of the national emission, and about 70 per cent of its 
CO
2
 emission is from the iron-making processes. Therefore, 
efforts to reduce CO
2
 in the iron-making processes are very 
important. Developing innovative technologies is essential to 
fulfi lling this goal. 
Innovative iron-making process
Original iron-making technologies to expand fl exibility in 
choice of raw materials and to reduce CO
2
 emission are 
sought for keeping competitiveness of the Japanese steel 
industry. Among various research and development activities, 
Advanced Research on Innovative Iron-making Process 
Project has implemented as a government project cooperated 
by JRCM, the Japanese steel industry and several universities 
starting in December 2006 (Naito, 2009) (Figure 15). One of 
the key technologies is a compound material made of ferrous 
oxide, carbon material and metallic Fe. Target of the research 
is set as ten per cent reduction in energy consumption, which 
is comparable to the improvement in the blast furnace shaft 
effi ciency when the thermal reserve zone temperature drops 
from 1000°C - 800°C. 
After three years of the research, important results, 
understandings of reduction behaviour of the compound 
material and a new manufacturing process of the compound 
material, have been obtained. By a high temperature softening 
test, the compound material showed improvement in the 
reduction rate and in permeability of the cohesive zone.
This technology is still being researched to bring to 
completion of a new Japanese original technology for energy 
conservation and CO2 emission control.
COURSE 50 (Matsuzaki et al, 2009)
The basic idea of COURSE 50
The project of CO
2
 Ultimate Reduction in Steelmaking process 
by innovative technology for cool Earth 50, COURSE 50, was 
proposed by JIFS with six major steel companies and related 
companies. The project is composed of two major technical 
developments. 
The low carbon, environmentally-accepted iron-making 
process development (research and development item 1) 
consists of two parts. One is development of a practical 
technology applying hydrogen to iron ore reduction, and 
the other is development of low cost hydrogen production 
technology (ex by reforming of coke oven gas (COG)). The 
CO
2
 separation and collection technology development 
(research and development item 2) consists of two parts. One 
is a technology of CO
2
 separation from blast furnace gas, and 
the other is a utilisation technology of waste energy for CO
2
 
separation and collection.
Possibilityof hydrogen reduction
To reduce CO
2
 emission, we will examine reformed COG 
gas (CO 35 per cent, H2 60 per cent) injection through 
 
FIG 13 - Schematic picture of experimental blast furnace (BF).
FIG 14 - Operation trends of no 2 blast furnace at East Japan Works (Keihin).
FIG 15 - The reduction behaviour change by existing carbon composite material.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
N TAKAMATSU
26
conventional tuyeres and also through trial nozzles set on the 
blast furnace shaft. The basic idea of hydrogen reduction in 
the blast furnace process using reforming COG is shown in 
Figure 16.
CONCLUSIONS
The environment surrounding iron-making is getting severer, 
and therefore the following issues are getting more important:
  degradation and imbalance in demand and supply of the 
iron ore and coal,
  aging plant facilities,
  fl exible in production and effi cient operation, and
  environmental problems.
The Japanese steel industry has been improving operations 
and conserving natural resources, and at the same time coping 
with aging equipments and long-term environment issues. 
We are convinced to create a leading edge of technologies to 
contribute the world’s iron and steel industries and to solve 
the problems ahead us.
REFERENCES
Doi, K, Noguchi, T, Fujikawa, H, Shioda, T, Yokomizo, M, Kato, K, 
Nakajima, Y and Uematsu, H, 2009. CAMP-ISIJ, 22:782.
JFE Steel Co, 2009. News release [online]. Available from: 
<http://www.jfe-steel.co.jp/release/2009/06/090615.html> 
(in Japanese) [Accessed: 15 June 2009].
Kasai, E, Komarov, S, Nushiro, K and Nakano, M, 2005. 
ISIJ Int, 45:538.
Matsuzaki, S, Yonezawa, K, Saito, K and Naito, M, 2009. 
CAMP-ISIJ, 22:268.
Mio, H, Komatsuki, S, Akashi, M, Shimosaka, A, Shirakawa, Y, 
Hidaka, J, Kadowaki, M, Matsuzaki, S and Kunitomo, K, 2009. 
ISIJ Int, 49:479.
Naito, M, 2009. JRCM NEWS, 268:1.
New Energy and Industrial Technology Development 
Organization (NEDO), 2007. Outline of NEDO, 2007-2008, 
October, First Edition, 114.
Nitta, M, Nakamura, H and Ishii, A, 2008. Nippon Steel technical 
report, 388:48.
Oyama, N, Sato, H, Takeda, K, Ariyama, T, Matsumoto, S, Jinno, T 
and Fujii, N, 2005. ISIJ Int, 45:817.
Sakaida, M, Awa, Y and Sugiura, M, 2006. Nippon Steel technical 
report, 384:63.
Ujisawa, Y, Snahara, K, Matsukura, Y, Nakano, K and Yamamoto, T, 
2006. Tetsu-to-Hagane, 92:591.
Yamamoto, K, Kashiwara, Y and Tsukiji, H, 2008. JFE Giho, 22:55.
 
Oven in 
works
Power 
generation 
in works
Power 
generation
For sale
Coke oven
Coke
Blast 
furnace
COG
Reforming COG
Gas treatment
Reforming COG
Hydrogen for society
H2;65%
Co;35%
COG: Coke oven gas
H2+CO H2
FIG 16 - Basic idea of hydrogen reduction in the blast furnace process using reforming coke oven gas.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 27
Developing the West Pilbara Iron Ore Project
K Watters1, A Priddy2 and D Ducler des Rauches3
ABSTRACT 
API Management Pty Ltd (API) is currently developing the West Pilbara Iron Project (the ‘Project’). The 
Project is a substantial greenfi elds mining and infrastructure development proposed for the west Pilbara 
region of Western Australia. The Project will establish API as a new and independent supplier of direct 
shipped iron ore. API manages the Australian Premium Iron Joint Venture (APIJV) on behalf of joint 
venture participants Aquila Resources Ltd and American Metals and Coal International Inc. The APIJV 
was formed in February 2005 with the objective of developing a long-term iron ore export operation in the 
west Pilbara region. 
Development of Stage 1 comprises a 30 million tonne per annum (Mt/a) channel iron deposit (CID) 
operation scheduled to commence in the second half of 2014 subject to approvals. JORC compliant CID 
resources total nearly 700 million tonnes and the Stage 1 mining area comprises eight iron ore deposits 
located between 30 km and 85 km south of Pannawonica. Run-of-mine ore from the eight deposits is 
transported to a central processing facility for crushing, screening and blending into a sinter fi nes product, 
which API has named West Pilbara Fines or WPF. Construction of a 282 km main line heavy haulage 
railway from the Stage 1 mining area to the proposed greenfi elds multi-user port at Anketell Point near 
Wickham, is capable of accommodating large Capesize bulk carriers and providing export facilities for 
other third parties. Capital investment in the order of $A6 billion includes essential support infrastructure 
including power, water, accommodation, roads and an airport at the mine. 
A feasibility study for the development of the Stage 1 mine, rail and port infrastructure has been completed 
along with environmental surveys of the mine, rail and port areas with Public Environmental Reviews 
(PERs) for each submitted for public review by end 2010. Consultation and cooperation with various 
stakeholders of the Project areas, and early engagement with traditional owners of the respective areas has 
commenced.
Various sampling and metallurgical characterisation programs combined with independent institute and 
steel mill based testing have confi rmed the positive sintering characteristics of WPF ore, which is higher in 
alumina (3.37 per cent) than typical Pilbara ores. 
The installed rail and port infrastructure will be the catalyst for further resource development within 
the west Pilbara region. There is considerable regional demand for a new multi-user port, given access to 
existing Pilbara facilities is severely constrained. 
While the expected Stage 1 mine life is less than 20 years, the infrastructure is planned to have a useful 
life of at least 50 years. Following on from successful exploration, acquisition and partnering, API plans to 
continue to develop and open up new mines and associated infrastructure in the west Pilbara. 
1. Project Director, API Management Pty Ltd, Level 2, 1 Preston Street, Como WA 6152. Email: kwatters@apijv.com.au
2. General Manager Marketing, API Management Pty Ltd, Level 2, 1 Preston Street, Como WA 6152. Email: apriddy@apijv.com.au
3. Senior Process Engineer, API Management Pty Ltd, Level 2, 1 Preston Street, Como WA 6152. Email: dducler@apijv.com.au
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 29
Development Options for the Eyre Peninsula 
Magnetite – An Owner’s Perspective
J White1 and B Hammond2
ABSTRACT
The Eyre Peninsula (EP) has been known for its iron ore for more than a century. Edward John Eyre 
noted the ironstone outcrop at what is now Iron Knob in the 19th century. Sir Douglas Mawson visited 
and reported on the geology of the lower EP in 1906, noting the widespread presence of iron stone and the 
likelihood of copper being associated with the iron stone. Despite this knowledge there has been only one 
successful iron ore mining region developed on the EP and that is Iron Knob and the Middleback Ranges 
which have been mined continuously since 1899. The Middleback Ranges can justly lay claim to be birth 
place of the Australian iron ore and steel industries.
The discovery of large scale haematite resources in Western Australia meant that there was no need to 
seek out iron ore that needed benefi ciation. The low price of iron ore precluded further development of 
the EP and many other areas in Australia and overseas. This situation changed rapidly in the mid-2000s 
with the emergence of China as the world’s industrial power house and the soaring demand and price 
for many resources including iron ore. This triggered a fl urry of activity across Australia and led to the 
rapid exploration and drives to develop the EP iron ore. A brief examination of the magnetic maps of 
the EP shows the wide spread presence of magnetite and associated smaller haematite deposits. A closer 
look reveals the magnetite deposits to vary in formation style and host rock lithology as well as being 
signifi cantly altered by regional metamorphism. 
Whilst developers of the traditional directship haematite ores of Australia have had limited control on the 
physical and chemical characteristics of the product they produce, the developers of benefi ciated magnetite 
ores such as on the EP must decide what product they will target along a continuum of processing options. 
These decisions must be made taking both the development constraints of the individual operation and 
orebody into account while balancing this with the end customer’s needs. Every decision in this respect is 
a trade-off between process effi ciency and product marketability. A comparison of the key trade-off factors 
of grind size, DTR recovery and concentrate chemistry (Fe per cent, SiO2 per cent) for the EP against 
current magnetite projects across Australia can be made showing in particular the potential of EP ores to 
produce highly saleable products at coarse grind sizes. 
Infrastructure provision has a large economic impact on magnetite mines and concentrators. Magnetite 
requires signifi cant amounts of water and power to drive the benefi ciation process and, in common with 
haematite, transport infrastructure to provide a route to market. The potential exists on the EP to access 
power from the east coast grid network and to derive the necessary water by using this power to produce 
desalinated water from the sea. All the EP ore deposits are relatively close to the coast and there are a 
number of proposals to develop ports close to the mine sites. In the case of the Centrex Metals projects, 
the port would be less than 100 km from the furthest deposit and slurry pipe lines for product transport 
preclude the need for extensive rail development and the associated high operating costs. In addition there 
are existing industrial towns in the region with some social infrastructure already in place.
A comparison of the attributes and location of the EP iron ore deposits give promise of the development 
of high quality magnetite concentrates that complement the large scale haematite developments elsewhere 
in Australia.
1. Managing Director, Centrex Metals Limited, Unit 1102, 147 Pirie Street, Adelaide SA 5000. Email: jwhite@centrexmetals.com.au
2. MAusIMM, Chief Operating Offi cer, Eyre Iron Pty Ltd, Level 9, 420 King William Street, Adelaide SA 5000. Email: ben.hammond@eyreiron.com.au
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 33
INTRODUCTION
The global mining and exploration database Intierra Resource 
Intelligence (Intierra Mapping, 2010) lists 154 iron ore 
deposits in Western Australia. Of these 154 deposits, only 
34 are currently being mined with most of the remainder at 
various stages of exploration, feasibility and development. A 
large proportion of the iron ore deposits expected to commence 
production in the next fi ve years comprise resources that 
will require some degree benefi ciation, such as magnetite or 
Channel Iron Deposits (CID).
Presently, there are very few sites around Australia that 
generate iron ore tailings from their operations. The sites 
that do generate tailings include Savage River in Tasmania, 
Middleback Range in South Australia and MESA J in Western 
Australia. Consequently, operational experience with iron ore 
tailings in the Australian, and particularly Western Australian, 
context is limited.
Li et al (2010) asserts that iron ore tailings have become one 
of the most hazardous solid wastes. Indeed, the collapse of 
one iron ore tailings storage facility in 2008 killed a reported 
254 people in China’s Shanxi Province. The failure contributed 
the largest number of fatalities to a list of not less than 11 major 
iron ore tailings dam failures that have occurred in the last 
35 years (ICOLD, 2001; WISE Uranium, 2011).
While the triggers for the geotechnical stability of tailings 
storage facilities are generally well understood, there 
are several other aspects of iron ore tailings that present 
challenges to the proponent. The geology of hard rock iron 
ore deposits in Australia means that they are often associated 
with the occurrence of asbestiform and sulfi de minerals. 
Consequently, the tailings generated by these projects can 
present a signifi cant hazard to the environment, both in terms 
of the potential for the liberation of respirable fi bres and the 
generation of Acid and Metalliferous Drainage (AMD). When 
coupled with sensitive environmental settings and large 
production volumes, these factors pose a signifi cant challenge 
for designers, regulators and producers.
Presently, there are several large magnetite mining and 
processing operations under consideration and development 
in Western Australia. These projects include the Sino 
Iron Project, Balmoral South Iron Ore Project, Jack Hills 
Expansion Project, Karara Iron Ore Project, Ridley Magnetite 
Project, Cape Lambert Iron Ore Project and Extension 
Hill Magnetite Project. Publically available information 
(shareholder announcements, Department of Mines and 
Petroleum (DMP) publications and company websites) 
suggests that these projects each have the capacity to generate 
1. Senior Tailings Engineer, Golder Associates Pty Ltd, Level 3, 1 Havelock Street, West Perth WA 6005. Email: danstey@golder.com
2. Civil Engineer, Golder Associates Pty Ltd, Level 3, 1 Havelock Street, West Perth WA 6005. Email: dreid@golder.com
Overcoming the Challenges of 
Managing Tailings from Iron Ore 
Mining in Western Australia
D R Anstey1 and D A Reid2
ABSTRACT
The growing demand for iron ore from Western Australia has led to an increased focus on the 
development of magnetite and lower grade iron ore deposits. The need to process and benefi ciate 
the ore from these projects can result in the generation of signifi cant quantities of iron ore tailings 
over the life of a project. Over the coming decades, the management of iron ore tailings will become 
a signifi cant aspect of the mining landscape in Western Australia. 
The tailings generated from iron ore projects differ signifi cantly from other mineral tailings 
currently produced in Western Australia. Iron ore tailings are generated on a scale not matched 
by any other mineral process and can often contain asbestiform or potentially acid forming 
components. New iron ore projects are coming under increased environmental scrutiny because of 
their large scale operations and the sensitive environments in which they occur. This combination 
of factors creates some unique challenges for the management of tailings generated by iron ore 
mining in Western Australia.
This paper presents the fi ndings of two studies conducted in recent years to overcome the 
challenges posed by iron ore tailings. The paper discusses the role of laboratory testing in the 
development of management strategies for asbestiform minerals in tailings. It also presents a 
numerical modelling approach for evaluating the rates of rise in the storage of magnetite tailings. 
The paper concludes by outlining how the fi ndings of these studies can be applied to reduce 
tailings management costs, lower risks and improve environmental outcomes for the management 
of tailings from iron ore mining.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
D R ANSTEY AND D A REID
34
between 18 and 67 Mt/a of iron ore tailings. The fi rst of these 
projects is scheduled to commence production in mid-2011. 
The amount of tailings generated by the largest of these new 
magnetite projects will be more than ten times the amount of 
tailings currently generated by the largest existing magnetite 
processing facility in Australia and four times the amount of 
tailings generated by the ‘Super Pit’ gold mining operations in 
Kalgoorlie.
The implications are that, over the next fi ve years, the rate 
of iron ore tailings generated in Western Australia will grow 
exponentially. With this comes the need to understand and 
manage iron ore tailings in an economically feasible, socially 
responsible and sustainable manner.
This paper describes two innovative approaches that 
have been recentlybeen used to support the design and 
environmental permitting of iron ore tailings storage facilities 
in Western Australia.
DESIGNING FOR THE CONTAINMENT OF 
ASBESTIFORM MINERALS
Regulatory context
The Western Australian government recently released a 
guideline on the management of fi brous minerals in Western 
Australian mining operations (DMP, 2010). The guideline 
acknowledged that:
Increasing demand for minerals has made mining 
and processing of previously uneconomic orebodies 
commercially viable. Consequently, fi brous minerals 
are now encountered in mining operations more 
frequently than in the past.
The DMP guideline makes particular reference to the 
occurrence of fi brous minerals within the banded iron 
formations of the Hamersley Basin and goes on to say that:
Processed tailings can contain similar concentrations 
of fi brous minerals to the feed ore, but the fi bre bundles 
are generally thinner and shorter and therefore more 
biologically active and more hazardous. Consequently, 
it is imperative that tailings are properly managed so 
that fi bres do not become airborne.
It is imperative that iron ore miners not only understand the 
presence and distribution of fi brous minerals in their orebody 
but also quantify the hazards these materials pose after 
processing. Where fi brous minerals are present in tailings, it is 
incumbent on the proponent to demonstrate to the Regulators 
and other stakeholders that these tailings can be stored safely.
One recognised method for the control of fugitive dust 
emissions from tailings is the use of water to prevent drying 
of the tailings. The DMP guideline suggests that mines should 
demonstrate how the tailings will be maintained in a wet state. 
However, it may not always be possible to maintain tailings 
in a wet state, particularly when dealing with large tailings 
areas, raises to tailings storage facilities, or mill shutdowns. 
Questions are also raised about the effi cacy of moisture control 
to contain fi bres and what other controls can be implemented 
to reduce the potential for respirable fi bres. Regulatory 
approval of mining operations may rest on the ability of the 
proponent to demonstrate that its proposed respirable fi bre 
controls are effective.
Wind tunnel testing
Wind tunnel simulations are a recognised method for 
estimating fugitive dust emissions generated by tailings 
(Neuman, Boulton and Sanderson, 2008). As part of the 
current study, wind tunnel experiments were carried out 
to examine the potential for iron ore tailings to generate 
respirable fi bres under several simulated fi eld conditions. The 
objectives of the study were to:
  assess the potential for tailings containing asbestiform 
minerals to generate respirable fi bres under a range of 
moisture conditions and wind speeds,
  examine the effect of crusting and crust rupture on 
respirable fi bre generation,
  examine the generation of fi bres from a tailings beach as a 
function of time, and
  compare the quantity of fi bres generated from a wind 
tunnel to the National Occupational Health and Safety 
Commission (NOHSC) guidelines.
Apparatus
Testing was carried out on a sample of Western Australian 
magnetite tailings generated from pilot plant testing of a 
bulk ore sample. The tailings were examined under Polarised 
Light Microscopy and were confi rmed as containing of both 
riebeckite and crocidolite particles (blue asbestos). A particle 
size distribution for the tailings is included as Figure 1.
Experiments were carried out in a purpose-built wind tunnel 
at Golder Associates’ Soils Laboratory in Perth, Western 
Australia. The wind tunnel has been specifi cally designed for 
the testing and containment of tailings containing asbestiform 
minerals. Airfl ow in the tunnel is generated from three variable 
speed Nilfi sk IVB 3-H vacuums, assembled in parallel. The 
vacuums are approved for the handling hazardous waste in 
0
10
20
30
40
50
60
70
80
90
100
0 1 10 100 1000
C
um
ul
at
iv
e 
%
 P
as
si
ng
 b
y 
M
as
s
Particle Size (μm)
FIG 1 - Tailings particle size distribution.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
OVERCOMING THE CHALLENGES OF MANAGING TAILINGS FROM IRON ORE MINING IN WESTERN AUSTRALIA
35
accordance with AS/NZS 60335.2.69 (2003). Staff involved 
in the testing wore appropriate personal protective equipment 
and adhered to strict wash-down, containment, monitoring, 
handling and waste disposal protocols.
The wind tunnel was designed and constructed using 
guidance from Mehta and Bradshaw (1979). The dimensions 
of the tunnel were restricted to facilitate testing at wind speeds 
higher than might otherwise be achievable. The tunnel is 
constructed from clear Perspex with internal cross-sectional 
dimensions of 150 mm (width) × 40 mm (height) and a 
length of 2.4 m. An aluminium condenser (approximately 
180 mm × 80 mm) feeds the tunnel and contains three wire 
mesh screens (each with four openings per 25 mm in lineal 
length) to assist in the creation of laminar air fl ow. The wind 
tunnel is capable of achieving maximum wind speeds of 
approximately 75 km/h (21 m/s).
Tailings were deposited directly into purpose built Perspex 
sample trays that slide into and sit fl ush with the base of 
the apparatus. The exposed tailings surface in the sample 
tray has dimensions 128 mm (width) × 325 mm (length) 
× 10 mm (depth). A photograph of a tailings sample that has 
been loaded into the wind tunnel in preparation for testing is 
included as Figure 2.
Two air sampling locations were used, one near the 
wind tunnel intake and another within the wind tunnel, 
approximately 150 mm downwind of the simulated tailings 
beach. The wind tunnel intake monitoring location was selected 
to measure respirable fi bre background concentrations during 
each test. The sampling location within the wind tunnel 
was selected as the primary measurement point for the test 
and comprised a 7.9 mm outer diameter plastic tube fi tted 
vertically through the top of the wind tunnel, and cut at a 
45 degree angle to face the wind tunnel intake. The sampling 
tube passed through a sample cowl fi tted with a 25 mm 
membrane fi lter and was connected to an air sampling pump. 
Air sampling was carried out in accordance with the NOHSC 
guidelines on the membrane fi lter method for estimating 
airborne asbestos fi bres (NOHSC:3003, 2005). The air 
sampling devices were calibrated prior to the commencement 
of each test.
Method
Tailings were tested at three different moisture conditions: 
a wet, intermediate moisture and dry state. The wet tailings 
simulated freshly deposited tailings, shortly after settlement. 
There was no free water on the surface of the tailings but the 
samples maintained a refl ective sheen. 
The intermediate moisture tailings were selected to simulate 
tailings that had undergone some drying on the tailings beach. 
This is typical of the drying that might occur as the result of 
rotation between the discharge locations during operation. 
The tailings were allowed to air dry over several days back to a 
moisture content of between 11 per cent and 18 per cent. 
The dry tailings samples were dried in a 50°C oven to 
achieve moisture content of less than one per cent. Air drying 
tests carried out on the same tailings indicate that the ‘dry’ 
moisture condition would be achieved after approximate eight 
days under simulated summer drying conditions.
Testing was carried out for each moisture condition at target 
speeds of 46 km/h and 75 km/h. A fresh tailings sample was 
used for each test. In addition, testing was carried out on three 
dry tailings samples that were disturbed prior to testing. Two 
of the samples were disturbed by crushing the surface of the 
tailings with a block to simulate the effect of traffi cking. The 
surface of a third sample was scoured with a metal brush.
Prior to each test, the wind velocity in the tunnel was set and 
measured using a blank sample tray. Theair sampling rate 
was calibrated for the target wind velocity, before the vacuums 
were switched off and the blank tray replaced with the test 
sample. The sample tray was sealed and the air sampling 
equipment turned on, followed by the vacuums. Each test was 
run for an air sampling duration of fi ve minutes.
Most of the test durations were continuous. However, one 
test was divided into fi ve time increments to measure changes 
in fi bre concentrations over time. Air fl ow was stopped after 
each time increment and a fresh sample cowl inserted into the 
air sampling device downstream of the tailings.
Following the completion of each test, the sample tray was 
removed from the wind tunnel and the fi nal moisture content 
of the tailings was measured by oven drying. The wind tunnel 
was thoroughly cleaned after each test to avoid contamination 
of subsequent tests. 
Results and discussion
The results of the wind tunnel testing are presented in Table 1.
Wind tunnel tests on the wet and dry tailings did not 
produce fi bres above the reportable detection limit at wind 
speeds up to 75 km/h. However, the intermediate moisture 
condition tailings did produce measurable numbers of 
respirable fi bres at both 46 km/h and 75 km/h wind speeds. 
This result is consistent with the fi ndings of Neuman, Boulton 
and Sanderson (2008), who suggest that it is the creation of 
electrochemical bonds (crusting) and not capillary forces that 
provide the greatest resistance to the wind erosion of tailings. 
Airborne particulates were not visible for any of the 
undisturbed tailings tests, despite the fact that measurable 
fi bre concentrations were recorded for the intermediate 
moisture content tests. The implication is that magnetite 
tailings may be vulnerable to respirable fi bre generation while 
it is drying, even if dust emissions are not visible.
By comparison to the undisturbed samples, airborne 
particulates were observed for the fi rst few seconds of all of the 
disturbed tests. Both the crushed and scoured tests produced 
respirable fi bre emissions that were approximately one order 
of magnitude higher than for the undisturbed samples tested 
at intermediate moisture contents.
Fibre concentrations were measured over different time 
intervals for a dry tailings sample, crushed and subjected to 
75 km/h wind speeds. The results of the incremental test are 
presented graphically, on a logarithmic scale, in Figure 3.
FIG 2 - Tailings sample loaded within wind tunnel for testing.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
D R ANSTEY AND D A REID
36
The results of the incremental testing support the hypothesis 
that the measured respirable fi bre concentration is a function 
of the test duration. For a single sample area and wind speed, 
fi bre concentrations will decay with time. This is consistent 
with the PM
10
 (dust particles with 1 μm < aerodynamic 
diameter <10 μm) emissions measured from discrete tailings 
samples by Neuman, Boulton and Sanderson (2008). The 
results indicate that the majority of the respirable fi bres 
become airborne within the fi rst fi ve seconds of being exposed 
to airfl ow.
Comparison with National Occupational Health 
and Safety Commission guidelines
The Australian occupational exposure standard for asbestos 
NOHSC:2002 (2005), specifi es that the time-weighted 
average asbestos fi bre concentration of the air breathed by a 
worker during a working shift should not exceed 0.1 fi bres per 
millilitre (fi bres/mL) for all forms or mixtures of asbestos. 
The fi bre concentrations derived from for the current 
experiments should not be directly compared to the 
occupational exposure standards. The results must be 
factored to take account of the relatively short duration of 
the air sampling and the location of the air sampling intake. 
The exposure standards relate to sampling within a 300 mm 
radius the worker’s breathing zone, while the measurements 
taken from the current tests have been sampled immediately 
downstream of the emission source.
The test results from this study can best be compared to a 
single exposure event that could occur during a working shift 
(eg a worker standing near a piece of working machinery on 
a tailings beach). Applying this logic, an employee working 
an 8 h shift could be exposed to a single fi ve minute event 
with a concentration of 9.6 f/mL, without exceeding the 
Australian occupational exposure standards. Compared on 
this basis, only the dry crushed tailings sample exposed to 
75 km/h wind speeds demonstrated the potential generate 
fi bre concentrations in excess of the exposure standard.
In reality, the fi bre concentrations that workers are exposed 
to will be infl uenced by several factors. These include 
the moisture content of the tailings, depth and degree of 
disturbance, wind gusts and boundary conditions, tailings 
storage geometry, exposed tailings area, saltation, dispersion 
and dilution. Notwithstanding limits on the sample size of the 
test area, the results of testing from the wind tunnel could be 
incorporated into a dispersion model that would take account 
of the factors listed above. Such a model could be used to 
estimate fi bre concentrations downstream of the source, at 
the location of sensitive receptors. The results of modelling 
could be compared with the NOHSC guidelines to provide 
indications of the exposure hazard for workers.
Tailings condition Wind speed (km/h) Final moisture content (%) Fibre concentration 
(fi bres/mL)
Background fi bre 
concentration (fi bres/mL)
Wet 48.8 22.7 BDL <0.01
Wet 74.6 24.3 BDL 0.01
Intermediate 45.9 11.4 0.4 0.2
Intermediate 75.0 17.9 1 0.01
Dry 46.2 0.1 BDL <0.01
Dry 75.1 0.2 BDL 0.01
Dry crushed 45.3 0.1 5 0.2
Dry crushed 75.2 0.1 11 0.2
Dry scoured 45.6 0.1 4 <0.01
Note: BDL: below reportable detection limit determined in accordance with NOHSC: 3003 (2005)
TABLE 1
Wind tunnel test fi bre concentrations.
1
10
100
1000
0 - 5 s 5 - 20 s 20 s - 1 min 1 - 2.5 min 2.5 - 5 min
Fi
br
es
 C
on
ce
nt
ra
tio
n 
(F
ib
re
s/
m
L)
Time Interval
FIG 3 - Comparison of fi bre concentrations measured over diff erent time intervals.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
OVERCOMING THE CHALLENGES OF MANAGING TAILINGS FROM IRON ORE MINING IN WESTERN AUSTRALIA
37
Study outcomes
Testing was conducted on tailings in wet, dry and intermediate 
moisture states, in disturbed and undisturbed conditions, at 
wind velocities of approximately 46 and 75 km/h. The study 
makes the following conclusions:
  wind tunnels can be used to measure respirable fi bre 
generation from magnetite tailings that contains 
asbestiform minerals and used to evaluate the relative 
effectiveness of control measures, such as moisture;
  tailings tested in both saturated and dry states did not 
demonstrate the potential to produce fi bres above the 
recordable detection limit;
  tailings that have undergone some drying and are in an 
intermediate moisture state may be vulnerable to the 
generation of respirable fi bres at high wind speeds;
  tailings that have been disturbed, whether through 
pulverisation or abrasion, have the potential to generate 
respirable fi bres at signifi cantly higher concentrations 
than undisturbed tailings;
  fi bre concentrations generated from a single disturbed 
source and subject to a constant wind speed will decay 
over time; and
  fi bre concentrations measured under wind tunnel 
conditions should not be directly related to national 
occupational exposure standards.
The implications of this study for the mining industry are 
that, where asbestiform minerals are present in tailings:
  tailings should not be used as a construction material, 
it should not be excavated, traffi cked or disturbed by 
machinery;
  tailings deposition should be controlled so as to maintain 
the tailings in a wet state to the extent possible during 
deposition; and
  fi xed and personal monitoring should be undertakenaround the tailings storage facility to confi rm that 
respirable fi bre concentrations do not exceed the national 
occupational exposure standard.
Limitations
Tailings properties, their particle size and mineralogy can 
vary signifi cantly between sites even for the same commodity 
and region. The results of this testing should, therefore, not be 
applied directly to other sites. It is recommended that testing 
be carried out that is specifi c to each project under conditions 
that best simulate those expected in the fi eld.
The current tests have not considered the effects of saltation. 
Saltation is a recognised mechanism for dust generation 
and could result in the generation of respirable fi bres that 
would not be generated in the absence of saltation. Further 
testing could be carried out using the current wind tunnel, 
introducing particles to the air intake stream to simulate the 
effects of saltation.
The current tests have concentrated on whole tailings 
samples. Tailings that have segregated across a tailings beach 
will have different particle size distributions, which may make 
them more or less susceptible to drying, wind erosion and 
respirable fi bre generation. Future tests could be carried out 
on tailings that are subsampled from a segregated mass, to 
assess the infl uence of segregation.
The dimensions of the wind tunnel and sample trays were 
limited in the current study because of the wind speeds 
required and the need to create a closed system, capable of 
safely containing respirable fi bres. The scale of the wind 
tunnel will affect the accuracy of the measurements. Ideally 
the wind tunnel and sample sizes should be as large as is 
possible within the constraints of the study.
Dust concentration has been demonstrated to vary across 
the vertical profi le downstream of a fl at bed emission source 
(Neuman, Boulton and Sanderson, 2008). A higher profi le 
wind tunnel and an increased number of sample points could 
be used to evaluate the concentration of respirable fi bres in 
profi le.
HIGHER FASTER – REDUCING THE 
DISTURBANCE FOOTPRINT
Current practice
It is common practice in Western Australia for tailings 
storage facilities to be raised using the upstream construction 
method. Upstream construction is characterised by using fi ll 
material to construct an embankment on the tailings beach, 
with the resultant centreline of the embankment moving in 
an upstream direction with each successive raise. Typically, 
tailings are deposited from the crest of a starter embankment 
to form a beach. Once the starter embankment has reached 
its capacity, fi ll is placed onto the tailings beach, which forms 
the foundation for the next embankment raise. In many cases, 
tailings are borrowed from the beach adjacent to the raise 
footprint and used as fi ll material. This process continues as 
the embankment increases in height.
The major advantage of upstream construction is its low cost, 
resulting from the small embankment fi ll volume required 
relative to the tailings storage volume achieved. Upstream 
construction uses signifi cantly less fi ll than alternative 
construction methods, such as downstream construction 
(Vick, 1990). However, it relies on the need for a competent 
tailings beach on which to construct each embankment 
raise. To improve foundation conditions, tailings can be 
mechanically disturbed, dried or compacted to provide 
suitable conditions for each raise. 
Upstream embankment construction is used throughout 
Western Australia for tailings generated from a range of 
commodities, principally gold and nickel. The physical 
properties of gold tailings, and the need for the tailings to 
form a competent foundation, means that the rate of rise of 
these facilities is typically limited to between 1.5 and 2.5 m/a. 
This rate of rise is consistent with experience in other semi-
arid areas of the world, and guidelines published by the 
International Committee on Large Dams (ICOLD, 1995).
Currently proposed magnetite projects are expected to 
produce many times more tailings solids per year than the 
largest gold mining operations in Western Australia. These 
operations will require the construction of tailings storage 
facilities on a massive scale. A magnetite mine producing 
50 Mt/a of tailings and achieving an average stored density 
of 1.8 t/m3, would require a minimum active tailings storage 
area of more than 1100 ha, to maintain an average rate of rise 
of 2.5 m/a. Many operations do not have an area of this size 
available for tailings storage.
The ability to raise a magnetite tailings storage facility using 
upstream methods at a rate of rise higher than 2.5 m/a may 
have signifi cant benefi ts in terms of cost, environmental 
impact and operational complexity.
Study approach
Mundle and Chapman (2010) propose a simple indicator test 
for estimating the rate of rise achievable for different tailings 
types based on the results of air drying tests. The test provides 
an indicative rate of rise for tailings but does not consider the 
effects of consolidation or the potential for the accumulation 
of excess pore pressures under loading. Excess pore pressure 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
D R ANSTEY AND D A REID
38
accumulation and its effect on stability is the limiting factor 
when considering upstream raises at high rates of rise. 
Air drying test results from the magnetite tailings suggest 
that the tailings can achieve 95 per cent of its maximum 
air dry density in approximately 3.6 days. Applying the test 
proposed by Mundle and Chapman (2010), it is estimated that 
a rate of rise of approximately 5.1 m/a may be achievable for 
the magnetite tailings. Based on this result, a rate of rise of 
5 m/a was selected for evaluation by the current study.
Numerical modelling was carried out to assess whether a 
tailings storage facility containing magnetite tailings could be 
constructed at a rate of rise of 5 m/a. Modelling was carried 
out using the two-dimensional fi nite element software code 
Plaxis v9.0 (Plaxis bv, 2008). Plaxis allows the simulation 
of non-linear, time-dependent and anisotropic soils. In this 
specifi c context, the software allows us to approximate the 
geotechnical stability, accumulation of excess pore water 
pressures, deformation and settlements of an embankment 
that is raised in an upstream direction over time. 
Model geometry
The model was developed to simulate an upstream-raised 
tailings storage facility constructed progressively over a 
period of 16 years to a maximum height of 80 m. The design 
concept is based on the rotation of tailings deposition between 
two active tailings cells to achieve an average rate of rise of 
5 M/a. Tailings deposition is periodically diverted into a 
single cell while fi ll is placed on the fallow cell to construct the 
embankment raise. 
The embankment raises and starter embankments are 
constructed from waste rock, diverted to the tailings storage 
facilities by the mining operation. The use of waste rock in the 
embankments allows more rapid and lower cost construction 
than might otherwise be achievable using borrowed tailings or 
locally available fi ll materials. 
The geometry of the embankment raises is based on the 
need for safe and effi cient access by the mining fl eet. The 
starter embankment is constructed to a maximum height of 
20 m, and is subsequently raised in 10 m increments every 
two years. The embankments are designed with a minimum 
crest width of 50 m. The geometry of the modelled perimeter 
embankment is indicated on Figure 4.
The embankments are modelled with 2.5H:1V upstream and 
downstream slopes, constructed on a horizontal foundation. 
For simplicity, the tailings beaching has also been modelled 
as a horizontal plane with no freeboard. This simplifi cation of 
the model geometry is not expected to signifi cantly infl uence 
the results.
The ‘standard fi xities’ available using Plaxis were applied 
to the model, with verticalboundaries constrained in the 
horizontal direction and the base of the model constrained 
vertically.
Material parameters
Three discrete material types were modelled: foundation, 
tailings and waste rock. The self weight, strength, stiffness and 
permeability parameters of each of the materials were based 
on the results of laboratory testing and experience with similar 
materials. The material properties assumed for the analysis 
are summarised, together with the soil model used for each 
of the materials, in Table 2. The table includes a summary of 
equivalent material parameters for a typical gold tailings for 
comparison (the relevance of the gold tailings properties are 
discussed later).
The Soil Hardening model was used to simulate the 
behaviour of the tailings. The Soil Hardening model is a 
hardening-plasticity model that includes both shear and 
volumetric yield criteria. It allows modelling of the increased 
shear and volumetric stiffness exhibited by a soil under 
increasing effective pressures (ie depth). The model also 
allows modelling of decreasing shear stiffness of a soil under 
shear loading, and the generation of excess pore water 
pressures under strain. These particular features of the model 
allow the simulation of the consolidation and shear of tailings 
under upstream raises.
The one-dimensional consolidation stiffness modulus 
(E
oed
) of the magnetite tailings has been based on the results 
of consolidation testing, as plotted in Figure 5 for a range of 
effective stresses. These values are based on oedometer and 
Rowe Cell testing of the magnetite tailings. The shear stiffness 
(E
50
) was based on the results of a series of triaxial tests on the 
tailings. The E
50 
stiffness was generally found to be larger than 
the E
oed
, by a factor of about 1.5. This is within the expected range 
for the relationship between these two parameters as proposed 
by Vermeer (2001), and validates the results of testing.
FIG 4 - Plaxis model geometry.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
OVERCOMING THE CHALLENGES OF MANAGING TAILINGS FROM IRON ORE MINING IN WESTERN AUSTRALIA
39
The Mohr Coulomb model has been used to simulate 
the behaviour of the waste rock and foundation materials. 
This model is appropriate for modelling the behaviour stiff 
materials that are not expected to yield in shear, or undergo 
signifi cant consolidation, as is the case for these materials.
Figures 5 and 6 show the modelled stiffness and permeability 
relationships for the magnetite tailings. The models are based 
on oedometer and triaxial test data for a single magnetite 
tailings sample generated from pilot plant testing. The 
relationships are compared to those of a typical gold tailings 
from the Western Australian Goldfi elds.
The magnetite tailings are stiffer, and more permeable 
than the typical gold tailings. This increased stiffness results 
in reduced displacement under loading, while the higher 
permeability allows the dissipation of excess pore water 
pressures to occur more quickly. As a consequence, the 
magnetite tailings exhibit far better engineering properties 
as a foundation material than comparable gold tailings. It is 
Parameter Magnetite tailings Typical WA gold tailings Waste rock Foundation
Soil model Soil hardening Mohr-Coulomb Mohr-Coulomb
Bulk density (t/m3) 2.0 1.6 to 2.0 2.3 2.5
Friction angle (degrees) 35 30 to 35 40 45
Young’s Modulus (MPa) Varies with eff ective stress, as shown in Figure 5 20 50
Vertical permeability (m/s) Varies with eff ective stress, as shown in Figure 6 1 × 10-5 m/s 1 × 10-6 m/s
Horizontal/vertical permeability ratio 5 5 to 10 1 1
TABLE 2
Key model input parameters.
0
5,000
10,000
15,000
20,000
25,000
30,000
0 200 400 600 800 1000 1200
E o
ed
 M
od
ul
us
 (k
P
a)
Effective Stress (kPa)
Modelled Iron Ore Tailings Modelled Gold Tailings
Iron Ore Tailings - Consolidation Test Gold Tailings - Consolidation Test
FIG 5 - Modelled tailings stiff ness relationship.
 
1.0E 10
1.0E 09
1.0E 08
1.0E 07
1.0E 06
0 200 400 600 800
Ve
rt
ic
al
Pe
rm
ea
bi
lit
y
(m
/s
)
Effective Pressure (kPa)
Modelled Iron Ore Tailings Modelled Gold Tailings
Iron Ore Tailings Laboratory Values Select Gold Tailings Laboratory ValuesGold Tailings Laboratory Values
FIG 6 - Modelled tailings permeability relationship.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
D R ANSTEY AND D A REID
40
the observation of these superior engineering properties that 
has prompted the current investigation into the potential for 
increased rates of construction. 
Modelling procedure
The model is divided into discrete layers, as indicated on 
Figure 4. Each layer was modelled as a separate construction 
phase to simulate the development of the facility over time. 
Tailings were deposited progressively into each 10 m layer 
over a period of 630 days. Tailings deposition then ceased 
and the next stage of the embankment was constructed over a 
period of 100 days.
Plaxis linearly increases the loading of new layers over the 
time interval in which they are placed. In this way, it simulates 
the progressive increase in loading from the embankments and 
tailings as they are constructed and deposited, respectively. 
This is an advantage over some other software codes that have 
been used to model upstream construction, which require that 
each new layer of material be placed instantaneously (Dai and 
Pells, 1999).
The model was run in stages up to a fi nal height of 80 m. 
Excess pore pressures in the model were then allowed to 
dissipate to simulate post depositional settlement of the 
landform. 
As a point of comparison, the model of the staged 
embankment construction was re-run under the same 
conditions using the properties from the typical gold tailings.
Results and discussion
Figure 7 shows the fi nal deformed mesh from the model. 
The model indicates signifi cant displacements of the waste 
rock and tailings but has maintained its integrity with no 
signifi cant shear failure of the materials. The model indicates 
maximum cumulative settlements of up to 15 per cent 
(11.7 m) within the waste rock. However, it is predicted that 
most of this settlement will occur during construction of the 
embankments and, as a consequence, will be compensated for 
by the placement of additional fi ll.
Some differential vertical settlement may be expected across 
the upstream wall raises, owing to the lower stiffness of the 
tailings material underlying the inside of the embankment 
relative to the stiffness of the waste material underlying the 
outside of the embankment. It is expected that differential 
settlements of the magnitude anticipated would be 
compensated for by use of well graded granular fi ll materials, 
such as waste rock, and the placement of additional fi ll, as 
required during construction.
By comparison to the magnetite tailings model, the model 
run using gold tailings parameters experienced signifi cant 
shear failure during construction of the fi rst raise to the 
facility. Figure 8 shows the incremental displacements within 
the model at the time that the embankment failed.
Study outcomes
The study has examined the material properties of typical 
magnetite tailings. Finite element modelling has been used to 
 
FIG 7 - Deformed mesh for tailings storage facility model.
 
FIG 8 - Failure of embankment raise constructed on gold tailings.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
OVERCOMING THE CHALLENGES OF MANAGING TAILINGS FROM IRON ORE MINING IN WESTERN AUSTRALIA
41
assess the potential to raise magnetite tailings storage facilities 
at rates of rise greater than those typically experienced for gold 
tailings storage facilities in Western Australia. The following 
conclusions can be drawn from the study:
  magnetite tailings exhibit high stiffness and permeability 
relative to other tailings types, giving them improved 
engineering properties as the foundation for upstream 
raisedembankments; and
  modelling suggests that it may be possible to construct 
safe and stable upstream raised embankments on 
magnetite tailings at a rate of rise at least twice that of the 
conventional guidance for other tailings types in Western 
Australia. 
Limitations
Plaxis cannot model the impact of evaporation and desiccation 
of the deposited tailings on consolidation. Evaporation and 
desiccation typically lead to improved stiffness and strength 
characteristics for tailings. By excluding the infl uence of these 
factors, the results of the model will be conservative.
The modelling does not examine the stability of the tailings 
under earthquake loading conditions. Earthquake loading 
can induce additional excess pore pressures resulting in a 
reduction of strength and the potential for liquefaction in 
tailings. Predicting the liquefaction risk of a tailings material 
prior to its deposition is challenging. In situ testing of an 
existing tailings, using experience based screening methods 
outlined by Youd and Idriss (2001) is generally considered the 
most reliable method. Cyclic testing is one available method 
for investigating liquefaction potential, when only tailings 
samples are available. Further testing could be carried out to 
provide input parameters to examine the effect of earthquake 
loading on the tailings using Plaxis.
Tailings storage facilities are usually developed over a period 
of decades. Changes to tailings properties, mine schedules, 
and the observations and monitoring of the actual behaviour 
of the materials deposited will necessitate frequent revisions 
and updates to the type of analysis presented in this study. 
Regular in situ testing is an essential tool to confi rm and 
update the inputs required to assess the stability of a tailings 
storage facility.
IMPLEMENTING THE THEORY
The studies described in this paper have been undertaken to 
support feasibility-level studies and environmental permitting 
of tailings storage facilities in Western Australia. The next 
step is the implementation and monitoring of these solutions 
in practice.
Wind tunnel testing has been used to demonstrate the 
conditions under which respirable fi bres can be generated by 
magnetite tailings. This information can assist designers and 
operators to make appropriate decisions about the conditions 
under which a tailings storage facility should be constructed 
and operated. 
The potential for release of asbestos fi bres poses a signifi cant 
risk to workers and the public. It is therefore important that 
the effi cacy of the controls implemented for respirable fi bres 
be confi rmed through appropriate monitoring in the fi eld. 
Personal and low volume sampling should be carried out in 
accordance with the NOHSC guidelines (NOHSC:2002, 2005) 
and monitored against Australian occupational exposure 
standard of 0.1 fi bre/mL (NOHSC:2002, 2005). This 
information should be supplemented with the results of fi xed 
location sampling and compared against site records of wind 
speed, temperature and tailings storage operation. 
Site monitoring data combined with emission source 
estimates from wind tunnel testing can be used to develop 
a particulate emission model for a site. Such a model could 
act to support an early warning system for respirable fi bres 
to indicate when site conditions (such as wind speed or 
temperature) may make working conditions unsafe or when 
additional controls (such as dust suppressants) may be 
required.
Similar to asbestiform minerals, it is important that 
appropriate monitoring be put in place to confi rm the 
safety and stability of tailings storage facility embankments. 
Vibrating wire and standpipe piezometers should be used 
to monitor pore pressure accumulation in the tailings and 
embankments as the as the level of the tailings storage 
increases. These pore pressures should be compared to 
trigger levels determined from numerical modelling (such 
as that undertaken for the current study) to confi rm that the 
embankment maintains an acceptable factor of safety against 
failure. Survey monuments and photogrammetry should also 
be used to confi rm that deformation of the embankment is 
maintained within acceptable limits and is consistent with the 
results of modelling.
CONCLUSIONS
The management of iron ore tailings poses signifi cant risks 
and challenges to the mining industry. These risks are 
expected to come into particular focus in Western Australia 
over the coming decades as several large magnetite and CID 
deposits commence production.
The current study has demonstrated that wind tunnel 
testing has a role in the measurement of respirable fi bre 
generation from tailings and the design of appropriate 
control measures. Tests carried out on a sample of magnetite 
tailings containing asbestiform minerals indicated little 
propensity for the material to generate respirable fi bres when 
in a saturated or dry state. However, tailings that have been 
disturbed, either through pulverisation or abrasion, have the 
potential to generate respirable fi bres at signifi cantly higher 
concentrations than undisturbed tailings.
The results of the wind tunnel testing can be applied to the 
development of operating procedures and control measures 
for the management of tailings that contain asbestiform 
minerals. They can also be used in the development of 
dispersion models to provide confi rmation that emissions will 
satisfy the national occupational exposure standards. In this 
way, wind tunnel testing can be used to improve the safety and 
environmental outcomes for the operation of iron ore tailings 
storage facilities.
Laboratory testing and numerical modelling of magnetite 
tailings has been carried out to justify acceptable rates of rise 
for magnetite tailings. The modelling suggests that magnetite 
tailings storage facilities can be constructed using upstream 
raises at rates of rise of 5 M/a, twice that of the typical guidance 
for tailings in Western Australia. The implications are that the 
footprint area, associated clearing, environmental impacts 
and construction costs for such a facility could be reduced by 
half. Where the footprint area available for tailings storage is 
restricted, the ability to sustain increased rates of rise using 
upstream construction methods provides signifi cant cost 
advantages over alternative construction methods, such as 
downstream construction.
It has been demonstrated that the application of two 
innovative technologies to iron ore tailings can assist in the 
development of improved environmental and economic 
outcomes for tailings management in Western Australia.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
D R ANSTEY AND D A REID
42
ACKNOWLEDGEMENTS
The authors would like to thank Mr Chad Mundle for the 
many hours spent designing, assembling and experimenting 
with the Golder Associates wind tunnel. We would also like 
to thank Dr Frank Fleer and Mr Peter Chapman for their 
advice and contribution to our understanding of the concepts 
presented in this paper.
REFERENCES
AS/NZS 60335.2.69, 2003. Household and similar electrical 
appliances – safety – particular requirements for wet and dry 
vacuum cleaners, including power brush, for industrial and 
commercial use, Standards Australia, 30 May.
Dai, C and Pells, P J N, 1999. Two-dimensional large strain 
consolidation prediction and incrementally deposited 
tailings, FLAC and Numerical Modelling in Geomechanics 
(eds: C Detourney and R Hart), pp 123-131 (Rotterdam).
Department of Mines and Petroleum (DMP), 2010. 
Management of fi brous minerals in Western Australian mining 
operations, report.
International Commission on Large Dams (ICOLD), 1995. 
Tailings dams, transport placement and decantation, Bulletin 
101, 57 p.
International Commission on Large Dams (ICOLD), 2001. 
Tailings dams risk of dangerous occurrences, Bulletin 121, 
Appendix 4. 
Intierra Mapping, 2010. Australian iron ore mining and exploration 
activity,AusIMM Bulletin, No 5 October 2010.
Li, C, Sun, H, Bai, J and Li, L, 2010. Innovative methodology 
for comprehensive utilization of iron ore tailings: Part 1, The 
recovery of iron from iron ore tailings using magnetic separation 
after magnetizing roasting, Journal of Hazardous Materials, 
174(1-3):71-77.
Mehta, R D and Bradshaw, P, 1979. Design rules for small low 
speed wind tunnels, The Aeronautical Journal of the Royal 
Aeronautical Society, pp 443-449.
Mundle, C A G and Chapman, P J, 2010. Selection of parameters 
for semi-arid tailings storage facility design, in Proceedings 
First International Seminar on the Reduction of Risk in the 
Management of Tailings and Mine Waste (University of Western 
Australia).
National Occupational Health and Safety Commission 
(NOHSC):2002, 2005. Code of practice for the safe removal of 
asbestos, second edition, April.
National Occupational Health and Safety Commission:3003, 
2005. Guidance note on the membrane fi lter method for 
estimating airborne asbestos fi bres, second edition, April.
Neuman, C M, Boulton, J W and Sanderson, S, 2008. Wind tunnel 
simulation of environmental controls on fugitive dust emissions 
from mine tailings, Atmospheric Environment, doi:10.1016/j.
atmosenv.2008.10.011.
Plaxis bv, 2008. Plaxis 9.0 professional version.
Vermeer, P A, 2001. On the parameters of the hardening soil model, 
Experienced Plaxis Users, 26 - 28 March.
Vick, S G, 1990. Planning, Design, and Analysis of Tailings Dams 
(Bitech Publishers: Vancouver).
WISE Uranium, 2011. Chronology of major tailings dam failures 
[online]. Available from: <http://www.wise-uranium.org/mdaf.
html> [Accessed: 1 February 2011].
Youd, T L and Idriss, I M, 2001. Liquefaction resistance of soils: 
summary report from the 1996 NCEER and 1998 NCEER/NSF 
Workshops on Evaluation of Liquefaction Resistance of Soils, 
Journal of Geotechnical and Geo-environmental Engineering, 
ASCE, 127(10):297-313.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 43
INTRODUCTION
Iron Ore Holdings Ltd (IOH) is an emerging junior iron ore 
company based in Perth that is actively exploring tenements in 
the Central and Western Pilbara region of Western Australia. 
The company has a number of iron ore deposits in the Pilbara 
region, and the Phil’s Creek Project was the f rst tenement 
to undergo the approvals process. The proposal attained 
approval, but the Project has now been placed on hold by IOH.
The Phil’s Creek iron ore resource is located approximately 
90 km north-west of Newman and 140 km east of Tom Price 
in the East Pilbara region of Western Australia. The project 
was to consist of a shallow open cut mine located above the 
water table, producing approximately seven million tonnes 
of crushed and screened iron ore product over a f ve year 
period (1.5 Mt/a). An ore crushing plant, associated mine 
infrastructure and a small accommodation village were also 
to be constructed on-site. URS Australia Pty Ltd (URS) has 
been assisting IOH since mid-2008 with preparation of 
environmental assessment documents.
IOH has faced several challenges in the environmental 
assessment process for the Phil’s Creek Project, some of 
which led to signi f cant delays in f nalising environmental 
approvals. Two major environmental challenges involved 
the environmental assessment of potential impacts on 
troglofauna and vertebrate fauna, and working with State and 
Commonwealth regulators through these processes.
As a result of meeting and overcoming these and other 
challenges in the environmental assessment process, IOH 
has gained useful knowledge of the current requirements 
of regulators, particularly in relation to developing 
environmental offsets and impact assessment of troglofauna.
Using the Phil’s Creek Project as a case study, this paper 
will discuss the challenges IOH faced in the environmental 
assessment process and the steps undertaken to overcome 
them. These challenges are generic to the mining industry 
and, undoubtedly, all proponents of signif cant developments.
BACKGROUND
The Phil’s Creek Project is located in close proximity to 
numerous iron ore mines in operation and exploration phases 
in the Pilbara region, but is remote from urban settlements. 
1. Senior Environmental Scientist, URS Australia Pty Ltd, Level 4, 226 Adelaide Terrace, Perth WA 6000. Email: hannah_fl etcher@urscorp.com
2. Principal Environmental Scientist, URS Australia Pty Ltd, Level 4, 226 Adelaide Terrace, Perth WA 6000. Email: jenny_moro@urscorp.com
Iron Ore Holdings Ltd – Case Study 
of the Environmental Approval 
Challenges for an Emerging Junior
H Fletcher1 and J Moro2
ABSTRACT
Iron Ore Holdings Ltd (IOH) is an emerging junior iron ore company that is actively developing 
projects in the Central and Western Pilbara region of Western Australia. IOH commissioned URS 
Australia Pty Ltd (URS) in mid-2008 to prepare environmental assessment documents for the 
Phil’s Creek Project, which was rapidly moving into the development phase. This project comprised 
a small open cut iron ore mine that was expected to deliver approximately seven million tonnes 
over a f ve year mine life.
IOH faced several challenges in the environmental assessment process for the Phil’s Creek Project. 
Although the Phil’s Creek Project attained approval but has now been placed on hold, it is a useful 
case study to discuss these challenges and how IOH addressed them. Several important lessons 
were learned through this process that can assist other proponents in addressing environmental 
challenges in a constructive and proactive manner.
One major challenge faced by IOH related to the endangered northern quoll ( Dasyurus 
hallucatus). Another challenge was the unusually high species richness of troglofauna recorded 
within the project area, with several species initially known only from within the proposed pit 
footprint.
Meeting and overcoming these challenges has provided IOH with an increased understanding of 
the level of input required for consultation with government departments regarding environmental 
assessments, and this should stand the company in good stead for developing future tenements. 
Challenges in regards to the environmental assessment process may be faced during future 
developments, but IOH’s experiences to date and the lessons learned from the Phil’s Creek Project, 
will help pave the way for successfully overcoming them.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
H FLETCHER AND J MORO
44
The iron ore deposit is located within a green f elds site which 
is typical for the East Pilbara region, being characterised by 
rocky hills, small gorges, seasonal watercourses and gravelly 
loams. The current land use is pastoral and approximately 
100 ha of vegetation was proposed be cleared within a 170 ha 
envelope for the Project.
In order for IOH to construct and operate the Phil’s Creek 
Project, IOH was required to seek environmental approval 
under the Environmental Protection Act 1986 (EP Act) and 
the Environmental Protection and Biodiversity Conservation 
Act 1999 (EPBC Act). A number of baseline environmental 
studies were undertaken for the Project for the environmental 
impact assessment, ranging from biological baseline studies 
of f ora and fauna, to noise and waste characterisation 
assessments.
Projects that are likely to have a signi f cant impact on a 
number of environmental factors or where there is strong 
public or community interest in the project, usually require 
formal assessment under Part IV of the EP Act. During initial 
scoping of the Project, it was thought unlikely that the Project 
would meet the criteria for formal assessment; however 
this needs to be determined via referral to the Western 
Australian Environmental Protection Authority (EPA). The 
EPA determined that the Project did not require formal 
assessment, instead requiring assessment under Part V of the 
EP Act as a works approval and licence, a clearing permit and 
assessment under theMining Act 1978 as a Mining Proposal.
IOH also undertook a desktop EPBC Act protected matters 
search in order to determine the requirement for referral 
under the EPBC Act. Whilst the protected matters search 
identif ed that a number of threatened species may occur in 
the vicinity of the Project area, (including the endangered 
northern quoll [ Dasyurus hallucatus]) the size and scope of 
the Project, as well as identi f cation of minimal impact on 
known northern quoll habitat, suggested that Project impacts 
would not be signi f cant. Nevertheless, IOH decided to refer 
the Project to the DSEWPaC in order to provide company 
certainty that IOH would not contravene the EPBC Act if a 
referral was not made, and also to mitigate against time delays 
if a third party referred the Project before IOH.
The Commonwealth Minister for Environment determined 
that the potential presence of the northern quoll was suff cient 
to make the Project a controlled action and require assessment 
under the EPBC Act. This decision was unexpected, and 
highlights the low level of tolerance of the DSEWPaC to even 
potential impacts on species having an endangered status 
at the Commonwealth level. This assessment became one of 
the major challenges of the Phil’s Creek Project in meeting 
company timelines.
APPROVAL CHALLENGES AND SOLUTIONS
IOH faced a variety of challenges during the environmental 
assessment process of the Phil’s Creek Project. The challenges 
discussed in this paper are:
  scoping environmental assessment studies at a time when 
the proposal is still under development,
  regulator perception of a new operator into the iron ore 
industry,
  ensuring that all project team members understand their 
contribution to facilitating the environmental assessment 
process,
  the State environmental assessment process in regards to 
potential impacts on subterranean fauna (troglofauna), 
and
  the Commonwealth environmental assessment process 
in regards to potential impacts on an endangered species 
(the northern quoll).
This section details how these challenges impacted IOH in 
regards to the Phil’s Creek Project, and the steps that IOH 
undertook to reach a solution to each of these challenges.
Challenge 1: project scoping
The f rst challenge of the Phil’s Creek Project was to scope the 
environmental assessment of the Project, at a stage when the 
Project description was not fully developed. IOH also required 
price certainty in relation to study costs and a relatively f rm 
assessment timeline in order for the Project to be ready for 
construction by a set date.
All projects are challenged by the need to initiate the 
environmental assessment process while the project 
engineering design is still being completed, and the two 
processes often occur in parallel. IOH commissioned URS 
to prepare an environmental scoping study to identify what 
the environmental factors of greatest risk to the Project were 
likely to be, as well as potential risks to the Project associated 
with obtaining environmental approval. The purpose of the 
environmental scoping study was to provide IOH with an 
initial understanding of the key environmental factors that 
needed to be addressed early in Project development. It also 
provided IOH with indicative costs and timing for a range of 
baseline studies to be undertaken.
A preliminary risk assessment was undertaken to identify all 
risks to the Project including, but not limited to, environmental 
approvals. The risk assessment was held as a workshop with 
personnel from different disciplines participating, which 
encouraged information sharing amongst the wider Project 
team. A bene f t of the risk workshop was that it identi f ed 
what the greatest time and budget constraints of the Project 
were likely to be and enabled IOH to establish which studies 
were on the critical path for development. Importantly, the 
risk assessment identi f ed that both subterranean fauna and 
vertebrate fauna posed a high to medium risk to the Project. 
These are discussed below as additional challenges to the 
Project.
Early identi f cation of costs enabled IOH to budget 
appropriately for the baseline studies as part of the overall 
Project costs. IOH also recognised during Project scoping 
that as the study of subterranean fauna is still a new science, 
additional budget for this baseline study may be required. In 
the event, additional troglofauna studies were subsequently 
required and the provision of a budget contingency by IOH 
allowed the assessment to proceed without delay.
The project scoping phase and risk assessment provides 
benef t to all companies as a f rst step in developing a project, 
even at an early stage of the project description, as it identif es 
key risks and establishes the critical path for development. 
The project scoping phase should identify the likely project 
environmental costs and timing to enable early budget 
planning and project scheduling. It is recommended that a 
budget and timing contingency is established for high risk 
environmental factors in case additional studies are required 
at a later date.
Challenge 2: company track record
Lack of track record in managing environmental issues is a 
challenge that is dif f cult for any new company to overcome; 
especially as environmental assessment documentation seeks 
information from companies regarding their environmental 
reputation and compliance record. For example, the DSEWPaC 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
IRON ORE HOLDINGS LTD – CASE STUDY OF THE ENVIRONMENTAL APPROVAL CHALLENGES FOR AN EMERGING JUNIOR
45
referral form requires the company proposing to develop a 
project to provide information regarding their environmental 
history. This includes whether the company has a satisfactory 
record of responsible environmental management, if the 
company has ever been subject to proceedings under a 
Commonwealth law for the protection of the environment, 
and if the company will ensure that the project is undertaken 
in accordance with its environmental policy and planning 
framework.
In the absence of any existing environmental record, 
IOH engaged URS to prepare environmental assessment 
documents and to initiate consultation with government 
regulators until IOH established relationships with key 
personnel. IOH fostered its relationship with government 
regulators by attending key meetings, which also showed the 
company’s commitment to the environmental assessment 
process. The role of experienced environmental professionals 
is essential to introducing a new proponent to environmental 
regulators and to facilitating the approvals process. A strong 
co-operative partnership that seeks contributions from all 
appropriate disciplines can foster better environmental 
outcomes.
During preparation of the Mining Proposal for assessment 
by the Department of Mines and Petroleum (DMP), the IOH 
Project team took the lead in developing environmental 
commitments for the Project. The involvement of IOH’s 
geology and engineering personnel in this process enhanced 
corporate understanding of the role that other disciplines 
can have in fostering better environmental outcomes. This 
contribution ensured that the environmental commitments 
which IOH developed were realistic and achievable for the 
company. The commitments also demonstrated that IOH was 
taking environmental matters seriously and took into account 
prior agreements with stakeholders. The involvement by 
other disciplines in this process can also encourage positive 
environmental attitudes that are carried forward into the 
development and operational phases of a project.
Ultimately, a company track record and reputation 
for responsible environmental management is built by 
involvement of the company project team, with support at 
management level. Management needs to demonstrate that 
environmental commitments will be taken seriously and 
these attitudes f lter down to other levels of an organisation.An added bene f t from the Phil’s Creek Project was that 
management were part of the Project team and therefore 
gained a f rst-hand experience of regulatory processes and 
regulator expectations.
Challenge 3: environmental knowledge
As a junior mining company moving from exploration to 
project development, IOH had a small exploration and 
development team with mainly geology and engineering 
backgrounds, but no in-house environmental team. This was a 
challenge to IOH because following completion of the scoping 
study the Project was rapidly progressing beyond exploration, 
and baseline studies were ready to be commenced.
To address this challenge, IOH engaged a single consultant 
to manage baseline surveys and commence preparation of the 
Project environmental assessment documents in order to f ll 
the gap in company environmental knowledge. By engaging 
URS who has experience with undertaking environmental 
assessments for mining projects, IOH was able to utilise 
URS’s prior experience with preparing environmental impact 
assessments for iron ore projects. URS was also able to 
initiate consultation with government regulators until IOH 
established a relationship with government staff.
During the early stages of commencement of baseline 
studies, IOH chose to second two URS environmental 
personnel to work full time on the Project. A secondment can 
provide many bene f ts, such as creating a team environment 
where environmental information is shared, and enable 
Project tasks to be re-prioritised according to changing 
business needs. Throughout this Project, the IOH Project 
team sought advice from URS on any changes to regulatory 
requirements that could affect the Project. This gave IOH an 
early and ongoing understanding of the complexities within 
the environmental assessment process and the changing level 
of effort required to meet regulators expectations.
A key aspect of the success of this approach is for proponents 
to select a single consultant as their environmental project 
manager. Experience during the previous mining boom was 
that a number of proponents retained multiple environmental 
consultants but without clear accountability. This meant that 
as project delays or cost overruns occurred it was sometimes 
diff cult to determine which consultant was accountable. By 
selecting a single consultant to project manage environmental 
approvals IOH were always able to establish for a single point of 
contact regarding any project delays or cost over-runs.
Understanding environmental constraints and requirements 
is an ongoing challenge, but seconding environmental 
consultants can f ll a gap in client knowledge, especially 
in the early stages of a project. For a small company, the 
incorporation of the wider project team in key meetings and 
workshops can lessen the likelihood of oversights in project 
design and creates a greater awareness of key environmental 
constraints. It can also be an opportunity to think about 
how a change to the project design could impact upon the 
environment and vice versa.
It is important for proponents to be mindful that the 
environmental assessment process is typically a dynamic 
process, and does not always follow a logical pathway like 
engineering. This can be a challenge to new companies, and 
requires a f exible approach by the project team as well as good 
communication between the project teams and management.
Challenge 4: State environmental assessment 
of troglofauna
Troglofauna are air-breathing subterranean fauna living 
underground in the air spaces within small f ssures and 
cavities of the underground matrix, and differ from the other 
type of subterranean fauna, stygofauna (Bennelongia, 2009). 
Stygofauna are aquatic and live in air spaces in the small f ssures 
and cavities in groundwater aquifers (Bennelongia, 2009). 
Troglofauna are most signif cantly impacted by loss of habitat 
(excavation and mining of an ore deposit), whilst stygofauna 
and are predominately impacted by changes to groundwater 
such as from mine dewatering. A troglofauna assessment was 
deemed necessary for this Project due to excavation of the pit 
which would remove and disturb troglofauna habitat. As no 
groundwater is proposed to be abstracted for this Project, a 
stygofauna assessment was not required.
Obtaining an approval decision from the State regulators 
in regards to potential impacts on troglofauna was a 
major environmental challenge to the Phil’s Creek Project 
proceeding. This is because the study and assessment of 
subterranean fauna (particularly troglofauna), is still a 
relatively new science and there is often uncertainly about 
what level of sampling effort is acceptable to regulators. There 
is also the additional challenge that proponents often require 
troglofauna studies to be initiated while the project description 
is still being developed. This may result in a change to the 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
H FLETCHER AND J MORO
46
project footprint and therefore a change in potential impacts 
to troglofauna later on in the project.
The EPA has issued a guidance statement that advises 
proponents on what time of year to undertake troglofauna 
surveys and the minimum number of samples required 
(Environmental Protection Authority, 2007). If sampling is 
undertaken in accordance with this guidance statement, the 
level of effort is usually accepted by the EPA as an adequate 
initial investigation of troglofauna. However, the body of 
scientif c data on troglofauna is still relatively small and there 
is uncertainty at the project scoping stage on whether or not a 
particular project site will encounter troglofauna conservation 
issues. This often leads to the requirement for multiple 
sampling phases over months or years.
For the Phil’s Creek Project, the initial troglofauna sampling 
recorded an unusually high species richness within the 
Project area, with several species known only from within the 
proposed pit footprint. IOH consulted with the Department of 
Environment and Conservation (DEC) after the preliminary 
troglofauna results and sought advice on how to proceed. IOH 
encountered a few site issues with the additional troglofauna 
sampling, such as exploration boreholes that were unsuitable 
for sampling and constraints imposed on sampling locations 
due to existing Aboriginal heritage agreements. However, 
consultation with the DEC resulted in agreement being given 
to IOH to obtain further samples from other IOH tenements 
within close proximity of Phil’s Creek, in order to support the 
environmental impact assessment. This challenge was f nally 
addressed when the DEC advised that no further sampling 
was required.
Troglofauna is currently a key environmental issue for 
regulators and in order for a company to undertake an 
acceptable troglofauna survey, early and ongoing consultation 
with the DEC is required. The DEC provides advice on whether 
the survey methodology is acceptable before time and money 
is spent, and provides comment on survey reports prior to 
inclusion in environmental assessment documents.
Proponents need to be aware at project scoping that 
subterranean fauna surveys, and troglofauna surveys in 
particular, can take considerable effort in order to demonstrate 
that a project will be environmentally acceptable. Studies 
should commence early in project development as sampling 
may take greater than a year to collect the required data – for 
example six rounds of sampling occurred over a 16 month 
period for the Phil’s Creek Project. The best case scenario is 
that initial troglofauna studies commence as early as possible; 
then if additional studies are undertaken they are able to 
be undertaken concurrently with other baseline studies, to 
ensure the overall timing of the project is not compromised.
Adequate money also needs to be allocated by proponents to 
undertake troglofauna sampling and consult with regulators. 
It is recommended that a contingencybudget is put aside in 
the event that additional studies are required, as was the case 
for the Phil’s Creek Project. 
Challenge 5: Commonwealth environmental 
assessment of the northern quoll
Obtaining an approval decision from the Commonwealth 
regulators in regards to vertebrate fauna (the northern quoll) 
was another major environmental challenge to this Project 
proceeding. During Project scoping, IOH identif ed that there 
was a high risk that a third party (such as a stakeholder) may 
refer the Project to the DSEWPaC following referral to the 
EPA, and this would lead to Project delays. 
Following a legal review, IOH chose to refer the Project to 
the DSEWPaC even though the Project did not appear to meet 
the criteria for assessment under the EPBC Act as a controlled 
action. IOH referred the Project to the DSEWPaC at the same 
time as the referral to the EPA was submitted. The decision 
for making the referral was both to ensure a greater level of 
company certainty that the Project did not require assessment 
by the DSEWPaC, and to avoid the risk of a third party 
referring the Project before IOH had to a chance to make a 
referral.
As discussed above, the EPA determined that the Project 
did not require formal assessment under Part IV of the EP Act 
and could be assessed under a lower level of assessment by 
the DMP and DEC. Projects that require assessment by the 
EPA are more likely to also be controlled actions requiring 
assessment under the EPBC Act. Such projects are usually 
large-scale developments with a greater likelihood of having 
signif cant environmental impacts.
For the Phil’s Creek Project, however, the assessment levels 
were different between the EP Act and the EPBC Act. The 
controlled action decision by the DSEWPaC was based on 
the likelihood of the Project having a signi f cant impact on 
the endangered northern quoll ( Dasyurus hallucatus ). The 
decision was appealed because the northern quoll was not 
recorded in or around the Project area during the baseline 
f eld survey and the vertebrate fauna report stated that there 
were no vertebrate fauna habitats of local signif cance present 
within the Project area (Western Wildlife, 2009).
However, the decision was upheld and IOH was required to 
development a speci f c research plan to minimise, manage, 
mitigate and offset potential impacts on the northern quoll. 
The decision to make the Project a controlled action was not 
foreseen by IOH or URS and this presented challenges to the 
approvals timeline.
IOH sought clari f cation from the DSEWPaC during the 
assessment process regarding the Project’s potential impact 
on the northern quoll, as the 100 ha proposed to be disturbed 
for the Project did not appear to constitute a signi f cant 
impact on a species that was not known to occur in the vicinity 
of the Project area. IOH assessed that only f ve ha of potential 
northern quoll habitat was present in the Project area (and 
therefore potentially disturbed), and sought comment from the 
DSEWPaC if this constituted a signi f cant impact. However, 
the DSEWPaC determined that the entire area of disturbance 
of the Project had the potential to impact upon the northern 
quoll. The DSEWPaC further advised during consultation that 
it was concerned about impacts on the northern quoll in the 
entire Pilbara region due to a retraction in the species range 
over the north of Australia, from impacts such as predation 
from the cane toad, infectious diseases and inappropriate 
burning regimes, which were threatening the species survival.
After the Project was approved, IOH and URS became 
aware of a draft internal DSEWPaC guideline for signi f cant 
impacts on the northern quoll, which likely formed the basis 
for the DSEWPaC’s decision. Although not published, it is 
highly likely that all future projects in the Pilbara region will 
be affected by this guideline as the DSEWPaC has determined 
that the entire Pilbara region is potential northern quoll 
habitat.
The outcome of this decision identif es that the DSEWPaC is 
exhibiting a low level of tolerance on impacts to endangered 
species from the development of projects. The DSEWPaC 
appears to be taking a conservative approach in regards 
to potential impacts to endangered species that may only 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
IRON ORE HOLDINGS LTD – CASE STUDY OF THE ENVIRONMENTAL APPROVAL CHALLENGES FOR AN EMERGING JUNIOR
47
potentially occur in an area, and making its assessment 
decision accordingly. Therefore, proponents should be aware 
that regardless of the scale of a development, if a project is 
located in the Pilbara region, it may meet the criteria to be 
a controlled action under the EPBC Act and this should be 
factored into project timing and cost. Indeed, wherever the 
desktop EPBC Act protected matters search identi f es the 
possible presence of endangered species, proponents should 
be aware of the potential for the project to be assessed as a 
controlled action regardless of the extent of any impact.
A contributing factor during environmental impact 
assessment was that the DSEWPaC identif ed a relatively high 
degree of uncertainty as to the behaviour and habitat of the 
northern quoll in the Pilbara. Therefore even though only 
f ve per cent of the disturbance area was identi f ed as being 
probable northern quoll habitat, the DSEWPaC recommended 
a precautionary approach to the other 95 per cent of the 
Project area in making their recommendation to the Minister. 
The DSEWPaC also advised IOH that the baseline vertebrate 
fauna survey undertaken for the Project did not meet its 
preferred timing, even though the survey was undertaken in 
accordance with the EPA’s guidance statement timing and 
methodology. It will therefore be important for proponents to 
discuss the proposed timing for vertebrate fauna surveys with 
the DEC and/or the DSEWPaC prior to undertaking any f eld 
work, to ensure that it also meets the DSEWPaC’s preferred 
timing.
Further challenges were met during assessment under 
the EPBC Act, with the greatest challenge being uncertainty 
regarding the total time that the assessment would take. 
IOH had to re-prioritise some other tasks undertaken by the 
Project team in order to focus on the assessment process. Also, 
some aspects of the assessment under the EPBC Act do not 
have statutory time frames, such as IOH being responsible for 
preparing advertisement of the Project for public comment, 
which took additional time.
To try and limit the time this assessment would take, IOH 
and URS followed up the DSEWPaC assessment of f cer on a 
regular basis to ensure that progress was being made on the 
Project and to address any issues at early stage. 
A win-win of the assessment process was that the DSEWPaC 
sought feedback from IOH and URS on the recommended 
draft conditions. Although this took some time to negotiate, it 
was benef cial to IOH to be given the opportunity to comment 
and make revisions on the draft recommended conditions for 
the Project. IOH was able to seek amendments to the wording 
of some approval conditions to ensure that the conditions 
of the Project did not unreasonably constrain the company, 
whilst still meeting the DSEWPaC’s expectations.
Importantly, the environmental offset required as 
conditional approval of the Project by the DSEWPaC did not 
involve a like-for-like land swap, and involved a number of 
actions to be undertaken throughout the life of the Project. 
The offset conditions required a f nancial commitment over 
a three-year period to implement a research study on the 
northern quoll speci f cally in the Pilbara region, as well as 
undertaking staff induction and environmental awareness 
programs, maintaining a threatened species register, and 
retaining buffers of undisturbed land along drainage lines 
(potential northern quoll habitat). Proponents therefore need 
to consult with regulators and species experts in order to 
determine the key factorsthat may be addressed as part of the 
offsetting process.
Overall, the main challenge of being assessed by the 
Commonwealth was that it was identi f ed during the project 
scoping phase that assessment under the EPBC was unlikely, 
as the Project did not appear to meet the EPBC Act assessment 
criteria. Although the Project was assessed at the lowest 
level of assessment by the DSEWPaC, the process was time 
consuming and took a total of ten months, which ultimately 
impacted on Project timelines. However, IOH could not have 
taken the risk that a third party would refer the Project to the 
DSEWPaC and could not have been aware of the DSEWPaC’s 
internal position on the northern quoll. For future projects, 
IOH will consult with the DSEWPaC early in the scoping 
process after having undertaken a desktop EPBC Act protected 
matters search. Where DSEWPaC advise that a controlled 
action may be required, IOH will allow appropriate time and 
budget for assessment under the EPBC Act, regardless of size 
and potential impacts.
LESSONS LEARNT
This paper has used the Phil’s Creek Project as a case study to 
discuss the challenges facing an emerging iron ore company. 
Through meeting the challenges that IOH has faced, the 
company has attained key lessons regarding the complexities 
of the Western Australian and Commonwealth environmental 
assessment processes, and identi f ed how to proactively and 
eff ciently overcome these challenges for future projects.
These lessons gained are not unique to the Phil’s Creek 
Project or to IOH, and apply to all iron ore companies. They 
include the following:
  Undertake an early multi-disciplinary risk assessment for 
all factors, including environmental, at the project scoping 
stage. Ensure appropriate time and budget contingencies 
are allocated.
  The use of established consultants and involvement of key 
project team members in consultation with government 
regulators can assist to allay regulator concerns. Consult 
early and often with regulators as this promotes buy-
in of the environmental aspects of the project to the 
government regulators, and builds the relationship with 
key government staff for future projects. It also ensures 
that the government assessment of f cer remains familiar 
with the project when future discussions are held.
  Proponents cannot pre-empt a regulator’s decision on 
whether or not to assess a project, and need to factors 
in contingencies. However, early proactive consultation 
with regulators especially following identi f cation of key 
environmental factors, should help to quantify contingency 
requirements.
  Where a proponent does not have internal expertise it is 
recommended that they appoint a single consultant as 
their environmental project manager. Secondment of 
consulting staff at an early stage allows project dedicated 
staff to build strong relationships with the company’s 
team.
  Maintain awareness that emerging environmental factors 
(such as species being placed on the endangered species 
list) may lead to project delays, until a suf f cient body of 
knowledge is available to allow some level of predictability 
in assessment outcomes.
  Develop environmental commitments at a company 
level with the support of environmental specialists 
where required, to ensure that the commitments are 
realistic and achievable for the company. Build a positive 
environmental awareness company-wide from project 
concept to development.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
H FLETCHER AND J MORO
48
  Proponents need to be adaptable to project changes 
following the development phase, such as modifying 
the project design at the suggestion of regulators during 
consultation or recommendations from subconsultants in 
baseline reports.
  Proactively consult with regulators and topic experts in 
order to understand key environmental factors. This will 
provide proponents with an understanding of the potential 
options for management, mitigation and offset of these 
key environmental factors.
It has been discussed that during the environmental 
assessment process there are often competing project 
demands. A proponent needs to design a project in enough 
detail in order for it to be acceptable for environmental 
assessment, as well as being f exible enough to make design 
changes as business demands change or as required during 
consultation with regulators and other stakeholders. Adaptive 
management therefore became a solution to many of IOH’s 
Project challenges. If proponents adequately scope a project 
for the required environmental studies and assessments, 
allow for realistic costs and time frames for undertaking 
studies and preparing assessment documents, and are f exible 
enough to modify aspects of the project design as potential 
environmental issues arise, this can ultimately streamline the 
environmental assessment process.
The challenges discussed are ones that exist for all mining 
companies, particularly in the Pilbara region, and the lessons 
learnt are intended to help all proponents successfully 
overcome these and other challenges.
ACKNOWLEDGEMENTS
The authors would like thank a number of people for their 
valuable assistance in helping to get this paper submitted. 
Firstly, past and current members of the IOH Project team for 
being supportive of this paper and their willingness to assist 
with the wording of the paper.
Thank you to the current URS Project team, many of who 
assisted with developing the structure of the paper and 
conducting reviews, particularly Karen Ariyaratnam, Donna 
Pershke, Mark Goldstone, Madolyn Morel and Beverley 
Turner. Thanks also to the previous members of the URS 
Project team for suggesting paper topics and undertaking 
reviews of information, particularly Diana Oates, Tanya 
Carpenter and Blair Hardman.
REFERENCES
Bennelongia Pty Ltd , 2009. Iron Ore Holdings Ltd Phil’s Creek 
Project: Troglofauna Assessment, report 2009/70, prepared 
for URS Australia Pty Ltd on behalf of Iron Ore Holdings Ltd, 
December 2009.
Environmental Protection Authority , 2007. Guidance for the 
assessment of environmental factors: sampling methods and 
survey considerations for subterranean fauna in Western Australia 
(draft) No 54Aa, technical appendix to Guidance Statement No 54 
(Environmental Protection Authority, August 2007).
Western Wildlife , 2009. Phil’s Creek Project Area: fauna survey 
2008, prepared for URS Australia Pty Ltd on behalf of Iron Ore 
Holdings Ltd, May 2009.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 49
INTRODUCTION
Reserves of high-grade iron ores are reducing all over the 
world; a result of the increase in demand for steel historically 
from Japan, India and more recently China. Benefi ciation 
techniques for iron ores are receiving world-wide attention 
due to development of new deposits of lower grade ore to 
satisfy the demand. The recovery of mineral value from the 
low-grade ores is more expensive than selling direct shipping 
ores (DSO) ores, due to higher energy costs of liberation 
and separation as well as capital costs, requiring the use of 
benefi ciation to produce high-grade concentrates from low 
grade ores.
Benefi ciation can be defi ned as a variety of processes 
whereby extracted ore is separated, by physical means, into 
mineral and waste streams. Most benefi ciation operations will 
result in the production of three materials:
1. a saleable concentrate; 
2. a middlings product (a low-grade stream that is either 
reprocessed or stockpiled); and 
3. a tailing stream, which is discarded.
To maintain a constant quality ore, companies need to 
differentiate ore grades more carefully and upgrade and blend 
grades produced in the mines. Because of the more diffi cult 
processing of this lower grade compared to the iron ores 
currently being processed, more modern technology will have 
to be considered.
The most energy intensive stage, the reduction step in the 
blast furnace, is mainly drivenby the requirement of carbon 
as a reducing agent. In order to diminish the environmental 
impact, careful control of the phases involved in the process is 
important for minimising the energy consumption.
Elemental analysis such as XRF or wet chemistry of samples 
are essential in the characterisation of iron ore feedstocks (and 
their sintered products), but since the metallurgist is almost 
always faced with the task of separating and concentrating 
minerals rather than elements, the mineralogy assumes great 
importance at an early stage.
However, the mineralogy of the sample is often not 
considered or at best inferred from visual geological logging 
of the sampled material. As such, mineral contaminants that 
cannot be identifi ed visually and/or by chemical analysis 
alone are often not quantifi ed. For example the identifi cation 
of gibbsite via chemical data is typically problematic, being 
complicated by the co-occurrence of kaolinite and aluminous 
goethite. The correct identifi cation of gibbsite is important 
as gibbsite can have a signifi cantly adverse impact on sinter 
effi ciency of the iron ore with gibbsite-bearing ores being 
likely to require longer sintering times and higher sintering 
temperatures to promote melt formation. This is due to the 
poor reactivity of this type of alumina and the high viscosity 
of the primary melt formed. As a result fuel rate is reported 
to increase and the sintering productivity to decrease as the 
alumina content increases (Yamaoka, 1974; Lu, Holmes and 
Manuel, 2007).
A further benefi t of resolving the mineralogical composition of 
iron ore is to optimise concentrator feed. Four gangue-related 
minerals are common to iron ores; kaolinite (Al
2
Si
2
O
5
[OH]
4
), 
quartz (SiO
2
), Gibbsite Al(OH)
3
 and aluminous goethite 
([Fe
1-2
Al
2
]O[OH]). Kaolinite, being typically hosted by shales 
within the ore, is the typical mineral liberated during typical 
wet concentration. Quartz and aluminous goethite are much 
less susceptible. It is therefore important to determine the 
1. Application Specialist XRD, PANalytical BV, Lelyweg 1, 7600 AA Almelo, The Netherlands. Email: uwe.konig@panalytical.com
2. Application Specialist XRD, PANalytical BV, Rua Jose de Carvalho, 55 04714-020 Sao Paulo SP, Brazil. Email: luciano.gobbo@panalytical.com
3. PANalytical BV, 1117 Flanders Road, Weatborough MA 01581, USA. Email: Kathy.macchiarola@panalytical.com
Using X-Ray Diff raction for Grade 
Control and Minimising Environmental 
Impact in Iron and Steel Industries
U König1, L Gobbo2 and K Macchiarola3
ABSTRACT
The use of high speed detectors made X-ray diffraction (XRD) an important tool for quality and 
process control in the mining industries. Besides knowledge about the mineralogy of a sample it 
provides useful information in terms of quantifi cation of the crystalline phases and the amorphous 
content. Cluster analysis of XRD data can facilitate multi-dimensional mapping of ore deposits 
and drill cores, identifying regions of favourable mineral compositions. The studies of this paper 
demonstrate how XRD and data clustering can be used for grade control, process optimisation 
and quality control of iron ores and iron ore sinters. These new developments have an enormous 
potential as an inexpensive, reliable tool, useful in the characterisation of iron and sinter materials 
implemented in an industrial environment.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
U KÖNIG, L GOBBO AND K MACCHIAROLA
50
proportions of kaolinite, quartz, gibbsite and aluminous 
goethite in the iron ore so that certain phases do not separate 
in the concentrator.
METHODS
X-ray powder diffraction (XRD) is a versatile, non-destructive 
analytical method for identifi cation and quantitative 
determination of crystalline phases present in powdered 
and bulk samples. The identifi cation of phases is achieved by 
comparing measured diffraction data to a reference database, 
the most comprehensive database is maintained by the 
International Centre for Diffraction Data (ICDD). This is a 
standard technique for qualitative analysis of mineralogical 
phases, but quantitative methods were often diffi cult since 
they require pure phase standards. Modern quantifi cation 
analysis techniques such as Rietveld analysis are attractive 
alternatives, as they do not require any standards or monitors 
(Rietveld, 1969). These methods offer impressive accuracy 
and speed of analysis. Modern XRD equipment is also capable 
of producing data of suffi cient quality for Rietveld analysis 
in just minutes, instead of an hour or more with traditional 
detectors, making it more amenable to process control, 
Macchiarola et al (2007).
A typical Bragg-Brentano confi guration as it is used for 
the analysis powder materials is show in Figure 1. The set-
up consist of an X-ray source, a spinning sample stage for 
optimising counting statistics, a high speed linear detector 
and several optics for optimising the quality of the data output.
To handle the large amount of data coming from the rapid 
data collection using the X’Celerator detector, ‘cluster analysis’ 
is a useful tool to group different grades of ore or materials 
into groups (clusters). This statistical method simplifi es the 
analysis of the data by:
  automatically sorting all scans from one or more ore 
samplings experiments into classes of closely related scans 
(clusters),
  identifying the most representative scan of each class,
  identifying the two scans of each class that differs most, 
and
  identifying outliers not fi tting into any class (non-
members).
The method can be used for sorting different grades of 
materials, such as ores with different mineralogy and thus 
predicted process behaviour.
In the current study, in order to determine the crystalline 
phase content of the iron ore feed and iron ore sinter samples 
analysed, pressed powder pellets were prepared. Sample 
preparation is an important issue to obtain correct results 
as it is a primary source of errors in the quantifi cation (Klug 
and Alexander, 1954). For that reason all samples were 
prepared using automatic sample preparation equipment for 
minimised preferred orientation and to maintain a constant 
preparation for all samples. All powder samples were milled 
for 30 seconds and pressed 30 seconds with ten ton into steel 
ring sample holders.
For the studies presented in this paper, a CubiX3 Minerals 
industrial diffractometer with Co anode and high-speed 
X’Celerator detector was used, featuring measurement times 
of less than ten minutes per scan generally.
Grade control of iron ores – analysis of iron 
ore feed
Over 300 minerals contain iron but there are only fi ve minerals 
considered commercial sources of iron: 
1. magnetite (Fe
3
O
4
),
2. haematite (Fe
2
O
3
), 
3. goethite (FeOOH), 
4. siderite (FeCO
3
), and
5. pyrite (FeS
2
). 
The best iron ore does not always have the highest iron 
content, but rather may have the least amount of impurities 
that are diffi cult to remove. Common impurities present in an 
iron ore are:
  silica (SiO
2
) (almost always present),
  phosphorous (P),
  alumina (Al
2
O
3
),
  sulfur (S),
  calcium carbonate (CaCO
3
), and
  manganese oxide (MnO
2
) (especially in haematite).
The quality of iron ore is defi ned by the total iron content 
as well as the above mentioned impurities, but also the 
quantitative content of iron containing phases and gangue 
minerals such as quartz (SiO
2
), kaolinite (Al
2
Si
2
O
5
(OH)
4
), 
gibbsite (Al(OH)
3
) or calcite (CaCO
3
). Quantifi cation of 
these minerals gives important information for downstream 
processing treatments (eg sintering). The analysis of Al 
containing phases, especially of Al containing goethite, are 
useful for the grade differentiation since the removal of Al 
from the ore is time, energy and cost intensive depending on 
which mineral the Al is bound.
Figure 2 shows a magnifi ed view of themeasurement of 
60 iron ore samples from one deposit in Western Australia. 
Besides different amounts of the main ore minerals haematite 
FIG 1 - X-ray diff raction setup for powder measurements, Bragg-Brentano geometry.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
USING X-RAY DIFFRACTION FOR GRADE CONTROL AND MINIMISING ENVIRONMENTAL IMPACT IN IRON AND STEEL INDUSTRIES
51
and goethite some samples also contain signifi cant amounts 
of kaolinite, quartz, gibbsite and calcite.
Before quantitative Rietveld analysis was applied on these 
samples, the data set was used for a cluster analysis to fi lter the 
different ore grades with varying mineralogy. The Principle 
Component Analysis (PCA) plot in Figure 3 indicates clearly 
three different clusters and some outliers. Together with the 
results of the Rietveld quantifi cation of all samples three 
different grades of iron ore material could be defi ned.
Table 1 shows the minimum, maximum and average modal 
abundances of the samples analysed. The three ore grades are:
1. high-grade – medium haematite, high goethite, low 
gibbsite, kaolinite, quartz;
2. low-grade – medium haematite, low goethite, high 
gibbsite, kaolinite, quartz; and
3. medium-grade – low haematite, high goethite, medium 
gibbsite, kaolinite, quartz. 
The results of the quantifi cation of all 60 samples are 
plotted against the cluster number to visualise the different 
grades (Figure 4). The defi nition of the different clusters with 
different ore materials and phase contents can be used for a 
quick automated grade control (pass/fail analysis). Once the 
cluster defi nitions are set-up, a new measurement can be 
sorted into one of the cluster and the corresponding material 
can be put on the corresponding stock pile in the mine.
An other problem for the processing of iron ores is the 
substitution of Fe with other cations in the crystal structure. 
Goethite is the best studied example of an isomorphously 
substituted iron oxide and of the various possible impurities 
that it can contain in natural samples. For example, although 
Al is 17 per cent smaller then Fe3+, up to one-third of the 
Fe3+ in goethite can be replaced by Al. The full range of 
substitution up to 33 mol per cent is found in natural goethites 
and up to 16 mol per cent in natural haematite. One reason 
for this is the abundance of Al in rocks and soils and its 
mobilisation together with Fe during weathering (Cornell and 
Schwertmann, 1996).
The Al-for Fe substitution was originally discovered with 
XRD in Jurassic, marine, oolitic iron ores (Correns and von 
Engelhart, 1941) and 20 years later found in soils by Norrish 
and Taylor (1961). Aluminium located in the structure of 
goethite and haematite in natural iron ores and bauxites 
cannot be extracted, both for Al metal production and iron ore 
benefi ciation.
A linear relationship exists between the Al substitution 
and the b and c parameters of the unit cell of goethite (Thiel, 
1963; Schulze, 1984) and subsequently on the peak position 
measured during a XRD experiment.
According to Schulze (1984), the c parameter can be used 
to determine the amount of Al substitution. The estimation 
of the Al substitution can be calculated from the relationship: 
mol per cent Al = 1730 - 572.0 c.
The unit cell parameters are refi ned during the quantitative 
Rietveld calculations and were used to estimate the amount 
of Al within the goethite of the 60 samples. The calculated 
concentration for the Al substitution in the analysed goethite 
varies from 1 - 6 mol per cent.
FIG 2 - Magnifi ed view on the main peaks of the ore minerals and impurities for 60 samples from one drill hole of an iron ore deposit.
FIG 3 - Principle Component Analysis plot of 60 iron ore samples from one 
deposit divided into three diff erent clusters (grades) with diff erent mineralogy 
(red = outlier), the three axes represent principle component 1, 2 and 3.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
U KÖNIG, L GOBBO AND K MACCHIAROLA
52
Figure 5 demonstrates the shift if the (111) refl ection of 
goethite with different amounts of Al substituted in the range 
from 0 - 35 mol per cent, whereas the (040) refl ection is not 
affected by the substitution of Al.
ANALYSIS OF IRON ORE SINTER
Iron ore sinter materials are an important feedstock material 
for the steel industry. Sintering is a high temperature 
agglomeration process. Since fi nes cannot be used in 
conventional blast furnaces because they impair the upward 
gas fl ow, fi nes are agglomerated in sinter plants (Ghosh 
et al , 1999). As a result of increased quality requirements, 
optimal energy use and minimum CO
2
 emissions, the phase 
composition and chemistry of iron ore sinters are important.
Especially the analysis of haematite and magnetite and 
subsequently the Fe2+/Fe3+ ratio affects the control of 
Cluster
Haematite 
(%)
Goethite 
(%)
Gibbsite (%)
Kaolinite 
(%)
Quartz (%)
Magnetite 
(%)
Mol % Al in 
Goethite
1 High-grade
Min 3.1 56.1 0.0 0.0 0.0 0.0 <5
Max 23.1 94.3 1.1 4.6 0.8 2.2 <5
Average 16.9 75.3 0.6 2.2 0.1 0.7 <5
2 Low-grade
Min 15.7 37.4 1.9 4.3 0.0 0.0 <5
Max 49.5 76.8 4.9 16.2 3.7 0.0 5 to 12
Average 30.3 55.2 3.0 9.9 1.6 0.0 5 to 12
3 Medium-grade
Min 5.9 50.7 0.5 0.0 0.0 0.0 <5
Max 39.1 91.2 4.1 8.8 4.8 0.0 <5
Average 19.5 74.2 1.9 3.5 0.9 0.0 <5
TABLE 1
Concentration ranges of the phase content of 60 iron ore samples obtained by Rietveld refi nement.
 
Goethite
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3
Cluster
%
Hematite
0
10
20
30
40
50
60
0 1 2 3
Cluster
%
Kaolinite
0
2
4
6
8
10
12
14
16
18
0 1 2 3
Cluster
%
Gibbsite
0
1
2
3
4
5
6
0 1 2 3
Cluster
%
FIG 4 - Phase concentration of the main phases versus the cluster number for 60 analysed iron ore samples.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
USING X-RAY DIFFRACTION FOR GRADE CONTROL AND MINIMISING ENVIRONMENTAL IMPACT IN IRON AND STEEL INDUSTRIES
53
the energy consumption and thereby the CO
2
 emissions. 
Magnetite is oxidised to haematite in an exothermic chemical 
reaction which produces heat and reduces the energy needed. 
Annually about 6.5 Mt of iron ore sinter are produced. The 
fuel consumption is approximately 60 kg coke or anthracite 
per tonne iron ore sinter. Small fuel saving (typically -1 kg/
tonne of sinter) already represents a signifi cant saving in 
energy and cost (Devitt, 1968).
Main minerals present in the sinter are haematite 
(Fe
2
3+O
3
), magnetite (Fe
2
3+Fe2+O
4
), wuestite (Fe2+O), larnite 
(Ca
2
SiO
4
), silico ferrites of calcium and aluminium (SFCA) 
and possibly a glass phase. The properties of an ideal iron 
ore sinter depend on these mineral phases. SFCA must be a 
major phase to ensure reducibility and strength. Magnetite 
is preferred to haematite because of volume changes during 
the reduction (Ghosh et al, 1999). The larnite content should 
be as low as possible since β-γ transformations can occur 
(Ishikawa et al, 1983). 
The process parameters for common sinter plants that need 
to be controlled are:
  the basicity CaO/SiO
2
 (usually 1.2 - 2.4),
  the return fi nes rate (usually 20 - 30 per cent), and
  the FeO content (usually 5 - 9 per cent). 
In a f rst step 55 samples from iron ore sinter plant over a 
production time of 100 days were measured using a CubiX3 
Minerals diffractometer. All scans are plotted together in 
Figure 6 to get an overview of the changes in the sinter 
produced over the time. Only minor differences are visible 
in the range of 72 - 74°2θ. In a second step, before further 
qualitative and quantitative investigations, a cluster analysis 
of all measured scans was applied.
Two different groups of sinter material can be identifi ed in the 
PCA plot shown in Figure 7. One cluster represents a relatively 
stable sinter product since the scans in the PCA plot are very 
close to each other. The other cluster shows a wider spread that 
indicates ahigher variation of the analysed sinter samples.
FIG 5 - Simulated shifts in peak position and intensity for Al substituted goethite.
 
0
5000
10000
Counts
Position [°2Theta] (Cobalt (Co))
65 70 75
FIG 6 - Zoomed view on 55 measurements of iron ore sinters over a production period of 100 days.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
U KÖNIG, L GOBBO AND K MACCHIAROLA
54
The most representative scans from each cluster were 
further analysed by X-ray fl uorescence spectroscopy (XRF) to 
compare the basicity CaO/SiO
2
 obtained and calculated from 
the XRF data, Figure 8. It was found that the two types of 
sinter correspond with two different levels of basicity, 1.7 and 
2.4. This refl ects a change in the composition of the sinter feed 
processed during the 100 days, later attributed to a change in 
supplier of one of the key raw materials. 
This example demonstrated that simply by statistical 
evaluation of the scan, the type of the sinter produced can be 
identifi ed. Changes in sinter composition that refl ect a different 
processing can be monitored and if necessary adjusted. Since 
this can be done within minutes after the sinter was produced, 
it has a major impact on the energy consumption and sinter 
quality.
The third step of the analysis contains the Rietveld quant-
ifi cation on all samples measured (Figure 9). All crystalline 
phases present such as haematite, magnetite, wuestite, larnite 
and the SFCA phases. SFCA-a refers to SFCA according to 
Hamilton et al (1989). SFCA-b is describes as SFCA-I by 
Mumme (Clout and Gable, 1998). 
From the phase concentrations the total FeO content can be 
calculated, as shown in Table 2. The amount of SFCA phase 
gives an indication for the hardness and subsequently the 
return of fi nes rate of the sinter material produced (Patrick 
and Pownceby, 2002).
The two groups of sinter material show differences in the 
amount of total metal iron. Also the Fe2+ (as FeO) content 
seem to be a bit higher in the sinter material of cluster 1. 
The amount of SFCA phase and wuestite is clearly increased 
in sinter material of cluster 2 whereas the haematite and 
magnetite content is lower compared to the samples of cluster 1.
CONCLUSIONS
In this article it is shown how X-ray diffraction can be used to 
provide valuable information for iron ore mining and iron ore 
sinter process control through standard-less quantifi cation 
and fast, statistical evaluation of large data sets through 
cluster analysis. Today’s optics, detectors, and software can 
provide rapid (within minutes) and accurate analyses, suitable 
for process control environments as well as research.
FIG 7 - Principle Component Analysis plot of 55 iron ore sinter samples 
representing two groups of sinter materials.
FIG 8 - Results of the data clustering versus the basicity CaO/SiO
2
.
Position [°2Theta] (Cobalt (Co))
35 40 45
Counts
0
5000
10000
Hematite 21.9 %
Magnetite 29.1 %
C2S - Belite 5.6 %
SFCA-a 25.1 %
SFCA-b 17.0 %
Wuestite-FeO 1.4 %
FIG 9 - Rietveld quantifi cation of the crystalline phase content of an iron ore sinter sample.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
USING X-RAY DIFFRACTION FOR GRADE CONTROL AND MINIMISING ENVIRONMENTAL IMPACT IN IRON AND STEEL INDUSTRIES
55
Cluster Haem % Mag % Lar % SFCA % Wue % Fe2+ Fe met
1 39.7 33.3 4.7 21.9 0.4 10.7 63.0
1 37.3 33.7 5.9 22.6 0.5 10.9 61.9
1 32.7 37.6 5.6 23.6 0.5 12.2 62.1
1 33.5 36.6 6.3 23.2 0.4 11.8 61.7
1 30.6 41.2 5.0 22.8 0.4 13.2 62.8
1 33.6 39.7 4.5 21.9 0.4 12.7 63.3
1 36.0 34.4 5.4 23.8 0.3 11.0 62.1
1 37.8 34.1 4.5 23.2 0.4 11.0 62.9
1 36.2 30.7 4.6 27.8 0.6 10.1 61.7
1 31.9 37.0 4.5 26.0 0.5 12.0 62.3
1 34.2 34.3 3.8 27.2 0.6 11.2 62.5
1 44.1 32.0 6.5 17.3 0.2 10.1 62.7
1 40.3 33.1 6.5 19.6 0.4 10.6 62.1
1 34.8 34.5 5.5 24.8 0.5 11.2 61.9
1 39.4 33.3 6.7 20.4 0.4 10.7 61.9
1 36.8 38.0 6.2 18.8 0.3 12.1 62.7
1 39.0 34.0 6.5 20.3 0.3 10.8 62.1
1 39.1 33.0 7.3 20.3 0.4 10.6 61.5
1 41.8 32.1 5.9 20.0 0.3 10.2 62.5
1 40.3 33.4 7.0 18.9 0.3 10.7 61.9
1 44.1 27.8 6.5 21.4 0.3 8.9 61.7
1 42.7 30.7 6.2 20.1 0.3 9.8 62.2
1 39.9 31.4 6.4 21.9 0.4 10.1 61.8
1 34.2 40.9 6.7 18.0 0.3 13.0 62.6
1 40.7 33.4 6.7 18.9 0.3 10.7 62.1
1 36.1 35.1 6.6 21.7 0.5 11.4 61.7
1 39.9 35.0 6.8 18.1 0.3 11.1 62.3
1 37.3 32.3 6.7 23.3 0.4 10.4 61.2
1 40.4 32.3 6.3 20.8 0.3 10.3 62.1
Min 30.6 27.8 3.8 17.3 0.2 8.9 61.2
Max 44.1 41.2 7.3 27.8 0.6 13.2 63.3
Average 37.7 34.3 5.9 21.7 0.4 11.0 62.2
2 19.9 35.8 6.8 36.3 1.2 12.3 58.5
2 21.9 29.1 5.6 42.1 1.4 10.4 58.1
2 19.3 31.7 8.1 39.4 1.5 11.3 56.9
2 20.8 31.2 7.8 39.0 1.3 10.9 57.2
2 24.2 26.1 6.2 42.2 1.3 9.4 57.5
2 21.3 33.4 6.4 37.7 1.2 11.6 58.5
2 22.2 29.4 5.8 41.3 1.3 10.4 58.1
2 23.3 28.2 6.4 40.9 1.3 10.0 57.7
2 21.2 29.9 5.9 41.7 1.3 10.5 57.9
2 18.2 30.8 7.2 42.4 1.4 10.9 56.9
2 20.4 31.6 6.2 40.6 1.3 11.0 58.0
2 18.9 34.5 5.4 39.9 1.4 12.0 58.8
TABLE 2
Results of the quantitative Rietveld analysis of 55 iron ore sinter samples (Haem = haematite, Mag = magnetite, Wue = wuestite, Lar = larnite) and 
calculated Fe2+ (as FeO) and total metal Fe content, data is separated into the two diff erent clusters identifi ed by the Principle Component Analysis.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
U KÖNIG, L GOBBO AND K MACCHIAROLA
56
REFERENCES
Cornell, R M and Schwertmann, U, 1996. The Iron Oxides , 573 p 
(Wiley).
Correns, C W and von Engelhart, W, 1941. Roentgenographische 
Untersuchungen ueber den Mineralbestand sedimentaerer 
Eisenerze, Nachr Akad Wiss Goettingen, Math Phys Kl, 2:131-137.
Devitt, T W, 1968. The integrated iron and steel industry air pollution 
problem, US department of health, education and welfare.
Ghosh, A and Chatterjee, A, 2008. Ironmaking and Steel Making: 
Theory and Practice, 492 p (Prentice Hall).
Hamilton, J D G, Hoskins, B F, Mumme, W G, Borbidge, W E 
and Montague, M A, 1989. The crystal structure and crystal 
chemistry of Ca2.3 Mg0.8 Al1.5 Si1.1 Fe8.3 O20 /SFCA): Solid 
solution limits and selected phase relationships of SFCA in the 
Si O2-Fe2 O3-CaO-(Al2 O3) system, Neues Jahrbuch Miner Abh, 
161(1):1-26.
Ishikawa, Y, Shimomura, Y, Sasaki, M and Toda, H, 1983. 
Improvement of sinter quality based on the mineralogical 
properties of ores, in Proceedings 42nd Ironmaking Conference, 
42:17-92.
Klug, H P and Alexander, L E, 1954. X-Ray Diffraction Procedures 
for Polycrystalline and Amorphous Materials , 992 p (Wiley-
Blackwell).
Lu, L, Holmes, J and Manuel, J R, 2007. Effects of alumina on 
sintering performance of hematite iron ores, ISIJ International, 
47(3):349-358.
Macchiarola, K, Koenig, U, Gobbo, L, Campbell, I, McDonald, 
A M and Cirelli, J, 2007. Modern X-ray diffraction techniques 
for exploration and analysis of ore bodies, in Proceedings Fifth 
Decennial International Conference on Mineral Exploration , 
Toronto.
Mumme, W G, Clout, J M F and Gable, R W, 1998. The crystal 
structure of SFCA-I, Ca3.18 Fe3+14.66Al1.34Fe2+0.82O28, 
a homologue of the aenigmatite structure type and new crystal 
structure refi nements of beta-CFF, Neues Jahrbuch Miner Abh , 
173(1):93-117.
Norrish, K and Taylor, R M, 1961. The isomorphous replacement 
of iron by aluminium in soil goethites, Journal of Soil Science , 
12:294-306.
Patrick, T R C and Pownceby, M I, 2002. Stability of SFCA in air solid 
solution limits between 1240 and 1390C and phase relationships 
within FCAS system, Metallurgical and Materials Transactions 
B, 33B:79-89.
Rietveld, H M, 1969. A profi le refi nement method for nuclear and 
magnetic structures, Journal of Applied Crystallography, 2:65-71.
Schulze, G D, 1984. The infl uence of aluminium on iron oxides VIII: 
Unit cell dimensions of Al-substituted goethite and the estimation 
of Al from them, Clays and Clay Minerals, 32(1):36-44.
Thiel, R, 1963. Zum System FeOOH-AlOOH, Z Anorg, Allg Chem , 
326:70-78.
Yamaoka, Y, 1974. Effects of gibbsite on sinter properties and 
quality, Transactions of the Iron and Steel Instituteof Japan , 
14:185-193.
Cluster Hem % Mag % Lar % SFCA % Wue % Fe2+ Fe met
2 19.4 33.2 6.2 40.0 1.3 11.5 58.1
2 20.2 30.1 6.3 42.0 1.4 10.7 57.6
2 22.7 27.7 6.7 41.6 1.3 9.9 57.3
2 24.2 25.4 6.4 42.7 1.4 9.3 57.3
2 25.2 25.2 6.4 42.0 1.2 9.0 57.4
2 23.1 28.1 6.9 40.8 1.2 9.9 57.4
2 22.1 26.2 6.7 43.6 1.3 9.4 56.8
2 20.8 31.5 7.5 38.9 1.4 11.1 57.4
2 23.2 29.4 6.4 39.8 1.2 10.3 57.9
2 21.7 29.2 7.1 40.7 1.3 10.4 57.3
2 21.0 30.2 6.6 40.9 1.3 10.7 57.6
2 18.4 32.2 7.4 40.7 1.4 11.3 57.1
2 18.5 29.6 7.3 43.2 1.5 10.6 56.6
2 18.2 32.8 7.5 40.2 1.3 11.5 57.2
Min 18.2 25.2 5.4 36.3 1.2 9.0 56.6
Max 25.2 35.8 8.1 43.6 1.5 12.3 58.8
Average 21.2 30.1 6.7 40.8 1.3 10.6 57.6
TABLE 2 CONT...
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 59
INTRODUCTION
Famous for its gold and high grade iron ores, the mining 
district of the southeastern Brazilian highlands known as 
Quadrilátero Ferrífero (QF) or Iron Quadrangle (Dorr, 1969) 
extends over an area of around 7500 km2, highlighted by 
ridges of Paleoproterozoic banded iron formation (Figure 1). 
Large scale iron ore production started in the QF after 
the World War II in the mid-1950s and has experienced a 
continuous growth ever since. The QF mining industry is 
today responsible for ca 70 per cent of the Brazilian and ca 
ten per cent of world iron ore production. 
The iron ores so far exploited in the QF were in general of 
higher grade (>55 per cent Fe) and, except for one category, the 
so-called hard haematite, derived from supergene enrichment 
of itabirites of the Paleoproterozoic Cauê Formation (Eichler, 
1968; Dorr, 1964, 1965, 1969; Melfi et al, 1976; Ramanaidou 
et al, 1996; Spier, Oliveira and Rosière, 2003, 2007; Rosière 
et al, 2008; Ramanaidou and Morris, 2010). Fresh, unenriched 
itabirite, viewed as protore, was never incorporated in the 
benefi ciation plants. However, the high current demand and 
consequent depletion of soft ores will cause a fast increment of 
1. Superintendent of Geology, Mine Planning and Environment, Mineração Usiminas Ltda, Rua Professor José Vieira de Mendonça, 3011, bairro Engenho Nogueira, Belo Horizonte, Minas Gerais, 
 31310-260, Brazil. Email: leandro.amorim@usiminas.com
2. Professor, Escola de Minas, Universidade Federal de Ouro Preto, Departamento de Geologia, Morro do Cruzeiro, Ouro Preto Minas Gerais 35.400-000, Brazil. Email: ff alkmim@gmail.com
New Ore Types from the Cauê Banded 
Iron Formation, Quadrilátero Ferrífero, 
Minas Gerais, Brazil – Responses to the 
Growing Demand
L Q Amorim1 and F F Alkmim2
ABSTRACT
The mining district of the Quadrilátero Ferrífero (QF) in southeastern Brazil contributes to 
70 per cent of the Brazilian iron ore production, which reached 370 Mt in 2010. Traditionally, four 
iron ore types were mined in the QF region: 
1. soft haematite, 
2. hard haematite, 
3. friable ‘itabirite’ (the Brazilian name for metamorphosed banded iron formation), and 
4. canga (an indurated iron-rich crust). 
With the exception of the hard haematite, these categories derive from supergene enrichment of 
the Paleoproterozoic Cauê Itabirite. High current iron ore demand means that the QF high grade 
supergene ores will be depleted in a few decades. Consequently, several mining projects are being 
developed to exploit unenriched, fresh itabirite. Among them, the Mineração Usiminas Serra Azul 
project in northwestern QF is one of the most advanced. Four major itabirite types occur in the 
QF region: 
1. siliceous (or standard), 
2. dolomitic, 
3. amphibolitic, and 
4. magnetitic. 
The dominant type in Serra Azul region is a fresh siliceous itabirite, composed essentially of 
haematite, martite and quartz that averages 36 per cent Fe (51.5 per cent Fe
2
O
3
) and 47 per cent 
SiO
2
. The second type, the magnetitic itabirite, consisting of magnetite, martite, grunerite-
cummingtonite, quartz, dolomite, stilpnomelane, and subordinate haematite, is typically a 
strongly magnetic greenish banded iron formation with the iron grade varying between 25 and 
35 per cent. Chemically, it differs from the siliceous itabirite due to the relatively high CaO, MgO 
and, consequently, high LOI contents. Together with supergene ores, fresh itabirites have been 
included in the ore reserves of the Serra Azul project. We anticipate that all the four types of fresh 
itabirite will be incorporated as ore in the QF region, causing a huge increase on the iron ore 
resources of the district, and representing enormous environmental and technological challenges 
for the local iron ore industry. 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
L Q AMORIM AND F F ALKMIM
60
iron ore production from fresh itabirite, which represents the 
new trend in the QF mining industry. 
Since fresh BIF was so far considered to be waste in the 
QF region, the unweathered section of the Cauê Formation 
was not drilled and investigated on a routine basis in the 
exploration programs carried out in the QF. As a fi rst in the 
region, Mineração Usiminas performed a comprehensive 
exploration program for hard itabirites in the locality named 
‘Serra Azul ‘, western QF (Figure 1), which included 67 km of 
drilling distributed in 542 boreholes. The new data obtained 
in this program forms the base for the descriptions and 
discussions presented in this paper. 
THE GROWING DEMAND ON IRON ORES
The trend of the world’s iron ore demand through time can be 
easiest visualised using crude steel production curves. Figure 2 
shows the total crude steel production of the world in the last 
100 years, plotted together with the sum of blast furnace iron 
(BFI) and directly reduced iron (DRI) productions. This sum 
represents the amount of steel produced from iron ore and the 
difference between the two curves corresponds to the amount 
of steel derived from scrap (Figure 2). 
At fi rst glance, the curves of Figure 2 record only the historical 
steel production growth based on iron ore consumption. This 
general tendency includes two major production increments, 
separated by a period relatively stability between the 
mid-1970s and the mid-1990s of the last century. A more 
careful examination of the plots reveals, however, that 
major events affecting the world in the last 100 years are 
reproduced by the curves. Critical periods – the 1929 crash, 
the World War II in the mid-1940s, the fi rst and second oil 
crisis, respectively in the mid-1970s and 1980s, the end of the 
USRR in the beginning of the 1990s, as well as the very recent 
2008 economic crisis – are represented by lows followed by 
recoveries in the years after.
In the aftermath of the World War II, the steel production 
experienced an enormous growth, rising from 200 Mt in 1950 
to 600 Mt in 1970. This increment was a consequence of several 
developments, but mainly of the further industrialisation of 
the fi rst world countries, the cold war, and reconstruction 
of Europe and Japan. In the following three decades, 
from 1970 - 2000, the steel production underwent only a 
vegetative growth, going from 600 - 800 Mt/a. Without any 
other historic equivalent, the second largest steel production 
increment, starting in the fi rst decade or the 21st century and 
reaching the present day, represents the economic booming 
FIG 1 - Digital topography model of the QF region showing the locations of the Curral Ridge and Serra Azul.
0
200
400
600
800
1000
1200
1400
1600
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
m
ill
io
n
to
nn
es
BFI+DRI crude steel
FIG 2 - World crude steel production in the last 100 years (data source: 
International Iron and Steel Institute – IISI; http://www.worldsteel.org).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
NEW ORE TYPES FROM THE CAUÊ BANDED IRON FORMATION, QUADRILÁTERO FERRÍFERO, MINAS GERAIS, BRAZIL
61
of China. The world steel production reached 1.4 Bt in 2010, 
a level that corresponds to the extraordinary increase of 
600 Mt/a in ten years. From the 1.4 Bt of crudesteel produced 
in 2010, around 1.1 Bt correspond to BFI plus DRI, which 
caused a consumption of 1.7 Bt of iron ore in one single year.
Brazilian iron ore production continuously increased after 
the mid-1950s, reaching 52 Mt in 1973 and the peak of 94 Mt in 
1976 (Figure 3). Several iron ore projects were implemented in 
Brazil in the late 1960s and early 1970s, leading to a production 
of almost to 100 Mt in the mid-1980s (Figure 3). All of these 
projects were developed in the Quadrilátero Ferrífero, the 
only signifi cant source of iron ore in Brazil until 1986, when 
Carajás, the second largest mining district of the country, start 
also to produce iron ores. During the 25 years period of the 
mid-1970s to the year 2000, the Brazilian iron ore production 
rose from near 100 Mt/a to approximately 200 Mt/a 
(150 Mt/a from the QF and 50 Mt/a from Carajás). From 2001 
to 2008, the annual production of the QF grew from 150 Mt to 
250 Mt. The production of other regions in the country (mostly 
from Carajás, northern Brazil) increased from 50 - 100 Mt/a, 
leading a national production of 351 Mt in 2008 (Figure 3) 
as reported by the Brazilian National Department of Mineral 
Production (DNPM). According to United States Geological 
Survey statistics, the 2010 Brazilian iron ore production was 
370 Mt.
The fi rst iron ore concentration plant of Brazil was set in 
the town of Itabira, Minas Gerais (QF), in 1973. Up to that 
time, only high grade iron ores were exploited in the QF. 
Immediately after, the Brazilian iron ore industry underwent 
signifi cant changes with the introduction of several other 
concentration plants, and the consequent incorporation of 
new iron types such as soft enriched BIF. Due to the enormous 
rise of the production in the last years, even the soft and rich 
BIF resources will be depleted in the QF in a few decades. 
Thus, as a response to present level of demand, new ore types, 
this time represented by former protores, ie fresh and hard 
itabirites, are to be introduced in the production system of the 
QF. The incorporation of these new ore types will result in an 
extraordinary increase of the QF iron ore resources, but also 
in new benefi ciation process and considerable smaller mass 
recovery, implying in larger tailings dams, larger pits and 
consequently greater environmental impacts.
THE QUADRILÁTERO FERRÍFERO AND ITS 
IRON ORES
Geological setting of the Quadrilátero Ferrífero
Geologically, the Quadrilátero Ferrífero is located in the 
southeast border of the São Francisco Craton, an Archean/
Paleoproterozoic stable block, surrounded by Neoproterozoic 
orogenic belts (Almeida, 1977; Alkmim and Marshak, 1998). 
The district is underlain by fi ve major lithostratigraphic units: 
1. the Archean basement, 
2. the Archean Rio das Velhas Supergroup, 
3. the Paleoproterozoic Minas Supergroup, 
4. the Itacolomi Group, and 
5. mafi c intrusive (Figure 4).
0
50
100
150
200
250
300
350
400
1950 1960 1970 1980 1990 2000 2010
m
ill
io
n
to
nn
es
Brazil Minas Gerais
FIG 3 - Brazilian iron ore production 1950 - 2008 (data source: National 
Department of Mineral Production – DNPM; http://www.dnpm.gov.br/). 
FIG 4 - Simplifi ed geologic map of the QF, emphasising the distribution of the main lithostratigraphic units and structures (Based on Dorr, 1969; Romano, 1989).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
L Q AMORIM AND F F ALKMIM
62
The Archean basement is made up of 3.2 - 2.9 Ga old gneiss 
complexes and two generations of Late Archean granitoids 
(Machado et al, 1992; Carneiro, Teixeira and Machado, 1994; 
Noce, 1995). The 2.78 - 2.61 Ga old Rio das Velhas Supergroup, 
a classical greenstone belt succession, comprises a volcano-
sedimentary package, mainly composed of komatiite, basalt 
and ryolitic lavas, intercalated and overlain by sedimentary 
units, including carbonate facies banded iron formations 
(Dorr, 1969; Machado et al, 1992; Zucchetti, Baltazar and 
Raposo, 1998; Noce et al, 2005). 
The Minas Supergroup, accumulated in a passive to 
convergent margin setting between 2.6 - 2.0 Ga, records the 
operation of a full Wilson cycle on the southern edge of the 
São Francisco craton (Alkmim and Marshak, 1998). Separated 
from the Rio das Velhas greenstone belt by a pronounced 
unconformity, the basal Caraça Group consists of alluvial 
conglomerates and sandstones that grade upwards into 
shallow marine pelites. Caraça sediments are overlain by the 
transgressive Itabira Group that contains the Cauê Banded 
Iron Formation, the marker bed of the QF that hosts all of its 
iron ore resources. The Cauê Formation grades upward into 
the shallow water carbonates of the Gandarela Formation, 
which is in turn overlain by a thick transgressive succession of 
sandstones and pelites of the Piracicaba Group. The youngest 
unit of the Minas Supergroup, the Sabará Group, made up of 
turbiditic conglomerates, sandstones and pelites, records the 
conversion of the Minas passive margin into a syn-orogenic 
(fl ysch) basin at ca 2.1 Ga (Dorr, 1969; Machado et al, 1996; 
Renger et al, 1995; Hartmann et al, 2006). 
Probably representing an intermontane molassa, the 
quartzites and conglomerates of the Itacolomi Group 
unconformably overlie Minas Supergroup strata in the 
southern QF (Dorr, 1969) (Figure 4). Intrusions cutting the 
previous units include mainly mafi c dykes, which dominantly 
trend N-S to NW-SE. One of these dykes was dated at 1.714 Ma 
(Silva et al, 1995). 
The regional geologic map pattern of the QF defi nes a 
dome-and-keel architecture, in which the Archean basement 
occurs in form of domes surrounded by ‘keels ‘ containing the 
metasedimentary units. Keels include fi rst-order synclines, 
such as the Moeda and Dom Bosco synclines, as well as the 
large Serra do Curral homocline (Marshak et al, 1992; Alkmim 
and Marshak, 1998) (Figure 4). Rocks of the supracrustal 
sequence adjacent to the domes contain a distinct high-T/
low-P metamorphic aureole (Herz, 1978; Jordt-Evangelista, 
Alkmim and Marshak, 1992; Marshak et al, 1992).
The rocks of the Rio das Velhas Supergroup have been 
strongly deformed prior to the deposition of the Minas 
Supergroup (Dorr, 1969). Furthermore, two sets of 
pronounced post-Minas structures that apparently do not 
show any relationship with the dominant dome-and-keel 
structure can be recognised in the region. One set trends 
NE-SW and comprises the Gandarela syncline, the Serra do 
Curral homocline, as well as the synclinoria of the Itabira and 
Monlevade regions (Figure 4). The second set includes a series 
of west-verging thrusts, which are particularly well developed 
in the eastern QF (Dorr, 1969; Chemale Jr, Rosière and Endo, 
1994; Chauvet et al, 1994; Alkmim and Marshak, 1998).
Second-order mesoscopic folds associated to penetrative 
tectonic fabrics (ie schistosity, mylonitic foliation, stretching 
lineation) also occur throughout the region. The stretching 
lineation and the mesoscopic fold hinges plunge preferentially 
towards S80°E with 30° (Dorr, 1969; Chemale Jr, Rosière and 
Endo, 1994; Chauvet et al, 1994; Alkmim and Marshak, 1998). 
A regional, syn-kinematic metamorphism affecting all units 
exposed in the QF region varies from lower greenschist to the 
west to upper amphibolite facies to the east (Dorr, 1969; Herz, 
1978; Pires, 1995).
The Cauê banded iron formation and 
associated ores
The Cauê Formation is composed of a variety of 
metamorphosed BIF currently referred to as ‘itabirite’, iron 
ores, and subordinated dolomites (Dorr, 1969; Pires, 1979; 
Pires, Aranha and Cabral, 2005; Rosière and Chemale Jr, 
2006, Rosière et al, 2008; Spier, Oliveira and Rosière, 2003, 
Spier et al, 2007). The name itabirite (Eschwege, 1833) is used 
to designate: 
... a laminated, metamorphosed oxide-facies iron 
formation in which the original chert or jasper bands 
have been recrystallised into granular quartz and the 
iron is present as haematite, magnetite or martite ... 
(Dorr, 1964).
Accordingto their mineralogical composition, the QF 
itabirites, as fresh rock, are classifi ed into four main types: 
1. siliceous (or standard), 
2. dolomitic,
3. amphibolitic, and 
4. magnetitic itabirite. 
The fi rst three types were described by several authors (Dorr, 
1969; Pires, Aranha and Cabral, 2005; Rosière and Chemale Jr, 
2006; Rosière et al, 2008, Spier et al, 2007). The last one, as 
a fresh rock, is described in the present paper. Their chemical 
and mineralogical compositions are presented on Table 1. The 
standard itabirite is the dominant type in the QF. The other 
types occur as beds or lenses within the Cauê Formation. 
Dolomitic itabirites are more frequent in the western part of 
the QF, whilst the amphibolitic itabirites are more common 
in the eastern part of the district. Magnetitic itabirites 
normally occur as a deeply weathered, ochreous, goethitic 
iron formation (‘yellow magnetitic iron formation’ according 
to Pires (1979). Fresh magnetitic itabirite, as recently found in 
the western portion of the QF, is a green banded iron formation, 
containing magnetite, martite, grunerite-cummingtonite, 
quartz, dolomite, stilpnomelane, and subordinate haematite 
(Alkmim, 2009). According to Pires, Aranha and Cabral (2005), 
chemical composition and observed macroscopic structures 
point towards a volcanic origin for the magnetitic itabirite. The 
presence of stilpnomelane in this type of itabirite is indeed an 
indicator for contamination with volcanic material (eg Pickard, 
2002, 2003). 
The maximum original thickness of the Cauê Formation is 
estimated to be 350 m (Dorr, 1964). However, as emphasised 
by this author, due to the plastic behaviour of Cauê BIF in 
the deformation processes affecting the QF region, the fi nal 
thickness of the unit exceeds 1500 m in many places. 
The metamorphic conditions experienced by the Cauê 
Formation can be expressed by temperatures obtained in 
oxygen isotope studies. The metamorphic temperatures 
show a general increase from west to east, varying from a 
minimum of 394°C in the Moeda syncline to a maximum 
of 780°C in the Itabira synclinorium (Müller, Schuster and 
Hoefs, 1982) (Figure 4). These values are in agreement with 
the metamorphic facies zoning of the QF, as described by Herz 
(1978) and Pires (1995). 
No direct age determinations are available for the Cauê 
Formation. Babinski, Chemale and Van Schmus (1995) 
obtained a Pb-Pb age of 2.42 Ga for the basal limestones of 
the overlaying Gandarela Formation. Considering that the 
maximum depositional age of the basal quartzites of the 
Minas Supergroup is 2.58 Ga (Hartmann et al, 2006), the 
accumulation of the Cauê BIF must be occurred between 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
NEW ORE TYPES FROM THE CAUÊ BANDED IRON FORMATION, QUADRILÁTERO FERRÍFERO, MINAS GERAIS, BRAZIL
63
2.58 and 2.42 Ga, a time interval that encompasses most of 
the Paleoproterozoic BIF deposition ages (Gole and Klein, 
1981; Trendall, 2002). 
The iron ores so far exploited in the QF belong to two 
genetic classes: hypogene and supergene. The fi rst category, 
also referred to as hard or compact haematite, comprise very 
rich (Fe >66 per cent), massive, banded or foliated haematitic 
orebodies. Exhibiting linear or less frequent tabular shapes, 
the haematite bodies can vary from less than fi ve to more than 
50 Mt. Their origin is matter of a long standing debate among 
various authors (for discussion see Guild, 1953; Dorr 1965, 
1969; Varajão et al, 1997; Rosière and Chemale Jr, 2006; 
Hagemann et al, 2006; Rosière et al, 2008; Rosière and Rios, 
2004; Clout and Simonson, 2005). After the benefi ciation, 
the hard ores produce a high proportion of lumps (more than 
50 per cent), making it a quite valuable material.
The QF ores generated by supergene processes include 
soft haematite, enriched itabirites, and laterite crusts. 
Morphotectonic processes acting upon the QF region probably 
at the beginning of the Paleocene (Spier, Vasconcelos and 
Oliveira, 2006) created conditions for vertical and lateral 
water circulation within the Cauê Formation layers, causing 
the leaching of silica and dolomite from the itabirites and 
thus forming supergene ores (Dorr, 1964, 1965; Eichler, 1968; 
Melfi et al, 1976; Ramanaidou et al, 1996; Spier et al, 2003). 
This process generated bodies of almost pure soft haematite 
(Fe ≥ 64 per cent) and an enormous volume friable itabirites. 
Two subtypes of laterite crusts, known as ‘canga’ in Brazil, 
occur in the QF region: 
1. the in situ; and 
2. the detrital canga (Dorr, 1969). 
The in situ or structural canga occurs in form of up to 
30 m hard caps on top of the Cauê itabirites and associated 
orebodies (Dorr, 1965, Ramanaidou et al, 1996, Ramanaidou 
and Morris, 2010), especially in high elevations (1200 m to 
1600 m above the sea level) (Varajão, 1994). Covering the 
Cauê Formation and adjacent units along hill slopes and 
small basins, the detrital canga corresponds to breccias or less 
frequent to conglomerates containing hard haematite and BIF 
clasts. Usually associated with high alumina (four per cent 
Al2O3 or more) and phosphorous (0.12 per cent P or more), 
both types of canga are iron rich (usually Fe >50 per cent) 
and contain, besides haematite, goethite, and limonite (Dorr, 
1969, Melfi et al, 1976; Varajão, 1994, Ramanaidou, 2009).
Due to the depletion of most of the high grade deposits, 
many projects considering the exploitation of the unenriched 
hard itabirites (former protores) are already launched in 
the QF, among them the Serra Azul project discussed in the 
next section. The new category of fresh rock ores (Figure 5) 
includes all types of previously mentioned itabirites that 
average around 36 per cent iron (approximately 50 per cent 
Fe2O3). In order to produce pellet feed fi nes, their main use, 
the fresh rock ores must be grinded before concentration. 
THE SERRA AZUL PROJECT
Geologic outline of the Serra Azul
The Mineração Usiminas mineral claims are located along 
the western segment of Curral ridge, locally referred to as the 
Serra Azul (literally Blue Ridge) (Figures 3 and 6). The Serra 
Azul, extending for ca 30 km in the NE-SW, is underlain by a 
Siliceous itabirite Dolomitic itabirite Amphibolitic itabirite Magnetitic itabirite
Major 
components
Accessories
Major 
components
Accessories
Major 
components
Accessories
Major 
components
Accessories
M
in
er
al
og
y Dark bands
haematite, 
martite
mt, se, qz, py, 
mnox
haematite, 
martite
mt, qz, dl, 
mnox 
hornblend, 
grunerite
hm, ma, mt, 
qz, dl, af
magnetite, 
haematite, 
martite, 
grunerite
hm, ma, qz, af, 
ca, cl, bi, sp, tc 
Light bands quartz
hm, ma, cl, se, 
dl, py, mnox
dolomite
hm, ma, qz, py, 
tc, mnox
tremolite, 
actinolite
hm, ma, mt, 
qz, dl
quartz, 
carbonate
mt, hm, ma, 
qz, af, ca, cl, bi, 
sp, tc
Ty
pi
ca
l c
h
em
ic
al
 
co
m
po
si
ti
on
Fe 30 - 40% 35% 35% 25 - 35%
SiO
2
40 - 60% <1% 45% 35 - 55%
CaO <0.1% 15% <1% 2 - 10%
MgO <0.1% 10% <1% 2 - 10%
LOI 1 - 2% >5% 1 - 2% >5%
af = amphibole, cl = chlorite, dl = dolomite, hm = haematite, mt = magnetite, ma = martite, Mnox = manganese oxide, py = pyrophyllite, qz = quartz, se = sericite, tc = talc, ca = carbonate, 
sp = stilpnomelane, bi = biotite.
TABLE 1
Chemical and mineralogical characteristics of the fresh itabirites.
FIG 5 - Genetic classifi cation of the Serra Azul iron ores, including the new 
category of fresh rock ores (unrenriched BIF) (modifi ed from Dorr, 1964).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
L Q AMORIM AND F F ALKMIM
64
continuous outcrop belt of the Cauê Formation, which hosts 
several iron ore deposits (Simmons, 1968; Romano, 1989; 
Romano et al, 1991; Alkmim, 2009), including the Western, 
Central and Eastern mines, as well as the Camargos and Pau 
de Vinho deposits of the Mineração Usiminas (Figure 6). As 
part of the Curral ridge, the Serra Azul is the morphologic 
expression of an overturned,SE-dipping homocline that 
involves Rio das Velhas and Minas strata (Dorr, 1969). Along 
the Serra Azul, the basal units of the Minas supergroup are 
very thin and Gandarela Dolomite is absent for over almost 
its whole length (Simmons, 1968; Alkmim, 2009) (Figure 6). 
The Serra Azul overturned homocline can be subdivided into 
three structural domains (Alkmim 2009) (Figure 6). Along 
the structural domain I, which encompasses the Western 
Mine and the major part of the Central Mine, the thickness 
of the overturned Cauê Iron Formation increases from west 
to east, varying from around 50 m, at the western end of the 
ridge, to 250 m at the Central Mine. The dips also increase 
from west to east, being 20o at Ponta da Serra and reaching 
45o at Central Mine (Figures 6, 7 and 8). Standard itabirites 
and their weathering products are the more abundant ore 
types in this domain. The amount of fresh magnetitic itabirite 
is negligible. However, its weathering product, a magnetite-
rich, ochreous itabirite occurs as a continuous layer on top 
(stratigraphic base) of the iron formation, especially in the 
Western Mine. A small lens of hard haematite was identifi ed 
near to the base (stratigraphic top) of the Cauê Formation in 
the Western Mine (Figure 7).
The ca 15 km long domain II (Figure 6) encompasses the 
central portion of Serra Azul, including the Camargos deposit. 
Minas Supergroup strata strike ENE-WSW and dip 40o to 80o 
to SSE. NW-verging folds and faults give rise to local structural 
complexities. Magnetitic itabirites occur in form of layers 
or lenses intercalated with siliceous itabirite, which in turn 
grades laterally into dolomites. A clay-rich, deeply weathered 
itabirite (called by the miners AIF, for ‘argillaceous iron 
formation’) is quite frequent close to the contact with the 
Cercadinho Formation, the basal unit of the Piracicaba Group 
(Figure 9).
Along the structural domain III, which hosts the Pau de 
Vinho deposit (Figure 6), layers of the Minas Supergroup 
describe a large curve with the concave side facing north. 
On the easternmost sector of the curve, the strata dip NW, a 
change probably caused by the uplift of the adjacent the Bação 
Dome (Figure 6). NW-trending sinistral to reverse-sinistral 
shear zones cut the homocline along its whole length. One of 
the largest sinistral faults runs along the border of a large dyke 
(~300 m wide) of gabbro that divided the domain (as well as 
the Pau de Vinho deposit) in two sectors (Figure 6). To the 
southwest of the dyke, layers of NW-striking fresh itabirites 
predominate along the ca 500 m-thick Cauê Formation. 
To southeast of the dyke, the Cauê Formation consists of a 
sequence of magnetitic itabirite, dolomite, friable ore and 
argillaceous iron formation (AIF) (Figures 6 and 10).
Serra Azul iron ores
During the Serra Azul exploration program, two basic criteria 
were used to classify the iron ores: 
1. the iron grade; and 
2. the ‘W1’ parameter, which represents the mass percentage 
above 6.35 mm after crushing the sample in 31.5 mm. 
According to the second variable, the itabirite (or haematite) 
is classifi ed as: 
FIG 6 - Digital terrain model and geologic map of the Serra Azul region (based on Simmons, 1968; Romano, 1989; Alkmim, 2009).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
NEW ORE TYPES FROM THE CAUÊ BANDED IRON FORMATION, QUADRILÁTERO FERRÍFERO, MINAS GERAIS, BRAZIL
65
FIG 7 - Cross-section 5100 – Western Mine (for location see Figure 6).
FIG 8 - Cross-section 8500 – Central Mine (for location see Figure 6).
FIG 9 - Cross-section 568 000 – Camargos Deposit (for location see Figure 6).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
L Q AMORIM AND F F ALKMIM
66
  ‘friable’, for W1 minor or equal to 30 per cent; 
  ‘semi-compact’ for W1 between 30 per cent and 55 per cent; 
  ‘compact’ for W1 between 55 and 70 per cent; and 
  ‘hard compact’ for W1 greater that 70 per cent. 
According to these criteria, three major iron ore types have 
been discriminated in Serra Azul mineral claims:
1. haematite and standard itabirite, 
2. magnetitic itabirite, and
3. argillaceous iron formation (AIF).
Haematites and standard itabirites
The iron grades and W1 values for 6385 itabirite and haematite 
core samples from Serra Azul are shown on the diagrams 
of Figure 11, which indicate a clear dominance of compact 
and hard compact ores in respect to other ore types. As 
consequence of the supergene enrichment process, the softer 
the material, the richer it is in average. Histograms of the iron 
grade for the different ore types as well as their statistics are 
shown on Figure 12 and Table 2. Under increasing degrees of 
weathering, the ore changes from hard compact, to compact, 
to semi-compact, and fi nally to friable. Accordingly, the 
iron grade increases and the ores become progressively 
heterogeneous. 
The iron grade of hard compact itabirite (fresh itabirite) 
from Serra Azul typically varies between 30 and 40 per cent, 
averaging 36 per cent, which corresponds to a haematite/
FIG 10 - Cross-section 2200 – Pau de Vinho Deposit (for location see Figure 6).
0,0%
0,0%
0,2%
0,4%
1,3%
6,5%
23,7%
27,4%
13,9%
8,6%
6,8%
6,9%
3,9%
0,5%
0% 10% 20% 30%
Relative frequency
2%
3,
5% 4
,3
%
4,
9%
4,
5%
3,
8%
3,
4%
3,
3%
3,
0%
2,
8% 3,
4% 4
,2
%
4,
6%
6,
6%
9,
5%
12
,5
% 1
4,
5%
7,
8%
1,
1%
0,
1%
0%
2%
4%
6%
8%
10%
12%
14%
16%
re
la
ti
ve
fr
eq
ue
nc
y
0
10
20
30
40
50
60
70
0 20 40 60 80 100
%
Fe
% > 6,35 mm
FIG 11 - Distribution of W1 and iron grade of the siliceous (standard) itabirites and haematites.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
NEW ORE TYPES FROM THE CAUÊ BANDED IRON FORMATION, QUADRILÁTERO FERRÍFERO, MINAS GERAIS, BRAZIL
67
martite (Fe
2
O
3
) content of around 50 per cent. The average 
silica content is 47 per cent, meaning that the hard compact 
itabirite is a rock composed basically of haematite/martite 
and quartz, with very low amounts of other components, 
which may include phosphorous, alumina and loss on ignition 
(Table 3).
Magnetitic itabirite
Less frequent in the western half of Serra Azul 
(Western and Central Mines), the magnetitic itabirite occurs 
mainly in the Eastern Mine, Camargos and Pau de Vinho 
deposits (Figure 6). 
The distribution of the iron grade and W1 for 375 core 
samples of the magnetitic itabirite is shown in the diagrams 
of Figure 13. Typically, the magnetitic itabirite it is a hard rock 
(W1 >80 per cent) with iron grade between 25 and 35 per cent. 
The iron bearing minerals are magnetite, haematite, goethite 
and grunerite. Calcite and dolomite are also present, as well 
as quartz. The presence of carbonates implies in relatively 
high contents of CaO and MgO and, consequently, high loss 
on ignition (LOI) (Table 4). The distribution of CaO, MgO and 
LOI is shown on Figure 14. The contents of these components 
tend to be similar, as indicated by the concentration of the plots 
in the centre of the diagram. Field and drill core observations 
indicate that the weathering product of the magnetitic itabirite 
is the goethitic, phosphorous-rich, ochreous (Dorr, 1969) or 
yellow iron formation (Pires, 1979, Pires, Aranha and Cabral, 
2005) widely known in the QF region.
Argillaceous iron formation 
A soft, iron bearing rock that looks like a phyllite occurs 
near to the stratigraphic top (close to the footwall, since the 
sequence is overturned) of the Cauê Formation in the eastern 
0,
0%
0,
3%
0,
6% 1,
2% 2,
2% 5
,5
% 1
0,
6% 14
,0
%
16
,2
%
17
,7
%
19
,6
%
11
,3
%
0,
8%
0%
10%
20%
30%
40%
50%
<
10
10
_1
5
15
_2
0
20
_2
5
25
_3
0
30
_3
5
35
_4
0
40
_4
5
45
_5
0
50
_5
5
55
_6
0
60
_6
5
>
65
re
la
ti
ve
fr
eq
ue
nc
y
% Fe
Friable
0,
0%
0,
1%
0,
7%
1,
3%
5,
9%
13
,8
% 17
,7
%
20
,5
%
15
,3
%
10
,4
%
9,
4%
4,
9%
0,
0%
0%
10%
20%
30%
40%
50%
<
10
10
_1
5
15
_2
0
20
_2
5
25
_3
0
30
_3
5
35
_4
0
40
_45
45
_5
0
50
_5
5
55
_6
0
60
_6
5
>
65
re
la
ti
ve
fr
eq
ue
nc
y
% Fe
Semi compact
0,
0%
0,
2%
0,
3% 1,
8%
8,
2%
22
,9
%
30
,6
%
19
,4
%
8,
3%
4,
1%
2,
4%
1,
5%
0,
2%
0%
10%
20%
30%
40%
50%
<
10
10
_1
5
15
_2
0
20
_2
5
25
_3
0
30
_3
5
35
_4
0
40
_4
5
45
_5
0
50
_5
5
55
_6
0
60
_6
5
>
65
re
la
ti
ve
fr
eq
ue
nc
y
% Fe
Compact
0,
0%
0,
2%
0,
3% 1,
1%
8,
2%
36
,8
%
38
,3
%
9,
6%
2,
3%
1,
0%
1,
1%
0,
7%
0,
5%
0%
10%
20%
30%
40%
50%
<
10
10
_1
5
15
_2
0
20
_2
5
25
_3
0
30
_3
5
35
_4
0
40
_4
5
45
_5
0
50
_5
5
55
_6
0
60
_6
5
>
65
re
la
ti
ve
fr
eq
ue
nc
y
% Fe
Hard Compact
FIG 12 - Histograms of the iron grade for diff erent ore types of the standard itabirites and haematites.
 Friable Semi-compact Compact Hard compact
n 1.478 1.018 984 2.905
Maximum 68.9 64.9 65.4 68.4
Percentile 95% 62.5 60.0 53.8 45.7
3rd quartile 56.6 49.9 42.3 38.3
Average 48.6 43.2 38.5 36.0
1st quartile 42.0 36.2 33.6 32.7
Percentile 5% 30.9 28.4 27.7 28.1
Minimum 10.6 11.3 13.2 11.7
Standard deviation (SD) 10.09 9.80 7.73 6.19
Variance 101.83 95.96 59.76 38.29
TABLE 2
Statistics of the iron grade for the diff erent types of haematite and standard itabirites (in per cent).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
L Q AMORIM AND F F ALKMIM
68
n = 2905 w1 Fe SiO
2
P Al
2
O
3
LOI CaO MgO
Maximum 97.7 68.4 81.5 0.440 7.84 11.19 1.37 1.22
Percentile 95% 88.6 45.7 58.7 0.077 1.17 3.66 0.12 0.18
3rd quartile 84.3 38.3 52.3 0.036 0.42 0.99 0.03 0.05
Average 80.2 36.0 47.0 0.030 0.37 0.87 0.03 0.07
1st quartile 75.9 32.7 43.5 0.014 0.14 0.17 0.01 0.05
Percentile 5% 71.4 28.1 32.0 0.007 0.05 0.03 0.01 0.05
Minimum 70.1 11.7 1.1 0.003 0.05 0.00 0.01 0.05
SD 5.4 6.2 9.2 0.028 0.49 1.22 0.07 0.09
Variance 29.3 38.3 85.2 0.001 0.24 1.49 0.00 0.01
TABLE 3
Statistics of the chemical characteristics of the hard compact itabirite (in per cent).
0,0%
0,3%
0,3%
2,4%
9,1%
26,7%
48,3%
10,4%
2,1%
0,0%
0,5%
0,0%
0,0%
0,0%
0% 20% 40% 60%
Relative frequency
0,
3%
0,
5%
0,
8%
0,
8%
0,
8%
0,
0%
0,
3% 0,
8%
0,
0%
0,
3%
0,
5%
0,
5%
0,
5%
0,
5%
6,
7%
11
,5
%
34
,7
%
35
,2
%
4,
0%
1,
3%
0%
5%
10%
15%
20%
25%
30%
35%
40%
Re
la
ti
ve
fr
eq
ue
nc
y
0
10
20
30
40
50
60
70
0 20 40 60 80 100
%
Fe
% > 6,35 mm
FIG 13 - Distribution of W1 and iron grade for the magnetitic itabirite.
n = 375 w1 Fe SiO
2
P Al
2
O
3
LOI CaO MgO
Maximum 98.3 53.2 71.4 0.242 6.96 32.22 20.53 14.30 
Percentile 95% 90.6 38.4 47.1 0.043 1.02 16.34 11.18 7.84 
3rd quartile 86.6 33.5 40.7 0.029 0.24 10.19 7.13 5.23 
Average 80.4 30.7 38.2 0.028 0.30 7.70 5.11 4.34 
1st quartile 80.3 28.3 35.5 0.021 0.05 4.55 2.68 3.02 
Percentile 5% 57.0 22.2 29.0 0.014 0.05 1.29 0.22 1.34 
Minimum 0.0 9.4 16.6 0.003 0.05 0.16 0.01 0.05 
SD 14.7 5.1 5.6 0.022 0.65 4.81 3.40 2.04 
Variance 215.9 26.4 31.4 0.000 0.42 23.14 11.57 4.17 
TABLE 4
Statistics of the chemical characteristics of the magnetitic itabirite (in per cent).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
NEW ORE TYPES FROM THE CAUÊ BANDED IRON FORMATION, QUADRILÁTERO FERRÍFERO, MINAS GERAIS, BRAZIL
69
half of Serra Azul (Eastern Mine, Camargos and Pau de Vinho 
deposits). 
The statistics of 441 core samples of this ore type is shown on 
Table 5 and the distribution of iron grade and W1 parameter 
on Figure 15. Differently from the hard itabirites, more than 
50 per cent of the iron grades are smaller than 30 per cent, 
typically averaging 29 per cent Fe.
Comparison between the three major ore types 
The SiO
2
,
 
Fe, and CaO+MgO contents of the hard compact 
itabirite, magnetitic itabirite and AIF are represented in the 
triangular diagram of Figure 16. Due to the absence of CaO 
and MgO in the hard compact itabirite and argillaceous iron 
formation, they plot along the Fe-SiO
2
 axis. Thus, an increase 
in Fe in these rock types implies in a decrease in SiO
2
 and vice 
versa. The magnetitic itabirite, on the other hand, containing 
CaO and MgO, plot in central sector of the diagram. Its 
distribution shows that an increase in the CaO+MgO content 
implies in an approximately equal decrease in both Fe and SiO
2
.
LOI
CaO MgO
FIG 14 - Triangular diagram CaO-MgO-LOI for the magnetitic itabirite.
n = 441 w1 Fe SiO
2
P Al
2
O
3
LOI CaO MgO
Maximum 78.7 57.1 83.0 0.620 16.66 12.92 0.15 1.44
Percentile 95% 55.3 45.6 73.8 0.203 6.58 7.78 0.06 0.44
3rd quartile 24.7 36.0 58.2 0.125 3.38 5.30 0.03 0.21
Average 17.5 28.6 48.5 0.097 2.81 4.21 0.03 0.18
1st quartile 4.7 21.5 40.1 0.054 1.54 2.90 0.02 0.05
Percentile 5% 0.8 11.7 21.9 0.028 0.65 1.24 0.01 0.05
Minimum 0.0 7.5 5.7 0.011 0.33 0.60 0.01 0.05
SD 16.9 10.5 14.9 0.069 2.16 2.09 0.02 0.18
Variance 284.4 109.7 222.6 0.005 4.66 4.37 0.00 0.03
TABLE 5
Statistics of the chemical characteristics of the argillaceous iron formation (in per cent).
0,0%
2,5%
10,7%
9,5%
12,5%
18,6%
19,0%
14,1%
7,3%
3,2%
1,8%
0,9%
0,0%
0,0%
0% 5% 10% 15% 20%
Relative frequence
27
,0
%
15
,9
%
13
,8
%
10
,7
%
8,
4%
6,
3%
3,
6% 4,
1%
1,
1% 1,
8%
2,
0%
1,
4%
1,
4%
1,
4%
0,
2% 0,
9%
0,
0%
0,
0%
0,
0%
0,
0%
0%
5%
10%
15%
20%
25%
30%
Re
la
ti
ve
fr
eq
ue
nc
y
0
10
20
30
40
50
60
70
0 20 40 60 80 100
%
Fe
% > 6,35 mm
FIG 15 - Distribution of W1 and iron grade for the argillaceous iron formation.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
L Q AMORIM AND F F ALKMIM
70
CONCLUSIONS
The enormous demand for iron ore in the last ten years is 
causing a rapid depletion of the ores traditionally exploited 
in the QF. In order to maintain the present status of the QF 
as global iron ore producing area, exploitation of low grade 
and compact ores, so far considered waste, will be necessary. 
In this scenario, the Mineração Usiminas project conducted 
in Serra Azul is pioneer in increasing the knowledge on fresh 
itabirites and incorporating them in the category of ore. Three 
new types of lower grade ores (~36 per cent Fe), namely 
standard itabirite, magnetitic itabirite and argillaceous iron 
formation, have been include in the Serra Azul reserves. 
The project is being very successful, even considering all 
technological, economic and environmental challenges to be 
faced in the next years. 
ACKNOWLEDGEMENTS
L Q Amorim is grateful to the Mineração Usiminas for permitting 
the publication of the Serra Azul project data. F F Alkmim 
benefi ted from the Conselho Nacional de Desenvolvimento 
Científi co e Tecnológico – CNPq research grant #307531/2009-
0. The manuscript of this paper benefi ted from comments and 
constructive criticism by an anonymous reviewer.
REFERENCES
Alkmim, F F, 2009. Geologic outline of the Curral ridge between the 
Paraopeba valley and Ponta da Serra. Mineração Usiminas, 54 p, 
unpublished technical report. 
Alkmim, F F and Marshak, S, 1998. Transamazonian Orogeny 
in the southern São Francisco craton region, Brazil: Evidence 
for Paleoproterozoic collision and collapse in the Quadrilátero 
Ferrífero, Precambrian Research, 90:29-58.
Babinski, M, Chemale Jr, F and Van Schmus, W R, 1995. The Pb/
Pb age of the Minas Supergroup carbonate rocks, Quadrilátero 
Ferrífero, Brazil, Precambrian Research, 72:235-245.
Carneiro, M A, Teixeira, W and Machado, N, 1994. Geological 
evolution of a sialic Archean crustal fragment from the Quadrilátero 
Ferrífero in eastern-central Brazil, based on U-Pb, Sm-Nd, Rb-Sr 
and K-Ar isotopic constrains, Terra Nostra, 2:12-13. 
Chauvet, A, Faure, M, Dossin, I and Charvet, J, 1994. A three 
stage structural evolution of the Quadrilátero Ferrífero: 
Consequences for the Neoproterozoic age and the formation of 
gold concentrations of the Ouro Preto area, Minas Gerais, Brazil, 
Precambrian Research, 68:139-167.
Chemale Jr, F, Rosière, C A and Endo, I, 1994. The tectonic evolution 
of the Quadrilátero Ferrífero, Minas Gerais, Brazil, Precambrian 
Research, 65:25-54.
Clout, JM F and Simonson, B M, 2005. Precambrian iron formations 
and iron formation-hosted iron ore deposits, Economic Geology 
and the Bulletin of the Society of Economic Geologists, 100th 
Anniversary Volume 1905 - 2005, pp 643-680.
Dorr, J and van, N II, 1964. Supergene iron ores of Minas Gerais, 
Brazil, Economic Geology, 59:1203-1240.
Dorr, J and van, N II, 1965. Nature and origin of the high-grade 
hematite ores of Minas Gerais, Brazil, Economic Geology, 
60(1):1-46.
Dorr, J and van, N II, 1969. Physiographic, stratigraphic and structural 
development of the Quadrilátero Ferrífero, Minas Gerais, Brazil, 
US Geological Survey Professional Paper, 641-A:1-110.
Eichler, J, 1968. O Residual and supergene enrichment of itabirites 
during weathering, Geologia (UFRJ), 1:29-40.
Eschwege, W L v, 1833. Pluto Brasiliensis, 375p (Reimer).
Gole, M J and Klein, C, 1981. Banded iron-formations through much 
of Precambrian time, Journal of Geology, 89:169-183.
Guild, P, 1953. Iron deposits of the Congonhas District, Minas 
Gerais, Brazil, Economic Geology, 48:639-676.
Hagemann, S, Rosière, C A, Lobato, L M, Baars, F J, Zuchetti, M and 
Silva, R C F, 2006. Controversy in genetic models for Proterozoic 
high-grade, banded iron formation (BIF)-related iron deposits – 
Unifying or discrete model(s)? Transactions of the Institutions of 
Mining and Metallurgy, Applied Earth Science, 115(4):B147-B151.
Hartmann, L A, Endo, I, Suita, M T F, Santos, J O S, Frantz, J C, 
Carneiro, M A, Naughton, N J and Barley, M E, 2006. Provenance 
and age delimitation of Quadrilátero Ferrífero sandstones based 
on zircon U-Pb isotopes, Journal of South American Earth 
Sciences, 20:273-285.
Herz, N, 1978. Metamorphic rocks of the Quadrilátero Ferrífero, 
Minas Gerais, Brazil, US Geological Survey Professional Paper, 
641-C:1-81.
Jordt-Evangelista, H, Alkmim, F F and Marshak, S, 1992. 
Progressive metamorphism and the occurrence of three 
Al2O3SiO5 polymorphs (kyanite, andalusite, sillimanite) in 
the Sabará Formation, Ibirité, Iron Quadrangle, Minas Gerais, 
Revista da Escola de Minas, 45(1/2):157-160.
Machado, N, Noce, C M, Ladeira, E A and Belo de Oliveira, O, 1992. 
U-Pb Geochronology of Archean magmatism and Proterozoic 
metamorphism in the Quadrilátero Ferrífero, southern São 
Francisco craton, Brazil, Geological Society of America Bulletin, 
104:1221-1227.
Machado, N, Schrank, A, Noce, C M and Gauthier, G, 1996. Ages 
of detrital zircons from Archean-Paleoproterozoic sequences: 
Implications for Greestone Belt setting and evolution of 
Transamazonian foreland basin in Quadrilátero Ferrífero, 
southeast Brazil, Earth and Planetary Science Letters, 
141:259-276. 
Marshak, S, Alkmim, F F and Jordt-Evangelista, H, 1992. Proterozoic 
crustal extension and the generation of dome-and-keel structure 
in an Archean granite-greenstone terrane, Nature, 357:491-493.
Melfi , A J, Pédro, G, Nalovic, L, Queiroz Neto, J P, 1976. Study about 
the chemical weathering of Brazilian Itabirites, Cahiers ORSTOM, 
Série Pédologie, XIV(3):179-192. 
Müller, G, Schuster, A and Hoefs, J 1982. Oxygen isotope variations 
in polymetamorphic iron ores from the Quadrilátero Ferrífero, 
Brazil, Revista Brasileira de Geociências, 12(1-3):248-355.
Noce, C M, 1995. Geochronology of the magmatic, sedimentary and 
metamorphic events recorded in the Iron Quadrangle region, 
Minas Gerais, PhD thesis (unpublished), Instituto de Geociências, 
Universidade de São Paulo, 128p.
Fe
SiO2 CaO+MgO
MtIb
AIF
Ib
FIG 16 - Triangular diagram Fe - SiO
2 
- CaO + MgO for standard itabirite (ib), 
magnetic itabirite (MtIb) and argillaceous iron formation.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
NEW ORE TYPES FROM THE CAUÊ BANDED IRON FORMATION, QUADRILÁTERO FERRÍFERO, MINAS GERAIS, BRAZIL
71
Noce, C M, Zucchetti, M, Armstrong, R, Dantas, E L, Baltazar, O 
F, Renger, F E and Lobato, L M, 2005. Age of felsic volcanism 
and the role of ancient continental crust in the evolution of 
the Neoarchean Rio das Velhas Greenstone belt (Quadrilátero 
Ferrífero, Brazil): U-Pb zircon dating of volcaniclastic graywackes, 
Precambrian Research, 141:67-82. 
Pickard, A L, 2002. SHRIMP U-Pb zircon ages of tuffaceous 
mudrocks in the Brockman Iron Formation of the Hamersley 
Range, Western Australia, Australian Journal of Earth Sciences, 
49:491-507. 
Pickard, A L, 2003. SHRIMP U-Pb zircon ages for the 
Palaeoproterozoic Kuruman Iron Formation, Northern Cape 
province, South Africa: Evidence for simultaneous BIF deposition 
on Kaapvaal and Pilbara cratons, Precambrian Research, 
125:275-315.
Pires, F R M, 1979. Structural geology and stratigraphy at the junction 
of the Curral anticline and the Moeda syncline, Quadrilátero 
Ferrífero, Minas Gerais, Brazil, PhD thesis, 220 p, Michigan 
Technological University.
Pires, F R M, 1995. Textural and mineralogical variations during 
metamorphism of the Proterozoic Itabira Iron Formation in the 
Quadrilátero Ferrífero, Minas Gerais, Brazil, Anais da Academia 
Brasileira de Ciências, 67(1):77-105.
Pires, F R M, Aranha, D M and Cabral, A R, 2005. Volcanic 
Origin of the Proterozoic Itabira Iron Formation, Quadrilátero 
Ferrífero, Minas Gerais, Brazil, in Proceedings Iron Ore 2005, 
pp 119-121 (The Australasian Institute of Mining and Metallurgy: 
Melbourne).
Ramanaidou, E R, 2009. Genesis of lateritic iron ore from banded 
iron-formation in the Capanema mine (Minas Gerais, Brazil), 
Australian Journal of Science, 56(4):603-618.
Ramanaidou, E R and Morris, R C, 2010. Comparison of supergene 
mimetic and supergene lateritic iron ore deposits, Transactions of 
the Institutions of Mining and Metallurgy, Applied Earth Science, 
119(1):B35-B37.
Ramanaidou, E R, Nahon, D, Decarreau, A and Melfi , A J, 1996. 
Hematite and Goethite from duricrusts developed by lateritic 
chemical weathering of Precambrian banded iron formations, 
Minas Gerais, Brazil, Clays and Clay Minerals, 44(1):22-31.
Renger, F E, Noce, C M, Romano, A W and Machado, N, 1995. 
Stratigraphic evolution of the Minas Supergroup: 500 Ma geologic 
record in the Iron Quadrangle, Minas Gerais, Brasil, Geonomos, 
2(1):1-11.
Romano, A W, 1989. Tectonic evolution of northwestern Iron 
Quadrangle, Minas Gerais, Brazil. PhD Thesis (unpublished), 
Université de Nancy I, Nancy, France.
Romano, A W, Bertrand, J-M, Michard, A and Zimmermann, J-L, 
1991. Contractional tectonics and Lower Proterozoic detachments 
(~2000 Ma Transamazonian orogeny) in the northern Iron 
Quadrangle (Minas Gerais, Brazil), Comptes Rendus Academie 
de Science de Paris, t.313, Série II, p 1195-1200.
Rosière, C A and Chemale Jr, F, 2006. Itabirites and high grade 
iron ores from the Iron Quadrangle – Overview and discussion, 
Geonomos, 8:27-43.
Rosière, C A and Rios, F J, 2004. The origin of hematite in high-
grade iron ores based on infrared microscopy and fl uid inclusion 
studies: The example of the Conceição mine, Quadrilátero 
Ferrífero, Brazil, Economic Geology, 99:611-624. 
Rosière, C A, Spier, C A, Rios, F J and Suckau, V E, 2008. The 
itabirite from the Quadrilátero Ferrífero and related high-grade 
ores: An overview, Reviews in Economic Geology, 15:223-254.
Silva, A M, Chemale Jr, F, Kuyjimian, R and Heaman, L, 1995. Mafi c 
dyke swams of the Qaudrilátero Ferrífero and Espinhaço range, 
MG, Brazil, Revista Brasileira de Geociências, 25:124-137.
Simmons, G C, 1968. Geology and Iron Deposits of the Western 
Serra do Curral, Minas Gerais, Brazil, US Geological Survey 
Professional Paper, 341C, 57 p.
Spier, C A, Oliveira, S M and Rosière, C A, 2003. Geology and 
geochemistry of the Águas Claras and Pico mines, Quadrilátero 
Ferrífero, Minas Gerais, Brazil, Mineralium Deposita, 38:751-754.
Spier, C A, Oliveira, S M, Sial, A N and Rios, F J, 2007. Geochemistry 
and genesis of the banded iron formations of the Cauê Formation, 
Quadrilátero Ferrífero, Minas Gerais, Brazil, Precambrian 
Research, 152:170-206.
Spier, C A, Vasconcelos, P M and Oliveira, S M B, 2006. 40Ar/39Ar 
geochronologicalconstraints on the evolution of lateritic iron 
deposits in the Quadrilátero Ferrífero, Minas Gerais, Brazil, 
Chemical Geology, 234:79-104. 
Trendall, A F, 2002. The signifi cance of iron-formation in the 
Precambrian stratigraphic record, International Association of 
Sedimentologists Special Publication, 33:33-66. 
Varajão, C A C, 1994. Supergene evolution of the Palladium-rich ore 
of the Maquiné Mine, Iron Qaudrangle, Minas Gerais, Brazil, PhD 
thesis (unpublished), Université de Marseille, 215 p. 
Varajão, C A C, Ramanaidou, E R, Colin, F and Nahon, D, 
1997. Genesis of hard hematite ore bodies: metasomatism, 
sedimentation or supergene alteration? Revista da Escola Minas, 
50:40-43.
Zucchetti, M, Baltazar, O F and Raposo, F O, (eds), 1998. Rio das 
Velhas Project – Explanatory note of the geologic map, 1:100.000 
scale, DNPM/CPRM Brazilian Geological Survey, 157 p.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 73
INTRODUCTION
Five fundamentally different processes can yield high-grade 
iron ore products from unmineralised iron formation, or its 
precursor sediment (Figure 1). From these, four are natural 
processes and one is industrial. 
Iron ore formation through syngenetic or early 
diagenetic processes
Syngenetic iron ore genesis models postulate that iron 
enrichment is either a primary depositional feature or an early 
diagenetic modifi cation of iron formation. A syn-diagenetic-
supergene concept of ore formation is best described in recent 
papers by Lascelles (2006a, 2006b) who advocates formation 
of a ‘chert free BIF’ during diagenesis (summarised in Figure 2). 
In this model, density currents, with ultra-fi ne colloidal iron 
oxides, iron silicates and minor iron carbonates deposited 
iron-rich sediments on to an unstable sea fl oor. Chert formed 
from the diagenetic dissociation of unstable iron-silicates into 
hydrous iron oxides and gelatinous silica. Unstable conditions 
induced fracturing of the iron oxide bands and the forced 
escape of gelatinous silica under pressure of the overlying 
iron-rich sediments. Subsequent collapse of the overlying 
sediment induced further fracturing with a ‘domino-effect’ 
to the top of the sediment, producing large volumes of chert-
free BIF, however with the original carbonates and some iron 
silicates preserved (Lascelles, 2006a, 2006b). Subsequently 
the iron formations and early ore deposits were completely 
oxidised and further leached, to depths exceeding the current 
mine and drilling depths. This deep oxidation gives the early 
ore deposits their high goethite contents. 
The silica that was removed from the orebodies could form 
cappings over high-grade ores, or occur on the contact of 
overlying shales. The location of syngenetic ore concentrations 
could also be controlled by syn-sedimentary structures, since 
these are the main cause of unstability in the sedimentary pile.
1. Principal Geologist, Rio Tinto Exploration Pty Ltd, 37 Belmont Avenue, Belmont WA 6104. Email: hilke.dalstra@riotinto.com
From Banded Iron Formation to 
Iron Ore – Genetic Models and Their 
Application in Iron Ore Exploration 
in the Hamersley Province, Western 
Australia
H J Dalstra1
ABSTRACT
High-grade iron ore may be derived from banded iron formation (BIF) through four fundamentally 
different processes:
1. syngenetic precipitation of chert free BIF, 
2. residual enrichment by removal of gangue minerals from BIF to form bedded (residual) iron 
ore, 
3. mechanical erosion of iron oxide and chert from bedded iron ore or BIF and re-deposition and 
upgrading as detrital iron ore, and 
4. chemical removal of iron from BIF and precipitation elsewhere as a secondary (channel) iron 
deposit. 
Finally, the BIF itself may constitute iron ore if a high-grade concentrate utilised for pellets or 
sinters can be derived from it through mechanical (industrial) separation of iron oxide by means 
of crushing/milling and mineral separation.
Historically, exploration for high-grade BIF hosted iron ores was focused on the residual (bedded) 
deposits and driven by supergene concepts. Exploration for detrital and channel type iron ores was 
mostly secondary to exploration for bedded deposits, but recently, a more focused exploration effort 
has led to the discovery of enormous moderate to low iron grade deposits in the Pilbara region of 
Western Australia. Exploration for high-grade iron ores using syngenetic concepts has been limited, 
but together with exploration using hypogene concepts offers the best hope of fi nding concealed 
deposits. Exploration for BIF sources suitable for concentrate ores commenced in the ‘old’ iron ore 
provinces after the high-grade resources were depleted and has only recently gained importance in 
Australian BIF provinces, mainly as a result of high iron ore prices in the last decade. 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
H J DALSTRA
74
Exploration using syngenetic concepts has been modest in 
the Hamersley province. Early exploration in the Hamersley 
province based on these concepts focused on the presence 
of syn-sedimentary structures which could have triggered 
the collapse of the sedimentary pile and on the presence of 
anomalous concentrations of silica above the target iron 
formations. Application of this concept led to the discovery 
of signifi cant concealed, high-grade, low P mineralisation, 
named Lens 2 and Lens 3 deep below the outcropping ores 
at Paraburdoo. However, mineralisation at these sites occurs 
within the weathered zone, and shows abundant wall rock 
alteration so supergene and hypogene concepts cannot be 
excluded for its origin. 
Syngenetic concepts offer possibly the best hope of fi nding 
large, concealed moderate to high-grade iron ore deposits in 
the Hamersley province. The syngenetic concept of Lascelles 
(2006b) postulates the presence of large moderate to high-
grade magnetite orebodies with small amounts of diagenetic 
carbonate and silicates below the weathered surface. Because 
of their early formation, these orebodies should have low 
porosities (comparable to primary BIF) and therefore high 
densities and should show signifi cant reduction in thickness 
compared to the unmodifi ed iron formation. However, they 
would not have signifi cant alteration haloes. Exploration 
targeting for such orebodies would be diffi cult, because they 
would have very little geological or geochemical surface 
expression. Geophysical methods, particularly gravity could be 
utilised to target zones of anomalously high density although 
the thinning of the host BIF at the site of mineralisation would 
probably obliterate much of the expected anomaly.
Residual iron ore formation through removal 
of gangue minerals – bedded ores
Residual ore formation means removal of all-, or most gangue 
minerals from the BIF, while preserving the iron oxides to form 
a high-grade residual or bedded iron ore. Although ore genesis 
models vary in fundamental aspects of the fl uids involved, 
and the timing and sequence of the removal of the gangue 
minerals, most agree that iron enrichment was predominantly 
the result of residual concentration, rather than iron addition 
to the mineralised sites. Hypogene genetic models propose 
that gangue minerals were mostly removed by ascending 
fl uids, either warm basinal brines (eg Taylor et al , 2001; 
Thorne, Hagemann and Barley, 2004) or hotter magmatic 
fl uids (Lobato et al, 2008). Supergene models postulate that 
descending meteoric or modifi ed meteoric waters were largely 
responsible for gangue removal (eg McLeod, 1966; Morris, 
1985). Several models propose multistage ore formation 
through sequential removal of gangue minerals from the BIF 
by a combination of supergene and hypogene processes (eg 
Taylor et al, 2001).
Supergene models for residual iron ore formation offer limited 
hope for discovery of concealed deposits. The prerequisite of 
the supergene concept, that ores are formed by descending 
solutions, implies that all orebodiesshould have a surface 
expression. The only exceptions are ancient supergene deposits 
which could occur below palaeo-unconformity surfaces, 
which are particularly prominent in the southern Hamersley 
province (eg the lower and upper Wyloo unconformities).
Exploration on the basis of recent supergene concepts 
led to rapid delineation of outcropping iron ore resources 
Precipitation as oxide-only BIF
or diagenetic removal of silica
BANDED 
IRON 
FORMATION
Magnetite or hematite concentrate
Mining, milling
and mineral separation
Chemical removal or replacement of 
gangue minerals from BIF
Supergene Hypogene
Hematite-goethite ores
Soft hematite ores
Enriched BIF ores
Magnetite-hematite ores
Magnetite-hematite-carbonate
proto-ores
Supergene overprinting
Hematite (goethite) 
ores
Pellets or sinter
Burial
Hematite ores
Erosion
Detrital ore
Cementation
Canga
BIF precursor sediment
Burial, diagenesis, 
metamorphism and deformation
Chert free BIF
Burial
Hematite 
conglomerate
Weathering
Chemical removal of iron oxide from BIF
iron in solution
Transport and reprecipitation
Channel Iron Deposit
BIF gravel
Supergene 
overprinting Detrital ore
Canga
Physical
Chemical
Pisolitic soil
Transport and 
cementation
Supergene overprinting
Hematite-(goethite) ores
Channel Iron 
Deposit
Processing
FIG 1 - Schematic presentation of processes leading from banded iron formation to high-grade iron ore.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
FROM BANDED IRON FORMATION TO IRON ORE – GENETIC MODELS AND THEIR APPLICATION IN IRON ORE EXPLORATION
75
in the Hamersley Province in the early 1960 ties. However, 
applying recent supergene concepts to iron ore exploration 
led to serious underestimation of the size of some deposits, 
particularly at Mount Whaleback. Although the deposit was 
discovered in 1957 by prospector Stan Hilditch, and brought 
to the attention of several mining companies including Rio 
Tinto, even after a considerable amount of drilling was carried 
out the resource was estimated at a modest 225 Mt (MacLeod, 
1966). Outcropping iron ore at Mt Whaleback occurred in 
relatively narrow bands up to 20 m in width, with signifi cant 
areas of unmineralised iron formation in between. Only after 
deep drilling was carried out, the presence of continuous and 
deep mineralisation below the narrow outcrops was recognised.
Morris’s refi ned supergene-metamorphic model allows for 
deeper, semi-concealed orebodies like Mt Whaleback, but 
postulates that, unless the deposits have gone through a post-
metamorphic leaching stage to remove the remnant goethite, 
and with it, the bulk of the phosphorous, the ores may not 
reach the quality to counter the economic problems of deep 
overburden (Morris, 1997). This model formed the backbone 
of a concealed orebody search (COBS) by Hamersley Iron 
from 1980 to 1994. Critical aspects of the exploration model 
were the presence of structural zones of permeability (faults, 
dykes, cross folds, etc) extending from surface to a possible 
mineralisation site at depth, and proximity to a major 
Palaeoproterozoic topographic feature, in this case the Mt 
McGrath Trough. This allowed for supergene upgrading of 
BIF in early Proterozoic times.
Although the program was unsuccessful in delineating a 
new large high-grade haematite resource in the Hamersley 
Province, several small occurrences of microplaty haematite 
were discovered in the Paraburdoo Western Ranges. These 
deposits occur underneath palaeoproterozoic unconformities 
deep below the present land surface, and are overlain by 
unmineralised BIF.
Supergene concepts offer little hope of fi nding large new 
deposits in areas of outcropping iron formation. However, 
signifi cant potential still exists in areas with abundant 
post-ore cover. Signifi cant recent discoveries along the 
margins of the Fortescue Valley including the down dip 
mineralisation in the Marra Mamba BIF along the Chichester 
Range, concealed mineralisation in the Brockman BIF at Iron 
Valley and Nyidinghu, and concealed mineralisation in the 
Boolgeeda BIF north of Newman testify to the exploration 
potential below cover.
H O2 H O + SiO2 2 H O2
A: Deposition of BIF as nontronite-
Fe-hydroxide turbiditic mud
B: Breakdown of nontronite to Fe-oxide and
colloidal silica 
D: De-watering and escape of silica leads to
localized formation of chert-free BIF 
Cherty BIF Cherty BIF Chert-free BIF 
E: Erosion exposes chert-free BIF,
supergene processes form weathered cherty BIF and 
 hematite or hematite-goethite ore 
C: Settling of Fe-oxide 
Nontronite
Fe-hydroxide
Silica
Fe-oxide
FIG 2 - Syngenetic/diagenetic ore genesis model presenting development of high-grade iron ore from banded iron formation (Lascelles, 2006b).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
H J DALSTRA
76
A hypogene concept for the genesis of residual, high-
grade haematite deposits was introduced by the Hamersley 
Iron Resources Task Force in 1995 (Figure 3). Fundamental 
aspects of this model were the presence of early normal faults 
to link deep carbonate aquifers with BIF, and the presence of 
alteration signatures in BIF (carbonate, microplaty haematite) 
and shales, dolerites or basalts (chlorite and talc), as well as 
evidence for deep weathering (Taylor et al, 2001). The latter 
is necessary for the fi nal upgrading of apatite and carbonate 
bearing protore to high-grade haematite ore. 
As for the COBS, application of this model was largely 
unsuccessful in the search for another large, high-grade 
haematite deposit in the Hamersley Province. In the period 
from 1996 to 2005 approximately 65 targets for high-
grade haematite had been generated and drill tested. This 
resulted in the intersection of high-grade (microplaty) 
haematite mineralisation at fi ve targets. Most signifi cant 
is the Wellthandalthaluna deposit (45 Mt @ 63.4 per cent 
Fe), a deposit which shares many characteristics with the 
Mt Tom Price deposit, particularly North deposit, be it on a 
much reduced scale. The deposit occurs in a faulted block of 
Brockman iron formation and contains microplaty haematite 
as well as magnetite-rich sections in the outcropping 
mineralisation. Across the main controlling structure, 
carbonate and talc alteration is common in protore in the 
base of the Dales Gorge member, in a downfaulted part of the 
deposit (Figure 4). This geometry is very similar to that of Mt 
Tom Price Southern Ridge or North Deposits, albeit in mirror 
image. A much larger fault juxtaposes mineralisation against 
unmineralised Fortescue Group rocks, resulting in a set-up 
with much less ore preserved than at Mt Tom Price.
Even though by now most outcrops of microplaty haematite 
in the Pilbara have been identifi ed and drill tested, hypogene 
concepts still offer signifi cant potential for discovery of large 
concealed deposits. Such deposits would be situated below 
unmineralised iron formation or below the Proterozoic 
EROSION AND WEATHERING STAGE:
Deep weathering results in removal of
apatite
Microplaty hematite-martite ore
OXIDATION STAGE:
Oxidized low salinity meteoric fluids
Overprint by microplaty hematite
Microplaty hematite-martite-ankerite 
(apatite) mineralization
 HYPOGENE STAGE:
Upward migration of brines
magnetite-carbonate (apatite) mineralization
Joffre Member 
Dales Gorge Member
Whaleback Shale Member
MT MCRAE SHALE
MT SYLVIA FORMATION
BROCKMAN IRON FORMATION
Stratigraphy
WEELI WOLLI FORMATION 
Bee Gorge Member
Paraburdoo Member
WITTENOOM FORMATION
Fault
Material Types
Hematite- (carbonate)
apatite
 
Magnetite -carbonate
-apatite
 
Hematite
 
 
Southern Ridge
meteoric fluids
Depth of weathering
Fluid flow
FIG 3 - Hypogene ore genesis model presenting development of high-grade iron ore from banded iron formation (modifi ed from Taylor et al, 2001).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
FROM BANDED IRON FORMATION TO IRON ORE – GENETIC MODELS AND THEIR APPLICATIONIN IRON ORE EXPLORATION
77
unconformities. The only geological indication for their 
existence could be in the form of a prospective geological 
structure or the presence of alteration minerals in outcropping 
BIF, shales or mafi c rocks.
Iron ore formation through redeposition – 
detrital ores
Until today, high-grade detrital iron ores have generally 
been mined as a by product of the much larger residual 
or bedded iron ore deposits. Canga (cemented haematite 
detrital) deposits of modest size, but of acceptable grade 
often immediately overlie residual deposits, and sometimes 
provide the fi rst iron ore in a new mining operation. Recently 
enormous (billions of tonnes) low to medium grade detrital 
iron deposits have been discovered in the central Hamersley 
province, an area with relatively few bedded high-grade 
deposits. It is possible that these will soon provide a signifi cant 
part of iron ore production from the province. Despite this 
imminent importance of detrital iron ores, they remain one of 
the least understood ore types of the province. 
Detrital iron ores of the Hamersley province can be 
subdivided into three distinct types:
1. Hard microplaty haematite conglomerates associated 
with the palaeoproterozoic Upper Wyloo unconformity. 
These are generally believed to be breakdown products 
of Proterozoic microplaty haematite deposits, and are 
generally too small to form economic deposits.
2. Canga associated with the Tertiary Hamersley Surface: 
cangas are goethite or limonite cemented haematite 
conglomerates, mostly well sorted, and often displaying 
crude bedding. These have been mined economically in 
for example the Brockman detrital deposit.
3. Loose haematite gravels: these gravels are moderate to well 
sorted masses consisting of rounded haematite pebbles in 
a silt or clay matrix. Bedding is generally absent. They are 
subclassed into mature (>60 per cent Fe) and immature 
(<60 per cent Fe) detrital mineralisation. A distinct group 
are the red-ochre detritals which are relatively fi ne grained 
iron-rich gravels characterised by abundant ochrous 
haematite. Excluding the hematite conglomerates, red 
ochre detritals are considered the oldest form of detrital 
material in the province, and often underlie channel iron 
deposits (CID). Mature, high-grade detrital ore was also 
mined in the Brockman detrital deposit.
Although the genesis of the Proterozoic haematite 
conglomerates as breakdown products of microplaty 
haematite deposits is considered relatively straightforward, 
the genesis of the latter two styles is less clear. 
Detrital iron ore deposits generally share a set of common 
features discussed below (see also Figure 5).
Canga is the second oldest form of detrital material and 
overlies deeply weathered bedrock at the base of the detrital 
pile. Canga’s were deposited on relatively steep topography, 
signifi cantly steeper than the slopes on which younger mature 
detritals were deposited. This may suggest that the terrain 
was stabilised by dense vegetation in a wetter climate. Rare 
collapse breccias or landslides formed when larger rock 
masses collapsed during erosion of the palaeo range fronts. 
The hard canga formed by cementation of a clay matrix 
(Killick, Churchward and Anand, 2008), and transformation 
of the clays to goethite and limonite.
The contact between canga and overlying uncemented 
mature detritals is generally unconformable, suggesting that 
this canga surface began to break down before deposition 
of the mature detritals. The canga surface was incised, and 
boulders of eroded canga, together with haematite pebbles 
from another high-grade haematite source, were deposited to 
form the mature detrital fans. 
The mature detrital fans are unique among normal alluvial 
fan systems in that they have a relatively uniform grain size, 
a monomineralic (Fe-rich) composition, a general absence of 
silty or sandy layers and a very homogeneous (non-bedded) 
HG
WS
DG3
DG2
DG1
FWZ
ENE
RL 350m
DDHWLT08
qtz veining
abundant
J6
J5
J4
J3
J2
J1
FWZ
HG
HG
DDHWLT01
DDHWLT02
HG
HG
FOR
Depth of Weathering
WSW
Joffre Member
Dales Gorge Member
Whaleback Shale Member
MT MCRAE SHALE
MT SYLVIA FORMATION
BROCKMAN IRON FORMATION
WITTENOOM FORMATION
 Hematite (magnetite)
Hematite-goethite (magnetite) Dolerite Dyke
N
o
rth
e
rn
 F
a
u
lt
FORTESCUE GROUP
Pillow lava & dolerite
Sheared rocks
Metres
0 50 100 150 200
MARRA MAMBA FORMATION
Diamond drillhole
Hematite/ magnetite carbonate (talc)
DDHWLT01
B
o
o
lg
e
e
d
a
 C
re
e
k
 F
a
u
lt
VV
V
V
V
V
V
V
V
V
V
V
VV
V
DDHWLT07
(proj.)
DDHWLT05
DDHWLT09
(proj.)
WESTERN
 AUSTRALIA
Perth
Section Area
FIG 4 - Cross-section through the Wellthandalthaluna deposit showing outcropping high-grade haematite (goethite, magnetite)
ore and concealed magnetite-carbonate-talc protore.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
H J DALSTRA
78
character. This suggests that they formed under special 
conditions, and that they have a very restricted source 
region. They are probably sheet wash deposits or seeve 
fans. Sporadic fl ash fl oods with a low mud and sand content 
fl owed evenly over previous deposited sediments on the fans. 
Due to the high porosity of the underlying detritals, water 
was quickly extracted from these fl oods, with the remaining 
gravels forming an evenly thick sheet over the entire fan. Low 
sediment loads of these fl oodwaters may indicate a dense 
cover of vegetation during deposition of the mature detritals.
Ancient ValleyAncient Valley
Bedded iron ore
Hamersley Surface
Erosion products forming mature
detrital fan on steep slope
Erosion of bedded ore
and Hamersley Surface material
Cementation and upgrading
of detritals to form canga
Further erosion of bedded ore and canga
 
Erosion products forming mature
detrital fan in trap site
Rapid incision of BIF bedrock
 
Erosion products forming colluvial cover
Canga outcrop
 
>45Ma
<5Ma
Oligocene to
 Miocene
Late Miocene
 to Pliocene
hardcap
 
Wittenoom Formation
 (dolomite)
MCRae/ Mt Sylvia
 Formations (shale)
Brockman
 Iron
 Formation
Mature detritals (concealed)
 
Hamersley surface partly
preserved below detrital sequence
 
FIG 5 - Six stage ore genesis model showing the formation of Brockman type high-grade detrital iron ore.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
FROM BANDED IRON FORMATION TO IRON ORE – GENETIC MODELS AND THEIR APPLICATION IN IRON ORE EXPLORATION
79
After deposition of the mature detritals, a renewed 
period of exposure and erosion started. This resulted in the 
formation of a hard cap on the detrital surface. The partly 
eroded mature detritals were subsequently overlain by lower 
grade detritals and colluvium. These formed more typical 
alluvial fan systems, with localised channels (fi ning up) and 
a predominance of mass-fl ow deposits or olistostromes which 
were generated by fl ood waters with higher mud content. The 
resulting sediments are mostly coarsening up conglomerates 
and pebbly mudstones.
The last (and presently continuing) stage of geological 
activity is intersection and erosion of the detrital and 
colluvium deposits, and continued erosion of the cangas as a 
result of a recent base-level drop.
There are two possible origins for the iron rich pebbles in 
detrital deposits:
1. mechanical breakdown of pre-existing residual (bedded) 
iron ores and redeposition of the ore clasts as high-grade 
detrital deposits; and 
2. mechanical breakdown of BIF, redeposition of the clasts as 
BIF gravels, followed by in situ upgrading of these deposits 
to high-grade detrital ores (see also Figure 5). 
In the Hamersley province the largest detrital deposits occur 
in the north, an area with relatively few residual iron ore 
deposits. Although some deposits may be spatially associated 
with bedded iron ore, many do not show this relationship.
Explorationby Rio Tinto for detrital iron ores in the 
Hamersley Province was initially focused on drilling of 
outcropping canga mineralisation, and scout drilling of cover 
sequences adjacent to these canga outcrops to target the 
concealed mature detrital fans. The presence of outcropping 
bedded mineralisation was not a signifi cant factor in 
targeting. Later the signifi cance of trap sites for hosting high-
grade detrital ores was recognised. These trap sites can form 
by erosion along suitable structures such as synclines or fault 
zones and can often result in narrow but deep ‘channels’ or 
pods of detrital material. Targeting for detrital ores was often 
assisted by narrow spaced ground gravity, because the detrital 
pods have good density contrast with the shaly host rocks 
(Flis, Butt and Hawke, 1998). 
Only recently the signifi cance of more widespread, lower 
grade detrital mineralisation in the northern parts of the 
Hamersley province has been recognised. These deposits 
often overlie channel iron deposits in the major drainage 
channels of the central Hamersley Province. Signifi cant 
new resources totalling several billions of tonnes have been 
identifi ed recently at Caliwingina, Serenity and Kings (Clarke 
et al, 2009; Dalstra et al, 2010), and in the Fortescue Valley 
at Koodaiderie and Marilana. Exploration for this style of 
mineralisation is currently still happening, and it is likely that 
more resources will be added in the near future.
Iron ore formation through secondary 
precipitation – channel iron deposits
Although secondary precipitation of iron at the site of the 
ore deposits is part of many ore genesis models, only in the 
formation of channel iron deposits it may have played a 
dominant role. 
For example, replacement of gangue minerals by goethite is 
part of the supergene metamorphic model of Morris (1985), 
but leaching of gangue resulting in volume loss and porosity 
increase appears to be the dominant factor of generating the 
bedded ore from BIF. Along the same line of arguing, some 
iron addition must have taken place during cementation of 
matrix material in canga by goethite.
The currently favoured ore genesis models see the genesis 
of CID as a combination of detrital and secondary enrichment 
processes (Harms and Morgan, 1964; Macleod, 1966; 
Morris and Ramanaidou, 2007; Dalstra et al, 2010) (see also 
Figure 6). The models propose that iron enrichment occurred 
as a result of normal weathering processes of the iron rich 
B: replacement stage
in headwaters iron in solution:
3+ 2+
Fe + organic = Fe (aq)
A: channel fill stage
silica taken out of the system:
 0
SiO (am)+ 2H O + organic = H SiO2 2 4 4
 in regolith:
pisoide formation
mechanical transport of
pisoids to channel
replacement of matrix and wood in sediment:
2+
Fe (aq) = Fe + e
3+ +
Fe + 3H O = Fe(OH) + 3H2 3
3+ -
C: inversion
e.g. Robe, Beasley
 
C: concealment
e.g. Bungaroo, Caliwingina, Kings
 
basaltic or shaly basement
colluviumdetrital ores
alluvium, clay, calcrete
basal conglomerate
vv
vv
vvv
v
v
vvvvv
vvv vv vv v
v
v
vvv
vv
vv BIF basement
FIG 6 - Three-stage ore genesis model presenting the formation of channel iron deposits type iron ore.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
H J DALSTRA
80
country rocks producing a deep regolith. During this phase, 
iron was transported in solution from the bedrock to the 
overlying soil profi le where it redeposited in the form of iron-
rich pelletoids. Erosion stripped these from this surface and 
slowly transported them to the channels, where they were 
redeposited. The initial channel fi ll was probably an iron 
rich mud, comprising pisoids, iron rich peloids and wood 
fragments, all in a fi ne grained silty matrix. Reduced iron rich 
groundwaters derived from source rocks in the headwaters of 
the channel carried iron in solution to these channel sediments 
possibly over a distance of hundreds of metres to kilometres. 
There, the iron was reprecipitated resulting in replacement 
of the silty matrix and the wood fragments by goethite. After 
formation of the CID, the channels were either inverted to 
mesas, or covered by younger detrital ores, colluvium and/or 
alluvium. Concealment occurred in areas where the bedrock 
comprised hard BIF, while inversion to mesas was favoured 
by softer basement rocks such as basalt or shale. The fi nal 
textures of the iron (hydroxides) of the CID do not resemble 
that of the BIF, because solution and reprecipitation has all 
but destroyed the initial texture of the iron (hydr)oxides.
Exploration for CID in the Hamersley province began in the 
early 1960s, even before recognition of the signifi cance of the 
high-grade haematite deposits. The fi rst deposits brought to 
the attention of RTZ by Hancock in early 1961 were in fact 
channel deposits, the Beasley River limonites and the Duck 
Creek CID’s, both of which remain unmined today. Recognition 
of the Robe River and Yandicoogina deposits followed soon. 
These early discovered deposits were mesas, channel systems 
where the CID profi le has been inverted through later erosion 
during the ongoing aridifi cation of the Australian continent. 
The Bungaroo Creek system was the fi rst CID system with a 
very large concealed resource, although it still has signifi cant 
outcrops in the upper reaches of the channel.
CID exploration by Rio Tinto from 2000 onwards focused 
on this concealed aspect of the CID, either below recent 
alluvials or below calcrete of the Oakover formation (Figure 6). 
The exploration model was simple, and encompassed an 
interpretation that CID deposits after their fi nal accumulation 
stage could be entirely concealed in their channels. Inversion 
to mesas would only occur in areas where modern drainages 
follow the ancient, and have signifi cantly incised their current 
bedrock. Application of this model by Rio Tinto and others 
recently led to discovery of an entirely new CID district in 
the Central Hamersley Ranges, more than 40 years after the 
fi rst mesas were discovered. This district encompasses the 
Caliwingina, Serenity and Kings CID systems. Total CID 
mineralisation in this area is of the order of several billions of 
tonnes and is of comparative size to the premining resources at 
Robe/Bungaroo Creek and slightly smaller than Yandicoogina 
(Dalstra et al, 2010).
The potential to discover signifi cant new outcropping CID 
systems in the Pilbara is limited. Any future exploration should 
focus on the concealed systems. The task is made more diffi cult 
because most large drainage channels of the Hamersley Range 
have now been tested by drilling. The discovery of the Solomon 
CID deposit in the northern Hamersley Range demonstrates 
that understanding landscape evolution and (re) interpreting 
palaeodrainage patterns is crucial for targeting new CID 
systems in the province (Clarke et al, 2009). 
CONCENTRATE IRON ORES THROUGH 
CRUSHING, MILLING AND BENEFICIATION OF 
BANDED IRON FORMATION
The appearance of mines dedicated to the mining of 
unmineralised iron formation as a source of magnetite or 
haematite concentrate is a relative new development in 
Western Australia. At the moment this activity is focused in 
the extreme northwestern part of the Hamersley province 
where proximity to the ocean avoids the high transportation 
costs of projects further inland. 
Geological factors that determine the economics of a 
concentrate operation from BIF include the amount of ore 
and ore to waste ratio during mining, the overall iron grade 
of the BIF, the minimum grind size to obtain a high-grade 
concentrate, the recovery of iron oxide from the BIF and the 
hardness of the ore. In general, soft, coarse grained ores with 
a relatively high in-ground iron grade are more likely to yield 
an economic concentrate than hard, fi ne grained and low iron 
grade ores.
Ore to waste ratios during mining are largely determined 
by the amount of overburden, the thickness and dips ofthe 
ore (BIF) bands and the relative percentage of intercalated 
waste bands. The minimum grind size is determined by the 
grain size of the oxides, their textures and the presence of 
inclusions such as apatite or pyrite. Recoveries are generally 
determined by the head grade of the ore (the amount of 
iron oxide in the rock), the mineralogy and texture of the 
ore and the grain size. Ore hardness on the other hand is 
largely determined by weathering, and to a lesser degree by 
mineralogy and metamorphic grade. Another important factor 
is the performance of the concentrating plant, which is largely 
defi ned by consistency of the ore feed. In extreme cases, 
feeding a coarse grained ore into a plant that was designed for 
fi ner grained material can lead to poorer plant performance.
Concentrate ore processing plants utilise either a magnetite 
circuit or a haematite circuit. Both haematite and magnetite 
circuits involve either: 
  three (and four) stage crushing followed by primary and 
secondary milling to the required grain size (for fi ne 
grained ores <25 μm); 
  primary crushing followed by semi-autogenous or 
autogenous milling; or 
  air swept autogenous milling (Dowson, Connely and Yan, 
2009). 
High pressure grinding rolls are a relatively new, but 
cost-effective technology to crush magnetite ores and is a 
technology based on interparticle breakage more than direct 
compression (Povey, 2009). Ultra-fi ne grinding is required for 
some ores, however, this is an energy intensive operation that 
needs to be well optimised and needs specialised magnetic 
separators (Li et al, 2009). 
In haematite circuits gangue minerals are rejected through 
a combination of fl occulation, desliming and fl otation 
processes. The concentrate is then thickened, fl ux is added 
and subsequently moulded into small pellets in a balling drum. 
These are baked in a kiln, before dispatch to the blast furnaces. 
Magnetite circuits are similar, but ore is concentrated 
through magnetic separation rather than fl otation. Magnetic 
separation can be achieved by low intensity magnetic 
separation (LIMS) using conventional or rare earth magnets 
in a rotating drum, or wet high intensity magnetic separation 
(WHIMS), using extremely strong magnetic fi elds to remove 
weakly magnetic particles from the processing stream. Recent 
innovations are machines that utilise the combined forces 
of magnetism, pulsating fl uid and gravity to continuously 
benefi ciate fi ne weakly magnetic minerals. Production of the 
pellets is similar to haematite circuits.
Benefi ciation operations will generally result in the 
production of three materials: a salable concentrate, a low-
grade product that is either reprocessed or stockpiled and 
tailings (Dowson, Connely and Yan, 2009).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
FROM BANDED IRON FORMATION TO IRON ORE – GENETIC MODELS AND THEIR APPLICATION IN IRON ORE EXPLORATION
81
Until today there has been relatively little exploration 
effort focused on BIF suitable for concentrate iron ore 
in the Hamersley province. Most activity has been in the 
northwestern part where weathering profi les are generally 
thinner and magnetite-rich BIF occurs closer to surface. 
The province is host to four major BIF’s, the Marra Mamba, 
Brockman, Weeli Wolli and Boolgeeda iron formations with a 
combined thickness of over 1000 m. Work by Rio Tinto and 
others (Ewers and Morris, 1980, 1981; McConchie, 1987) 
demonstrates that BIF’s (excluding shales bands) of the Dales 
Gorge Member have the highest iron contents (approximately 
32.5 per cent Fe), compared to Joffre (28 - 31.3 per cent Fe), 
Marra Mamba Newman Member (25.1 per cent Fe) and Weeli 
Wolli (21 - 25.6 per cent Fe). No data is currently available 
for the Boolgeeda BIFs. The average magnetite content of the 
Dales Gorge and Joffre BIFs is comparable at about 28 - 31 per 
cent, which is signifi cantly higher than the Marra Mamba BIF, 
which contains more iron in silicates and carbonates. 
The Brockman Iron Formation covers an outcrop area of 
approximately 10 000 km2. Even if less than ten per cent of 
the outcropping material could be mined and benefi ciated, 
this would still imply 1000 Bt of potential BIF source for 
magnetite concentrate and potentially 300 Bt of concentrate 
(at weight recoveries of approximately 30 per cent). 
The Balmoral project in the northwestern part of the 
Hamersley Province has reported a primary iron grade 
in the Joffre member of the Brockman Iron Formation of 
31.3 per cent Fe, producing a high-grade concentrate of about 
69 per cent Fe. Recovery of the iron is relatively low (approx 
75 per cent of total iron recovered for a Davis Tube recovery 
of 32.1 per cent), possibly because of the presence of iron in 
non-magnetic minerals. Pilot studies by Rio Tinto and others 
in other parts of the Province suggest that total iron recoveries 
of other iron formations are generally lower (<50 per cent), 
and the produced concentrates of poorer quality than those 
at Balmoral.
Exploration for magnetite or magnetite concentrate by Rio 
Tinto in Western Australia has been modest. Two projects 
resulted from analysis of regional magnetic data in the 
Hamersley Province; B26E in the Western Turner syncline 
targeted the Dales Gorge member and Silvergrass Peak in 
the northwestern Hamersley province targeted both the 
Joffre and Dales Gorge members. Both projects focused 
on BIF with anomalously high magnetic susceptibility. A 
third project targeted concealed eastern extensions to the 
Southdown deposit in the Albany Fraser Belt. This project 
resulted in delineation of signifi cant magnetite-concentrate 
resources, which were subsequently sold to Grange Resources. 
Opportunistic work was carried out on areas of the Hamersley 
province where fresh diamond core obtained in the search for 
high-grade deposits was available, such as North Deposit and 
Southern Ridge at Mt Tom Price, the Turner Syncline, Giles 
Mini and other parts of the Rhodes Ridge area.
The Silvergrass Peak project incorporated drilling two deep 
diamond holes on separate intense magnetic anomalies (up 
to 140 000 nT) associated with a late NNE trending gabbroic 
dyke (Figure 7). The Silvergrass Dyke is an approximately 
30 - 40 m thick hypersthene-gabbro, with a SHRIMP zircon 
age of 764 ± 10 Ma, and is a member of the Neoproterozoic 
Mundine Well dyke swarm of NW Australia (Wingate and 
Giddings, 2000). 
Drilling of the magnetic targets identifi ed signifi cant 
recrystallisation of the Joffre Member BIF adjacent to the 
gabbro dyke (Figure 7), however the lateral extent of this away 
from the contact was only 10 m. In the contact zone, the BIF 
is recrystallised into a mosaic aggregate of quartz, magnetite 
and amphiboles (magnesioriebeckite and cummingtonite). 
The latter comprise poikiloblastic textures, with grain sizes up 
to 5 mm. The main iron oxide was magnetite with a polygonal 
texture, which equated about 29 per cent by weight, which 
is similar to the magnetite content of other Joffre BIF in the 
province (McConchie, 1987). Up to half the reduced Fe was in 
DD00SGP002
0 40 80m
Silvergrass Dyke
(764+/-10Ma) Contact zone
Joffre J6
Joffre J1
Joffre J2
Joffre J3
Joffre J4
Joffre J5
Joffre Sill
100 RL
200 RL
300 RL
400 RL
7
,5
5
6
,0
0
0
 m
N
482,000 mE480,000 mE
480,000 mE
478,000 mE
7
,5
5
6
,0
0
0
 m
N
7
,5
5
8
,0
0
0
 m
N
7
,5
5
8
,0
0
0
 m
N
7
,5
6
0
,0
0
0
 m
N
DD00SGP002
0 0.5 1km
WESTERN
 AUSTRALIA
Perth
Map Area
437.3m
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
FIG 7 - Map and cross-section of the Silvergrass Peak magnetite project. Transparent magnetic image (ASVI, analytical signal
of the vertical integral) draped over aerial photography.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
H J DALSTRA
82
silicates and carbonates. A 15 cm interval of massive coarse 
magnetite occursat the contact with the dyke. Because the 
drilling failed to identify signifi cant high-grade magnetite-
rich BIF, the project was abandoned. Diamond drilling at the 
B26E project also failed to identify magnetite enriched BIF 
associated with the Dales Gorge member. 
Opportunistic mineralogical and metallurgical test work has 
been carried out on BIF of the Dales Gorge and Joffre members 
from a number of prospects in the Hamersley province. 
This work has shown that the mineralogy of unoxidised 
iron formation is dominated by silica and magnetite, with 
accessory haematite, siderite, riebeckite/crocidolite and 
ankerite/dolomite and traces of stilpnomelane and talc. 
Results from magnetic separation (Davis Tube) of BIF from 
these prospects are very similar to those reported at Balmoral 
and indicate that high quality concentrates suitable for blast 
furnace pellets with iron grades of around 69 per cent with 
low Al and P contents can be achieved from normal BIF at a 
grinding size of about 45 microns. Iron recoveries are relatively 
low (around 70 per cent, resulting in a total mass recovery of 
about 30 per cent) which is most likely due to the presence of 
iron in non-magnetic minerals such as siderite, iron silicate 
and haematite. Other problems associated with benefi ciation 
of Brockman BIF to a high-grade concentrate product are the 
great hardness of the rocks, and the widespread presence of 
fi brous minerals, particularly crocidolite.
Future exploration for magnetite-concentrates in the 
Hamersley province could focus on the enriched magnetite-
carbonate-silicate proto-ores predicted by both syngenetic 
and hypogene ore genesis models. However, work carried out 
to date suggests that these occurrences are small, and that 
in general the BIF’s show little chemical variation across the 
province. If future exploration fails to fi nd such deposits, focus 
should go to the mineralogically most attractive protores. 
These could include softer or coarser grained parts of the 
Brockman Iron Formation which would limit grinding costs, 
areas where iron occurs dominantly in the form of oxides 
increasing recovery, or areas where the absence of riebeckite 
in BIF limits the costs and hazards associated with fi brous 
minerals. 
ACKNOWLEDGEMENTS
Special thanks to members of the RTX iron ore exploration 
team and the former Hamersley Iron Resources Task Force 
for stimulating discussions which were essential to form the 
ideas in this paper. Rio Tinto Exploration Pty Ltd, and Rio 
Tinto Iron Ore are thanked for permission to publish this 
paper.
REFERENCES
Clarke, N, Kepert, D, Simpson, C and Edwards, D, 2009. Discovery 
of the Solomon iron deposits, in Proceedings Iron Ore 2009, 
pp 51-57 (The Australasian Institute of Mining and Metallurgy: 
Melbourne).
Dalstra, H J, Gill, T, Faragher, A, Scott, B and Kakebeeke, V, 
2010. Channel iron deposits, a major new district around the 
Caliwingina Creek, central Hamersley Ranges, Western Australia, 
Transactions of the Institutions of Mining and Metallurgy , 
Applied Earth Science, 119(1):B12-B20.
Dowson, N, Connely, D and Yan, D, 2009. Trends in magnetite 
ore processing and test work, in Proceedings Iron Ore 2009, 
pp 231-241 (The Australasian Institute of Mining and Metallurgy: 
Melbourne).
Ewers, W E and Morris, R C, 1980. Chemical and mineralogical 
data from the uppermost section of the upper BIF member of the 
Marra Mamba Iron Formation, CSIRO report no FP23, Perth.
Ewers, W E and Morris, R C, 1981. Studies of the Dales Gorge 
Member of the Brockman Iron Formation, Western Australia, 
Economic Geology, 76:1929-1953.
Flis, M, Butt, A L and Hawke, P J, 1998. Mapping the range front with 
gravity, are the corrections up to it?, Exploration Geophysics , 
29:378-383.
Harms, J E and Morgan, B D, 1964. Pisolitic limonite deposits in 
northwest Australia, in Proceedings Australasian Institute of 
Mining and Metallurgy, 212:91-124.
Killick, M F, Churchward, H M and Anand, R R, 2008. Regolith 
terrain analysis for iron ore exploration in the Hamersley 
province, Western Australia, fi nal report; CRC LEME open fi le 
report no 214, Bentley, Western Australia, 94 p.
Lascelles, D F, 2006a. The Mount Gibson banded iron formation 
hosted magnetite deposit, two distinct processes for the origin of 
high-grade ore, Economic Geology, 101:651-666.
Lascelles, D F, 2006b. The genesis of the Hope Downs Iron Ore 
Deposit, Hamersley Province, Western Australia, Economic 
Geology, 101:1359-1376.
Li, Q, Tong, Z, Wang, X and Gao, M, 2009. A new magnetic drum 
separator for superfi ne magnetite, in Proceedings Iron Ore 2009, 
pp 255-258 (The Australasian Institute of Mining and Metallurgy: 
Melbourne).
Lobato, L M, Figueiredo e Silva, R C, Hagemann, S, Thorne, W 
and Zucchetti, M, 2008. Hypogene alteration associated with 
high-grade banded iron formation related iron ore, Reviews in 
Economic Geology , pp 107-128 (Society of Economic Geologists 
Inc: Littleton).
MacLeod, W N, 1966. The geology and iron deposits of the 
Hamersley Range area, Western Australia: Geological Survey of 
Western Australia, Bulletin 117, 1:170.
McConchie, D M, 1987. The geology and geochemistry of the Joffre 
and Whaleback Shale Members of the Brockman Iron Formation, 
Western Australia, Precambrian Iron Formations (eds: P W 
Appel and G L LaBerge), pp 541-597 (Theophrastus Publications: 
Athens).
Morris, R C, 1985. Genesis of iron ore in banded iron-formation by 
supergene and supergene-metamorphic processes: A conceptual 
model, Handbook of Strata-bound and Stratiform Ore Deposits 
(ed: K H Wolff), pp 73-235 (Elsevier: Amsterdam).
Morris, R C, 1997. Iron ore: Is there ‘new’ life for old models, in 
Advances in Understanding the Hamersley Province , short 
course, pp 1-9 (The University of Western Australia: Nedlands, 
Western Australia).
Morris, R C and Ramanaidou, E R, 2007. Genesis of the channel iron 
deposits (CID) of the Pilbara of Western Australia, Australian 
Journal of Earth Sciences, 54:735-759.
Povey, B C, 2009. The use of high pressure grinding rolls for 
crushing magnetite, in Proceedings Iron Ore 2009 , pp 301-308 
(The Australasian Institute of Mining and Metallurgy: Melbourne).
 Taylor, D, Dalstra, H J, Harding, A E, Broadbent, G and Barley, M E, 
2001. Genesis of high-grade hematite orebodies of the Hamersley 
Province, Western Australia, Economic Geology, 96:837-873.
Thorne, W S, Hagemann, S G and Barley, M E, 2004. Petrographic 
and geochemical evidence for hydrothermal evolution of the 
North Deposit, Mt Tom Price, Western Australia, Mineralium 
Deposita, 39:766-783.
Wingate, M T D and Giddings, J W, 2000. Age and palaeomagnetism 
of the Mundine Well dyke swarm, Western Australia: Implications 
for an Australia – Laurentia connection at 755 Ma, Precambrian 
Research, 100:335-357.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 83
INTRODUCTION
In this pilot study, we apply (U-Th)/He dating in combination 
with other analytical methods to investigate channel iron 
deposits (CID) in Western Australia. Despite their economic 
importance, models of CID genesis are the subject of debate 
due to a paucity of geochronological data that could constrain 
the timing of their formation (MacLeod et al , 1963; Harms 
and Morgan, 1964; Campana et al , 1964; Butler, 1976; Hall 
and Kneeshaw, 1990; Morris, Kneeshaw and Ramanaidou, 
2007; Stone, 2005; Heim et al , 2006; Morris and 
Ramanaidou, 2007).
Age estimates based on palinological data could not provide 
a precise age determination and it is currently accepted that 
f lling of the existing paleochannels could have taken place 
from Late Eocene to Middle Miocene times (MacPhail and 
Stone, 2004; Balme in Morris, Ramanaidou and Horwitz, 
1993; Morris, 1994; Morris and Ramanaidou, 2007; Morris, 
Kneeshaw and Ramanaidou, 2007). The only radiometric 
age constraints were presented by Heim et al (2006) who 
applied combined (U-Th)/He and 4He/3He methods to date 
late stage authigenic goethite fromten samples collected from 
1. Senior Research Fellow, The University of Waikato, Department of Earth and Oceanic Science, Faculty of Science & Engineering, Private Bag 3105, Hamilton 3240, New Zealand. 
 Email: m.danisik@waikato.ac.nz
2. MAusIMM, Research Stream Leader - Geometallurgy of Carbon Steel Materials, Minerals Down Under Flagship, Commodity Leader for Iron Ore, CSIRO Earth Science and Resource Engineering, 
 ARRC PO Box 1130, Bentley WA 6102. Email: Erick.Ramanaidou@csiro.au
3. Principal Research Scientist, CSIRO Earth Science and Resource Engineering, ARRC PO Box 1130, Bentley WA 6102. Email: Noreen.Evans@csiro.au
4. Technical Offi cer, CSIRO Earth Science and Resource Engineering, ARRC PO Box 1130, Bentley WA 6102. Email: Brad.Mcdonald@csiro.au
5. Technical Offi cer, CSIRO Earth Science and Resource Engineering, ARRC PO Box 1130, Bentley WA 6102. Email: Celia.Mayers@csiro.au
6. Research Professor, Director, John de Laeter Centre for Isotope Research, John de Laeter Centre for Isotope Research, Department of Applied Physics, Curtin University, GPO Box U1987, Perth WA 6845. 
 Email: B.McInnes@curtin.edu.au
(U-Th)/He Geochronology of Channel 
Iron Deposits, Robe River, Hamersley 
Province, Australia – Implications for 
Ore Genesis
M Danišík1, E R Ramanaidou2, N J Evans3, B J McDonald4, C Mayers5 
and B I A McInnes6
ABSTRACT
Two drill core samples of haematite/goethite from the Robe River (Western Australia) channel 
iron ore deposits (CID) were dated using (U-Th)/He methods in order to constrain the timing of 
iron oxide formation and thereby provide a temporal context for CID genesis. (U-Th)/He ages (He 
ages) from these samples range from the late Oligocene to Late Miocene and despite a high degree 
of scatter, they corroborate relationships expected from the internal ooidal stratigraphy: For 
individual ooids, the ages from haematitic core are older than or indistinguishable from the ages 
of the surrounding goethitic cortex. The goethitic cortices are, in turn, older than the ferruginised 
wood fragments recovered from the cementing goethitic matrix. 
The data suggest the following succession of ore formation phases: 
  Haematitic cores in ooids of both samples formed in the Early to Middle Miocene.
  Goethitic cortices of both samples formed in the late Middle to early Late Miocene. This 
appears to be the age of goethite cortex formation regardless of depth in the core, which does 
not support the top-down inf ll model, genetic model proposed by Heim et al (2006).
  Wood fragments form a prominent component of the matrix and were ferruginised during the 
Late Miocene. Thus the data suggest that the unique environmental conditions for the CID 
formation existed during the Miocene.
A methodological implication of this study is that the temperature utilised for He-extraction 
from iron oxides has a critical impact on the mobility of parent nuclides. The typical ~1100°C laser 
heating used for crystalline minerals like apatite or zircon induces loss of U and Th and results in 
erroneously old ages.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
M DANIŠÍK et al
84
a vertical section through the Yandi CID in the Hamersley 
Province (Western Australia). These authors found that the 
He ages were all Miocene and progressively decreased with 
depth, suggesting that precipitation of goethite cement was 
controlled by water table drawdown during the Neogene 
(Heim et al , 2006). However, despite having reported ages 
consistent with the generally accepted Neogene formation age 
of the CID, the model of Heim et al (2006) was challenged by 
Morris, Kneeshaw and Ramanaidou (2007), who questioned 
the sampling strategy and statistical treatment of the data (for 
details see Morris, Kneeshaw and Ramanaidou, 2007 and 
reply by Vasconcelos et al, 2007).
Here we aim to derive new age constraints on the formation 
of CID in the Hamersley Province and re-evaluate the existing 
genetic models in the light of the new data.
SAMPLES AND METHODS
Two CID samples (G-82 and G-328) were collected from 
a diamond drill core (J1136 generously made available to 
CSIRO by Robe River Mining) located at Mesa J in the Robe 
River Valley from vertical depths of 8.2 and 32.8 m below the 
present surface, respectively (Figure 1). 
Polished thin sections of the samples of the samples were 
prepared for vital petrological work underpinning the 
dating analysis. Textural and in situ chemical analyses were 
respectively completed with optical and scanning electron 
microscopes equipped with an EDS system.
For (U-Th)/He dating, the samples were coarsely crushed 
to isolate intact ooids, haematitic core and goethite cortex 
shards, and ferruginised wood fragments from the matrix; 
these were then handpicked under a binocular microscope and 
ultrasonically cleaned. The samples were then analysed for 
4He, 238U and 232Th using isotope-dilution mass spectrometry 
(quadrupole and ICP-MS, respectively) at the John de Laeter 
Centre for Isotope Research in Perth (Australia) following the 
procedures described in Danišík et al (submitted). The total 
analytical uncertainty (TAU) was calculated as a square root 
of sum of squares of weighted uncertainties on U, Th, and He 
measurements. TAU was typically less than ~6 per cent (1 
sigma) and was used to calculate the uncertainty of He ages. 
Given the large size and polycrystalline character of dated 
samples, (U-Th)/He ages were not corrected for alpha ejection 
(Farley, Wolf and Silver, 1996). The ages were not corrected 
for diffusive loss (Shuster et al , 2005) that might result to a 
slight underestimation of true age.
RESULTS
Sample characterisation: microscopy 
The petrological study of the two samples revealed textures 
typical of a granular CID as de f ned by Ramanaidou, Morris 
and Horwitz (2003). The samples depict an ooidal-dominated 
texture with haematitic nuclei surrounded by a goethitic 
cortex and cemented by a goethitic matrix containing wood 
fragments (Figure 1). Under re f ected light, the ooids (0.25 
- 2 mm) haematitic nuclei can be either simple or complex; 
cracks are widespread and the internal texture shows compact 
white or more porous haematite, some nuclei contain 
haematitised wood fragments (Figure 2). The combination of 
ref ected and transmitted light (Figure 3) highlights the three 
major components:
1. dark red nuclei in transmitted light,
2. yellow cortices, and 
3. matrix goethite appearing grey in ref ected light.
FIG 1 - (A) Ooidal texture of a sample from the investigated J1136 drill core in Mesa J where three typical components are visible: haematitic core (1), goethitic cortex 
(2) and matrix (3); (B) ferruginised wood in the matrix; (C) SEM image of the wood with a typical porous structure; (D) SEM results with major element contents.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
(U-TH)/HE GEOCHRONOLOGY OF CHANNEL IRON DEPOSITS, ROBE RIVER, HAMERSLEY PROVINCE, AUSTRALIA
85
The matrix contains goethitised wood fragments 
(Figures 1 to 3).
(U-Th)/He results
Our preliminary (U-Th)/He data suggest the following four 
relationships:
1. The He ages of the cores are older than those of the 
corresponding cortices, which is in agreement with their 
internal stratigraphy.
2. The wood fragment from the matrix of the deeper sample 
(G-328) yielded Late Miocene ages, clearly younger 
than He ages of the cores and cortices from that sample. 
The Late Miocene ages of wood from the matrix of the 
shallower core are also consistently younger than the 
cores and cortices from this depth. This age relationship 
corroborates internal stratigraphic relationships, where 
the matrix should be the last phase of the CID to form.
3. There is a difference between the samples in terms of the 
He ages of ooid cores. Core He ages of the shallower sample 
G-82 are Middle Miocene, younger than Early Miocene He 
ages yielded by the deeper sample G-328.
4. He ages of cortices from both samples are indistinguishable,clustering at Middle/Late Miocene boundary.
INTERPRETATION AND PRELIMINARY 
CONCLUSIONS
Assuming the He ages represent the time of mineral formation, 
our He dating results suggest the following succession of ore 
forming processes: 
  formation of ooid cores of the deeper sample (G-328) in 
the Early Miocene,
  formation of ooid cores of the shallower sample (G-82) in 
the Middle Miocene,
FIG 2 - Microphotograph in refl ected light of sample G - 82 portraying the ooidal texture, the haematitic nuclei with some wood fragments (W - Hm), the goethite 
cortices (Co - Go) and matrix (Ma - Go) with goethitised wood fragment (W - Go). 
FIG 3 - Microphotograph in refl ected light and transmitted light of a selected 
area of sample G – 82 showing the white and reddish haematite nuclei in 
refl ected light and the dark reddish colour in transmitted light. The goethitic 
cortex and matrix are also highlighted. 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
M DANIŠÍK et al
86
  formation of goethitic cortex that coated already existing 
ooids of both samples at Middle/Late Miocene boundary, 
and
  ferruginisation of the wood fragments in the Late Miocene.
The ages of the cores imply that the dated ooids formed 
in the Early to Middle Miocene. The inverse age-depth 
relationship between the samples may suggest that the ooids 
represent components derived either from two sources of 
different age or from the same regolith pro f le which became 
inverted during the erosion, transport and re-sedimentation. 
The deeper (older He ages) and shallower (younger He ages) 
ooids thus represent the older, near-surface and younger, 
deeper regolith levels, respectively.
The formation of the goethitic cortex in the late Middle to 
early Late Miocene, in contrast, appears to be a fast process 
lasting maximally a few million years as inferred from 
indistinguishable He ages. Despite the fact that the timing 
seems to be well constrained by the He data, the location of 
goethitic cortices formation is somewhat arguable. Whereas 
Morris and Ramanaidou (2007) favoured formation of 
the goethitic cortices during the colluvium stage, before 
accumulation of the ooids in the river channels, Heim et al 
(2006) proposed that the goethitic cortex formed after the 
channel aggradation had cemented the cores. Heim et al 
(2006) also found a downward younging trend in He ages 
measured over a ~30 m deep discontinuous pro f le. These 
authors argued that the process of iron cementation occurred 
at the groundwater-atmosphere interface and was driven 
by water table drawdown during the Neogene. Although we 
cannot discern from our data which model is correct, we can 
say that our data do not support the interpretation of Heim et 
al (2006). The two cortex samples, with a vertical separation 
of 23 metres, collected in the same borehole do not show any 
variation in He age. Therefore, the top-down in f ll model 
driven by water table drawdown cannot apply to the locality 
investigated in this study.
The last process recorded by He data is the ferruginisation 
of wood fragments, in which the organic wood tissues were 
replaced by mobilised ferrous iron. Although the mechanism 
of this process is not well understood (Morris, Ramanaidou 
and Horwitz, 1993; Morris and Ramanaidou, 2007; Morris, 
Kneeshaw and Ramanaidou, 2007), it most likely occurred in 
the Late Miocene to Early Pliocene, provided the ages record 
the time of goethite/haematite replacement. This result clearly 
shows that river aggradation must have been completed in the 
Late Miocene. 
Our data f t well the Miocene climatic optimum for the CID 
formation proposed by Morris and Ramanaidou (2007).
ACKNOWLEDGEMENTS
We thank Michael Verrall for help with the SEM, C Scadding 
and A Thomas for assistance with ICP-MS, and Richard Morris 
for constructive suggestions on the CID genesis. This project 
was funded by the Minerals Down Under Flagship, CSIRO. 
MD received f nancial support from the WA Geothermal 
Centre of Excellence.
REFERENCES
Butler, R J T, 1976. Geology of the Tertiary ironstones in the middle 
and upper Robe River area, Pilbara Region, Western Australia, 
MSc thesis (unpublished), University of Western Australia, Perth.
Campana, B, Hughes, F E, Burns, W G, Whitcher, I G and 
Muceniekas, E, 1964. Discovery of the Hamersley iron deposits 
(Duck Creek-Mt. Pyrton-Mt. Turner Area), The AusIMM 
Proceedings, 210:1-30.
Danišík, M, Ramanaidou, E R, Evans, N J, McDonald, B J, Mayers, 
C and McInnes, B I A, submitted. (U-Th)/He geochronology of 
channel iron deposits, Hamersley Province, Western Australia: 
Implications for ore genesis and landscape evolution, Submitted 
to Chemical Geology. 
Farley, K A, Wolf, R A and Silver, L T, 1996. The effects of long alpha-
stopping distances on (U-Th)/He ages, Geochim Cosmochim 
Acta, 60:4223-4229.
Hall, G C and Kneeshaw, M, 1990. Yandicoogina-Marillana pisolitic 
iron deposits, in Geology of the Mineral Deposits of Australia 
and Papua New Guinea (ed: F E Hughes), pp 1581-1586 (The 
Australasian Institute of Mining and Metallurgy: Melbourne).
Harms, J E and Morgan, B D, 1964. Pisolitic limonite deposits in 
northwest Australia, The AusIMM Proceedings, 212:91-124.
Heim, J A, Vasconcelos, P M, Shuster, D L, Farley, K A and Broadbent, 
G C, 2006. Dating palaeochannel iron ore by (U-Th)/He analysis 
of supergene goethite, Hamersley Province, Australia, Geology, 
34/3: 173-176.
MacLeod, W N, de la Hunty, L E, Jones, W R and Halligan, R, 1963. 
A preliminary report on the Hamersley Iron Province, North-
West Division, Geological Survey of Western Australia, Annual 
Report for 1962, pp 44-54.
MacPhail, M K and Stone, M S, 2004. Age and palaeoenvironmental 
constraints on the genesis of the Yandi channel iron deposits, 
Marillana Formation, Pilbara, northwestern Australia, Australian 
Journal of Earth Sciences, 51:497-520.
Morris, R C, 1994. AMIRA Project P75G – Detrital iron deposits 
of the Hamersley Province, CSIRO Exploration and Mining 
Restricted Report 76R.
Morris, R C, Kneeshaw, M and Ramanaidou, E R, 2007. Dating 
paleochannel iron ore by (U-Th)/He analysis of supergene 
goethite, Hamersley province, Australia: Comment, Geological 
Society of America, e118.
Morris, R C and Ramanaidou, E R 2007. Genesis of the channel 
iron deposits (CID) of the Pilbara region, Western Australia, 
Australian Journal of Earth Sciences, 54/5:733-756.
Morris, R C, Ramanaidou, E R and Horwitz, R C, 1993. Channel iron 
deposits of the Hamersley Province, CSIRO Australia, Exploration 
and Mining Report 399R.
Ramanaidou, E R, Morris, R C and Horwitz, R C, 2003. Channel 
iron deposits of the Hamersley Province, Western Australia, 
Australian Journal of Earth Sciences, 50:669–690.
Shuster, D L, Vasconcelos, P M, Heim, J A and Farley, K A, 2005. 
Weathering geochronology by (U-Th)/He dating of goethite, 
Geochim Cosmochim Acta, 69:659-673.
Stone, M S, 2005. Depositional history and mineralisation of 
Tertiary iron deposits at Yandi, Eastern Pilbara, Australia, PhD 
thesis (unpublished), University of Western Australia, Perth.
Vasconcelos, P M, Heim, J A, Farley, K A, Shuster, D L and 
Broadbent, G, 2007. Dating paleochannel iron ore by (U-Th)/He 
analysis of supergene goethite, Hamersley province, Australia: 
Reply, Geological Society of America, e118.
87IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
INTRODUCTION
Large, high-grade (>55 wt per cent Fe) iron ore deposits 
in the Hamersley Province and Yilgarn Craton of Western 
Australia are commonly the combined product of several 
distinct types of Fe ore. For example, the Mt. Tom Price 
deposit in the Hamersley Province hosts goethite-haematite 
ore overlying magnetite ± haematite ore (Taylor et al , 2001), 
whereas Fe deposits in the Koolyanobbing Fe camp, Yilgarn 
Craton, contain a combination of goethite-, haematite-, and 
magnetite-rich ores (Angerer and Hagemann, 2010). Each 
ore type has distinct physical and chemical characteristics, 
suchas ore and gangue mineralogy, grain size, Fe grade, 
and contaminant elemental suites, which together causes 
substantial heterogeneity within a single deposit and 
inf uences benef ciation processes and exploration strategies. 
For these reasons, it is vital to understand the key relationships 
between different Fe ore types.
The Weld Range greenstone belt, in the Murchison Domain 
of the Yilgarn Craton, hosts two Archean, high-grade Fe 
deposits, Madoonga (68 Mt resource at 57.7 wt per cent Fe) 
and Beebyn (62 Mt resource at 59.6 wt per cent Fe, ASX 
announcement 2008). These deposits contain up to f ve 
different types of Fe ore that are heterogeneously distributed 
in the deposits, but are important contributors to the overall 
resource estimate for the deposits. This paper summarises the 
geological setting of the Weld Range greenstone belt before 
documenting the location, properties, and relative timing of 
the f ve types of Fe ore. These results are then discussed in 
terms of their implications for exploration within the Weld 
Range greenstone belt. 
GEOLOGICAL SETTING OF THE WELD RANGE 
GREENSTONE BELT
The Weld Range greenstone belt is located in the Murchison 
Domain of the Youanmi Terrane, Yilgarn Craton, Western 
Australia. The oldest supracrustal rocks exposed in the 
1. Assistant Professor, Centre for Exploration Targeting, The University of Western Australia, Crawley WA 6009. Email: paul.duuring@uwa.edu.au
2. Professor, Centre for Exploration Targeting, The University of Western Australia, Crawley WA 6009. Email: steff en.hagemann@uwa.edu.au
Contrasting Styles of High-Grade Iron 
Mineralisation at Weld Range, Western 
Australia
P Duuring1 and S G Hagemann2
ABSTRACT
Late-Archean banded iron formation (BIF)-hosted deposits in the Yilgarn Craton of Western 
Australia are less well understood compared to the larger Paleoproterozoic, Superior-type BIF-
hosted deposits of the Hamersley Province. The Weld Range greenstone belt, in the Murchison 
Domain of the Yilgarn Craton, hosts two Archean, high-grade deposits, Madoonga (68 Mt resource 
at 57.7 wt per cent Fe) and Beebyn (62 Mt resource at 59.6 wt per cent Fe, ASX announcement 
2008). Five main types of high-grade (>55 wt per cent Fe) iron mineralisation at Weld Range each 
display characteristic grades, tonnage, and contaminant levels, which affect exploration strategies 
and benef ciation methods: 
1. ‘Residual’ ore formed as a result of two generations of hypogene alteration of BIF. The f rst 
alteration phase replaced silica-rich bands with siderite and/or Fe-rich dolomite; the second 
phase involved the removal of carbonate gangue minerals and the concentration of residual Fe 
oxide-rich bands in the BIF via volume reduction. The product is a high-grade, high-tonnage 
Fe orebody, with minor contaminants. 
2. Magnetite-bearing shear and fault zones cut BIF along lithological contacts. These zones of 
secondary magnetite formed as a result of the addition of hypogene magnetite via the circulation 
of Fe-rich, hypogene hydrothermal f uids through BIF. This type of mineralisation produces 
narrow (low-tonnage), high-grade magnetite-rich ore zones with high levels of contaminants 
(eg SiO2, Al2O3) due to the diff culty in mining these zones. 
3. Specular haematite ± quartz-bearing shear and fault zones cut residual orebodies and magnetite-
bearing structures. They share similar characteristics to magnetite-bearing structures. 
4. Goethite-haematite supergene ore zones are controlled by brittle faults that cut BIF. These 
faults promoted the f ow of supergene f uids through BIF and resulted in large ore zones with 
variable Fe grades and contaminants. 
5. Detrital deposits comprise transported BIF fragments and are intensely goethite-haematite 
supergene-altered. These orebodies are locally extensive, with moderate Fe grades. 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
P DUURING AND S G HAGEMANN
88
Weld Range greenstone belt include steeply S-dipping, 
S-facing komatiite, komatiitic basalt, and tholeiitic basalt 
that together represent part of the 2800 to 2730 Ma Polelle 
Group (stratigraphic associations def ned by Van Kranendonk 
and Ivanic, 2009). These komatiite f ows are overlain by 
andesitic to rhyolitic volcanic and volcaniclastic rocks of 
the Greensleeves Formation, which are in turn overlain 
by BIF, tuffaceous siltstones, and felsic rocks of the Wilgie 
Mia Formation that are intruded by dolerite to gabbro sills. 
The Wilgie Mia Formation coincides with a 3 - 5 km-wide, 
~70 km-long, series of parallel ridges that trend mainly ENE 
but curves to a more northerly trend in eastern areas of the 
district. The Madoonga and Beebyn Fe deposits are hosted 
by discrete BIF sequences within the formation (Figure 1). 
Farther south, younger felsic volcanic and volcaniclastic 
rocks are overlain by pelite and psammite of the Ryansville 
Formation, which are intruded by gabbronorite, gabbro, and 
dolerite of the Yalgowra Suite (Ivanic, 2009). 
Supracrustal rocks in the Weld Range greenstone belt are 
metamorphosed to upper-greenschist to lower-amphibolite 
facies (300 ± 50°C) at pressures of <2 - 3 kbars (Gole, 1980). 
District-scale structures in the belt include ENE-trending, 
isoclinal, f rst-generation (F1) folds that have an axial planar 
foliation oriented subparallel to bedding in supracrustal rocks 
(Spaggiari, 2006) (Figure 1). The folding of bedding contacts 
and a bedding-parallel foliation (axial planar to F 1 folds) in 
rocks of the Ryansville Formation (Ivanic, 2009) de f ne a 
moderately SW-plunging F 2 syncline located to the south 
of the main series of ENE-trending ridges (Figure 1). The 
NNE-trending Carbar fault (Spaggiari, 2006) truncates the 
northwestern limb of the regional F2 syncline.
METHODOLOGY
Detailed lithological, structural, and alteration outcrop 
mapping were performed at a 1:2000 scale over the individual 
~8 km-strike lengths of the Beebyn and Madoonga deposits. 
At each deposit locality, outcrop is contained within 
<150 m-wide, <100 m-high, ENE-trending ridges. Core 
from 12 diamond drill holes that cover the respective strike 
lengths of each deposit was studied with emphasis placed on 
the relationship between unweathered rock types, structures, 
hydrothermal alteration, and iron mineralisation. The holes 
were mostly drilled from surface to the NNW at moderate to 
steep angles. Examples of least-altered/weathered rocks were 
collected in conjunction with their more altered and weathered 
variants for the purpose of thin-section, carbonate-staining, 
geochemical, and spectral (analytical spectral device) studies.
GEOLOGICAL OVERVIEW OF THE BEEBYN AND 
MADOONGA DEPOSITS
Beebyn deposit
The Beebyn deposit stratigraphy trends ENE, is steeply 
SSE-dipping, and includes BIF, basalt, dolerite, and gabbro, 
with minor chloritic shale or siltstone interbedded with the 
BIF. Three main, <80 thick, BIF units (informally labelled 
the ‘North’, ‘Central’, and ‘South’ BIF) in the deposit are 
 
FIG 1 - Solid geology map and cross-section for the Weld Range study area. Volcanic, volcano-sedimentary, and sedimentary rocks defi ne a regional, moderately 
SW-plunging syncline, with the Weld Range area defi ning the northeast limb (modifi ed after Van Kranendonk and Ivanic, 2009). Note the location of the Beebyn and 
Madoonga Fe deposits.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
CONTRASTING STYLES OF HIGH-GRADE IRON MINERALISATION AT WELD RANGE, WESTERN AUSTRALIA
89
bounded by ma f c volcanic rocks (Figure 2a). All major 
lithological contacts between BIF and maf c volcanic rocks are 
deformed. Only narrow (<3 m-thick) basalt or dolerite sills 
in the North BIF preserve their intrusive contacts. Chloritic 
shale or siltstones are minor constituents of the North and 
Central BIF. Stratigraphic younging directions are not clearly 
preserved in any rock types at Beebyn. The North BIF is the 
thickest BIF and hosts the widest zones of high-grade iron 
mineralisation.These zones are most commonly located 
along the stratigraphic footwall contact of the North BIF. The 
depth of weathering varies over the strike length of the Beebyn 
deposit, averaging about 120 m below the present surface, 
but extending to considerably greater depths (~200 m) along 
lithological contacts or within steeply-dipping fault zones. 
Least-altered BIF mostly comprises equal proportions of 
0.5 - 3 cm-thick, alternating bands of quartz and Fe oxide 
minerals (Figure 3a).
Five main deformation events are preserved at Beebyn. The 
f rst deformation event (D 1) involved E-W shortening and 
resulted in the formation of isoclinal, recumbent F 1 folds that 
presently have axial planes oriented subparallel to the ENE-
trending, multiply-folded stratigraphy. F 1 fold axes observed 
within the BIF have variable plunge directions, however, they 
mainly plunge moderate to steeply to the NE. Reverse shear 
zones or faults occur along folded limbs and locally displace 
fold hinges. The second deformation event (D 2) coincided 
with N-S shortening and caused the refolding of F 1 folds. 
These F 2 folds are tight, upright, ENE-trending, and mainly 
plunge shallow to moderately to the WSW. Rare centimetre-
scale examples of refolded F 1 folds are exposed in the North 
BIF. Asymmetric, Z-shaped parasitic F 2 folds (in plan) are 
commonly def ned by mesobands in BIF and correspond with 
asymmetric, north-limb-down F2 folds in cross-section. These 
parasitic fold relationships are consistent with the existence of 
a district-scale, synclinal fold hinge to the south of the broadly 
S-facing Beebyn deposit stratigraphy (perhaps corresponding 
with the synclinal fold hinge within the Ryansville Formation 
mapped to the south of the Weld Range by Ivanic, 2009). 
A third deformation event (D 3) involving E-W shortening 
resulted in re-folding of the stratigraphy and the formation 
of F 3 folds in BIF. These folds are open to tight, trend N-S, 
and plunge steeply to the N or S. The F 3 folds contribute to 
the large variation in the orientation of F 1 and F 2 fold hinges 
at Beebyn. Subsequent deformation at Beebyn (D 4 and D 5) 
corresponds with the transition from ductile to brittle styles 
of deformation. North-south shortening during D 4 resulted 
in NNW- to NNE-trending, subvertical, brittle faults that 
cut the folded Beebyn stratigraphy. These faults commonly 
host extensional quartz veins and are spatially associated 
with zones of intense goethite-weathering of BIF horizons. 
Displacement indicators are rare, but where they exist they 
indicate mostly dextral displacements of up to 20 m. Based 
on the interpretation of aeromagnetic data, it is possible that 
dextral fault displacements may be even greater (<400 m). 
A resumption of E-W shortening during D 5 resulted in the 
centimetre-scale, dextral displacement of N-trending faults 
and extensional quartz veins by ENE-trending, bedding-
parallel, brittle faults.
Madoonga deposit
The Madoonga deposit stratigraphy strikes ENE and 
dips steeply SE. From oldest to youngest, the SE-facing 
sequence includes rhyolite, dolerite, and gabbro overlain 
by a <60 m-thick BIF (hereafter referred to as the ‘North 
BIF’), a volcaniclastic or (volcanogenic)-sedimentary rock, a 
<150 m-thick southern BIF (ie ‘South BIF’), pyritic mudstone, 
and dolerite/gabbro. These steeply dipping rocks are 
unconformably overlain by a f at-lying, poorly-sorted, BIF 
clast-rich breccia along the southern margin of the main 
ridge at Madoonga (Figure 2b). All major lithological contacts 
between BIF and surrounding rocks are deformed. The limit 
of supergene alteration extends to vertical depths of up to
300 m below surface in some areas of the deposit. 
Deformation events interpreted for the Madoonga deposit 
are comparable to those described for the Beebyn deposit. 
Isoclinal, f rst-generation (F1) folds are rare at Madoonga, 
whereas magnetite-bearing shear zones are commonly located 
along the margins of the North and South BIF. Locally, the 
shear zones anastomose and transgress banding in the BIF 
as well as lithological contacts at an acute angle, clockwise 
to contacts. Magnetite-bearing shear zones, bands, and 
lithological contacts are folded in response to N-S shortening 
during the regional D2 event. These F2 folds are tight, upright 
and mainly plunge moderately (40° to 60°) to the W. The folds 
have asymmetric, S- or Z-shaped geometries in plan view, 
and south- or north-side-down geometries in cross-section. 
Specular haematite ± quartz veins cut banding in BIF at acute 
angles and are folded by a third generation of folds (F3). East-
west shortening resulted in N-trending, open, low-amplitude, 
F3 folds that warp the previously folded tectono-stratigraphy. 
 
FIG 2 - Representative cross-sections for the (A) Beebyn; and (B) Madoonga deposits; showing the relationships between rock types and ore zones.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
P DUURING AND S G HAGEMANN
90
The F3 folds are cut by NNW- to NNE-trending, subvertical, 
conjugate faults and quartz veins (D 4). East-trending, brittle 
faults occur along major stratigraphic contacts and result 
in the disruption of folds and the reactivation of existing, 
bedding-parallel structures (D5).
CONTRASTING STYLES OF HIGH-GRADE IRON 
ORE AT WELD RANGE
Hypogene ‘residual’ magnetite-martite ore
The type locality for this style of mineralisation is the Beebyn 
deposit, where magnetite-martite-rich ore occurs beneath the 
base of supergene weathering to vertical depths of at least 
250 m in the North and Central BIF units (Figure 2a). The 
ore zones are preferentially located along the sheared footwall 
contact of the North BIF, but may also occur at various 
stratigraphic positions within the BIF. In the centre of the 
Beebyn deposit, high-grade magnetite-martite ore zones 
are <50 m-thick and extend continuously for about 500 m 
along strike. In more distal areas, the ore zones are narrower 
(<15 m-thick), shorter (<50 m-long), and are sporadically 
distributed up to one kilometre from the main magnetite-
martite orebody. The ore zones are genetically associated 
with banding-parallel shear zones that are folded by F 2 folds. 
In hand specimen, high-grade magnetite-martite ore is black, 
thinly-banded (<3 mm-thick), and comprises >70 volume 
percent Fe oxide bands (Figure 3b). Locally, the rock displays 
brecciation of Fe oxide-rich bands and the centimetre-scale 
crenulation of bands. In thin-section, primary Fe oxide bands 
comprising subhedral magnetite are rimmed by secondary, 
euhedral magnetite. Minor (<30 vol per cent) gangue minerals 
include f ne-grained primary quartz and coarse-grained, 
hypogene siderite and ferroan dolomite. Relative to least-
altered BIF, magnetite-martite ore is enriched in Fe and P, and 
depleted in SiO2. The product is a high-grade (>55 per cent Fe), 
high-tonnage orebody, with minor contaminants. An outer 
alteration halo of hypogene siderite and ferroan dolomite 
surrounds the magnetite-martite ore.
A two-stage, hypogene hydrothermal alteration process is 
interpreted for the genesis of the ‘residual’ magnetite-martite 
ore type. The f rst stage of hydrothermal alteration involved 
the replacement of existing silica-rich bands in least-altered 
BIF by hypogene siderite and ferroan dolomite, with minor 
addition of hypogene magnetite. The product is a BIF that 
has retained its primary texture and iron content, but has 
experienced substitution of quartz with carbonate gangue 
 
 FIG 3 - Photoplate displaying examples of the least-altered banded iron formation and main ore types at Weld Range; (A) least-altered banded iron formation, 
(B) hypogene ‘residual’ magnetite-martite ore with minor preservation of earlier ferroan dolomite alteration, (C) hypogene magnetite-rich vein with fragments of 
magnetite-altered, banded iron formation wall rock, (D) crystalline specular haematite intergrown with quartz in a vein cutting banded iron formation,(E) supergene 
goethite-haematite alteration replacing the primary and hypogene minerals and textures in banded iron formation; and (F) detrital ore containing banded iron 
formation fragments cemented by goethite.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
CONTRASTING STYLES OF HIGH-GRADE IRON MINERALISATION AT WELD RANGE, WESTERN AUSTRALIA
91
minerals (Figure 3b). This rock has lower silica abundance 
than least-altered BIF, but has about the same concentration 
of Fe. This Stage 1 hypogene alteration style is preserved in 
the outer halo surrounding high-grade magnetite-martite 
ore at Beebyn. Although the Fe content of this carbonate-
altered BIF has not changed signi f cantly, the replacement of 
quartz bands by carbonate minerals aids benef ciation. Hence, 
this type of altered BIF is considered to be low-grade (40 to 
55 wt per cent Fe) ore at Weld Range. The reactivation of 
shear zones along the footwall contact of the North BIF at 
Beebyn encouraged the repeated focusing of hydrothermal 
f uids along the contact, resulting in the (partial) replacement 
of Stage 1 alteration minerals by Stage 2 alteration minerals 
to form magnetite-martite ore. The timing of this ore-forming 
event is pre-D2.
Fault or shear zone-hosted, hypogene 
magnetite-martite ore
Magnetite-martite veins associated with faults or shear zones 
cut BIF at the Madoonga and Beebyn deposits. At Madoonga, 
<3 m-thick, brittle-ductile shear zones are located along major 
lithological contacts, including footwall and hanging wall 
margins of the North and South BIF units. These magnetite-
martite-rich shear zones are oriented subparallel to banding, 
apart from in dilational jogs, where subsidiary shear zones are 
oriented up to 30° clockwise from the main ENE-trending, 
en echelon shear zones. Within the jogs, magnetite-martite 
veins de f ne a network amongst magnetite-martite-altered 
BIF, which is now largely oxidised to haematite (Figure 3c). 
Magnetite-martite veins, faults, and shear zones are locally 
folded by F 2 folds and thus predate the D 2 event. At the 
Beebyn deposit, <1 m-thick, mylonite zones are located along 
the footwall contact of the North BIF. The mylonites contain 
hypogene magnetite enveloped by zones with f attened and 
rotated clasts of BIF wall rock. No cross-cutting relationships 
were observed between magnetite-martite veins and the 
hypogene ‘residual’ magnetite-martite ore at Beebyn, although 
both are folded by F2 folds.
Magnetite-martite ore associated with veins, faults, and 
shear zones indicate the addition of Fe to least-altered BIF, 
without signi f cant removal of silica-rich bands from BIF. 
Consequently, high-grade ore zones are narrow and have the 
potential to be contaminated by SiO 2 from nearby BIF-wall 
rock. The thickest ore zones coincide with dilational jogs and 
splays to the ENE-trending shear and fault zones. 
Hypogene specular haematite ore
Specular haematite ± quartz veins cut least-altered BIF, 
hypogene ‘residual’ magnetite-martite ore zones, and fault 
or shear zone-hosted hypogene magnetite-martite ore zones 
at Madoonga and Beebyn. The thickest specular haematite 
± quartz veins are at Madoonga, where they are 10 - 20 m 
thick and coincide with dilational jog zones to ENE-trending, 
fault-hosted veins. The proportion of haematite relative to 
quartz varies in the veins. At one locality at Madoonga, the 
proportion of quartz in veins increases with distance from 
a dominant specular haematite vein. In hand specimen 
examples, randomly oriented specular haematite crystals are 
<4 cm long, euhedral and intergrown with crystalline quartz 
(Figure 3d). At the Beebyn deposit, randomly-oriented, 
euhedral specular haematite crystals are concentrated in the 
hinges of asymmetric F 2 folds. These cigar-shaped lenses 
strike ENE and plunge moderately to the WSW, that is, in the 
same plunge direction as F 2 folds. The limbs of the folds and 
specular haematite veins are refolded by F 3 folds. Elsewhere 
at Beebyn, specular haematite and chlorite- f lled shear zones 
oriented at a shallow angel to banding cut hypogene ‘residual’ 
magnetite-martite ore. Structural relationships indicate that 
the specular haematite ore formation event took place syn- to 
post-F2 folding, but before F3 folding.
Hypogene specular haematite ore zones have similar 
characteristics to the magnetite-martite veins in that they 
formed by the addition of Fe to BIF. Specular haematite high-
grade zones are narrow, with variable SiO 2 contamination 
due to the presence of coeval hypogene quartz and the 
incorporation of silica-rich, BIF wall rock in the vein. Locally, 
ore zones are thicker where they coincide with dilational jogs 
associated with ENE-trending faults.
Supergene goethite-haematite ore
Goethite-haematite ore is best developed at the Madoonga 
deposit where it is 400 m long × 150 m wide, and extends 
to depths of about 300 m (Figure 2b). The ore is structurally 
controlled in that it coincides with the intersection between 
the sheared hanging wall contact of the South BIF and later, 
cross-cutting NNW- to NNE-trending, subvertical brittle 
faults. Within these highly deformed areas, least-altered BIF 
and earlier hypogene magnetite-martite veins are replaced by 
strong goethite-haematite alteration, which destroys existing 
bands in the BIF and results in a cavity-rich, mottled texture 
(Figure 3e). Compared with least-altered BIF, the goethite-
haematite-rich ore is enriched in Fe and P, with depletion in 
SiO2. Elsewhere at Weld Range, goethite-haematite-altered 
BIF shows similar genetic associations with NNW- to NNE-
trending, subvertical faults, but without the same magnitude 
of alteration. These subvertical faults cut F 3 folds and offset 
the ENE-trending strata with dextral displacements of up to 
200 m.
Goethite-haematite ore zones are interpreted to be the 
product of supergene alteration, whereby near-surface 
oxidised f uids circulated down and along NNW- to NNE-
trending, subvertical brittle faults during the D 4 event. The 
supergene f uids removed SiO2 from the BIF via leaching of 
silica-rich bands, resulting in the consolidation of Fe oxide-
rich bands and an upgrade in Fe. The intensity and magnitude 
of supergene enrichment of BIF is primarily controlled by 
f uid permeability through the BIF via the late faults. The 
signif cance of existing hypogene magnetite-martite veins as 
contributors to the size and quality of the goethite-haematite 
ore is probably minor due to their narrow widths.
Detrital accumulation of goethite-haematite 
ore
Accumulations of goethite-haematite-rich detrital sediments 
occur along the southern side of the main ENE-trending 
ridge at the Madoonga deposit (Figure 2b). These poorly-
consolidated concentrations contain large fragments of 
least-altered and hypogene-altered BIF that are cemented 
by supergene goethite and haematite (Figure 3f). The 
concentrations extend for about 3 km in strike, are <200 m 
wide, and <100 m deep. The detrital ore zones vary in grade 
and width considerably along strike and with depth. 
Detrital accumulations of goethite-haematite ore formed 
as a result of the erosion and transport of Fe-enriched BIF 
exposed along the ridge at Madoonga. These fragments were 
transported short distances due to gravity and f uvial processes, 
and were deposited in palaeotopographic depressions adjacent 
to the ridge. Once accumulated, the poorly-sorted fragments 
were then partly cemented by goethite and haematite during 
the circulation of supergene f uids.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 201192
P DUURING AND S G HAGEMANN
IMPLICATIONS FOR EXPLORATION
The most economically signif cant ore types in the Weld Range 
greenstone belt are hypogene ‘residual’ magnetite-martite ore 
(ie at the Beebyn deposit) and supergene goethite-haematite 
ore (Madoonga deposit). These two ore types are responsible 
for stand-alone orebodies that have low contamination 
from SiO 2, Al 2O3, or P. Thesmall Fe prospects in the belt 
generally host narrow, hypogene magnetite veins or specular 
haematite veins, whereas comparatively larger Fe prospects 
have coincident ore zones. For example, specular haematite 
ore zones are locally cut by N-trending faults near Madoonga, 
resulting in the supergene modif cation of hypogene ore zones.
Each ore type has a speci f c physical and chemical 
characteristic (eg Fe grade, contaminants, magnetism, 
density), relative timing with respect to regional deformation 
events, and controls for ore formation. These differences have 
direct implications for exploration, such as when interpreting 
geophysical anomalies in BIF (ie in the recognition of 
magnetic/dense magnetite-rich ore versus less-magnetic/
less dense supergene goethite-haematite ore), or during the 
identif cation of important structures in the belt (ie mylonite 
zones verses NNW- to NNE-trending brittle faults). The early 
identif cation of ore type from a newly identi f ed prospect 
helps to anticipate the likely Fe ore quality, orebody size, and 
controls on ore formation for the occurrence. The ranking 
of exploration targets based on their ore type is an effective 
strategy for reducing the number of potential targets in a 
district and for focusing exploration activities.
CONCLUSIONS
The Weld Range greenstone belt hosts f ve main types of 
high-grade (>55 wt per cent Fe) mineralisation. Each ore 
type displays key differences in physical and chemical 
characteristics, such as Fe grade, tonnage, and contaminant 
levels. In order of formation: 
1. Hypogene ‘residual’ magnetite-martite ore formed as 
a result of two generations of hypogene alteration of 
BIF. The f rst alteration phase replaced silica-rich bands 
with siderite and/or Fe-rich dolomite; the second phase 
involved the removal of carbonate gangue minerals and 
the concentration of residual Fe oxide-rich bands in the 
BIF via volume reduction. The product is a high-grade, 
high-tonnage Fe orebody, with minor contaminants of 
SiO2, Al2O3, and P.
2. Magnetite-martite veins and associated faults or shear 
zones cut BIF along lithological contacts. These zones of 
secondary magnetite formed as a result of the addition of 
hypogene magnetite via the circulation of Fe-rich, hypogene 
hydrothermal f uids through BIF. This type of 
mineralisation produces narrow (low-tonnage), high-
grade magnetite-rich ore zones with high levels of 
contaminants due to the diff culty in mining these zones.
3. Hypogene specular haematite±quartz veins cut residual 
orebodies and magnetite-bearing structures. They share 
similar characteristics to magnetite-bearing structures. 
4. Supergene goethite-haematite ore zones are controlled by 
brittle faults that cut BIF. These faults promoted the f ow 
of supergene f uids through BIF and resulted in large ore 
zones with variable Fe grades and contaminants.
5. Detrital deposits comprise transported BIF fragments and 
are intensely goethite-haematite supergene-altered. These 
orebodies are locally extensive, with moderate Fe grades. 
The contrasting properties of the f ve ore types have direct 
relevance to exploration strategies in the belt and impacts on 
benef ciation methods for the ore. The ranking of exploration 
targets based on their ore type is an effective strategy for 
reducing the number of potential targets in a district and for 
consolidating exploration activities.
ACKNOWLEDGEMENTS
This study is f nancially supported by Sinosteel Midwest 
Corporation. Weld Range geologists and f eld assistants are 
gratefully acknowledged for their generous f eld support.
REFERENCES
Angerer, T and Hagemann, S G, 2010. The BIF-hosted high-grade 
iron ore deposits in the Archean Koolyanobbing greenstone 
belt, Western Australia: Structural control on synorogenic- and 
weathering-related magnetite-, hematite-, and goethite-rich iron 
ore, Economic Geology, 105:917-945.
Gole, M J, 1980. Mineralogy and petrology of very-low-metamorphic 
grade Archaean banded iron-formations, Weld Range, Western 
Australia, American Mineralogist, 65:1-2.
Ivanic, T, 2009. Madoonga, WA Sheet 2444 - 1:100 000 Geological 
Series, Geological Survey of Western Australia.
Spaggiari, C V, 2006. Interpreted bedrock geology of the northern 
Murchison Domain, Youanmi Terrane, Yilgarn Craton, Western 
Australian Geological Survey, Record 2006/10.
Taylor, D, Dalstra, H J, Harding, A E, Broadbent, G C and Barley, 
M E, 2001. Genesis of high-grade hematite orebodies of the 
Hamersley Province, Western Australia, Economic Geology , 
96:837-873.
Van Kranendonk, M J and Ivanic, T J, 2009. A new lithostratigraphic 
scheme for the northeastern Murchison Domain, Yilgarn Craton, 
Geological Survey of Western Australia, Annual Review 2007-08.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 93
INTRODUCTION
The McPhee Creek iron deposit lies 30 km north of the 
town of Nullagine and 220 km southeast of Port Hedland 
in the northeastern Pilbara region of Western Australia 
(Figure 1). The deposit is located adjacent (5 km east) to the 
Marble Bar road, which links to the Great Northern Highway. 
The road from Port Hedland to the Nullagine turn-off is sealed 
while the remainder of the road to the project is unsealed 
(approximately 60 km). 
Giralia holds a number of exploration licences in the 
McPhee Creek area with the McPhee Creek deposits situated 
1. MAusIMM, General Manager – Business Development, Gascoyne Resources Ltd, PO Box 1449, West Perth WA 6872. Email: julian@gascoyneresources.com.au
2. MAusIMM, Former Director, Giralia Resources NL, PO Box 1665, West Perth WA 6872. Email: mjoyce@zenithminerals.com.au
3. Project Geologist, Atlas Iron Limited, PO Box 7071, Cloisters Square WA 6850. Email: Paul.Bonato@Atlasiron.com.au
4. Project Geologist, Zenith Minerals Limited, PO Box 1426, West Perth WA 6872. Email: alan@zenithminerals.com.au
Discovery and Geology of the McPhee 
Creek Iron Deposit, Northern Pilbara, 
Western Australia
J D Goldsworthy1, R M Joyce2, P Bonato3 and A D’Hulst4
ABSTRACT
The McPhee Creek bedded iron deposit is a recent discovery made by Giralia Resources NL (Giralia) 
in the northern Pilbara of Western Australia. The deposit lies 30 km north of the town of Nullagine 
along the southeastern margin of the East Pilbara Granite-Greenstone Terrane (EPGGT).
The McPhee Creek area is located in the Kelly Greenstone Belt in the southeastern part of the 
EPGGT. The Archaean aged greenstone belt consists of volcanic and sedimentary sequences 
including the Warrawoona Group and the Gorge Creek Group. The McPhee Creek iron ore 
deposit lies within banded iron formation (BIF) rocks of the Gorge Creek Group. The structure 
of the McPhee Creek area is dominated by a northeast trending, upright synform (Sandy Creek 
Syncline) and associated folding that is truncated by northeast trending faults in the east. Iron 
ore mineralisation is predominantly stratabound and follows the dip of the bedding – however, 
in some areas, particularly in the southwest of the deposit, mapped steep to moderate dips of 
bedding planes are at odds with underlying f at mineralised zones in drill sections. Zones within 
the deposit that contain substantial thickness of iron ore correspond to areas of complex folding 
suggesting structural control and iron enrichment in fold hinges. Petrographic analyses con f rm 
mineralisation is goethite dominant and the deposit is strongly supergene enriched.
Giralia initially identif ed iron ore potential in the McPhee Creek area through the presence of a small 
known channel iron deposit (CID) mesa. Giralia began exploration in 2008 by re-drilling a portion 
of the CID (subsequently named Crescent Moon) and reported a small low-grade Inferred Mineral 
Resource. Target generation in the wider area utilising Geological Society of Western Australia maps 
and Ikonos satellite images identif ed strongly ferruginous outcrops associated with major structures 
along an 8 km long range located 2.5km to the west north-west of the Crescent Moon CID deposit. In 
late July 2008, a total of 62 rock chip samples were collected during a helicopter supported sampling 
program conducted at McPhee Creek. Samples taken from haematite-goethite enriched BIF and 
canga zones outcropping intermittently over the 8 km length of the western and central areas of the 
prominent range assayed up to 63 per cent Fe. These zones represent the discovery outcrops and 
surface expression of the McPhee Creek main range iron ore deposit.
Follow up sampling, geological mapping, bulldozing of access tracks and ramps and reverse 
circulation (RC) drilling commenced in September 2009. A maiden Inferred Mineral Resource 
was announced by Giralia to the Australasian Securities Exchange (ASX) on 15 December 2009.
The current Indicated and Inferred Mineral Resource at December 2010 stands at 260 Mt 
grading 56.2 per cent Fe, 0.12 per cent P, 6.7 per cent SiO2, 2.3 per cent Al2O3 and 9.5 per cent LOI 
at a 50 per cent Fe cut-off. McPhee Creek is one of the largest iron deposits found in Archaean BIF 
in the northern Pilbara.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
J D GOLDSWORTHY et al
94
on exploration licence 46/733-I. The company’s tenement 
package is located on both Vacant Crown Land and the Bonnie 
Downs pastoral lease.
EXPLORATION HISTORY
The McPhee Creek region has been subject to prospecting, 
mining and exploration for over 100 years. The focus of past 
explorers and prospectors has been predominantly for gold, 
base metals and asbestos.
The Lionel Mining Centre is located 6 km to the south of the 
McPhee Creek iron deposit where a number of small asbestos 
(chrysotile) occurences have been mined. The asbestos occurs 
in layered ultrama f c rocks associated with a complex fault 
system between the Kelly and McPhee greenstone belts. 
Further south, copper has been mined at the Lionel and 
Hendersons copper deposits associated with narrow veins 
in ultramaf c and gabbroic rocks. To the east, gold has been 
found in quartz veins at the McPhee’s and Gold Show Hill 
gold prospects. At Quartz Circle (12 km SE of McPhee Creek, 
Figure 2) explorers from the 1970s onwards have intersected 
in drilling, zones of massive and vein associated Cu-Zn-Ag 
mineralisation in predominantly felsic rocks.
The only documented prior iron ore exploration in the area 
was in 1979 - 1980 when Amoco Minerals Australia Company 
(Amoco) held an iron ore Temporary Reserve No 7173H which 
covered an area of 93 km 2 at McPhee Creek. Amoco’s main 
focus was a small east-west trending Channel iron deposit 
(CID) mesa where it conducted rock chip sampling, reporting 
values between 51 per cent and 62 per cent Fe, and shallow 
vacuum drilling. A total of 22 holes were drilled along a 1350 m 
long section of the CID mesa with the deepest hole reaching 54 
m. The drilling intersected goethite and maghemite enriched 
pisolites with characteristic fossil wood fragments. Iron 
enrichment was reported in the top 10 - 14 m of the CID with 
 
FIG 1 - Location of the McPhee Creek iron ore deposit with respect to other known iron deposits (and owners) in the region.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
DISCOVERY AND GEOLOGY OF THE MCPHEE CREEK IRON DEPOSIT, NORTHERN PILBARA, WESTERN AUSTRALIA
95
modest grade intersections returned including 12 m @ 53.7 
per cent Fe, 10 m @ 53.4 per cent Fe and 14 m @ 51.5 per cent 
Fe. Amoco conducted minor reconnaissance sampling in the 
wider area away from the CID mesa including an area in the 
northwest of the Temporary Reserve (TR), outside Giralia’s 
current tenure, where ten rock chip samples were described 
as CID type ore. Seventeen samples were also collected from 
the southwest of the TR with values returned varying between 
43 per cent and 51 per cent Fe. The samples were described as 
being associated with banded iron formations (BIFs). Amoco 
concluded that although its assessment of the potential of the 
TR was not completed due to rugged terrain and unseasonal 
FIG 2 - McPhee Creek regional geology.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
J D GOLDSWORTHY et al
96
rainfall, its exploration work had not identif ed large tonnages 
of iron ore (Purkait, 1980) and it subsequently relinquished 
the TR in 1980.
DISCOVERY OF MCPHEE CREEK IRON DEPOSIT
In November 2007, exploration licence E46/733-I was granted 
to Giralia. A review of the GSWA geology maps in particular 
the Nullagine 1:100 000 sheet revealed the presence in the 
tenement area of BIFs of the Gorge Creek Group which hosts 
iron ore deposits elsewhere in the West and East Pilbara 
Granite-Greenstone Terrane. The GSWA 1:100 000 Nullagine 
sheet also recorded the small McPhee Creek CID hosted iron 
occurrence. Research of the GSWA WAMEX database located 
Amoco’s 1980 McPhee Creek iron exploration report, which 
documented the vacuum drill testing of the CID mesa and the 
potential presence of iron enriched BIF’s 2 - 4 km further west.
A f eld reconnaissance trip was undertaken in late November 
2007 when the CID was visited and sampled. Outcrops of 
Gorge Creek Group iron formations in the ranges to the west 
were brie f y inspected, but oncoming bush f res thwarted 
detailed examination.
In 2008, Giralia began exploration by drilling the CID 
(subsequently named Crescent Moon). A total of 43 RC holes 
were drilled targeting the central 1.6 km section of the Crescent 
Moon CID, which was partially tested by Amoco in 1980. The 
Crescent Moon mesa is an elongate east-west crescent shaped 
body of channel iron up to 140 m wide. To the west a number 
of smaller mesas dissected by drainages occur over a combined 
length of 900 m. Drilling intersected pisolitic, goethite and 
maghemite-rich CID from surface down to a vertical depth 
of 16 m; below which a mixture of clayey goethite and chert-
silica-rich zone was encountered. At a depth of between 
20 - 24 m, a zone of ‘puggy’ white to purple clay was intersected 
at the base of the channel, with residual clay zones grading into 
saprolitic basement Archaean ultrama f c rocks below. Assay 
results from the goethite – maghemite zones (from surface 
to 8 - 16 m in depth) average >50 per cent Fe; below this 
point Fe grades are signif cantly reduced. Better intersections 
included, 14 m @ 55.9 per cent Fe and 12 m @ 55.3 per cent 
Fe. The drill density at 100 × 50 m spacings and the f at 
lying continuous nature of the Crescent Moon mesa allowed 
the estimation of an Inferred Mineral Resource of 5.17 Mt 
grading 53.6 per cent Fe, 0.03 per cent P, 7.2 per cent SiO 2, 
6.1 per cent Al2O3 and 11.3 per cent LOI (Louw, 2008).
In late July 2008, a helicopter-supported sampling program 
was conducted at McPhee Creek. Targets for the survey were: 
1. an approximately 8 km long range 2 - 3 km west northwest 
of the Crescent Moon CID – where newly acquired Ikonos 
satellite images identi f ed strongly ferruginous outcrops 
associated with major structures in mapped (BIFs), and
2. reported 1980 Amoco Fe-rich rock chip samples from 
the area – although their exact location was not known 
(preGPS).
A total of 62 rock chip samples were collected, assaying up to 
63 per cent Fe. The samples were taken from haematite-goethite 
enriched BIF and canga zones outcropping intermittently 
along the western and central areas of the prominent range. 
These zones represent the discovery outcrops and surface 
expression of the McPhee Creek main range iron ore deposit. 
Outcrop examination of the mineralised zones indicated a 
northeast strike with shallow to steep southeast dips.
After a period of follow up sampling, geological mapping, 
bulldozing of access tracks and ramps, RC drilling on the main 
range commenced in September 2009. An initial program of 
47 angled RC holes was completed on nominal line spacings 
of 200 m targeting mineralised outcrops on the western side 
of the main range. Very encouraging results were obtained 
from almost every drillhole. Better intersections included 
90 m @ 58.6 per cent Fe, 62 m @ 57.3 per cent Fe and 58 m 
@ 59 per cent Fe. A further drill program commenced in 
November 2009 and 24 previously permitted angled RC drill 
holes were completed. The drilling returned similar haematite 
– goethite rich Fe intersections and con f rmed the continuity 
of iron ore mineralisation between and along the lines of 
drillholes. Results from the drilling at this point con f rmed 
that the mineralisation dipped shallowly to the southeast and 
extended south into a synformal hinge zone. Based on the f rst 
71 drill holes completed, a maiden Inferred Mineral Resource 
of 52.1 Mt grading 56 per cent Fe, 0.08 per cent P, 6.7 per cent 
SiO2, 3.2 per cent Al 2O3 and 9.2 per cent LOI was estimated 
(Louw, 2010a) and announced by Giralia to the ASX on 15 
December 2009. Subsequent mineral resource estimates were 
publicly reported on 26 July 2010 and 9 September 2010, with 
the current 260 Mt estimate reported on 3 December 2010. 
On 21 December 2010, Atlas Iron Limited announced an 
off-market takeover bid for Giralia.
REGIONAL GEOLOGY
The McPhee Creek deposit is located in the Kelly Greenstone 
Belt in the southeastern part of the East Pilbara Granite 
Greenstone Terrane (EPGGT). The Archaean aged greenstone 
belt consists of volcanic and sedimentary sequences including 
the Warrawoona Group and the Gorge Creek Group (Figure 2). 
Unassigned ultrama f c rocks intrude the southern area of 
the Kelly Greenstone Belt and are mapped as serpentinised 
metadunites by the GSWA in the Lionel Mining Centre. West 
of the Kelly Greenstone Belt is the Corunna Downs Granitoid 
Complex. Abutting to the east is the domal McPhee Greenstone 
Belt, which is intruded by prominent gabbro, and dolerite 
dykes. To the south of the McPhee Dome lies sediments of 
the Mosquito Creek Basin. Unconformally overlying the Kelly 
Greenstone Belt, Mosquito Creek Basin and the McPhee Dome 
are sediments and volcanics of the Neo Archaean Fortescue 
Group. A series of northeast faults occur between the Gorge 
Creek and Warrawoona Groups in the Kelly Greenstone Belt, 
and form terrane boundaries between the Gorge Creek Group 
and McPhee Greenstone Belt on the western margin of the 
McPhee Dome (Bagas, 2005).
MCPHEE CREEK IRON DEPOSIT
Geology
T he McPhee Creek bedded iron ore deposit lies in the faulted 
Sandy Creek Syncline within rocks of the Gorge Creek Group. 
The Gorge Creek Group has an age estimate of c.3020 Ma (Van 
Kranendonk et al, 2006). The Gorge Creek Group is subdivided 
into the Farrel Quartzite and Cleaverville Formation. In the 
McPhee Creek area the Farrel Quartzite consists of banded 
chert, black carbonaceous shale and siltstone. The Cleaverville 
Formation, which hosts the McPhee Creek main range deposit, 
consists of thinly bedded BIF and ferruginous chert. Outcrops 
of haematite and goethite enriched BIF and canga form the 
surface expression of the McPhee Creek deposit. The western 
margin of the deposit is marked by a package of shale and chert 
and is interpreted to be part of the Farrel Quartzite (Figure 3). 
The shale-chert sequence appears to be conformable with the 
BIF, and a similar sequence also underlies the BIF to the east, 
however in the southeastern area of the main range deposit, 
east of a prominent northeast trending fault, high magnesium 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
DISCOVERY AND GEOLOGY OF THE MCPHEE CREEK IRON DEPOSIT, NORTHERN PILBARA, WESTERN AUSTRALIA
97
basalt has been intersected in drilling at depth below the BIF. 
This unit does not outcrop in the near vicinity. A large area 
of silcrete, which contains fragments of BIF, chert and shale, 
outcrops in the central area of the range. Further zones of 
haematite are mapped in the southwest of the range.
Structural geology
The structure of the McPhee Creek area is dominated by a 
northeast trending, upright synform and associated folding 
that is truncated by a northeast trending fault de f ning the 
eastern edge of the Cleaverville Formation (Noble and Beeson 
2010). The upright folding deforms earlier formed recumbent 
folds. Bedding usually strikes northeast-southwest with 
moderate dips to both the west and east. Northwest-southeast 
trending faults, with associated joints, fractures and open 
folding, overprint and cross-cut the McPhee Creek area (Noble 
and Beeson, 2010).
Early recumbent folding (D1)
Detailed geological mapping identi f ed several inclined to 
recumbent folds in the project area (Figure 4). The folds are 
tight to isoclinal with transposed limbs and in some cases 
form intrafolial folds. The recumbent folds are observed at 
the outcrop scale but not mappable at the project scale. If 
present however, large recumbent folds would be responsible 
for substantial thickening and local overturning of the BIF 
sequence (Noble and Beeson, 2010).
Upright folding (D2)
North-northeast trending upright folds occur throughout 
the project area (Figure 5). The folds vary from open to tight 
and have subvertical to steeply inclined axial planes and 
plunge moderately to both the northeast and southwest. 
The overall structure of the McPhee Creek deposit is a large 
northeast trending upright synform where the east limb has 
been truncated by the eastern fault zone. The hinge zone of 
the synform is best observed in the south of the project where 
it has a shallow plunge to the northeast (approximately 20° 
towards 042°). The upright folds refold earlier recumbent 
folding (Noble and Beeson, 2010).
 
FIG 3 - McPhee Creek geological interpretation.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
J D GOLDSWORTHY et al
98
Nor theast trending faults (D1 and D2)
A large northeast trending fault zone marks the eastern 
margin of the range. The fault zone is up to 50 m wide with 
silicif cation of the host rocks and minor quartz veining and 
is characterised by strong shearing and variably developed 
lineations. There is an increase in strain towards the eastern 
fault with zones of complex folding seen in BIF outcrops just 
to the west of the fault suggesting the upright and possibly the 
early recumbent folding are related or contemporaneous with 
movement along the eastern fault (Noble and Beeson, 2010).
Other parallel northeast trending fault zones occur in the 
north and south of the project (Figure 3).
Northwest trending faults and jointing (D3)
Cross-cutting the project area and earlier structures is a 
series of northwest trending faults with associated joints 
and fractures. The faults have steep dips to the northeast 
and southwest and appear to be responsible for a number of 
breaks and offsets in the iron ore mineralisation.
Iron ore mineralisation
Goethite-haematite iron ore mineralisation on the main 
range at McPhee Creek is hosted in BIF and in ferruginous 
laterite/canga. Mineralised outcrops occur along almost the 
entire western margin of the BIF (approximately 8 km) near 
the contact with underlying shale and chert. Outcropping 
mineralisation is also present in the south of the project in the 
position of the synformal hinge zone. Iron ore mineralisation is 
predominantly stratabound and follows the dip of the bedding 
– however in some areas, particularly in the southwest of the 
deposit, mapped steep to moderate dips of bedding planes are 
at odds with underlying f at mineralised zones in drill sections 
(Figure 6), indicating near surface supergene enrichment 
of the iron rich BIF. This supergene remobilisation and 
enrichment has created mineralised domains orientated 
subparallel to a now partially dissected palaeosurface, evident 
in the distribution of the canga – ferruginous laterite. This 
style of mineralisation maybe akin to ‘crustal ores’ that occur 
in the Nimingarra – Shay Gap – Yarri deposits in the northeast 
Pilbara (Podmore, 1990).
Areas within the deposit that contain substantial thickness of 
iron ore (drill intersections include 154 m @ 57.8 per cent Fe) 
correspond toareas of complex folding suggesting structural 
control and iron enrichment in fold hinges. 
Northeast trending faults bound and offset the iron ore 
mineralisation. The major eastern fault zone bounds the 
 
(A) (B) 
FIG 4 - Recumbent folding: (A) small recumbent folds (203226 mE, 7612937 mN); (B) view to southeast of west verging recumbent folds (201256 mE, 7609719 mN).
 
FIG 5 - Upright folding. Large upright synform. View looking to the north (photo taken from 201505 mE, 7609948N).
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
DISCOVERY AND GEOLOGY OF THE MCPHEE CREEK IRON DEPOSIT, NORTHERN PILBARA, WESTERN AUSTRALIA
99
eastern extent of mineralisation. Another northeast fault 
in the south of the deposit bounds and offsets the western 
mineralised zone. Other breaks in the deposit correspond to 
mapped and inferred northwest trending faults (Figure 3).
Petrography
Petrographic and mineragraphic analyses of a number of 
samples of PQ drill-core from various locations within the 
McPhee Creek main range deposit indicate that most samples 
 
FIG 6 - Geological cross-sections through the McPhee Creek Deposit. Sections show measured dip and dip directions along the drill section lines.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
J D GOLDSWORTHY et al
100
are strongly goethite rich with goethite up to 90 per cent 
(Teale, 2011).
Goethite occurs in many textural styles and habits. 
It can be massive, banded, colloform to botryoidal and 
pseudomorphous after martitised magnetite (Figure 7). 
The martitised magnetite is observed occurring in bands 
intercalated with goethite only or goethite-haematite bands 
(Teale, 2011). The magnetite is interpreted as a signi f cant 
component of the original primary BIF. In some samples 
goethite is intercalated with lepidocrocite in colloform, 
rhythmically banded aggregates. Lepidocrocite dominates in 
some samples (Teale, 2011).
The haematite occurs as described above in relict martitised 
magnetite as earthy aggregates, as platy to bladed grains, 
as coarse angular grains and micro-veins cutting goethite. 
Wherever present, haematite is ‘cemented’ by goethite (Teale, 
2011) 
The petrographic analysis indicates that the McPhee Creek 
main range deposit is a strongly supergene enriched iron 
deposit, with several phases of enrichment of the BIF evident.
Current resource
Subsequent to the initial Inferred Mineral Resource 
announced in December 2009, exploration drilling has been 
ongoing. The current Mineral Resource at 3 December 2010 
stands at 260 Mt grading 56.2 per cent Fe, 0.12 per cent P, 
6.7 per cent SiO2, 2.3 per cent Al2O3 and 9.5 per cent LOI at a 
50 per cent Fe cut-off (Louw, 2010b) based on 361 RC holes 
and three diamond drill holes and is classi f ed as Indicated 
and Inferred (Table 1). The resource was estimated by CSA 
Global Pty Ltd using Ordinary Kriging with model block sizes 
of 20 m × 20 m × 5 m with subcells down to 4 m × 4 m × 2 m. 
The majority of samples were collected as 2 m composites; 
the modelled mineralisation was grouped into zones bounded 
by interpreted structures for statistical analyses. The results 
of the grade estimation were validated by means of visual 
comparison along sections, statistical analyses and trend plots 
comparing the estimated block grades and drill hole sampling 
grades (Louw, 2010b). A bulk density of 3.1 t/m 3 was applied 
to the mineralised zones based on analyses of results of 
48 density measurements taken from within mineralised 
zones
DEVELOPMENT PLAN
Giralia commissioned Promet Engineering Consultants 
(ProMet Engineers Pty Ltd, 2010) to prepare a scoping 
study for the McPhee Creek project in early 2010, soon after 
the de f nition of the maiden JORC resource of 52.1 Mt in 
December 2009. A start-up mining rate of 2 Mt/a was selected 
as the Base Case based on public road haulage (approximately 
265 km) to Port Hedland, with the current 260 Mt resource 
clearly able to support signif cantly higher mining rates.
A number of variations on the base case mining, processing 
and transport options were considered, with superior f nancial 
model results returned from alternatives incorporating 
increased production rates and private haul roads or rail 
 
 
(A) (B) 
(C) (D) 
FIG 7 - Photomicrographs – refl ected light: (A) colloform and botryoidal goethite containing intercalations of lepidocrocite (white); (B) colloform banded goethite 
with coarse grained acicular goethite crystals perpendicular to banding; (C) goethite, haematite and martitised magnetite (former BIF with greater than 20 per cent 
magnetite); (D) goethite pseudomorphing martitised magnetite surrounded by haematite (white) and goethite.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
DISCOVERY AND GEOLOGY OF THE MCPHEE CREEK IRON DEPOSIT, NORTHERN PILBARA, WESTERN AUSTRALIA
101
spurs linking to third party rail. The scoping study indicated 
that the McPhee Creek project has potential for attractive 
returns under a number of different development scenarios. 
Further drilling, metallurgical test work and environmental 
studies as part of a prefeasibility study are needed to con f rm 
these results, to enable selection of the optimum development 
model.
CONCLUSIONS
The McPhee Creek bedded iron ore deposit in the northern 
Pilbara has progressed rapidly from the f rst identif cation of 
mineralised outcrops in late 2008, to the delineation of the 
current resource of 260 Mt just over two years later.
The iron ore enrichment at McPhee Creek is similar to 
the haematite-goethite mineralisation of other documented 
iron deposits in the Archaean east and west Pilbara Granite-
Greenstone Terrain. However, at McPhee Creek main range 
the combination of moderately dipping stratigraphy and 
low angle plunging folds cut by near vertical structures is 
interpreted to have enhanced the enrichment process and 
created exceptionally thick zones of relatively f at-lying, near-
surface mineralisation.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the hard work of the f eld 
and off ce personnel of Giralia Resources NL, consultants CSA 
Global Pty Ltd and Jigsaw Geoscience Pty Ltd.
REFERENCES
Bagas, L, 2005. Geology of the Nullagine 1:100,000 sheet, 33 p, 
Western Australia Geological Survey, 1:100,000 Geological Series 
Explanatory Notes.
Louw, G, 2008. Crescent Moon Mineral Resource Estimate Summary 
Report, unpublished consulting report for Giralia Resources NL, 
CSA Global Pty Ltd.
Louw, G, 2010a. McPhee Creek Mineral Resource Report, 
unpublished consulting report for Giralia Resources NL, CSA 
Global Pty Ltd.
Louw, G, 2010b. McPhee Creek Mineral Resource Estimate, 
Technical Summary, unpublished consulting report for Giralia 
Resources NL, CSA Global Pty Ltd.
Noble, M and Beeson, J, 2010. Observations and results of geological 
mapping in the McPhee Creek project, Pilbara Region, Western 
Australia, Australia, unpublished consulting report for Giralia 
Resources NL, Jigsaw Geoscience.
Podmore, D C, 1990. Shay Gap-Sunrise Hill and Nimingarra iron 
ore deposits, in Geology of the Mineral Deposits of Australia 
and Papua New Guinea (ed: F E Hughes), pp 137-140 (The 
Australasian Institute of Mining and Metallurgy: Melbourne).
ProMet Engineers Pty Ltd, 2010. McPhee Creek Iron Ore Project 
– Scoping Study, unpublished consulting report for Giralia 
Resources NL.
Purkait, P K, 1980. Final report McPhee Creek, Temporary Reserve 
7173, Western Australia, GSWA open f le report item 1312, A9416.
Teale, G S, 2011. A mineragraphic investigation of a suite of drill-core 
samples from the McPhee Creek iron deposit, Western Australia, 
Teale and Associates, unpublished consulting report for Giralia 
Resources NL.
Van Kranendonk, M J, Hickman, A H, Smithies, R H, Williams, I R, 
Bagas, L and Farrell, T R, 2006. Revised lithostratigraphy of 
Archaean supracrustal and intrusive rocks in the northern Pilbara 
Craton, Western Australia: Western Australia Geological Survey,Record 2006/15, 57 p.
Deposit 
cut-off grade
Category Tonnes (Mt) Fe % P % SiO
2
% Al
2
O
3
% LOI% Fe calcined %
Total >50% Fe Indicated 65.3 56.3 0.11 6.2 2.6 9.7 62.3
Total >50% Fe Inferred 194.7 56.2 0.13 6.9 2.2 9.4 62.1
Total >50% Fe Combined 260 56.2 0.12 6.7 2.3 9.5 62.1
TABLE 1
Mineral resource estimate – results for McPhee Creek deposit.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 103
INTRODUCTION
The McCamey’s North project is located 50 km east of Newman 
and just 3 km from the Jimblebar Mine (Figure 1). The project 
is wholly contained within the exploration Lease E52/2303 
(the tenement), which was granted to Warwick Resources 
Ltd on 2 November 2009. During December 2009, Warwick 
Resources Ltd and Atlas Iron Limited (Atlas) successfully 
merged. To date 38.9 Mt of haematite-martite and goethite 
iron mineralisation has been identi f ed within the Boolgeeda 
Iron Formation at McCamey’s North. The Boolgeeda Iron 
Formation stratigraphy is well exposed in the southern 
portion of the tenement but disappears beneath colluvial and 
alluvial cover to the north.
Atlas commenced exploration at McCamey’s North in 
December 2009, with 1:10 000 scale mapping and rock chip 
sampling. Drilling commenced in August 2010 with 184 
reverse circulation drill holes totalling 16 886 m drilled as of 
3 February 2011, when a maiden inferred resource was released 
(Table 1). Drilling has been completed at 50 m centres on 
100 - 200 m spaced lines (Figure 2).
Previous work in the area includes the GSWA 1:250 000 
Robertson Sheet SF51-13 (Williams and Tyler, 1991); which 
shows the tenement area as being dominated by north-dipping 
1. MAusIMM, Senior Exploration Geologist, Atlas Iron Limited, Level 9, Alluvion, 58 Mounts Bay Road, Perth WA 6000. Email: Paul.Howard@atlasiron.com.au
2. Exploration Manager, Atlas Iron Limited, Level 9, Alluvion, 58 Mounts Bay Road, Perth WA 6000. Email: Pip.Darvall@atlasiron.com.au
Realising the Potential of the 
Boolgeeda Iron Formation – 
Stratigraphy and Iron Mineralisation at 
McCamey’s North, Hamersley Province, 
Western Australia
P J Howard1 and P Darvall2 
ABSTRACT 
Atlas Iron Limited’s McCamey’s North project lies at the eastern end of the Hamersley Basin, 
approximately 50 km east of Newman, Western Australia, and 3 km northeast of BHP Billiton’s 
Jimblebar Mine. The presence of signi f cant surface iron mineralisation identi f ed during f eld 
reconnaissance, has subsequently been further def ned by detailed mapping and RC drilling.
The project area is dominated by outcrops of moderate to steeply north and south dipping 
banded iron formation, cherts and shales of the Boolgeeda Iron Formation overlying rhyolites and 
rhyodacites of the Woongarra Volcanics. In lower lying areas recent colluvium and alluvium mask 
the Archaean geology. Drilling and downhole geophysics has allowed a composite stratigraphic 
section to be compiled through almost the entire Boolgeeda Iron Formation.
Surface orientations and subsurface stratigraphic correlations using natural gamma logs indicate 
large scale open folding about east-west trending fold axes across the project area; while small 
scale tight to isoclinal folding with similar fold axis orientations are visible at the outcrop scale. 
These folds are cut by a number of moderate to steeply south-dipping normal faults de f ning a 
series of north-dipping fault blocks which repeat the stratigraphy across the project area. 
Bedded iron mineralisation consists mainly of haematite-martite and goethite ores with best 
results of 78 m @ 60.5 per cent Fe. Vitreous goethite is observed at the surface of the mineralised 
horizons. The mineralisation exhibits a strong stratigraphic control and extend down-dip up to 
120 m below surface. The strike extent of the mineralisation is still to be determined. An inferred 
resource of 39 Mt at 58 per cent Fe has been def ned to date (ASX: AGO 3 February 2011).
The f ndings at McCamey’s North indicate that there is potential for economic iron mineralisation 
to exist in other areas of Boolgeeda Iron Formation within the Hamersley Basin.
Class Mt
Grade %
Fe SiO
2
Al
2
O
3
P LOI S
Inferred 38.9 57.96 4.94 4.81 0.169 6.27 0.012
Fifty-three per cent fe cut-off .
TABLE 1 
McCamey’s North Inferred Resource – February 2011.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
P J HOWARD AND P DARVALL
104
units of the Archaean-Proterozoic Woongarra Volcanics and 
the Boolgeeda Formation extending north under Recent 
cover material. Also visible on the 1:250 000 mapping are a 
signif cant number of northeast trending faults extending 
onto the tenement area from the vicinity of the Jimblebar 
iron ore mine. Such structures (including the Whaleback 
and nearby Wheelarra Faults) are regarded as important in 
controlling the location of iron enrichment in the Newman 
area (eg Kneeshaw, 2004).
REGIONAL SETTING
The McCamey’s North project is located at the eastern end 
of the Hamersley Province, Western Australia (Figure 1). 
Archaean granitoid-greenstone sequences of the Sylvania 
Inlier, the Fortescue Group and Hamersley Group successions 
dominate the region (Tyler, 1991), covered by Quaternary and 
Recent cover. 
Up to f ve deformation events have been recognised across 
the Hamersley Province, with the Capricorn and Ashburton 
Orogenies being the main episodes in the southern part 
of the Province (Tyler, 1991; Kneeshaw, 2004). In this 
southern region of the Hamersley Province the Capricorn or 
Ophthalmian Orogeny is characterised by south over north 
directed thrusting and folds with tight inter-limb angles and 
southerly dipping fold axes. The later Ashburton Orogeny is 
characterised by large-scale, upright E-W trending folds that 
def ne the regional outcrop pattern.
Approximately 6 km south of the tenement area, granitic 
basement of the Sylvania Inlier is unconformably overlain 
by volcanic and sedimentary units of the Archaean Fortescue 
Group, conformably overlain by the entire Hamersley 
Group succession. The Hamersley Group stratigraphy dips 
and youngs to the north, with only the uppermost units; 
the Woongarra Volcanics and Boolgeeda Iron Formation 
(‘Boolgeeda’) exposed in the tenement area (eg Tyler, 1991; 
Williams and Tyler, 1991). 
STRATIGRAPHY
A schematic stratigraphic column for the McCamey’s North 
project is presented in Figure 3.
Woongarra Volcanics
Units correlated with the Archaean Woongarra Volcanics form 
the lowermost stratigraphic unit encountered at McCamey’s 
North. These volcanics are seen to outcrop along the southern 
margin of the project area and where uplifted in normal fault 
blocks in the central portion of the tenement (Figure 2). This 
unit consists of approximately 300 m of massive to locally 
f ow banded rhyolite and rhyodacite, with several 1 - 5 m thick 
jaspilitic banded iron-formation (BIF) and chert horizons 
and discontinuous 2 - 5 m thick dolerite sills in the central 
portion. The felsic units typically have a pale orange to brown 
microcrystalline groundmass with 1 - 2 per cent 1 - 2 mm 
euhedral phenocrysts of feldspar ± quartz. A f nely layered 
tuffaceous unit was observed in one location suggesting an 
extrusive origin for at least some of the unit. The reported 
thickness for the Woongarra Volcanics in the area is 260 m 
(Williams and Tyler, 1991).
Boolgeeda Iron Formation
Conformably overlying these volcanic units is a package 
of BIF, cherts and shales of the Boolgeeda Iron Formation, 
which dominate the project area. 
Immediately overlying the Woongarra Volcanics are 5 - 10 m 
of chaotically folded and brecciated BIF, commonly jaspilitic 
(see BG1); overlain in turn by a f nely laminated, blocky chert 
and BIF unit approximately 50 m thick, and 30 - 40 m of 
calcareous/dolomitic shales capped by ochreous goethite 
and calcrete. A series of cherts, BIF and shales make up the 
remainder of the unit, with an estimated overall thickness 
of 250 m. A total thickness of up to 340 m has been inferredFIG 1 - McCamey’s North location plan, surrounding geology and infrastructure.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
REALISING THE POTENTIAL OF THE BOOLGEEDA IRON FORMATION – STRATIGRAPHY AND IRON MINERALISATION AT MCCAMEY’S NORTH, WA
105
for the Boolgeeda Formation in this location. Williams and 
Tyler (1991) give an approximate thickness of 200 m but 
acknowledge the top of the formation was not observed. 
A 254 m composite stratigraphic drill hole through a large 
part of the Boolgeeda Iron Formation at McCamey’s North 
has allowed the identi f cation of 11 distinct units (BG1-11) 
as outlined in Figure 3. The lowermost two units were not 
intersected in drilling and are based on f eld mapping only, 
while the remainder were intersected in RC drilling and 
extensive downhole natural gamma data. A combination 
of geological logging, assay interpretation and downhole 
geophysics, have allowed these units to be correlated across 
the project area. The following is a more detailed description 
of the 11 units outlined in Figure 3 and discussed above:
  BG1 – brecciated and chaotically folded BIF. This 5 - 10 m 
thick unit is a distinct marker horizon and will be referred 
to as a jaspilite for it exhibits a brilliant red haematite 
banding, interbedded with a white silica rich chert, 
very different to the other BIF horizons seen within the 
Boolgeeda at McCamey’s North.
  BG2 – blocky chert and BIF unit approximately 50 m thick.
  BG3 – a shaley-BIF; the lowermost unit encountered 
in drilling, displaying an increased silica and alumina 
geochemical signature and has been logged as thinly 
laminated BIF interbedded with shale. Iron values are 
generally 35 - 40 per cent within this unit. It has an 
estimated total thickness of 40 m, although only 15 m of 
this unit was intersected in drilling.
  BG4 – logged as an oxidised BIF where magnetite has 
been oxidised to martite. This commonly mineralised unit 
is up to 70 m thick with iron mineralisation consistently 
above 62 per cent. Increased shale bands towards the base 
of this unit decreases iron values to ~57 per cent for the 
lowermost 20 m. 
  BG5 – this thin shale unit is approximately 10 m thick and 
is high in alumina (7 - 12 per cent Al 2O3). It has a cream/
beige, f ne grained appearance with minor fragments of 
silica present.
  BG6 – A 20 m thick BIF with numerous thin interbedded 
shale and chert bands showing up in the ‘high background’ 
natural gamma log.
  BG7 – is a 40 m thick cherty-shale band. This unit has 
been correlated with a green chert zone mentioned by 
Kneeshaw (2004). It is easily recognisable in drilling and 
is a useful marker horizon throughout the project area.
  BG8 – a 20 m shaley-BIF unit, logged as interbedded 
haematite-goethite and shale. Alumina is elevated through 
this unit and iron values are typically 53 - 54 per cent. 
Minor jaspilite bands are common; red and rich in silica.
  BG9 – a 30 m thick BIF unit, with little or no interbedded 
shale and chert bands, and a very ‘clean (low background)’ 
natural gamma log, which indicates that this BIF is not a 
structural repetition of BG4. Commonly exhibiting strong 
iron enrichment to haematite-martite. 
  BG10 – similar to BG8, this 20 m shaley-BIF is mineralised 
and could be bene f ciated. Geochemical and natural 
gamma signatures differ enough to conclude that this unit 
is not a structural repetition of BG8.
  BG11 – is the uppermost Boolgeeda Iron Formation unit 
encountered at McCamey’s North. It is a f ne grained 
yellow/orange shale with a thickness of at least 35 m thick. 
The top of this unit has not yet been identif ed.
Quaternary and recent sediments
At McCamey’s North, the more resistant units of the Boolgeeda 
Iron Formation form low, east-west trending ridges, with 
lower lying shalier areas (BG9) are blanketed in Quaternary 
colluvium and alluvium. The colluvium, consisting of 
 
FIG 2 - Geological map of the McCamey’s North Project.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
P J HOWARD AND P DARVALL
106
FIG 3 - Schematic stratigraphic column for the McCamey’s North project.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
REALISING THE POTENTIAL OF THE BOOLGEEDA IRON FORMATION – STRATIGRAPHY AND IRON MINERALISATION AT MCCAMEY’S NORTH, WA
107
subangular to subrounded fragments of adjacent lithologies, 
typically drapes the lower portions of slopes. Alluvium occurs 
as well sorted, f ne grained sands and soils along the margins 
of the present Jimblebar Creek drainage.
STRUCTURE
Units at McCamey’s North strike generally east-west, 
swinging to a more northwest-southeast orientation at the 
eastern extent of the project area. The exposed ridges towards 
the south exhibit a moderate to steep dip to the north, while 
isolated ‘islands’ of outcrop among the Quaternary cover in 
the north display a moderate to steep dip to the south.
Natural gamma logs from numerous downhole surveys, 
coupled with assay and geological interpretation, support the 
surface observations, which def ne an open syncline/anticline 
pair (Figure 4). Numerous small-scale, tight to isoclinal folds 
are also visible at the outcrop scale. 
Brittle deformation represented by large-scale normal faults 
with up to 350 m of offset are also seen throughout McCamey’s 
North. These east-west trending, south dipping, normal faults 
expose the underlying Woongarra Volcanics (Figure 4). The 
monotonous nature of the Boolgeeda outcrop (no obvious 
marker horizons) means that there may be several other 
similarly oriented smaller scale faults not noted at the scale of 
mapping. These faults may be observed in diamond drilling or 
more detailed prospect scale mapping.
The change in strike orientation across the project area, 
mentioned above, is interpreted to be due to a younger, 
northeast trending dextral strike-slip fault cutting the prospect 
at the location of the Jimblebar Creek. Offset across this fault 
is inferred to be approximately 200 m.
MINERALISATION
We have de f ned iron mineralisation at McCamey’s North 
as material with greater than 53 per cent Fe and consists 
of haematite-martite and goethite ores. The high-grade 
mineralisation (~60 per cent Fe) is generally limited to the 
upper (BG9) and lower (BG4) BIF units (Figure 3) with the 
thinner middle (BG6) BIF unit containing subeconomic DSO 
grade material (45 - 53 per cent Fe). Table 2 contains some of 
the better intercepts from McCamey’s North.
The lower BIF unit (BG4) is overlain by barren shales 
and is approximately 70 m thick. Mineralisation extends 
to depths greater than 120 m below surface. The upper BIF 
(BG9) is bound by mineralised shaley-BIF. These shalier 
units are approximately 20 m thick but have a higher alumina 
content of 7 - 12 per cent than the high grade unit, which is 
approximately 30 m thick and has an alumina content of 
three to six per cent. 
The bedded haematite-martite ore is capped by a goethite-
rich hydrated zone. The ore exhibits a tight stratigraphic 
control and is open at depth towards the core of the interpreted 
syncline. 
Iron mineralisation at surface is characterised by siliceous 
or vitreous goethite with minor ochreous goethite and clay, 
and vuggy and breccia textures. There are also minor detrital/
canga outcrops at the base of slopes. These outcrops comprise 
angular to subangular ore clasts in a haematitic clay matrix 
(Archer, 2009; Darvall, 2010).
At depths of up to 120 m below surface the mineralisation 
appears as dense haematite-martite ore. Dark grey/
blue haematite and chocolate brown goethite alternate in 
mineralised bands of varying thicknesses ranging from 1 cm 
to 15 m. Conchoidal fracturing is apparent on larger goethite 
chips.
DISCUSSION
These f ndings show that the Boolgeeda Iron Formation 
holds signi f cant potential for economic iron mineralisation 
at McCamey’s North and throughout the Hamersley Basin. 
Further work to extend the current resource via RC and 
diamonddrilling, and more detailed prospect scale mapping 
will aid in unravelling the structural complexities of 
McCamey’s North. 
The available data suggests a much thicker Boolgeeda Iron 
Formation section than previously documented. However, 
there is a strong possibility that some structural thickening 
of the section has occurred. The additional work planned 
will enable this to be con f rmed or otherwise. A more 
comprehensive and detailed stratigraphic and natural gamma 
section for the unit will be compiled once the information 
becomes available.
 
FIG 4 - Schematic cross-section on 212850mE at the McCamey’s North project.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
P J HOWARD AND P DARVALL
108
CONCLUSIONS 
Extensive exploration completed to date at the McCamey’s 
North project has resulted in the de f nition of signi f cant 
Boolgeeda Iron Formation-hosted Fe mineralisation. 
Additional work designed to test the undercover potential 
has commenced and will assist in an even more detailed 
understanding of the stratigraphy and potential of the 
Boolgeeda as a whole across the Hamersley Province.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the valuable contributions 
made by the entire Atlas Iron Exploration team and in 
particular Steve Warner who contributed signi f cantly to 
the development of the geological model. The shareholders, 
Board and Management at Atlas are acknowledged for their 
ongoing support of Exploration activities and the production 
of this paper. 
REFERENCES 
Archer, D A, 2009. McCamey’s North Project – DSO potential initial 
f eld review, Warwick Resources consultant report, internal 
company report, pp 9-13.
Darvall, P, 2010. McCamey’s North mapping report, Atlas Iron 
Limited, internal company report, pp 3-5.
Kneeshaw, M, 2004. Guide to the Geology of the Hamersley and 
North East Pilbara Iron Ore Provinces (‘The Blue Book’) (BHP 
Billiton Iron Ore Exploration Western Australia). 
Tyler, I M, 1991. The Geology of the Sylvania Inlier and the Southeast 
Hamersley Basin (Geological Survey of Western Australia: Perth).
Williams, I R and Tyler, I M, 1991. Robertson 1:250 000 Sheet SF51-
13 (Geological Survey of Western Australia: Perth).
Hole_Id mFrom mTo Intercept (Fe %) SiO
2
% Al
2
O
3
% P % S % LOI %
JMRC044 36 92 56 m @ 58.61 % 3.36 3.77 0.21 0.01 8
JMRC052 18 68 50 m @ 60.26 % 3.08 4.2 0.17 0.01 5.56
JMRC058 16 94 78 m @ 60.50 % 2.25 4.49 0.18 0.01 5.67
JMRC064 0 78 78 m @ 59.02 % 5.1 4.04 0.15 0.01 5.36
JMRC065 24 76 52 m @ 60.03 % 4.14 4.62 0.13 0.01 4.52
JMRC144 46 120 74 m @ 59.82 % 3.63 4.58 0.19 0.01 5.33
JMRC208 18 86 68 m @ 59.76 % 3.75 4.71 0.2 0.01 5.09
JMRC209 64 128 64 m @ 61.69 % 2.33 4.05 0.2 0.01 4.6
JMRC210 48 114 66 m @ 58.92 % 3.77 4.56 0.2 0.01 6.29
JMRC221 4 64 60 m @ 58.94 % 4.53 4.35 0.15 0.01 5.79
JMRC228 20 70 50 m @ 57.36 % 5.26 5.56 0.2 0.01 5.87
JMRC229 52 108 56 m @ 57.81 % 5.4 5.03 0.18 0.01 5.84
 Fifty-three per cent Fe cut-off , 2 m max internal waste.
TABLE 2 
Best intercepts at McCamey’s North as of February 2011.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011 109
INTRODUCTION
Geophysics assists iron ore exploration where the effect of the 
varying physical properties of the rocks can be detected in a 
geophysical measurement. Iron ore deposits may be strongly, 
weakly or non-magnetic. They may or may not be electrically 
conducting or associated with electrically conductive or 
polarisable rocks. Generally, however, they can be relied on 
to have a positive density contrast and, consequently, gravity 
techniques are valuable in iron ore exploration.
Historically, many iron ore deposits have been discovered 
as a result of their extensive outcrop combined with relatively 
thin soil cover and sparse vegetation (Blockley, Reid and 
Trendall, 1990; Morris, 1998). Undiscovered outcropping 
deposits may still exist but the growth in demand and the 
preference for deposits near existing infrastructure means 
that there is growing opportunity for discovery under cover 
using geophysics (Butt, Hawke and Flis, 2001).
In the past, ground gravity has been used but very good 
quality terrain corrections are required (Flis, Butt and Hawke, 
1998) and this is dif f cult in the generally rugged terrains 
associated with iron ore provinces. In any case, it is slow and 
laborious to collect.
Airborne gravity gradiometry has been used since late 1999. 
This offers the advantages of rapid and complete coverage, 
excellent terrain corrections if light detection and ranging 
(LIDAR) data are collected simultaneously, and, usually, 
coincident aeromagnetics.
DEPOSIT STYLES
The iron ore mineral fact sheet (Geoscience Australia, 2010b) 
says:
In the Hamersley Province in the Pilbara district 
of Western Australia there are three main types of 
deposit: iron oxide enrichments within BIFs; iron 
oxides deposited along ancient, mainly Tertiary age 
river channels (palaeochannels); and iron oxide 
deposits formed from the erosion of existing orebodies 
(detrital iron ore deposits).
BIF enrichment iron ore deposits can be split into three 
broad groups comprising leached BIF, haematite-goethite 
and martite-microplaty haematite. However, leached BIF 
deposits tend to be found in areas of high rainfall, such as 
Brazil and India, and not the arid Hamersley Basin where the 
1. Senior Geoscientist, Fugro Airborne Surveys Pty Ltd, 435 Scarborough Beach Road, Osborne Park WA 6017. Email: rmiller@fugroairborne.com.au
2. Chief Geophysicist, Fugro Airborne Surveys Pty Ltd, 435 Scarborough Beach Road, Osborne Park WA 6017. Email: mdransfi eld@fugroairborne.com.au
Airborne Gravity Gradiometry and 
Magnetics in the Search for Economic 
Iron Ore Deposits
R Miller1 and M Dransfi eld2
ABSTRACT 
The enrichment of magnetite to haematite and haematite-goethite leads to a coincident magnetic 
low and gravity high within the banded iron formation. Palaeochannel deposits of pisolitic limonite 
can be expected to provide a relative magnetic high and gravity low in a sinuous geometry. Detrital 
iron deposits of conglomerate, scree or ‘canga’ might have indistinct magnetic response but will 
usually have a relative gravity high.
These signatures provide clear discriminators detectable by a joint gravity and magnetic survey.
The combination of airborne gravity gradiometry and magnetometry is particularly powerful 
in detecting and discriminating targets, as well as mapping lithologies and structure, across a 
wide variety of iron ore deposit styles. It provides detection capabilities in a platform capable of 
rapid, complete coverage with minimal access issues. This approach has been used in FALCON™ 
airborne surveys for iron ore exploration in the Great Lakes, USA; the Middleback Ranges, South 
Australia; and the Hamersley Basin, Western Australia.
The complex magnetic properties of iron formations (particularly remanence, anisotropy and 
depth of weathering) prevent this approach from being simple. The use of vector residual magnetic 
intensity (VRMI), calculated from the measured total magnetic intensity (TMI), assists by reducing 
the complexity due to the different directions of the remanent and induced magnetic f elds.
Estimations of the density to magnetisation ratio via the pseudo-lithology calculated from the 
VRMI and gravity gradients provide a useful method of capturing the discriminatory power of joint 
magnetic and gravity gradient data.
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
R MILLER AND M DRANSFIELD
110
principal deposits (eg Mount Whaleback, Yarrie, Mount Tom 
Price, Paraburdoo, Channar) are of the supergene enrichment 
haematite-goethite and martite-microplaty haematite 
style (Blockley, Reid and Trendall, 1990; Blockley, 1990; 
Morris, 1998). The martite-microplaty haematite ore type is 
particularly important in regard to resources and production 
as a result of the low phosphorous content often preferred for 
export specif cations (Morris, 1998).
Extensive discussionsof the formation of supergene BIF 
enrichment deposits in the Hamersley Basin can be found 
in literature (eg Blockley, Reid and Trendall, 1990; Blockley, 
1990; Morris, 1998) and a detailed discussion will not be 
presented here. However, it has been shown that supergene 
enrichment of BIF and, in some cases, subsequent burial 
metamorphism is the only mechanism that can explain the 
available data (Morris, 1998).
Haematite-goethite ore is produced by the oxidation of 
magnetite and metasomatic replacement of gangue minerals 
by goethite. Continual groundwater leaching or surface 
exposure can modify this ore producing a range of different ore 
types, from porous to extremely dense. This type of Mesozoic 
aged, non-metamorphosed ore constitutes over 90 per cent of 
the resources in the Hamersley Province (Morris, 1998).
The high-grade ores, such as Mount Tom Price and Mount 
Whaleback, are thought to have been buried at a depth of 
approximately 4 km, resulting in low grade metamorphism 
and generating haematite-rich ores from goethite. A second 
stage of groundwater leaching, resulting from erosion in the 
Mesozoic, removed the bulk of the residual goethite leaving 
high-grade, low phosphorous, haematite ore (Morris, 1998).
Channel iron deposits (CIDs) or palaeochannel deposits 
comprised of pisolitic limonite (eg Pannawonica (Mesa A 
and Mesa J), Yandicoogina) are the next in importance. 
The CIDs formed in predominantly Early to mid-Tertiary 
palaeochannels with the concentration of possible Oligocene 
to Miocene aged f uvial iron particles from weathered iron-
rich source rocks (Ramanaidou, Morris and Horwitz, 2003).
The palaeochannels typically occur as mesas or benches 
preserved by preferential weathering (Blockley, Reid and 
Trendall, 1990; Blockley, 1990; Ramanaidou, Morris and 
Horwitz, 2003) and though deposits range from <1 m to 
100 m thick with channel widths generally less than 1 km they 
can extend for several kilometres, the CID of the Robe River 
Formation (formerly the Robe Pisolite) can be traced for 80 km 
along the Marillana palaeochannel near Yandi (Blockley, 
1990; Ramanaidou, Morris and Horwitz, 2003). Whilst 
regionally iron concentration ranges from 40 - 60 per cent Fe, 
mining is generally restricted to deposits of anomalously high 
Fe, averaging 57 - 58 per cent (Blockley, 1990). Although the 
iron concentration tends to be lower than that of enrichment 
ores, low impurities (particularly phosphorous), lack of 
cementation, and shallow depth of burial make CIDs attractive 
targets.
Detrital iron ore deposits (eg Brockman Syncline No 2 
Deposit, Brockman Southwest Extension), composed of 
deposits of conglomerate, scree and ‘canga’, are derived 
from, and are usually found close to, enriched BIF deposits 
(Blockley, Reid and Trendall, 1990; Butt, Hawke and Flis, 
2001). Leaching f uids that result in the ferruginisation of the 
matrix also tend to reduce the phosphorous content, making 
this type of deposit more attractive than it would otherwise 
be (Blockley, Reid and Trendall, 1990; Butt, Hawke and Flis, 
2001).
Butt, Hawke and Flis (2001) present a geological cross-
section through Brockman Syncline No 2 Deposit Pit 5 showing 
a sequence of immature detrital material (partially cemented, 
poorly sorted subangular clasts of predominantly BIF, chert 
and haematite) overlying and masking the underlying mature 
detritals. These mature deposits tend to be subrounded to 
rounded clasts of predominantly haematite with minor BIF 
and chert in a matrix of ferruginous clay. Pisolites are identif ed 
by a higher proportion of ‘small ball-bearing type clasts’ in 
comparison to the mature detritals. ‘Canga’ is comprised of 
cemented detrital haematite clasts forming a hard rock at 
the base of the detrital material whilst a thin basal layer of 
limonite (mainly hydrated ‘canga’) separates the deposit from 
the underlying hydrated bedrock.
GEOPHYSICAL SIGNATURES OVER KNOWN 
IRON ORE DEPOSITS
One of the earliest geophysical successes in Australia was the 
discovery by gravity and magnetic surveys of the buried Iron 
Princess (Race Course) deposit in the Middleback Ranges 
(Bubner et al , 2003). The same authors note that the ‘iron 
formations in the Middleback Ranges are both dense and 
highly magnetic when compared with other lithologies in the 
area’ but the ‘iron ores are in general denser, but less magnetic, 
than the iron formation’.
Examples of the geophysical responses over iron ore regions 
and known iron ore deposits can be found in a number of 
publications (eg Dentith, Frankcombe and Trench, 1994; 
Kerr et al, 1994; Hawke and Flis, 1997; Flis, Butt and Hawke, 
1998; Flis, Hawke and McMillan, 1998; Butt, Hawke and Flis, 
2001; Bubner et al , 2003; Dentith, 2003). These examples 
predominantly discuss the results of magnetic or gravity 
surveys though Hawke and Flis (1997) discuss the application 
of electrical techniques for exploration in the Hamersley Basin 
and Flis, Hawke and McMillan (1998) present results from a 
helicopter-borne frequency domain electromagnetic survey.
Regional mapping is extremely important for exploration 
and airborne magnetic data can be used to map the structural 
and stratigraphic controls on mineralisation, providing 
targets for more detailed exploration (Dentith, Frankcombe 
and Trench, 1994; Kerr et al , 1994; Morris, 1998; Bubner 
et al, 2003).
The host iron formations tend to be the most magnetic units 
present, the iron ores being much less magnetic (Bubner et al, 
2003). As a result mineralisation can sometimes be associated 
with magnetic lows within an otherwise high amplitude 
background (Dentith, Frankcombe and Trench, 1994; Kerr 
et al , 1994; Bubner et al , 2003). However, identifying such 
responses can be dif f cult due to the large dynamic range 
usually present in the data, requiring careful processing 
to enhance the subtle features that might be present (Kerr 
et al , 1994), whilst anisotropic magnetic susceptibilities 
and the effects of remanence can further complicate the 
response (Dentith, Frankcombe and Trench, 1994). Magnetic 
responses can also be dependent on the depth of weathering 
as haematite is the dominant iron oxide in the weathered zone 
with magnetite dominating below it (Bubner et al, 2003).
The iron formations are generally remanently magnetised 
and the varying direction of this magnetisation with respect 
to the induced magnetisation means that the same formation 
can have a positive, negative or zero magnetic intensity as the 
formation’s orientation changes.
Iron ores tend to be the densest materials present, and in 
general the presence of dense haematite ore makes gravity an 
attractive exploration tool (Morris, 1998; Bubner et al, 2003) 
with positive gravity anomalies anticipated over signi f cant 
IRON ORE CONFERENCE / PERTH, WA, 11 - 13 JULY 2011
AIRBORNE GRAVITY GRADIOMETRY AND MAGNETICS IN THE SEARCH FOR ECONOMIC IRON ORE DEPOSITS
111
orebodies. In the BIF-enrichment deposit style, the density 
of iron formation host rocks can overlap with that of the 
orebodies so that the gravity method is a good detector but 
can be limited as a discriminator.
Thus the combination of magnetic and gravity data are 
particularly important in exploration for the enriched BIF 
style of mineralisation.
Buried palaeochannel deposits of pisolitic limonite can be 
expected to provide a relative magnetic high and gravity low 
in a typical sinuous geometry. Detrital iron deposits (DIDs) 
of conglomerate, scree or ‘canga’ are considered to have an 
indistinct magnetic response but a relative gravity high 
(Butt, Hawke and Flis, 2001). Underlying geology can further 
complicate the geophysical response.
Butt, Hawke and Flis (2001) describe the discovery of a 
detrital iron orebody using gravity.
The application of airborne gravity gradiometry technology 
to iron exploration provides several advantages in comparison 
with current gravity data. Ground access is limited by the 
rugged terrain