alta 2009 nickel/cobalt conference - alta ... 2009 nickel/cobalt conference may 25-27, 2009 sheraton...

ALTA 2009

NICKEL/COBALT

CONFERENCE

MAY 25-27, 2009

SHERATON HOTEL

PERTH, AUSTRALIA

ALTA Metallurgical Services Melbourne, Australia

www.altamet.com.au

ALTA Free Library www.altamet.com.au

PROCEEDINGS OF

ALTA 2009 NICKEL-COBALT CONFERENCE

25-27 May 2009Perth, Australia

ALTA Metallurgical Services Publications

All Rights Reserved

Publications may be printed for single use only. Additional electronic or hardcopy distribution without the express permission of ALTA Metallurgical Services is strictly prohibited.

Publications may not be reproduced in whole or in part without the express written permission of ALTA Metallurgical Services.

The content of conference papers is the sole responsibility of the authors.

To purchase a copy of this or other publications visit www.altamet.com.au

ALTA Metallurgical Services was established by metallurgical consultant Alan Taylor in 1985, to serve the worldwide mining, minerals and metallurgical industries.

Conferences: ALTA conferences are established major events on the international metallurgical industry calendar. The event is held annually in Perth, Australia. The event comprises three conferences over five days: Nickel-Cobalt-Copper, Uranium-REE and Gold-Precious Metals.

Publications: Sales of proceedings from ALTA Conferences, Seminars and Short Courses.

Short Courses: Technical Short Courses are presented by Alan Taylor, Managing Director.

Consulting: High level metallurgical and project development consulting.

ALTA Metallurgical Services Level 13, 200 Queen Street, Melbourne, Vic, 3000, Australia T: +613 8600 6909 | F: +613 9686 3008 | www.altamet.com.au


http://www.altamet.com.au/

http://www.altamet.com.au/

Elemental Engineering provides specialist process development and simulation services to the minerals processing industry.Our Perth and Sydney-based metallurgical consulting teams have extensive experience in the hydrometallurgical design of Nickel, Cobalt, Copper and Uranium circuits.

Core Services

• Flowsheetdevelopmentandevaluation

• ProcesssimulationutilisingSysCAD,Metsimandothersoftware

• Testworkdefinition,interpretation&management

• Pilotprogrammanagement

• ScopingandPrefeasibilityStudies

• ProjectFinancialEvaluation

Our key experience covers

• SolventExtraction

• CarbonatePrecipitation

• MixedHydroxidePrecipitation

• SulphidePrecipitation

• PressureMetallurgy

• HeapLeach

• Acid&AlkaliLeach

• Flotation

• IonExchange

• Electrowinning

• HydrogenReduction

• UraniumOxideProduction

Contact Us PERTHUnit 114/396Scarborough Beach RoadOSBORNE PARK WA 6017Ph: 08 6363 5298

SYDNEYLevel 6 - 69 Reservoir StreetSURRY HILLS, NSW 2010Ph: 02 8218 2138

i n fo@elementa l . ne t . au | www.e lemen ta l . ne t . au


CONTENTS

Conference Opening Address - Cobalt: Future Challenges David Weight, Cobalt Development Institute, UK

Trends in Nickel-Cobalt Processing Alan Taylor – ALTA Metallurgical Services - Australia

Identifying Opportunities to Reduce the Consumption of Energy Across Mineral Processing and Recovery Plants David Stirling – Schneider Electric - Australia

Production of Nickel Pig Iron in China Xian Jian Guo – Ramu NiCo (MCC) Ltd - China

Pressure Acid Leaching

The Mindoro Nickel Project HPAL Piloting and Process Development Niels Verbaan and James Brown – SGS Minerals – Canada, Boyd Wiilis – Independent Consultant to Aker Solutions – Australia, and Debbie Marshall – Aker Metals - Canada

Manganese Removal from the Gladstone Leach Residue Slurry and Solution by Precipitation W. Zhang, W. Wang, and Z. Zhu – CSIRO Minerals and P. Mason – Gladstone Pacific Nickel Ltd - Australia

Challenges of Material Selection for PAL Plants Naresh Balakrishnan - Consultant – Australia

Controlled Rapid Autoclave Blowdown Kevin Jackson – Mogas Industries – USA

Titanium: A Solution for Highly Corrosive Hydrometallurgical Applications- Alloy Selection, Cladding and Fabrication John Banker – Dynamic Materials Corp. - USA and Bruce Craig – MetCorr – USA

Field Performance Review of Autoclave Valves John Williams – Mogas Industries – USA

Benefits of TIMETA ® PGMA™ in Nickel Laterite Refining James Grauman and Eliana Fu – TIMET – USA and Ian Flower – Minara Resources Limited - Australia

Treatment of Sulphides

Engineering Aspects of the Platsol™ Process

Mike Wardell-Johnson and G Steiper – Bateman Engineering – Australia and D Dreisinger – PolyMet Mining - Canada

Nickel and Cobalt Recovery from Mesaba Concentrate Keith Mayhew, Rob Mean, Lisa O’Connor and Trevor Williams – CESL - Canada

X-Ray Sorting and Other Techniques for Upgrading Nickel Ore Allison Allen and Hilton Gordon – UltraSort - Australia

Process & Equipment Design

SO2/Air Oxidation - An Engineering Perspective Nitin Goel – GRD Minproc Ltd – Australia and Bill Bagulay – Mixtec - Australia

The Development of a DSX Process for the Recovery of Nickel and Cobalt from Laterite Leach Solutions - from Batch Tests to Pilot Plant Operation C.Y. Cheng, W. Zhang, D.J. Robinson, Y. Pranolo, Z. Zhu, L. Zeng and W. Wang – CSIRO Minerals, and G. Boddy and M. Godfrey – Rio Tinto Technology and Innovation - Australia

Agitator Design for Large Scale Hindered Settling Slurry Tanks Richard Kehn, Graham Seal, Robert Stewart and Bernd Gigas - SPX Process Equipment/ Lightnin – Australia/USA

The Evolution of Thickeners Ron Klepper – FLSmidth Minerals - USA


Click to navigate to the papers.

CONTENTS (CONT.)

Exploring Synergies in Western Australian Nickel Projects Hermann Scriba, Sam Spencer and Jeff Connor – SNC-Lavalin Australia

Laterite Heap Leaching

Nickel Laterite Processing – A Shift Towards Heap Leaching.

Bruce Wedderburn – Malachite Consulting – Australia

The Development of Nickel Laterite Heap Leach Projects Mark Steemson and Mark Smith - Vector Engineering, Ausenco Group – Australia/USA

Development of Heap Leaching and its Integration into the Murrin Murrin Operations D.J. Readett and J. Fox - Minara Resources, Murrin Murrin Operations - Australia

Cobalt Extraction & Refining Symposium

The Use of Bioleaching for Cobalt/Arsenic Tailings Remediation in Ontario Canada Paul Miller – BacTech Mining Corp. – Canada

The Application of Molecular Recognition Technology (MRT) in the Purification of Cobalt Process and Electrowinning Streams Steven Izaat, Neil Izaat, Ronald Bruening and John Dale - IBC Advanced Technologies Inc. - USA

Design of Copper-Cobalt Hydrometallurgical Circuits Graeme Miller – Miller Metallurgical Services Pty Ltd - Australia


Click to navigate to the papers.

Presented by

David Weight

General Manager

The Cobalt Development Institute, UK

[email protected]

ALTA 2009 OPENING ADDRESS

COBALT: FUTURE CHALLENGES


COBALTCOBALTCOBALT

27

58.93


3

You are solely responsible for evaluating the accuracy and

completeness of any content appearing in this presentation. Whilst the Cobalt Development Institute (CDI) has endeavoured to provide

accurate and reliable information, it does not make any

representations or warranties in relation to the content of this

presentation. In particular, the CDI does not make any

representations or warranties regarding the accuracy, timeliness or

completeness of the content of the presentation or in respect of its

suitability for any purpose. No action should be taken without

seeking independent professional advice. The CDI will not be

responsible for any loss or damage caused by relying on the

content contained in this presentation.

DISCLAIMERDISCLAIMERDISCLAIMERDISCLAIMERDISCLAIMERDISCLAIMERDISCLAIMERDISCLAIMER


4 Mission StatementMission StatementMission StatementMission StatementMission StatementMission StatementMission StatementMission Statement

““““““““Promoting the Sustainable Promoting the Sustainable Promoting the Sustainable Promoting the Sustainable Promoting the Sustainable Promoting the Sustainable Promoting the Sustainable Promoting the Sustainable

and Responsible Use of and Responsible Use of and Responsible Use of and Responsible Use of and Responsible Use of and Responsible Use of and Responsible Use of and Responsible Use of

Cobalt in all its FormsCobalt in all its FormsCobalt in all its FormsCobalt in all its FormsCobalt in all its FormsCobalt in all its FormsCobalt in all its FormsCobalt in all its Forms””””””””

� Support the sustainable

development and use of Cobalt,

Cobalt compounds and products

� Promote the use of and

knowledge about Cobalt in all

its forms


5 ActivitiesActivitiesActivitiesActivitiesActivitiesActivitiesActivitiesActivitiesActivitiesActivitiesActivitiesActivities

• Health, Safety and Environment

• Pro-active scientific work

• Regulation and Legislation

• Represent Members opinions

• Product Stewardship

• Information and data

• REACH management

• Representing Cobalt’s interests globally


6

Cobalt BasicsCobalt Basics


7 Occurence & PropertiesOccurence & Properties

•• DiscoveryDiscovery: Stockholm in1735 by G. Brandt (GpVIII Element)

•• Atomic Weight = 59Atomic Weight = 59 (stable and naturally occuring)

•• Stable in atmospheric oxygenStable in atmospheric oxygen (less reactive than Fe)

•• 3333rdrd most abundant elementmost abundant element

•• Mainly byMainly by--product of Ni & Cu miningproduct of Ni & Cu mining. Over 70

minerals but most commonly associated with minerals such as:

• Arsenides: CoAs2-3 (Smeltite and Safflorite)

• Arsenosulphides: CoAsS (Cobaltite)

• Sulphides: Co3S4 (Linnaeite)

• Found in: Ni-bearing laterites or Ni-Cu sulphide deposits

•• ResistantResistant: to many mild corrosive agents

•• Metallic CoMetallic Co: surface stability & strength


8 Crystal Lattice StructuresCrystal Lattice Structures

Body Centred Cubic Body Centred Cubic

Arrangement (BCC)Arrangement (BCC)

Face Centred Cubic Face Centred Cubic

Arrangement (FCC)Arrangement (FCC)


9

COBALT

APPLICATIONS

COBALT

APPLICATIONS


10

Cobalt Consumption by End-Use (2008) Superalloys

Hardmetal

Magnets

HS Steel

Batteries

Catalysts

Pigments

Organics

Other

Cobalt EndCobalt End--UseUse

Chemical Chemical

ApplicationsApplications

(56%)(56%)

Metallurgical Metallurgical

ApplicationsApplications

(44%)(44%)


11

Name of Substance Formula CAS EINECS

Cobalt Co 7440-48-4 231-158-0

Cobalt di(acetate) C2H3O2.½Co 71-48-7 200-755-8

Cobalt(2+) propionate C3H5O2.½Co 1560-69-6 216-333-1

Cobalt bis(2-ethylhexanoate) C8H15O2.½Co 136-52-7 205-250-6

Cobalt(2+) isononanoate C9H17O2.½Co 84255-52-7 282-603-0

Neodecanoic acid, cobalt salt C10H20O2.xCo 27253-31-2 248-373-0

Stearic acid, cobalt salt C18H35O2.xCo 13586-84-0 237-016-4

Oleic acid, cobalt salt C18H33O2.xCo 14666-94-5 238-709-4

Naphthenic acids, cobalt salts - 61789-51-3 263-064-0

Tall-oil, cobalt salts - 14666-96-7 238-711-5

Fatty acids, tall-oil, cobalt salts - 61789-52-4 263-065-6

Resin acids and Rosin acids, cobalt salts - 68956-82-1 273-321-9

Cobalt Compounds

About 30 known commercial substances:About 30 known commercial substances:


12

Name of Substance Formula CAS EINECS

Cobalt, borate propionate complexes - 91782-61-5 295-033-2

Cobalt, borate (2-ethylhex.complexes) - 91782-60-4 295-032-7

Cobalt, borate neodecanoate complexes - 68457-13-6 270-601-2

Cobalt, borate (2-ethylhex. neodec.complexes) - 92502-53-9 296-340-4

Cobalt oxalate C2O4.Co 814-89-1 212-409-3

Cobalt (II) 4-oxopent-2-en-2-olate C5H7O2.½Co 14024-48-7 237-855-6

Cobalt dichloride CoCl2 7646-79-9 231-589-4

Cobalt dinitrate Co(NO3)2 10141-05-6 233-402-1

Cobalt sulphate CoSO4 10124-43-3 233-334-2

Cobalt carbonate CoCO3 513-79-1 208-169-4

Cobalt oxide CoO 1307-96-6 215-154-6

Dicobalt trioxide Co2O3 1308-04-9 215-156-7

Tricobalt tetraoxide Co3O4 1308-06-1 215-157-2

Cobalt dihydroxide Co(OH)2 21041-93-0 244-166-4

Cobalt trihydroxide Co(OH)3 1307-86-4 215-153-0

Cobalt hydroxide oxide CoO(OH) 12016-80-7 234-614-7

Cobalt sulphide CoS 1317-42-6 215-273-3

Cobalt lithium dioxide CoO2.Li 12190-79-3 235-362-0

Cobalt Compounds Cont/-


13 Rechargeable BatteryRechargeable Battery

Typical raw material for LIB

• Cobalt metal or oxide/hydroxide

• Converted to lithiated cathode

material

2008 cobalt consumption

• >16,000 MT

Use

• Active cathode material

• Application requirements dictate the

chemistry and physical properties

Comment

• New chemical systems developed

during high price period


14 Hardmetal & Diamond ToolHardmetal & Diamond Tool

Typical raw material

• Cobalt oxalate

• Converted to fine cobalt powders


• >7,000 MT

Use

• Binder for cemented carbides (WC)

• Powder metallurgical parts processing

Comment

• New alloy alternatives in development

• Recycle plays an important role


15 PigmentPigment


• Predominantly cobalt oxide

• Some cobalt sulphate, carbonate,

and hydroxide


• 5-6,000 MT

Use

• Colourant additive

• Ceramic tile, glass, and plastic

Comment

• High cobalt price limits demand

• Had seen strong growth in China,

Asia, and Eastern Europe


16 Catalyst Catalyst –– Polymerisation/OxidationPolymerisation/Oxidation


• Cobalt acetate and hydroxide

• Converted to carboxylates in some

cases


• 4-5,000 MT Cobalt (PTA driver)

Use

• Increases polymerisation rates in:

� PTA – Purified Terephthalic Acid

� UPR – Unsaturated Polyester Resin

Comment

• Centered in Asia/Pacific

• Recycle is important


17 Rubber AdhesionRubber Adhesion


• Cobalt hydroxide and chloride

• Converted to cobalt carboxylates


• <2,500 MT

Use

• Wet blended adhesion additive

• Strengthens bond between rubber

and steel cord

Comment

• Competitive market – move of activity

to Asia


18 Catalyst Catalyst –– HydroprocessingHydroprocessing


• Cobalt carbonate and hydroxide

• Cobalt nitrate solution


• <1,500 MT Cobalt

Use

• Synthesised with other metals

(typically molybdenum)

• Promoter in Hydrodesulphurisation

(HDS) Catalyst

Market trends

• Driven by tighter sulphur legislation

• Cyclical with catalyst change-outs


19 CoatingsCoatings


• Cobalt hydroxide and metal

• Converted to cobalt carboxylates


• <1,000 MT

Use

• Wet blended drier additive

• Accelerates drying of alkyd based

paint, ink, and varnish

Market trends

• Movement towards low VOC

formulations or water emulsified

systems


20 Catalyst Catalyst –– Fischer Tropsch for GTLFischer Tropsch for GTL


• Cobalt metal converted to cobalt

nitrate


• Several hundred tonnes

Use

• Proprietary synthesis

• Key catalyst for gas-to-liquid

conversion at large plants in Qatar

(Shell, Sasol-Chevron)

Market trends

• GTL projects are more attractive

with high oil prices

• Recycle will be important

• Long term growth potential


21 SuperalloysSuperalloys

~11,000 tonnes in these ~11,000 tonnes in these

applications. Mainly USAapplications. Mainly USA


22 Casting AlloysCasting Alloys


23 Medical ProstheticsMedical Prosthetics


24 Remaining Significant ApplicationsRemaining Significant Applications

Recording MediaRecording Media

Copper Refining Copper Refining

Animal Feed Animal Feed

Electronics Electronics

Surface TreatmentSurface Treatment

MagnetsMagnets


25

Cobalt MarketCobalt Market


26

Cobalt Supply/Apparent Demand and Price 1980-2008

0

10000

20000

30000

40000

50000

60000

70000

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

Year

To

nn

es

$0

$5

$10

$15

$20

$25

$30

$35

$40

$45

US

$/l

b

SUPPLY DEMAND PRICE

Historical DataHistorical Data


27

COBALT PRICE 1995 - 2009(1st Quarter)

0

10

20

30

40

50

60

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

Years

US

$/l

b

High Grade Low Grade Spread

Historical Price DataHistorical Price Data


28 Cobalt ProductionCobalt Production

Nickel Industry ~ 50%

Copper Industry & Other ~ 35%

Primary Cobalt Operations ~ 15%

Current Production Sources:Current Production Sources:


29

Cobalt Consumption by End-Use 2008

19%

14%

7%4%27%

9%

10%

6% 4%

Superalloys

Hardmetal

Magnets

HS Steel

Batteries

Catalysts

Pigments

Organics

Other

Cobalt EndCobalt End--UseUse


30

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2001 2002 2003 2004 2005 2006 2007 2008

Year

ROW

Europe

China

Africa

Refined Production by Geographical AreaRefined Production by Geographical Area


31

0%

20%

40%

60%

80%

100%

2001

2002

2003

2004

2005

2006

2007

2008

Other

Oceania

Americas

China

Asia

Europe

Africa

Refined Cobalt Apparent Demand by RegionRefined Cobalt Apparent Demand by Region


32

Hazard

&

Risk

Hazard

&

Risk


33 Cobalt in the EnvironmentCobalt in the Environment

•• Naturally occurring cobalt is Naturally occurring cobalt is ““CobaltCobalt--5959””• Cobalt - 59 is NOT radioactive

•• Naturally occurring inNaturally occurring in: • Sea, Surface and Ground water

•• Other naturally occurring cobalt sourcesOther naturally occurring cobalt sources:• wind-blown dusts

• weathering of rocks and soils

• volcanoes, forest fires

• plants, continental and marine biogenic emissions.


34 Cobalt in the EnvironmentCobalt in the Environment

Environmental Compartments:Environmental Compartments:

•• Aquatic:Aquatic: Cobalt +2 (divalent) substances are toxic to many fresh-water plants and animals

•• Soils and SedimentsSoils and Sediments:: Compared to aquatic effects, divalent cobalt substances appear to be less toxic to soil and sediment organisms.

•• MarineMarine:: A data-poor area. The effects of cobalt substances are

currently being investigated by CDI-sponsored research projects.

NB: The compartmental effects of Co+3 (trivalent) substances are being investigated by CDI-sponsored research projects.


35 Human ExposureHuman Exposure

•• Naturally:Naturally: air, water, soils and dusts, plants and other foods.

•• ManMan--made sourcesmade sources:: burning of fossil fuels, sewage

sludge, phosphate fertilizers, mining/smelting of cobalt containing ores, industrial processing of cobalt substances and alloys,

consumer products.

•• Most significant source of exposureMost significant source of exposure toto cobalt cobalt

for the general public is from foodfor the general public is from food:: Humans

require Vitamin-B12 which contains cobalt as the central co-factor. Many animals and plants require small amounts of divalent cobaltdirectly for growth and vitality.


36 Cobalt in FoodCobalt in Food

(Source: Le Blanc et al, 2004)


37 Health EffectsHealth Effects

Results of over-exposure:

•• AsthmaAsthma

•• Allergic dermatitisAllergic dermatitis

•• Hard metal lung diseaseHard metal lung disease (Co+WC)

•• Occupational lung cancerOccupational lung cancer (Co+WC, limited

evidence?)

•• Thyroid hyperplasiaThyroid hyperplasia (goitre)

•• Heart effects including cardiomyopathyHeart effects including cardiomyopathy

•• PolycythemiaPolycythemia (> red blood cells)


38

CobaltCobalt

Oxide

Cobalt

Sulphide

Cobalt

Chloride

Cobalt

Sulphate

Cobalt

Carbonate

Cobalt

Nitrate

Cobalt

Acetate

CAS: 7440-48-4 CAS: 1307-96-6 CAS:1317-42-6 CAS: 7646-79-9 CAS : 10124-43-3 CAS: 513-79-1 CAS: 10141-05-6 CAS: 71-48-7

AND AND AND AND

Cobalt

Chloride

Cobalt

Sulphate

Cobalt

Nitrate

Cobalt

Acetate

6H20 7H20 6H20 4H20

CAS: 7791-13-1 CAS: 10026-24-1 CAS: 10026-22-9 CAS: 6147-53-1

(NO EINECS) (NO EINECS) (NO EINECS) (NO EINECS)

Physical Properties None None None None None None None None

Acute Oral None Xn; R22 None Xn; R22 Xn; R22 None None None

Acute Inhalation None None None None None None None None

Dermal Irritation R42/43 R43 R43 R42/43 R42/43 R42/43 R42/43 R42/43

Eye irritation None None None None None None None None

Dermal Sensitization R42/43 R43 R43 R42/43 R42/43 R42/43 R42/43 R42/43

Respiratory Sensitisation R42/43 None None R42/43 R42/43 R42/43 R42/43 R42/43

Chronic Toxicity None None None None None None None None

Reproductive Toxicity None None None Cat 2; R60 Cat 2; R60 Cat 2; R60 Cat 2; R60 Cat 2; R60

Mutagenicity None None None Cat 3; R68 Cat 3; R68 Cat 3; R68 Cat 3; R68 Cat 3; R68

Carcinogenicity None None None Cat 2; R49 Cat 2; R49 Cat 2; R49 Cat 2; R49 Cat 2; R49

Aquatic Environment R53 N; R50-53 N; R50-53 N; R50-53 N; R50-53 N; R50-53 N; R50-53 N; R50-53

Indications of Danger Xn; R42/43 Xn, N Xi, N T, N T, N, T, N, T, N, T, N,

S-Phrases2, 22, 24,

37, 61

2, 24, 37,

60, 61

2, 24, 37,

60, 61

53, 45, 60,

61

53, 45, 60,

61

53, 45, 60,

6153, 45, 60, 61 53, 45, 60, 61

Endpoint

EINECS: 231-

158-0

EINECS: 215-

154-6

EINECS: 215-

273-3EINECS: 200-755-8

EINECS: 231-

589-4

EINECS: 233-334-

2

EINECS: 208-

169-4EINECS: 233-402-1

Classification and LabellingClassification and Labelling

R22 = Harmful if swallowed

R43 = May cause sensitization by skin contactR42/43 = May cause sensitisation by inhalation and skin contactR49 = May cause cancer by inhalationR50/53 = Very toxic to aquatic organisms, may cause long term adverse effects

in aquatic environmentR53 = May cause long term adverse effects in the aquatic environmentR60 = May impair fertilityR68 = Possible risk of irreversible effects

S2 = Keep out of the reach of childrenS22 = Do not breathe dustS24 = Avoid Contact with skinS37 = Wear Suitable GlovesS45 = In case of accident or if you feel unwell, seek medical

advice immediately (show the label where possible).S53 = Avoid exposure – obtain special instructions before use.S60 = This material and its container must be disposed of as hazardous waste.S61 = Avoid release to the environment. Refer to special

instructions/safety data sheets

Xi = Irritating

Xn = HarmfulT = ToxicN = Dangerous for the environment

Note: Substances

with specific effects on human health (see chapter 4 of Annex VI of Directive 6715481EEC) that

are classified as carcinogenic, mutagenic, and/or toxic for reproduction

in categories 1 or 2 are ascribed Note E if they are also classified as very toxic (T+), toxic (T) or

harmful (Xn). For these substances, the risk phrases R2O, R21, R22, R23, R24, R25. R26. R27, R28,

R39, R68 (harmful), R48 and R65 and all combinations of these risk phrases shall be preceded by the word

'Also'.


39

•• European UnionEuropean Union• REACH – Self classification (hazard) and risk assessment

• Toys – no CMR substances

• Water Framework Directive – Co under consideration to be a priority substance (if a priority substance then progressive emission reduction required)

•• North AmericaNorth America• Canada – Domestic Substance List evaluating environmental and

health effects. Also at Provincial level: Toxics Reduction Law (Ontario)

• USA: EPA – Safe Drinking Water Act. Also State Proposition 65 (California) relating to ‘safe’ levels of substances (Cobalt included!)

•• AsiaAsia• Japan – Currently undertaking human health risk assessment

• China – Implementing chemical regulations

Need for the industry to respond proactively!

International RegulationInternational Regulation


40 Hazard Versus RiskHazard Versus Risk

Strict Control!


41 CHALLENGESCHALLENGES

• Market Factors

• Hazards and Risks

• Regulation


TRENDS IN NICKEL-COBALT PROCESSING

By

Alan Taylor

ALTA Metallurgical Services

Presented by

Alan Taylor

[email protected]


2

RECENT INDUSTRY TRENDS

General slow down due to financial situation.

• Some production taken off line or reduced.

• Some new projects slowed down.

• Some new projects put on hold.


3

INDUSTRY TRENDS - LATERITES - PAL

• Two PAL plants closed – Cawse and Ravensthorpe (BHPB reviewing future of Yabulu operation which was to receive MHP from Ravensthorpe).

• Vale Inco’s Goro operation in New Caledonia due on stream by the end of the 2009.

• Expansion of Sumitomo’s Coral Bay (Rio Tuba). operation in Philippines due on line 2009.

• Sherritt’s Ambotovy Project construction slowed, due to be completed by end of 2010.

• MCC/Highlands Pacific Ramu Project in PNG due on stream at end of 2009.

• Expansion project for Moa Bay slowed.

• Vale Brazil’s Vermelho Project on hold.

• Gladstone Project in QLD has received approval for EIS and are seeking financing to proceed.


4

INDUSTRY TRENDS – LATERITES – HEAP LEACHING

• Minara at Murrin Murrin in Western Australia are progressively extending their successful scats heap leaching operation to process ore.

• European Nickel have received their final operating permit for their heap leaching project in Turkey, and are targeting start of construction in second half of 2009.

• European Nickel/Rusina developing Acoje Project in

Philippines. Heap leach trials at site planned for 2009.

• NORNICO, QLD, Australia, Metallica Minerals Limited. Feasibility study program deferred.


5

INDUSTRY TRENDS – LATERITES - SMELTING

• Vale Onça Puma ferronickel project, Brazil, slowed by

at least one year, now due on stream in January 2010.

• Koniambo ferronickel project SMSP/Xstrata due on line

2011.

• Large reduction in nickel pig iron production in China due to lower metal price.


6

INDUSTRY TRENDS – SULPHIDES - HYDROMET

• Construction of Voisey’s Bay pressure oxidation

operation due to start in 2009 and be complete 2013.

• Tati Nickel Activox Project in Botswana suspended by

Norilsk.

• Talvivaara heap bioleaching operation in Finland now

on stream.

• PolyMet Mining is in the late stages of the

environmental review process for their NorthMet Project in Minnesota USA using the PLATSOL™ pressure

oxidation process.


7

WHAT ARE THE OVERALL EFFECTS?

• Fundamentals are largely unchanged.

• Nickel demand is likely to increase again due to consumption by China, India and other developing countries.

• Most of future new nickel will have to come from laterites.

• Environmental issues will continue to promote hydromet for new sulphide projects.


8

LATERITE PROCESSING DEVELOPMENTS

• Most new projects will have to process ore too low grade and with unsuitable mineralogy for smelting.

• Thus they will have to use some form of leaching process.

• Commercially applied processes are:

- PAL

- PAL with an associated AL circuit

- Heap leaching

- Reduction roast-ammonia leach (Caron) process


9

LATERITE PROCESSING DEVELOPMENTS (CONT.)

PAL:

• Currently selected for all large laterite leaching projects.

• Experience now gained from a number of commercial operations.

• The least sensitive to mineralogy.

• Sensitive to sulphuric acid price.

• Relatively insensitive to climatic conditions


10


PAL with associated AL:

Compared with PAL only –

• Extends operation to economically process higher acid consuming portions of deposit.

• Reduces consumption of neutralizing agent.

• Reduces overall capex per unit of nickel production.

• Reduces PAL equipment requirements.


11


Heap leaching:

• Lower capex.

• Potentially more suitable than PAL for small and medium size projects.

• Can be applied as a satellite to a PAL operation to treat suitable portions of the deposit.

• Generally has higher acid consumption than PAL, which increases opex and exposure to acid price.

• Lower metal extractions and slow leach kinetics.

• Relatively sensitive to mineralogy.

• More sensitive to climatic conditions than PAL.

• Limited commercial experience to date.


12


Reduction roast-ammonia leach:

• Proven technology.

• Mild operating conditions: reduction < 800C, high pH solution leach not aggressive.

• High energy required for ore drying.

• Nickel recovery only moderately high.

• Cobalt recovery relatively low.

• Sensitive to ore mineralogy.

• No recent new plants.


13

OPPORTUNITY FOR NEW LATERITE PROCESSES

Ideal specifications::

• High recoveries of nickel and cobalt.

• Lower capex and opex than existing processes

• No initial energy intensive drying step.

• Atmospheric pressure operation.

• Low net reagent consumption.

• Ability to provide separate nickel and cobalt products.

• Suitable for large and small projects.

• Low to moderate corrosivity.

• Low environmental impact.


14

PROCESSES UNDER DEVELOPMENT

• Atmospheric sulphuric acid leaching (AL)

• Sulphation atmospheric leach (SAL Process)

• Atmospheric chloride leaching


15

AL - POTENTIAL ADVANTAGES

• Low capex than PAL.

• Less aggressive leach conditions than PAL.

• Lower maintenance cost than PAL.

• Atmospheric pressure operation.

• Simpler operation than PAL.

• Higher on-stream availability – individual leach tanks can be by-passed.


16

AL - CHALLENGES

• High acid consumption.

• Longer retention times than PAL.

• Need to process leach solutions with high impurities, especially iron.

• Lower recoveries than PAL.

• Aggressive conditions – though less so than PAL.


17

AL - STATUS

• No stand alone commercial operations as yet.

• Operated at Ravensthorpe for treating saprolite ore in parallel with PAL for limonite (EPAL Process). Benefits from availability of “free” acid from PAL discharge.

• A number of projects using sulphuric acid are at various stages of development.


18

SAL PROCESS

• Skye Resources, Canada, have piloted a sulphation atmospheric leach process (SAL Process) at SGS Lakefield as a possible option for the Fenix Project in Guatemala.

• Features initial sulphation of limonite ore by pugging with strong sulphuric acid, which generates 140C temperature - thus accelerating leach kinetics. Iron forms ferric sulphate.

• Second stage consists of adding crushed saprolite ore and water and grinding in ball mill, then leaching for about 24 hrs in series of agitated tanks with temp. maintained at 95-105C.

• Limestone is added and iron is hydrolyzed to ferric hydroxide. Acid is generated which leaches saprolite. SO2 is added to reduce cobalt and enhance extraction.


19

SAL PROCESS (SGS PRESENTATION AT ALTA 2007)

ROM Crushed

Limonite Saprolite

H2SO4

H2O

SO2

CaCO3

Wash Leach Residue

SIR/SP Solids

recycle

CaCO3

MgO MHP Solids

Ca(OH)2

Ca(OH)2 MR Solids

MgSO4 liquor

Leach Slurry

Preparation (LSP):

Limonite Sulphation

Saprolite Grinding

Upstream Circuit:

Atmospheric Leaching (AL)

Primary Iron Removal (PIR)

Counter Current Decantation (CCD)

Downstream Circuit:

Secondary Iron Removal (SIR)

Mixed Hydroxide Precipitation (MHP)

Scavenger Precipitation (SP)

Manganese Removal (MR)


20

SAL PROCESS (CONT.)

• Pilot plant achieved nickel and cobalt leach extractions of 85-89%.

• Acid consumption 600 kg/t (for 9-10% Mg content)

• Downstream process was conventional MHP circuit.


21

ATMOSPHERIC CHLORIDE LEACH

• Chloride leaching tested for the Young Laterite Project for Jervois Mining, NSW, Australia, using magnesium chloride-HCl lixiviant.

• Iron is precipitated as hematite and lixiviant regenerated.

• Tests indicate high pulp filtration rate allowing use of belt filters, which is much more favourable than sulphuric acid leach residue solid-liquid separation

• Chloride leaching has also been tested by Chesbar/ Jaguar Nickel in Canada and Intec in Australia.


22

CHLORIDE LEACH

FLOWSHEET TESTED FOR

YOUNG PROJECT

(BRYN HARRIS

PRESENTATION AT ALTA 2006)


23

PREVIOUSLY PROPOSED LATERITE PROCESSES

Numerous processes proposed over years involving

pyrometallurgy, hydrometallurgy, vapour metallurgy and combinations including:

• Segregation roasting

• Sulphation roasting

• Republic Steel HSO - HTCP Process

• Aqueous chlorination


24

PREVIOUSLY PROPOSED LATERITE PROCESSES (CONT.)

• Nitric acid leaching

• Carbonyl extraction (Inco)

• Sulphur dioxide leaching

• Reduction roast - sulphuric acid leaching

• Gas phase sulphation with SO2/air - water leach

• Submerged lance smelting (Ausmelt)


25

FIELD STILL WIDE OPEN

No one has come up with a

real winner to date.

So opportunity still KNOCKS


26

FURTHER DEVELOPMENTS OF PAL

In the meantime possible further developments of PAL include:

• Flotation for upgrading some saprolitic ores.

• Treatment of blended laterite/sulphide feed for generating acid and heat in autoclaves.

• Extraction and reuse of residual leach acid by SX/IX/membranes.

• Development of more selective SX extractants, including synergistic mixtures.

• Application of RIP to reduce the number of CCD thickeners (or eliminate totally) and reduce soluble metal losses.


27

SULPHIDE PROCESSING DEVELOPMENTS

Present trend in hydromet processing appears to be:

• Pressure oxidation for concentrates.

• Heap bio-oxidation for low grade ores.

Other possibilities include:

• Tank bioleaching.

• Chloride leaching.


28

VOISEY’S BAY PRESSURE OXIDATION FLOWSHEET(FROM PUBLISHED PROJECT EIS)


29

TALVIVAARA HEAP BIOLEACHING FLOWSHEET(

(PAPER AT ALTA 2007)


30

FURTHER DEVELOPMENTS FOR SULPHIDE FLOWSHEETS

Additional developments could include:

• Treatment of blended laterite/sulphide feed for generating acid and heat in autoclaves.

• Application of more selective SX extractants, including synergistic mixtures.


Identifying opportunities to reduce the consumption of energy

across mineral processing and recovery plants.

David Sterling

Manufacturing Execution Systems

Process and Machine Business

Schneider-Electric

Brisbane, QLD 4009

[email protected]

http://www.schneider-electric.com

Abstract

In addition to focusing on meeting Government Policy on Energy

Efficiency Opportunities (EEOs), mining and mineral processing

companies are trying to reduce energy use to reduce costs in the current

financial conditions. One of the major issues with EEOs is that there is

often insufficient data available on energy use, and more importantly the

energy events linked to the energy use, to identify opportunities to reduce

energy use.

As well as energy reduction, mining and mineral processing companies

often struggle with the prediction of energy use, and are often penalised

for under or over forecasting. Once again it is the lack of timely

information that makes this prediction difficult.

This paper looks at expanding the use of a Manufacturing Execution

Systems, by integrating with Energy Solutions. This will provide

automatic, timely information, at a granularity that makes it easier to

identify EEOs, reduce energy costs, and better predict energy use in a

mining and mineral processing operation.

1. Energy Efficiency Opportunities (EEO)

Both the financial and atmospheric climate are key discussion points in 2009. Mining and

mineral processing companies are looking at both reducing energy use to lower costs, and

to reduce emissions, particularly in light of any carbon emissions scheme that may be

introduced. To achieve this, companies must have a clear understanding of their current

energy use, and therefore require tools that will empower people to make decisions.


The Australian Government has recognised the potential for energy efficiency,

particularly in the mining sector. In recent years they introduced the Energy Efficiency

Opportunity (EEO) program. The Department of Resources, Energy and Tourism

administers the EEO program, which encourages large energy-using businesses to

improve their energy efficiency by identifying, evaluating, and reporting publicly on cost

effective energy savings opportunities.

Participation in Energy Efficiency Opportunities is mandatory for corporations that use

more than 0.5 petajoules (PJ) of energy per year. There are approximately 210

corporations registered for the Energy Efficiency Opportunities program.

One of the reasons behind the EEO program is that the Australian Government believe

that increased uptake of cost-effective energy efficiency technologies and process, will

help Australian businesses maintain competitiveness under a Carbon Pollution Reduction

Scheme, as Australia moves towards a low carbon economy.

As shown below in the Rio Tinto Diagram (from a Rio Tinto presentation at an EEO

forum) there is a close relationship between emissions and energy use.

Figure 1: Rio Tinto’s link between Energy and Emissions (How Rio Tinto Addresses

Energy Management. Energy efficiency opportunities –22nd

May, Melbourne, 2008.)

Current Examples of EEOs in Australian Mining and Minerals Processing

As part of the EEO Program in Australia, large energy users must complete EEO reports.

These are posted publically on the company web sites. There are also links to the

company web sites and the specific EEO reports on the Department of Resources, Energy

and Tourism web site.


One of the issues with EEO reporting is that there is an increasing “sameness” about the

type of EEOs that mining companies are starting to identify in their operations.

For example, a sample of common EEOs are:

• Reducing the lighting (lighting control systems)

• Reducing the energy used in air conditioning by changing the air conditioning

• Changing to variable speed drives to reduce energy

• Reducing energy related to compressed air leaks

There is not an issue with the four opportunities suggested above. They are indeed quick

win opportunities in the mining and mineral processing environment. The question in the

next few years will be where to go next to reduce energy use.

Figure 2: Extracts from BHPB, Xstrata and Rio EEO reports showing “lighting

reductions” as a possible EEO.

We believe that the operations teams will play a major role in the next level of energy

reduction. Operations teams will be challenged to achieve the target recoveries at the

lowest energy consumption. The next level of energy savings will come from process

changes and operations management changes, rather than more energy efficient drives or

lighting changes.


2. Energy Forecasting and Events of Excess Energy Use

Introduction

In discussions with Australian mining and mineral processing companies it became

apparent that they are facing common problems as they attempt to do more using less

energy. The challenge for mining and mineral processing companies is that they currently

don’t have sufficient information to make decisions to reduce energy (identify EEOs), or

forecast energy use accurately.

Complex energy forecasting tools are available to better forecast energy use, but if they

don’t take into account the context of excess energy use, or the historical and future

production data, they will not be accurate.

Mining and mineral processing companies are investing in meters and enterprise Energy

Management (EM) software to monitor and visualise/report on energy use. Providing

information on energy use is only partially solving the problem of determining new

opportunities to reduce energy use, and in turn comply with EEOs. Along with energy

use there needs to be information on what is actually happening in the plant at the time, in

other words there has to be context supporting the energy use. For example near real time

information like kWh/tonne or kWh/ounce of production.

Mining and mineral processing companies tend to have information systems that provide

production and delay accounting (downtime) information, but do not combine or

integrate these systems with energy systems to provide more useful information,

providing accuracy to forecast models and perhaps new EEOs. Production and downtime

are obviously key factors in energy use, therefore integration of these systems with

energy systems have potential benefits.

The integration of Energy Management (EM) with Manufacturing Execution Systems

(MES) can be given the general term Plant Energy Optimisation (PEO).

The key problem that these companies face is timely access to accurate energy

information. They either:

• do not have the information available at all

(example: either it is not captured, or a manual process and not available to

everyone)

• the information is not granular enough, or

(example: there is insufficient power information for the entire facility)

• provide information that is isolated and has no context in with which it was used.

(example: energy figures do not link to what was actually happening in the plant)


Added to this is the lack of ability to accurately forecast energy use. Under forecasting or

over forecasting energy use often results in financial penalties from utility companies or

dedicated energy providers.

There is certainly sufficient evidence to support the capacity for mining and mineral

processing operations to reduce their energy use. The two following graphs from the US

Department of Energy certainly indicate the opportunity exists.

Figure 3 looks at the top 10 energy intensive processes in coal, minerals and metals. It is

no surprise that grinding is the largest energy intensive process.

Figure 4 is interesting in that it looks at the potential for energy savings for the coal,

metals and mineral industries within the USA. The key observation from the graph is that

where coal and metals have the potential to reduce their energy use by 17% and 21%

respectively due to best practice, minerals has the potential to reduce it by 27%.

Figure 3: Energy-Savings Opportunity in US Mining Industry for Top 10 Energy-

Intensive Processes


Figure 4: U.S. Mining Industry Energy Bandwidth for

Coal, Metal, and Mineral Mining (Source: Mining Industry

Energy Bandwidth Study, June 2007 by the U.S.

Department of Energy)

A similar conclusion was made in a 2005 report by the Natural Resources of Canada

where they benchmarked the energy consumption of open cut iron ore and gold mines.

In investigating iron ore in an open cut mine, it was reported that there was potential to

reduce the mining costs (based on energy savings) by 36% and milling costs by 47%. For

gold mining this estimate was reduced to a staggering 53% for milling.

Examples of Energy Events

If an MES is already collecting information on downtime of large mining assets, along

with production figures, then it is efficient to link or integrate this with energy

information to provide useful “energy eventing” information that can be easily reported

or visualised together. See figure 5.

There are many reasons why an increase in mill energy used could spike. For example if

the same throughput is required within a mill, and the feed material is a harder or more

competent ore, energy use can increase by over 10%. This information about ore

competency could be available historically in the MES.

Similarly if the grind size had to reduce by half to maintain recovery rates during

flotation (due to liberation issues with an ore), energy will increase in excess of 10%,


more likely to be a factor of four. Once again this information about liberation changes

could be available historically in the MES.

In both these cases the integration of MES with EM systems or Plant Energy

Optimisation (PEO), could provide visualisation in near real time with such parameters as

energy/tonne, cost/tonne. PEO can also visualise downtime with energy information and

allow drilling down of information to see the reasons. Therefore delivering context with

energy use. The real advantage comes with energy events, such as sudden events where

more or less than normal energy is used, without a reason.

Plant Energy Optimisation systems support:

1. Automatically capturing events:

a. Start and end time, duration, excess energy used

b. Plus context information – material, product, grade, crew, shift,

c. Automatically or manually split events

2. Knowing when your equipment is consuming too much energy and by

how much

a. When demand (kW) is over a target

b. When kWh/ton is over target

Figure 5: Energy Events Graph

Figure 5 illustrates an “energy event” where over consumption of energy

occurred. The MES captured context around this event and poor feed was

attributed to the over consumption of energy.


Energy Forecasting

When failing to meet a forecast (either over or under) incurs monetary fines, tracking

energy consumption against forecasts becomes very important. Mining and mineral

processing companies need to not only be able to accurately forecast energy

consumption, but to have timely access to actual energy consumption so that decisions

can be made.

Sophisticated mechanisms for modeling energy consumption are needed to

generate a forecast. These models are available as part of an Energy

Management solution and use regression and correlation based on ASHRAE

Guideline 14. They assist in identifying the factors/drivers that affect energy

consumption, and then develop algorithms that can forecast energy

consumption. To do this type of modeling more effectively requires the

production and downtime data that an MES provides.

Figure 6: Predicted Energy Use using Regression and Correlation Model

By integrating MES and EM systems, the energy model can easily be derived from

historical process data, with the MES providing future process data like production plans,

shift targets, and upstream feed grade.

3. Recommended Solution

Importance of Operations Data (already captured in MES)

The importance of the data already captured in the MES cannot be under estimated. The

value lies in linking it with energy data. Many vendors are looking at providing energy

solutions; however the improved value proposition is not having an MES and an EM

system, but an integrated Plant Energy Optimisation solution.


Recommended Approach

By combining automatically captured data on production and downtime in the

MES, and energy use in the EM system delivers business benefits. To

maximise the benefits it is suggested that the information be not only

combined but also displayed together. This type of integrated solution we

have already referred to as called Plant Energy Optimisation.

The best approach would be for the Plant Optimisation Solution to enable:

● Real time Energy Consumption Reporting

● Forecasting of energy consumption based on certain

parameters

● Establishment of optimal energy consumption target for each

section of an operation

● Identification and quantification of all consumption above the

target

● Discovery of root causes of over consumption

● Report on shift or daily consumption and over consumption

events

● Real time calculation of sustainability KPIs such as kWh/tonne

● The provision of validated actual data to justify future capital

expense and/or process changes

To systematically, reduce energy consumption in the minerals processing environment

we recommend the following process:

● Identify the energy drives for your plant. E.g. material grades,

recovery rates etc

● Report in real –time the energy consumption as well as the

energy drivers

● Use an energy model to forecast the energy consumption based

on the forecasts of your energy drivers

● Use this forecast to determine an energy target

● Identify and qualify all consumption above the target

● Analyse results to determine root causes of overconsumption

Some over consumption may be caused by failing equipment and could be

another trigger used to optimise maintenance programs.

Real-time metrics can also help to drive behavior. KWh/ounce can be easily

calculated and displayed so that operators know how the plant is performing

in terms of production and energy.

By collecting accurate records on the causes of energy consumption above

target, this data can be used to support capital expenditure required to change

process equipment.


Figure 7: Proposed architecture for integration of MES with Enterprise

Energy solution.

Potential Energy Savings

In mining operations there is a need for real-time reporting and analysis. As with delay

accounting, there is a requirement to empower operations people by providing the

information in a form that they can easy interpret and act upon on a daily basis.

Operations people know their plant and can help identify EEOs if they have timely

information.

In looking at Plant Energy Optimisation in the MMM industries, accurate, timely

information is needed to create opportunities to reduce energy use. The US research

illustrated graphically in Figures 3 and 4, is a great place to start as it identifies some

areas for energy savings through best practice.

Better information will make it easier for mining companies to identify these potential

savings.


Forecasting Energy Use

Improved forecasting of energy use can provide two advantages:

1. Reduction in penalties for over or under forecast (usually based upon a shared

grid)

2. Reduction in the gap between the provided power from a power plant, and the

actual used (dedicated power plant), often called the “spinning excess”

1. Improved Energy Forecasting to Reduce Costs

• Using forecasting algorithms, and by combining the energy information with

production or downtime information, produce more accurate forecasts of energy

use. This supports reductions in penalties for under or over forecasting.

• Maximise the use of existing power infrastructure capacity and avoid

overbuilding power infrastructure by having more accurate information.

2. Forecasting and Understanding Energy Use to Reduce the Spinning Excess Created by

the Power Plant

The example of a mining and mineral processing site that requires approximately 1.8

MW of power, where the dedicated power plant (run by a 3rd

party) produces 2.7MW and

therefore a spinning excess of 0.9 MW. The company only pays for the 1.8MW but has to

pay for maintenance and fuel for the 2.7MW.

Figure 8: Reducing Spinning Excess Using Plant Energy Optimisation Example

If sufficient information was known about why the peak loads occur (for example the

starting sequence for the mills), the spinning excess could be reduced, therefore reducing

the fuel and maintenance costs.

Energy events may also be the reason for the size of the spinning excess. A better

understanding of their root cause and prevention may also support the reduction of costs

associated with spinning excess.


4. Conclusion

Manufacturing Execution Systems (MES) are commonly used in mining and mineral

processing companies, reporting key data and KPIs around production and downtime.

Their information empowers people to make decisions that may improve business

excellence.

Energy Management (EM) solutions are increasingly being installed in mining and

minerals processing to provide information for tracking operational consumption,

forecasting consumption and providing information around energy quality. Like MES,

the information provided is useful for better understanding energy consumption. The

information empowers people to make decisions that may improve business outcomes by

reducing energy use or better forecasting energy use.

Both MES and EM systems have their use in mining, by providing automated, accurate

information in a timely manner. It is however the integration of both systems that will

provide more value when endeavors are made to better understand energy use, and

therefore potentially reduce energy use. The integration of MES and EM solution to form

a Plant Energy Optimisation (PEO) provides added value.

Using PEO energy events can be identified, the reasons for them understood, and perhaps

preventative action put in place to reduce or stop them recurring as frequently.

PEO providing context with abnormal energy events are critical to understanding energy

use and over/under performance. It is also useful in providing better forecasting in the

high energy mining and mineral processing environments. In isolated mining

environments a potential quick win on reducing energy could be found by using better

forecasting and understanding of energy use to reduce the spinning excess of a power

plant.


References

[1] How Rio Tinto Addresses Energy Management. Energy efficiency opportunities –22

nd May, Melbourne,

2008.

[2] EEO Public Report 2007 Rio Tinto Ltd.

[3] Xstrata Holding Pty Ltd Energy Efficiency Opportunities, Annual Public Report, December 2008.

[4] BHPB EEO Report 2008.

[5] EEO Public Report Gold Fields Australia Pty Ltd, 2008.

[6] Mining Industry Energy Bandwidth Study, June 200. The U.S. Department of Energy: Industrial

Technologies Program

[7] Benchmarking the Energy Consumption of Canadian Open-Pit Mines. Natural Resources of Canada,

2005.

[8] Benchmarking the Energy Consumption of Canadian Underground Bulk Mines. Natural Resources of

Canada, 2005.


1

Production of Nickel Pig Iron in China

Xian Jian Guo， You fu Shi

Ramu NiCo Management (MCC) Ltd

A-25th Floor, Beijing Global Trade Center, 36 North Third Ring East, Beijing, China 100013,

[email protected]

Abstract

This paper introduces the production of nickel pig iron in China including processes

and economical analysis. The production of nickel pig iron boomed in the past several

years in China. In 2007, production output of nickel in pig iron reached its peak,

109,000 tonnes while nickel mined in China was 104,000. In 2008, it decreased to

61,000 tonnes of nickel in pig iron. The processes of producing pig iron are blast

furnace smelting and electric furnace smelting. The volumes of blast furnaces are

from 100 cubic meters to 500 cubic meters and the power of electric furnace is from

5 MGW to 20 MGW. The laterite ore is mostly imported from Philippines, Indonesia

and New Caledonia. The cost mainly depends on the price and grade of imported ore.

The other factors affecting cost are the prices of coke and electricity. For blast

furnace and electric furnace plants that treat ore with 1.8-2.0% Ni content, the cash

cost is respectively 800 US$ and 750 US$ per tonne of product containing 6-7%Ni.

Based on the data from operating plants, the total cost estimated is respectively 960

US$ and 900 US$ per tonne of product. The price of pig iron containing 6% nickel was

2000 US$ per tonne in 2007 and dropped to 800 US$ per tonne at the end of 2008.

With nickel pig iron price at 900 US$ per tonne, there is little profit for nickel pig iron

producers. This paper also compares the cost between pig iron production and the

hydrometallurgical process of laterite treatment -HPAL process.

1. Introduction

The production of pig iron containing nickel boomed in the past several years in

China. Nickel output in pig iron reached 109,000 tonnes at its peak in 2007. It

surpassed the nickel mined (104,000 tonnes) in China. However, it decreased to

61,000 tonnes in 2008 because nickel price went down. The main reasons for the

booming of nickel pig iron production in China are as follows:

Rapid increase in nickel demand in recent years because of massive increase

in stainless steel production;

Limited domestic nickel resources and heavily reliant on the import of nickel;

Surging nickel price in recent years.

Figure 1 shows nickel mined and consumed in the past years in China, based on the

statistics published by Chinese Nonferrous Metals Industry Association [1]. As shown


mailto:[email protected]

2

in Figure 1, the gap between nickel consumed and mined becomes larger and larger

since year 2000. In 2005, the difference of nickel consumed and mined was 115,000

tonnes, but reached 220,000 tonnes in 2008. The demand for nickel from stainless

steel industry creates a huge market for nickel pig iron.

Figure 2 Indicates nickel price in the past ten years [1]. Nickel price started to rise

rapidly in 2005 and reached its peak in 2007. Figure 3 shows nickel contained in pig

iron from imported laterite ore [2]. As shown in Figure 3, laterite ore was imported to

make nickel pig iron from Philippines in 2005 and began to import from Indonesia

and New Caledonia in 2006. It reached 109,000 tonnes of nickel contained in pig iron

in 2007. As the price of nickel began to drop in 2008, the imported ore was reduced

with the decrease of nickel pig iron production. It is known from Figure 2 and 3 that

the production of nickel pig iron is based on nickel price, thus surging nickel price is

the main cause for the booming nickel pig iron production.

The other cause for nickel pig iron production growing in China is that many small

blast furnaces have been shut down since 2000 due to high energy consumption and

environment issues. Without much capital cost, these blast furnaces could be easily

restarted up to produce nickel pig iron. This advantage attracted many small private

companies to get into nickel pig iron production business. There were totally more

than 300 plants producing nickel pig iron at the end of 2007 and their production

capabilities varied from hundreds of tonnes to thousands of tonnes of nickel. Most of

these plants were located in Shanxi and Shandong provinces with plenty of coal

supplies, Sichuan province with the advantage of cheap power and Zhejiang and

Jiangsu provinces that have large consumers.

Figure 1, Chinese nickel mined and consumed


3

Figure 2, Historical nickel price

Figure 3, Nickel in pig iron from imported laterite ore

2. Process

There are two processes to make nickel pig iron in Chinese nickel pig iron plants. One

is blast furnace smelting and another is electric furnace smelting. Figure 4 shows the

simplified flowsheet. Electric furnace smelting (RKEF) is widely used to treat saprolite

type of laterite ore and is a popular process in ferronickel production. The operation

of electric furnace smelting includes ore preparation, dry, reduction roasting and

smelting. The electric furnaces used in China are small and the operation is also

simplified. The power of furnaces ranges from 5 MGW to 20 MGW. Some plants do

not have refinery to produce final ferronickel. They produce crude alloy that is similar

to nickel pig iron.


4

The blast furnace smelting was used in making ferronickel one hundred years ago.

The operation was almost the same as iron making, including dry, sintering and

smelting. The blast furnaces used in nickel pig iron making are small. Their volumes

were from 100 cubic meters to 500 cubic meters. Blast furnace nickel pig iron

smelting was patented [3]. The main improvement was in slag chemistry to increase

nickel recovery and operate furnace smoothly.

The operating parameters of the processes are shown in Table 1. The operating

temperatures of furnace are respectively 1450-1550 C and 1500-1550 C for blast

furnace and electric furnace. The nickel recovery is more than 80% for a blast furnace

and more than 90% for an electric furnace. The coke rate is 200-250 kg/t.ore for blast

furnace and 80-100 kg/t.ore for electric furnace. The fluxes of electric furnace

smelting include silica and lime, and fluxes added in blast furnace smelting are mostly

limestone, fluorite and dolomite.

Laterite ore

Pig Iron Crude ferronickel (Pig iron)

Figure 4, Simplified flowsheet of pig iron making

Table 1, Operation parameters of blast furnace and electric furnace*

Parameters Blast furnace Electric furnace Note

Operating Temp. C 1450-1500 1500-1500

Ni recovery, % >80 >90

Coke, Kg/t.ore 200-250 80-100

Limestone, kg/t.ore 30-50 CaO>50%

Lime, Kg/t.ore 200-220 CaO>85%

Silica, Kg/t.ore 30-50 SiO2>95%

Fluorite Kg/t.ore 100-120 CaF>85%

Dolomite，kg/t.ore 10-30 MgO>80%

Power, Kwh/t.ore 120 550

*Producing 5-7% Ni pig iron

Dry Dry

Sintering

Blast furnace smelting

Reduction roasting

Electric furnace smelting


5

Table 2 and 3 show respectively the composition of nickel pig iron making by blast

furnace and electric furnace at different periods with feed ore of Ni: 1.68%; Fe:

16.36%; P: 0.035%; S: 0.10%; SiO2: 39.00%; CaO: 4.09: MgO: 19.09% [4]. It can be

seen in Table 2 and 3 that the composition of pig iron making using two different

processes is similar. The nickel content in the products has more than 6% Ni, and the

average content of nickel in electric furnace reaches 8.08% Ni. The product of blast

furnace has higher silicon and carbon content, and phosphor content is slightly

higher. Using these products to make stainless steel causes the problem of too high

phosphor content. Thus, refining process is installed in some electric furnace plants

to remove phosphor and other impurities.

Table 2, The content of pig iron produced by blast furnace.

No. Ni% Si% C% P% S%

1 6.08 2.60 4.50 0.070 0.090

2 7.14 2.58 4.30 0.067 0.080

3 6.23 2.63 3.20 0.067 0.080

4 7.00 2.60 3.00 0.050 0.056

5 7.20 2.57 4.60 0.068 0.072

6 6.84 2.64 4.00 0.068 0.075

Average 6.72 2.60 3.93 0.065 0.075

Table 3, The content of crude ferro-nickel produced by electric furnace.

No. Ni% Si% C% P% S%

1 7.76 4.65 2.98 0.054 0.070

2 8.32 5.37 2.36 0.104 0.205

3 8.15 4.48 3.38 0.051 0.036

Average 8.08 4.84 3.24 0.056 0.053

3. Economical analysis

The high profit is the impetus for so many companies to get involved in nickel pig iron

production. The profit mainly depends on the prices of pig iron product, imported

ore, coke and electricity. Figure 5 shows the price of nickel pig iron containing 6% Ni

and 8% Ni in year 2008[2]. As shown in Figure 5, the price was 2550 US$ per tonne of

pig iron containing 8% NI in January 2008 and dropped to 1040 US$ per tonne in

December. The price of pig iron with 6% Ni was 1830 US$ per tonne at the beginning

of 2008 and dived to 830 US$ per tonne at the end of 2008. The iron value in the

alloy was included in the prices.

Figure 6 shows the imported nickel laterite ore price for different grades at Chinese

harbor during 2008[2]. The data in Figure 6 show the same trend of the prices as

those of pig iron. For ore of 1.8-2.0% Ni, the price dropped from 145 US$ per tonne


6

in January 2008 to 29 US$ per tonne in December 2008. As for ore of 1.5-1.7% Ni, the

price dropped from 116 US$ per tonne to 25 US$ per tonne during 2008.

Figure 5, Price of nickel pig iron in 2008

Figure 6，Imported laterite ore price in 2008

The estimated cash cost based on the operating data of some plants for blast furnace

smelting and electric furnace smelting to produce nickel pig iron of 6-7% Ni is shown

in Figure 7, 8, 9 and 10. The prices of materials consumed in the estimate are listed

in Table 4. The labor and other plant costs are the actual operating data of some

nickel pig iron plants and are respectively 45 US$ and 55 US$ per tonne of nickel pig

iron for blast furnace plant and electric furnace plant. In the estimation, ore price

changes with ore grade. The basic price of ore is the price of ore less than 1.6% Ni

content. For ore of 1.6-2.0% Ni, the factor is 1.1 of the basic price while it is 1.2 of the

basic price when ore grade is higher than 2.0% Ni. As shown in Figures7, 8, 9 and 10,


7

although the price of ore increases with the increase of nickel content in ore, the cost

decreases with the increase of ore grade in general. However, high grade ore is

limited in market. Most of ore treated in pig iron plants is in the range of 1.5-2.0% Ni.

Table 4, Materials price in the estimate

Item Ore Coke Limestone Lime Fluorite Power, US$/kwh

US$/t 25 230 10 25 120 0.08

Figure 7 shows the estimated cash cost per tonne of pig iron at different laterite ore

prices for blast furnace plant. As indicated in Figure 7, the cost increases with the

increase of ore price. For the same grade ore, for example, 1.6% Ni, the cost is 825

US$ per tonne of pig iron with ore price at 15 US$ per tonne and 1134 US$ with ore

price at 60 US$ per tonne.

Figure 8 shows the estimated cash cost per tonne of pig iron at different coke prices

for blast furnace plant. As show in Figure 8, the effect of coke prices on the cost is

remarkable. For the same ore grade, say, 1.6% Ni, the cost increases from 826 US$ to

1191 US$ per tonne of pig iron when the coke price increases from 180 US$ to 450

US$ per tonne of coke.

In February of 2009, the ore price was about 25 US$ for ore of 1.8-2.0% Ni, the cash

cost was around 800 US$ per tonne of pig iron containing 6-7% Ni.

Figure 7, Estimated cash cost at different ore prices for blast furnace plant


8

Figure 8, Estimated cash cost at different coke prices for blast furnace plant

Figure 9 shows the estimated cash cost per tonne of pig iron at different laterite ore

prices for electric furnace plant. It is similar to blast furnace operation. The cost

increases with the increase of ore price. For ore of 1.6% Ni, the cost is 758 US$ per

tonne of nickel pig iron with ore price at 15US$ per tonne, and when the price is at

60 US$ per tonne, the cost increases to 1033 US$ per tonne of pig iron. In general,

the cost of electric furnace operation is lower than that of blast furnace operation.

Figure 10 shows the estimated cash cost per tonne pig iron at different electricity

prices for electric furnace plant. As show in Figure 10, there is obvious impact of

power prices on the cost. When the power price increases from 0.05 US$ to 0.25

US$ a KWH, the cost increases from 735 US$ to 1230 US$ per tonne of alloy to treat

1.6% Ni ore.

With ore (1.8-2.0% Ni) price at 25 US$ per tonne in February 2009, the cash cost of

an electric furnace plant is about 750 US$ per tonne of alloy.

As indicated above, treating ore of 1.8-2.0% Ni based on the material prices shown in

Table 4, the cash cost is about 800 US$ and 750 US$ per tonne of product containing

6-7%Ni respectively for blast furnace and electric furnace. The total cost is 1.2 factor

of the cash cost according to the data of operating plants, so, the total cost is

respectively 960 US$ and 900 US$ per tonne of product. If the price of nickel pig iron

is lower than the total cost, the producers sustain losses. This is why most of the

nickel pig iron plants in China were shut down at the end of 2008 when the pig iron

price dropped to less than 900 US$ per tonne. However, there are a few pig iron

plants with electric furnaces still in operation. These plants are located in the areas


9

where cheap power can be supplied, and they also choose to treat ore with as high

grade as possible to reduce the cost.

Figure 9, Estimated cash cost at different ore prices for electric furnace plant

Figure 10, Estimated cash cost at different coke prices for electric furnace plant


10

The cost of hydrometallurgical plant- a typical HPAL plant to produce nickel/cobalt

hydroxide intermediates is assessed in order to compare with that of nickel pig iron

production. Although the type and grade of laterite ore treated by HPAL process are

different from those of nickel pig iron production, the cost of HPAL process can be a

reference benchmark for nickel pig iron production. The parameters used in the

cost assessment of HPAL plant are listed in Table 5.

Table 5, the parameters used in cost assessment of HPAL plant

Parameters Value Note

Capacity, t.ore/year 3,200,000

Capital cost, KUS$ 1,500,000 Including mine and ore preparation

Sulphur cost, KUS$/Year 21,000 280 kg sulphuric acid per tonne ore

Fuel cost, KUS$/Year 46,100

Manpower, KUS$/Year 35,000

Maintenance, KUS$/Year 40,000

Other cost, KUS$/Year 41,200

Nickel recovery, % 89

Cobalt recovery, % 88

Nickel /cobalt ratio of ore 10/1

Figure 11 shows that comparison of direct operating cost between nickel pig iron

production and HPAL plant. The data in Figure 11 indicate the direct operating cost of

HAPL plant is much lower than that of nickel pig iron production. At different ore

grade, the cost of HPAL plant is 1.46-2.93 US$/lb. Ni without cobalt deduction. With

cobalt deduction, it is 0.12-1.58 US$/lb. Ni. But the direct operating cost of blast

furnace plant and electric furnace plant to produce nickel pig iron is respectively

4.78-7.51 US$ /lb. Ni and 4.53-6.57 US$/lb. Ni.

Ore treated by HPAL plant is mostly limonite that has low magnesium and aluminum,

thus low sulphuric acid consumption. Its nickel content is relatively low. But nickel pig

iron plants prefer to treat saprolite type ore that has high magnesium and high nickel

content. If these plants treat ore feeding to a HAPL plant, the cost is even higher.

However, if HPAL plant accepts high magnesium ore and sulphuric acid consumption

increases, the cost is still lower than that of nickel pig iron plant. For example, in

the assessment, sulphuric acid consumption increases from 280kg to 450kg per

tonne of ore, the cost of sulphuric acid increases from 21.5 million to 34.6 million per

year. The operating cost increases about 0.23 US$/lb. Ni.

Figure 12 shows the comparison of total cost between nickel pig iron production and

HPAL plant. Based on the data from some new pig iron plants, it is assumed that the

capital cost for blast furnace plant and electric furnace plant to produce 10,000

tonnes of nickel in pig iron is respectively 150 million US$ and 180 million US$. As

shown in Figure 12, the trend is the same as direct operating cost. The total cost of

HAPL plant is much lower than that of nickel pig iron plant though the capital cost is


11

higher. With different ore grades, the cost of HPAL plant is respectively 2.66-5.32

US$/lb. Ni and 1.31-3.97 US$/lb, Ni without and with cobalt deduction, while the

cost for blast furnace plant and electric furnace plant to produce nickel pig iron is

respectively 5.46-8.19 US$ /lb. Ni l and 5.35-7.39 US$/lb, Ni.

Figure 11, Comparison of direct operating cost between nickel pig iron production

and HPAL plant.

Note: the cost for purchasing ore is not included in the Figure for blast furnace and

electric furnace plants

4. Conclusion

The production of nickel pig iron began in 2005 in China when the nickel price was

soaring. The output reached 109,000 tonnes of nickel at its peak in 2007. It surpassed

nickel mined (104,000 tonnes). But it decreased to 61,000 tonnes of nickel in pig iron

because of the decline in nickel price. The main reasons that resulted in the booming

of nickel pig iron production in China are:

Rapid increase of nickel demand in recent years because of massive increase

in stainless steel production;

Limited domestic nickel resources and heavily reliant on the import of nickel;

Surging nickel price in recent years.


12

Figure 12, Comparison of total cost between nickel pig iron production and HPAL

plant

Note: the cost for purchasing ore is not included in the Figure for blast furnace and

electric furnace plants

The cost of producing nickel pig iron mainly depends on the price and grade of

imported ore. For blast furnace plant and electric furnace plant, other main factors

affecting the cost are the prices of coke and electricity. The cash cost is about 800

US$ and 750 US$ per tonne of product containing 6-7%Ni respectively for blast

furnace plant and electric furnace plant that treat 1.8-2.0% Ni ore. The total cost is

1.2 factor of the cash cost based on the data of operating plants, thus it is

respectively 960 US$ and 900 US$ per tonne of product. The price of pig iron

containing 6% nickel was 2000 US$ per tonne in 2007 and dropped to 800 US$ per

tonne at the end of 2008. With nickel pig iron price at 900 US$ per tonne, there is

little profit for nickel pig iron producers. That’s why the majority of nickel pig iron

plants in China were shut down at the end of 2008, especially those blast furnace

plants.

The direct operating cost of nickel pig iron production is much higher than that of

HAPL plant even though HPAL plant treats ore of high sulphuric acid consumption.

The total cost has the same trend as direct operating cost though the capital cost of

HPAL plant is higher. With different ore grade, the total cost of HPAL plant is

respectively 2.66-5.32 US$/lb. Ni and 1.31-3.97 US$/lb. Ni without and with cobalt

deduction, while the cost for blast furnace and electric furnace plants to produce

nickel pig iron is respectively 5.46-8.19 US$ /lb. Ni and 5.35-7.39 US$/lb. Ni.


13

Reference

[1] Data from Chinese Nonferrous Metals Industry Association

[2] Zhong, Yongqi, The trend of nickel pig iron production in China, Presentation on

Symposium of 2009 Chinese Nickel Industry, Beijing, China, February, 2009

[3] Liu, Shengyie, The process of blast furnace to treat laterite ore, Chinese patent,

2005

[4] Chui, Xianyun, Development of process of nickel pig iron, Ferro-Alloys, No.4, 2008,

pp 1-7 (In Chinese)


ALTA 2009 NICKEL/COBALT

PRESSURE ACID LEACHING


THE MINDORO NICKEL PROJECT

HPAL PILOTING AND PROCESS DEVELOPMENT

By

Niels Verbaan and James Brown SGS Minerals Services, Lakefield Site

Boyd Willis

Independent Consultant to Aker Solutions Australia

Debbie Marshall Aker Metals (Toronto), a division of Aker Solutions Canada Inc.

Presented by

Niels Verbaan

[email protected]

and

Boyd Willis [email protected]

CONTENTS

ABSTRACT 2

1. INTRODUCTION 3

2. BENCH PROGRAM FINDINGS 3

3. PILOT PLANT FLOWSHEETS 6

4. PILOT PLANT RESULTS 10

5. PILOT PLANT INTERPRETATION FOR PROCESS DESIGN CRITERIA AND METSIM® MODEL 19

6. CONCLUSIONS 34

7. ACKNOWLEDGEMENTS 34

8. REFERENCES 35

APPENDIX 1-A 36

APPENDIX 1-B 37

APPENDIX 2 38

APPENDIX 3 39


2

ABSTRACT

The Mindoro Nickel Project is located in the Philippines and is being developed by Intex Resources. SGS Minerals was contracted by Intex in April 2008 to provide testing services in support of a Definitive Feasibility Study. The metallurgical testing program commenced with laboratory and continuous pilot plant beneficiation programs, followed by a laboratory hydrometallurgical program. The program culminated in the successful 8 day operation of an integrated continuous pilot plant consisting of: high pressure acid leaching (HPAL), primary neutralisation (PN), counter current decantation (CCD) and secondary (solution) neutralisation (SN). The pilot plant provided feed (SN thickener overflow) for a mixed sulphide precipitation (MSP) laboratory program. This paper discusses the design, operation and results of the integrated HPAL pilot plant. Key findings of 98% nickel and cobalt recoveries at 30 minutes residence time and >60% CCD underflow densities, are highlighted. In addition, this paper discusses the interpretation of the pilot plant data and how it was used by Aker Solutions Australia in the development of process design criteria and a METSIM® model to facilitate the design of the full-scale plant.


3

1. INTRODUCTION

SGS Minerals Services was contracted by Intex Resources in April 2008 to provide testing services in support of a DFS for the Mindoro Nickel Project. The original scope of testing services included beneficiation and hydrometallurgical bench scale programs followed by a beneficiation pilot program and an integrated High Pressure Acid Leach (HPAL), Pre-CCD Neutralisation (PN), CCD and Solution Neutralisation (SN) pilot plant. The current scope has widened to include mixed sulphide precipitation (MSP) bench testing, a metals refinery program (re-leach of MSP, iron/copper removal, zinc/cobalt SX) and a saprolite leach program. The scope of the current paper focuses on the results of the HPAL pilot plant and shows how process design criteria were developed from this leading towards the development of a METSIM® model.

SGS Minerals Services (SGS) is a global leader in the development and demonstration of bankable flowsheets and pilot plant programs, and has completed numerous nickel laterite pilot programs. The work carried out in this program was performed in the Lakefield, Canada facilities.

Aker Solutions is part of the Aker group, and is a leading global provider of engineering and construction services, technology products and integrated solutions serving several industries including mining and metals, oil and gas, refining and chemicals, and power generation.

The Mindoro Nickel Project is being developed by Intex Resources, an independent diversified mineral exploration company based in Oslo, Norway, and listed on the Oslo Stock Exchange (OSE:ITX). The project is located in the Philippines, in the central region of Mindoro Island, and is listed as one of the Philippine Government’s priority projects. The resource is estimated to contain over 200 million tonnes of limonite ore averaging approximately 1% nickel, with an underlying saprolite resource currently not defined to the same level of confidence.

The Mindoro nickel deposit consists of limonite, transitional ore and saprolite ore horizons. The work performed to date at SGS Minerals has focused on the processing of the limonite ore horizon, although metallurgical testwork has been initiated on the saprolite ore horizon. Two separate limonite samples as well as a local limestone sample were received in the summer of 2008. The first sample (700 kg received December 2007) was used for the beneficiation testing program and produced a sample used for the first part of the HPAL bench program. The second sample (21.4 tonnes received August 2008) was used in the beneficiation pilot program which ran parallel to the HPAL bench program. As soon as a representative sample was produced from the beneficiation pilot plant, it was used in the HPAL bench program and subsequently in the HPAL pilot program. Compositions of both samples are shown in Table 1. The magnesium and silicon grades were significantly different and led to different behaviour (physically and chemically).

Table 1 Ore Head Grades

Sample H2O -38 µm Ni Co Cu Zn Fe Mg Al Cr Mn Ca Si S

% % % % % % % % % % % % % %

Bench 43.5 86.6 1.12 0.09 0.01 0.06 40.5 3.6 3.7 2.1 0.65 0.11 5.1 0.1

Pilot 45.5 91.4 1.08 0.15 0.01 0.02 47.2 1.0 3.8 2.38 0.84 0.05 2.1 0.1

2. BENCH PROGRAM FINDINGS

Block process flowsheets for the full scale plant are shown at Appendix 1. Hydrometallurgical bench scale testing focused on determining optimum HPAL, PN, CCD and SN conditions prior to the pilot plant campaign. Part of the program was carried out with the “bench scale” beneficiated samples, but as soon as the beneficiation pilot plant was carried out, HPAL testing switched to using representative samples from the pilot plant. Optimum conditions were established for each unit operation as discussed below.


4

2.1 HIGH PRESSURE ACID LEACHING (HPAL)

Key variables investigated were acid addition, temperature and residence time, the latter by means of taking kinetic samples throughout the test. A series of 20 tests were carried out in 2L and 4L Parr autoclaves at 25% solids feed density.

Acid additions varied from 300 kg/t to 360 kg/t, equivalent to free acid concentrations (before flashing) of 30 to 50 g/L free acid. Excellent extractions (>98% Ni and Co), accomplished at 255°C, 30 minutes and >35 g/L free acid, indicated that the Mindoro ore was very reactive (see Figure 1). Excursions below 30 g/L free acid led to significant drops in nickel extraction (94-96%) and to a lesser extent in cobalt extraction (95-97%).

At the lower temperature of 250°C similar high recoveries (98% Ni and Co) were accomplished at 50-60 minutes and free acid concentrations of 35 g/L. At 30 minutes, nickel and cobalt extractions decreased to around 97%.

Initial bench scale tests indicated that approximately 2.5 g/L of Fe(II) was formed. The ORP in these tests was approximately 450 mV (vs. Ag/AgCl sat. KCl electrode). In subsequent tests an air overpressure of around 200 kPa was applied, increasing the ORP to about 600 mV. This generally led to terminal ferrous levels of 50 to 150 mg/L Fe(II), sufficient to prevent formation of hexavalent chromium.

Average compositions of PLS and residue, as well as average extractions (all at 30-35 minutes) are included in Table 2. The effect of shorter or longer retention time on nickel extraction is shown in Figure 1. The pilot plant ran at 44 minutes retention time (equivalent to 30 minutes in the 4th compartment) and initially targeted 45 g/L free acid concentration (before flashing) in the autoclave discharge stream.

Table 2 Average PLS, Residue and Extraction of Bench HPAL Tests (340 kg/t acid, 255°C, 30-35 minutes)

Ni Co Cu Zn Fe Mg Al Cr Mn Ca Si Na Fe(II) S

PLS, mg/L 3338 396 25 85 2403 3248 871 71 2670 130 467 1120 137

Residue, % 0.017 0.004 0.002 0.002 47.3 0.17 3.47 1.63 0.039 0.042 2.08 0.97 4.5

% Dissolution 98.4 98.1 77.2 89.8 1.5 83.3 6.6 1.1 95.6 NA 1.3 NA

90

91

92

93

94

95

96

97

98

99

100

10 15 20 25 30 35 40 45 50

Free Acid, g/L H2SO4

% N

i E

xtr

ac

tio

n

250C 25-35 min

250C 40-45 min

250C 50-60 min

255C 20-25 min

255C 30-32 min

255C 35 min

250°C, 40-45 min

250°C, 25-35 min

255°C, 20-25 min

255°C, 35 min

255°C, 30-32 min250°C, 50-60 min

Figure 1 Nickel Extraction versus Free Acid in ACD


5

2.2 PRE-CCD NEUTRALISATION (PN)

PN bench scale testing was carried out using freshly produced HPAL pulps and local (Philippine) limestone prepared in seawater. The original objective of the tests was to neutralise the majority of the acid without precipitating any of the iron, which was thought to lead to the presence of difficult to settle iron hydroxides. Hence 5 g/L terminal free acid was selected as the initial target.

The correlation between free acid concentration and dissolved iron and aluminium levels is shown in Figure 2, indicating that iron precipitation commences at free acid levels significantly higher than 5 g/L. In fact, iron precipitation started as soon as limestone was added to the autoclave discharge pulp. Further analysis of the soluble sodium balance indicated that sodium co-precipitated with iron, suggesting the formation of a jarosite type compound. As natro-jarosite was assumed to be more crystalline than ferric hydroxide, this was no longer seen as a concern for CCD performance. For the pilot plant a free acid level of 5 g/L H2SO4 was selected as start-up target. Average limestone consumption in the bench tests was determined to be approximately 110 kg/t ore feed.

0

500

1000

1500

2000

2500

3000

0 10 20 30 40 50


mg

/L F

e,

Al

Fe(III)

Al

Fe(III)

Al

Figure 2 Iron and Aluminium Correlation with Free Acid Concentration

2.3 COUNTER CURRENT DECANTATION (CCD)

Preliminary flocculant scoping and settling tests (2 L cylinder scale) were performed using pulp from one of the bench scale tests. Optimum results were obtained with CIBA’s Magnafloc 455 and produced thickener unit areas of 0.25 m

2/t/d and 25% solids underflow

densities at 115 g/t of Magnafloc 455 flocculant dosage and feed dilution to 3% solids. The pilot plant CCD circuit was designed based on this data. It was designed conservatively to prevent CCD performance from interfering with downstream circuits. Live CCD feed samples were submitted to a thickener vendor, Outotec (Canada) Ltd, and flocculant suppliers, who conducted testing in parallel to the pilot plant at Lakefield.


6

2.4 SOLUTION NEUTRALISATION (SN)

Only 2 bench scale SN tests were carried out prior to the commissioning of the pilot plant. Based on these tests, a final pH of 3.2 (by means of limestone addition) was selected. At pH 3.2, recoveries (percent precipitated) of 76% iron, 70% aluminium and 64% chromium were measured.

2.5 WORK IN PROGRESS

The bench scale testwork program is ongoing, with the following work underway to complement the pilot plant program findings:

• Mixed sulphide precipitation (MSP) testwork, using the SN thickener overflow solution (PLS) from the pilot plant as feed;

• Final neutralisation testwork, using CCD tailings from the pilot plant and synthetic MSP barren solution as feeds;

• Environmental and Geotechnical testwork on a bulk sample of neutralised tailings produced during the final neutralisation testwork

• Pressure oxidation (POX) testing for the re-leach of the mixed sulphide precipitates at the front of the metals refinery;

• Impurity removal testwork to precipitate and separate iron and copper from the POX discharge stream;

• Solvent extraction testwork to separate zinc, cobalt and nickel from the iron/copper free solution;

• Atmospheric leaching testwork for the saprolite ore underlying the limonite ore that was processed in the HPAL pilot plant campaign.

3. PILOT PLANT FLOWSHEETS

Detailed flowsheets of the beneficiation and HPAL circuits piloted at SGS are included in Figure 3 and in Figure 4. The flowsheets, as depicted, were tested in two separate pilot plant setups due the different throughput requirements of the beneficiation pilot plant (which operated up to 700 kg/hr) and the HPAL pilot plant (12 kg/hr). Material from the beneficiation plant was collected in drums and decanted prior to the start of the HPAL pilot plant to adjust the pulp density to the target value.


7

SGS Minerals Beneficiation Pilot Plant: Intex Resources

SGS Project: 11926-003 LEGEND

Synth salt water Client: Intex Resources - Mindoro ProjectCampaign: PP1 liquid

Scrubber Classification Revision: rev9

Screen Date revised: pulp

Ore

2.5' x 5' Scrubber ore (wet)

(fitted w 3/8" grate)

+ 1 mm -1 mm

Beneficiation Plant Final classification screen

(desliming simulation at 400 mesh)

+ 400 mesh - 400 mesh

Rougher Gravity Table

Ro conc.

Ro middlings Cleaner Gravity Table

Ro tails

12" x 12" ball mill Chromite

Concentrate

Cleaner tails

Ball Mill

+ 65 mesh Classification - 65 mesh

Screen

Cleaner Middlings

November 4, 2008

chromite

HPAL

Feed

Figure 3 Beneficiation Pilot Plant Flowsheet

Although for an 8 day HPAL campaign only about 4 tonnes of ore is required, much more is needed to operate, debug and optimise a beneficiation pilot plant due to the significantly higher throughputs. For this project approximately 20 tonnes of material was processed through the beneficiation pilot plant. Even more would have been required to fully pilot the de-sliming stage using cyclones. Instead, fine screening at 400 mesh (38 µm) was applied in the pilot plant. Subsequently about 2.5 tonnes (dry basis) of the beneficiated ore was processed through the HPAL pilot plant.


8

SGS Minerals Pilot Plant: Intex Resources

SGS Project: 11926-004 LEGEND

Client: Intex Resources - Mindoro Project

Campaign: PP1 liquid

Revision: rev9

Date revised: pulp From Beneficiation Pilot Plant

ore (wet)

gas

400 kg/t

Limestone 96% H2SO4 200 kPa Air

25% solids off-gas overpressure (on

C1-C6)

C6 C1

Primary Neutralization

32 L (at 255C) Ti Autoclave

T: C 6 compartment

Letdown/ACD Surge High Pressure Acid Leaching

T: C

T: C

T: C

pH: ( 3 g/L H2SO4)

Plant Air

M455 Flocculent, 0.25 g/L

Wash Water

u/f

CCD1 mL/min CCD2 CCD3 CCD4 CCD5 CCD6 CCD7

10% H2O2

25% Limestone

floc

ORP Adjustment

T: 85 C

pH: as is

ORP: 550 mV T: C

Plant Air pH:

T: C

pH:

T: C

pH:

T: C

Secondary Neutralisation pH:

300 % Seed Recycle

November 4, 2008

80

85

90

1.8

3.2

3.1

3

3

85

95

85

85

85

PN2

PR1

SN1

SN2

SN3

SN4

Feed 2000LFeed 2000LFeed 2000LFeed 2000L

H2SO4

Balance

Balance

Balance

Balance

MSP Feed

HPAL

FeedHPAL

Feed

Balance

PN3

PN4

PN1

Figure 4 HPAL Pilot Plant Flowsheet

The key pieces of equipment are described below:

Scrubbing of moist ore was accomplished using an overflow mill (5 ft x 2.5 ft) equipped with six equally spaced angle lifter bars. The mill was equipped with a 3/8 inch grate to keep the few rocks present in the feed inside the scrubber to improve scrubbing. During the plant operation, parameters such as retention time and density were optimised.


9

Gravity separation was simulated using Wilfley tables. Again more material would have been required to pilot this stage using spirals (approximately 1 t/h).

Leaching was carried out using a 6 compartment submarine type titanium (grade 12) autoclave. Concentrated sulphuric acid was injected into the first compartment and 200 kPa air over pressure was applied for ORP adjustment. Each compartment was mixed using Lightnin A320 hydrofoil impellers.

The PN, CCD and SN circuits consisted of custom built polypropylene reactors equipped with baffles, lids, steam coils and industrial quality pH and ORP electrodes. A photograph of the PN circuit is shown in Figure 5.

Figure 5 Photograph of PN Circuit


10

4. PILOT PLANT RESULTS

The HPAL pilot plant (consisting of HPAL, PN, CCD and SN) operated for 8 days. Overall the pilot plant operation was smooth and uneventful, with autoclave on-time operability of about 90% and just one stoppage to replace a valve.

The primary purpose of this pilot plant was to:

• Demonstrate the key sections of the process flowsheet (Appendix 1) under continuous closed circuit conditions in order to facilitate a high level of confidence in producing a ±15% cost estimate for the purposes of “bankability”;

• Confirm process conditions as determined in bench program studies (255°C, >35 g/L free acid in HPAL, <5 g/L free acid in PN, pH 3.2 in SN);

• Obtain process design data under “steady state” continuous conditions, with major process recycle streams in place. In particular, precipitation circuits (such as laterite leaching, i.e. hematite precipitation) need to be tested under continuous conditions as different solid liquid separation properties are encountered under batch versus continuous conditions;

• Produce “live” samples for testing by equipment vendors such thickener manufacturers and reagent suppliers;

• Produce bulk samples for subsequent tailings neutralisation and environmental and geotechnical testing.

Some interesting observations and results of the pilot plant are highlighted below.

4.1 FEED SETTLING AND RHEOLOGY

Product from the beneficiation pilot plant was settled in 200 L product drums and decanted for use in the pilot plant. It was soon realised that the feed sample did not settle beyond the targeted 25 % solids, despite the fact the bench sample settled to significantly higher values. This was attributed to the high difference in Mg and Si grades between the bench and pilot samples (see also Table 1).

As per normal SGS practice prior to the start of a large pilot plant, a short 24 hour autoclave only campaign was conducted to mechanically test the continuous autoclave systems. After less than one hour of operation at temperature with ore and acid addition, it quickly became apparent that there was an issue around the autoclave feed pumps when it became impossible to pump the feed slurry. The problem centred around plugging of the pump check valves. After replacing the valves multiple times, failure occurred within a matter of minutes after re-starting the pumps. To combat the slurry flowability problem the pilot plant ore feed was diluted from the target 25% solids to 23-24% solids and at this density the commissioning campaign was able to proceed for several hours. At full scale the slurry is diluted by steam in the direct contact preheaters and the rheological properties are expected to significantly improve at elevated pulp temperatures.

Offline rheology testing completed on pilot plant ore feed slurries revealed a critical solids density of approximately 23% solids. The critical solids density is defined as the point above which small increases of the solids density cause a significant decrease of the flowability of the slurry

1. The determination of the critical solids density is graphically shown in Figure 6.

The implication of this indicates that in order to maintain pumpability, the solids density should be less than 23% solids. The remainder of the pilot plant campaign used 22-24% solids feed density and a single set of pump check valves lasted throughout the campaign.

Similar samples were also tested by Outotec (Canada) Ltd using their 100 mm diameter Supaflo high rate thickener test unit leading to an underflow density of 33.5% solids.


11

0

5

10

15

20

25

18 19 20 21 22 23 24 25

Solids density, % wt.

Yie

ld S

tress, P

a

Unsheared

Sheared

CSD

Figure 6 Critical Solids Density - Feed Pulp

4.2 LEACH KINETICS AND EFFECT OF TERMINAL FREE ACIDITY

The bench program identified the reactive nature of the Mindoro ore and that high nickel and cobalt extractions were achieved with relatively short retention times. Based on the bench scale test results the fourth autoclave compartment was used to gauge metallurgical performance at 30 minutes retention time during the pilot plant. The overall retention time (entire 6 compartment autoclave) was 44 minutes.

Figure 7 and Figure 8 show the rate of nickel and cobalt extraction across the autoclave as determined from average data taken from twice daily autoclave profile sample sets. Extraction is calculated based on an iron tie as:

( )nsol

res

fd

fd

resFe

Fe

Fe

Ni

NiExtractiontieFe

'%11 −××−=−

As can be seen the dissolution of nickel is rapid, with more than 95% extraction by the second compartment (approximately 17 minutes RT). Very little additional leaching is achieved beyond the fourth compartment (equivalent to 30 minutes).

94.5

95.0

95.5

96.0

96.5

97.0

97.5

98.0

98.5

1 2 3 4 5 6 7

Compartment (C7 = ACD)

Ni E

xtr

actio

n, %

45 g/L Target FA

40 g/L Target FA

36 g/L Target FA

Figure 7: Nickel Extraction Kinetics – Various Free Acid levels


12

95.5

96.0

96.5

97.0

97.5

98.0

98.5

1 2 3 4 5 6 7

Compartment (C7 = ACD)

Co

Extr

actio

n, %

45 g/L Target FA

40 g/L Target FA

36 g/L Target FA

Figure 8: Cobalt Extraction Kinetics – Various Free Acid levels

Figure 9 shows the extractions for nickel and cobalt after 30 minutes of leaching and the corresponding acid additions. For 97.5% nickel and cobalt extraction, an acid addition of 342 kg/t was required; resulting in a leach liquor containing about 34 g/L H2SO4. Increasing the acid addition resulted in marginally better nickel extraction to just over 98%; however, cobalt extraction was unchanged.

0.405

0.342

0.377

95.0

95.5

96.0

96.5

97.0

97.5

98.0

98.5

99.0

99.5

100.0

32 34 36 38 40 42 44


Extr

actio

n, %

0.33

0.34

0.35

0.36

0.37

0.38

0.39

0.40

0.41

Acid

Ad

ditio

n

A/O

Ni

Co

Figure 9: Compartment 4 (30 min RT) Free Acid vs. Ni/Co Extractions

Figure 10 illustrates the iron precipitation through the back half of the autoclave, with PLS soluble iron concentration plotted against free acid for the last 5 compartments. For a given free acidity the iron concentration decreases with increasing retention time, as expected, through iron hydrolysis reactions. However, after 30 minutes (C4) the decrease in dissolved iron is less significant. The same analysis is made for hydrolysis of aluminium as seen in Figure 11, with little additional aluminium precipitation taking place beyond 30 minutes retention time.

For a HPAL discharge liquor of 34 g/L free acid (acid addition of approximately 342 kg/t), iron and aluminium tenors would be expected to be 1800 mg/L and 800 mg/L respectively in 30 minutes retention time.


13

1000

1500

2000

2500

3000

3500

4000

4500

30 32 34 36 38 40 42 44 46


Fe

, m

g/L

C2

C3

C4

C5

C6

Figure 10: Free Acid vs. Fe in PLS

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

30 32 34 36 38 40 42 44 46


Al, m

g/L

C2

C3

C4

C5

C6

Figure 11: Free Acid vs. Al in PLS

4.3 IRON PRECIPITATION IN PRE-CCD NEUTRALISATION

Pre-CCD neutralisation (PN) was conducted in the pilot plant using local (Philippine) limestone slurried (25% solids) in synthetic seawater. A little more than two thirds of the iron present in the leach liquor was precipitated in the PN circuit with approximately 85% acid neutralisation.


14

0

500

1000

1500

2000

2500

1 2 3 4 5 6 7

PN4 FA, g/L H2SO4

PN

4 L

iq [F

e], m

g/L

Figure 12: PN Terminal Free Acid vs. Fe Concentration

Table 3: Metal Precipitation in PN

FA, g/L Fe Na Al Cr Cu Zn Si

Overall 70.5% 5.6% 3.9% 6.4% 3.2% 2.7% 11.9%

up to Nov 10 4-5 70.5% 7.0% 6.2% 8.9% 1.7% 4.8% 7.7%

Nov 10-12 3-4 70.3% 2.8% -4.1% -1.7% -2.0% 1.3% -0.2%

If all of the 5.6% reduction in soluble sodium was precipitated as natro-jarosite (NaFe3(SO4)2(OH)6) it would account for roughly 55% of the iron precipitated in the PN circuit. This is a benefit of the background salt concentration in the limestone slurry (seawater preparation at 11 g/L Na), as it is assumed that significant precipitation of ferric hydroxide in the PN circuit would have a detrimental effect on the solid-liquid separation properties of the PN discharge slurry. A higher free acidity in the PN discharge slurry favoured the formation of natro-jarosite. When the free acid in the PN discharge slurry was between 4 and 6 g/L the reduction in soluble sodium was approximately 7% and when the PN circuit was operated at a free acid of about 3 g/L, the soluble sodium decreased by only about 3%.

Figure 13 shows the particle size distribution for HPAL and PN discharge solids. It can be seen that the precipitate produced in PN was considerably larger in size when compared to the precipitates formed in the autoclave.


15

0

10

20

30

40

50

60

70

80

90

100

0.1 1 10 100 1000Size, µm

Vo

l. %

Un

de

r

HPAL D/C

PN D/C

Figure 13: PSA for HPAL Discharge Solids Compared to CCD Feed Solids

4.4 LEACH RESIDUE SETTLING IN CCD

A custom 7 stage CCD circuit was designed for this project. Feed streams consisted of PN discharge and SN underflow (introduced in CCD-3) and a synthetic wash feed simulating a barren solution recycle stream which was introduced in CCD-7. Flocculant was added to each thickener.

Bench scale settling work applied on a batch produced feed sample (HPAL and PN), indicated that, at 115 g/t of Magnafloc 455 flocculant dosage and feed dilution to 3% solids, a thickener unit area of 0.25 m

2/t/d was required, leading to 25% solids underflow density. This

data was input to an Excel based mass balance model which estimated the required wash feed flow and SN feed flow for the pilot plant.

Underflow density trends for 3 out of the 7 thickeners are shown in Figure 14, indicating CCD-1 densities of around 60% solids and CCD-7 densities of around 70% solids. These values were achieved using 60 g/t of Magnafloc 455 flocculant dosage and feed dilution to 4% solids. The difference in underflow densities between bench scale HPAL/PN produced data and pilot plant results is remarkable and can largely be attributed to the difference between continuous iron precipitation (in both the autoclave and PN circuit) and batch iron precipitation. An overview of the settling performance from bench to pilot scale is shown in Table 4.

Conducting a shake down campaign (say 1-2 days duration consisting of HPAL and PN) prior to the design stages of the pilot plant would have allowed settling testing to be carried out on continuously produced CCD feed and could have avoided operating an over-sized CCD circuit. Nevertheless, the excellent settling behaviour observed in the pilot plant was confirmed by Outotec (Canada) Ltd, who also reported about 60% solids in CCD underflows, at even lower flocculant dosages.

No adverse effects of recycling underflow from the SN circuit (SNU) were observed. In fact densities in CCD-3 and CCD-4 increased from 60 to 70% within 24 hours. Rheology tests revealed that the critical solids density (CSD) was around 62% solids (see Figure 15), indicating that the pilot plant operated above the critical solids density for the latter part of the campaign. This explains the frequent mechanical problems that were encountered with the rake shafts and drive mechanisms and which ultimately led to CCD-3 being taken off-line, effectively changing it into a 6 stage CCD circuit.


16

10

20

30

40

50

60

70

80

0 24 48 72 96 120 144 168 192

CCD Operating Time, h

Un

de

rflo

w D

en

sit

y, %

so

lid

s (

w/w

)

CCD1

CCD4

CCD7

SNU Recycling to CCD3/4

Figure 14 CCD Underflow Density Trends

Table 4 Overview of CCD Performance

Pre Pilot Pilot Outotec Unit

Bench Plant (pilot sample)

Thickener feed density 4 4 5 % solids (w/w)

CCD1 floc dosage 115 60 7* g/t

Underflow Unit Area 0.25 0.05 0.05 m2/t/d

Avg CCD1 Density 25 58 ~ 60 % solids (w/w)

Avg CCD7 Density 70 % solids (w/w)

Wash Ratio 0.6 m3/t

Wash Efficiency 99.3 %

*Note: using CCD2 O/F liquor from pilot plant with residual floc


17

0

20

40

60

80

100

120

140

160

57 58 59 60 61 62 63 64 65 66 67


Yie

ld S

tres

s,

Pa

Unsheared

Sheared

CSD

Figure 15 Critical Solids Density Determination – CCD4 Underflow

4.5 SOLUTION NEUTRALISATION UNDERFLOW (SNU) RELEACHING IN CCD

Evaluating SNU re-leaching was accomplished by comparing the Fe-tie calculated extractions in CCD feed with the CCD-7 underflow solids stream. Figure 16 shows the overall nickel recovery trends as well as measured CCD pH. It is clear that the gap between nickel recovery in CCD feed and CCD-7 underflow widens after the start of SNU recycling. In fact the average difference between CCD feed and CCD-7 underflow recovery (from Nov 10

th AM

to Nov 12th PM), was 0.5% Ni and 0.2% Co (not shown in the graph) and can be attributed to

losses occurring in the SN circuit. The overall average SN losses were 1.2% Ni and 0.6% Co (from Nov 7

th AM to Nov 12

th PM). Hence it can be determined that the estimated average

nickel re-leach efficiency should be ((1.2-0.5)/1.2)*100) = 62%. Using the same calculation it can be determined that other elements re-leached include 72% cobalt and 40% manganese. Iron and aluminium re-leaching is difficult to determine due to the large quantity of iron and aluminium units in CCD feed relative to the amount of iron and aluminium in SNU. However, using the Fe-tie comparison method, aluminium re-leaching appears to be negligible.


18

96.0

96.5

97.0

97.5

98.0

98.5

7-Nov 8-Nov 9-Nov 10-Nov 11-Nov 12-Nov 13-Nov

Me

tal

Ex

trac

tio

n,

%

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

pH

Ni - CCDF

Ni - CCD7U

Start SNU to CCD3

pH CCD4

pH CCD7

Figure 16 Metal Recovery Trends in CCD Feed and CCD-7 Underflow versus CCD pH

4.6 SOLUTION NEUTRALISATION (SN) PERFORMANCE

In the SN circuit, residual levels of iron, aluminium and chromium from CCD1 overflow were precipitated by the pH controlled addition of 25% limestone slurry. Acceptable precipitation (filtrate levels below 50 mg/L Fe) was achieved at an average pH of 3.2 and temperature of 80°C. Average nickel and cobalt precipitation extents in the SN circuit were roughly 1.0% and 0.5%; however, during the last 2.5 days of piloting, which operated at slightly lower pH (just below 3.2), this was reduced to 0.5% Ni and 0.3% Co. Limestone consumption was measured to be 41.1 kg/t limonite ore feed. The average analysis of the SN thickener overflow (SNO) is presented in Table 5.

Table 5 SNO Average Liquor Assays

Ni Co Fe Mg Na Al Cr Mn Cu Zn Ca Si Cl

mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L

Average SNO 3871 472 36 5869 5046 142 4 3214 20 81 749 61 19367

Off-line settling testing (2 L cylinder scale) using live samples from the pilot plant yielded a thickener underflow unit area of 0.34 m

2/t/d at 67 g/t of Magnafloc 455 flocculant. The pilot

plant thickener resulted in underflow densities as high as 53% solids, although off line rheological testing revealed a critical solids density value of approximately 51% solids (see Figure 17). Additional settling tests were performed by Outotec (Canada) Ltd (see section 5.5.5).


19

0

2

4

6

8

10

12

14

16

46 47 48 49 50 51 52


Yie

ld S

tres

s, P

a

Unsheared

Sheared

Unsheared Plug

Flow >51% Solids

CSD

Figure 17 Critical Solids Density Determination – SN Thickener Underflow

5. PILOT PLANT INTERPRETATION FOR PROCESS DESIGN CRITERIA AND METSIM® MODEL

During the course of the pilot plant campaign a number of small adjustments were made to key operating variables and process set points in response to trends and observations. These adjustments enabled the optimum operating conditions for the full scale plant to be established; however, in most cases this also meant that the pilot plant averages could not be used directly as inputs to the process design criteria. Consequently another level of data manipulation was necessary to extract from the pilot plant results a set of process design criteria that better represented the optimised operating conditions selected for the full scale plant. SGS provided detailed testwork results in native format (MS Excel) to make this data manipulation exercise possible.

A METSIM® model was developed prior to the pilot testwork. This was utilised during the pilot plant campaign to investigate the impacts of various process changes on recycles and on operations downstream of the piloted sections of the process flowsheet.

5.1 FEED SLURRY SETTLING

The beneficiation product (fine slurry) settled poorly and bench scale (cylinder) settling tests performed during the beneficiation pilot plant campaign achieved, at best, 25% underflow solids. In addition, the beneficiation pilot product stored drums (200 L each) did not compact further than 25% solids after several days of settling.

Outotec (Canada) Ltd was commissioned to perform dynamic thickening tests using a 100 mm bench scale high rate thickener. The results of this work were more encouraging, producing 33.5% solids by high rate settling and 34.8% solids in a paste simulation. The key Outotec findings are summarised in Table 6.


20

Table 6 Outotec Leach Feed Slurry Thickening Test Results

High Rate Paste Simulation

Feed Density (% w/w) 5.1 5.1

Flocculant SNF 905 VHM SNF 905 VHM

Floc Dose (g/t) 126.5 123.4

Unit Area (m2/t/d) 0.17 0.28

Underflow % Solids 33.5 34.8

Based on advice from Outotec Global Technology Support that full-scale operation would add at least 2% solids to the reported underflow densities, it was concluded from the test results that 35% solids could be achieved in a high rate thickener, and this value was adopted for the process design criteria. The remaining process design criteria were selected from the high rate thickening data presented in Table 6.

5.2 HIGH PRESSURE ACID LEACHING (HPAL)

5.2.1 Leach Residence Time

The bench scale HPAL tests suggested that autoclave residence time as short as 30 minutes was a possibility. To simulate this in the pilot plant, the autoclave was operated with 44 minutes overall residence time, which translated to a residence time of 30 minutes at compartment 4 (Table 7).

Table 7 Pilot Autoclave Residence Time Distribution

Compartment No. Residence Time (min) Cumulative R.T. (min)

1 10 10

2 7 17

3 7 24

4 6 30

5 6 36

6 8 44

The nickel extraction profiles for the pilot autoclave are presented in Figure 7 (section 4.2). The continuous testwork results demonstrated that 98% nickel extractions could indeed be achieved by compartment 4 of the pilot autoclave, corresponding to 30 minutes, and this was adopted as the design criterion for autoclave residence time.

5.2.2 Residual Free Acid and Nickel and Cobalt Extraction

Based upon a leach residence time of 30 minutes in the pilot autoclave, a plot of nickel and cobalt extraction against residual free acid at compartment 4 was constructed (Figure 18).


21

95

96

97

98

99

100

30 35 40 45

Ex

tra

cti

on

(%

)

Free Acid (g/L)

Nickel

Cobalt

39.5

Figure 18 Compartment 4 Nickel & Cobalt Extraction vs. Free Acid

From Figure 18 a target terminal free acid concentration of 39.5 g/L was selected as the design criterion, corresponding to 98% nickel extraction and 97.7% Co extraction.

5.2.3 Iron, Aluminium and Sodium Chemistry

The continuous pilot plant demonstrated that residual concentrations of iron and aluminium could be less than 2 g/L and 1 g/L respectively at 44 minutes residence time (Figure 10 and Figure 11). It is well known however that the hydrolysis of iron and aluminium is dependent upon both residence time and residual free acid concentration

2, and a small penalty in terms

of residual concentrations was accepted in lowering the residence time to 30 minutes (refer Figure 10, C4 vs. C6 data). A plot of residual iron and aluminium concentrations against free acid at compartment 4 (30 minutes) was constructed (Figure 19).

0

500

1000

1500

2000

2500

3000

3500

30 35 40 45

Co

nc

en

tra

tio

n (

mg

/L)

Free Acid (g/L)

Iron

Aluminium

39.5

Figure 19 Compartment 4 Iron and Aluminium Tenors vs. Free Acid


22

At 39.5 g/L residual free acid, the residual concentrations of iron and aluminium are 2.6 and 1.4 g/L respectively, and these were adopted as the process design criteria.

As seawater was used for ore preparation, the leach feed slurry contained an average background sodium concentration of 4.2 g/L. This promoted precipitation of sodium alunite and natro-jarosite species in the autoclave. It was expected that aluminium would preferentially precipitate as sodium alunite, and as there was more than enough sodium present to achieve this, it was assumed that sodium alunite accounted for all of the precipitated aluminium.

This left the task of interpreting the sodium and iron behaviour. By tabulating the compartment 4 (30 minutes) data from the pilot plant and calculating the amount of sodium associated with sodium alunite, and allocating the balance of the sodium precipitation to natro-jarosite, the proportion of iron precipitated as natro-jarosite was determined. A plot of the iron deportment to natro-jarosite against residual free acid was constructed (Figure 20).

3.0%

3.5%

4.0%

4.5%

5.0%

5.5%

6.0%

30 35 40 45

Fe

Pre

cip

as

Na

-Ja

ros

ite

Free Acid (g/L)

39.5

Figure 20 Iron Deportment in Compartment 4 Residue

From Figure 20 an extent of 4.8% was selected as the design criterion for iron precipitation as natro-jarosite in HPAL.

A number of potential relationships describing sodium behaviour were investigated:

• Residual Na tenor vs. FA

• Na in residue vs. FA

• Residual Na tenor vs. Na in residue

• S:Na ratio in residue vs. FA

• (Fe+Al):Na ratio in residue vs. FA

• Al:Na ratio in residue vs. FA

• Fe:Na ratio in residue vs. FA

• Overall Na precipitation vs. FA


23

While several of these produced reasonable correlations using the pilot plant data in isolation, overlaying the bench scale HPAL results demonstrated that these correlations did not hold for variations in the feed composition. Ultimately the most reliable correlation proved to be residual free acid vs. overall sodium precipitation. The sodium precipitation extent was largely between 70 and 75% in the pilot plant and between 70 and 80% in the bench scale testwork. Figure 21 shows the pilot plant data at compartment 4 (30 minutes).

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

30 35 40 45

So

diu

m P

rec

ipit

ati

on

Free Acid (g/L)

39.5

Figure 21 Compartment 4 Sodium Precipitation vs. Free Acid

From Figure 21 an extent of 73% was selected as the design criterion for sodium precipitation in HPAL.

5.2.4 Extraction of Other Metals in HPAL

Most of the ore components are substantially dissolved during the leach process and reaction extents can be calculated from the pilot plant results. The major exceptions are iron, aluminium, calcium, silicon and chromium, which establish residual concentrations. In order to establish extractions at 30 minutes residence time, plots of extraction against residual free acid at compartment 4 were prepared for magnesium, manganese and zinc (Figure 22).


24

82

84

86

88

90

92

94

96

98

100

30 35 40 45

Ex

tra

cti

on

(%

)

Free Acid (g/L)

Magnesium

Manganese

Zinc

39.5

Figure 22 Compartment 4 Mg, Mn and Zn Extractions vs. Free Acid

From Figure 22, extractions of 96%, 95% and 88.5% were selected as the design criteria for manganese, zinc and magnesium respectively.

5.2.5 Residual Concentrations in HPAL Discharge

Iron and aluminium have been previously discussed in 0. In order to establish residual concentrations at 30 minutes residence time, plots of concentration against residual free acid at compartment 4 were prepared for calcium, silicon and chromium (Figure 23).

0

100

200

300

400

500

600

700

800

900

1000

30 35 40 45

Co

nc

en

tra

tio

n (

mg

/L)

Free Acid (g/L)

Chromium

Calcium

Silicon

39.5

Figure 23 Compartment 4 Cr, Ca and Si Concentration vs. Free Acid

From Figure 23, residual concentrations of 860, 190 and 80 mg/L were selected as the design criteria for silicon, calcium and chromium respectively.


25

5.2.6 Redox Potential and Fe(II)

During the early bench scale HPAL tests high Fe(II) concentrations (~2.5 g/L) were observed in the leach solution. A decision was taken to apply an air overpressure of 200 kPa in the autoclave and this resulted in an increase in the ORP from approximately 450 mV to 600 mV (vs. Ag/AgCl sat. KCl electrode). This lowered Fe(II) levels to <200 mg/L in the remaining bench scale HPAL tests. A 200 kPa air overpressure was subsequently applied in the continuous leaching campaign, resulting in an average Fe(II) concentration in leach discharge of 70 mg/L. An autoclave air overpressure of 200 kPa was subsequently adopted for the full scale design criterion.

5.2.7 Dissolution of Precipitates During Pressure Letdown

It has often been observed that some re-dissolution, particularly of iron and aluminium, occurs during pressure letdown due to the increase in free acid concentration as steam is flashed off. This is difficult to replicate in the pilot plant as the autoclave discharge is collected in a letdown pot with negligible flashing (although significant evaporation takes place in the ACD surge tanks). Two offline attempts were made to investigate this phenomena.

The first attempt involved heating a quantity of autoclave discharge slurry to 90°C and evaporating approximately 30% of the contained water to simulate flashing losses. This resulted in an increase in solution iron tenor of 2.7 times the simple concentrative effect. It was postulated that jarosite species were dissolving; however, it was noted that sodium and sulphate concentrations did not increase beyond the concentrative effect. It was concluded that in the time it took to evaporate the required amount of water some leaching of hematite had in fact taken place and the results were therefore deemed unreliable.

An alternative method was then attempted. Sufficient sulphuric acid was added to a sample of the autoclave discharge slurry to rapidly increase the final acid concentration to 1.44 times the starting level, equivalent to the concentrative effect of flashing (from 255°C) on free acid concentration. Note that for all of the metal salts in solution, as well as for the pulp density, this experimental method caused dilution rather than concentration, so equilibrium changes due to the concentrative effect of flashing were not adequately addressed. Performing a mass balance around solids and solutions suggested that minor dissolution of solids had taken place (less than 2% of the iron and aluminium), although the results were within experimental error.

In lieu of more accurate data, the small dissolution extents estimated were applied in the METSIM® model in order to make an allowance for this behaviour.

5.2.8 METSIM® Modelling of the HPAL Circuit

METSIM® offers a set of modules (unit operations) ideally tailored to modelling a HPAL circuit. The autoclave was modelled using the AUT unit operation which calculates the acid heat of mixing, simulates the chemistry (including heats of reaction), sets the required HP steam flow to achieve the setpoint discharge temperature and calculates the flow and composition of the autoclave discharge and vent streams. Residual concentrations, where relevant, are achieved by controlling reaction extents to meet a target elemental concentration in autoclave discharge. Acid addition is controlled to achieve the desired free acid concentration in the autoclave discharge (before flashing) and the rate of consumption is determined by the model based on the specified feed solids composition and the specified reactions and extents.

Flash vessels (4 stages) are modelled using the FLA unit operation which calculates the evaporation rate based on slurry pressure drop, including compensation for boiling point elevation. Direct contact preheaters (4 stages) are modelled using the HTD unit operation which employs the same functionality as the FLA unit operation. Dissolution during pressure letdown, as discussed in 5.2.7 above, is modelled via reactions in the flash vessels. Even the HPAL vent scrubber is modelled, using the DCV unit operation, a venturi scrubber module that simulates a gas/solids stream being scrubbed with liquor.

Parasitic water streams such as pump gland seal water and autoclave agitator seal flush water are incorporated in the model.


26

In the absence of detailed mineralogy for the Mindoro limonite ore, the chemical components were generally simplified for ease of controlling the extraction of each major element. The following chemical reactions were assumed for the purposes of developing the process METSIM® model.

Dehydration of gibbsite to boehmite during autoclave feed slurry preheating:

Al(OH)3 → AlOOH + H2O

Leaching of metal oxides:

NiO + H2SO4 → NiSO4 + H2O

CoO + H2SO4 → CoSO4 + H2O

2 FeO(OH) + 3 H2SO4 → Fe2(SO4)3 + 4 H2O

FeO + H2SO4 → FeSO4 + H2O

MgO + H2SO4 → MgSO4 + H2O

2 AlOOH + 3 H2SO4 → Al2(SO4)3 + 4 H2O

ZnO + H2SO4 → ZnSO4 + H2O

CuO + H2SO4 → CuSO4 + H2O

Cr2O3 + 3 H2SO4 → Cr2(SO4)3 + 3 H2O

MnO + H2SO4 → MnSO4 + H2O

CaO + H2SO4 → CaSO4 + H2O

SiO2.H2O → SiO2 + H2O

SiO2 + 2 H2O → H4SiO4

Iron precipitation:

3 Fe2(SO4)3 + 2 NaCl + 12 H2O → 2 NaFe3(SO4)2(OH)6 + 5 H2SO4 + 2 HCl

3 Fe2(SO4)3 + 14 H2O → 2 (H3O)Fe3(SO4)2(OH)6 + 5 H2SO4

Fe2(SO4)3 + 3 H2O → Fe2O3 + 3 H2SO4

Aluminium precipitation:

3 Al2(SO4)3 + 2 NaCl + 12 H2O → 2 NaAl3(SO4)2(OH)6 + 2 HCl + 5 H2SO4

3 Al2(SO4)3 + 14 H2O → 2 H3OAl3(SO4)2(OH)6 + 5 H2SO4

During the pilot plant campaign a version of the METSIM® model was created that mirrored the pilot HPAL circuit and this was used to fine tune the chemistry employed in the model.

A copy of the METSIM® flowsheet for the HPAL circuit is included as Appendix 2.

5.3 PRE-CCD NEUTRALISATION (PN)

5.3.1 Terminal Free Acidity and Iron Precipitation

The continuous pilot plant campaign commenced with a free acid target of 5 g/L in PN discharge slurry, with the objective of maximising CCD performance by avoiding iron hydroxide precipitation. Early in the pilot campaign it was observed that terminal iron concentrations in PN discharge were much lower when the terminal free acid was less than 4 g/L (see Figure 12) and in the latter part of the pilot campaign the free acid target was lowered to 3.0-3.5 g/L to maximise iron precipitation. As these conditions were too acidic for significant precipitation of hydroxides, it was concluded that natro-jarosite was being formed. This was a coincidental benefit of employing seawater for limestone preparation which introduced a supplementary source of aqueous sodium ions.


27

It was observed that acid neutralisation was essentially complete after the first tank and iron precipitation was 80% complete after the second tank. Roughly another 10% precipitated in each of the 3rd and 4th tanks. To avoid allowing for unnecessary residence time, the data for tank 3 was selected to construct a plot of iron concentration against terminal free acid (Figure 24).

10

100

1000

10000

1 2 3 4 5 6 7

Fe

(m

g/L

)

Free Acid (g/L)

3.5

Figure 24 Fe Concentration vs. Terminal Free Acid in PN3

Based on the final 3 days of operation of the pilot plant PN circuit, a terminal free acid concentration of 3.5 g/L was selected as the target value for the design criterion. From Figure 24, a residual iron concentration of 650 mg/L was selected as the design criterion for iron in PN discharge.

5.3.2 Co-Precipitation of Nickel and Cobalt

An advantage of running PN at relatively high terminal free acidity is that the conditions do not favour co-precipitation of nickel and cobalt. Based on estimated recoveries to CCD feed, the nickel and cobalt precipitation extents were 0.2% and 0.3% respectively and these extents were adopted for the process design criteria.

5.3.3 Precipitation of Other Species

During the final 3 days of operation of the pilot plant the PN circuit terminal free acid concentration averaged approximately 3.5 g/L. The only other species precipitated to any notable extent during this period was sodium (Table 3), with 2.8% precipitated as natro-jarosite, and this extent was adopted for the process design criterion.

5.3.4 Key Process Design Criteria for the PN Circuit

Key process design criteria for the PN circuit are summarised in Table 8.

Table 8 PN Circuit Design Criteria

Terminal Free Acid Concentration 3.5 g/L

Residual Iron Concentration 650 mg/L

Precipitation Extents: Ni 0.2%

Co 0.3%

Na 2.8%


28

5.3.5 METSIM® Modelling of the PN Circuit

The PN reactors are modelled using the FLA unit operation which calculates the evaporation rate based on slurry temperature, including stripping of water by evolved carbon dioxide and compensation for boiling point elevation. Limestone is modelled using two CaCO3 components, one that reacts and one that does not, to accurately simulate reactivity.

The following chemical reactions were assumed for the purposes of developing the process METSIM® model:

H2SO4 + CaCO3 + H2O → CaSO4·2H2O + CO2

3 Fe2(SO4)3 + 2 NaCl + 12 H2O → 2 NaFe3(SO4)2(OH)6 + 5 H2SO4 + 2 HCl

3 Fe2(SO4)3 + 14 H2O → 2 (H3O)Fe3(SO4)2(OH)6 + 5 H2SO4

Fe2(SO4)3 + 3 CaCO3 + 9 H2O → 2 Fe(OH)3 + 3 CaSO4·2H2O + 3 CO2

2 FeSO4 + 5 H2O + ½ O2 → 2 Fe(OH)3 + 2 H2SO4

5.4 COUNTER-CURRENT DECANTATION (CCD)

5.4.1 Settling Performance

The CCD settling performance was outstanding, with only CCD-1 averaging below 58% solids in underflow slurry. Underflow densities across the circuit are summarised in Table 9.

Table 9 Pilot Plant CCD Thickener Underflow Densities

CCD-1 CCD-2 CCD-3 CCD-4 CCD-5 CCD-6 CCD-7

Average * 58% 63% 67% 62% 62% 70% 70%

Min. 54% 55% 63% 52% 56% 56% 64%

Max. 63% 69% 70% 72% 74% 73% 74%

* Low values due to washing out following bogged rakes excluded (7 out of 88 results)

The trends for CCD’s 1, 4 and 7 in Figure 14 illustrate that aside from the rake bogging events performance was very consistent.

The introduction of Solution Neutralisation (SN) thickener underflow to CCD-3 (CCD-4 after CCD-3 was decommissioned due to rake failure) caused no detriment to settling performance. This is probably due to two key factors:

(i) use of a 300% seed recycle in SN to improve the PSD of the precipitates; and

(ii) a lower iron precipitation load in SN due to the high iron precipitation extent in PN, which resulted in SN underflow solids containing predominantly gypsum (>70%) with less than 5% iron.

Outotec (Canada) Ltd attended the pilot plant and performed dynamic thickening tests using a 100 mm bench scale high rate thickener. The results of this work confirmed the observations from the pilot plant CCDs. The key Outotec findings are summarised in Table 10.

Table 10 Outotec CCD Thickening Test Results

CCD-1 CCD-3 (with SN U/F

solids recycle)

Feed Density (% w/w) 5.0 5.0

Flocculant Magnafloc 455 Rheomax 1010

Floc Dose (g/t) 7.0 9.0

Unit Area (m2/t/d) 0.05 0.07

Underflow % Solids 61.5 64.2


29

The process design criteria for CCD-1 and CCD-2 were based on the high rate thickening data for CCD-1 and the process design criteria for CCD-3 onwards were based on the high rate thickening data for CCD-3 (due to the presence of SN underflow solids in these CCDs).

5.4.2 Wash Efficiency

The pilot plant CCDs operated with an overall wash efficiency∗ of 99.3%. This was essentially achieved with a 6 stage circuit as bogged rakes meant that there were few occasions when all 7 CCDs were in service. The wash ratio, expressed as volume of wash liquor per unit volume of liquor in CCD-7 underflow, averaged 1.5.

The CCD wash liquor used in the pilot plant was a synthetic solution designed to represent the barren solution discharged from mixed sulphide precipitation, adjusted to pH 2.7. The composition of the barren solution prepared was predicted using METSIM®.

Exploratory METSIM® modelling confirmed that 6 CCDs could achieve a soluble nickel loss to CCD tailings of <0.5%, based on the pilot plant wash ratio and wash efficiency data. Process design criteria adopted from the pilot plant results are summarised in Table 11.

Table 11 CCD Washing Design Criteria

Number of CCDs 6

CCD Wash Ratio 1.5

CCD Wash Liquor pH 2.7

CCD Wash Efficiency (Overall) 99.3%

Soluble Nickel Loss to CCD Tailings <0.5%

5.4.3 Re-Leaching of SN Precipitates

With the CCD circuit running under acidic conditions (CCD-3 slurry feed at approximately 1 g/L free acid and CCD wash liquor at pH 2.7) and providing considerable residence time, some re-leaching of hydroxides precipitated in the SN circuit was observed (refer section 4.5). Good recoveries of nickel and cobalt co-precipitated in SN were achieved, with negligible re-dissolution of the precipitated iron and aluminium (as discussed in section 4.5). Key re-leach extents adopted for the process design criteria are presented in Table 12.

Table 12 Re-Leach of SN Precipitates in CCD

Element Re-Leach Extent

Nickel 62%

Cobalt 72%

Manganese 40%

Iron Negligible

Aluminium Negligible

∗ Wash Eff = [(CCD-1 O/F Liq. Ni Units) - (CCD-7 U/F Liq. Ni Units)] / (CCD-1 O/F Liq. Ni Units)


30

5.4.4 METSIM® Modelling of the CCD Circuit

METSIM®’s countercurrent washing module, the CCW unit operation, was used to model the CCDs. Multiple CCW modules with consecutive unit operation numbers are treated as one group of unit operations and are calculated as a block. Due to this functionality, however, CCW modules cannot deal with evaporation and heat losses. An iterative MS Excel spreadsheet was developed to calculate CCD evaporation losses and heat losses. A phase splitter (SPP) was placed on the overflow stream from each CCW, each set up with reactions to convert the required mass flow of water to water vapour and with a heat loss as a fraction of heat input. The design case METSIM® model includes the total heat losses but excludes the evaporation mass losses, thus generating the maximum hydraulic flowrates for equipment sizing purposes.

In order to set the individual mixing efficiencies used in each CCW unit operation in the METSIM® model, a version of the model was created that mirrored the pilot plant CCD circuit. The model parameters were fine tuned by entering stream data from the pilot plant into the CCD circuit input streams and adjusting mixing efficiencies until a satisfactory match in output streams was achieved.

A copy of the METSIM® flowsheet for the CCD circuit is included as Appendix 3.

5.5 SOLUTION NEUTRALISATION (SN)

5.5.1 Selection of Target Terminal pH

Downstream of SN, nickel and cobalt will be recovered by mixed sulphide precipitation (MSP). The precipitation reactions in MSP generate acid and nickel recovery decreases sharply when the MSP discharge pH falls below 1.6

3. The original basis for selection of the terminal pH in

SN was therefore to be high enough to maintain pH >1.6 in MSP discharge. However it became apparent from METSIM® modelling that, due to the quantity of acid produced by the MSP reactions, a pH as high as 4 in SN could not keep the MSP discharge at pH >1.6. A decision was taken to dose a neutralising agent into the MSP reactors and to select the SN target pH based on the optimal balance between iron precipitation and nickel co-precipitation.

It was observed during the first two days of the pilot plant campaign that the nickel assay in the SN precipitates was approximately 0.1% at SN discharge slurry pH values from 3.0 - 3.2. When the slurry pH drifted up to 3.3 – 3.4 the nickel assay in the precipitates increased rapidly to almost 0.2%, corresponding to >1% nickel precipitation. This observation is supported by Figure 25, which was constructed using pilot plant data. It was therefore decided to set a pH target of <3.2 for the latter part of the pilot plant campaign and based on this a pH range of 3.1 - 3.2 was selected for the process design criterion.


31

0.0

0.1

0.2

0.3

0.4

2.9 3.0 3.1 3.2 3.3 3.4 3.5

Nic

ke

l in

SN

Re

sid

ue

(%

)

SN Discharge pH

Figure 25 Nickel in SN Precipitates vs. Terminal pH

Co-precipitation of nickel and cobalt averaged 1.0% and 0.5% respectively over the full pilot plant campaign, but just 0.5% and 0.3% over the final 2.5 days with the lower pH target. The latter were taken as process design criteria inputs.

5.5.2 Iron and Aluminium Precipitation

Residual concentrations of iron and aluminium in SN discharge are plotted against pH in Figure 26. During the final 2.5 days of the pilot plant campaign, at target pH <3.2, approximately 89% of the iron and 84% of the aluminium in SN feed were precipitated. Over this period the residual iron and aluminium concentrations averaged 66 mg/L and 166 mg/L respectively. Process design criteria values of <200 mg/L for aluminium and <80 mg/L for iron were selected from an intersect drawn at pH 3.1.

1

10

100

1000

2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6

Co

nc

en

tra

tio

n (

mg

/L)

SN Discharge pH

Aluminium

Iron

pH 3.1

Figure 26 SN Iron & Aluminium Concentrations vs. Terminal pH


32

5.5.3 Silicon and Chromium Precipitation

Residual concentrations of silicon and chromium in SN discharge are plotted against pH in Figure 27 and Figure 28 respectively. During the final 2.5 days of the pilot plant campaign, at target pH <3.2, approximately 92% of each of the silicon and aluminium in SN feed were precipitated. Process design criteria residual concentrations of <70 mg/L for silicon and 5 mg/L for chromium were selected from intersects drawn at pH 3.1.

0

20

40

60

80

100

120

2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6

Si C

on

ce

ntr

ati

on

(m

g/L

)

SN Discharge pH

pH 3.1

Figure 27 SN Silicon Concentration vs. Terminal pH

0

5

10

15

2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6

Cr

Co

nc

en

tra

tio

n (m

g/L

)

SN Discharge pH

pH 3.1

Figure 28 SN Chromium Concentration vs. Terminal pH


33

5.5.4 Seed Recycle

SN thickener underflow was recycled to the first tank of the SN circuit as seed at an average recycle rate of just under 300% (solids basis) and this has been adopted as the process design criterion. This produced SN discharge solids with a very narrow particle size distribution and with a small average particle size (~6 µm), but this was expected with just 5 days of seed recycling.

5.5.5 SN Discharge Slurry Settling

The SN discharge settling performance was better than anticipated, averaging 43% solids in underflow slurry.

Outotec (Canada) Ltd attended the pilot plant and performed dynamic thickening tests using a 100 mm bench scale high rate thickener. The results of this work confirmed the observations from the pilot plant thickener. The key Outotec findings adopted as process design criteria are summarised in Table 13.

Table 13 Outotec SN Discharge Slurry Thickening Test Results

SN Discharge Slurry

Feed Density (% w/w) 16

Flocculant Magnafloc 455

Floc Dose (g/t) 18.0

Unit Area (m2/t/d) 0.07

Overflow Clarity (mg/L) 88

Underflow % Solids 57

5.5.6 METSIM® Modelling of the SN Circuit

The SN reactors are modelled using the FLA unit operation which calculates the evaporation rate based on slurry temperature, including stripping of water by evolved carbon dioxide and compensation for boiling point elevation. Limestone is modelled using two CaCO3 components, one that reacts and one that does not, to accurately simulate reactivity.

The following chemical reactions were assumed for the purposes of developing the process METSIM® model:

H2SO4 + CaCO3 + H2O → CaSO4·2H2O + CO2

Fe2(SO4)3 + 3 CaCO3 + 9 H2O → 2 Fe(OH)3 + 3 CaSO4·2H2O + 3 CO2

2 FeSO4 + 5 H2O + ½ O2 → 2 Fe(OH)3 + 2 H2SO4

Al2(SO4)3 + Na2SO4 + 12 H2O → 2 NaAl3(SO4)2(OH)6 + 6 H2SO4

Al2(SO4)3 + 3 CaCO3 + 9 H2O → 2 Al(OH)3 + 3 CaSO4·2H2O + 3 CO2

Cr2(SO4)3 + 3 CaCO3 + 9 H2O → 2 Cr(OH)3 + 3 CaSO4·2H2O + 3 CO2

NiSO4 + CaCO3 + 3 H2O → Ni(OH)2 + CaSO4.2H2O + CO2

CoSO4 + CaCO3 + 3 H2O → Co(OH)2 + CaSO4.2H2O + CO2

CuSO4 + CaCO3 + 3 H2O → Cu(OH)2 + CaSO4.2H2O + CO2

ZnSO4 + CaCO3 + 3 H2O → Zn(OH)2 + CaSO4.2H2O + CO2

n H4SiO4 → (SiO2)n + 2n H2O


34

6 CONCLUSIONS

1. An 8 day integrated HPAL pilot plant (HPAL, PN, CCD, SN) was successfully completed following the completion of beneficiation and HPAL bench scale testwork programs and a beneficiation pilot plant. Approximately 20 wet tonnes of material was processed through the beneficiation pilot plant and 2.5 dry tonnes of material through the HPAL pilot plant.

2. Pilot plant test results demonstrated that a bulk Mindoro limonite ore sample was treated very successfully using the nominated process flowsheet. Optimised process design criteria have been developed for the key sections of the process that were piloted.

3. Developing in advance, and having available during the pilot plant campaign, a detailed working process simulation (METSIM® model) yielded benefits both to decision making during the pilot plant and fine tuning and verification of the model parameters.

4. In some instances the process design criteria developed for the full scale plant design differ from direct pilot plant results, illustrating the importance of obtaining detailed testwork results in native format to facilitate additional data manipulation. This also highlights the need for good cooperation between the test facility and the engineering contractor.

5. The remarkable difference in solid liquid separation results between the “bench” and the “pilot” feed ore samples stresses the importance of sample selection and is a strong indication that a significant ore variability testing program is required at the next stage of design.

6. It was found that the ore from the Mindoro deposit was very reactive with 98% Ni and 97.7% Co extractions possible in 30 minutes at 255°C and 39.5 g/L terminal free acid (before flashing).

7. One benefit of using seawater for ore processing is that aluminium precipitates as sodium alunite in the autoclave. At 39.5 g/L free acid, the residual Al tenor was 1.4 g/L and the residual Fe tenor was 2.6 g/L.

8. The free acid target in the PN circuit was lowered to 3.0-3.5 g/L, at which it was concluded that natro-jarosite was being formed because these conditions were too acidic for significant precipitation of hydroxides and because a supplementary source of aqueous sodium ions had been introduced through the use of seawater for limestone preparation.

9. The CCD settling performance was outstanding with only CCD-1 averaging below 62% solids in underflow slurry. This is also a benefit of the formation of jarosite, rather than hydroxide, in the PN circuit.

10. Nickel and cobalt co-precipitation during SN increased significantly at pH levels above 3.3. Although a higher pH is required to offset acid generation during the subsequent mixed sulphides precipitation (MSP), it was concluded that higher pay metal recoveries would be achieved by keeping pH in SN below 3.2 and incorporating neutralisation during MSP. Ni and Co losses in SN at these conditions were 0.5% and 0.3% respectively.

7 ACKNOWLEDGEMENTS

The authors wish to express their thanks to Intex Resources, Aker Solutions and SGS Minerals Services for their permission to publish this paper. In addition the diligent work by numerous SGS plant operators is acknowledged.


35

8 REFERENCES

1. A. Mezei, M. Ashbury, I. Todd, Rheological Aspects of Nickel Hydrometallurgy, ALTA 2002: Nickel/Cobalt Conference, May 2002, Perth, Australia.

2. B.I. Whittington and D. Muir, Pressure Acid Leaching of Nickel Laterites: A Review, Mineral Processing and Extractive Metallurgy Review, Volume 21, Issue 6 October 2000, pages 527 – 599.

3. N. Tsuchida, Y. Ozaki, O. Nakai, and H. Kobayashi, Development of Process Design for Coral Bay Nickel Project, International Laterite Nickel Symposium – 2004, TMS, March 2004, Charlotte, North Carolina, USA.


36

APPENDIX 1-A

Mindoro Nickel Project - Beneficiation Plant Block Flowsheet


37

APPENDIX 1-B

Mindoro Nickel Project - Process Plant Block Flowsheet


38

APPENDIX 2

METSIM® Flowsheet for Mindoro Nickel Project HPAL Circuit


39

APPENDIX 3

METSIM® Flowsheet for Mindoro Nickel Project CCD Circuit


MANGANESE REMOVAL FROM THE GLADSTONE LEACH

RESIDUE SLURRY AND SOLUTION BY PRECIPITATION

By

W. Zhang*, Peter Mason#, W. Wang* and Z. Zhu*

*The Parker Centre, CSIRO Minerals, Australia #Gladstone Pacific Nickel Ltd, Australia

Presented by

Wensheng Zhang

[email protected]

CONTENTS

1. INTRODUCTION 2

2. EXPERIMENTAL 3

3. RESULTS AND DISCUSSION 4

4. LIME CONSUMPTION 14

5. CONCLUSIONS 15

6. ACKNOWLEDGEMENT 15

7. REFERENCES 15


2

1. INTRODUCTION

Manganese is an impurity in laterite hydrometallurgical processes and has to be removed to an acceptable level from the waste effluents to comply with environmental guidelines for discharge. One of the main issues associated with the removal of manganese by neutralisation is the co-precipitation of magnesium, which can result in substantial lime usage to the extent that it becomes a significant consumable and cost in a nickel laterite hydrometallurgical process and generates substantial volumes of neutralization residue. Oxidative precipitation of manganese allows the reactions to occur in a lower pH range and therefore minimises the co-precipitation of magnesium; however when a slurry is treated, care must be taken not to oxidise and dissolve chromium from the associated residue solids. In recent years, the Parker Centre / CSIRO Minerals has investigated and developed methods and processes for the recovery and removal of manganese from various types of solutions using oxidative precipitation (Zhang 2000; Zhang et al. 2002, Zhang and Cheng, 2007a, b, c).

Recently, investigation for removal of manganese from the laterite leach residue slurry and solution of Gladstone Pacific nickel Ltd (GPNL) has been carried out at the Parker Centre CRC and CSIRO Minerals. The process selected for the Gladstone nickel Project (GNP) is high pressure acid leaching (HPAL) followed by mixed sulphide precipitation (MSP) to recover nickel/cobalt as mixed sulphide product. The GNP circuit options tested for removal of manganese are shown in Figure 1, and include manganese removal by precipitation from (1) the underflow of counter-current decantation (CCD) and (2) the excess barren solution from the MSP. The objective is to economically minimise dissolved manganese and chromium in the treated liquor leaving the process.

The current project was proposed to investigate both carbonate and oxidative precipitation with SO2/air for removal of manganese from both the slurry and the solution. The objectives of the project were to test kinetics and selectivity of the precipitation methods for reducing the soluble manganese to below 10 mg/L, whilst holding the soluble chromium below 1 mg/L in the final discharge, with minimum co-precipitation of magnesium. The parameters examined for the oxidative precipitation of manganese included the solution pH, SO2/air ratios and temperature. While the main focus was the removal of manganese, the behaviours of other metals of concern such as nickel, cobalt, and zinc were also closely monitored during the oxidative precipitation. This paper summarises the test results and discusses the optimum conditions for the removal of manganese.

Figure 1: Routes tested for removal of manganese from the GPNL circuit

LEACH

CCD Mn PPTN 1 RSF

MSP

Mn PPTN 2

Lime/Carbonate or SO2/air

Lime/Carbonate or SO2/air

Treated Barren

<500 mg/L Mn

<10 mg/L Mn

~6000 mg/L Mn

RSF: Residue Storage Facility

Product

Decant


3

2. EXPERIMENTAL

2.1 TEST RIG SET-UP

A photograph of the test rig set-up is shown in Figure 2. All the tests for the oxidative precipitation of manganese with SO2/Air were carried out in a two-litre acrylic reactor with a water jacket. A Digital overhead stirrer and a disk turbine impeller (6 cm in diameter) were used for mixing. Air was supplied from laboratory compressed air and SO2 from a BOC cylinder. The gas flow rates were measured with calibrated flow meters and manually adjusted to the desired values with the regulators and valves. A conical shaped flask was attached for mixing gases before entering the reactor through a sparger with fine holes positioned right under the impeller. Slurry temperature was achieved and maintained by circulating hot water through the water jacket from a Julabo hot water circulator. Solution pH was continuously monitored by a thermo pH probe connected to a Walchem pH controller and automatically adjusted by adding lime or limestone slurry with a dosing pump. The amount of lime added was continuously weighed on a balance and recorded when a sample was taken. The system potential (E) was monitored continuously by a platinum electrode with Ag/AgCl reference electrode.

Figure 2: A photograph of the test rig set-up

2.2 MATERIAL AND METHOD

The original slurry supplied by SGS was diluted with a spiked solution to meet the specifications set by the METSIM model. The spike solution was made by dissolving technical grade metal sulphate salts into the sea water taken from Floreat beach in Perth. The final slurry contained about target 28% solid (w/w) with the metal concentrations close to the target values from the GPNL METSIM modelling (Table 1).

Table 1: Solution Metal concentrations in the synthetic slurry

Concentration (mg/L) Ni Co Zn Cu Cr Al Fe Mn Mg Ca Cl

Target value 94.7 9.6 60.1 0.08 113.5 1108 2827 3696 13417 637 Before NEU* 59.9 8.08 7.38 0.67 113.8 1208 3066 3397 15236 563 20834 After NEU* 35.4 5.05 2.66 0.73 0.19 0.46 13.4 2969 13893 557 18918 *NEU: Neutralisation


4

The synthesised slurry was neutralised to about pH 5 and 60°C with limestone slurry. Air was continuously sparged into the slurry with sufficient agitation to accelerate the oxidation of ferrous ions. About half of the neutralised slurry was filtered to obtain the feed solution for the parallel tests for the manganese oxidative precipitation under the same other conditions. Chemical assay results for the aqueous phase of the neutralised slurry are given in Table 1.

For each oxidative precipitation test, the slurry or the solution was preheated to a desired temperature with constant stirring. Timing was started when a gas mixture of SO2 and air was introduced into the reactor. Slurry samples were taken at regular intervals. At the end of each test, the slurry or the solution was collected and weighed for estimation of water loss by evaporation during the test.

2.3 SAMPLE TREATMENT AND CHEMICAL ANALYSIS

Slurry samples went through filtration or centrifugation to separate solution from solid. The wet cakes were dried in an oven at 110°C to determine the percentage of solids in the slurry. Metal concentrations in the aqueous samples were assayed by Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES) at the analytical laboratory of CSIRO Minerals. Some typical aqueous samples at different stages were also selected and sent to Australian Laboratory Services Pty Ltd (ALS) for assaying cadmium, chromium, nickel and cobalt to 0.001 mg/L detection limits by the Inductively Coupled Plasma Mass Spectroscopy (ICP-MS). Free acid in the slurry solution was determined by the magnesium-EDTA titration method (Rolia and Dutrizac 1984).

3. RESULTS AND DISCUSSION

3.1 PRECIPITATION OF MANGANESE WITH LIME/MAGNESIUM CARBONATE

This test aimed at examining the feasibility to precipitate manganese as manganeseCO3 using magnesiumCO3. Based on thermodynamic solubility constants (Lide and Frederikse 2007), the reactions between magnesiumCO3 and Mn(II) at appropriate pH are expected to proceed according to equation (1).

MgCO3 + Mn2+

= MnCO3 + Mg2+

(Log K = 5.48) (1)

Large value of Log K suggests that the system is highly selective for manganeseCO3 over magnesiumCO3. Laboratory grade (BDH) hydrated basic heavy magnesium carbonate, Mg(OH)2·nMgCO3 (42-45% MgCO3), was used. The pre-neutralised slurry was heated to 60°C and the initial pH was adjusted to 7 with lime slurry. At this pH, 1.2 times of the stoichiometric amount of magnesiumCO3 based on equation (1) was added in the form of magnesiumCO3 slurry mixed with water at 1:3 ratios. This addition resulted in about 0.2 pH unit increase, but no further change in pH was observed. It was expected that the formation of manganeseCO3 would need pH >7.5 at 60°C. Therefore, the slurry pH was gradually increased with the lime slurry to pH 7.8 where a significant precipitation occurred. After equilibrium and sampling, another 1.2 times of the stoichiometric amount of magnesiumCO3 was added. Again, a slight increase in pH was observed. In order to test the effect of pH on the degree of the manganese precipitation, the slurry pH was further increased to about pH 8 with lime slurry and sampled again after equilibrium.

The variation of manganese and magnesium concentrations as a function of pH is shown in Figure 3. Manganese concentration decreased to 10 mg/L with addition of 2.4 times of the stoichiometric amount of magnesiumCO3 as the pH increased from 7 to 8. Based on equation (1), magnesium was expected to be replaced by manganese and entered into the solution. However, there was no increase in magnesium concentration while the manganese precipitated. Instead, significant amounts of magnesium were precipitated at pH > 7.2. This suggests that hydroxide precipitation was significant under the test conditions. Although reaction (1) was thermodynamically favourable with a large equilibrium constant K, the reaction could be kinetically limited by reaction (2) due to the less soluble nature of the magnesiumCO3 used.

MgCO3 + 2OH- = Mg(OH)2 + CO3

2- Log K = 6.08 (2)

In practical application, using more soluble carbonate salts such as Na2CO3 is expected to be more reactive and efficient.


5

The behaviours of other metals are shown in Figure 4. Small amounts of Fe, Al and chromium were reduced to below the detection limit (0.2 mg/L) at pH > 7 while cobalt needed pH above 7.8. Nickel concentration was decreased to about 1 mg/L at pH 8.

0.1

1.0

10.0

100.0

1000.0

10000.0

4.0 5.0 6.0 7.0 8.0 9.0

pH

Mn

co

ncen

trati

on

(m

g/L

)

0

10

20

30

40

50

60

70

80

90

100

Mg

pre

cip

itati

on

(%

)

Mn

Mg

Figure 3: Carbonate precipitation with MgCO3 and lime at 60°C.

0.1

1.0

10.0

100.0

4.0 5.0 6.0 7.0 8.0 9.0

pH

Meta

l co

ncen

trati

on

(m

g/L

)

Al

Fe

Ni

Co

Cr

Figure 4: Precipitations of metals using MgCO3 and lime at 60°C.


6

3.2 PRECIPITATION WITH SO2/AIR USING THE SLURRY

A number of tests were conducted using the neutralised slurry to examine the effects of pH, SO2/air ratio and temperature on metal precipitation.

3.2.1 Effect of pH

The rate of manganese precipitation increased significantly with increasing pH in the range of 5-7 (Figure 5). For example, it took less than 60 minutes at pH 6.5 and 7 to obtain <10 mg/L manganese compared with 130 minutes at pH 6 and 270 minutes at pH 5. As shown in Figure 5, the initial rate of manganese precipitation at pH 6 was almost the same as that at pH 6.5 and 7 with about 50% precipitation, but it then slowed down with reaction time and decreasing manganese concentrations in the solution (Figure 5). This may offer a useful strategy for pH control in operation to use different pH for different stages based on the range of manganese concentrations. In this way, the co-precipitation of magnesium, and usage of both SO2 and lime reagents would be expected to be minimised at a lower pH while the rate of manganese precipitation be ensured. These results indicate that pH is one of the key parameters for optimisation, but further increase in pH above 7 would be limited by a significant co-precipitation of magnesium. The results indicate that the SO2/air oxidising system performs well in a wide pH range for removal of manganese.

0.1

1.0

10.0

100.0

1000.0

10000.0

0 30 60 90 120 150 180 210 240 270 300

Time (min)

Mn

co

ncen

trati

on

(m

g/L

)

pH 5

pH 6

pH 7

pH 6.5

Figure 5: manganese precipitation at different pHs, 60°C and 3.8% SO2.

Unfortunately the oxidising conditions promote dissolution of chromium; thus it would to be necessary to limit the extent of oxidation to avoid exceeding soluble chromium limitations. As shown in Figure 6, chromium concentration started to increase when the manganese concentration decreased below 1000 mg/L and continued to increase with reaction time. The dissolution rate increased with increasing pH in the range tested. Nevertheless, it can be seen in Figure 5 that at pH 6.5, manganese can be reduced to below 10 mg/L without raising chromium above 1 mg/L (Figure 6). A novel approach, which is currently the subject of a patent application, is being developed to overcome this issue.

Co-precipitation of magnesium fluctuated in the range of ± 3% in the pH range of 5 - 6.5 during the manganese precipitation, which was well within experimental and analysis errors. There was an increased tendency for co-precipitation of magnesium at higher pH 7 after the completion of manganese precipitation.


7

The variation of nickel, cobalt and zinc concentrations with reaction time and pH are shown in Table 2. The trends of nickel precipitation with time at different pH very much resembled those for manganese precipitation (Figure 5 and Figure 7). This suggests that nickel was likely to co-precipitate together with manganese oxides mainly by adsorption mechanism in this pH range. In comparison, the concentration of zinc more depended on pH and readily decreased to below 0.2 mg/L at pH above 5. Cobalt behaved similar to zinc with faster and more complete precipitation (<0.2 mg/L) at pH above 6.

0.1

1.0

10.0

100.0

0 30 60 90 120 150 180 210 240

Time (min)

Cr

co

ncen

trati

on

(m

g/L

)

pH 5

pH 6

pH 6.5

pH 7

Figure 6: Chromium dissolution at different pHs, 60°C and 3.8% SO2.

Table 2: Variations of nickel and cobalt at different pHs, 60°C and 3.8% SO2

Time pH 5 pH 6 pH 6.5 pH 7

(min) Ni Co Zn Ni Co Zn Ni Co Zn Ni Co Zn

0 37.0 5.34 4.17 38.9 5.39 4.58 39.6 5.54 3.23 38.9 5.39 0.27

30 33.3 2.71 3.10 32.1 1.76 2.85 4.3 0.60 <0.2 4.93 0.97 0.25

60 30.9 1.14 2.30 8.58 0.36 1.54 1.09 <0.2 <0.2 1.63 0.2 <0.2

90 30.5 0.49 4.01 4.82 <0.2 0.25 0.51 <0.2 <0.2 1.14 <0.2 <0.2

120 24.3 0.22 1.50 3.36 <0.2 0.28 0.37 <0.2 <0.2 0.48 <0.2 <0.2

180 22.0 <0.2 1.24 1.45 <0.2 <0.2 0.26 <0.2 <0.2 0.43 <0.2 <0.2

240 18.4 0.41 1.01 0.60 <0.2 0.21 0.61 <0.2 <0.2 <0.2 <0.2 <0.2


8

0.1

1.0

10.0

100.0

0 30 60 90 120 150 180 210 240 270 300

Time (min)

Ni

co

ncen

trati

on

(m

g/L

)pH 5

pH 6

pH 6.5

pH 7

Figure 7: Nickel precipitations at different pHs, 60°C and 3.8% SO2.

3.2.2 Effect of SO2/air ratio

Oxidising power of the SO2/air system depends on the actual ratio of dissolved SO2 to that of O2, corresponding to a SO2/air ratio in the gas mixture. In this project, the effect of SO2/air ratio (v/v) was investigated in the range of 1.9-7.4% SO2 by varying the flow rates of SO2 and air at pH 6 and 7 at 60°C. The results clearly showed that an optimum SO2/air ratio for maximum rate of manganese oxidative precipitation existed (Figure 8).

At lower pH 6, the oxidative precipitation rate dramatically increased from 1.9% SO2 to 3.8% SO2, indicating that the rate was in the region of SO2 input control. A further increase to 7.4% SO2 resulted in a significant decreased rate, indicating that the system became less oxidising with excess input of SO2, i.e. in the region of O2 mass transfer control. Excess SO2 dissolved in the solution would react with oxygen to produce H2SO4 and need more base reagent for neutralisation. The results suggested that optimum SO2/air ratio in terms of maximum rate of manganese oxidative precipitation could be further tuned in the range of 1.9% - 3.8% and 3.8 – 7.4% SO2.

At higher pH 7, the rate of manganese precipitation with 7.4% SO2 also considerably decreased compared to that with 3.8% SO2 (Figure 8). Therefore, the SO2/air ratio at pH 7 could be similarly further optimised around 3.8% SO2.

For chromium re-dissolution, in addition to its dependency on the manganese concentration and pH as discussed earlier, it is interesting to note that the chromium re-dissolution also significantly depended on the SO2/air ratio or the oxidising power (Figure 9). For example, at pH 7, the re-dissolved chromium with 3.8% SO2 reached about 1 mg/L at the time for obtaining the target 10 mg/L manganese compared to 0.4 mg/L chromium with 7.4% SO2. Similarly, at pH 6, the re-dissolved chromium with 3.8% SO2 reached above 1 mg/L at the time for obtaining the target 10 mg/L manganese compared to that well below 1 mg/L chromium with 7.4% SO2. However, any less oxidising condition would be at the expense of kinetics of manganese oxidation and utilisation of SO2 and base reagents for neutralisation (see section 4).

As expected, the variations of magnesium precipitations with different SO2/air ratios were very small in the pH range tested with only a few percent positive or negative fluctuations. No significant variation of cobalt and zinc concentrations with different SO2/air ratios was observed at a fixed pH. Nickel precipitations were faster with 3.8% SO2 at both pH 6 and 7, corresponding to faster rates of manganese precipitation, again suggesting that adsorption mechanism was dominant.


9

0.1

1.0

10.0

100.0

1000.0

10000.0

0 30 60 90 120 150 180 210 240

Time (min)

Mn

co

ncen

trati

on

(m

g/L

)

1.9% SO2, pH 6

3.8% SO2, pH 6

7.4% SO2, pH 6

3.8% SO2, pH 7

7.4% SO2, pH 7

Figure 8: Manganese precipitation with different SO2/air ratios at 60°C, and pH 6 and 7.

0.1

1.0

10.0

0 30 60 90 120 150 180 210 240

Time (min)

Cr

co

ncen

trati

on

(m

g/L

)

1.9% SO2, pH 6

3.8% SO2, pH 6

7.4% SO2, pH 6

3.8% SO2, pH 7

7.4% SO2, pH 7

Figure 9: Chromium dissolution with different SO2/air ratios at 60°C, and pH 6 and 7.


10

3.2.3 Effect of temperature

The effect of temperature was investigated at 50°C, 60°C and 80°C, which well covered the ranges in a typical counter-current decantation (CCD) circuit in a nickel laterite operation. The rate of manganese precipitation increased with increasing temperatures in the range of 50-80°C (Figure 10). The target < 10 mg/L manganese concentration was obtained within 80 minutes at 50°C compared to 50 minutes at temperature range of 60-80°C. These results demonstrated that the SO2/air system worked well in a wide temperature range for the treatment of the slurries or solutions from a CCD circuit.

The chromium re-dissolution consistently increased with increasing temperatures from 50°C to 80°C (Figure 11). However, the chromium concentrations at different temperatures for obtaining the target <10 mg/L manganese were almost the same at about 1 mg/L.

The precipitation of magnesium below 60°C was insignificant (Figure 12), but at 80°C, it consistently increased to about 10% with time up to 60 minutes at which the manganese precipitation completed, but little increase with extended reaction time. It appears that this fact is more related to the adsorption or ion exchange mechanism rather than hydroxide precipitation.

0.1

1.0

10.0

100.0

1000.0

10000.0

0 30 60 90 120

Time (min)

Mn

co

ncen

trati

on

(m

g/L

)

50°C

60°C

80°C

Figure 10: Manganese precipitation at different temperatures, pH 7 and 3.8% SO2/air.


11

0.1

1.0

10.0

0 30 60 90 120

Time (min)

Cr

co

ncen

trati

on

(m

g/L

)

50°C

60°C

80°C

Figure 11: Chromium dissolution at different temperatures, pH 7 and 3.8% SO2/air.

0

10

20

30

40

50

60

70

80

90

100

0 30 60 90 120

Time (min)

Mg

pre

cip

itati

on

(%

)

50°C

60°C

80°C

Figure 12: Magnesium precipitations at different temperatures, pH 7 and 3.8% SO2/air

The variations of nickel and cobalt concentrations at different temperatures during the precipitation were generally similar to the trends of the manganese precipitation. However, a significant enhancement of both nickel and cobalt precipitations was observed at 80°C. In addition to the adsorption mechanism as discussed earlier, a higher temperature may also favour the formation of nickel and cobalt hydroxide precipitate due to an improved solubility.


12

3.3 PRECIPITATION USING THE SOLUTION WITH SO2/AIR

Two parallel tests using the solution separated from the neutralised slurry were carried out at pH 6 and 7 under the same other conditions. The results are compared in Figure 13. At both pH 6 and 7, the rates of manganese precipitation with the solution were slower than that with the slurry, but the difference in time for obtaining the target < 10 mg/L manganese was less than 30 minutes. The possible cause could be related to the variations of dissolved gas ratios in the presence and absence of solids and the presence of large amounts of Fe and Al precipitates in the slurry which could favour adsorptive precipitations.

Chromium concentration in the feed solution filtered from the neutralised slurry was <0.2 mg/L. It unexpectedly increased to the range of 0.3-0.45 mg/L with reaction time (Figure 14). This could result from small amount of residual solid adhered to the reactor walls in the previous slurry tests or some very fine colloidal hydroxides passed through the filtration. In practical application, the clarity of the solution may need to be controlled to minimise dissolution of chromium from the solid carried over.

The behaviour of magnesium co-precipitation in both cases was similar. Cobalt readily decreased to <0.2 mg/L at about 90 minutes in both cases. Nickel precipitations appeared to be faster to approach <0.2 mg/L using the solution than that using the slurry at pH 7.

0.1

1.0

10.0

100.0

1000.0

10000.0

0 30 60 90 120 150 180 210 240

Time (min)

Mn

co

ncen

trati

on

(m

g/L

)

Mn using slurry, pH 6

Mn using solution, pH 6

Mn using slurry, pH 7

Mn using solution, pH 7

Figure 13: Manganese precipitation using the slurry and the solution at 60°C and 3.8% SO2.


13

0.1

1.0

10.0

0 30 60 90 120 150 180 210 240

Time (min)

Cr

co

ncen

trati

on

s (

mg

/L)

Cr using slurry, pH 6

Cr using solution, pH 6

Cr using slurry, pH 7

Cr using solution, pH 7

Figure 14: Chromium dissolution using the slurry and the solution at 60°C and 3.8% SO2.

3.4 ASSAYS FOR CADMIUM, NICKEL AND COBALT TO LOWER LIMITS

Typical aqueous samples from different precipitation stages were selected and sent to Australian Laboratory Services (ALS) for assaying cadmium, nickel and cobalt to 0.001 mg/L detection limits with ICP-MS. The assay results are given in Table 3. The variations of the metal concentrations using the slurry and the solution were essentially similar. The possible explanations for the assay results are discussed below.

Table 3: Assay results by ICP-MS at ALS

Solution Concentration (mg/L)

Cd Co Ni

S1: Original slurry (59% solid) 0.004 13.2 139.5

S2: Slurry neutralised to pH 5 0.780 6.47 35.5

S3: Oxidative pptn at pH 7 for 60 min using the slurry 0.040 0.040 0.267

S4: Oxidative pptn at pH 7 for 120 min using the slurry 0.010 0.005 0.230

S5: Oxidative pptn at pH 7 for 60 min using the solution 0.050 0.008 0.218

S6: Oxidative pptn at pH 7 for 120 min using the solution 0.005 0.008 0.168

3.4.1 Variations of cadmium concentrations

Based on the solubility of Cd(OH)2, the precipitation of a very low level of cadmium would occur at pH above 10. Therefore, any change in the cadmium concentration at pH <7 should result mainly from the mechanism of adsorption on or desorption from metal (Fe, Al, and manganese) hydroxides and oxides.

The cadmium concentration in the solution (S1, pH 1.6) from the original slurry (59% solid) was 0.004 mg/L, but it became 0.78 mg/L in the solution S2 from the slurry (28% solid) which was neutralised to pH 5 with limestone. This indicates that the dilution of the original slurry from 59% to 28% solids (w/w) with the addition of required metal salts in sea water might cause interactions of the species and desorption of cadmium from the iron and aluminium hydroxides, resulting in more soluble cadmium into the solution.


14

The cadmium concentration was decreased to 0.04-0.05 mg/L at 60 minutes during the oxidative precipitation of manganese with SO2/air (solution S3 and S5) and further to 0.005-0.01 mg/L at 120 minutes (solution S4 and S6). This suggests that the manganese oxides (MnO2, Mn2O3, MnO(OH)) formed during the oxidative precipitation with SO2/air were effective adsorbents for removal of soluble cadmium species in solution.

3.4.2 Variation of cobalt and nickel concentrations

The solution (S1) from the original slurry (59% solid) contained about 13 and 139 mg/L cobalt and nickel, respectively. After dilution to about 28% solids, it contained about 8 mg/L cobalt and 60 mg/L nickel based on the assay by ICP-AES at CSIRO Minerals. The pre-neutralisation with limestone to pH 5 resulted in 6.5 mg/L cobalt and 35 mg/L nickel remaining in the solution S2. The oxidative precipitation with SO2/air effectively decreased the cobalt to 0.040 mg/L at 60 minutes (S3), and further down to 0.005-0.008 mg/L at 120 minutes (S4 and S6). This could result from both actions of adsorption and precipitation of hydroxide or oxide in the SO2/air oxidising system. In comparison, the nickel decreased to the range of 0.218-0.267 mg/L at 60 minutes (S3 and S5) when the oxidative precipitation of manganese was completed, but there was insignificant further decrease with an extended time to 120 minutes (S4 and S6), confirming a dominant adsorption mechanism as discussed earlier.

4. LIME CONSUMPTION

In the SO2/air oxidative precipitation system, lime neutralisation reagent is mainly consumed by neutralisation of acids produced by the oxidation of Mn(II) to MnO(OH)/Mn2O3 (reaction 3) and manganeseO2 (reaction 4), catalysed side reaction (5) and hydroxide co-precipitation of magnesium (reaction 6):

2Mn2+

+ SO2 + O2 + 3H2O → 2MnO(OH) + SO42-

+ 6H+ (3)

Mn2+

+ SO2 + O2 + 2H2O → MnO2 + SO42-

+ 4H+ (4)

SO2 + 1/2O2 + H2O → SO42-

+ 2H+ (5)

Mg2+

+ 2H2O → Mg(OH)2 + 2H+ (6)

The lime consumption with different pHs, SO2/air ratios and temperatures are compared in Figure 15. At each pH with a fixed 3.8% SO2 and 60°C, the lime usage generally increased with reaction time, but it was not always proportional to the amounts of manganese precipitated. The rate of manganese precipitation and the time for obtaining the target manganese concentration was crucial for minimum usage of lime. For example, the lime usage at pH 7 for obtaining the target manganese concentration was about 10 g/L at 50 minutes compared with about 16 g/L at pH 6 and 130 minutes, and 26 g/L at pH 5 and 270 minutes. In fact, the lime usage at pH 6 was almost the same as that at pH 7 per unit time, but the total usage was higher resulting mainly from an extended reaction time. The lime consumption was slightly higher at higher pH and higher temperature, which could be due to small amount of co-precipitation of magnesium according to reaction (6).

Among the parameters tested, SO2/air ratios showed the greatest effect on the lime consumption. The amount of lime consumption with 7.4% SO2 was more than triple that with 3.8% SO2. This can be attributed to the slower oxidation rate due to less oxidising condition or longer reaction time needed for obtaining < 10 mg/L manganese (reactions 3-4) and excess SO2 to reaction with O2 to produce H2SO4 (reaction 5).

These results suggest that the optimum conditions for minimum usage of lime are consistent to those for maximum rate for oxidative precipitation of manganese, involving in an overall optimisation of pH, SO2/air ratio and temperature parameters.


15

0

5

10

15

20

25

30

35

40

45

50

0 30 60 90 120 150 180 210 240 270 300

Time (min)

Lim

e c

on

su

mp

tio

n (

g/L

slu

rry)

pH 5, 3.8% SO2, 60°C

pH 6, 3.8% SO2, 60°C

pH 7, 3.8% SO2, 60°C

pH 7, 7.4% SO2, 60°C

pH 7, 3.8% SO2, 80°C

Dashed line crossing:

< 10 mg/L Mn obtained

Figure 15: Lime consumption at different pHs, SO2/air ratios and temperatures

5. CONCLUSIONS

Removal of manganese from both GPNL slurry and solution with SO2/air has been investigated. T he results have shown that the oxidative precipitation with SO2/air is kinetically efficient under approaching optimum conditions in the pH range of 6-7, about 1.9 -7.4% SO2 in air and temperature of 50-80°C for using both the slurry and the solution. Under these approaching optimum conditions, a manganese concentration below 10 mg/L could be obtained within 50 minutes from initial about 3 g/L manganese with minimal dissolution of chromium (<1 mg/L) and little co-precipitation of magnesium, although a slightly increasing trend of magnesium (< 10%) has been observed for a higher pH 7 and higher temperature 80°C. The oxidative precipitation has been shown to favour the removal of other metals such as nickel, cobalt and cadmium to very low levels. Conditions for minimum consumption of lime and maximum oxidative precipitation rate could be consistently optimised.

One test using magnesiumCO3 for manganese carbonate precipitation has been conducted and shown less efficient in terms of co-precipitation of magnesium at relatively high pH and consumption of lime. It is expected that further investigation with soluble carbonate salt such as sodium carbonate would improve the efficiency for partial removal of manganese from the circuit.

6. ACKNOWLEDGEMENT

The authors would like to thank the management of GPNL for review and approval for publishing this paper and Dr Dave Robison for internal review with valuable comments.

7. REFERENCES

1. Lide, D. R. and Frederikse, H.P.R. 2007, CRC Handbook of Chemistry and Physics, CRC Press, Inc., 88th Edition, London.

2. Rolia, E. and Dutrizac, J.E. 1984, The determination of free acid in zinc processing solutions, Canadian Metallurgical Quarterly, 23(2), 159-167.

3. Zhang, W., 2000. SO2/O2 as an oxidant in hydrometallurgy, PhD thesis, Murdoch University.


16

4. Zhang, W. and Cheng, C. Y., 2007a. Mn metallurgy review, Part I: Leaching of ores/secondary materials and recovery of electrolytic/ chemical manganese dioxide, Hydrometallurgy, 89(3-4), 137-159.

5. Zhang, W. and Cheng, C. Y., 2007b. Mn metallurgy review, Part II: Mn separation and recovery from solution, Hydrometallurgy, 89(3-4), 160-177.

6. Zhang, W. and Cheng, C. Y., 2007c. Mn metallurgy review, Part III: Mn control in zinc and copper electrolytes, Hydrometallurgy, 89(3-4), 178-188.

7. Zhang, W., Singh, P. and Muir, D 2000, SO2/O2 as an oxidant in hydrometallurgy, Minerals Engineering, Vol. 13, No. 13, 1319-1328.

8. Zhang, W., Singh, P. and Muir, D., 2002, Oxidative precipitation of Mn with SO2/O2 and separation from cobalt and nickel, Hydrometallurgy, 63, 127-135.


CHALLENGES OF MATERIAL SELECTION FOR PAL PLANTS

By

Naresh Balakrishnan

Consultant

Presented by

Naresh Balakrishnan

[email protected]

CONTENTS

1. INTRODUCTION 2

2. FEED PREPARATION 2

3. PRESSURE ACID LEACH 4

4. THICKENING AND NEUTRALIZATION 7

5. DOWNSTREAM PROCESSING 9

6. CONCLUSION 11

7. REFERENCES 12


2

1. INTRODUCTION

The gradual exhausting of sulphide ores has lead to reliance on lateritic sources of nickel and cobalt to meet demand. Lateritic sources exist close to the surface and while there is an inherent economic advantage in mining, there is requirement to use involved techniques like Pressure Acid Leaching (PAL) to successfully extract the metal economically from the deposit. PAL plants have operational parameters that require materials of construction to have a combination of wear resistance, corrosion resistance and strength at high temperatures. This combination of required properties makes the selection of materials a challenge because it is unrealistic to expect one material to meet all the necessary requirements. Material selection in some cases involves trade-offs between the competing requirements of corrosion resistance, wear resistance and strength, coupled with economics and ease of manufacture. The typical processing stages in PAL Operations are:

• Feed Preparation.

• Pressure Acid Leach.

• Thickening and Neutralization.

• Downstream Processing. Construction materials in typical PAL operations could include ordinary steels, high alloyed steels, titanium and other corrosion resistant exotic materials, brick lining, ceramics and plastics. Each of these materials has specific fabrication and inspection guidelines. Some of the materials are used plant-wide while others are limited to specific process areas. Material selection is guided by a combination of factors including intrinsic material properties, established engineering principles and operational or historical plant data. This paper attempts to look at material selection to address degradation aspects of equipment (corrosion and erosion) in some PAL operations. The Fitness for Purpose of plant equipment is governed by the ability of the materials of construction meeting required degrees of degradation resistance and retention of mechanical properties. The paper follows the above mentioned stages in PAL operations and looks at some material requirements and material selection decisions at the different processing stages.

2. FEED PREPARATION

OVERVIEW

Feed preparation involves processes like:

• Crushing

• Grinding if needed.

• Slurry preparation and transport to the processing facility. Material selection for equipment in the above mentioned processes is guided predominantly by material properties of wear resistance – with corrosion resistance being a secondary consideration in some cases. Wear conditions for operations like crushing and grinding are 3-body wear conditions involving 2 wear surfaces with a deliberate break down of material. Material strength requirements are high and both equipment construction and selection of wear resistant materials are based on proprietary designs, experience and testing. Wear conditions on slurry containments (pipelines, chutes, launders, pumps) are 2-body wear conditions wherein there is one wearing surface between the two bodies in contact. Size breakdown of the transported material is more a consequence of energy transfer and not the main intention. The discussion in this paper is mainly concerned with material selection for situations of 2-body wear in slurry transport.


3

2.1 WEAR CONSIDERATIONS

Wear of the slurry containment is influenced by the flow of slurry through it. Selection of wear resistant material may require a compromise to be made between the opposing requirements of toughness and hardness of the materials used for the containment. Flow of slurry imparts kinetic energy to the solid particles in the slurry. The particle velocity associated with this kinetic energy can have components perpendicular and parallel to the wall that contains and guides the slurry. The wear mechanism depends on the nature of interaction between the solid particles of the slurry and the walls of the containment. If the velocity is predominantly perpendicular to the wall of the containment then the energy transfer mode is one of impact. Repeated impacts work hardens the surface of the containment and in time can cause wear of the containment by chipping action. If the velocity is predominantly parallel to the wall then the wear is one of an abrasive type in which the slurry particles grind away the containment. The difference in hardness between the individual slurry particles and the hardness of the slurry containment is one of the factors that influence the wear rate for abrasive wear conditions. Apart from the energy considerations and particle hardness, the hydro-dynamical aspects of slurry flow also influence the wear characteristics. Some hydro-dynamical aspects that influence wear resistance are the volume concentration of solids (Cv) and the settling characteristics of the slurry. Non-settling slurry has a constant particle distribution over the cross-section of flow. The slurry behaves like a continuum. If slurry is settling then the slurry distribution over the cross-section is not uniform and slurry profile is one of a settled solid and a supernatant liquid. The wear resistance of some materials to slurry flow is listed in Figure 1 below. The numerical values represent wear resistance with respect to Mild Steel which has been assigned a value of 1. A higher value denotes a more wear resistant material.

Material of Slurry Containment Set 1 Set 2 Set 3 Set 4 Set 5

Hi Chromium White Cast Iron - 17-18.9 - - -

Sintered Tungsten Carbide - 16 - - -

Zirconia/Alumina Ceramic 10.9 20.9 13 - -

Ni-Hard Steel 8.6 3.8 - - -

Polyurethane 7.4-8.2 - 7.7-19 3.7-33 2.5-3.9

Rubber 2.7-12.5 7.2 - - -

Ceramic-Epoxy - 5.0 - 8 -

HDPE 1.9-2.4 - - 0.24 0.6

Stainless Steel 1.3 - 3.2 - -

Figure 1: Wear Resistance of materials relative to Mild Steel

(Sourced from “The Design of Slurry Pipeline Systems”, Short Course Notes, Paterson and Cooke Consulting Engineers, 14-16 March 2001.)

Slurry composition for each set is given below:-

Set1: 15µ, 150µ, 1500µ emery – Cv up to 15%

Set2: 750µ-4000µ diamond bearing heavy metal concentrate - Cv 12%

Set3: Sand Slurry d50= 1000µ, Cv=20%

Set4: 8000µ crushed gold ore. Cv =20%

Set5: 210mm granite gravel d50=1500µ, CV=20% d50 value indicates that more than 50% of the particle size is lesser than the value indicated. Cv value in percentage is the volume concentration of solids in the slurry. 2.2 MATERIAL SELECTION FOR WEAR RESISTANCE

Material characteristics that are needed for each type of wear mechanism are different. Impact wear resistance requires materials that are tough and have energy absorbing abilities. Tough steels and elastomers that are resilient are options available to combat impact wear. Abrasive wear requires hard materials to combat the grinding action at the interface of the slurry particle and the


4

containment. Induction hardened pipe or pipes with abrasive resistant weld overlay are some options available for abrasive wear conditions. The corrosive nature of some slurry dictates that material selection should consider the need for corrosion resistance in addition to wear resistance. Materials such as rubber lined carbon steel, fibreglass, HDPE lined pipe and plastic piping are commonly considered for corrosive slurry applications. Applicability of commonly available materials/material combinations for different types of slurries are listed below. The ultimate choice of a particular material depends on material resistance to degradation (erosion and corrosion), manufacturing methods and expected life of the system.

• Mild Steel: Mild Steel Pipe has low surface hardness (120-150 BHN), good impact resistance and strength. It does not have corrosion resistance properties and so is suitable for non–abrasive and non-corrosive slurries.

• Abrasion Resistant Pipe: The surface hardness of Abrasion Resistant Pipe is higher than Mild Steel (200-220 BHN). It has moderate strength and moderate impact resistance. It has no corrosion resistance and can be considered for mildly abrasive and non corrosive slurries. The average life expectancy of abrasion resistant pipe is about 1.5-2 times that of a Mild Steel pipe.

• Induction Hardened Pipe: Has a high surface hardness (400-600 BHN) and moderate impact resistance making it suitable for abrasive slurries. It does not have corrosion resistance and so has the limitations on use for corrosive slurries.

• Fibre Glass Pipe: Does not have bulk hardness. The wear resistance for the pipe comes from having abrasive articles impregnated into the inner surface layer. It has low strength but has corrosion resistance and hence finds use in low pressure abrasive and corrosive slurry applications.

• Rubber Lined Pipe: Rubber Lined Pipe makes use of Rubber to provide the degradation resistance (corrosion and erosion) and a carbon steel backing pipe to provide the strength. The surface of the rubber lining is rougher than steel piping and so use of a rubber lined pipe may involve more friction loss than an equivalent steel pipe. Depending on the erosive and corrosive properties of the slurry a suitable grade of rubber can be selected. Operational temperature limitations and the fact that rubber lined pipe must be spooled into individual flanged lengths requiring many flanged joints are a disadvantage.

In summation, selection of wear resistant materials involves considering material properties of slurry containment, bulk properties of the slurry and the hydro-dynamical properties of slurry flow.

3. PRESSURE ACID LEACH

OVERVIEW

The typical leach operation can be summarised as below:

Leaching operations are staged processes with varying pressure and temperature conditions. The requirements guiding material selection are a combination of wear resistance, corrosion resistance and material strength at conditions of high temperature and pressure for the following reasons:

• Slurries are abrasive so wear resistance is needed.

• The leaching makes use of concentrated sulphuric acid and so corrosion resistance is needed.

Ore Slurry

+

Steam

+

Acid

Leach

Process

Leach

Solution

+

Residual

Acid


5

• Pressure acid leaching is typically carried out at temperatures in excess of 250°C for reduction in acid consumption.

The commonly used material to satisfy the mix of requirements is either a noble metal like Titanium or acid brick lining. Carbon Steel is not suitable for acid leach operations because of its lack of corrosion resistance at leaching temperatures. Figure 2 below shows the corrosion resistance of materials in a combination of acid and chloride environment. The temperature effects are not considered in this graph.

Tantalum

Zirconium

Hastelloy B

Ti- 0.15 Pa

Ti-12

CP Titanium

Hastelloy C

Monel

Zirconium

Hastelloy C

Monel

Inconel

Stainless Steels

Oxidising Acid Strength Reducing Acid Strength

No

n C

hlo

ride

C

hlo

ride

Figure 2: Range of Corrosion Resistance of Materials

Source: Hinshaw, E and Moser, K; Understanding Hydrogen, Tantalum and Niobium Materials of Construction, H.C Stack Inc

Certain inferences can be derived from the above:

• Tantalum has the best corrosion resistance across a wide range of acid strengths in a chloride environment.

• Zirconium and Hastelloy B accommodate high chlorides in a reducing acid environment but are not suitable for an oxidising environment.

• Titanium alloys have good corrosion resistance in moderately high chloride environments across a range of acid strengths and are the preferred choice for use in aggressive environments. Titanium alloys with Palladium addition have more corrosion resistance than those with Molybdenum and Nickel addition. Alloyed Titanium also has good wear resisting properties.

• Stainless Steels have good resistance in some oxidising acid environments in the absence of chlorides.

The following table (Figure 3) shows the evolution of various Titanium alloys for use in the Chemical and Mining Processing Industries. The existing grades of commercially pure Titanium are modified by the addition of alloying elements to enhance the corrosion resistance. The Strength Ratio is the ratio of the strengths at 260°C (close to the upper limit of Leaching Temperatures) and ambient temperatures.


6

Material Grade

Basic Grade

Additional Element(s)

Yield Strength

Ultimate Strength

Elongation at Rupture

Strength Ratio

Ti-1 Ti-1 - 170 240 24% 0.47

Ti-11 Ti-1 0.15 Pd 170 240 24% 0.47 Ti-17 Ti-1 0.05 Pd 170 240 24% 0.47 Ti-27 Ti-1 0.10 Ru 170 240 24% 0.47

Ti-2 Ti-2 - 275 345 20% 0.53

Ti-7 Ti-2 0.15 Pd 275 345 20% 0.53 Ti-16 Ti-2 0.05 Pd 275 345 20% 0.53 Ti-26 Ti-2 0.10 Ru 275 345 20% 0.53

Ti-9 Ti-9 - 483 620 15% 0.70

Ti-18 Ti-9 0.05 Pd 483 620 15% 0.70 Ti-28 Ti-9 0.10 Ru 483 620 15% 0.70

Ti-12 0.3 Mo, 0.8 Ni 345 483 18% 0.65

Figure 3: Titanium Composition and Strengths

Source: ASME Boiler and Pressure Vessel Code: Materials: Section II Part B: Non-Ferrous Material Specifications

Comparison of the information in Figure 2 and Figure 3 shows that there are a variety of Titanium alloys available that can meet the dual requirements of strength and degradation resistance. The suitability of a Titanium alloy for a particular application is assessed by review of the specific leach operation. Unalloyed Tantalum which has the best corrosion resistance does not have the strength of Titanium (172 MPa.Tensile strength and 103 MPa. Yield Strength at room temperature) and so its use in leach operations is restricted to situations requiring extreme corrosion resistance. Commonly used equipment for leach operations include tanks, pressure equipment, piping, valves and pumps. Tanks are used to store slurry either prior to leaching in the autoclave or store the leached slurry prior to thickening and neutralization. Carbon Steel and Stainless Steels are common choices for construction of storage tanks. Slurry pumps transport the slurry to various processing equipment within the leach operations. Design of slurry pumps, their material selection and manufacture is based on proprietary design and often decided by slurry pump suppliers themselves. The following discussion highlights some important classes of other equipment for leach operations and the material selection principles involved in their construction. Pressure Equipment and Pipe Spooling: Pressure equipment like autoclaves, heater and flash vessels make use of clad materials. The clad material is typically Carbon Steel backing + Titanium lined plate wherein, the Carbon Steel provides the strength and the Titanium lining (commercially pure or alloyed grade of Titanium) takes care of the corrosion and abrasion requirements. Titanium-1 based alloys with lower strengths are widely used for these applications. Cladding of carbon steel is commonly done by explosion bonding that plastically deforms the Titanium and so the low yield strengths of the Titanium-1 family of alloys is an inherent advantage in forming of the clad plate. The use of Titanium alloy for pipe spooling including pipes and fittings makes it necessary to select a grade of Titanium that meets the dual requirements of strength and resistance to material degradation (by both corrosion and erosion.) Alloyed versions of Titanium Grade 9 or Titanium Grade 12 are options that have worked successfully in practice. Low pressure equipment with relatively thin walls make use of solid Titanium with material compositions that meet specific resistance to the processing media.

Brick Lined Pressure Equipment: Brick Lined Pressure Equipment was a precursor to Clad Equipment and finds use in some leach operations. The construction is based on having acid resistant bricks inside a pressure vessel to provide both corrosion resistance and refractory properties. The design of brick lined pressure equipment needs to consider the stress on the shell that come about because of thermal swelling of the bricks in addition to the pressure and mechanical stresses in the shell. Fabrication of the shells is also critical with rolling tolerances for the shell more stringent


7

than that required for normal pressure equipment. In addition, grinding of weld reinforcements on the inside of a vessel flush to enable a smooth surface is to be carried out. Acid Bricks are selected by considering the requirements of acid resistance and resistance to spalling of bricks. The construction of the brick lined vessel consists of a steel shell, a membrane (lead lining, epoxy) and acid resistant bricks. Selection of the brick and the thickness of the brick lining (number of brick courses) is based on the refractory properties of the brick to ensure that the membrane temperature between bricks and the shell does not exceed the melting point of either the lead or the glass transition temperature of the epoxy resin. Choke Tubes and Valves: Pressure acid leach operations make use of flash steam to pre-heat slurry feed to the autoclave. Leached slurry from the autoclave is flashed progressively into a series of flash vessels. The steam from slurry depressurisation is used to preheat the slurry going into the autoclave. Depressurisation of leached slurry is done by devices known as choke tubes. The choke tube is subjected to extreme pressure stresses due to depressurisation and conditions of wear and corrosion because of the slurry. Material selection for the choke tubes to meet these conditions is critical and a combination of ceramic and titanium has been used successfully. A typical choke tube construction consists of a ceramic inner tube encased within a Titanium tube. Since the choke tube is within a pressure vessel both the inner and outer surfaces are subject to corrosive environments and so the outer Titanium is selected on corrosion resistance requirements. The ceramic is selected for both corrosion and wear requirements. Ceramics have good compressive strength but are weak in tension. So the design of the choke tubes is based on keeping the stresses in the ceramic compressive in nature. Valves used for slurry service are subjected to aggressive corrosion and erosion conditions. The Valve Body is usually made from a grade of titanium similar to that of the associated process piping. Trims are the internal parts of a valve. The trim for a typical valve like a ball valve include the ball, valve seat and the spindle. Trims are subject to more erosive wear than the main valve body and need more erosion protection than the valve body. The trims are usually made of same material as the main valve body but have their wear resistant property enhanced by ceramic coatings. The selection of wear resistant ceramic coating is an involved process requiring immersion testing to ensure corrosion resistance – in addition to mechanical testing to check bonding of the ceramic to the metal substrate. Immersion testing includes immersing a ceramic coated specimen in a medium that simulates the actual operational medium. Mechanical testing to check bonding of the ceramics to the substrate involves test methods like ASTM C633. Acid Lances: An acid lance is a pipe that delivers the sulphuric acid close to the agitator in the autoclave compartment. This ensures contact between the slurry and the acid to initiate the leaching process. Dilution of the concentrated acid and the presence of steam results in a local area of high temperature in the vicinity of the lance. Material for the lance should meet the requirements of acid corrosion resistance at high temperature and Tantalum is preferred choice of material for this onerous requirement. Tantalum has good corrosion resistance but does not have the strength at high temperatures. This results in the acid lance being of a bi-metallic construction with a relatively strong material like Titanium on the inside and a corrosion resistant material like Tantalum on the outside. Leaching is the primary step in the extraction of Nickel and Cobalt from laterite deposits and so leach facilities are the heart of a PAL plant. Process conditions in leaching are the most onerous and improper selection of materials can lead to expensive downtime and repairs.

4. THICKENING AND NEUTRALIZATION

OVERVIEW

The output from the leach stage is an acidic solution that contains leached metal solution and undissolved residue and probably some excess acid. The processes that happen in this stage are thickening to separate the leach solution from the leach residue. The excess acid may need to be neutralized before downstream processing can take place. This stage may also involve removal of metals that would be detrimental to downstream operations.


8

Material selection for equipment in this stage of the processing are mainly governed by corrosion requirements and to a lesser extent wear requirements. The temperatures encountered in this stage are generally not as high as leaching and so temperature dependence of allowable material strengths is not a deciding factor in material selection. While temperature may not be an issue, material strength is still a deciding factor. Carbon and low-alloyed steels cannot meet the corrosion requirements and so the common choice of materials for this stage of processing (because of their material strength) is Austenitic and higher-alloyed Steels. Thickening and neutralization operations make use of media that may have varying degrees of acid concentration and may contain halides. The combined effects of acid and halides leads to increased corrosion resistance requirements and so material selection may include a variety of High Alloyed Steels. The table below (Figure 4) lists the strengths and compositions of some of the commonly used materials for equipment associated with thickening and neutralization operations.

Material Grade

Werkstoffe No

Yield Strength

Ultimate Strength

Elongation at Rupture

Cr Ni Mo Cu

S 30400 205 515 40 18 8

S 31600 205 515 40 16 12

S 31753 1.4439 240 550 40 18 14 4.5

S 31803 450 620 25 22 5 3

S 32750 550 795 15 25 7 4

S 32760 550 750 25

N 08904 1.4539 220 490 35 20 25 4 1.5

N 08020 240 550 30 20 34 2.1 3.5

N 08367 310 690 30 20 25 6

N 08926 1.4529 295 650 35 21 25 6.5 0.9

C-276 2.4819 310 750 30 16 57 16

Alloy 59 2.4605 340 690 40 23 59 16

Figure 4: Strength and Composition of Stainless and High-Nickel Steels Source: ASME Boiler and Pressure Vessel Code: Materials: Section II Part A: Ferrous Material Specifications

The following Table (Figure 5) shows possible material selections for combinations of acid strength and halides. The Werkstoffe numbers in each bounded region show the type and family of alloy that is suited to the combination of acid strength and halide concentration.

6.5 4.5 3.0 2.0 1.5 1.0

>100 ppm 1.4439

<500 ppm 1.4435

>1000 ppm 1.4539

<5000 ppm 1.4529

>10000 ppm 1.4562

<50000 ppm

>100000 ppm 2.4819

<200000 ppm 2.4605

pH

Low Halides

High Halides

Very High

Halides

Acid Strength Weakly acidic Acidic Very Acidic

Figure 5: Suitability of Materials for Acidic and Halide combinations. Source: VDM Report No 26- High Alloy Materials for aggressive environments.

Based on the above two Tables (Figures 4 and 5) and established industry norms, some guidelines for material suitability can be deduced:

• Austenitic Stainless Steels have moderate strength and are good for general corrosion resistance in non halide environments.


9

• Duplex Stainless Steels have High Strength and provide good corrosion and pitting resistance.

• Materials like N 08904 and N 08020 with copper additions have more resistance to acidic media particularly sulphuric acid.

• The 6 Molybdenum alloys like N 08367 and N 08926 have good applications for pitting situations with high halides.

• High Nickel alloys like Alloy 59 and C-276 have good resistance for high halide and very acidic solutions but are not as strong as Duplex Stainless Steels.

5. DOWNSTREAM PROCESSING

OVERVIEW

The leach solution that has been neutralised after leaching is processed in a variety of ways to obtain a usable end product. Some of the options available for downstream processing are:

• Sulphide precipitation for further refining.

• Mixed hydroxide precipitation followed by ammoniacal leach.

• Reduction after treating with hydrogen and ammonia.

• Solvent Extraction This discussion is limited to the Solvent Extraction option. Solvent extraction makes use of organic solvents that dissolve metal from the neutralized process leach solution and form organometallic compounds. These organometallic compounds are later on processed to yield the metal products. Solvent extraction operations occur at relatively low temperatures (60-120°C) and corrosion resistance is the requirement that drives material selection. Plastics and Glass Reinforced Plastic (GRP) are preferred choice of materials for solvent extraction equipment. Both these materials have adequate mechanical strength at ambient operating temperatures while the mechanical strength falls off for elevated temperatures. The reduction in material strength with rise in temperatures is lesser for GRP than for plastics and so GRP is the more preferred option of the two for process equipments like Tanks, Pressure Equipment, Filter Vessels and Piping. GRP is composed of glass in the form of mat, rovings or thread and a corrosion resistant resin to hold the glass together. The structure of GRP is in the forms of multiple layers on top of each other and is also referred to as a laminate. Glass gives the mechanical strength while the resin is both the binder and provider of corrosion resistance. The body of a GRP component (pipe, fitting, tank) consists of a corrosion resistant layer and a structural layer. Types of Glass typically used in GRP constructions are:

• C glass which is chemically resistant glass that is used in the corrosion resistant layer of the laminate.

• E-CR glass is glass that combines the dual properties of chemical resistance and strength.

• E glass is glass that is mainly for mechanical strength. Typical resins used are epoxy and vinyl ester. Vinyl ester resins are commonly used to provide corrosion resistance for a wide range of media.

The challenges associated with selection and fabrication using GRP are as follows:

• The design temperature of the equipment depends on the temperature limitations of the resin and is governed by the glass transition temperature of the resin. This is the temperature at which the resin turns brittle. Design codes usually dictate that a margin be maintained between the maximum operating temperature and the design temperature (approx. 20°C) Selection of resin depends on the ability of the resin to meet both the temperature and corrosion resistance criteria.


10

• Manufacture of equipment is a combination of both mechanised processing and manual operations. This can lead to a wide fluctuation in material properties and quality of the finished laminate. While shop manufacture can be automated and controlled to reduce the fluctuation in laminate quality, site work like joints on pipe spooling have to be completely hand made and so additional care is to be taken to ensure that the overall integrity of the fabrication is not compromised by varying degrees of processing.

• The processing liquids that one encounters in solvent extraction have low conductivities and there is the problem of build-up of electrostatic charge. This requires that the laminate be electrostatically dissipative. Laminates have to be engineered and installed to meet this requirement.

• A direct consequence of charge build-up is the need to have fire-retardant laminates in some areas of the plants. Again, laminates have to be engineered to meet this requirement. This is normally achieved by the use of fire retardant resins and the addition of fire retardant chemicals like Antimony Trichloride in the outermost layers of the finished laminate.

• For two phase liquids like aqueous-organic mixtures having low conductivity, the transport velocity is limited to 1 m/s. This has an important bearing on the size of the equipment and with size there is also the additional challenge of installation and site based fabrication.

• Identification of glass and resin to provide the required corrosion resistance are also important. Corrosion resistance data is normally available from resin suppliers. If corrosion data for a specific requirement is not available, tests have to be carried out and corrosion data evaluated before final selection. Such tests are usually time-intensive and this aspect has to be factored into any final resin selection.

6. CONCLUSION

A typical PAL plant has a wide variation in process conditions throughout the plant. This requires consideration of a variety of materials. This paper set out to identify material requirements of typical equipment and look at some of the material selection decisions. The table below is provided as a summary of the issues raised in the above paper including, material requirements, operating conditions, commonly used materials of construction and their manufacturing complexity

Processing

Stage

Materials of

Construction

Pressure Temperature Manufacturing

Complexity

Materials

Requirements

Feed Preparation Abrasive Resistant

Steels, Lined

Steel,Ceramics

Low to

medium

Ambient Low to Moderate High wear

resistance,

moderate

corrosion

resistance, low

to medium

strength

Pressure Acid

Leach

Alloyed Steel, Clad

Carbon Steel, Titanium

and other exotic

materials, Acid Bricks,

Ceramics

High High High High wear

resistance, high

corrosion

resistance,high

strength

Thickening and

Neutralization

Alloyed Steel, Carbon

Steel

Low to

medium

Ambient Moderate High Corrosion

Resistance,

Moderate wear

resistance,

Moderate

strength

Solvent Extraction GRP, Plastics Low Medium Moderate to High High Corrosion

Resistance.


11

11. REFERENCES

1. ASME Boiler and Pressure Vessel Code: Materials: Section II Part A: Ferrous Material Specifications, The American Society of Mechanical Engineers, NY.

2. ASME Boiler and Pressure Vessel Code: Materials: Section II Part B: Non- Ferrous Material

Specifications, The American Society of Mechanical Engineers, NY. 3. Bacon, G and Mihaylov, I; “Solvent Extraction as an Enabling Technology in the Nickel

Industry”, The Journal of the South African Institute of Mining and Metallurgy, Vol 102 (8) pp 435-443.

4. Cooke, R and Patterson A; The Design of Slurry Pipeline Systems, Short Course Notes by

Patterson and Cooke Consulting Engineers, 14-16 March 2001. 5. Heubner,U and Kohler,M; High Alloy Materials for Aggressive Environments, VDM Report No

26, ThyssenKrup VDM GmbH, Werdohl Germany, July 2002 . 6. Hinshaw, E and Moser, K; Understanding Hydrogen, Tantalum and Niobium Materials of

Construction, H.C Stack Inc.. 7. Klemm, Robert E; Abrasive Resistant Steel Piping Systems for Slurry Transport in Mining

Applications, Ultra Tech. Port Washington WI 53074-308 USA. April 1999. 8. Louie, D.K; Handbook of Sulphuric Acid Manufacture, DKL Engineering Inc.,2005. 9. Mellor, B.G, Surface Coatings for Protection against Wear, Woodhead Publishing Limited,

Cambridge,England,2006. 10. Metals Handbook: Vol.13 Corrosion, 9

th Ed,American Society for Metals, Metals Park

Ohio,1987. 11. Robinson, J and Chipman, S; Using Structural Ceramics in Reactive Metal Control Valves to

Increase Valve Service Life, Caldera Engineering, LC. 12. Vaughan, J. and Alfantazi, A; “Corrosion of Titanium and its Alloys in Sulphuric Acid in the

Presence of Chlorides”, Journal of Electrochemical Society, Vol 153 (1) B6-B12, 2006.


CONTROLLED RAPID AUTOCLAVE BLOWDOWN

By

K.Jackson

Mogas Industries, USA

Presented by

Kevin Jackson

[email protected]

CONTENTS

1. BACKGROUND 1

2. APPLIED TECHNOLOG 3

3. SIZING 5

4. CONCLUSION 6


2

(a)

(b)

Fig 1

(a)

Quench Vessel

Autoclave Vent Line

1. BACKGROUND

Traditionally brick lined autoclaves used in the High Pressure Acid Leach (HPAL) and Pressure Oxidation (POX) leaching process of minerals such as Nickel, Gold and Copper, have a requirement to be depressurized or “blown down” due to process upsets, emergency situations, allow access for essential maintenance, equipment replacement or failure.

The autoclave is fitted with a vent line that maintains the pressure in the autoclave. The vent line should not be confused with the “discharge” line that maintains the liquid level in the autoclave.

The vent line consists of isolation valves (fig 1 a) and a control valve (fig 1 b). Typical vent line sizes are 4” - 8” ANSI 600 class that contain two 8” high integrity severe service metal seated isolation ball valves that hold pressure in the autoclave when required and a 4” or 6” angle pattern globe control valve with actuator and 4/20mA control signal that controls the release of the elevated gas pressure from the autoclave into the quench vessel. This maintains desired pressure and prevents over pressurization of the autoclave.

Autoclave blow down takes place using the vent line. The blow down rate is governed by two factors 1) the brick lining is subject to thermal shock if the temperature inside the autoclave changes too quickly and 2) the capacity of the control valve.

Generally the limiting factor for autoclave blow down has not been the thermal shock of the brick lining as you would expect, but, the capacity of the globe control valve.

The initial design pressure in the POX autoclave is approximately 3600 kPag at 180

oC – 240

oC and HPAL operates

at approximately 6000 kPag at 240oC – 270

oC and pressure in

the quench vessel is atmosphere, so the capacity of the control valve to handle flowrates at this ∆P needs to be small, around 25 to 35 Cv.

When the autoclave needs to be blown down the angle pattern control valve releases pressure in the autoclave as it does in normal operating conditions, however as the pressure in the autoclave reduces the ∆P (difference in pressure in the autoclave and the quench vessel) the capacity required to pass the flowrate increases beyond the capacity of the angle pattern globe valve and therefore takes longer to depressurize the autoclave completely.

Globe control valves by design have limited rangeability (the ratio of the maximum controllable flow to the minimum controllable flow), this is because of their linear construction. Their ability to handle large scale differences in flow and/or pressure parameters inhibits their use in blowdown and other applications that require rangability above 40:1

The rangeability for the rapid autoclave blowdown is approximately 70:1

Typical angle pattern valve layout


3

2. APPLIED TECHNOLOGY

Within the oil & gas industry it has long been the practice to let pressure down across a multiple number of stages. There are a number of reasons for this, but primarily it is done when controlling gas flow to reduce velocity, noise, erosion and vibration. Of course this multi-stage letdown or torturous path technology has been applied in the past to relatively clean gases.

The principle operation of the torturous path for gas flow is to control the pressure drop at each right angle turn, thus limiting the velocity and noise. With less velocity there is less erosion and vibration.

The number of right angel turns in the FlexStream® trim can vary from 2 to 36, the number is dependent on the ∆P across the valve and flowrate.

The autoclave blowdown gas has some solid particle carryover, so by controlling the velocity the FlexStream® trim is limiting the speed at which the particles collide with the trim and therefore limiting the amount of potential erosion. It is crucial that velocity in the flow passages is not reduced to a level where solids can stall in the trim because they don’t have sufficient speed to pass through the path.

It is important that the size of the path is sufficient to allow the volume of particles to pass through the trim. The solids content for this particular application is estimated at a maximum of 200kg/hr or <2% by volume. Any particle that is too large to travel through the trim will gather at the bottom of the ball and be discharged when the valve is in its fully open position.

Because of the manufacturing techniques used in the production of the FlexStream® trim, path sizes can be changed by simply changing the thickness of the plates that make up the trim element.

Rangeability is the key issue as stated before, any additional control valve added into the piping arrangement must be capable of handling the changing pressure drop and flow rates associated with the depressurization of the autoclave. The unique design benefit of FlexStream® is to have

rangeability in excess of 300:1

Unlike a linear globe valve the FlexStream® trim is not linear and therefore not held by a bonnet assembly arrangement that compresses the trim and/or seat in place. The trim or seat is not screwed but is independently held within the ball by a removable retention ring. This means that the ball ID can be filled from 10% to 100% to ensure that the right amount of trim can be inserted to deal with the high ∆P and flowrate where required. As ∆P reduces the ball is opened up producing the hydraulic diameter that allows low ∆P and high flow rates.

Transition across the flow cases may require less pressure letdown turns due to the reduced ∆P, this can be accommodated within the FlexStream® trim element but by changing the number of pressure letdown turns across

each torturous path, in some cases just down to the drilled exit holes.


4

The advantage with the trim element being in the ball is that the surrounding valve body, stem, seals, gaskets and actuators are interchangeable with the existing isolation valves, this means less training for site or workshop based maintenance staff.

Therefore there is no introduction or concern surrounding the viability of the base valve design as it is already proven in use.

The revised vent line system layout would include for an additional 8” pipe run connecting the new rapid blow down line to the quench vessel, the FlexStream® control valve, actuator, 4/20mA positioner and additional I/O counts in the PLC/DCS.

The additional FlexStream® rotary control valve would be fitted with an actuator capable of accepting a 4/20mA control signal for precise position control, this actuation can be achieved with pneumatic, hydraulic or electric power sources.

The normal mode of operation would be exactly the same as before, the control valve © would be closed and control valve (b) would be regulating pressure in the autoclave. If then the autoclave was called upon to depressurize, the control valve (c) would position itself using the control loop and based upon temperature the valve would position its self to allow gas and pressure to pass into the quench vessel. This process would continue until the autoclave had reached atmospheric pressure and the normal safety procedures for entry could be made.

This system is best incorporated in initial plant design and layout, but can also be retrofitted.

(a)

(b)

Fig 2

(a)

Quench Vessel

Autoclave Vent Line

(c)


5

3. SIZING

The FlexStream® sizing that has been carried out in accordance with ISA S75-01. The chart above shows the 5 basic operating conditions of the valve. At each case the flowrate, ∆P, Cv, noise, kinetic energy and velocity is calculated to ensure they are not above the limits allowed by the standard or best practices for the industry.

The sizing shows that the Max and Norm cases have a Cv of 30 or less at a flow rate of between 3,573 and 1,989 kg/s with declining inlet and outlet pressures. These flow cases are passing through the 8 turn torturous paths with velocity and noise control taking place due to the significant ∆P present.

This can be seen on the Cv – Stroke graph below

The inlet pressure reduces as the pressure in the autoclave decreases and the outlet pressure to the quench system also decreases. The flowrate decreases but has less impact on Cv as the ratio of inlet to outlet pressure remains steady. As the pressure in the autoclave decreases the gas volume increases so it takes longer to release the gas through the restricted orifice in the angle control valve. With the high rangeability in the additional valve it simply opens forth to allow the higher Cv to pass the increased gas volume.

At the Min case the ∆P is 2kPag and requires a Cv of 1303 to be able to pass the flowrate of 1,989 kg/s, in this position, i.e. with the valve fully open, some of the flow will pass through the torturous paths but the majority will flow down the hydraulic diameter.

FLUID Autoclave Gas/Vapor

UNITS Max Norm Norm Norm Min

FLOW RATE kg/s 3.573 3.573 2.871 1.989 1.989

INLET PRESSURE KPa(g) 3618 3352 2352 1352 100

OUTLET PRESSURE KPa(g) 0 2352 1352 100 98

INLET TEMPERATURE Deg. C 230 230 165 100 100

MOL. WEIGHT 32 32 32 32 32

RATIO OF SPECIFIC HEAT 1.54 1.54 1.42 1.42 1.3

COMPRESSIBILITY 0.850 0.850 0.925 1.000 1.000

INLET DENSITY lbm/ft3 2.091 1.941 1.455 0.936 0.130

KINETIC ENERGY Psi 61.00 3.44 3.58 8.47 0.31

CALCULATED CV 20 30 29 27 1303

STEM TRAVEL % 32.1% 42.6% 42.0% 40.2% 98.6%

SPL AT 1 METER dBA 89.4 75.0 75.0 75.0 75.0

OUTLET PIPE VELOCITY Mach 0.144 0.011 0.015 0.060 0.084

VALVE OUTLET VELOCITY Mach 0.148 0.011 0.015 0.061 0.087

P

R

O

C

E

S

S

.

D

A

T

A

Cv - Trim / Open / Turns 30 1302 8 30 1302 8 30 1302 8 30 1302 8 30 1302 8


6

4. CONCLUSION

By using the controlled rapid autoclave blow down rotary control valve, blow down times for the autoclave can be reduced from 24+ hours to approximately 13 hours. This represents a significant reduction in turnaround times for the autoclave and can have dramatic impacts on run times and costs.

Irrespective of the process, HPAL or POX, the impact of this reduced depressurization time is reflected in increased production and will always be significant to the plants efficiency. In the case of gold (our example) this is magnified due to the base metal price.


TITANIUM: A SOLUTION FOR HIGHLY CORROSIVE

HYDROMETALLURGICAL APPLICATIONS- ALLOY SELECTION, CLADDING

AND FABRICATION

By

John Banker

Dynamic Materials Corp., USA

and

Bruce Craig

MetCorr, USA

Presented by

John Banker

[email protected]

CONTENTS

ABSTRACT 2

1. INTRODUCTION 2

2. TITANIUM ALLOYS FOR HYDROMETALLURGICAL APPLICATIONS 2

3. CORROSION CONSIDERATIONS IN PAL AND POX 4

4. COST CONSIDERATIONS 10

5. CLAD MANUFACTURE AND FABRICATION 10

6. TITANIUM AND STEEL ALLOY CONSIDERATIONS FOR CLAD 11

7. TITANIUM AND TITANIUM CLAD EQUIPMENT DESIGN AND FABRICATION 12

8. CONCLUSIONS 14

9. REFERENCES 14


2

ABSTRACT

Hydrometallurgical processes can offer a cost effective, environmentally friendly solution for recovery of many

metals from their ores. These processes involve leaching the ores with aqueous solutions of common industrial

acids. They can range from atmospheric heap leach operations to more sophisticated pressure acid leaching

processes (PAL) and pressure oxidation leaching processes (POX). Titanium is the metal of choice for

autoclaves, piping, and other vessels for many of these aggressively corrosive processes. When there are

additional requirements for high temperatures and/or high pressures, titanium-clad is a reliable and cost effective

alternative. Corrosion considerations, titanium alloy options, cladding processes, equipment fabrication, and

other factors necessary to assure reliable equipment are addressed.

1. INTRODUCTION

Historically, pyrometallurgical processes have been the dominant method for recovering many metals from their ores. These processes have tended to be both energy intensive and environmentally unfriendly. Hydrometallurgical processes provide a lower cost, environmentally friendlier alternative for production of many metals. These processes involve leaching the ores with aqueous solutions of common industrial acids. Acid leaching processes have been proven commercially viable for extraction of copper, gold, nickel, uranium, molybdenum, zinc and other metals [1]. These hydrometallurgical processes can range from heap leach operations to more sophisticated high pressure acid leaching (PAL) and pressure oxidation leaching (POX).

The aggressive acid environments of most leaching processes present significant corrosion concerns. Low temperature-pressure processes, such as heap leaching, are commonly contained with plastics or other non-metallics. The high temperatures and pressures of the PAL and POX operations require containment in metal pressure vessels, commonly called autoclaves. Corrosion protection in the autoclaves is typically achieved using corrosion resistant alloys (CRA) or acid resisting brick; however, clad metals offer potentially significant cost saving as temperature, pressure and size increase. Proper selection of appropriate corrosion resistant alloys requires both a good knowledge of operating environment and strong knowledge of corrosion mechanisms. Alloy testing in the actual slurry environment is frequently performed in order to determine which alloys may be suitable for a specific application. Alloys commonly used in hydrometallurgical process equipment range from austenitic stainless steels and duplex stainless steels to nickel based alloys (ie. Inconels, Hastelloys, etc) to titanium alloys. This paper focuses on the use of Ti alloys in PAL and POX applications and presents the corrosion considerations that must be addressed for proper alloy selection.

2. TITANIUM ALLOYS FOR HYDROMETALLURGICAL APPLICATIONS

For reasons explained in the next section the process conditions encountered in PAL and POX are so severe that steel and other alloys such as stainless steels are not suitable for these applications. Furthermore, the requirements to contain relatively high pressure at temperatures of 250 to 300

oC require that large vessels such

as the autoclaves be internally clad to keep the costs down. This means that some alloys must be able to be capable of being cladded onto steel and others must be available as pipe and fittings. The most common grades of Ti used for PAL and POX are listed in Tables 1 and 2.


3

Table 1: Titanium alloys commonly used in PAL and POX equipment.

Titanium Grade* Composition Features Cost Ratio

1 Chemically Pure (CP) Low cost & good fabricability 1.05

2/2H CP strengthened w/O & Fe Similar to Gr 1 - higher strength 1.00

5 Ti + 6.0% Al + 4.0% V Much higher strength + erosion resistance 1.5 to 2.0

7/7H Gr 2 + 0.2% Pd Better corrosion performance than Gr 2 1.50

11 Gr 1 + 0.2% Pd Better corrosion performance than Gr 1 1.55

12 Ti + 0.3% Mo + 0.8% Ni Higher Strength 1.20

16/16H Gr 2 + 0.05% Pd Corrosion performance of Gr 7 but lower cost 1.30

17 Gr 1 + 0.05% Pd Corrosion performance of Gr11 but lower cost 1.35

36 Ti + 45% Nb Oxidation Resistance 5.0

* ASTM B265 Grade Designations

Table 2: Titanium product forms commonly used in Hydromet applications.

Titanium Grade Product Form Application in Hydromet Processes

2/2H Plate Low pressure autoclaves

7/7H Sheet Preheaters

12 Pipes Piping

16/16H Fittings Transfer lines

1 Autoclaves

11 Preheaters

17

Clad on Steel

Heat Exchanger Tubesheets

5 Bar, Plate Agitator Shafts,

7 Castings Hubs and Blades

36 Pipe and Tube Spargers


4

3. CORROSION CONSIDERATIONS IN PAL AND POX

The process conditions for PAL and POX are addressed separately here since the absence of oxygen in the PAL and its presence in POX has an important effect on Ti that must be considered.

Figure 1: Simplified Diagram of a Pressure Acid Leaching (PAL) circuit.

3.1 PRESSURE ACID LEACHING (PAL)

Figure 1 shows a simplified diagram of a PAL circuit [2]. The most significant part of the entire plant and also the most expensive are the high pressure autoclaves. These autoclaves are typically constructed from heavy wall carbon steel plate that had been clad with Grades 1, 17 or 11 titanium using the explosive bonding process. The autoclave environment is extremely aggressive consisting of concentrated sulfuric acid injected into the autoclave at 250

oC which dissolves the ore releasing metals such as Ni and Co. At first glance the use of any Ti alloys in

concentrated H2SO4 would be a poor choice as demonstrated in Figure 2 [3]. The corrosion rates would be

extremely high and Ti would not be a suitable material for these autoclaves. However, of great importance to the overall corrosion reactions is that during PAL the ore also releases ferric, cupric and chromic ions that are all strong oxidizing ions that help passivate Ti as shown in Table 3 [Ref 3]. The importance of these oxidizing ions in the use of Ti alloys for autoclave construction cannot be over emphasized. Without the oxidizing potential from these elements, Ti alloys could not resist the corrosive attack in the autoclaves. These oxidizing ions aid in the passivation of Ti alloys in reducing environments by driving the potential of the Ti into the passivation range.


5

Figure 2 - Isocorrosion diagram for titanium alloys in pure, aerated sulfuric acid.

Nickel-based alloys are not passivated by oxidizing ions, and in fact the presence of these ions accelerate the corrosion rate of nickel alloys much the same as happens for stainless steels. Table 3 shows the combined results of coupon testing of various alloys in simulated PAL autoclave test environments with a variety of nickel laterite ores [4]. The test durations ranged from 6 to 15 days. There are actually two phases in the high pressure autoclaves, the slurry phase and the vapor phase. It can be seen that alloys behave differently in the vapor phase compared to the slurry phase. As indicated by the data in Table 4, the chloride content of the slurry is quite important to the choice of alloys for an PAL plant, except of course for Ti alloys that show no dependence on the chlorides.

Table 3: Effect of Certain Multivalent Ions on the Corrosion of Titanium in Boiling 10% Sulfuric Acid [3].

Inhibiting Ion Concentration(ppm) Corrosion Rate (mm/y)

0 >76.2

100 0.208 Fe3+

500 0.069

0 >76.2

100 0.419 Cu2+

500 0.361

0 >76.2

100 0.001 Cr6+

500 0.001


6

Table 4: Simulated Autoclave Corrosion Tests at 250oC in Potable Water (Low Chloride) and Saline Water

(150 g/l NaCl) Slurries from Various Nickel Laterite Ores

Potable Water Slurry Saline Water Slurry

Material Corr Rate (mm/y) Crevice Attack/ U-Bend Cracking

Corr Rate (mm/y) Crevice Attack/ U-Bend Cracking

Vapor Phase Corrosion Tests Results

Inconel 686 1.5 No/No 6.6 No/No

Zeron 100 17.8 No/No --- ---

Ti Gr 2, 7, 11, 12, 17

0.01-0.03 No/No 0.01-0.02 No/No

Ti Grade 26* <0.001 --- --- ---

TiGrade 27** 0

Tantalum --- --- 0 No/No

Zr 702 --- --- 31.2 Destroyed

Slurry Phase Corrosion Test Results

Inconel 686 47.6 Yes/No 120.3 Destroyed

Zeron 100 67.8 Destroyed --- ---

Ti Gr 2, 7, 11, 12, 17 0.01-0.03 No/No 0.01-0.02 No/No

Ti Grade 26* 0.01 --- --- ---

TiGrade27** 0.01 --- --- ---

Zr 702 --- --- 0.01 No/No

* Tested at 260oC, ** Tested up to 275

oC

It is apparent that the nickel-based alloys are not suitable for service in the autoclave environment and that even Ta corrodes at an unacceptable rate. Titanium alloys are the only alloys that consistently show resistance to corrosion in the slurry and vapor phase of the high pressure acid leach. Additional testing on slurries from other potential ore bodies has consistently demonstrated that in addition to Ti Grades 7, 11, 12 and 17 shown in Table 4, Grades 1 and 2 also are resistant to corrosion in the autoclave environment with corrosion rates less than about 0.025 mm/y with no evidence of crevice corrosion or stress corrosion cracking (SCC). On the contrary, all of the stainless steels and nickel-based alloys show very high corrosion rates and are unsuitable for the PAL autoclave. A few tests performed at temperatures in excess of 250

oC have shown all of the stainless steels and

nickel-based alloys (ie, Alloys C276 and C-22 ) corrode at rates in excess of 100 mm/y and even Ta displays a considerable corrosion rate on the order of 2 mm/y.


7

Figure 3 – Crevice Corrosion Susceptibility as a Function of pH and Temperature

The primary concern with chlorides and Ti at high temperature is the potential for crevice corrosion. Figure 3 shows the standard curves for the onset of crevice corrosion in Ti alloys. It can be seen in Figure 3 that the commercially pure alloys (Grades 1 and 2) have low resistance to crevice corrosion while the alloyed grades such as Grade 12, 11 and 17 have better resistance. Since the autoclave temperatures exceed 250

oC and the pH is

about 1, Figure 3 would indicate that crevice corrosion is indeed a potential problem in PAL autoclaves. However, this has not proven to be the case in practice nor found in laboratory testing. Electrochemical polarization testing in a simulated autoclave environment using creviced Grades 1 and 17 electrodes were tested and showed no tendency for crevice corrosion for either alloy in a very high chloride, high temperature environment [5].

Similarly electrochemical polarization testing was carried out by J. Vaughan [6] to determine the pitting potentials for Ti Grades 7, 12 and 18 in 30 g/l H2SO4 at 225

oC. As shown in Figure 4 there was no significant hysteresis on

the reverse scan indicating that pitting corrosion was not a problem. Vaughan did not perform crevice corrosion tests nor did he include chlorides in the testing but other polarization work (ref 5) appeared very much the same confirming that localized corrosion (including crevice corrosion) of Ti alloys in these high temperature high pressure environments is not a problem.


8

Figure 4 – Cyclic polarization of Ti Grade 12 in 30 g/l H2SO4 at 473 and 498K.

-1.0

-0.5

0.0

0.5

1.0

1.5

1.E-08 1.E-07 1.E-06 1.E-05

Current (A/cm2)

E (

V)

vs N

HE

473 / 1930

498 / 2760

Temperature / Pressure

(K / kPa)

However, this is not to say that the Ti alloys are completely immune to corrosion or SCC regardless of how they are manufactured or handled. If these alloys are not properly fabricated they can become susceptible to corrosion and SCC. A number of the PAL nickel plants have had problems with SCC of the agitators in the autoclaves that are made from a combination of Ti Grades. Shafts made from Gr 7 have been fine but hubs and blades made from Gr 5 that were welded to the shafts without post weld heat treatment have failed from SCC. Repair welds on these agitators that were subsequently stress relieved did not crack. Other plants have been supplied agitators with Ti Grade 26 hubs.

One other important and problematic component associated with the PAL auotclaves are the acid injection tubes. Many of the plants have had serious corrosion problems with these injection or dip tubes. The dip tubes are used to inject concentrated sulfuric acid (96-99.5%) into the autoclaves at temperatures to 200

oC and sometimes

higher not by design. Alloy 59 (a nickel-based alloy) was initially used in the vapor space above the slurry but failed rapidly from corrosion when back flow occurred. The only alloy that has sufficient corrosion resistance under these conditions is tantalum but Ta is relatively soft and not available in tubular form of large enough diameter from which to manufacture dip tubes. Therefore, one solution has been to internally line Ti tubes with Ta or Ta-W alloys. Another approach has been to spot weld Pd buttons on the inside of Ti tubes.

3.2 PRESSURE OXIDATION LEACHING (POX)

While nickel ores are generally processed using PAL, copper and gold are more often processed using oxygen injection to achieve high dissolved oxygen levels in the leaching slurry, referred to as POX. Large POX autoclaves have been used in the gold industry for over 20 years and have typically been lead membrane and acid brick lined steel construction utilizing titanium internal components for agitators, baffles, valves, and nozzle liners. The titanium has performed well from the corrosion perspective but infrequent ignition events have occurred with the titanium internals [7]. The classical titanium ignition studies were performed by Littman and Church in the 1960’s under sponsorship of US nuclear and space projects [8]. The graph in Figure 5 shows the relationship that they observed between ignition potential, total pressure and oxygen partial pressure. The typical operating conditions for PAL and POX autoclaves are overlaid onto this graph [9].


9

Their work demonstrated that spontaneous ignition of titanium would not occur anywhere near autoclave operating conditions. It further indicated that sustained ignition could not be achieved due to fracturing of thin sections of titanium in the POX atmosphere range. However, it did show that sustained ignition could be achieved if sufficient preheating was available from external heat sources. Titanium internal components continue to be extensively used in these autoclaves in conjunction with implementation of well developed risk management and mitigation programs. For cost and performance reasons [10] there is strong interest within the industry to understand the risks associated with use of titanium clad and solid titanium autoclaves as alternatives to brick lining. Recently, Banker and McMaster

[Ref 9] reported on results from a study on ignition of solid Ti and Ti clad steel

and concluded that heavier sections of titanium are not easily ignited under POX autoclave operating conditions. Ignition could not be achieved with a 20,000v spark charge and mechanical impacts, even in a pure oxygen atmosphere. The heat input of burning 0.25g of magnesium on top of the titanium surface was required to initiate a sustained fire in a 100 bar pure oxygen atmosphere. When steam was introduced to more realistically simulate autoclave oxygen partial pressure conditions, ignition was not achieved. The titanium surface could not be initiated under simulated Gold autoclaves conditions with the heat input of a 2g magnesium fire on the surface. This work further confirms that non-standard events or upset conditions are necessary for titanium ignition to occur in POX autoclaves. Once ignition is achieved, the vertical orientation of components in an autoclave, and their location relative to gasket areas, will have significant influence upon the event outcome. Although not confirmed due to test equipment design conditions, the data indicate that an ignition event at the wall of a solid titanium autoclave is likely to result in a through-wall ignition failure and catastrophic decompression. It further indicates that a clad autoclave is less likely to experience a vessel wall failure during an ignition event. In a clad autoclave it is likely that a significant amount of the titanium cladding may be removed by burning and melting. If undetected and nor corrected, subsequent attack by the corrosive process liquor would likely follow, likely resulting in a failure later.

Aside from the potential ignition issue, Ti alloys exposed to POX autoclave environments show very good corrosion resistance, often less than 0.05 mm/y [11], and no indication of crevice corrosion when tests were

Figure 5: Littman and Church Graph showing the regions of titanium ignition overlayed with the

approximated pressure and oxygen partial pressures of the three main PAL and POX operating ranges.

Ignition was sustained in the regions shown by the arrows.

Promoted Ignition in O2 + Steam

Static O2

Flowing O2

Low temperature POX

autoclaves, Cu

Medium temp POX

autoclaves, Gold

High temp PAL

autoclaves, Ni


10

performed with crevice coupons. Stainless steels and nickel-based alloys showed corrosion rates well in excess of 20 mm/y under these same conditions.

4. COST CONSIDERATIONS

Titanium has a reputation as being exotic and expensive. Leveraging on this, our financial industry frequently ranks Titanium credit cards right up there their Gold and Platinum cards, even though titanium is less than 0.1% the price of gold. Titanium’s broad use in high end aircraft and space vehicles further fosters the high price image. Not to be out done, the sports industry promotes titanium as a premier metal for improving our golf game and cycling performance, attaching a high price to match the image, much more than justified by the differential metal cost. Then, try buying a titanium hip replacement. In reality, the cost of titanium process equipment is much closer to that of the other CRA’s than many designers imagine. This is partly due to titanium’s high strength to weight ratio. The density of titanium is roughly half that of the austenitic stainless steel and nickel alloy CRA’s, while the design strength is generally higher (as much as double). When the cost of design, fabrication, testing and transport are added to the cost of the metal, the cost multiplier is often surprisingly low for titanium process equipment. This is addressed in a number of excellent papers presented over the years [12]. Table 1 shows the relative cost of the titanium alloys commonly used in hydromet applications. The values are presented as a ratio of the alloy price to that of the lowest cost titanium alloy, Grade 2.

The cost difference relative to other CRA’s was further reduced by the recent reclassification of several titanium alloys by ASME. In 2007 the “H” variant of alloys was added to the ASME Code and allowed for pressure equipment in Code Case 2497 [13]. This family includes Grade 2H and its palladium alloyed variants 7H and 16H. The minimum tensile strength, and the design allowable strength, for these alloys is 16% higher than their non “H” sisters. The change was made after an industry-wide review of mechanical property data for all products manufactured and certified as Grade 2 family alloys [14]. The data indicated that well over 99% of all product being manufactured was compliant with the higher values being proposed for the “H” family. In other words, nothing changed but the name and the allowable design strength. The potential cost reduction for solid titanium equipment via use of the 2H alloy family is in the 10% to 20% range when compared to the Grade 2 family [15]. Savings can be even higher when standard plate or pipe thicknesses allow a lesser standard thickness to be used.

When thicker plates are required, further and very significant cost reduction is achievable through use of clad [16] and [Ref 10]. Cladding a thin layer of titanium onto a low cost base metal, such as steel, generally provides a lower cost product when pressures and/or temperatures mandate heavy wall equipment. It is generally accepted that solid titanium construction is lower cost than clad when the titanium thickness is under about 15mm and that clad construction is lower cost than solid titanium when thicknesses exceeds 25mm. In the 15 to 25mm range, the cost comparison depends upon a broad range of design factors. This is well addressed by Bower and Banker [17].

5. CLAD MANUFACTURE AND FABRICATION

5.1 CLAD MANUFACTURE

Three manufacturing technologies are commonly used for manufacture of large clad plates and clad vessel components: explosion cladding, hot roll bonding, and weld overlay. Due to the significant metallurgical dissimilarities between titanium and steel, weld overlay is not a viable technology for manufacture of titanium-steel clad. These same metallurgical issues highly limit the options for hot roll bonding. Consequently virtually all titanium clad used in process equipment construction is manufactured by explosion cladding.

Explosion cladding is a solid state metal-joining process that uses explosive force to create a metallurgical weld between two metal components. Although the explosive detonation generates considerable heat, there is no time for heat transfer to the component metals; therefore, there is no appreciable temperature increase in the metals. Due to the absence of heating, the microstructures, mechanical properties and corrosion properties of the


11

wrought parent components are not significantly altered during explosion bonding. There are no heat affected zones, and brittle intermetallic layers are not formed. For these reasons explosion cladding is ideally suited for bonding of virtually any combination of metal [18, 19].

Figure 6 presents a schematic of the explosion welding process. During the cladding operation the titanium plate is accelerated toward the steel plate by the explosive detonation energy. The plates collide under high velocity oblique conditions generating a jet of metal being stripped from the colliding metal faces. Surface contaminants which normally prevent welding are removed in this jet. The result is a high strength, ductile metallurgical bond over the full plate surface area.

Figure 6: Schematic of the explosion cladding process.

It is standard practice to stress relieve titanium/steel clad in the 538°-635°C (1000°-1175° F) range to optimize bond toughness prior to proceeding with fabrication. When proper fabrication procedures are employed, titanium clad will not disbond during forming and fabrication processes [20]. The thermal induced bond stresses in autoclaves, resulting from the hot operating conditions and differing thermal expansion coefficients, are below the bond strength stresses and will not induce bond failure in service.

Titanium clad plates with widths of 4.5m (176in) and lengths of 8m (315in) are commonly produced. The titanium cladding thickness typically ranges between 2mm (0.08in) and 19mm (0.75in), dependent upon the application. The steel base metal typically ranges between 12mm (0.5in) and 500mm (20in), dependent upon design requirements.

Titanium clad is typically produced to ASTM Specification B898. This specification mandates ultrasonic testing to assure high bond integrity. Bond strength is determined by a shear strength test. B898 requires a minimum shear strength of 137mpa (20,000 psi). Bond shear strengths for titanium-steel are typically in the 200 MPa (30,000 psi) to 340MPa (50,000 psi) range with average values of 270 MPa (39,300).

6. TITANIUM AND STEEL ALLOY CONSIDERATIONS FOR CLAD

Table 1 lists the titanium alloys which are commonly used in the process industries. All of these alloys can be clad using the explosion welding process. The optimum clad mechanical properties and optimum plate sizes are produced when the yield strength the cladding is below 300 MPa (45 ksi) and that of the base metal is below approximately 350 MPa (50 ksi). Consequently the optimum strength and toughness of titanium cladding results from a combination of Titanium Grade 1 (or the Gr 17 and 11 variants) clad to a moderate strength pressure vessel steel, such as ASME SA-516 Grade 70. The higher strength titanium alloys, Grades 2, 7, 12 and 16 and


12

their “H” variants, can also be clad to steel. However, due to the higher strength and lower ductility, the maximum sizes that can be manufactured reliably are smaller. Consequently, costs for clad of these alloys is higher than for clad of the Grade 1 family, even though the opposite is true for un-clad titanium plates.

7. TITANIUM AND TITANIUM CLAD EQUIPMENT DESIGN AND FABRICATION

Due to certain basic material properties, mechanical design of titanium equipment is typically different than design of the same equipment in steel or stainless steel. The elastic modulus of titanium is roughly half that of steel and stainless steel. The coefficient of thermal expansion of titanium is approximately 70% that of steel and even lower compared to stainless steel. These factors affect structural reinforcement and expansion joint considerations considerably. Substituting titanium into a piece of equipment that was previously designed in steel or CRA, without consideration of these factors can result in reduced performance or possible failure.

6.1 TITANIUM WELDING

Titanium welding is not difficult, but the concerns are very different than with steel welding [21,22]. At high

temperatures, titanium will react with virtually everything except the inert gases. Above 425°C (800°F), titanium will readily absorb oxygen; the absorbed oxygen can make the titanium brittle and mechanically useless. Even very small amounts of oxygen (0.8%) will significantly embrittle titanium. Cleanliness, freedom from moisture, and protection with an inert gas cover are absolutely critical for successful titanium welding. Since titanium will react with oxygen well below the melting point, it is important to maintain inert gas shielding of the weld area until it has cooled below at

least 300°C (575°F). For clad batten strip welds, backside purging with inert gas is mandatory to prevent oxygen contamination of the weld from the root side. Note the backside purge hole in Figure 4.

Titanium welds that have been contaminated by oxygen exhibit two readily discernable features. They exhibit obvious discoloration and increased hardness. Good titanium welds are shiny silver to slightly straw color. Contaminated welds are blue to grey to white. Welders must learn to read the color of their welds, and to immediately reject areas where weld color indicates a defective weld. Areas of contaminated weld cannot be repaired simply by welding over the defective area; the contaminated area must be fully removed before a successful replacement weld can be made. With proper training, equipment, attention to detail, and a strong quality culture, reliable high-quality titanium welds can be assured.

7.2 CLAD FORMING AND FABRICATION

Titanium clad equipment can be reliably constructed and has proven service reliability. However, due to differences in metallurgical characteristics, thermal expansion, modulus, and other aspects, special considerations must be taken in design, fabrication, welding, and testing to insure a reliable product [Ref 20].

7.2.1 Clad Weld Joint Design

Titanium and steel cannot be directly fusion welded to each other due to brittle intermetallic formation. Clad fabrication is typically accomplished using a batten strip technique as depicted in Figure 4. The cladding is striped back from the edges to be welded and steel welds are made. The vessel is then cleaned and prepared for titanium welding. A strip of titanium, the batten strip, is then placed over the steel weld area. The batten strap is welded along the edges.


13

Figure 7: Schematic of batten strip welding concepts for Titanium-Steel clad welds\

When the titanium cladding metal layer is thin, it is common to place the batten strip on top of the cladding surface, as depicted in Figure 7(a). When this technique is used, a filler-metal strip is inserted into the space where the titanium has been removed. The choice of filler is dependent upon proprietary fabrication preferences; commonly used materials include copper, steel, and titanium. Surface mounted batten strips have been extensively used in titanium clad equipment for chemical process applications over the past 40 years [Ref 21]. This is the simplest and easiest way to install batten strips; however, when the cladding metal is thick or when there are erosion concerns, the surface mount batten strip can lead to problems.

When erosion is expected, recessed batten strip designs are preferable. Figure 7(b) presents the partially recessed concept. This design offers benefits with regards to stresses, welding design, and inspectability. Figure 7(c) shows fully recessed batten strip design. This option potentially offers even better stress distribution and freedom from erosion. However, welding process control, particularly on the titanium root pass, can be challenging.

Although the batten strip depictions in Figure 7 are quite simple, the precise design of welds and batten straps for autoclaves is not. One critical concern results from the high difference in thermal expansions of titanium and steel combined with the high operating temperatures of the equipment. The coefficient of thermal expansion (CTE) of titanium is 30% lower than that of carbon steel. Since the thickness of titanium is typically only 5% to 10% of the steel thickness, the titanium cladding layer is essentially forced to move at the steel CTE rate. This strain results in a tensile stress in the titanium cladding layer. Over the main body of the autoclave, the continuous clad bond transfers the thermal expansion stresses between the two metals uniformly. In the weld region, all of the differential stress must be borne by the batten strip and the welds along each edge. Incorrect design can result in stresses in the welds that could potentially result in weld failure in service. It is critical that detailed stress analyses and experimental verification be performed to assure proper sizing and configuration of the batten strips and the welds joining them to the clad plate surface. Specific details are typically highly protected proprietary designs of the equipment fabricators. For successful vessel performance the importance of this aspect of vessel design cannot be over emphasized. Experience has shown that properly designed weld systems are problem free in service [Ref 20].

7.2.2 Forming of Heads and Cans

During vessel fabrication the clad plates are formed into cylinders and heads. Cylinder rolling can be performed

near ambient temperature or in the temperature range of 540°C to 650°C (1000°F to 1200° F), dependent upon

forming equipment capability. Hot pressing is the preferred head forming technique. Forming in the 540°C to

650°C (1000°F to 1200° F) range is preferred. At these temperatures intermetallic formation is sluggish; the operator can concentrate on producing a good head, not on getting it done quickly. If needed, higher

temperatures can be used, but forming temperatures must be maintained below 815°C (1500°F) and hold times at this temperature must be minimized to about 2 hours to avoid excess intermetallic formation and bond failure [23,24].


14

7.2.3 Clad Equipment Fabrication Precautions

The fabrication of titanium clad vessels requires additional precautions. As with most steel fabrication, the steel portion of the clad fabrication process is typically dirty, producing large amounts of metal dust, swarf and sparks. The titanium portion of the fabrication process must be clean. During steel fabrication titanium surfaces should be protected from grinding, welding and cutting sparks to avoid localized iron contamination of the clad surface. After this aspect of fabrication, and before the steel post weld stress relief heat treatment (if required), all iron surface contamination should be removed by grinding or similar processing.

After steel fabrication and heat treatment, and prior to titanium welding, it is best to transfer the vessel to a clean shop. Alternatively, since vessels are normally clad on the inside, the interior of a clean clad steel vessel can provide suitable conditions for titanium welding, once the openings are sealed off. However, safety precautions must be taken to prevent accumulation of deadly levels of inert gas in the work area inside the enclosed vessel.

7.2.4 Clad Equipment Testing and Inspection

Clad pressure vessels require testing of the pressure containing components and welds in accordance with the applicable design Code requirements. Upon completion of fabrication and testing, cladding metal surfaces must be cleaned to eliminate any residual iron contamination which may cause accelerated localized corrosion. Ferrolxy testing, ASTM B380, can be used to confirm the absence of iron contamination. Batten strips and related welds in titanium clad equipment are typically subjected to Hot Gas pressure cycling to simulate operational stresses on the welds [Ref 20]. Batten strip welds are examined for surface defects using penetrant testing. They are further examined for pin hole defects using helium leak testing in which helium is fed into the area behind the batten strip via the weld-purge holes, and defects are detected using helium sniffers on the inside vessel surface.

8. CONCLUSIONS

Titanium is the metal of choice for construction of PAL and POX autoclaves, vessels, and piping. In those environments, titanium exhibits unique corrosion resistance which is unmatched by any other common engineering metal. Titanium has provided proven corrosion performance in these environments for well over 30 years. Although considered to be an expensive metal, titanium has proven to be cost effective when compared against brick lining and less effective metal options. In POX autoclaves, potential ignition conditions between titanium and oxygen present unique and manageable risk management concerns. In heavy equipment, such as autoclaves, titanium clad provides considerable cost and performance benefits. Explosion welded titanium-steel clad is a proven, robust product which is reliably manufactured, fabricated and utilized in PAL autoclaves.

9. REFERENCES

1. R.W. Schutz., L.C. Covington, “Hydrometallurgical Applications of Titanium” Industrial Applications of Titanium and Zirconium, ASTM STP830, R.T.Webster & C.S.Young. American Society for Testing and Materials, 1984, pp29-47.

2. 2. R. Francis, G. Byrne and G. Warburton, "The Use of Super Duplex Stainless Steels in the Nickel Mining Industry", Nickel-Cobalt '97, Sudbury, Ontario, Canada, 1997.

3. R. Schutz and D.E. Thomas, Corrosion of Titanium and Titanium Alloys", p 682, ASM Metals Handbook, Vol 13, Corrosion, 1987.

4. B. D. Craig, “High Pressure Acid Leaching- A Materials Challenge”, Keynote Lecture, Corrosion Applications Conf, Wah Chang, Coeur d’Alene, Idaho, Sept, 2003.

5. MetCorr, unpublished research, 1997. 6. J. Vaughn, “ Corrosion of Titanium and its Alloys at High Temperatures and High Pressures”, Master’s Thesis,

McGill U., 2001. 7. Oxygen Fire Study, Hermann Pieterse, Pieterse Consulting Inc., Proprietary Publication, April 28, 2004.


15

8. Littman, F.E, Church, F.M, and Kinderman, E.M., “A study of metal ignitions, the Spontaneous Ignition of Titanium”, The Journal of Less Common Metals, Vol 3, 1961, pp 367-378.

9. J.G. Banker and J.A. McMaster, “Titanium and Titanium Clad Ignition Studies in Simulated Pressure Oxidation Leaching Environments”, Wah Chang Conf, 2007.

10. L. Zunti, M. Pearson , “A comparison of refractory lined and metal-clad process vessels for specific operating conditions”, Pressure Hydrometallurgy 2004, Collins and Papangelakis ed, CIM, pp 639-649.

11. MetCorr, unpublished research, 2008. 12. J.S.Grauman and B. Willey, “Shedding New Light on Titanium in CPI Construction”, Chemical Engineering,

August 1996, pp. 106-111. 13. J. McMaster, “ASME Code Case 2497-2 - Unalloyed Grade 2, 7, 16, and 26 Titanium with 58 ksi Minimum

UTS”, Stainless Steel World, November 2005 Titanium Special. 14. J. McMaster, B. Greene, S. Kirsch, “Review of Mill Test Reports Leads to Code Case 2497-2 and ASTM H

Grades” Wah Chang 2007 Corrosion Solutions Conference, Sunriver, Oregon, September 13, 2007. 15. J. McMaster, “Rationalization of Unalloyed Titanium Material Specifications to Current Production Capabilities

Offers Opportunities for the Titanium Industry”, NACE Corrosion 2003, March 2003. 16. Banker, J. G., “Titanium Clad – A Proven Material for Sulfuric Acid POX Processes”, Sulfur Magazine,

October 2007.

17. L.Bower and J.G.Banker, “Large Titanium Heat Exchangers, Design, Manufacture and Fabrication Issues”, Corrosion Solutions Conference 2003, Wah Chang, September 2003

18. A. Pocalyko, “Explosively Clad Metals,” Encyclopedia of Chemical Technology, Vol. 15, Third Edition, John Wiley & Sons, 1981, pp 275-296.

19. J.G. Banker, E.G. Reineke, “Explosion Welding”, ASM Handbook, Vol. 6, Welding, Brazing, and Soldering, 1993, pp 303-305.

20. J. Laermans and J. Banker, “Large Titanium Clad Pressure Vessels, Design, Manufacture and Fabrication Issues”, Corrosion Solutions Conference 2003, Wah Chang, September 2003.

21. J. McMaster, “Practical Titanium Welding”, Practical Welding Today, pp. 22-28, Vol. 4, No. 4, July-August 2000.

22. J. McMaster, “Inert Gas Shielding and Purging for Titanium Welding ”, Canadian Welding Association Journal”, Fall 2008 Issue.

23. C. Prothe, J. Banker, “Metallurgical Considerations in the Manufacture of Clad Heads”, Corrosion Solutions 2007 Proceedings, ATI Wah Chang, Oregon, pp 83-90.

24. C. Prothe, S. Pauly, C. Toth, “Effects of Heat Treatment time an dTEmperature on the Properties of Titanium/Steel and Zirconium/Steel Clad”, Corrosion Solutions 2005 Proceedings, ATI Wah Chang, Oregon, pp 59-6.


FIELD PERFORMANCE REVIEW OF AUTOCLAVE VALVES

By

John Williams

MOGAS Industries, Inc., USA

Presented by

John Williams

[email protected]

CONTENTS

1. INTRODUCTION 2

2. OBJECTIVES 3

3. HISTORICAL VALVE PERFORMANCE 3

4. INTERACTION OF VALVE DESIGN, MATERIALS AND OPERATION 3

5. RESULTS 4

6. CONCLUSIONS 6

7. RECOMMENDATIONS 7

8. REFERENCES 7


2

INTRODUCTION

Figure 1 – MOGAS valves installed at RNO Nickel Operation

MOGAS Industries is in the business of developing and manufacturing ball valves for severe service. A necessary component of success is the accurate collection of information regarding the field performance of valves, understanding the limiting factors associated with field performance and adjusting the design to provide better performance. While many properties may be individually measured using standardized laboratory tests, there is no substitute for the information gained from actual field installations. Field performance identifies the interactions of material, designs and operations. Of the many MOGAS Autoclave customers, the Ravensthorpe Nickel Operation (RNO) plant is the most documented and examined mining operation. The data and results gained at RNO were planned and organized during the pre-bid phase of the project. Systems were predetermined and set up to methodically gather performance information and analyze against key performance indicators. Project operations personnel and service groups were dedicated to accurate performance measurements. Information regarding field performance of POx service valves at Lihir Gold, Oceana Gold and Sepon Copper are qualitatively compared to the measurements and data collected at RNO.


3

Figure 2 - Map of selected autoclave sites

1. OBJECTIVES

1. Review the collected data for RNO from October 2007 to January 21, 2009 2. Compare the HPAL valve performance to historical performance. 3. Compare current HPAL performance to anticipated performance.

2. HISTORIC VALVE PERFORMANCE

In 1999, the goal expressed to MOGAS by Cawse Nickel Operations for a “good run” with an autoclave discharge valve was to achieve either 60 days or 60 cycles of service.

An independent study conducted in 2001 of HPAL plants indicated that autoclave valve costs were six times greater than expected and represented 30 percent to 40 percent of the operating budget.

3. INTERACTION OF VALVE DESIGN, MATERIALS AND OPERATION

Valves for HPAL and POx service must be designed with the severe conditions created by abrasive solids and the highly corrosive elevated temperature environment in mind. An excerpt from 2003 ALTA paper “Ball Valves with Nanostructured Titanium Oxide Coatings for High Pressure Acid Leach Service: Development to Application” details the relative importance and interaction of various attributes associated with valves in Table 1 below.


4

Table 1 – Issues Relating to HPAL Ball Valves

Factor Result Control Method Perceived

Importance

Verification

I. Solids

Effect on mechanics:

Pack-in increases loads and

stresses

Geometry

****

Field

Inability to operate and

Over stress of coating

Robustness of design **** Model testing

Effect on wear:

Sliding 2-body

Loss of seal

Geometry / materials

properties

***

In-house fixture testing

Abrasive 3-body Loss of seal Geometry / materials

properties

*** ASTM G65-91

Slurry erosion Loss of seal Geometry / materials

properties / operation

** Lab testing

Gross damage Geometry / materials

properties / operation

**** Field

II. Environmental

Corrosion resistance:

Gross corrosion causes loss of

seal and mechanical integrity

Corrosion resistant

materials

****

Lab testing / field

Corrosion at coating bond line

leads to coating & seal losses

Corrosion resistant

materials/reduce or fill

coating pores/ add dense

intermediate layers

**** Lab testing / field

Temperature: Differences in CTE results in

different thermal expansions of

internal components, increasing

loads and stopping valve

operation

Geometry **** Field

Differences in CTE between

coating and substrate produces

bond line stresses that may result

in delamination and loss of seal

Select compatible

materials/maximize bond

strength

*** Field/ ASTM C 633-79

Rapid changes may cause

coating cracking / delamination

Maximize bond

strength/select high strain-

fracture resistant coating

* Field

Scaling: Silicates & particles adhere to

components and promote wear

and/or bridging

Geometry / operating

practices, i.e., purging

*** Field

III. Operation

Speed, Frequency,

Differential Pressure:

Effects velocities, exposure of

critical seal surfaces, and heating

and cooling rates

Operating practices /

accessories / purges

***

Field

Note… The greater number of * depicts the relative importance.


5

4. RESULTS

4.1 RNO FIELD DATA

Table 2 – Performance Life of Key Severe HPAL Valve Applications

Application Mean Cycles Mean Days

Slurry Inlet Inboard 313 327

Slurry Inlet Outboard 331 327

Slurry Outlet Inboard 148 281

Slurry Outlet Outboard 148 281

Vent 1061 -

General Service (4 inch valve) 603 -

General Service (2 inch valve) 1216 -

5.1.1 The most concrete data available from the RNO operations is when a valve is determined to not be operating properly either because of leakage through bore or external or the inability to operate or cycle. This information formed the primary performance measurement. 5.1.2 RNO utilized a hyper saline water wash to maintain plant cleanliness. 5.1.3 The applications that are traditionally considered the most critical are the slurry inlet, slurry discharge and vent isolation valves. See schematic below.

Figure 3 - Schematic of typical autoclave

5.1.4 Slurry Inlet 10 inch ASME 600 class PL Ni valves Both inboard and outboard valves were cycled a total number of 3,513 cycles. The average cycle life before valve repair was 313 cycles for the outboard valves and 331 cycles for the inboard valves. It is noted that two inboard valves were never replaced so their cycle life has not been defined. The mean installed time between repairs for the slurry inlet valves was 327 days.


6

5.1.5 Slurry Discharge 10 inch ASME 600 PL Ni valves Both inboard and outboard valves were cycled a total of 592 times. The average life before repair was 148 cycles. All outlet valves were repaired at least one time. The mean installed time between repairs for the slurry discharge valves was 281 days.

5.1.6 Vent valves 6 inch ASME 600 PL Ni valves Vent valves accumulated a total of 2,511 cycles. The longest life was 1,906 cycles, the shortest was 216 cycles. One vent valve was never replaced so its life was never defined. The average cycle life for vent valves was 1,061 cycles. The mean installed time between repairs for the slurry discharge valves was 490 days.

5.1.7 General Service valves General Service valves control steam injection, water flush, as well as line bleeds and drains. The average cycle life for all 4 inch ASME 600 class valves requiring repair was 603 cycles. The average cycle life for all 2 inch ASME 600 class valves was 1,216 cycles.

5.1.8 Observed field performance of ball and seat coatings The nano-structured titanium coating employed on all valves for the RNO project was the result of much development work. Laboratory tests indicated it possessed superior wear and erosion resistance to other ceramic coatings in HPAL usage. In general, the nano-structured titanium performed exceptionally well as evidenced by the high cycle life achieved by all valves. Valves that operated in vent, steam or general service experienced remarkably high cycle life before removal. It was common to only touch-up lap these valves rather than to recoat and resurface the ball and seats. Valves that experienced the shortest coating life were those that were exposed to pressure reversals in service and actually sealed in two directions. Examples are the slurry inlet and discharge valves. It was also observed that the exposure of valves to pressure reversals could be greatly curtailed by operational practices. 5.1.9 Packing-in Valves at RNO utilized specific geometry and clearances to resist solids packing into the valve and restricting operation. No valves failed to operate or develop any apparent torque increases due to solids packing-in.

5.1.10 Operating torque Instrumentation to monitor actuation pressures and consequently operating torques in service was not functional during this period of plant operation and thus service increases in torques (if developed) were not measured.

5.1.11 The hyper saline water wash resulted in corrosion development on the mounting bracket, stem bushing and bolting that had not been previously observed in other field installations.

5.1.12 Reasons for removal of valves RNO contained 200 installed valves. Valves were removed from service if they could no longer operate (open and close), leaked externally or had leaked through. The causes and frequency of valve removal are as follows:


7

• Stem leak, fourteen valves serviced: Valves on the RNO project were provided without live loading or a specification requirement. Repairs were affected with live loading.

• Valve Passing or through leakage, nine valves serviced

• Stem bushing corrosion; more than eight valves were serviced: An inferior coat (out-of-specification) protective coating on stem guide bushings and stem adaptors resulted in one 6 inch valve not being able to operate. All valves, as a matter of proactive measures, were equipped with an improved design during routine repairs. MOGAS proactively implemented this change to remedy the stem bushing corrosion.

5.1.13 The slurry inlet and outlet valves both experienced a weakness in a previously qualified seat retention design. Investigation indicated both geometry modifications and operating practice changes would eliminate the ill effects of this retention weakness.

POx GOLD AND COPPER SITES Valves utilized in POx gold and copper extraction are constructed with similar design principals and materials. These sites differ in that the valves in place were ordered at different times and constructed to variations in design specifications. Additionally, the data collection regarding installed time, cycles and reasons for removal was not typically well documented. This has led to subjective determinations of performance rather than the objective measurements achieved at RNO. The following are general statements reported by Optimum Control regarding the servicing and maintenance of valves at Lihir Gold, Oceana Gold and Sepon Copper. “Valves are lasting a consistently dependable length of time allowing reasonable production campaigns to be set. These campaigns are driven by other maintenance issues and not valves.” “The typical POx plant has valves with much lower cycling frequency than at start-up as existed at RNO. The durability of valves may be reduced by precipitation of solids inside the piping and valves. Thus the frequent cycle rate of valves at RNO may have prevented unwanted precipitation of solids that can tend to happen if valves remain dormant.” “While difficult to measure, it does appear repair costs are generally decreasing from earlier years.” “The driving factor associated with HPAL / POx plants is having confidence in valves to bridge set maintenance intervals. Vent and discharge valves are generally removed for preventative maintenance at each major shut down.”

5. CONCLUSIONS

1. The dependability of slurry discharge valves has increased significantly over the acceptable 60-

cycle or 60-day standard at Cawse 1999. 2. The dependability of slurry inlet valves and vent valves are two and seven times greater than

slurry discharge valves. 3. The most severe applications in HPAL and POx are first the slurry discharge valves followed by

the slurry inlet valves.

Titanium vent valve design and materials used in HPAL have improved significantly so that their life is equally as long as the general purpose steam and drain valves.

Vent valves used in POx applications remain a high maintenance item primarily due to corrosion of the primary material of construction duplex stainless steel.


8

4. Organized, objective data collection of valve performance is a useful tool to validate designs as well as identify areas to consider for improvement.

5. The identification and corrective actions encountered in the seat retention design emphasize the

value and benefit of field monitoring and detailed inspection of removed valves.

6. RECOMMENDATIONS

1. Plant operators should consider the sequence and conditions applied to slurry inlet and outlet

valves with valve engineers to optimize valve life. 2. New designs and materials should be considered for POx vent applications. A material with

successful application to consider is TiNb. 3. Operations consider establishing organized, systematic tracking of valve performance in

agreement with both service and manufacturers as a means of improving plant efficiency.

7. REFERENCES

1. Data Collected Through MOGAS, “Valve Management Program” for RNO Nickel from July 1, 2007 through January 21, 2009.

2. Questionnaire provided by Brian Wood of Optimum Control April 19, 2009. 3. “Ball valves with Nanostructured Titanium Oxide Coatings for High Pressure Acid Leach Service:

Development to Application” 2003.


BENEFITS OF TIMETAL ® PGMA ™ IN NICKEL LATERITE REFINING

By

James Grauman*, Eliana Fu* & Ian Flower+

*TIMET HTL, USA +Minara Resources Limited, Australia

Presented by

Eliana Fu

[email protected]

CONTENTS

ABSTRACT 2

1. CORROSION RESISTANCE 2

2. APPLICATION OF TIMETAL ® PGMA™ 4

3. EXPERIMENTAL TRIALS OF PGMA ON AUTOCLAVES 5

4. AUTOCLAVE CORROSION HISTORY 7

5. RESULTS OF FIELD TRIALS 9

6. CONCLUSION 10

7. REFERENCES 11

8. ACKNOWLEDGMENTS 11


2

ABSTRACT

The Titanium Metals Company (“TIMET”) has developed a unique solution to corrosion in the

chemical process industry: PGMA™ (platinum group metal appliqué), wherein a patch of titanium alloy containing a platinum group metal can be applied to a base metal of commercially pure titanium such as ASTM grade 1. This provides for enhanced corrosion resistance at a substantially lower cost as compared to an entire component constructed from the traditional titanium-palladium grades 7 and 11. Field trials of the technique have been conducted by Minara Resources Ltd, specifically for autoclave lining application in nickel laterite refining at the Murrin-Murrin facility. Results of the testing to date have shown promising indications that PGMA does provide protection such that the level of pitting corrosion has been reduced dramatically in titanium ASTM grade 1 liner components within an 18 month trial period. Autoclaves, which Minara Resources use in the laterite nickel process, with titanium ASTM grade 1 liners, are very susceptible to pitting and crevice corrosion. In Minara’s experience, after only a short time in service, the titanium ASTM grade 1 liner surface suffered from corrosion in the form of a large amount of shallow pits, typically less than 1 mm deep. This trend has continued over the past 5 years, until in 2004, when a reduction in the number of individual pits was observed. However the severity of the corrosion increased, with the individual pits becoming much deeper, typically 2-3 mm. This climaxed in 2007 when the titanium liner was breached and a void 72 mm long by 45 mm wide by 15 mm deep, was observed in the carbon steel shell of the vessel. Minara Resources commenced installation of PGMA in compartment one of autoclave 4 in 2006 and continued installation of PGMA in the remaining three autoclaves during 2007 and 2008. During the same period the level of free acid added to the process has been reduced from 70 to 60 grams per liter. As a result of these actions, the number of corrosion pits and the severity of the corrosion attack have reduced significantly, so much so that the interval between internal inspection has been increased from 3000 hr to 4500 hr and finally to 6000 hr, increasing available autoclave hours by 720 hrs per year. Moreover, the greatest economic gain has been in terms of total life of the autoclaves at Murrin-Murrin, which consist of a 102 mm thick carbon steel shell with a 6 mm thick titanium ASTM grade 1 explosion bonded liner. Due to the nature of construction and code requirements, a major defect in the shell will have a devastating impact on the autoclave and under extreme conditions, may result in the equipment being damaged beyond repair. The installation of PGMA has significantly reduced pitting corrosion and also reduced the risk of a titanium liner and catastrophic autoclave failure.

1. CORROSION RESISTANCE

1.1. THE CORROSION RESISTANCE OF TITANIUM

Titanium has become an increasingly important structual material used in a wide range of industries from aerospace to automotive, architecture and consumer products. Equally important, its use is expanding in the chemical process industry. Its unique properties include a high strength-to-weight ratio and an unparalleled corrosion resistance, due to the tenacious inherent oxide film. When this oxide film is stable, corrosion of the titanium metal is extremely low and in many environments essentially zero. However, if the film is destabilized, such as in hot reducing acid conditions, then the titanium, which is very reactive, can corrode at an exceedingly rapid rate. Fortunately, titanium oxide is extremely stable to oxidizing environments, even in the presence of high acid concentrations. This has been the basis for its successful use in HPAL processing of metal ores. Titanium has been used very successfully in hydrometallurgical applications for more than 30 years. A detailed overview of the corrosion resistance of titanium can be found elsewhere. [1]

The disadvantage with the titanium grades which contain platinum group metal elements are their inherently high cost. These metals are bought and sold on the open metal exchanges and futures market and consequently fluctuate in price, with the tendancy to be on the high side. As a result, material is difficult to obtain and distributors are often unwilling to stock a material with such a high inventory price.


3

TIMET’s Henderson Technical Laboratory (HTL) has developed a new process, TIMETAL ® PGMA (Platinum Metal Group Appliqué) which can provide the same corrosion resistance as the more traditional palladium alloyed grades, with a reduced cost, since the platinum group metal is only contained on the surface of the metal in very small attached patches. The process, being remarkably simple, consists of a “patch” or appliqué, a small piece of a special alloy in coupon form, which contains the alloying element of a platinum group metal attached to the base metal component (grade 1 or 2 titanium). The appliqué provides an ennobling effect to the surrounding substrate metal and thus provides the enhanced corrosion resistance so long as the patch contact is maintained. The additional benefits with the PGMA system are the use of a much more readily available grade of titanium and the flexibility to apply the appliques anywhere along the manufacturing chain or as a retrofit.

1.2. CORROSION RESISTANCE OF TITANIUM IN HPAL

In the strong acid environment of HPAL systems, titanium could possibly exhibit several different forms of corrosion. General corrosion attack would be possible if the source of oxidizing metal ions, which come from the ore itself and provides for passivation in reducing acid, was lost or masked, since hot sulfuric acid is known to attack titanium. In addition, localized corrosion could be possible in tight gasket to metal or underdeposit crevices. The localized corrosion could take the form of shallow surface etching or deep discreet pitting. Much of this corrosion morphology would depend on the composition of the solution trapped beneath the creviced area. Stress corrosion cracking (SCC) has been observed in a few instances where high strength titanium alloys were being utilized, such as agitator impellers. Commercially pure titanium has not demonstrated any latent susceptibility to SCC under typical HPAL environments. A more extensive review of the corrosion behavior of titanium in HPAL systems is available.[2]

Grades 1, 2, 12, and 11/17 comprise the bulk of the titanium used for components in the nickel laterite processing facilities. Generally, the corrosion performance of the titanium has been excellent especially considering the aggressive nature of the high temperature sulfuric acid environment. In particular, though, there appears to be some difference in corrosion performance between the autoclave lining materials used by the various facilities, which has been grade 1 and grade 11/17. It is known that the presence of the oxidizing ions in the ore can act in a similar manner to the effect of palladium additions to the titanium such that in strongly reducing acids, commercially pure titanium can often times display a 100 to 1000 fold reduction in corrosion rates, giving it the same or very similar corrosion behavior as its palladium enhanced relatives. Obviously, the oxidizing metal ion and its concentration have an important part in determining how closely the commercially pure titanium can mimic the behavior of the palladium enhanced grades. However, we now know that the palladium enhanced (grade 11/17) materials have withstood the rigors of this service better than the commercially pure grade 1 titanium. As described later on in this paper, the grade 1 material appeared to start exhibiting significant corrosion attack after a period of 3-5 years service.

This noticeable difference in corrosion resistance, then, is most likely due to the starvation of the beneficial oxidizing metal ions from the oxidation of the ore slurry. Typically, this scenario occurs most readily beneath tenacious deposits where oxygen deprevation can occur, even in the highly oxygen rich environment present in these units. The metal ions can only act to inhibit corrosion attack while they exist in their highest oxidation state, which is maintained by the air pressure within the autoclave. Once the metal ions (nickel, cobalt, and iron) have been reduced from their highest oxidation state due to this lack of oxygen within a deposit, any and all corrosion protection afforded the titanium by these ions is gone, leaving the titanium to fight off the acid conditions on its own. Commercially pure titanium grade 1 cannot handle this acid concentration and will corrode. At this point in time, the presence of the platinum group metal can make all the difference in maintaining low corrosion rates on the titanium.

1.3. CORROSION RESISTANCE OF TIMETAL PGMA IN HPAL

The superior performance of the palladium enhanced titanium grades in combating the corrosion scenario described above led to the decision to test the PGMA system in the autoclave environment at Minera. PGMA studies indicated that in a pure reducing acid environment, the corrosion protection system, when properly installed, performed on an equal basis with grade 7 titanium. [3] Prior excellent experience with the palladium enhanced grades at various HPAL facilities gave additional credence to this foregoing corrosion theory and supported selection of the PGMA for the trial.


4

The PGMA corrosion protection acts equally well to protect the substrate titanium from localized attack as well as uniform or general corrosion. Thus, whether the corrosion attack being experienced on the autoclave lining was due to general corrosion or (more likely) some form or localized attack (crevice or pitting), it was possible that the application of PGMA could help remedy the corrosion. Also, the advantage of using the PGMA as a retrofit option here lent itself readily to installation within strategic locations of highest corrosion of the autoclave.

This paper provides a resume of TIMETAL ® PGMA field trials in the nickel laterite refining process performed by Minara Resources at Murrin-Murrin.

2. APPLICATION OF TIMETAL ® PGMA™

2.1. GENERAL CORROSION OF TITANIUM

In those environments where the metallic corrosion rate is ≥ 0.125 mm/yr, e.g. hot HCl, titanium is not suitable as the corrosion rate can be excessively high at concentrations > 0.5 wt%. Using titanium ASTM grade 7 (CP grade 2 + 0.15% Pd) however, means that structural components in titanium can

be used in concentrations up to 5 wt% or more, resulting in corrosion rates ≤ 0.125 mm/yr.

Figure 1: Corrosion rate comparison for PGMA (Pt) in HNO3 [3]

2.2. ENNOBLING EFFECT AND THROWING POWER OF TIMETAL ® PGMA

TIMETAL ® PGMA’s benefit in providing an ennobling effect on the titanium base metal is based on cathodic depolarization of the titanium surface oxide film. In the same way that using a platinum group metal as an alloying element in bulk titanium alloy lowers the corrosion rate by effectively polarizing the surface of the titanium. The oxide now having no propensity to react to ions present in surrounding media, maintains its integrity such that the titanium base metal is effectively protected as being “corrosion-resistant”. In particular, ASTM Grade 7 (CP Titanium Grade 2 plus 0.15% Pd) is the most widely used corrosion resistant titanium grade in the chemical process industry. In previous laboratory testing, testing of platinum PGM patches on grade 2 titanium, regardless of the method of attachment, corrosion resistance was equal to the Grade 7 and superior to PGM containing ASTM grades 16 and 26. [3]

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

Grade 2 Grade 7 Grade 12 Grade 2 + PGMA

Titanium Grade

Co

rro

sio

n R

ate

(m

m/y

r)

40% Nitric @

Boiling


5

Figure 2: Titanium Column protected by PGMA patches

The TIMETAL ® PGMA patches are able to protect a surface area of base metal many times greater than their own surface area. As mentioned above, previous laboratory testing was performed at TIMET and has demonstrated a protective “throwing power” of at least 1:250 in terms of patch surface area to base metal surface area in 5% hydrochloric acid at boiling point. [3]

Figure 2 shows a typical PGMA patch arrangement for protecting the interior of a pipe structure.

2.3. OTHER INDUSTRIAL APPLICATIONS OF TIMETAL ® PGMA

TIMET currently has PGMA in field trials in other industrial applications. These include:

• Polymer fabrication plant, Spain

• Chemical process column, Kingdom of Jordan

The company responsible for a chemical process facility in Jordan, is the Albemarle Corporation, who have successfully utilized the PGMA technology to both extend the service life of existing titanium equipment and to safeguard new titanium equipment in very aggressive chemical application. [4] In the first set of patch installations, a titanium ASTM grade 2 process column was selected, as this component is exposed to highly acidic halide at 250°F (121°C). The reducing acid environment allows for a high hydrogen uptake. The 42” (1067 mm) diameter column was experiencing active-passive breakdown in one region of the column. After welding PGMA patches very close to the previously-attacked region, no further metal loss or corrosion of any kind was observed and this unit is currently still in service with PGMA patches still providing protection 3 years after installation. In the second set of PGMA patch installations, Albemarle replaced a titanium ASTM grade 7 component - the upper section of a 36” (914mm) diameter column - with a lower cost option - titanium grade 2, plus PGMA patches installed. This unit is part of a halogen scrubber operating at 140°F (60°C). N o corrosion has been observed after over a year of service and the component is still corrosion-free to this date.

3. EXPERIMENTAL TRIALS OF PGMA ON AUTOCLAVES

3.1. ACID LEACHING AUTOCLAVE OPERATION

Minara’s Murrin-Murrin facility was chosen for the field testing; this is the largest facility for laterite nickel and cobalt refining in Australia. After mining from nearby open pits, ore is dumped on the run of mine (ROM) pad, where it is sorted according to grade and blended to ensure consistent feed to the feed preparation circuit. The pressure acid leach circuit consists of four giant autoclaves, each one approximately the size of a small submarine (4.95 m diameter by 35 m long) consisting of a carbon steel shell 102 mm thick, with a titanium (ASTM grade 1) liner, 6 mm thick, attached by explosive welding. [5] The chemical reactions which are involved in this metal-winning process are some of the harshest conditions in the chemical process industry worldwide, which is the reason that titanium was chosen for the lining material, as the carbon steel shell alone would not have provided sufficient protection for the autoclave. The overall corrosiveness of this process is so harsh that even stainless steels and nickel alloys alone cannot be considered, as corrosion would occur almost instantly on contact with the hot acid medium. This is the reason why autoclaves must be lined with titanium in the first instance.


6

However, taken together, the steel is used for the strength of the component as well as being chosen for a low price compared to the high corrosion resistant titanium alloy plate. The autoclaves themselves were fabricated by ASC Engineering. [6, 7] Their function in the HPAL process, is to carry the hot acid-leached slurry to the next component in the metal-winning circuit. The entire process is as follows: Firstly, the nickel and cobalt are leached out of the ore slurry, grading approximately 1.35% nickel and 0.09% cobalt, and into solution, by raising the pressure up to 44 atmospheres (4.45 MPa), where it is dosed with highly concentrated sulfuric acid (H2SO4) at a temperature of 255ºC (437ºF). The acid concentration was originally set at 98% with a pH of 2.5. [7] This process generates substantial quantities of heat and acid, which are later recycled throughout the plant. The ore is leached, with the valuable nickel and cobalt in a soluble form, which must be separated from the residue waste material. This solution is then "washed" to remove waste materials and is recycled back into the slurry ore preparation before eventually being pumped out to the tailings dam as neutralized and inert waste. Leached ore solution is then neutralized with calcrete, which is a mixture of gravel and sand cemented together with calcium carbonate. The solution is then passed into the mixed sulfides precipitation circuit, where H2S gas is added to convert the solution into a mixed nickel cobalt sulfide. The mixed nickel cobalt sulfide enters another autoclave where pure oxygen converts the solution from a mixed sulfide into a metal sulfate. At this point the nickel and cobalt molecules remain attached to each other. Impurities such as iron and zinc are removed, before the cobalt is separated using an organic reagent. The nickel sulfate solution then enters five parallel autoclaves, known as the hydrogen reduction circuit, where the hydrogen is added, liquids are separated and the remaining solids are converted into a dry powder of pure nickel. Finally, the powder is formed into a pillow-shaped briquette, sintered in a furnace and then packaged for transportation, with the entire process taking approximately 14 days.

3.2. SET-UP OF AUTOCLAVE PGMA PATCHES

The TIMETAL ® PGMA patches were produced by TIMET in sheet product form, from commercially pure titanium ASTM grade 1 + 1%Pd. The sheet dimensions are shown in Table 1. The sheet material was cut into coupons of size 16 mm x 32 mm.

PGM Plate ID Thickness (mm) Width (mm) Length (mm)

B12510 1.2 196 171

B12511 1.2 202 171

B12513 1.2 204 171

Table 1: PGMA Sheet and Coupon Dimensions

The coupons were placed on the intersection of a grid of squares of 750 mm size (Figure 3), then tack welded using titanium ASTM grade 2 (MIG) welding consumable.

Figure 3. Layout of PGMA patches in 750 mm grid pattern


7

In this study, four HPAL (High Pressure Acid Leach) Autoclaves, referred to as Trains 1, 2, 3 and 4 were targeted for corrosion protection. In particular, each autoclave has 6 compartments, where historically compartments 1 and 6 were shown to suffer from the highest level of pitting corrosion. The compartment liner, consisting of a titanium ASTM grade 1 liner, is the target structure to be protected from corrosion by exposure to the hot H2SO4 environment.

4. AUTOCLAVE CORROSION HISTORY

4.1. PITTING HISTORY

Since their initial commissioning, pitting surveys have been carried out in all HPAL autoclaves during internal inspections. During the first 5 years of operation, pitting in compartment 1 and 6 has been more wide spread and numerous than the other compartments. Typically compartment 1 would have as much as 600 individual pits identified and mapped in the course of an internal inspection; however in these early years, the depth of the pits were quite shallow, usually <1.0 mm. All pits over 0.3 mm deep were repaired, either by blending (smoothing out rough edges and corrosion damage by abrasion using hand held grinder and as abrasive "flap wheel", or a die grinder) or repaired by welding. In 2006 the repair consumable was changed to titanium ASTM grade 7 in order to utilize the added corrosion resistant benefit of alloyed Pd inherent in grade 7 into the repair welding carried out on the liner.

4.2. CHANGE IN CORROSION PIT CHARACTERISTICS

In 2004, the pattern of pitting in compartment 1 changed dramatically, with a sudden drop in the number of pits, the downside of this was that severity of the pitting increased. This was mainly seen in the depth of the pits, which increased from < 1.0 mm to 2 – 4 mm, a depth of pitting not normally seen (Figure 4). The reason for the changes in the pitting characteristics is not clear, but may be due to a range of issues such as ore blend, additives to the process and other variables.

Figure 4: Typical Corrosion observed after 2004

4.3. TITANIUM LINER BREACHED

In April 2007 a corrosive pit breached the 6 mm titanium liner in Autoclave 1 and exposed the carbon steel to the acidic slurry. This breach resulted in a cavity in the carbon steel 70 mm long, 45 mm wide by 20 mm deep (Figure 5); however at this point the corrosion of the carbon steel ceased. The reason for the limited corrosion appears to be due to scale buildup on the titanium liner which prevented contact between the slurry and the liner. The scale takes a few weeks to form and build the thickness required to isolated the liner, this may limit the amount of corrosive “fuel” available to the pit.

Titanium

grade 2

liner

Corroded

carbon steel

shell

scale


8

Figure 5: Liner breach in Autoclave 1

Over the next 9 months, two other autoclaves experienced liner breaches, however none were as destructive as the pit in Autoclave 1.

4.4. INSTALLATION OF PGMA

In 2006, the first 200 PGMA patches were installed in compartment 1 of Autoclave 4, the appliqué were set out in a 750 mm grid covering the lower third of the compartment 1. This “floor” area of the compartment 1 contains the highest population of corrosion pits and is usually the site of the most severe pitting. Installation of the PGMA patches in the other three autoclaves followed, with appliqué being tack welded to the liner of the autoclave (Figure 6), during each full internal inspection and de-scale occurring every 6000 operating hours. Installation of PGMA patches in to the compartment 6 of each train did not take place until 6 months after the initial PGMA installation in compartment 1. The reason for this being that the compartment 6 areas did not experience as significant corrosion as the corresponding compartment 1 in each train. However, the compartments were inspected for corrosion to the same schedule as the other compartments.

Figure 6: PGMA patch tack welded to autoclave liner


9

5. RESULTS OF FIELD TRIALS

5.1. RESULTS OF CORROSION PROTECTION BY TIMETAL ® PGMA

The results of the testing are summarized in Table 2. Initially, Train 4 compartment 1 had approximately 200 coupons welded onto the Grade 1 liner. After 3000 hours, an inspection was carried out where it was observed that there was no adverse effect of having these coupons installed. There also appeared to be minimal pitting in that compartment.

Placement of

patches

2006 inspection

2007 inspection

2008 Inspection

Train 4, Compartment 1

PGMA installed,

1 Pit - Repaired

Two small areas of pitting, – weld repaired

Minor pitting,

blended out

Train 1

Compartment 1

PGMA installed,

4 pits – weld repaired

1 area of shallow pitting, blended out

Train 1 Compartment 6

PGMA installed *

Shallow Pitts blended out No pits

Train 2

Compartment 1

PGMA installed,

1 large pit and small area of shallow pitting

4 areas of shallow pitting, blended out


PGMA installed *

No Pitting No Pitting

Train 3

Compartment 1

PGMA installed,

5 areas of shallow pitting

Minor Pitting,

blended out


PGMA installed,

Very minor pitting No pitting

* PGMA installed in compartment 6, approx. 6 months after corresponding compartment 1 in trains 1, 2, 3

Table 2: Results of TIMETAL ® PGMA corrosion protection

Figure 7. Typical Corrosion observed after 2008


10

5.2. FIELD REPAIR OPTIONS OF PGMA

The advantage of using PGMA process for additional corrosion protection is that the patches can be attached to the target component at any stage of the component’s life, including the time when the component is in service. Other engineering materials such as steels might experience corrosion during their service life and have to be replaced, whereas when using a titanium component in conjunction with PGMA this replacement requirement is negated. Due to the throwing power of PGMA to cover a certain area, if a patch becomes dislodged by erosion (movement of corrosive medium) or else removed by physical means such as during a cleaning operation, then re-application of the patch should present no problems. Patches are of a suitable size that can be held easily by a gloved hand during manual welding, which is probably the easiest way to re-attach a dislodged patch. As mentioned previously in 2.1 plating is another method of attaching the patches that could be easily accomplished on-site by any plant personnel using inexpensive equipment [1]. Maintaining a supply of additional patches will give some security in the case that future repair work is required.

5.3. PITTING SURVEY RESULTS

Although it is early in the trial (approximately two years), there appears to be very significant reduction in the severity of pitting corrosion in compartments 1 and 2 of all four autoclaves. During 2008, Trains 2 and 3 ran for 6000 hrs with no practical shutdown, which usually occurs at 3000 hrs. After descaling of the compartments, it was observed that the corrosion pits in these autoclaves have reduced dramatically in number and severity, (Figure 8). Given this data, Minara Resources Limited has now extended the run time between autoclave shutdowns to a nominal 6000 hrs. Increasing the hour between shutdows has increased autoclave availability by 720 hours per year; this will have a significant effect on the operation as downtime decreases and productivity increases.

Figure 8. Typical Pitting Corrosion observed after 2008

5.4. FREE ACID LEVELS

It was believed that the level of free acid in the process may have had some relationship with the extent of corrosion in the autoclaves, particularly in compartment 1, where acid is introduced into the process. Initially, the Murrin-Murrin site used acid rate of 400 kg per tonne of ore and 98% concentration [5]. However a review of free cid levels over the last four years [8], has found that although the free acid has level has dropped about 10 g/l from 70 to 60 g/l, the reduction does not appear to have a significant effect on corrosion. This is confirmed by an independant study commissioned by Minara Resources Limited which found there were no deleterious effects from free

acid (H2SO4) on titanium ASTM grade 1 up to 120 g/l at 255°C. This review of free acid level has recently been conducted and was not available until very recently (early 2009).

6. CONCLUSION

6.1. PGMA

PGMA has been shown to drastically reduce the amount of corrosion experienced by titanium components in what is probably the harshest environment in the chemical process industry today.


11

With PGMA protection, no further autoclave breaches have taken place and what corrosion there is, has been limited to a very small amount of shallow pitting, which is easily blended or weld repaired. Minara Resources Limited intends to continue the maintenance and installation of PGMA throughout all compartments of the four HPAL titanium lined autoclaves on site. The company is hopeful that PGMA will afford similar corrosion resistance to that found in autoclave manufacture with titanium ASTM grade 17 liners.

7. REFERENCES

1. J.S. Grauman et al, p. 26, Corrosion Resistance of Titanium, TIMET Handbook, Denver 1999

2. J.S. Grauman & T. Say, “Understanding the Behavior of and Preventing Problems with Titanium in Hydrometallurgical Pressure Leaching Equipment”, Proc. Conf. ALTA, Nickel/Cobalt Pressure Leaching & Hydrometallurgy Forum, Perth, Western Australia, May 11-12 1999.

3. J.S. Grauman, " PGMA - A Corrosion Protection Method Well Suited for Use of Titanium Within the CPI ". Proc. Conf. Corrosion Solutions, Coeur d’Alene, ID, USA, 2003.

4. H.T. Wells, PGMA Case Study, Albemarle Corporation, February 2009.

5. J. G. Banker, “Titanium clad autoclave performance in nickel laterite hydrometallurgy”, Proc. Conf. Randol Gold and Silver Forum, 2000, Randol Intl, April 2000, pp.253-257.

6. J.G. Banker & J.O. Winsky, “Titanium/steel explosion bonded clad for autoclaves and vessels”, Proc. Conf. ALTA Autoclave Design & Operation Symposium, Perth, Western Australia, May 1999.

7. R. Mayze, “An engineering comparison of the three treatment flowsheets in WA nickel laterite projects”, Proc. Conf. ALTA 1999, Nickel /Cobalt Pressure Leaching & Hydrometallurgy Forum.

8. Review of free acid levels in autoclave compartments, Internal Document, Minara Resources Ltd, 2009.

8. ACKNOWLEDGMENTS

The authors wish to thank Suresh Divi and Jim Miller for their assistance in research on PGMA as well the lab staff of K-53 at TIMET for preparing the PGMA patch test material. In addition, the test results are gratefully received from Minara Resources. Finally the authors wish to extend thanks to Tony Say and Titanium International for all their help and assistance in putting the test study together.



TREATMENT OF SULPHIDES


ENGINEERING ASPECTS OF THE PLATSOL™ PROCESS

By

M Wardell-Johnson*, G Steiper* and D Dreisinger **

* Bateman Engineering Pty Ltd, Australia

** PolyMet Mining, Canada

Presented by

Mike Wardell-Johnson

[email protected]

CONTENTS

1. ABSTRACT 2

2. INTRODUCTION 2

3. METALLURGICAL TESTWORK 11

4. CONCLUSION 20

5. ACKNOWLEDGMENTS 20

6. REFERENCES 20


2

ABSTRACT

Bateman Engineering has completed a Definitive Feasibility and Basic Engineering Package for a copper / nickel project based on the Platsol™ leaching technology. The fundamental process chemistry of Platsol™ is summarised together with other base and precious metal recovery circuits. A brief description of PolyMet’s NorthMet project is provided highlighting some engineering requirements imposed by the process chemistry. Reagents used in the leaching of bulk sulphide concentrates and management of Platsol™ leaching waste streams are discussed in addition to water balance and potential impurity build-up issues. Metallurgical testwork is described particularly in those unit processes that differentiate Platsol™ from other flotation concentrate leaching technologies. General plant layout and materials of construction of the Platsol™ autoclaves are examined.

1. INTRODUCTION

1.1 BACKGROUND AND HISTORY

Bateman Engineering provides considerable expertise in nickel, cobalt, copper and platinum group metals hydrometallurgy and specifically in pressure leaching technologies. Several current hydrometallurgical processing technologies offer viable alternative to smelting for the treatment of sulphide concentrates that contain low values in gold, platinum group metals (AuPGM) and base metal sulphides. The Platsol™ leach technology was developed in 1998 for PolyMet’s NorthMet project in north eastern Minnesota, USA in conjunction with Lakefield Research in Canada. The process involves addition of chloride ions to a total pressure oxidation (POX) autoclave employing high pressure and temperature leaching. The unique feature of Platsol™ is establishment of operating conditions where copper, nickel, cobalt, zinc and AuPGMs are placed into solution in a single operation and then sequentially recovered in the form of AuPGM precipitates, electrowon copper cathode and nickel-cobalt mixed hydroxide precipitate (MHP).

Bateman Engineering has completed the Basic Engineering Phase (FEED) of the project after successfully demonstrating the technical feasibility of the individual unit processes during the Definitive Feasibility Study (DFS) undertaken in 2006. The FEED package consisted of (but not limited to) the following:

• Project introduction;

• Process and Utility description;

• Preliminary process plant control philosophy;

• Process design criteria;

• Process plant mass and energy balance;

• Process plant equipment list;

• Electric motor list;

• Process flow diagrams;

• Preliminary piping and instrumentation diagrams;

• Preliminary plant layout drawings (Appendix G);

• General equipment specifications;

• Process equipment datasheets;

The package was delivered to PolyMet’s US based engineering and construction contractor in early 2007 and detailed engineering efforts commenced shortly thereafter. Construction of the project will immediately follow receipt of environmental and operating permits from the local, state and federal regulating authorities. Bateman is working actively with the detailed engineering company to ensure the intent of the DFS and FEED packages are being rigorously followed.


3

1.2 PLATSOL™ FUNDAMENTALS

Copper production from concentrates involves employing either pyrometallurgical or hydrometallurgical processes. Hydrometallurgical processing of concentrates using Platsol™ provides a viable alternative to smelting for extraction and recovery of sulphide concentrates containing base metals, gold and platinum group metals (AuPGMs). The process is relatively insensitive to deleterious metals such as bismuth, arsenic, selenium, mercury and lead in contrast to smelters that may have a limited tolerance to these metals. Further, hydrometallurgical processes do not generate dusts or gases such as sulphur dioxide which must be managed separately. Chrome and magnesium are also found in many ores and high concentrations of each have limited solubility in slag causing them to smelt at higher temperature. Platsol™ may offer lower capital and operating cost compared with a traditional smelter especially if the reserve is relatively small.

In total pressure oxidation environments (>200 oC, 700 kPa oxygen overpressure), base metals are

placed into a simple sulphate system in which AuPGMs have very low solubility. This can sometimes be advantageous as the residual AuPGM are concentrated into a comparatively smaller mass in the autoclave discharge in the same way pre-oxidation of refractory gold ores prepare the residue for subsequent leaching and recovery of metals from solution. It may be also possible to deliberately produce an AuPGM enriched residue that is amenable to concentration through flotation or a gravity technique then sold to third party refiners for final extraction of metals.

Introducing approximately 10 g/l of chloride (in the form of hydrochloric acid) into the autoclave feed slurry provides the chemical environment by which AuPGMs are placed into solution as chloride complexes according to the following simplified equations:

Chalcopyrite Dissolution 4CuFeS2 + 4H2O + 17O2 � 4CuSO4 + 2Fe2O3 + 4H2SO4

Cubanite Dissolution 2CuFe2S3 + 4H2O + 13O2 � 2CuSO4 + 2Fe2O3 + 4H2SO4

Pentlandite Dissolution NiFeS2 + 4O2 � NiSO4 + FeSO4

Pyrrhotite Dissolution 2Fe7S8 + 2H2O + 31O2 � 14FeSO4 + 2H2SO4

Linnaeite Dissolution 2Co3S4 + 2H2O + 15O2 � 6CoSO4 + 2H2SO4

Sphalerite Dissolution ZnS + 2O2 � ZnSO4

Gold Dissolution 4Au + O2 + 2H2SO4 + 16NaCl � 4Na3AuCl4 + 2Na2SO4 + 2H2O

Palladium Dissolution 2Pd + O2 + 2H2SO4 + 8NaCl � 2Na2PdCl4 + 2Na2SO4 + 2H2O

Platinum Dissolution 2Pt + O2 + 2H2SO4 + 8NaCl � 4Na2PtCl4 + 2Na2SO4 + 2H2O

Rhodium Dissolution 4Rh + O2 + 2H2SO4 + 24NaCl � 4Na3RhCl6 + 6Na2SO4 + 6H2O

Ferric Hydroxide Dissolution 2Fe(OH)3 + 3H2SO4 � Fe2(SO4)3 + 6H2O

Hematite Precipitation Fe2(SO4)3 + 3H2O = Fe2O3 + 3H2SO4

Hematite has the advantage of being “benign” with no ability to generate acid over time and presents a suitable material for long term disposal. Sodium jarosite may form if sodium chloride is used as the chloride ion provider which may decompose over time releasing acid according to the following reaction:


4

Na2SO4 + 3Fe2(SO4)3 + 12H2O � 2NaFe3(SO4)2(OH)6 + 6H2SO4

Excessive acid in the autoclave feed slurry resulting from recycled streams within the downstream base metal refinery reporting back to the autoclave may result in the formation of basic iron sulphate that will decompose over time releasing acid.

Fe2(SO4)3 + 2H2O � 2FeOHSO4 + H2SO4

As mentioned earlier, the key feature of the Platsol™ processes relates to high temperature and pressure leaching in the presence of oxygen gas and chloride ion. Leaching is done in an autoclave where such equipment is commonly used in the POX industry as it ensures maximum oxygen uptake and rapid chemical kinetics early during the leaching process.

The autoclaves are usually constructed from carbon steel which is lined with a corrosion resistant membrane overlain with two layers of acid resistant bricks that protect the shell from the aggressive chemical environment within. Autoclave wetted parts such as agitators, baffles and compartment dividers are constructed in titanium alloys to resist corrosion by the acidic leach liquor. Oxygen is injected into all autoclave compartments (usually five or six) in each autoclave at a controlled rate to ensure complete oxidation of all sulphide sulphur in the autoclave feed.

1.3 NORTHMET PROCESS DESCRIPTION

PolyMet plans to exploit a large, polymetallic, disseminated sulphide mineral deposit by conventional open pit methods at a rate of approximately 29,000 tonnes per day of run-of-mine ore and to extract base and precious metals. The selected ore processing route involves four stages of crushing followed by rod and ball milling with conventional froth flotation to produce a bulk sulphide flotation concentrate. Crushing and grinding will occur in a reactivated former taconite iron ore processing plant while a new flotation plant will be retrofitted into the same taconite facility. The bulk sulphide flotation concentrate will then be subjected to Platsol™ autoclave pressure oxidation leaching followed by hydrometallurgical extractive processes to produce high purity copper cathode, mixed nickel and cobalt hydroxides and a sulphide enriched AuPGM precipitate. An existing tailings basin will be reactivated for flotation tailings and hydrometallurgical residue disposal. The basic flowsheet is shown in Figure 1 with a brief unit process description thereafter.

1.3.1 Total Pressure Oxidation Autoclaves

The autoclaves chosen for PolyMet consist of two mild steel autoclaves each 26.1m long with an internal diameter of 4.47m equating to 325 m

3 of effective volume. Each autoclave is brick lined for

abrasion resistance and contains five internal compartments to moderate the flow rate of slurry through the pressure vessel to maintain residence time a little over one hour. Corrosion of the shell is prevented using a membrane type barrier between the shell and inner most courses of acid resistant bricks. To ensure complete oxidation, all compartments are fitted with agitators to thoroughly mix the slurry with sparge oxygen as it moves towards the discharge end of the autoclave and ensure maximum uptake of oxygen gas into solution. As oxidation of sulphide is exothermic, operating temperature within the claves is held at 225°C using neutralised raffinate as cooling solution. Oxygen is produced by a single on-site 660 t/d cryogenic oxygen plant which also supplies nitrogen gas for process plant blanketing in some areas and waste gas purging in others. Figure 2 shows a portion of the autoclave area general arrangement.


5

FIGURE 1 Simplified Process Flow Block Diagram

HCl, O2

Residues

SO2 AuPGM

Precipitate

CaCO3 CaSO4

Copper

Cathode

air, CaCO3 CaSO4

Fe(OH)3

Cooling water Al(OH)3

NaHS, N2

CuS Recycle

Mg(OH)2 Mixed Ni / Co

Hydroxide

Ca(OH)2

Residual MHP Recycle

Ca(OH)2 Mg(OH)2Mg Removal

Residual Cu Recovery

Stage 1 MHP

Stage 2 MHP

AuPGM Recovery

Solution Neutralisation

Cu SX / EW

Raffinate Neutralisation

Pressure Oxidation

HydroMet Tails DamFlotation Concentrate

FIGURE 2 General Layout of Autoclaves and Feed Systems


6

Slurry retention time was determined by metallurgical testwork to ensure complete sulphur oxidation and each autoclave has a dedicated single stage pressure let-down vessel. Hot pressurised slurry is discharged from each autoclave in a flash process which generates steam. All vented gas is directed to a dedicated gas scrubbing system where waste heat is recovered for heating process streams, as required and cleaned vent gases are emitted directly to atmosphere. Base metal extraction is rapid and is essentially complete within the first half of the autoclave while complete sulphur oxidation and AuPGM dissolution occurs towards the back half.

The temperature of hot slurry discharging from the flash vessels is reduced to 60 °C using spiral heat exchangers and subsequently pumped to a leach residue thickener. Cooling solution for the heat exchangers is the feed solution for residual copper recovery as this unit process requires additional heat to replace that which is removed prior to copper solvent extraction. Approximately one-quarter of the 55% solids (w/w) thickener underflow reports directly to a vacuum belt filter where the final residue is washed, re-pulped in acidified process water and pumped to the Hydrometallurgical Residue Facility. The remaining three-quarters is recycled back to the autoclave feed tank where it is used to moderate the exothermic oxidation process, maximise sulphide oxidation and provide an additional chance to ensure a high degree of AuPGM extraction. Design provides for up to 250% recycle around the autoclave though pilot testing showed that acceptable oxidation could be achieved with as little as 100% recycle.

The key process parameters and their main influence on autoclave operation include:

• Concentrate feed sulphur content – required for autothermal conditions to achieve target leaching temperature;

• Autoclave feed pulp density – determines oxygen uptake rate and slurry rheology;

• Degree of recycle employed – residence time decreases as slurry pulp density increases;

• Autoclave feed rate – residence time and leaching temperature;

• Initial acidity of the autoclave feed slurry – formation of basic iron sulphate (excess acid) or slower leaching kinetics (insufficient acid);

• Chloride concentration in the autoclave feed – AuPGM extraction;

• Oxygen flowrate, over-pressure, presence of inert gas and amount required above stoichiometric – degree of sulphur oxidation, mass flow of waste gas to the vent system;

• Agitator type and speed – transfer of oxygen from the head space into the slurry.

1.3.2 Gas Scrubbing

Autoclave vent gas scrubbing ensures gases such as nitrogen, carbon dioxide and oxygen which discharge from the autoclave via pressure control vents or as a result of the flash cooling process are cleaned of residual solid particles and acidic liquors in two stages. Firstly by dedicated venturi type packaged scrubbers and finally by a common packed tower scrubber. A portion of the partially scrubbed gas maybe utilised for heat recovery in the mill process water heaters or for general building heating requirements.

1.3.3 AuPGM Precipitation

AuPGM recovery from solution is achieved using synthetic covellite (CuS) produced in residual copper recovery while maintaining base metals in solution. Prior to cementation, ferric iron is reduced to ferrous using sulphur dioxide gas which is sourced from a dedicated packaged bullet. The saleable concentrate contains AuPGM associated with copper sulphide and is suited to treatment in an offsite AuPGM processing facility. Salient and simplified reactions for the process include:

Reduction of Ferric Iron SO2 + H2O � H2SO3

Fe2(SO4)3 + H2SO3 + H2O � FeSO4 + H2SO4

Cementation of AuPGM 2Na3AuCl4 + CuS + H2SO4 � 2Au + 6NaCl + 2HCl + S + CuSO4


7

Na2PtCl4 + CuS + H2SO4 � Pt + 2NaCl + 2HCl + S + CuSO4

Na2PdCl4 + CuS + H2SO4 � Pd + 2NaCl + 2HCl + S + CuSO4

AuPGM enriched copper sulphide is thickened with a portion of the underflow reporting to a filter while the balance is recirculated to the cementation stages. Vent gas is directed to a central plant scrubber where residual acidic liquor is removed before discharge to atmosphere.

1.3.4 Solution Neutralisation & Gypsum Production

The aim of the solution neutralisation is to remove residual acid using limestone to precipitate gypsum. This reduces the process liquor stream acidity to a level appropriate for subsequent copper solvent extraction while avoiding any precipitation of iron or base metals. Gypsum produced in this process is thickened with a portion of the underflow reporting to a filter where a saleable, high quality, synthetic grade product of 96.8% purity which may be packaged and exported from site should there be a buyer of such a product or repulped and directed to the tails dam if there is not. The balance is returned to the first stage of solution neutralisation as seed material.

Precipitation of Gypsum CaCO3 + 2H2SO4 � CaSO4 + H2O + CO2

CaSO4 + 2H2O � CaSO4.2H2O

1.3.5 Copper Solvent Extraction

Solvent extraction (SX) is a proven and widely practised process for removing copper ions from neutralised solutions to produce concentrated solution suitable for recovery of copper metal by electrowinning (EW). Three stages of copper extraction using an organic solvent produce a Cu rich organic phase which is washed in a single wash stage to remove residual chloride. Two stages of stripping using acidic spent electrolyte from EW is employed to remove copper from the organic which increases copper tenor in the electrolyte before being returned to EW. Raffinate from the SX extraction stages is depleted in copper and pumped to the combined raffinate neutralisation/iron and aluminium removal circuit.

Extraction / Stripping of Copper from Organic (R) CuSO4 + RH2 � RCu + H2SO4

1.3.6 Copper Electrowinning

Copper metal is electro-deposited from filtered electrolyte onto stainless steel cathode blanks to meet London Metals Exchange quality specifications. Cathodes are harvested on a rotation basis every seven days using an overhead crane, a cathode washing system and an automated stripping machine. Stripped metal is strap-packaged onto pallets and shipped for sale. Spent electrolyte is returned to the SX strip circuit to act as stripping solution.

1.3.7 Raffinate Neutralisation

After SX has removed essentially all copper from solution, the raffinate is partially neutralised to reduce the acidity produced during copper extraction. Solution pH is increased using limestone and air is added to ensure ferric iron is present to a point whereby iron and aluminium precipitate as hydroxides. Co-precipitation of gypsum occurs with the combined precipitate treated with similar equipment to that used in the solution neutralisation circuit. A portion of the thickener underflow is directed to a filter where the precipitate is washed in process water and repulped while the balance reports back to the raffinate neutralisation tanks as seed material.

Precipitation of Iron and Aluminium Fe2(SO4)3 + 6H2O � 2Fe(OH)3 + 3H2SO4

Al2(SO4)3 + 6H2O � 2Al(OH)3 + 3H2SO4

Gypsum produced in this unit process is not of sufficient quality for sale and must be directly to the on-site tailings facility.


8

1.3.8 Residual Copper Recovery

Copper SX is operated in such a way as to pass a small amount of copper in solution to the downstream unit processes. The objective of the residual copper recovery stage is to reclaim soluble copper from the raffinate neutralisation discharge liquor to produce synthetic covellite (along with minor amounts of other base metals) for use as a cementation agent in AuPGM recovery. This is achieved by reacting residual copper feed solution with sodium hydrosulphide. Excess CuS is directed to the POX autoclave as required. Solution temperature is increased to a point whereby precipitation kinetics is optimised using waste heat from the POX autoclave.

Precipitation of Covellite 2CuSO4 + 2NaHS � 2CuS + Na2SO4 + H2SO4

A single thickener is used to increase the CuS slurry density with a portion of the underflow being directed to AuPGM recovery while the balance is returned to the head of the residual copper recovery circuit. Hydrogen sulphide gas may be evolved in the process and all vent gases containing such off-gases report to a central packed bed scrubber. Nitrogen gas is introduced into each reaction tank to exclude air and maintain solution potential.

1.3.9 Base Metal Mixed Hydroxide Precipitation

The base metal hydroxide precipitation circuit allows recovery of nickel and cobalt from solution as a saleable product. This is achieved using two discrete precipitation stages whereby milk of magnesia is used in the primary stage while lime is used to scavenge remaining base metals from solution in the second stage. Precipitation with milk of magnesia produces high purity nickel and cobalt hydroxide with little unreacted or co-precipitated magnesium. The precipitate is thickened with a portion of the underflow directed to a product filter while the balance is returned to the first stage precipitation tanks to minimise magnesia consumption and act as seed material.

First Stage Base Metal Precipitation NiSO4 + Mg(OH)2 � Ni(OH)2 + MgSO4

CoSO4 + Mg(OH)2 � Co(OH)2 + MgSO4

Addition of lime milk increases solution pH further in the second stage thereby recovering all remaining nickel and cobalt albeit with significant magnesium co-precipitation. This precipitate is thickened with a portion of the underflow directed to the solution neutralisation circuit to re-leach the precipitated base metals and reduce consumption of limestone while the balance reports to the head of the second stage precipitation tanks as seed material. Any manganese present in the solution will also precipitate from solution at this stage.

Second Stage Base Metal Precipitation NiSO4 + Ca(OH)2 � Ni(OH)2 + CaSO4

CoSO4 + Ca(OH)2 � Co(OH)2 + CaSO4

MnSO4 + Ca(OH)2 � Mn(OH)2 + CaSO4

1.3.10 Magnesium Removal

Following recovery of base metals and AuPGMs from upstream processes, magnesium sulphate remains in any significant solution concentration. Magnesium is partially precipitated as a hydroxide using lime milk together with gypsum to prevent development of a recirculating load of magnesium within the hydrometallurgical plant water balance. The combined precipitate is directed to the hydrometallurgical tailings facility without thickening or filtration.

Magnesium Removal MgSO4 + Ca(OH)2 � Mg(OH)2 + CaSO4

1.3.11 Hydrometallurgical Residue Disposal

The Platsol™ process yields hydrometallurgical residues that are produced as separate process residue streams from:

• the autoclave discharge (predominantly hematite, minor silicates and other gangue species);

• solution neutralisation (gypsum, with minor unreacted limestone and silicates);

• crud from SX (silicates with minor organic);


9

• raffinate neutralisation (gypsum, iron and aluminium hydroxides, minor silicates and unreacted limestone);

• Magnesium removal (gypsum and magnesium hydroxide, unreacted lime milk).

Such residues are benign though are typically stored separately from beneficiation plant tailings unless the materials of construction upstream of a Platsol™ refinery can tolerate solutions containing chloride. Hydrometallurgical residue disposal facility design must satisfy local, state and federal environmental requirements which aim to prevent release of substances resulting in adverse impacts on natural resources. Residue cells may be lined with natural clay, geosynthetic clay or synthetic membranes to reduce seepage with the supernatant water being returned to the hydrometallurgical plant for re-use.

Factors that should be considered during preliminary design of the liner and cover systems include:

• Compatibility with residue chemical characteristics;

1.3.12 Hydraulic conductivity and containment ability;

• Construction season limitations;

• Tolerance of differential settlement;

• Freeze / thaw cycles if the mine area is located in a cold climate;

• Ease of installation;

• Wet transport and disposal as opposed to dry transport and disposal methods;

• Ability to accommodate vertical cell development;

1.4 REAGENT SUMMARY

The Platsol™ hydrometallurgical process treats acidic liquors and potentially acid generating species by converting active materials such as mineral sulphides to hematite and sulphuric acid to gypsum. Acid neutralisation forms a considerable part of the total operating cost through the use of limestone, lime and milk of magnesia. It is therefore important to consider local suppliers for these reagents to minimise supply and transport costs.

The principal plant reagents used in a typical Platsol™ process in order of decreasing consumption rate are:

• Limestone;

• Milk of Magnesia (61% w/w);

• Hydrochloric Acid (32% w/w);

• Quicklime;

• Sulphuric Acid (93% w/w);

• Sulphur Dioxide;

• Sodium Hydrogen Sulphide;

• Flocculent;

• SX Diluent;

• Caustic Soda (50% w/w);

• SX Extractant;

• Guar Gum;

• Cobalt Sulphate.


10

Reagents should be received and stored according to local regulations and following the information provided from their respective Materials Safety Data Sheet. Liquid sulphur dioxide is usually stored in a pressurised storage bullet adjacent to the point of injection to minimise vapour pressure losses within process piping. Copper solvent extraction reagents are combustible in nature and must be received, stored and used with this in mind.

1.5 HYDROMETALLURGICAL REFINERY WATER BALANCE

Bateman Engineering developed a complete site water balance for the NorthMet Process Plant providing a quantitative description of all input and output water flows required for the effective operation of the facility. The water used in the facility is physically separated into two distinct plant areas:

• Concentrator Water Balance;

• Hydrometallurgical Plant Water Balance;

Testwork, modelling and simulations have shown the Concentrator will require a continuous, net make-up of fresh water with no requirement for purging. In the hydrometallurgical refinery, water losses occur as steam is vented to atmosphere from the autoclave pressure let-down system. As a result, the hydrometallurgical plant will require continuous fresh water make-up without requirement to periodically purge water from the hydrometallurgical system. The major streams considered when simulating the hydrometallurgical plant water balance are:

• Water entrained in incoming concentrate feed;

• Fresh water make-up from an aquifer, lake or river;

• Rainfall with seasonal snow and ice melt;

• Water entrained in residues and return water from tailings disposal cells;

• Various process plant vents though principally as steam from the autoclave pressure flash let-down vessels, scrubbers and thickeners.

Strong emphasis is placed on maximising water recycling within all stages of the hydrometallurgical process. From the processing point of view, it is important to avoid build-up of impurities in process streams that might require purging to the environment at some stage. To ensure impurity build-up is avoided, water and mass balance simulations were carried out with particular reference to impurities such as chlorides, magnesium and sodium. The simulation results suggest no situation was found that would require purging and discharge to the environment, a conclusion which was supported by the absence of impurity build-up during the integrated pilot plant.

Water inputs to the Hydrometallurgical plant were identified as:

• Concentrate Product – The amount of water contained in the concentrate fed to the Hydrometallurgical plant from the Concentrator as determined from the pilot test work data;

• Concentrate Water for Heating – The amount of Concentrator water that is transferred to the Hydrometallurgical Plant for non-contact heating;

• Reagents – The amount of water accompanying various reagents added to the Hydrometallurgical Plant. Reagent consumption was determined from the pilot test work data with typical results shown below:

Reagent % Water Content

Magnesium hydroxide 70

Sodium hydrosulphide 70

Hydrochloric acid 68

Caustic 50

Sulphuric acid 7

• Process Air – The water vapour content of the process air is determined from a standard calculation, which sets the equilibrium saturated water vapour composition (from a steam table) at an estimated air blower discharge temperature of 40

oC;


11

• Precipitation – The amount of rain and/or snow falling onto the project as described by nearby meteorological monitoring stations;

• Raw Water Input – The amount of water required to balance the Hydrometallurgical Plant and is sourced from a nearby raw water supply.

The water outputs from the Hydrometallurgical Plant were identified as:

• Plant Vents – The amount of water vapour vented to the atmosphere from the Final Autoclave Gas Scrubber, Plant Scrubber and Electrowinning Mist Scrubber. This value is calculated from the equilibrium saturated water vapour composition (from a steam table) at the exit gas stream temperature;

• Products – The amount of water contained in the various products from the Hydrometallurgical Plant including copper cathode (minimal), AuPGM precipitate, and base metal mixed hydroxide;

• Reactive Residue – The amount of water contained in the slurry that is ultimately pumped to the Reactive Residue Dam. The amount of water is calculated by multiplying the mass flow of slurry from each stream by the slurry water content;

• Hydrometallurgical Plant Evaporation – The amount of water evaporated from various thickeners within the facility. The value is calculated from the slurry temperature, thickener surface area and equilibrium saturated water vapour composition (from a steam table);

• Evaporation – The amount of water lost to the atmosphere due to natural evaporation as determined by nearby meteorological monitoring stations;

• Chemically Consumed Water – The amount of water that is consumed in the various chemical reactions that occur in the Hydrometallurgical Plant. These reactions and their extent were determined by the bench and pilot plant test work. An example reaction, involving the neutralisation of sulphuric acid with limestone is shown below.

CaSO4 + 2H2O � CaSO4.2H2O

2 METALLURGICAL TESTWORK

Development of the Platsol™ metallurgical process has taken place via several integrated pilot plant testwork campaigns from as early as 1999. A Platsol™ style base metal and AuPGM refinery was fully piloted as part of the PolyMet DFS testwork programme in 2005. Subsequent testwork in 2007 focused on the impact of recycling different portions of leach residue thickener underflow back to the autoclave to maximise AuPGM extraction. Both programs included bench scale and continuous piloting at SGS Lakefield in Ontario, Canada and aimed to develop and demonstrate the effectiveness of a complete process flowsheet for treatment of NorthMet sulphide material. Further, it aimed to generate data of sufficient quality to allow engineering of the flowsheet from ROM ore receipt to production of saleable products. The technical viability of Platsol™ was demonstrated on an ore type that contains values in copper, nickel, cobalt, platinum, palladium and gold.

The flowsheet arising from this testwork subsequently served as the basis on which the plant was designed by Bateman Engineering to process 29,030 metric tonnes per day or 10.6 million metric tonnes per year of ROM ore to produce approximately 35,000 tonnes per year of copper cathode.

2.1 HYDROMETALLURGICAL PLANT TESTWORK IN 2005-2006

The 2005-2006 pilot plant program was designed, supervised and interpreted by Bateman, SGS Lakefield and PolyMet Mining to confirm the entire metallurgical flowsheet from ore comminution, beneficiation, Platsol™ hydrometallurgical leaching performance, solution purification unit processes to final product recovery. It also was designed to provide the basis for the process plant, to collect extensive environmental data and to optimise aspects of the process, in particular:

• To recycle a portion of the leach residue to the autoclave for improved AuPGM extraction and autoclave design optimisation (reduced autoclave sizing);

• To allow precipitant selection and optimisation for iron reduction and AuPGM recovery;


12

• To investigate an option to separate Co and Zn by solvent extraction prior to Ni hydroxide precipitation, as an alternative to precipitation of a mixed Ni-Co-Zn hydroxide product.

The metallurgical testwork program considered processing approximately 40 tonnes of bulk sample consisting of PQ diamond drill core for flotation testwork and subsequent production of concentrate for hydrometallurgical pilot plant testing. Another eight tonnes of drill core was made available from PolyMet in April 2006 for additional pilot scale testwork. Metallurgical test composites were prepared from individual core samples to be generally representative of the expected life of mine ore copper feed grade range.

Pre-piloting bench testwork was conducted to optimise circuit conditions for the pilot-scale testwork, in particular temperature, residence time and reagent additions for a number of unit operations. The results of this testing were then incorporated into the pilot plant design and operating philosophy. A number of bench programmes were undertaken to provide important information for final design.

Outcomes and conclusions from hydrometallurgical pilot plant work are summarised in Table 1

TABLE 1 Pilot Plant Test Outcomes and Conclusions

Flowsheet Area Pilot Plant Conditions and Outcomes

Pressure Oxidation Optimum autoclave operating parameters included: operating at 225

oC, ~3,100kPag total pressure, ~800kPa O2, 10g/l chloride and a 1.1

hour first pass residence time. Metals extractions were shown to improve by the introduction of a 200% residue recycle stream (i.e. a 2:1 ratio of leach residue to fresh feed). Average extractions for metals at optimum conditions were: Cu 99%, Ni 99%, Co 98%, Au 89%, Pt 93% and Pd 94%.

AuPGM Recovery AuPGM were precipitated from solution by adding CuS, recycled from the residual Cu recovery circuit. Recoveries were excellent with below detection limit values for AuPGM remaining in solution.

Further testwork led to the re-introduction of sulphur dioxide (SO2) in the final flowsheet as a reductant for iron prior to CuS addition. This reduces the consumption of CuS and limits the elemental S content of the concentrate. The SO2 pre-reduction system was originally tested and piloted in the year 2000 pilot plant at SGS-Lakefield.

Solution Neutralisation – gypsum recovery

This circuit operated to pH 1.3 using ground limestone addition while gypsum thickener underflow was recycled as seed to the first reactor. Analysis of the gypsum residue reported insignificant base metal content and low residual carbonate (0.07%) making it a synthetic gypsum of marketable quality.

Copper Solvent Extraction

Copper was extracted at 40° C in 3 counter current stages, scrubbed in 1 stage (to prevent chloride transfer to Cu electrowinning) and stripped in 2 stages. Two organic extractants, Acorga® M5640 and LIX® 973NS LV, were pilot tested. Orfom® CX80CT diluent was used in each case.

Recovery of Cu to the strip liquor averaged 95.5% for both extractants, producing raffinate with Cu <1.0 g/L from PLS ranging 18-25 g/l Cu. No evidence of crud formation during testing was noted.

Copper Electrowinning

A total of 69 kg of copper metal was produced. Cathodes were harvested twice during the program.

Four cathodes were sampled for purity; two from each extractant cycle. Cathodes from Cycle 2 met LME grade A specifications while cathodes from Cycle 1 showed minor contamination of Pb and S attributed to erratic temperature control during test start-up.

Raffinate Neutralisation

Raffinate is neutralised prior to recycling a portion as cooling solution back to the autoclave. This is necessary to reduce the free acid level in the autoclave product solutions and prevent the formation of basic ferric


13


sulphate.

The pH set points for raffinate neutralisation varied between 1.2-1.5 and were controlled via limestone slurry addition. Loss of Ni and Co to the residue was minimal.

Iron Removal The balance of neutralised raffinate solution was directed to nickel and cobalt recovery. The first step in the Ni and Co recovery circuit is iron removal by oxidation and neutralisation. Ferrous iron was oxidised to ferric iron by addition of gaseous oxygen and removed from solution (along with aluminium) by hydroxide precipitation. Limestone was added to achieve the target pH of 4.2.

Iron removal residue consisted predominantly of gypsum with low levels of iron and aluminium hydroxides. Ni and Co losses in the residue were minimal.

Iron and aluminium removal efficiencies were 99.9% and 94.1% for this circuit.

Aluminium Removal A separate stage of aluminium removal was included in the pilot plant circuit. In practice, this circuit did not consume limestone, as pH naturally rose to 4.6-4.7 due to an excess of alkalinity from the iron removal stage.

Iron and aluminium removal efficiencies were 71% and 96% respectively (to give overall precipitation efficiencies of nearly 100% after two stages).

Residual Copper Recovery

Residual copper was precipitated as copper sulphide (CuS) using sodium hydrosulphide (NaHS), and collected for use in AuPGM recovery. A fraction over stoichiometric addition of NaHS was required for 92% Cu precipitation with insignificant co-precipitation of Ni and Co.

Mixed Hydroxide Precipitation

Stage 1 (MHP-1)

Ni and Co were precipitated as a mixed hydroxide using magnesium hydroxide slurry to a target efficiency of 85%. The mixed hydroxide precipitates collected during the pilot plant analysed 31.5-36.3% Ni, 1.67-1.92 % Co, 0.31-0.37% Cu, 0.51-0.59% Fe, 4.27-4.84% Zn and 0.62-1.04% Mg.

Mixed Hydroxide Precipitation

Stage 2 (MHP-2)

This circuit recovered residual nickel and cobalt from solution by precipitation with hydrated lime slurry at pH 8.

Precipitate was thickened and recycled to the neutralisation circuit (where the residual metal hydroxides redissolved).

Removal efficiency of residual Ni and Co from the feed solution averaged 93% and 92% respectively giving overall precipitation efficiencies through the two stages of hydroxide precipitation of nearly 100% for both Ni and Co.

Magnesium Removal

Magnesium was removed from the barren solutions after Ni and Co recovery by addition of hydrated lime slurry to pH 9. Mg precipitation was close to the target 50%. The magnesium hydroxide – gypsum product slurry was thickened, with overflow used as process water and underflow directed to tails.

The absence of pay metals in the feed to magnesium precipitation resulted in negligible Ni and Co losses (0.14% and <0.02% respectively).


14


Co/Zn Solvent Extraction

The cobalt and zinc solvent extraction circuit was run as part of the program to produce purified metal hydroxides (rather than mixed hydroxide precipitation). Bulk Co/Zn extraction was achieved in 4 stages at pH 5.0-5.5 and 55° C, using 5 %v/v Cyanex® 272 extractant in Orfom® SX80CT diluent. The higher temperature favoured Co extraction and displacement of co-extracted Mg.

Co stripping took place in 3 stages at pH 3 and 45oC, before Zn

stripping in 2 stages at pH <1 and 40oC. Co extraction rates greater than

96% were achieved, with raffinate grades of below 10 ppm Co. Zinc extraction was greater than 99.9%. No evidence of crud formation during testing was noted and the circuit operated smoothly.

This unit process was later abandoned from the process flowsheet as the nickel / cobalt hydroxide product was of sufficient quality to make it saleable without attracting significant penalties.

Process Plant design criteria were derived or obtained from a number of sources including testwork reports, vendor testing and/or recommendations, previous Bateman projects, consultants, contractors and from PolyMet. As the unique features of the Platsol™ process involves the autoclave and AuPGM unit processes specifically, only these processes are described further.

2.1.1 Recycle Rate and Metal Extraction

The commercial plant process design strategy involves recycling a portion of the autoclave discharge residue back to the autoclave feed to maximise sulphur oxidation and AuPGM extraction. This has the effect of increasing the solids residence time without the capital expense of a larger autoclave. This recycle concept had been tested at the laboratory scale in 2005 but was not incorporated in the full 2005 pilot plant.

The recycle concept was tested in a further, short autoclave only pilot plant run. It is crucial especially with a short testwork campaign to have sharp and definitive changeover of recycle conditions in the autoclave feed makeup. This was accomplished by SGS Lakefield using a small mix pot ahead of the autoclave metering feed pumps to blend the various components of the feed slurry. Peristaltic pumps were used to feed a calculated blend of concentrate, autoclave discharge recycle slurry and synthetic raffinate dilution solution each of which was stored in separated mix tanks. The recycle slurry was made up in batches every hour of operation using the latest discharge filter cake and filtrate analyses. A target of 55% solids was used in making up the recycle slurry to simulate leach residue thickener underflow as would be found in the eventual commercial plant.

Extractions for each recycle condition tested are listed in Table 2 with graphical representation shown in Figure 3 and 4. Assay values from which the core metal (Au & PGM, Cu, Ni, Fe) extractions are based are average steady state values. The extractions without any recycle were 99.5%, 97.0%, 97.2%, 91.3%, 87.7% and 84.7% for copper, nickel, cobalt, gold, platinum and palladium respectively.

Recycling autoclave discharge residue results in an increase in base and AuPGM metal extraction of approximately 7% for platinum and palladium extraction, 1% for gold and 1% for nickel with copper and cobalt extractions remaining high (99.4% and 97.5% respectively). The pilot plant was designed and developed by Bateman using a scaled version of the full metallurgical flowsheet with mathematical modelling using MetSim mass and energy balancing software. MetSim is an industry standard metallurgical simulation and design computer software package.


15

TABLE 2 Results of Recycle Testwork

Recycle Rate, % Cu Co Ni Fe Au Pt Pd S2-

0 99.5 97.2 97.0 3.8 91.3 87.7 84.7 98.5

75 99.4 97.5 98.1 8.8 92.4 94.1 91.2 100.0

200 99.4 97.5 98.4 19.4 88.8 92.7 93.6 98.9

250 99.2 97.0 98.4 20.3 90.5 92.7 93.3 98.5

with respect to new feed

Extractions and Sulphide Oxidation Efficiency

FIGURE 3 Effect of Recycle Rate on Base Metal Extraction

96.0

96.5

97.0

97.5

98.0

98.5

99.0

99.5

100.0

0 50 100 150 200 250

Recycle Rate, %

Ba

se

Me

tal E

xtr

actio

n,

%

0

5

10

15

20

25

Fe

Extr

actio

n,

%

Cu Co Ni Fe

FIGURE 4 Effect of Recycle Rate on AuPGM Extraction & Sulphur Oxidation

80.0

85.0

90.0

95.0

100.0

0 50 100 150 200 250

Recycle Rate, %

Base M

eta

l Extr

actio

n, %

Au Pd Pt S(2-)

Autoclave sizing calculations are used to combine data obtained from metallurgical testwork programs with mass and energy balance output such as the example shown in Table 3, thermodynamic data, “guesstimates for aeration volumes, brick lining specifications and the like”, physical information such as solids and solutions specific gravity to yield the required values for vessel size and shape.


16

TABLE 3 Example Autoclave Volume Calculation Output

Parameter Unit 0 100 150 200 250

New Feed

Sulphur Content %S 22% 22% 22% 22% 22%

Solids t/h 42.1 42.1 42.1 42.1 42.1

Solution t/h 22.7 22.7 22.7 22.7 22.7

Hydrochloric Acid t/h 2.6 2.6 2.6 2.6 2.6

Neutralised Raffinate t/h 330.5 277.2 251.8 227.8 205.0

Total t/h 397.7 344.5 319.1 295.1 272.3

Volume Flowrate m3/h 327.6 280.4 258.0 236.7 216.5

Recycle to Autoclave

Solids t/h 0.0 42.1 63.1 84.1 105.2

Solution t/h 0.0 63.1 94.6 126.2 157.7

Total t/h 0.0 105.2 157.7 210.3 262.9

Volume Flowrate m3/h 0.0 61.1 91.7 122.3 152.9

Combined Autoclave Feed

Sulphur Content % S 22% 11% 9% 7% 6%

Solids t/h 42.1 84.1 105.2 126.2 147.2

Solution t/h 22.7 85.8 117.3 148.9 180.4

HCL t/h 2.6 2.6 2.6 2.6 2.6

Dilution t/h 330.5 277.2 251.8 227.8 205.0

Total t/h 397.7 449.6 476.9 505.4 535.2

Calculated Volume Flowrate m3/h @STP 327.6 341.6 349.7 358.9 369.3

Calculated Volume Flowrate m3/h @temp 407.2 434.2 431.4 444.0 453.7

RecycleRate, %

2.1.2 Iron Reduction and AuPGM Recovery

The stability of the leached AuPGM species in the autoclave discharge were tested by timed sampling of slurry taken from the pilot plant discharge. This was important to confirm that the AuPGM would not re-precipitate and be lost during the post autoclave solid-liquid separation steps. The stability of AuPGM in solution was proven to be independent of agitation and temperature within the range of conditions tested. Agitating the solution also did not result in the precipitation of the dissolved AuPGM’s. The test results indicated that the dissolved Au, Pd and Pt in the autoclave discharge would remain in solution through the solid/liquid separation stage. Rheology tests were carried out on slurry samples recovered from the Pilot Plant operation.

Autoclave testwork was conducted to upgrade AuPGM content of the AuPGM concentrate. This was done by selective re-leaching of base metals and sulphur from the AuPGM concentrate product, at both high and low temperatures. This work confirmed an optimum average temperature of 195°C to upgrade the AuPGM precipitate from approximately 1,000 g/t AuPGM to 16,000 g/t AuPGM. This particular processing route was subsequently abandoned due to the high capital cost associated with separate autoclaving and recovery systems.

AuPGM Precipitation tests were conducted to determine if CuS could be effective at both reducing ferric iron and removing AuPGM by a cementation type process from the same solution. The test results indicated that CuS could be used in this dual purpose at an addition rate between 1 and 10 g/l. Where the feed solution contained elevated ferric iron required a larger amount of CuS (≥ 5 g/L) to completely remove Pt from solution.

Solution Pt concentration was reduced to <0.01 mg/L after reaction time of 5 minutes for the tests with CuS additions of greater than 5 g/L. The reaction solution ORP value was used as a measure of the degree of ferric iron reduction. All tests achieved ORP values close to the target value of 400 mV (vs. Ag/AgCl), where the ferric iron concentration is vanishingly small. The test data indicated that CuS, at addition levels approximately 5 g/L, would be affective in a dual role as both a reductant and a precipitant. Tests confirmed that the iron reduction/AuPGM precipitation steps, if necessary, could be conducted in two stages. The precipitation of the AuPGM’s could still be achieved using CuS in the second stage. The residual reductant (Na2S2O5) from the first stage did not reduce effectiveness of CuS as a precipitant.

The influence of temperature and agitation on the stability of the dissolved AuPGMs in the autoclave discharge was studied. The addition of agitation and heat did not destabilise the AuPGMs in the autoclave discharge as highlighted in Table 4 and Figure 5.


17

TABLE 4 Effects of Agitation and Time on AuPGM Stability

Time, h Au Pt Pd

0.00 0.31 0.43 2.14

0.00 0.31 0.42 2.16

0.25 0.34 0.47 2.14

0.25 0.31 0.46 2.09

0.50 0.34 0.47 2.24

0.50 0.35 0.51 2.37

1.00 0.33 0.45 2.19

1.00 0.38 0.54 2.57

2.00 0.29 0.44 2.11

2.00 0.30 0.45 2.21

4.00 0.25 0.34 1.68

4.00 0.36 0.52 2.54

Average 0.32 0.46 2.2

Std.Dev. 0.01 0.02 0.07

Solution [M], mg/L

FIGURE 5 Effects of Agitation on AuPGM Stability

0.3

0.4

0.5

0.6

0.7

0.0 1.0 2.0 3.0 4.0

Time, h

Co

ncen

trati

on

Au

, P

t

(mg

/l)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Co

ncen

trati

on

Pd

(m

g/l

)

Au Pt Pd

A variety of thickener and filtration equipment supply vendors were present during piloting to perform bench tests on slurry samples withdrawn from the operating pilot plant. The results of this testing have been used to provide equipment design parameters.

5.1 PLANT ARRANGEMENT

The ROM size reduction process areas make use of the existing primary, secondary, tertiary and quaternary crusher houses with associated feeding and conveying equipment. The existing Concentrator building will be reactivated as necessary to house the Grinding and new Flotation areas, including the flotation reagents. A new building will be constructed to the immediate east of the Concentrator to house the POX autoclaves and hydrometallurgical refinery as shown in Figure 6. The existing General Shop building will be used to house most of the refinery reagents. The SX plant will be located in a new building to the south of the General Shop with the EW plant located in a new building adjacent to the SX plant.


18

FIGURE 6 Proposed Plant Layout

The following considerations were made in determining the final plant layout:

• Site relief contours and elevations across the plant areas were considered to maximise flow by gravity and minimise the requirement to pump solutions and slurries;

• The existing roadways and railways that can be reasonably used to import consumables and reagents and export products have been earmarked for refurbishment;

• Disturbing previously un-utilised land was minimised;

• Vent stack locations are cognisant of prevailing wind direction to minimise the potential for emissions impacting on neighbouring structures and offices;

• Due to low winter temperatures and for the comfort of operating and maintenance personnel, most equipment is housed inside buildings;


19

• The SX circuit is located separately from the main hydrometallurgical building due to the potential for fire. The EW tank-house is also housed separately from the SX circuit;

• The EW tank-house features access to sufficient lay-down area and truck access to allow for product handling, storage and loading. This are is located to the east of the building;

• Unit processes within the flotation area and hydrometallurgical building are grouped in a logical order and generally reflect the process flow. Inter and intra-connecting pipe length has not been optimised and will be studied in more detail during the engineering phase of the project;

• The AuPGM product area is located within the new hydrometallurgical building, though separate from the other facilities to provide higher level of security in this area. This is necessary to protect the high value product from theft;

• Allowances have been made for future expansion of some major equipment;

• The oxygen plant is located separately to the plant buildings for occupational, health and safety reasons, chiefly noise abatement;

• The leach residue, neutralisation and raffinate neutralisation belt filters are located together to allow the sharing of filter cake handling facilities;

• Orientation of tanks was optimised to maximise gravity discharge of slurry/solution thereby minimising the number of pump sets;

• Maintenance access was considered for all major equipment items;

• A mobile ten tonne gantry style crane services the entire hydrometallurgical refinery equipment with the exception of the POX autoclave flash vessels.

5.2 CORROSION COUPON TESTING PROGRAM

The Platsol™ process introduces challenges with respect to choosing appropriate materials of construction particularly within the autoclave and associated heat recovery areas. Bateman has consulted extensively with various equipment and materials suppliers regarding candidate metallic alloys that may withstand the hot, acidic and chloridic nature of hydrometallurgical solution typically produced in a Platsol™ process. A corrosion coupon testing program was undertaken as part of the fully integrated pilot plant campaign in 2006 using a selection of materials.

The objective of the test program was to examine the compatibility of various materials of construction with respect to the process conditions that existed in the various circuits of the plant. Corrosion testing coupons representing a total of 16 different alloys and metals of various grades were placed in different circuits of the pilot plant prior to commencement of operation. The coupons were kept in the circuits for the duration of all pilot plant runs. Unfortunately the fibre reinforced plastic (FRP) coupons were too big for placement in the small pilot tanks and thus were not tested.

The coupons were placed in process tanks with the help of polytetrafluoroethylene nuts, bolts and spacer rings. All coupons received were photographed and pre-tared with their weights recorded. At the end of the pilot plant operation, the coupons were recovered, washed with dilute nitric acid, rinsed with deionised water, dried, photographed and re-weighed. The difference in weight and photographic comparison provides some direction of whether the material should be short-listed for further investigation during the detailed engineering phase of the project.

Some of the key observations made from the work included the extremely poor performance of all coupons except for Ti in the autoclave head as well as the poor performance of 316 stainless and 317-L in almost all the circuits tested. Pitting corrosion was dominant which can be attributed to the presence of chlorides for most of the hydrometallurgical refinery.

The use of titanium alloys in applications such as oxygen injection tubes has generally been met with reserve due to the propensity for spontaneous ignition of freshly exposed titanium surfaces in the presence of oxygen. These surfaces may develop where gas velocity is high usually at the tip of the tube or where an agitator blade, for instance, has come free and collided with the tube. Materials are available to alleviate some potential for fire such as titanium – niobium alloys or ceramics, however ceramic materials are too brittle to withstand the intense agitation within the autoclave without extensive and expensive bracing.


20

Ceramic materials may find use as a fire break located near the autoclave oxygen injection nozzle. Although this does not prevent a fire from initiating, it may restrict the fire front from burning to the nozzle linings which may ultimately lead to autoclave damage. Oxygen may be injected using a small ceramic nozzle penetrating the autoclave from the bottom to directly below the compartment agitator.

6 CONCLUSION

Bateman Engineering has been involved with PolyMet’s NorthMet project from mid 2004 until the present day. Following successful completion of separate DFS and FEED packages, the project has moved into detailed engineering and construction will begin upon receipt of environmental permits. The Platsol™ process provides a technically viable alternative to smelting especially for small resource projects containing economical values in base metals and AuPGMs. Bateman has designed supervised and interpreted extensive bench-scale and fully integrated pilot campaigns between 2005 and 2007 in conjunction with PolyMet and SGS Lakefield, based on a Platsol™ concentrate leaching stage followed by sequential recovery of gold, platinum, palladium, copper, nickel, cobalt and zinc from solution into saleable products. A fully Hazoped plant design including reagent and utility areas has been delivered for detailed engineering and ultimately for construction.

7 ACKNOWLEDGMENTS

The authors would like to thank the management of Bateman Engineering Pty Ltd and PolyMet Mining Corp. for permission to publish this paper. Further thanks are extended to the individuals and companies that contribute to the successful completion of the DFS and FEED packages.

8 REFERENCES

1. Dreisinger, D., et al, 2006, “Metallurgical Processing of PolyMet Mining’s NorthMet Deposit for Recovery of Cu-Ni-Co-Zn-Pd-Pt-Au”, ALTA Nickel / Cobalt Conference, pp 11-18.

2. Ferron, C.J., et al, 2001, “Platsol™ Treatment of the NorthMet Copper-Nickel – PGM Bulk Concentrate – Pilot Plant Results”, ALTA Nickel / Cobalt Conference, pp 1-34.

3. Fleming, C., 1999, “Platsol™ Process Provides a Viable Alternative to Smelting”, Africa Mining Journal, pp1-4.

4. Hamilton, W.R., Wooley, A.R. and Bishop, A.C., 1981, “The Hamlyn Guide to Minerals, Rocks and Fossils”, Hamlyn Publishing Group pp 20, 76.

5. Hanks, J. and Barratt, D., 2002, “Sampling a Mineral Deposit for a Metallurgical Testing and Design of Comminution and Mineral Separation Processes”, SME Mineral Processing Plant Design, Practice and Control, pp 99-116.

6. Lamb, K. and Gulyas, J., 2002, “Selection of materials and Mechanical Design of Pressure Leaching Equipment”, SME Mineral Processing Plant Design, Practice and Control, pp 1511-1514.

7. Milbourne, J., Tomlinson, M. and Gormely, L., 2003, “Use of Hydrometallurgy in Direct Processing of Base Metal/PGM Concentrates”, Hydrometallurgy 2003, pp 625.


NICKEL AND COBALT RECOVERY FROM MESABA CONCENTRATE

By

K Mayhew, R Mean, L O’Connor and T Williams

CESL, Canada

Presented by

Rob Mean

[email protected]

CONTENTS

ABSTRACT 2

1. INTRODUCTION 2

2. STAGED CESL NICKEL PROCESS DEVELOPMENT 3

3. MESABA CESL REFINERY FLOW DIAGRAM AND PILOT PLANT RESULTS 8

4. CONCLUSIONS 16

5. REFERENCES 16


2

ABSTRACT

CESL has developed a hydrometallurgical process for treating Copper and Nickel sulphide concentrates, including bulk concentrates. Both Cu and Ni are leached efficiently into solution, using a proprietary pressure oxidation process, along with other base metals, notably Co. CESL developed the technology in 2001-02 using a low grade Cu-Ni bulk concentrate sample from an undeveloped property owned by Teck in Minnesota, USA called Mesaba. Ni and Co recovery from the leached solution was contemplated at the time, but not finalized. This paper will describe the latest work at CESL on Ni and Co recovery from a recent concentrate obtained from the same ore body. A discussion is also presented regarding the required impurity removal steps and quality of the final mixed hydroxide product.

1. INTRODUCTION

1.1 PROJECT BACKGOUND

The Mesaba Copper-Nickel Project of Teck Resources is one of several copper-nickel sulphide deposits within the Duluth Complex of Northern Minnesota. The deposit contains a geologic resource in excess of 1 billion tonnes at approximately 0.43% Cu and 0.09% Ni with minor Co and PGM values. This region consists of several large Cu-Ni deposits. Access and infrastructure (power, road, rail and the town of Babbitt) is excellent as the property is located immediately adjacent to Cliffs’ Northshore Iron Ore Mine.

Minnesota

100 km

Babbitt

1 km

Mesaba Property Access

Duluth

Northshore

Mine

Mesaba

Property

Figure 1: Location of Teck’s Mesaba Property

The mineralogy of the Mesaba ore does not allow for effective production of separate copper and nickel concentrates, however, a low grade bulk concentrate may be produced with reasonable copper and nickel recoveries. Due to the capability of the CESL Process to handle bulk concentrates, Teck Resources initiated studies to determine the amenability of the Mesaba bulk concentrate samples to hydrometallurgical testing.


3

1.2 PREVIOUS TESTING ON THE MESABA CONCENTRATE

In September 2001, CESL started performing pilot plant work on the bulk Mesaba concentrate. The results from this campaign have been documented previously [1] and will only be briefly reviewed. Considerable work went into developing the copper portion of the flow sheet, and by January 2002 it had been decided that Process 4 version of the CESL Process (leaching of all metals in the autoclave without the need for a second stage leach) provided the best results. While the leached copper was recovered effectively using traditional SX/EW technology, recovery of nickel and cobalt from the leach solution was examined, but the flow sheet was not finalized.

In 2002, the scope of the Mesaba project was redefined from making refined nickel and cobalt metal products, to the production of a nickel – cobalt intermediate hydroxide product via precipitation with lime and subsequent elutriation. The decision to produce an intermediate for Mesaba simplified the nickel flow sheet by keeping the solution in a sulphate matrix and thus did not require ammoniacal circuits. The flow sheet was operated at pilot scale from February through July 2002. The design basis for the pilot plant was 1.3 kg/d nickel to a mixed hydroxide intermediate.

Table 1 outlines the key results from the pilot testing program.

Table 1: Major Process Parameters from 2001-02 CESL Pilot Test Program [1]

Parameters Result Parameters Result

Copper Extraction 95.4% Gross Oxygen Ratio 0.21

Nickel Extraction 90.8% Autoclave Solids Density 235 g/L

Sulphur Oxidation 5.2% Autoclave Retention Time 60 min

In the fall of 2002, the Mesaba Project was delayed due to exceptionally low metal prices. Due to an emphasis on other projects, minimal nickel process development work at CESL occurred over the next five years.

2. STAGED CESL NICKEL PROCESS DEVELOPMENT

Further development of the CESL Nickel Process, specifically the purification and recovery stages, was initiated beginning in 2006. This was driven by both a renewed interest in the Mesaba property as an economically viable copper-nickel deposit, and also by attractive nickel prices which predominated at the time.

The development of the nickel purification and recovery flow sheet for the Mesaba concentrate was divided into six distinct stages, with each stage having its own objective. A brief description of each stage is presented in Table 2.


4

Mg Removal Product (To PO)

Sulphide PPT (Cu Removal)

Fe/Al Removal

Mg Removal

Zn Recovery

MHP w/

CaO

Ni Scavenging

Cu SX Raff

(From SX)

Table 2: Mesaba Nickel Flow Sheet Development

Project Stage Project Description Objective

1 Bench Scale Develop metallurgy & evaluate OPEX

2 Semi-Continuous Process Optimization

Confirm flow sheet, metallurgy, METSIM

3 Scoping Study Confirm economics

4 Marketing Study Determine intermediate payable metal value

5 Pilot Decision Pilot construction

6 Pilot Scale Mass balance. prove metallurgy, produce marketing & environmental samples

In designing the nickel recovery circuits that comprise the Mesaba flow sheet, a review of the previous 2001-2002 Mesaba pilot plant campaign was performed to see if opportunities existed for optimizations and improvements. During the 2001-2002 campaign, a bleed of raffinate from Cu SX fed the nickel plant. The simplified nickel plant consisted of the following circuits as illustrated below:

This flow sheet produced a low-grade mixed Ni/Co hydroxide which contained elevated calcium tenors. When flow sheet development restarted in 2006, this intermediate nickel product was not anticipated to be a saleable product due to the impurity levels and low nickel grade.

With aggressive worldwide development of nickel laterite properties around the world early in the twenty-first century, many references became available on process development for nickel properties that faced similar challenges for impurity removal from acid leach streams and the production of saleable nickel products. Of the hydrometallurgical plants, most flow sheets produced a solid nickel intermediate prior to metallic nickel production. There was no particularly favored intermediate amongst the flow sheets investigated—mixed Ni/Co sulphides, mixed hydroxides, and mixed carbonates were all seen. Each intermediate was suited to a particular refinery flow sheet. Sulphides are suitable for processing in smelters and hydrometallurgical refineries and also minimize co-precipitation of certain impurities such as iron, aluminum, and manganese. Hydroxides and carbonates are most suited to hydrometallurgical refineries which could involve either ammoniacal or acid leaching to resolubilize the nickel and cobalt. Subsequent investigations indicated that some pyrometallurgical facilities could also accept mixed hydroxides as a feed, in part because of recent modifications to equipment and processes in anticipation of an increasing world supply of nickel intermediate products from new mine sites.

The industrial production and marketing trends assisted in the definition of a nickel bleed flow sheet that was developed from bench evaluations, progressing to small scale semi-continuous mini-pilot operations, and culminating in a fully continuous pilot campaign.

2.1 BENCH TEST WORK – NICKEL BLEED CIRCUIT EVALUATIONS

Bench work was commenced to evaluate possible unit operations within a flow sheet that would produce a saleable solid nickel intermediate as a product. This flow sheet was designed for leach

Figure 2: 2001-02 Mesaba Nickel Recovery Circuits


5

solution from Mesaba concentrate (specifically a bleed of raffinate from copper solvent extraction) which was expected to contain 20-40 g/L nickel with significant levels of iron, aluminum, zinc, copper, and cadmium as primary impurities. Synthetic Mesaba leach solution (based on the 2002 pilot campaign data) was prepared and used for bench test work which evaluated the performance of the following impurity removal circuits:

• Iron/Aluminum Removal

• Zinc/Copper/Cadmium Removal

• Nickel/Cobalt Product Precipitation

The impurity removal circuits were required to remove impurities to a level that the resulting mixed nickel/cobalt product would be saleable with minimal treatment penalties. Other requirements of the impurity removal circuits were to achieve acceptably low nickel losses from the plant, minimize nickel recycles within the plant and minimize operating and capital costs.

The product precipitation stage needed to be able to selectively precipitate nickel and cobalt in the presence of impurities such as calcium, magnesium, sodium and manganese. Other challenges in the production of the mixed nickel product were selection of an intermediate product with mixed hydroxides, mixed sulphides and mixed carbonates as the primary options.

Key Findings:

• Iron and Aluminum Removal should be performed in two stages to minimize the re-circulating load of nickel. Co-precipitation of nickel with Fe/Al leads to high nickel losses if performed in one stage at pH 5.0 and 40°C. Nickel loss in the second stage at pH 5.0 is dependent on the Al concentration in the feed to that stage because of co-precipitation. At the pH of the second stage, nickel precipitation was found to correlate directly with the aluminum concentration in the feed solution. This effect was illustrated with aluminum spiking tests during bench test work (see Figure 3).

0

2

4

6

8

10

12

14

16

0 500 1000 1500 2000 2500

Ni P

recip

itate

d (%

)

Al in Feed (ppm)

Figure 3: Effect of Al in Feed Solution on Ni Precipitation in Al Removal

• Sulphide Precipitation could selectively precipitate copper, cadmium and zinc to below 1 mg/L with less than 0.2% Ni co-precipitation.

• An evaluation was performed to determine whether to produce a mixed Ni-Co hydroxide or sulphide at Mesaba. It was decided that a mixed hydroxide would be produced as it would result in lower up-front risk (startup costs and safety) and is suitable for the leach process liquor (i.e. low Mn), but came with a greater back-end risk (marketing). The decision was, in part, influenced by the potential for future on-site refining using the CESL Nickel Process.


6

• Several types of magnesia were tested for mixed hydroxide precipitation with varying degrees of reactivity and contained impurities. From a leach liquor containing 20 g/L Ni, a magnesia utilization of 93% could be obtained while precipitating 95% of the nickel.

• Manganese removal prior to MHP was a technical success using a variety of process routes. However, an economic review of the process’s feasibility excluded it from the final process flow sheet.

2.2. MESABA SEMI-CONTINUOUS TESTING

Based on the 2007 bench results, a semi-continuous nickel plant was constructed and operated in Q4-2007 through Q1-2008. The objective of the mini-pilot plant was to have a defined flow sheet for the Mesaba pilot plant operations scheduled for the fall of 2008. Process solution generated from the Mesaba concentrate confirmed a conceptual nickel flow sheet for Mesaba, which consisted of the following unit operations: Iron and Aluminum Removal, Sulphide Precipitation, Mixed Hydroxide Precipitation (with magnesia), Nickel Scavenging, and Magnesium Precipitation.

Key Findings:

• Limestone precipitation stages for iron and aluminum removal would precede the sulphide precipitation stage as there was a clear correlation between copper concentration in the feed solution and subsequent iron precipitation (see Figure 4).

0

10

20

30

40

50

60

70

80

90

100

0 100 200 300 400 500 600

Fe in

Pro

duct S

olu

tio

n (p

pm

)

Cu in Feed (ppm)

Figure 4: Effect of Copper Tenor in Feed Solution on Iron Precipitation

• Pressure filtration was a critical parameter for the successful operation of sulphide precipitation. Negative pressure filtration (i.e. pan filters) passed air over the filters resulting in unwanted 1-2 mg/L Zn re-dissolution.

• Three different MHP flow sheet configurations were tested during semi-continuous testing. The flow sheet variations were implemented with the objective of maximizing magnesia utilization and MHP nickel grade, while attempting to minimize manganese precipitation. A basic illustration of the three flow sheets is provided in Figure 5. Flow sheet C was chosen for use in the pilot plant campaign as it produced a product at the required specifications at reduced capital.


7

Cu SX Raff (From SX)

Mg Removal Product (To PO)

Fe/Al Removal

Mg Removal

MHP w/ MgO

Ni Scavenging

Sulphide PPT (Cu, Zn Removal)

FLOWSHEET C FLOWSHEET A FLOWSHEET B

MHP Stage 1

S L

Sulphide Precipitation Product

Ni(OH)2

Co(OH)

MgO

S

L

MHP Stage 2

MgO

MHP Product

MHP Stage 1

S L


Ni(OH)2

Co(OH)

MgO

MHP Product

S

L

MHP Stage 2

MHP

S L


Ni(OH)2

Co(OH)

MgO

MHP Product

Figure 5: MHP Flow Sheets Tested

2.3 MESABA INTEGRATED PILOT PLANT CAMPAIGN

Based on the success of the semi-continuous testing, an integrated Mesaba Copper-Nickel pilot plant operated for a three month period. The major objectives of the operations were to:

• Confirm and improve upon process metallurgy;

• Confirm flexibility of the CESL flow sheet by processing Mesaba bulk concentrates at varying mineralogical composition and grade;

• Operate plant in a fully integrated manner ensuring all recycles were continuous;

• Produce Ni-Co intermediate samples for potential customers;

• Collect environmental samples for long-term residue stability testing; and

• Optimize and collect solid-liquid separation engineering data.

The simplified nickel pilot plant consisted of the following circuits as illustrated below:

Figure 6: 2007-2008 Mesaba Nickel Recovery Circuits

The process flow sheet and key findings from the 2008-09 Mesaba pilot plant campaign are discussed in the following section.


8

3. MESABA CESL REFINERY FLOW DIAGRAM AND PILOT PLANT RESULTS

The final integrated Mesaba flow sheet is presented in Figure 7. The plant processes a bulk concentrate through to copper cathode and a nickel-cobalt hydroxide intermediate.

PO

Residue

Washing

3°°°° SX

4°°°° SX

S L

3°°°° Neutralization

Cu EW

Fe/Al Removal

Sulphide PPT

MHP

Ni Scavenging

Mg Removal

Concentrate

Leach Residue

(To Tails)

Fe Solids (To Tails)

Al Slurry (To 3° Neut)

Zinc Product

(To Market)

Mixed Hydroxide Precipitate

(To Market)

Ni Slurry

(To 3° Neut)

Mg Slurry

(To Tails)

Copper Cathode

(To Market)

80%

20%

Gypsum

(To Tails)

Figure 7: Simplified Mesaba CESL Process Flow Diagram


9

The following section provides a brief description of the results from the 2008-09 CESL Mesaba pilot plant campaign.

3.1 FEED MATERIAL

A sample of Mesaba bulk Cu-Ni-Co ore was gathered from the Mesaba property in the fall of 2008. From this sample, nine tonnes of concentrate were produced at the Coleraine Minerals Research Laboratories in Coleraine, Minnesota. Various grades of concentrates were produced; with the following table presenting the concentrate feed composition from three distinct periods of operation.

Table 3: Concentrate Composition of Bulk Concentrate (2008-09 Mesaba Campaign)

Period Cu (%) Ni (%) Co (%) Fe (%) S (%)

A 13.9 1.64 0.08 26.7 21.3

B 19.0 2.30 0.10 30.6 26.2

C 21.7 2.36 0.10 31.6 28.7

The mineralogy of the 2008 concentrate sample was considerably different to that processed at CESL in 2001, with an average chalcopyrite to cubanite ratio of 1.8. The change in mineralogical composition was due to a different sampling location of the bulk ore sample, which enabled further examination of the leaching characteristics of concentrate produced from a different area of the deposit. The following table presents the mineralogical composition of the concentrate processed in 2001 and the high-grade concentrate processed during period C of the 2008-09 campaign.

Table 4: Mesaba Concentrate Mineralogy

Concentrate Chalcopyrite

(%)

Cubanite

(%)

Pentlandite

(%)

Bornite

(%)

Pyrrhotite

(%)

Gangue

(%)

2001 10 45 5 Trace 5 35

2008 50 28 7 0.30 2 12

3.2 PRESSURE OXIDATION

One of the key unit operations in the CESL Process is Pressure Oxidation (PO) in an autoclave, where the copper and nickel sulphide minerals are oxidized to form soluble forms of copper and nickel. During Pressure Oxidation the primary objective is to maximize metal extraction while minimizing sulphur oxidation.

The copper in the Mesaba concentrate is mainly contained in chalcopyrite and cubanite. The oxidation reactions of these minerals are shown below:

12 CuFeS2 + 15 O2 + 12 H2SO4→ 12 CuSO4 + 6 Fe2O3 + 24 S° + 12 H2O

CuFe2S3 + 2 O2 + H2SO4→ CuSO4 + Fe2O3 + 3 S° + H2O

The nickel in the Mesaba concentrate is mainly contained in pentlandite, with the oxidation reaction of pentlandite shown below:

8 (Ni,Fe)9S8 + 45 O2 + 36 H2SO4 → 36 NiSO4 + 18 Fe2O3 + 36 H2O + 64 S°

Pyrrhotite in the concentrate is oxidized almost quantitatively to hematite and elemental sulphur, without any sulphate formation. This reaction is very fast, has only a minor effect on the overall processing


10

cost, and generates an ideal “gangue” material or dilution mineral for the concentrate. The oxidation of pyrrhotite is shown below (for simplicity, it is assumed that the iron to sulphur ratio is one):

4 FeS + 3 O2 → 2 Fe

2O

3 + 4 S°

In addition to the sulphide minerals, 5-10% of the elemental sulphur oxidizes:

2 S° + 2 H2O + 3 O2 → 2 H2SO4

The entire plant process liquor inventory has a chloride concentration of approximately 8-12 g/L. The chloride increases reaction kinetics in the autoclave and works as a catalyst to ensure complete oxidation of the sulphide minerals, whilst minimizing sulphate formation.

There are two main feed streams to the Pressure Oxidation autoclave: acid feed and concentrate slurry. The acid feed solution, which is made up of Mg Removal product and recycled acid in the copper SX raffinate, is stored in the acid feed tank. The concentrate slurry is fed to the autoclave from the PO slurry feed tank.

After a retention time in the autoclave of 60 minutes, at 1,380 kPag and 150 °C, the slurry is depressurized in one flash stage. The letdown brings the contents to atmospheric pressure and a temperature of 95-100 °C.

Oxygen

Pressure Oxidation

Thickener Underflow

(To Wash Circuit)

Acid Feed

(Cu SX Raff + Mg Removal Product)

Concentrate

Slurry Feed

3°PLS

(To SX)

Flash

Steam

(To Vent)

Thickener

Figure 8: Mesaba Pressure Oxidation Flow Sheet

Operating conditions in the Pressure Oxidation circuit were based on results from the 2001-02 Mesaba Campaign.

Metal extraction

To determine if there was an opportunity to increase nickel extraction from the Mesaba concentrate, bench leach tests were performed that evaluated several operating parameters. It was determined that the acid feed Ni tenor impacts overall nickel extraction; where an increased re-circulating load of nickel suppressed overall extraction. This effect is illustrated in Figure 9, where the effect of nickel concentration in the PO acid feed is graphed versus the nickel content of the residue for a concentrate containing less than 100 ppm Ni. An increase in the acid feed nickel tenor from 10 g/L to 40 g/L increased the residue nickel content from 0.04% to 0.12%. This suggests that a higher concentration re-circulated to the autoclave may lead to increased nickel losses due to adsorption into the residue. Based on this data and a sulphate balance generated from the Mesaba Metsim model, the design for the Mesaba pilot plant was 24 g/L nickel in the acid feed liquor.


11

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0 10 20 30 40 50

Ni C

onentr

atio

n

in P

O R

esid

ue (

%)

Ni in PO Acid Feed (g/L)

Figure 9: Effect of [Ni] in Acid Feed Solution on Ni in Residue

The following table presents the key metallurgical results from the Mesaba pilot plant campaign. The results confirm the bench findings that decreasing the re-circulating load of nickel would increase nickel extraction.

Table 5: Key Metallurgical Results from 2008-09 CESL Pilot Test Program

Parameters Result

Copper Extraction 95%

Nickel Extraction 95%

Cobalt Extraction 97%

Sulphur Oxidation 6.9%

Sulphur oxidation is defined as the percentage of sulphur in the concentrate that is oxidized to sulphate. The amount of sulphur oxidized affects numerous aspects of the flow sheet, and low sulphur oxidization is usually desirable. Most importantly, the sulphur oxidation has a large impact on the autoclave heat balance, and subsequently the autoclave size and all associated equipment. In addition to the effect on capital, the amount of sulphur oxidation affects the overall autoclave oxygen consumption, as well as the amount of acid and soluble iron that must be neutralized with limestone.

3.3 COPPER SX/EW

The Solvent Extraction (SX) circuit is the purification step in the CESL copper flow sheet. The SX circuit enables the production of high purity cathode copper in electrowinning by selective ion transfer of copper from a pregnant leach solution (PLS) to a copper electrolyte. Numerous commercial applications of this process exist, and the technology is well proven.

Mesaba operations require two extraction circuits, 3° and 4°. The purpose of each of the extraction steps is as follows:

• The 3° SX circuit extracts most of the copper leached in the PO circuit. Of the resultant raffinate, 80% is recycled to the autoclave and 20% is sent to 3° Neutralization.

• The 4° SX circuit processes the solution that feeds the nickel purification and recovery circuits. The circuit targets a low raffinate copper tenor to minimize copper losses.


12

During operations, the 4° SX partially loaded organic (pLO) was fed to the 3° SX circuit to fully load the organic stream. Any iron loaded on the partially loaded organic stream was crowded off by fully loading the organic with copper. A simplified schematic of the SX flow sheet is presented in Figure 10.

4°SX E3/E4

SO

Solvent

Strip S1/S2

pLO

SO

LO

3° Raff

(To PO & 3°Neut )

3°KPT3°SX E1/E2

3° PLS

(From PO)

Cu: 46 g/L

FA: 1.5 g/L

Cl: 8.9 g/L

3° Raff

(To PO & 3°Neut )

Cu: 8.9 g/L

FA: 57 g/L

Cl: 8.7 g/L

4° PLS

(From 3°Neut)

Cu: 9.2 g/L

FA: 2.6 g/L

Cl: 8.2 g/L

4° Raff

(To Ni Plant)

Cu: 0.46 g/L

FA: 15 g/L

Ni: 23 g/L

SE

Cu: 37 g/L

FA: 182 g/L

PE

Cu: 49 g/L

FA: 164 g/L

Electrowinning

LME Grade A

Copper Cathode

Figure 10: Mesaba Copper SX and EW Flow Sheet

During the campaign, 1245 kg of copper cathode were produced at a current efficiency of 96.8%.

3.4 NEUTRALIZATION

The Neutralization circuit processed 20% of the 3° raffinate stream prior to advancing to the nickel and cobalt purification and recovery circuits. The purposes of this stage are to remove sulphuric acid (H2SO4) from solution with limestone and to re-leach nickel and cobalt metal hydroxides recycled from the Nickel Scavenging and Aluminum Removal circuits. The composition of the gypsum produced in 3° Neutralization is shown in Table 6 below.

Table 6: 3° Neutralization Solids Composition

Cu (%) Ni (%) Fe (%) Al (%) Ca (%) S (%)

<0.02 <0.02 0.09 0.03 22.3 17.3

The gypsum produced was virtually free of contaminants, indicating that all elements within the Nickel Scavenging and Aluminum Removal recycles were re-leached into solution. These recycles, along with a 2 hour retention time and the internal recycle of thickener underflow, resulted in 99.2% limestone efficiency.

The addition of Nickel Scavenging and Aluminum Removal recycles to the 3° Neutralization provides a mechanism whereby potential losses of valuable metals are recovered through re-leaching. Table 7 shows the recirculating load of various metals within the nickel plant.


13

Table 7: Re-circulating Loadsr of Various Metals from Nickel Clean Up and Aluminum Removal

Recycles

Ni Co Al Fe Mg Mn Si Zn

1.07 1.03 2.34 2.28 1.17 1.15 1.54 1.10

r mass in product solution / mass in initial solution

From the ratios, it can be seen that there is twice as much iron and aluminum in the 3° Neutralization product stream as a result of these metals re-leaching. Nickel, cobalt and zinc also re-leach, representing a recovery of valuable metals.

Fe and Al REMOVAL

Iron and aluminum are removed from solution to less than 5 mg/L in a two-stage precipitation circuit. The impurities are removed to minimize contamination of the MHP product and to minimize magnesia consumption.

Each stage has a 2-hour residence time and operates at 40°C. The pH is controlled using limestone powder to a pH of 3.7 in stage 1 and 5.0 in stage 2. The residue solids from the first stage contain the bulk of the iron and aluminum with minimal nickel and cobalt values, and are a waste stream from the process.

A stepwise improvement in the stage 1 solids settling and filtration rates was seen at elevated recycle ratios. As shown in Figure 11, with minimal solids recycle, gypsum particles settled first followed by iron hydroxide particles producing two layers of solids at the bottom of the settling cylinder. With an elevated solids recycle, a homogeneous sludge was formed as the iron hydroxide seeded onto the gypsum crystals.

Figure 11: Left – gypsum and metal hydroxide layers; right – homogenous sludge

The second stage of Fe/Al Removal can tolerate some nickel loss for the sake of MHP product quality as the residue solids are recycled to 3° Neutralization. The recirculating load of nickel, cobalt and copper in the thickener U/F is included in Table 8 as determined by the mass balance. Note that recycling the copper improves overall copper recovery within the plant as the 3° Neutralization product feeds 4° SX.

Sludge

Gypsum

Metal Hydroxide


14

Table 8: Recirculating Load of Metals from Fe/Al Removal Circuit Stage 2

Metals Deportment Metal Precipitated

Nickel 7%

Cobalt 3%

Copper 79%

3.5 SULPHIDE PRECIPITATION

Zinc, copper, and residual cadmium are precipitated selectively with hydrogen sulphide gas prior to Mixed Hydroxide Precipitation. The circuit effectively removed copper, cadmium and zinc to target levels of <1ppm, while averaging 0.15% nickel co-precipitation.

The solids in Sulphide Precipitation had a typical zinc grade in excess of 40% (see Table 9). Preliminary discussions indicate that these solids may be marketable to a third party.

Table 9: Sulphide Precipitation Solid Composition

Ca (%) Cd (%) Co (%) Cu (%) Ni (%) S (%) Zn (%)

0.72 0.53 0.58 10.0 1.4 33.0 43.4

3.6 MIXED HYDROXIDE PRECIPITATION

Nickel and cobalt are recovered from the process solution through the addition of calcined magnesia. The MHP circuit consisted of five cascading reactors that operated at 50°C.

Magnesia addition to the circuit is primarily based on stoichiometric addition, with secondary process control being aqueous nickel assays and the pH of reactor five. As shown in Figure 12, at varying magnesia dosages, the pH in reactor 5 correlates well with the product solution nickel tenor. The correlation between the product nickel tenor and reactor 1 pH indicates that higher concentrations of nickel left in the product stream are predictable through pH control, but becomes difficult when the nickel tenor is less than 1 g/L.


15

0

500

1000

1500

2000

2500

3000

3500

6.8 7 7.2 7.4 7.6 7.8

Ni

in P

rod

uc

t S

olu

tio

n (

pp

m)

pH

Reactor 1

Reactor 5

Figure 12: pH in Reactor 1 & 5 vs. Product Solution Nickel Concentration

Figure 13 presents the MHP circuit reactor profiles at a 4.1 hour retention time. Precipitation of nickel is rapid with most of the precipitation occurring in reactor 1 and slowing between reactors 2 and 5; manganese precipitation was relatively linear with time. Overall magnesia utilization through the circuit is 96%.

0

10

20

30

40

50

60

70

80

90

100

1 2 3 4 5

Extr

actio

n(%

)

Reactor

%Ni

%Co

%Mn

Figure 13: MHP Reactor Profile at a 4.1 h Retention Time

The composition of the MHP solids produced during the campaign is shown in Table 10. The high-quality MHP product will be sold to a third party refinery. Note that arsenic, cadmium, and chromium tenors were all less than the detection limit.


16

Table 10: Average MHP Composition from Mesaba Pilot Plant Campaign

Ni (%) Co (%) Mg (%) Ca (%) Cu (%) Fe (%) Mn (%) Zn (%)

46.0 2.0 0.8 0.6 0.02 0.10 0.68 0.02

Operating the MHP circuit with a solid recycle ratio of 200% increased magnesia utilization and nickel grade, as well as improved the settling and filtration characteristics of the solids.

3.7 NICKEL SCAVENGING

In order to minimize impurities such as manganese and magnesium in the MHP intermediate, a residual tenor of nickel and cobalt was left in the MHP product solution. The Nickel Scavenging stage precipitates these metals with hydrated lime and recycled them back to 3° Neutralization. Overall recovery of nickel and cobalt within this circuit was greater than 97%.

3.8 MAGNESIUM REMOVAL

Magnesium is both leached from the concentrate in Pressure Oxidation and added via magnesia in the Mixed Hydroxide Precipitation circuit. The purpose of the Magnesium Removal circuit is to bleed magnesium from the process yet leave enough magnesium to support the chloride to minimize the formation of calcium chloride complexes.

Hydrated lime is used to precipitate magnesium, with the product solid sent to tailings for disposal. In the pilot plant, 65% of the magnesium was precipitated per pass with the remainder recycled to PO.

4. CONCLUSIONS

The application of the CESL Process to the Mesaba property is being examined due to difficulties in the production of saleable concentrate(s). Extensive pilot plant testing has shown extractions in excess of 95% for both nickel and cobalt, with overall nickel recovery of 98% from the leach liquor. The low autoclave sulphur oxidation allows for economic recovery of metals from a bulk concentrate.

During the pilot campaign, a saleable mixed nickel/cobalt hydroxide intermediate product was produced that exceeded the grades modeled from bench testing. Samples of the intermediate were sent out for marketing purposes to various third party refiners.

A suite of Mesaba residues were collected throughout the fully integrated pilot plant campaign for residue stability testing.

With the campaign complete, CESL continues to support the advancement of the Mesaba Project. A scoping study is underway and is scheduled for completion later in 2009.

5. REFERENCES

1. D.L. Jones and R. Moore, CESL, “CESL Process: Application to a Bulk Copper-Nickel Concentrate”, Proceedings of ALTA 2002 Conference, ALTA Metallurgical Services, Perth, Australia, May 2002.


X-RAY SORTING AND OTHER TECHNIQUES FOR UPGRADING

NICKEL ORES

By

A Allen, H Gordon

UltraSort Pty Ltd, Australia

Presented by

Allison Allen

[email protected]

CONTENTS

1. INTRODUCTION 2

2. SORTING PRINCIPLES 2

3. ELECTROMAGNETIC SORTERS 4

4. NEW X-RAY SORTERS 6

5. COMPARISON BETWEEN EM AND X-RAY TECHNOLOGY 7

6. CURRENT NICKEL SORTING MACHINE THROUGHPUT 7

7. CONCLUSION 9

8. REFERENCES 9


2

1. INTRODUCTION

Ore Sorting is used to upgrade nickel ores in the mining industry. Electromagnetic sensors can identify the magnetic properties of nickel. The sorter uses this information to separate the non nickel bearing particles from the nickel bearing particles. There have been a number of technological advancements in ore sorting technology which have some positive repercussions in upgrading nickel ore.

Recent advancements in electromagnetic sensors mean that more finely disseminated ores can now be identified. Current nickel sorters mostly sort on the conductive and magnetic properties of nickel, but there is an alternative option for sorting nickel ores based on the atomic density of the material. This is achieved by using a dual energy x-ray sorter.

UltraSort and Commodas have been active in sensor-based sorting since 1993. In 2008, UltraSort joined the Titech Group which includes Commodas, and is wholly owned by Tomra. UltraSort is a strong strategic fit with Commodas (TiTech's unit within the mining segment), and the two units combined form the sole market leader in the segment. The Tomra/Titech Group is a leading supplier of both recycling and mining sorting machinery around the world.

2. SORTING PRINCIPLES

2.1 OPERATION

The material is fed into a feeder bin at the top of the machine. This then passes onto a vibrating feeder, where it can be sprayed with water and screened to remove water and fines generated in the hopper.

The material falls from this feeder into the machine. At this stage, the material is spread so as to be only one rock deep. The material is then transported to the scanning and separating area. The scanning system can consist of a number of different sensors that detect the properties of the rock.

Based on the information from the sensors the processor can make a decision whether to accept or reject the material. This decision will cause an ejector to open and a blast of air to change the trajectory of the rock. The rationale behind whether the good material is blasted from the waste material, or vice versa depends on the ore/waste concentration; usually the segment with the smaller amount is blasted to reduce air consumption.

Figure 1: Particle Sorting

2.2 PARTICLE PRESENTATION AND SEPARATION

Particle presentation is the way in which the material is delivered to the sensory equipment. To be able to correctly analyse each rock the material must be presented in a stable single layer within the correct particle size range and in some cases washed so that dust and other particles do not interfere with the signal. To achieve this, a specially developed materials handling system is used. There are two types of systems commonly used:

- The chute or free-fall principle

- The conveyor-belt machine

Ejected Rock

Normal Trajectory


3

Particle separation involves removing the ore material from the waste material. Separation systems that have been implemented include air blasting, a gate system, a pusher system, and a finger system.

2.2.1 Freefall Equipment

The free fall or chute equipment has a very simple feeding mechanism without any rotating componets. It is a compact and reliable feed system and is generally cheaper than a belt machine. It also allows for a simple double side scanning layout and has efficient separation using valves and nozzles close to the feed.

Figure 2: Free Fall Materials Handling

2.2.2 Belt Based Equipment

The conveyor-belt based machine allows for a constant feed of heterogenous feeds with better efficiency for sensing techniques such as electromagnetic sensors. It is also a rugged design able to handle high tonnage and large particles. Multiple sensory devices can also be combined in the one machine.

The separation system can be via valves and nozzles underneath or above the material.

Figure 3: Belt Materials Handling System with Upward Ejectors

Feed

Dual Scanners

Scanning System

Ejectors


4

Figure 4: Belt Materials Handling System with Downward Ejectors

3 ELECTROMAGNETIC SORTERS

The electromagnetic sensors detect the conductivity and magnetic permeability of a particle. Depending on the ore, either or both of these properties may be analysed to allow for waste removal.

There are two properties of materials that the sensor needs to detect, these are:

• Magnetic permeability - when a rock with permeability greater than air passes through the field it changes the inductance.

• Conductivity - when a conductive object is placed in the field generated by the sensor, a current flows in the object, which produces its own magnetic field out of phase to the primary field.

Figure 5: Detector Operation

UltraSort’s electromagnetic ore sorters have been used in the nickel mining industry since 2006 (Allen, 2008). The sorters have successfully upgraded low grade nickel feed (less than 0.5%) to over 4% nickel (Goode, 2006).

Conveyor

belt

Coil n n+1 Coil n+2

Scanning System

Ejectors


5

3.1 EM SENSOR RESEARCH FOR FINELY DISSEMINATED ORES

Research conducted at the Department of Mineral Processing (AMR) at RWTH Aachen University in Germany has investigated the response of the available sensors to detecting finely disseminated ores (Wortruba et al 2006). A set of artificial sulphide ore samples were produced from a nickel sulphide ore containing pentlandite, pyrrhotite and pyrite with a nickel grade of 5% in massive sulphide particles. The massive sulphide rocks were crushed and milled and then added to a plaster/quartz matrix in discrete size distributions to form tablets with a diameter of 42mm and thickness of 20mm.

Although not a real ore sample (where particle size and distributions are much more irregular), the artificial samples provide a way of investigating the correlation between sulphide particle size and detector response.

Six samples with different sulphide particle sizes were produced for six different nickel grades. The average sulphide particle size was from 0.08 to 1.5mm and grade from 0.1 to 3.2% nickel

The sensor was mounted below a flat conveyor belt which was set to a speed of 2m/s. Each sample was run via the belt over the detector coil and the signal produced was recorded.

The following table lists all samples with their respective mixed-in nickel grades and measured detector sensor readings. There is a strong correlation between the detector response and the sulphide particle size. Samples with an average sulphide size of 1.5mm produced a signal at all grades. The smallest sulphide particle size was only measurable from 0.8% nickel.

Ni Grade [%] EM Response

Avg sulphide particle size

[mm]

0.10% 0.20% 0.40% 0.80% 1.60% 3.20%

0.08 0 0 0 40 80 200

0.18 0 0 40 100 160 480

0.38 0 0 40 100 220 530

0.63 0 0 40 120 300 780

0.88 0 60 80 160 320 800

1.5 80 80 140 260 400 1040

Table 1: Artificial Nickel Sample EM Response

Detector Response (by sulphide particle size)

0

200

400

600

800

1000

1200

0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 3.50%

Ni Grade [%]

Dete

cto

r R

esp

on

se 0.08

0.18

0.38

0.63

0.88

1.5

Figure 6: EM Detector Response (by sulphide particle size)


6

4 NEW X-RAY SORTERS

The same technology used in x-ray baggage scanning equipment and in recycling sorters is now being applied to sorting in the mining industry. X-ray transmission equipment allows material to be sorted based on their atomic densities. This technique means that the finely disseminated ores which current electromagnetic sensors may not identify can now be successfully sorted.

While travelling along a belt, the material passes under a broad band x-ray source. The x-ray sensor system below the belt produces a digital image of the material using two different energy bands. The x-ray attenuation through the material is different in the two bands and can be used to determine the density without being affected by the material thickness.

This technology is already in use in recycling in Europe where it is provien in sorting materials such as ceramics, demolition rubble, domestic waste, tv screens etc. In the mining industry, feasibility studies at Mintek have shown success in sorting run-of-mine coal and torbanite with recoveries of up to 98% pure coal (MRA 2006).

The figure below shows an example of a base metal sample with vairable metal content ranging from low to high grade. The image on the right displays the color-encoded classification of the pixels based on densities. The classification is based on a reference density (atomic number) to which the system has been calibrated.

Figure 7: Dual Energy X-Ray System Processing

The sensor’s high spatial resolution of 0.8mm or 1.6mm also permits the evaluation of a particle’s shape and size. Testing of finely disseminated nickel ores using the tablets described above has shown a better discrimination in the lower nickel grades and sulphide particle size than the EM testing.

Ni Grade [%] X-Ray Response

Avg sulphide

Particle Size [mm]

0.10% 0.20% 0.40% 0.80% 1.60% 3.20%

0.08 137 138 129 123 96 96

0.18 136 139 127 119 104 69

0.38 136 133 127 115 97 65

0.63 137 135 124 115 91 62

0.88 133 133 125 114 92 61

1.5 132 130 123 114 86 50

XRT Image Processing

Low Channel

High Channel

Broad Band X-Ray Tube

DE-XRT Sensor

Classified Image

Z > Z ref

Z ~ Z ref

Z < Z ref

Z: Atomic Number


7

Table 2: Artificial Nickel Sample Dual Energy X-Ray Response

X-Ray Response (by sulphide particle size)

0

20

40

60

80

100

120

140

160

0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 3.50%

Ni Grade [%]

X-R

ay R

esp

on

se 0.08

0.18

0.38

0.63

0.88

1.5

Figure 8: X-Ray Response (by sulphide particle size)

The x-ray sensor is able to get a measurement on all the artificial samples but no discrimination can be made between 0.1 and 0.2% nickel. A cut can be made at 0.4% or higher irrespective of the sulphide particle size. The x-ray technique is less reliant on the particle size and more directly linked to the overall nickel grade than the EM sensor for this sample.

5 COMPARISON BETWEEN EM AND X-RAY TECHNOLOGY

Even though the X-Ray detection system is more sensitive than the current EM system, there is still a case for using EM in lieu of an x-ray system depending on the sorting context.

The current EM technology is still improving with new more sensitive sensors being developed. Electromagnetic sorting of nickel ores is successfully proven in the mining industry and is a cost effective solution. The EM system does not require the shielding and health and safety controls that the x-ray system requires. It also is a cheaper sensory device.

For very disseminated samples which cannot be detected using electromagnetic techniques, a dual energy x-ray system can be used. The X-ray system is a more complex solution to the electromagnetic system and may also have difficulty with larger particle size and waste materials high in other metallic elements.

Dual energy x-ray sensing may be able to upgrade ore bodies that have not responded to electromagnetic testing, such as nickel laterite deposits.

6 CURRENT NICKEL SORTING MACHINE THROUGHPUT

6.1 THROUGHPUT

One of the main considerations when investigating sorting options is the required throughput of the plant and the number of sorting machines needed. The current range on UltraSort/Commodas machines can sort a range of size fractions at a range of throughputs.


8

Figure 9: UltraSort/Commodas ROM Secondary Sorter on Site

The table below shows the specifications for the UltraSort/Commodas ROM Sorters. These machines are specifically designed for run of mine applications where reliability, robustness, high feed rates and constant operation are required. The machines can operate at feed rates of up to 300 tonnes per hour. For an optimum sort, the material is usually prescreened into a 2 to 1 size fraction. The sensory options include a combination electromagnetic and optical systems or dual energy x-ray scanning.

Type Material

Handling

Sensor Resolution Width

[mm]

Size

range

[mm]

Capacity

[tph]

ROM Primary EM Belt EM/PMT 20/0,8 1200 40..300 Up to 300

ROM Secondary EM Belt EM/PMT 20/0,8 800 5…80 Up to 80

ROM Secondary XRT Belt DE-XRT 0,8 800 10..65 Up to 80

Table 3: UltraSort/Commodas ROM Sorters for Nickel Applications

The UltraSort/Commodas PRO sorters are simple and efficient freefall sorters with dual energy x-ray sensing. There a two models suited to nickel ore sorting – the PRO Secondary which sorts material from 20 to 80mm and the PRO Tertiary which sorts material from 15 to 40mm.

Type Material

Handling

Sensor Resolution Width

[mm]

Size

range

[mm]

Capacity

[tph]

PRO Secondary XRT Chute Dual Energy XRT 0,8/1,6mm 1100 20..80 40..70

PRO Tertiary XRT Chute Dual Energy XRT 0,8/1,6mm 1100 15..40 20..40

Table 4: UltraSort/Commodas PRO Sorter for Nickel Applications


9

Figure 10: UltraSort/Commodas PRO XRT Sorter

7 CONCLUSION

The upgrading of Nickel Ores using Ore Sorting equipment can be a beneficial inclusion in nickel mining operations. Electromagnetic sensors are becoming more effective in identifying low grade disseminated ores and there are alternative sensing solutions including dual energy x-ray systems which can be implemented.

8 REFERENCES

Allen, AB 2008, “High Speed Nickel and Sulphide Ore Sorters using Electromagnetic Principles” Alta, Perth, Western Australia Goode, K. 2006 “Ore sorter to shake up nickel production” Paydirt Australia Mining Review Africa, Issue 5 2006, “X-Ray Imaging World First in Sorting Technology”. Wortruba, H. and Riedel, F. 2006, “Sorting of Disseminated Sulphide Ore by X-Ray Transmission and Electromagnetic Sensors”, Sensor-Based Sorting 2006, Aachen, Germany.



PROCESS & EQUIPMENT

DESIGN


SO2 AIR OXIDATION – AN ENGINEERING PERSPECTIVE

By

Nitin Goel

GRD Minproc Limited, Australia [email protected]

Bill Baguley

Mixtec Australia

[email protected]

Presented by

Karel Osten [email protected]

CONTENTS

1. INTRODUCTION ............................................................................................................. 2 2. THEORY OF IRON AND MANGANSE REMOVAL BY AIR/SO2 ...................................... 2 3. GAS LIQUID MASS TRANSFER: .................................................................................... 5 4. DESIGN OVERVIEW....................................................................................................... 6 5. TEST PROGRAM ............................................................................................................ 7 6. SCALE-UP....................................................................................................................... 7 7. REFERENCES .............................................................................................................. 10


2

1. INTRODUCTION

In recent years, there has been an increased interest in the application of SO2/Air mixtures for the oxidation of leach liquors, Applications include:

• Oxidation and pH adjustment to precipitate and remove iron and manganese from solution in base metal recovery circuits,

• Oxidation of ferrous iron in uranium circuits either within the leach itself or in recirculated raffinates to promote dissolution

A number of development programs are being undertaken to evaluate applications and engineering firms are actively completing technical-economic studies based on the results. Projects planning to use this technology are in the detailed design phase and one has recently started operation.

The objectives for process engineering are to ensure that the system has an adequate residence time to achieve the desired degree of oxidation, whilst providing a chemical environment conducive to the oxidation. To attain these goals, it is necessary for the engineer to select an appropriate reactor system, evaluate the stoichiometry, kinetics and thermodynamics of the proposed reactions, while taking into account the gas dispersion and solid suspension mixing requirements.

2. THEORY OF IRON AND MANGANSE REMOVAL BY AIR/SO2

2.1. CHEMISTRY OF FE/MN REMOVAL BY AIR/SO2

2.1.1. Fe Oxidation

The chemistry of iron oxidation by air/SO2 and subsequent hydrolysis is generally accepted to be as follows:

2FeSO4 + O2 + SO2 + 6H2O → 2Fe(OH)3 + 3H2SO4 [1]

The stoichiometric SO2 requirement is therefore 0.5 mol SO2/mol Fe oxidized

2.1.2. Mn Removal

Zhang (1)

showed that manganese is removed as MnO2 at a pH levels below 4 and as Mn2O3 at a pH of 5-6. The pH range 4-5 is a transition region where both MnO2 and Mn2O3 can be formed.

The reaction at pH levels below 4 can be represented as:

MnSO4 + O2 + SO2 + 2H2O → MnO2 + 2H2SO4 [2]

The stoichiometric SO2 requirement is therefore 1.0 mol SO2/mol Mn removed

The reaction at pH greater than 5 can be represented as:

2MnSO4 + O2 + SO2 + 3H2O -> Mn2O3 + 3H2SO4 [3]

The stoichiometric SO2 requirement is therefore 0.5 mol SO2/mol Mn removed

2.2. MECHANISM OF FE/MN REMOVAL BY AIR/SO2

The conversion of SO2 to acid proceeds via several reaction paths, both in the gaseous and aqueous phase. The reported conversion of aqueous SO2 in the presence of excess oxygen via a peroxo-monosulphate type intermediate is of particular interest. This reaction is believed to be catalyzed by Mn

2+ and/or Fe

3+:

SO2 (g) �� SO2 (aq)

SO2 (aq) + O2 (aq) + H2O � peroxo-monosulphate (in presence of Mn2+

or Fe3+

)

Peroxo-monosulphate + H+ � H2SO4

The oxidation of Fe(II) and Mn(II) with SO2/Air mixtures has been investigated separately and it is postulated that both oxidation reactions follow the same oxidation mechanism, which includes the formation of peroxo-monosulphate species

(3). The removal of manganese is most efficient under

conditions that promote the rate of Mn(II) oxidation at the expense of Mn(IV) reduction and direct acid formation.


3

Zhang (1)

proposed that the oxidation of soluble Fe and Mn by means of an air/SO2 gas mixture proceeds via an intricate mechanism that also includes the formation of peroxo-monosulphate species. The proposed mechanism involves the slow initial formation of a ferric sulphite complex (or mangenic sulphite complex in the case of Mn oxidation) and decomposition to produce the

sulphite radical SO3• -

. This is followed by a fast reaction with O2 to form a peroxo-monosulphate

species SO5• -

, and subsequent HSO5 –

which is believed to be responsible for the oxidation of Fe(II) and Mn(II).

Krause (3)

investigated the removal of Fe via O2/SO2 gas mixtures and concluded that the efficiency of such systems is dependent on the dissolved O2:dissolved SO2 ratio. Krause identified an ‘oxygen excess’ regime (dissolved O2: dissolved SO2 > 60) which is associated with efficient oxidation and an ‘oxygen starvation’ regime (dissolved O2: dissolved SO2 < 60) where SO2 also exhibits reducing properties. Under the latter condition, manganese (for example) is first oxidised (it is the ‘forward reaction’) and then partially reduced (it is the ‘back reaction’) in an apparent inefficient net reaction.

2.2.1. ‘Net Oxidation’ of Mn by Air/SO2

The net oxidation of Mn by air/SO2 at pH below 4 can be shown to be the sum of the following two reactions:

Forward Reaction: 2MnSO4 + 2O2+ 2SO2 + 4H2O → 2MnO2 + 4H2SO4

Back Reaction: MnO2 + SO2 → MnSO4

Net Reaction: MnSO4 + 2O2+ 3SO2 + 4H2O → MnO2 + 4H2SO4

Determination of the optimum conditions in which the forward reaction is maximized and the back reaction is minimized is required for the reactor system design.

Krause’s work emphasised the following:

• The SO2 content of an air/SO2 mixture can act as both oxidizing and reducing agent.

• In order to ensure efficient oxidation by air/SO2 mixtures, oxidation should be performed under operating conditions that minimize the reducing behaviour of SO2. Such conditions can be identified by the following:

- The use of gas mixtures of moderate SO2 content, typically 0.5% SO2 in air (should be determined by actual testwork as per section 5.0).

- Addition of oxidant in a controlled manner as to avoid unnecessary side reactions and excessive reagent consumption in the system.

- Operating at the highest allowable pH levels (with consideration of co-precipitation losses of products such as cobalt).

- Operation at the lowest practical temperature.

• As the air/SO2 reaction is influenced by dissolved levels of O2 and SO2, careful consideration should be given to the gas-liquid mass transfer properties of the reaction vessels. Equations [4] and [5] show that dissolved gas levels (DO2, DSO2) are influenced by both gas composition and agitator performance (kLa).

Oxygen Transfer Rate = kLaO2 (C*O2 – DO2) [4]

SO2 Transfer Rate = kLaSO2(C*SO2 – DSO2) [5]

• It should be noted that C* is influenced by pressure (according to Raoult’s Law: P = P0 x, where P is the partial pressure of the component above solution, P0 is the vapor pressure of the pure component, and x is the mole fraction of the component in the solution) – this introduces ‘reactor depth’ as a further variable for the engineering design. Careful consideration should be given to reactor depth in the design of the system as lower SO2 in air concentrations would be required if ‘deep’ reactors are used to counter higher SO2 solubility (‘back reactions’ could become more prominent).


4

Pa

H

Gas Flow, Q Nm3/h

h

Partial Pressure Calculation Nomenclature

Assuming the gas composition to be x %v/v SO2, y %v/v O2, and z %v/v N2, the formula for the gas partial pressure for the various components at the gas injection point is given by:

• PSO2 = (x/100) * (Pa + �gH)

• PO2 = (y/100) * (Pa +�gH)

• PN2 = (z/100) * (Pa + �gH)

Where;

• H is the height of the liquor/slurry above the gas injection point in m

• Pa is the pressure on the tank top in N/m2 (normally site pressure)

• � is the density of the liquor/slurry

Table 1 below lists the variation in gas partial pressure for different liquid heights.

Table 1: Partial Pressure for Different Liquid Heights

Height Total

Pressure PSO2 PO2 PN2

m kPa.a kPa kPa kPa

0.5 106.2 2.1 20.2 83.9

1 111.1 2.2 21.1 87.8

2 120.9 2.4 23.0 95.5

3 130.7 2.6 24.8 103.3

4 140.5 2.8 26.7 111.0

5 150.3 3.0 28.6 118.7

6 160.1 3.2 30.4 126.5

7 169.9 3.4 32.3 134.2

8 179.7 3.6 34.1 142.0

The above table assumes the atmospheric pressure to be 101 kPa and a liquid density of 1,000 kg/m

3. The variation in partial pressure is shown for a fixed inlet gas composition of 2%v/v

SO2, 19%v/v O2, and 79%v/v N2.

As is evident for the above table, oxidation results obtained from a tank with 0.5 m liquid level can be significantly different from a tank of 8 m liquid level for the same gas composition.

Key observations from other work include the following:


5

• The oxidation of iron (the ‘forward reaction’) is much faster than the oxidation of Mn (1)

. As a result:

- Iron is less sensitive to the detrimental effect of the ‘back reaction’ and therefore more tolerant to fluctuations in air/SO2 concentration.

- Iron oxidation is characterized by efficient SO2 utilisation to the ‘forward reaction’.

- Mn oxidation is more sensitive to fluctuations in air/SO2 concentration.

• Mn oxidation rates are slow below pH of 4, but increases significantly at higher levels (1)

.

• Net Mn removal rate is inversely dependent on Mn tenor; Mn removal is faster and more efficient in terms of SO2 utilization at higher Mn tenors. The removal of Mn at low concentrations (typically less than 150 mg/l) is very inefficient.

• Manganese removal at low tenors is greatly improved in the absence of MnO2 (1)

; this principle is exploited by a Mintek patent whereby the Mn oxidation rate and efficiency is improved by inter-stage removal of manganese precipitates.

• The presence of ferric iron promotes manganese oxidation (1)

.

Messing (2)

investigated on a laboratory scale the effect of dissolved oxygen levels on the rate of manganese oxidation. The results illustrated that the rate of manganese precipitation was first order with respect to the dissolved oxygen concentration.

The influence of gas composition on the rate of manganese precipitation was also investigated. The rate of manganese removal increased linearly with an increase in SO2 content of the gas mixture, up to 12% (SO2 in an SO2/O2 mixture). There-after the rate dropped with a further increase in the SO2, probably due to the reduction of the product manganese oxides by excess dissolved SO2.

The rate of manganese removal decreased as the dissolved SO2 increased. This effect was more pronounced in tests that used a gas mixture rich in SO2. It was shown that the rate at which SO2/O2 gas mixtures precipitate manganese is dependent on the dissolved sulfur attributed to the dissolution of SO2 from the gas mixture. The drop in the rate of manganese removal was not seen for gas mixtures with a SO2 content of less than 9%.

Since SO2 is readily soluble in solution at the prevailing operating conditions, the design challenge is to provide adequate oxygen mass transfer to solution to match the oxygen uptake for oxidation reactions and maintain the dissolved oxygen level in solution.

3. GAS LIQUID MASS TRANSFER:

Gas-liquid mass transfer is described as follows:

TR = kLaT (Ci* - DC)

With

TR the gas transfer rate

kLaT the mass transfer coefficient at temperature T

Ci* the solubility of gas i under the prevailing operating conditions

DC the bulk dissolved gas concentration

The temperature dependence of kLa is as follows (T in °C):

kLaT = kLa20 . 1.024 (T-20)

In an agitated vessel, the mass transfer coefficient is dependent on power input and superficial gas velocity:

kLa20 = C1 (Pg/V)C2

(vs)C3

with

Pg the gassed power input to the reactor contents

V the pulp volume

Vs the superficial gas velocity

C1,C2,C3 constants describing the mass transfer characteristics of a specific impeller


6

The gas solubility at various temperatures and pressures can be determined from Henry’s Law;

yiP = H(T) xi

with

yi the concentration of component i in the gas phase (mole fraction)

P the operating pressure (atm)

H(T) Henry’s constant for component I at temperature T

Xi the soluble concentration of component i (kg/m3)

For ideal gases, the above relation can be written as:

pi = H(T) xi

where

pi is the partial pressure of the component i in the gas stream.

From the above relation it can be seen that the solubility of a component in solution at a particular temperature is directly proportional to the partial pressure of that component in the gas stream.

For sparingly soluble gases (oxygen in this case) the process of mass transfer is considered to be liquid film controlled.

The requirement to maintain a higher dissolved oxygen level in solution gives a lower driving force (C* - DC) for the oxygen mass transfer. Therefore, to maintain the same oxygen transfer rate, the constant kLa has to be changed, which is dependent on the agitator design.

When air is introduced below the surface, the equilibrium concentration C* is not a constant but varies through the tank. At the bottom, the equilibrium concentration is increased by the hydrostatic pressure of the liquid head. At the top, the equilibrium concentration is reduced by the stripping of oxygen from the air as the bubble rises. A logarithmic mean of these values provides a consistent model for the design of the oxygen mass transfer system.

A higher oxygen partial pressure in the gas stream, (a function of the oxygen concentration in the gas stream) and the hydrostatic head in the tank, will give a higher oxygen equilibrium concentration (as per Henry’s law). A higher equilibrium oxygen concentration would give a greater differential (C* - DC) for oxygen mass transfer and make the design of the oxygen mass transfer system more efficient.

Clues to the Kinetic Regime from Solubility Data:

For the reactions which occur in the film, the phase distribution coefficient H can suggest whether the gas phase resistance is likely to be significant or not.

Gas phase resistance controls for highly soluble gases.

Liquid film resistance controls for slightly soluble gases.

4. DESIGN OVERVIEW

The following points cover the major activities in the selection and sizing of equipment for the design process:

• Review the continuous test program showing the relationship of residence time to degree of oxidation, reagent requirements, and overall metallurgical balances.

• Define the required extent of reaction and determine the stoichiometry.

• Calculate process flows for the desired plant scale using demonstrated results in the testwork.

• From the stoichiometry and testwork results, estimate the SO2 and oxygen requirements and the heat of reaction.

• From the process flows and required residence time, size the reactors.

• From the SO2 and O2 demand, define air flow and dispersion requirements, and select the agitation system.

• From the heat of reaction and optimum target temperature for the reactions, calculate heat balance and size heat exchangers.


7

5. TEST PROGRAM

Because of the lack of operating examples SO2/Air systems used for Fe/Mn oxidation, laboratory testing for kinetic evaluation is a prerequisite for a credible design, even at a prefeasibility stage of a project. The testing should consist of initial batch amenability runs, followed by operation of a continuous lab scale reactor system under varying conditions to determine the reactor configuration, effect of recycle, optimum pH, residence time, feed gas requirement, optimum SO2/O2 ratio, dissolved O2 limitations, optimum DSO2/DO2 ratio, and neutralizing reagent requirement. It is imperative for the successful plant design that tests are conducted at varying partial pressures of the SO2 and O2 in the feed gas covering an entire envelope to enable scale-up and determination of the optimum partial pressures in the actual plant scale reactor.

If the scope allows, kinetic equations should be developed to predict the behaviour of the oxidation process under varying feed and operating conditions. Also, the testwork should aim to establish the following parameters:

• Determination of the optimum SO2 requirement by variation in SO2 dose rate.

• Determination of the optimum dissolved oxygen in the solution.

• Determination of the optimum ratio of dissolved SO2 to dissolved O2.

• Determination of the optimum SO2 concentration in the feed gas by variation in the SO2 to O2 ratio in the feed gas.

• Determination of SO2 and O2 utilisation at varying feed gas conditions.

Due to the challenge in designing gas dispersion equipment for SO2/Air applications, it is recommended that agitator vendors should be engaged early during the testwork to ease subsequent scale-up during the engineering design of the reactors.

6. SCALE-UP

6.1. AGITATOR DESIGN

Previous researchers studying gas dispersion have been more concerned about energy levels than the mixing profile. However a simple test increasing and reducing the impeller size shows the impact that the impeller diameter to tank diameter (D/T) ratio has on the impeller flooding point and hence the efficiency of the agitator. Therefore the D/T and the flooding point of the system must be determined as part of the testwork

Von Essen commented on the determination of the flood point as one of the most important design parameters in gas-liquid mixing. This is because the impellers can become flooded if agitation is not sufficiently intense when gas is sparged to a mixing vessel. The empirical correlation below can help to avoid this problem by predicting the transition point to flooding.

The correlation was developed with the aid of data over a wide range of vessel sizes and can be used in scale-up of agitators from laboratory testing. Geometric similarity is important and the correct sizing must be developed from the laboratory. To minimise experimental error it is advised that the test tanks used exceed 500 mm in diameter. In addition experience has shown that it is better to use the gas rising velocity and not the volumetric flow rate when establishing scale criteria. Previous researchers studying gas liquid agitation had only used relitively small range of mixing volumes. Von Essen work focued on a broad range of volumes including site experience. This resulted in a far better correlation for predicting the flooding point and hence best operating point for the agitator.

6.2. THE FLOOD POINT

It is interesting to note that if a mixer starts flooding and the gas is still sparged into the tank, then the mass transfer can still occur. However, impeller pumping rates fall off sharply which can lead to such process problems as:

• Solids falling out of suspension and building up in the tank.

• Formation of dead zones, resulting in poor mass transfer.

• incomplete blending can occur resulting in short circuiting through the tank.


8

A typical reaction by the operator, who witnesses a fall off in mass transfer, is to increase the gas flow rate. As we shall see this can have the reverse effect by pushing the impeller firmly into the flooded zone and further reducing the extraction efficiency.

Figure 1 Weak Flow Pattern

and Poor Mass Transfer

Figure 2 Poor Flow Pattern and Mass Transfer

Figure 3 Better Flow Pattern and Mass Transfer

Figure 4 Best Flow

and Mass Transfer

The flooding point has been defined as any transition between the conditions shown in Figure 1 and 2. However, most engineers are mainly interested in bringing about a change from a gas-controlled system (Figure 2) to an impeller-controlled one (Figure 3).Note a flooded system is easy to identify by the reversal in flow direction at the surface , (Figure 1 and 2) The liquid flow from the centre of the tank to the tank side makes an excellent indicator of the flooding transition, especially in large vessels or at high gas flowrates

Predictions of the flooding transition point in agitated gas-liquid systems have been offered for many power-based correlations including Nienow and Lu

(4, 5). These reduce, at constant D/T ratios, to

N3D

4/Qa

(4) and N

3.3 D

4/ Qa

(5), respectively, the indices should remain constant upon scale-up.

However, use of these equations on large gas-dispersion vessels predicts agitator designs that are too big.

In other cases power per unit volume has been found too conservative, a correlation using torque per unit volume has been successful. Here, the ratio of torque per unit volume to superficial gas velocity reduces to N’

2D

4/ Qa. Again, however, use of this correlation for design of large vessels has

resulted in problems, which in this case result in undersized agitators.

A more successful method to predict flooding transition followed extensive testing and a combination of the previous approaches. The resulting correlating giving the Gas Transition factor GTF = Np N

2.67D

4/Qa.


9

Figure 5: Flooding Transition (Impeller Controlled Flow)

Although the equation GTF = Np N2.67

D4

/ Qa holds over a wide range of fluid volumes, the D/T ratio is a important process variable. The value of the factor falls exponentially at high values of D/T for both HA790 S turbines (radial-flow, disk type) and wide-blade high-solids hydrofoils (HA736 axial-flow impellers).

The factor can be used to calculate the power requirements per unit volume to avoid flooding de-clines slightly with increasing volume. Von Essen reports that this correlation predicts power per unit volume to decrease as the — 1/12

th power of volume. This happens to be the same scale-up rule as

that for dilute solids suspension. It makes sense that agitation for settling solids and rising bubbles should scale-up in the same manner.


10

Table 2: Gas Transition Factor – Fluid DynamicsLaboratory Scale-Up Procedure

Example Site Laboratory

Tank Size 12 x 12 m 0.6 x 0.6 m

Tank Area 113 m2 0.283 m

2

Tank Volume 1300 m3

160 m3

Gas Rate 5 m3/m

3pulp/hr

Gas Flow Rate 6,500 m3/hr 0.8 m

3/hr

At Pressure 3,250 m3/hr 0.74 m

3/hr

Rising Velocity 0.48 m/min 0.04 m/min

Required Velocity for Laboratory Test 0.48 m/min

Required Flow for for Laboratory Test 8.15 m3/hr

As can be seen in the above example the gas rising velocity is important when making predictions on scale up. Most impellers can disperse gas in the 0.04 m/min gas rising velocity range and the lower this number is, the easier it is for any type of impeller to perform well. However when the gas rising velocity is high only special dispersion impellers can be used. It is therefore important to test the impeller performance at the correct gas rising velocity and not the desired velocity based on reaction rates. Compensation has to be allowed for scale up under these conditions

7. REFERENCES

1. Zhang et al. Oxidative Precipitation of Manganese with SO2/O2 and separation from Cobalt and Nickel. Hydrometallurgy 63 (2002) 127-135.

2. Messing et al. An empirical rate equation for the partial removal of manganese from solution using a gas mixture of sulphur dioxide and oxygen. Hydrometallurgy 86 (2007) 37-43.

3. E. Krause. The Oxidation of Ferrous Sulphate Solutions by Sulphur Dioxide and Oxygen. PhD Thesis, University of Waterloo, Ontario, Canada, 1988.

4. Nienow et al. On Flooding/loading transition and complete dispersal conditions in aerated vessels agitated by Rushton turbines. Proceedings from Fifth European Mixing Conference, Germany, 1985.

5. Lu, W and Chen,H. Flooding and critical impeller speed for gas dispersion in aerated turbine- agitated vessels. Chemical Engineering Prog 65, 8, 1986.


THE DEVELOPMENT OF A DSX PROCESS FOR THE

RECOVERY OF NICKEL AND COBALT FROM LATERITE

LEACH SOLUTIONS - FROM BATCH TESTS TO PILOT

PLANT OPERATION

By

C.Y. Cheng1, G. Boddy

2, W. Zhang

1, M. Godfrey

2, D.J. Robinson

1, Y.

Pranolo1, Z. Zhu

1, L. Zeng

1 and W. Wang

1

1 The Parker, CSIRO Minerals, Australia.

2 Rio Tinto Technology and Innovation, Australia.

Presented by

Chu Yong Cheng

[email protected]

CONTENTS

ABSTRACT 2 1. INTRODUCTION 2 2. BATCH TESTS 3 3. SEMI CONTINUOUS TESTS 6 4. FULLY CONTINUOUS TESTS 8 5. CONCLUSIONS 12 6. ACKNOWLEDGEMENT 12 7. REFERENCES 13


2

ABSTRACT The synergistic solvent extraction system consisting of Versatic 10/LIX63/TBP was used for the separation and purification of nickel and cobalt from synthetic laterite leach solution after iron removal. Batch, semi and fully continuous tests were conducted to optimise operating conditions. Very encouraging results were obtained for separating nickel and cobalt from manganese, magnesium and calcium. In batch tests, the organic composition of the synergistic system was optimised and very large synergistic shifts for nickel and cobalt were observed. The ∆pH50(Mn-Co) and ∆pH50(Mn-Ni) values were as large as 1.65 and 2.40, respectively. The extraction kinetics of nickel and cobalt were fast with 91% nickel and 97% cobalt extracted within 30 seconds. The stripping kinetics of nickel and cobalt was also fast: after 2 minutes of mixing, 92% nickel and 98% cobalt were stripped. Semi continuous extraction tests were conducted using the synergistic organic system consisting of 0.50 M Versatic 10, 0.45 M LIX63 and 1.0 M TBP in Shellsol D70. With a pH profile of 5.5/5.9/6.3 in three stages at 40°C, the nickel and cobalt extractions reached 99.9% with only 5 mg/L nickel and <1 mg/L cobalt left in the raffinate. With two stages of scrubbing, two minutes of residence time, a pH profile of 5.4/5.0 at 40°C, about 2 mg/L manganese and less than 1 mg/L magnesium and calcium were left in the scrubbed organic solution. With two stripping stages, two minutes residence time and an O/A ratio of 10 at 40°C using 50 g/L H2SO4 as strip solution, the stripping efficiencies of Ni and Co reached over 95%. A fully continuous pilot plant operation was conducted successfully for 280 hours. With an O/A ratio of about 2 and a pH profile of 5.5/5.8/6.0/6.3 for the four stages EX1/EX2/Ex3/EX4 at 40°C, both nickel and cobalt were almost completely extracted. The nickel and cobalt concentration in the raffinate was lower than detection limit of 0.2 mg/L. The manganese, magnesium and calcium concentrations in the loaded organic solution were 34, 8 and 1 mg/L, respectively. Using a pH profile of 5.4/5.0 for SC1/SC2 at an O/A ratio of 10 and 40°C, the manganese scrubbing efficiency was over 96% and the concentrations of manganese and magnesium in the scrubbed organic were <5 mg/L and that of calcium 1 mg/L. Using three strip stages and a strip solution containing 50 g/L sulphuric acid and 55 g/L Ni at an O/A ratio of 10 and 40°C, over 98% Ni and 99% Co were stripped with only 64 mg/L Ni in the stripped organic solution. The nickel concentration in the loaded strip liquor reached 83 g/L, indicating a ∆Ni of over 28 g/L. The pH of the loaded strip liquor was 3.1 and the loaded strip liquor contained less than 1 g/L acid. The loaded strip liquor would be suitable as a feed to a conventional Cyanex 272 solvent extraction circuit in which Co, Zn and Mn would be separated from Ni.

No gypsum precipitation was observed, even though the feed solution was deliberately saturated with calcium when the majority of the organic was saponified with NaOH and pH was adjustment by Na2CO3.

1. INTRODUCTION

Novel synergistic solvent extraction (SSX) systems have been developed by the solvent extraction (SX) team of the Parker Centre (CSIRO Minerals) to separate nickel and cobalt from manganese, magnesium and calcium in leach solutions using direct solvent extraction processes (DSX) without intermediate precipitation and re-leach steps [1–6]. At the request of Rio Tinto Technology and Innovation, a project was carried out to test the SSX systems and to develop DSX process flowsheets to purify and recover nickel and cobalt from a synthetic leach solution. This paper presents the test results from a series of shake out, semi-continuous and fully continuous tests (piloting).


3

2. BATCH TESTS

The objective of the batch tests was to optimise the composition of the SSX system consisting of Versatic 10 acid, LIX63 and TBP (with Shellsol D70 used as diluent) for the separation of nickel and cobalt from manganese, magnesium and calcium for a specified synthetic laterite leach solution after iron removal at pH 5.3 (Table 1).

Table 1 – Composition of the feed solution to SX tests

Synthetic feed solution elemental concentrations (mg/L)

Ni Co Zn Mn Mg Ca Si Na Cl Cu Fe Al Cr

5048 271 118 1573 30422 486 15.6 2839 8207 LD* LD* LD* LD*

*LD: Lower than detection limit, 0.2 mg/L. 2.1 EXTRACTION PH ISOTHERMS A series of tests with variable concentrations of Versatic 10, LIX63 and TBP were conducted. Based on the recovery of nickel and cobalt, and the rejection of manganese and calcium, the best organic composition was determined to be 0.50 M Versatic 10, 0.45 M LIX63 and 1.0 M TBP, used at an O/A ratio of 2, at 40°C. Compared to the organic system containing Versatic 10 alone (Figure 1), this synergistic organic system resulted in large synergistic shifts for nickel and cobalt with their ∆pH50 values of 3.20 and 2.47 pH units, respectively (Table 2 and Figure 2). The ∆pH50 (Mn-Ni) increased substantially from 0.35 to 2.40 pH units and ∆pH50 (Mn-Co) from 0.33 to 1.65 pH units. By the use of this SSX system, 98.3% nickel and 79.7% cobalt were extracted at pH 5.5 with only 84 mg/L nickel and 53 mg/L cobalt remaining in the raffinate after a single contact (Table 3). The extractions (or entrainment) of manganese, magnesium and calcium were small, with the loaded organic solution containing only 6 mg/L manganese and magnesium and 0.5 mg/L calcium.

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

4.5 5.0 5.5 6.0 6.5 7.0 7.5

Equilibrium pH

Extr

acti

on

(%

)

Ni

Co

Zn

Mn

Mg

Ca

Figure 1 – Extraction isotherms with 0.50 M Versatic 10 in Shellsol D70 operating at

an O/A ratio of 1:1.5 and a temperature of 40°C.


4

Table 2 – Comparison of pH50 and ∆pH50 values of nickel and cobalt over

manganese with the organic solution containing 0.33 M Versatic 10 alone and the

SSX system.

V10 (M)

L63 (M)

TBP

(M) O/A

pH50

(Mn) pH50

(Co) pH50

(Ni)

∆pH50

(Mn-Co)

∆pH50

(Mn-Ni) SF

(Co/Mn) SF

(Ni/Mn)

0.33 0.0 0.0 1.5 7.35* 7.02 7.00 0.33* 0.35* 6 15

0.50 0.45 1.0 2.0 6.20 4.55 3.80 1.65 2.40 534 7720

∆pH50(Me) 1.15* 2.47 3.20 --- --- --- ---

• Estimated.

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0

Equilibrium pH

Extr

acti

on

(%

)

Ni

Co

Zn

Mn

Mg

Ca

Figure 2 – Extraction isotherms with the SSX system operating at an O/A ratio of 2

and 40°C.

Table 3 – Metal concentrations in the aqueous and organic solutions with the SSX

system operating at pH 5.0 with O/A ratio of 2.0

Concentrations (mg/L)

Ni Co Zn Mn Mg Ca

Feed solution 4920 240 120 1640 32480 480

Loaded organic 2389 105 39 6 6 0.5

Raffinate 84 53 67 1689 34873 520

Extraction (%) 98.3 79.7 45.9 0.73 0.03 0.19

2.2 EXTRACTION KINETICS

It can be seen from Figure 3, the extraction kinetics of cobalt and nickel were very fast with 91% nickel and 97% cobalt extracted in the first 30 seconds. At the same time, about 13% manganese was extracted. In the next 30 seconds, most of the co-extracted manganese was crowded out by nickel and its extraction decreased to 2%. The majority of the Co was recovered, and the small amount crowded out by Ni would be extracted in the next extraction stage of a counter-current operation.

2.3 STRIPPING KINETICS

The SSX system was loaded at pH 5.5, an O/A ratio of 2 and 40°C. The concentrations of nickel, cobalt and manganese in the loaded organic solution were 2.25, 0.112 and 0.033 g/L, respectively. The stripping tests were conducted using the loaded organic


5

solution and a synthetic spent nickel electrolyte containing 6 g/L nickel and 7 g/L sulphuric acid at an O/A ratio of 1 and 40°C. Very fast cobalt and manganese stripping kinetics and reasonable fast nickel stripping kinetics were observed (Figure 4). After 2 minutes of mixing, the stripping efficiencies of nickel, cobalt and manganese reached 92%, 98% and 99%, respectively.

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0

Time (minutes)

Extr

acti

on

(%

)

Ni

Co

Mn

Figure 3 – Extraction kinetics with the SSX system at an O/A ratio of 2 and 40°C.

2.4 PHASE SEPARATION

It was found that during extraction the phase disengagement time (PDT) under organic continuity was almost three times longer than under aqueous continuity both for primary and secondary phase disengagement (Table 4). This indicates that organic continuity should be avoided during extraction operation. For the stripping operation, the PDT under aqueous continuity was shorter than that under organic continuity for both primary and secondary phase disengagement. However, the PDT difference was relatively small. Therefore, both continuities could be used for stripping.

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0

Time (minutes)

Str

ipp

ing

(%

) Co

Ni

Mn

Figure 4 – Stripping kinetics with the SSX system at 40°C.


6

Table 4 – Phase disengagement times during extraction and stripping

Operation Continuity 1st PDT (s) 2

nd PDT(s)

A/C 140 165 Extraction

O/C 360 490

A/C 79 121 Stripping

O/C 157 174

3. SEMI CONTINUOUS TESTS

The objectives of the semi-continuous tests were to determine the optimum conditions for extraction, scrubbing and stripping separately in counter current operations for the separation of nickel and cobalt from impurities. Three and four stages of mixer-settlers (0.2 L mixer and 0.4 L settler) were used for the semi-continuous tests. A three-stage semi continuous extraction flowsheet is shown in Figure 5.

Figure 5 – A schematic of semi-continuous test flowsheet with three extraction

stages

3.1 SEMI CONTINUOUS EXTRACTION

Semi continuous extraction tests were conducted using the synergistic organic system consisting of 0.50 M Versatic 10, 0.45 M LIX63 and 1.0 M TBP in Shellsol D70. With a pH profile of 5.5/6.1/6.5 (EX1/EX2/EX3) in three stages at 40°C, the nickel and cobalt extractions reached 99.1% with only 12 mg/L nickel and <1 mg/L cobalt left in the raffinate (Table 5). The concentration of manganese in the loaded organic solution was 43 mg/L. With four extraction stages and a pH profile of 5.5/5.8/6.0/6.3 (EX1/EX2/EX3/EX4), both nickel and cobalt extractions reached over 99.9% with only 1 mg/L nickel and <1 mg/L cobalt left in the raffinate (table 6). The manganese concentration in the loaded organic solution was only 17 mg/L, suggesting better performance with four extraction stages and a pH profile of 5.5/5.8/6.0/6.3.

EX2

Raffinate

EX3

NaOH

Control

pH 6.3

Fresh

organic

Aq. recycling

Aq. recycling

EX1

Feed

.

Aq. recycling

Loaded

organic

Control

pH 5.5 Control

pH 5.9


7

Table 5 – Metal concentrations and recovery with 3-extraction stages and pH

profile 5.5/5.9/6.3 for EX1/EX2/EX3 at A/O ratio 1:2 and 40°C

Element Metal in

aqu. feed (g/L)

Metal in loaded org.

(g/L)

Metal in raffinate

(g/L) Total recovery

(%)

Ni 5.769 3.128 0.005 99.92

Co 0.280 0.149 <0.001 >99.9

Zn 0.161 0.101 <0.001 >99.6

Mn 1.883 0.096 1.899 8.35

Mg 34.52 0.109 35.01 0.47

Ca 0.462 0.001 0.444 0.00

Cl 7.000 0.010 5.925 0.00

Table 6 – Metal concentrations and recovery with 3-extraction stages and pH

profile 5.5/5.8/6.0/6.3 for EX1/EX2/EX3/EX4 at A/O ratio 1:2 and 40°C

Element Metal in

aqu. feed (g/L)

Metal in loaded organic (g/L)

Metal in raffinate

(g/L)

Total recovery (%)

Ni 5.840 3.250 0.001 99.99

Co 0.237 0.137 <0.001 99.91

Zn 0.149 0.118 0.001 99.17

Mn 1.850 0.017 0.782 1.82

Mg 37.90 0.003 35.40 0.00

Ca 0.486 0.005 0.430 0.00

Cl 6.251 0.010 6.019 0.00

3.2 SEMI CONTINUOUS SCRUBBING

The goal of semi-continuous scrubbing tests was to scrub the co-extracted manganese, magnesium and calcium from the loaded organic solution, so that high grade nickel and cobalt products can be obtained in the subsequent processes. With two scrub stages, two minutes of residence time, a pH profile of 5.4/5.0 for SC1/SC2, a scrub solution containing 600 mg/L nickel at 40°C and an O/A ratio of 5, the manganese scrubbing efficiency reached 97% and only 2 mg/L of manganese (and less than 1 mg/L magnesium and calcium) was left in the scrubbed organic solution. Only about 2% nickel and no cobalt were scrubbed (Table 7). In the batch scrubbing test, it was found that some 9% cobalt was scrubbed. In a two-stage counter current continuous operation, the scrubbed cobalt in SC2, where the scrub solution entered and the organic solution left the scrub circuit, was extracted by the loaded organic solution in SC1, with the overall results that no cobalt was scrubbed.


8

Table 7 – Scrubbing with two stages and a pH profile of 5.4/5.0 using a scrub

solution containing 600 mg/L nickel at an O/A ratio of 5

Concentration (g/L)

Element In loaded organic

In scrubbed organic

Scrubbing Efficiency

(%)

Ni 2.900 2.960 -2.07

Co 0.112 0.112 0.00

Zn 0.049 0.044 10.02

Mn 0.070 0.002 96.88

Mg 0.008 <0.001 87.50

Ca 0.001 <0.001 0.00

3.3 SEMI CONTINUOUS STRIPPING The objectives of semi continuous stripping tests were to obtain loaded strip liquors suitable for downstream processing (eg nickel/cobalt separation via SX and nickel electrowinning) and to generate stripped organic solutions containing low residual metals suitable for recycling to the extraction stages. With two stripping stages, 2 minutes of residence time, an O/A ratio of 10 at 40°C using a strip solution containing 50 g/L H2SO4 and 80 g/L nickel, the stripping efficiencies of 95-96% were achieved for nickel, cobalt and zinc (Table 8). The ∆Ni in the loaded stripped liquor reached 24 g/L and the loaded strip solution contained over 100 g/L nickel. The pH of resultant loaded strip liquor was found to be 2.1, suggesting a residual acid of 0.4 g/L, which was reasonably low.

Table 8 – Metal concentrations and their stripping efficiencies with two stripping

stages and a strip solution containing 50 g/L H2SO4 and 80 g/L nickel

Concentration (g/L) Element In scrubbed

organic In stripped

organic

Stripping efficiency

(%)

Ni 3.518 0.128 96.37

Co 0.114 0.005 95.63

Zn 0.050 0.002 96.71

Mn 0.010 <1 94.42

4. FULLY CONTINUOUS TESTS

The objectives of the fully continuous tests were to further optimise the operating conditions for extraction, scrubbing and stripping in a fully counter current operation mode for the separation of nickel and cobalt from impurities based on the data obtained in the semi continuous tests, and to accumulate data for plant design and operation. The SX pilot plant consisted of up to nine water-jacketed mixer settlers (with individual active volumes of 0.5 L for mixer and 1.2 L for settler). A photograph is shown in Figures 6 and a schematic flowsheet in Figure 7. The pH values in all extraction and scrubbing stages were adjusted using pH controllers and dosing pumps with diluted base solutions. The stripped organic solution was batch-saponified with 30% (w/w) NaOH solution in some tests. The pilot plant was operated for over 280 hours. The parameters tested include O/A flowrate ratios, number of stages and pH profiles. Surveys of operating parameters were conducted at two-hour intervals and samples taken at four-hour intervals. Most of the samples were sent to a commercial analytical laboratory for assay.


9

Figure 6 – The fully continuous pilot plant

4.1 EXTRACTION

4.1.1 Gypsum formation

In other operations using a DSX process, such as in the Bulong process [7], gypsum formation caused serious operation problem. Therefore, the potential for a similar gypsum formation problem was closely observed under strict conditions. The aqueous feed solution was deliberately saturated with CaSO4 by addition of CaCl2 at 40˚C before it entered extraction Stage 1. During the three weeks of operation, no gypsum precipitation was observed under the testing conditions.

Figure 7 – A schematic flowsheet of the fully continuous pilot plant operation

SC

Raff.

EX

Saponification with

30% NaOH

Sapon.

organic

ST

Feed

Stripped

organic

Scrub solution

Strip solution

LSL


10

4.1.2 Crud formation

Cruds, consisting mainly of basic magnesium sulphates, formed in the extraction stages when 10% (w/w) NaOH solution was injected directly into the mixer. Separate batch tests were carried out and it was found that the use of sodium hydroxide as neutralisation reagent was responsible. Locally high pH occurred in the mixer when the 10% (w/w) NaOH solution was injected directly to adjust the pH values.

Although the use of ammonia solution could solve the problem, it was felt that adding ammonia to the system would necessitate more complex and expensive final tailings treatment. An alternative method of adding NaOH was tested. About 90% of the base required to saponify the organic for the extraction of all Ni, Co and Zn was added as 30% (w/w) NaOH solution directly to the stripped organic before the extraction operation. The remaining 10% of base required was automatically and directly injected to the mixer by dosing pumps of the pH controllers with lower NaOH concentration of 5% (w/w). This method proved to be effective in reducing crud formation. 4.1.3 Effect of number of extraction stages and pH profiles

In all the tests conducted, the Co extraction was over 99.9%, with only <1 mg/L Co left in the raffinate. The Ca concentration in the loaded organic was typically <5 mg/L and that of Mg <30 mg/L, indicating that the separation of Ni and Co from Mg and Ca was very good, especially when taking into consideration of the Mg concentration of over 30 g/L in the feed. The effect of number of stages and pH profile on the extraction of nickel and cobalt, and the separation of nickel and cobalt from manganese and magnesium are shown in Table 9. The results were obtained during the piloting campaign with three and four extraction stages. With three extraction stages and a pH profile of 5.4/6.0/6.4 for EX1/EX2/EX3 at an O/A ratio of about 2 and 40˚C, the nickel extraction reached 99.8% and cobalt over 99.9% with only 8 mg/L nickel left in the raffinate. The manganese, magnesium and calcium concentration in the loaded organic solution was 75, 10 and 1 mg/L, respectively. With three extraction stages and a pH profile of 5.3/5.8/6.0/6.3 for EX1/EX2/EX3/EX4 at an O/A ratio of about 2 and 40˚C, both nickel and cobalt were almost completely extracted. The nickel and cobalt concentration in the raffinate was lower than detection limit of 0.2 mg/L. The manganese, magnesium and calcium concentrations in the loaded organic solution were 34, 8 and 1 mg/L, respectively. Although a wider pH control range was used in three-stage extraction to scrub the co-extracted manganese and magnesium and to extract residual nickel in the raffinate, a four-stage extraction with narrow pH control range performed better due to the scrubbing effect of nickel on the co-extracted manganese and magnesium in a large number of stages. Table 9 – Effect of number of stages and pH profile on extraction at an O/A ratio of

2 and 40˚C

Extraction (%) In Raff. (mg/L)* In loaded organic (mg/L) No of

Stages pH

profile Ni Co Ni Co Mn Mg Ca

3 5.4/6.0/6.4 99.8 >99.9 8 <0.2 75 10 1

4 5.5/5.8/6.0/6.3 >99.9 >99.9 <0.2 <0.2 34 8 1

* 0.2 mg/L was the detection limit 4.2 SCRUBBING

The two scrubbing stages were pH controlled with diluted sodium hydroxide solution. The pH profile was varied to maximise scrubbing efficiencies of manganese, magnesium and calcium. The scrub solution was made by diluting the aqueous feed solution 10-fold and mixing with a portion of the loaded strip liquor to give a total nickel concentration in the range of 1.2-1.3 g/L. The flowrate ratio of the organic solution to the scrub solution


11

(O/A ratio) was controlled at about 10. The choice of scrub solution makeup is dictated by the need to have sufficient ionic strength to obtain acceptable phase separation, and the desire to avoid using large amounts of loaded strip liquor. Industrially this is achieved by having a high O/A in the scrub stage, but this would make the time for the pilot plant to reach steady state very long. As a compromise, for most of the pilot plant the ionic strength of the scrub solution was increased by including some feed solution in the scrub solution and lowering O/A ratio in the scrub circuit. As a result some impurities such as manganese, magnesium and chloride came into the scrub circuit. These impurities entered the strip circuit by aqueous entrainment in the scrubbed organic solution. In the last test, the flowrate ratio of the organic solution to the scrub solution increased to 20 with a scrub solution made from loaded strip liquor by diluting 40-fold, to demonstrate that the main source of manganese, magnesium and chloride in the loaded strip liquor was from the aqueous entrainment in the scrubbed organic solution. The phase separation was very good. Therefore, in practise, the scrub solution should be made from loaded strip liquor by dilution. With a pH profile of 5.4/5.0 for SC1/SC2, in most cases, the manganese scrubbing efficiency was over 95% and the concentrations of manganese in the scrubbed organic solution were below 4 mg/L (Table 10). The scrubbed organic solution contained <5 mg/L Mg and 1 mg/L Ca. Since the scrub solution was diluted from the feed solution and therefore contained nickel and cobalt, some cobalt was extracted by the organic solution, which is shown by the negative scrub efficiency. Table 10 – Typical cobalt and manganese scrubbing efficiencies and manganese,

magnesium and calcium concentrations in the scrubbed organic solution

pH Scrub efficiency (%) In scrubbed org. (mg/L)

SC 1 SC 2

O/A ratio Co Mn Mn Mg Ca

5.32 5.01 9.18 10.08 96.21 1 1 1

5.41 5.01 10.19 -5.71 96.97 2 2 1

5.52 4.92 10.99 -4.87 95.42 3 4 1

5.61 5.10 10.83 6.58 91.71 4 4 1

5.29 4.87 10.98 7.2 96.70 2 5 1

5.47 4.97 9.58 -10.18 95.79 3 4 1

5.47 5.02 9.93 12.79 97.49 2 3 1

4.3 STRIPPING

The main objective was to maximise metal stripping efficiencies while minimising the residual metal concentrations in the stripped organic solution and keeping the residual acid in the loaded strip liquor as low as possible. Initially two strip stages with a strip solution containing 50 g/L H2SO4 and 75 g/L Ni were tested, and later three strip stages with a strip solution containing 50 g/L H2SO4 and 55 g/L Ni were tested.

Table 11 lists typical strip results with two and three strip stages. With two strip stages, over 96% of nickel and cobalt were stripped and the residual nickel concentration in the stripped organic solution was 112 mg/L. The nickel concentration in the loaded strip liquor reached 106 g/L, indicating a ∆Ni of over 30 g/L. The pH of the loaded strip liquor was 2.4, suggesting a low acidity and minimum amount of acid to be neutralised in the down stream Cyanex 272 SX circuit.

With three strip stages, over 98% Ni and 99% Co were stripped and the residual nickel concentration in the stripped organic solution was 64 mg/L. The nickel concentration in the loaded strip liquor reached 83 g/L, indicating a ∆Ni of over 28 g/L. The pH of the loaded strip liquor was 3.1 and the loaded strip liquor contained <1 g/L acid. The loaded strip liquor would be highly suitable as a feed to a conventional Cyanex 272 solvent extraction process in which Co, Zn and Mn would be separated from Ni.


12

Compared with two strip stages, the nickel strip efficiency was higher and the residual nickel concentration in the stripped organic solution was lower when three strip stages were used. Most importantly, with two strip stages, when the circuit was disturbed, the stripped organic solution turned green, indicating high residual nickel concentration. With three strip stages, the circuit was more stable.

Table 11 – Typical stripping efficiencies, residual Ni concentrations in the stripped

organic solution and metal concentrations and pH in the loaded strip liquor

Strip eff. (%) In loaded strip liquor

(g/L) No of stage

s Ni Co

pH in ST1

Acid in ST1 (g/L)

Res. Ni

(mg/L) Ni Co Zn

2 96.1 96.9 2.38 --- 115 106 1.21 0.63

3 98.0 99.5 3.06 0.96 64 83.4 1.29 0.76

5. CONCLUSIONS

The synergistic system consisting of 0.5 M Versatic 10, 0.45 M LIX63 and 1.0 M TBP in Shellsol D70 generated very large synergistic shifts for nickel and cobalt with the ∆pH50(Mn-Co) and ∆pH50(Mn-Ni) values of 1.65 and 2.40, respectively, and provided separation factors for cobalt and nickel over manganese as large as 534 and 7720, respectively. The extraction and the stripping kinetics of nickel and cobalt were fast and the phase separation was suitable for industrial operations. In semi continuous extraction with three stages, the nickel and cobalt extractions reached 99.9% with only 5 mg/L nickel and <1 mg/L cobalt in the raffinate. After scrubbing in two stages, only 2 mg/L manganese and <1 mg/L of both magnesium and calcium were in the scrubbed organic solution. The stripping efficiencies of nickel, cobalt and zinc were very high with the residual nickel occupying only 4% of organic capacity. Fully continuous pilot plant operation was conducted successfully for over 280 hours. Using four extraction stages and a pH profile of 5.5/5.86.0/6.3 for EX1/EX2/EX3/EX4 at an O/A ratio of about 2 and 40°C, the nickel and cobalt were almost completely extracted. Using a pH profile of 5.4/5.0 for SC1/SC2 at an O/A ratio of 10 and 40°C, the manganese scrubbing efficiency was over 96% and the concentrations of manganese and magnesium in the scrubbed organic were <5 mg/L and that of calcium 1 mg/L. Using three strip stages and a strip solution containing 50 g/L sulphuric acid and 55 g/L Ni at an O/A ratio of 10 and 40°C, over 98% Ni and 99% Co were stripped with only 64 mg/L Ni in the stripped organic solution. The nickel concentration in the loaded strip liquor reached 83 g/L, indicating a ∆Ni of over 28 g/L. The pH of the loaded strip liquor was 3.1 and the loaded strip liquor contained less than 1 g/L acid. The loaded strip liquor would be suitable as a feed to a conventional Cyanex 272 solvent extraction circuit in which Co, Zn and Mn would be separated from Ni.

No gypsum precipitation was observed, even though the feed solution was deliberately saturated with calcium. The potential precipitation of basic magnesium sulphates in the extraction stages caused by NaOH addition was avoided by the use of saponification of the majority of the organic system with 30% NaOH (w/w) solution with pH adjustment by Na2CO3 solution.

6. ACKNOWLEDGEMENT

The authors would like to thank CSIRO Minerals and Rio Tinto Technology and Innovation for the permission to publish this paper.


13

7. REFERENCES

1. C.Y. Cheng, SX application for nickel and cobalt: pros and cons of existing

processes and possible future development., Proceedings of ALTA SX/IX World Summit, ALTA Metallurgical Services, Perth, Australia 2003.

2. C.Y. Cheng, and M.D. Urbani, Purification of laterite leach solutions by direct solvent extraction, Yazawa International Symposium on Metallurgical and Materials Processing: Principals and Technologies, TMS Annual Meeting, San Diego, USA, 2003, Vol. 3, 251-265.

3. C.Y. Cheng, and M.D. Urbani, Solvent extraction process for separating cobalt and/or nickel from impurities in leach solutions. International Patent Publication, 2005, WO 073416A1.

4. C.Y. Cheng, and M.D. Urbani, Solvent extraction process for separating cobalt and/or manganese from impurities in leach solutions. International Patent Publication, 2005, WO 073415A1.

5 C.Y. Cheng, Solvent extraction of nickel and cobalt with synergistic systems consisting of carboxylic acid and aliphatic hydroxyoxime. Hydrometallurgy, Vol. 84, 2006, 109-117.

6 C.Y. Cheng, W. Zhang and Y. Pranolo, Separation of cobalt and zinc from manganese, magnesium and calcium using synergistic solvent extraction consisting of Versatic 10 and LIX63. Solvent Extraction Fundamentals and Industrial Applications, Proceedings of ISEC 2008, Tucson, USA, 63-168.

7. C.Y. Cheng, Separation of Nickel from Calcium by Synergistic Solvent Extraction. ALTA SX/IX – 1, ALTA Metallurgical Services, Adelaide, Australia, October 2000.


AGITATOR DESIGN FOR LARGE HINDERED SETTLING SLURRY TANKS

By

Richard Kehn1, Graham Seal

2, Robert Stewart

3, Bernd Gigas

4

1. LIGHTNIN – An SPX Brand, Rochester, New York USA 2. SPX Process Equipment (LIGHTNIN), Sydney, Australia 3. SPX Process Equipment (LIGHTNIN), Sydney, Australia 4. LIGHTNIN – An SPX Brand, Rochester, New York USA

TABLE OF CONTENTS:

1. ABSTRACT – Page 2

2. INTRODUCTION – Page 3

3. SLURRY RHEOLOGY – Page 5

4. AGITATOR DESIGN – PROCESS DESIGN CONSIDERATIONS – Page 9

5. AGITATOR DESIGN – MECHANICAL DESIGN CONSIDERATIONS – Page 10

6. TANK DESIGN CONSIDERATIONS – Page 17

7. CONCLUSIONS – Page 19

8. REFERENCES – Page 19


Abstract: Designing agitators for large storage tanks has always been a challenge to agitator vendors. Almost all processing plants for mines utilize storage vessels that require agitation. Additional markets such as pulp and paper, power plants and water treatment plants use high weight percent slurries for their processes as well. The level of suspension (whether it be off bottom suspension or uniform suspension) is dictated by the process requirements downstream from the vessel. The required impeller design and agitator power level is not only based upon the settling characteristics of the slurry but also based on the viscosity level, as most slurries above 50% wt. typically exhibit non-Newtonian viscosity behavior. Most slurries contain a wide range of particle sizes and thus the particle interaction becomes important as well. If the application is viscosity controlled, the impeller style and number of impellers chosen for a given application becomes important. Lab scale mixing studies are an excellent way to predict the required agitator design for the full scale and lead to an optimum design. Power level and impeller selection are not the only parameters that are important to the success of the installation: proper mechanical design and proper tank design are also required to yield a successful installation. The purpose of this paper is to draw key information from known sources and experience and discuss the key process and mechanical parameters that will directly affect agitator design for large hindered settling slurry tanks.


Introduction: There has been quite a bit of work completed regarding free settling solids and the mixing levels required. One of the earlier known works on this was completed by Zwietering, where the “just suspended” impeller speed was correlated to settling velocity, particle size and densities and impeller size and type (p. 558 Handbook of Industrial Mixing). This correlation and the agitator sizing procedures that have ensued are best applied to free settling solid-liquid mixtures where the particles settle in a random order. The problem agitator vendors run into with high weight percent mining slurries processed is they do no exhibit free settling behavior. Most slurries fall into the “hindered” category and have the following key characteristics:

� The particles in a typical slurry can vary quite a bit in size and shape. A typical particle size distribution is shown in Figure 1.

� The particles do interact and thus “hinder” each other as they settle. This reduces the expected settling velocity. Most slurries typically have very low settling velocities (<< 0.3 m/min).

� If a large amount of smaller particles exist (<50 microns) and the percent solids are high enough (>50%), the viscosity of the slurry can increase significantly above water (>> 1 cps).

Figure 1 – Particle Size Distribution (Iron Ore)

Regardless of the ore type, these slurries are typically referred to as “hindered” slurries (p. 553 Handbook of Industrial Mixing). Prior to agitator design, it is required to determine whether the slurry will be free settling or hindered. There are a number of correlations that can be used to estimate settling velocity of a single free settling particle. However as stated previously, the settling velocity of a single particle does not predict what occurs with a hindered slurry and thus the settling velocity will be different than predicted for a single particle. Empirical correlations exist that can be used to predict the actual settling velocity of a hindered particle, one being the following:

n

tts VV )1( χ−= (p. 553 Handbook of Industrial Mixing)


where Vts is the hindered settling velocity of the particle, Vt is the free settling velocity of the particle, χ is the volume fraction of solids in the slurry and n is based on the Reynolds number of

the particle.

µ

ρ ptliquid

particle

dV=Re (p. 551 Handbook of Industrial Mixing)

where ρ is the liquid density, Vt is the free settling velocity of the particle, dp is the particle size

(diameter) and µ is the viscosity of the fluid.

As reported by Oldshue, when the settling velocity of the particles is less than 1 fpm, the slurry is generally considered hindered. A graphical correlation that is used to determine whether a slurry is free settling or hindered is shown in Figure 2 (p. 116 Fluid Mixing Technology). This correlation relates specific gravity of the solid/liquid medium to percent solids to predict whether the slurry will be free or hindered.

Figure 2 – Hindered Settling vs. Free Settling Determination


Slurry Rheology: Prior to agitator design, it is important to fully understand some key physical properties of the slurry. These include the following:

� Particle Density (i.e. solid specific gravity) � Liquid Density (typically water) � Percent Solids (weight percent is typical) � Particle Size Distribution � Slurry Viscosity

Particle/Liquid Densities: The need for data on these properties prior to agitator selection is obvious. These values will directly affect the settling velocity of the particles and thus will affect whether the slurry is hindered or free settling. These properties are easily obtainable. Percent Solids: This property is typically expressed as a weight percent and is simply the following:

LiquidSolids

Solids

WeightWeight

WeightWt

+

=%

This is an important property as this will tell the agitator vendor in part whether the slurry will be hindered or free settling. It’s very important to understand how this will fluctuate during operation as it can affect agitator design. In Figure 3 below, the general effect of weight percent solids on agitator power required is shown (p. 120 Fluid Mixing Technology).

0 10 20 30 40 50 60 70 80 90

% wt. Solids

Ag

itato

r kW

Figure 3 – Agitator Power vs. % wt. Solids

The effect of hindered settling on agitator power is graphically depicted between 40-70% solids in Figure 3. The value of the weight percent solids at the transition points varies from slurry to slurry. In general, the point at which the slurry becomes hindered will vary between 35-50% solids and the point at which the power requirement significantly increases as weight percent increases is somewhere between 65-75% solids.


An agitator that will operate near these transition points must be carefully designed as the requirements for mixing in each zone are quite different from one another. As weight percent increases from 40 to 60% in Figure 3, the slurry becomes more hindered and the power required to achieve a given level of suspension decreases. If the agitator is designed to operate at a weight percent above the first transition point yet can operate below this point, it is possible that the slurry will become free settling and the power required may increase. If the agitator is to operate near the second transition point (around 65-70%), a very small change in weight percent solids could lead to a very large increase in power required, due to the substantial increase in viscosity Particle Size Distribution: The particle size is typically known prior to agitator design, although a single particle size is not enough to properly design the agitator for suspension duty. The best data to obtain is a particle size distribution such as that shown in Figure 4.

Figure 4 – Comparison of Two Particle Size Distributions

The distributions shown in Figure 4 are of two different slurries for different mines. Both are at the same weight percent solids and solid specific gravity and thus the same slurry specific gravity. Although these distributions have similar P80 values, the one slurry (right) has a fair amount of small particles or fines versus the second slurry. Therefore with just percent solids, specific gravity and a P80 value, the same agitator selection would be predicted for identical tanks utilizing each slurry. However the slurry with the high level of fines has a far higher viscosity than the slurry with a more uniform particle size distribution and thus the agitator selections are quite different for each. Particle distributions can easily be generated for any slurry with a very small sample utilizing current technology (Figure 5).

Figure 5 – Particle Size Analyzer (Microtrac)


Slurry Viscosity: This property is the one that can be the most difficult to obtain. The slurry viscosity drives not only the power level selection but also affects the number of impellers required and the style of impeller chosen. When a fluid is continuously deformed under the action of a shear stress, the shear rate (du/dy) and shear stress (τ ) are related to each other as follows:

y

u

y

u

y

uk

n

∂

∂=

∂

∂

∂

∂=

−

µτ

1

where µ is the viscosity of the fluid (p. 29 Introduction to Fluid Mechanics). A fluid is said to be

non-Newtonian when the value of n is not 1 and thus viscosity will vary with shear rate. The majority of slurries requiring agitation are non-Newtonian in nature and therefore it is important to obtain apparent viscosity versus shear rate data prior to agitator design. There exists different types of non-Newtonian fluids and the most common to mining slurries are shown in Figure 6 (pp. 29-30 Introduction to Fluid Mechanics):

Figure 6 – Viscosity Vs. Shear Rate

� Pseudoplastic: Apparent viscosity decreases with increasing shear rate (n<1). Most hindered slurries fall into this category.

� Yield Stress Fluid: In this case, the viscosity changes with shear only after a specific shear stress or yield stress has been achieved. Pulp slurries, clay slurries and some mineral slurries act as Yield Stress Fluids.

10

100

1000

10000

0.1 1 10 100 1000

Shear Rate [s-1]

Vis

co

cit

y[c

P]

Newtonian

Pseudoplastic

Yield Stress Fluid

10

100

1000

10000

0.1 1 10 100 1000

Shear Rate [s-1]

Vis

co

cit

y[c

P]

Newtonian

Pseudoplastic

Yield Stress Fluid


Viscosity data is valuable to the agitator vendor. The magnitude of the viscosity and the slope of the viscosity versus shear rate curve will affect the type of impeller and number of impellers required for a given application. Viscosity versus shear rate data is typically generated during rheology studies of ore bodies and thus should be available. This data can easily be estimated in the laboratory using a small sample of actual slurry and a Brookfield Viscometer (see Figure 7) provided the slurry does not settle too quickly.

Figure 7 – Brookfield Viscometer

Many hindered slurries exhibit some level of viscosity; viscosity levels below 2000 cps at 2.3 s

-1

are typical for many slurries. When the viscosity level climbs above this value, multiple impellers need to be considered to produce complete motion throughout the vessel. At viscosities approaching 10,000 cps, impeller type becomes very important and high solidity impellers / close clearance impellers need to be considered.


Agitator Selection - Process Design Considerations:

Off Bottom Versus Uniform Suspension: In addition to fully defining the slurry’s rheology, it is important to determine the Customer’s requirements for suspension (off bottom versus uniform suspension). In off bottom suspension, all particles are moving above the bottom of the tank at some vertical velocity, with the larger particles moving at a lower vertical velocity than the finer particles. In uniform suspension, all particles are suspended relatively uniformly throughout the tank with the exception of the top one to six inches near the surface (p. 96 Fluid Mixing Technology). Uniform suspension is a relative term as there will be some variation in solids concentration within the tank. Basically, for the case of uniform suspension, further increase in power will not yield an appreciable improvement in solids concentration variability. For free settling solids, it is typically easy to distinguish a difference in off bottom versus uniform suspension (see Figure 8).

Figure 8 – Off Bottom vs. Uniform Suspension For hindered slurries, especially those that are viscous, the difference in uniform versus off bottom suspension is hard to distinguish. In those cases, the viscosity will drive the agitator selection and number of impellers rather than the level of suspension. Typically when bottom outlets are being specified, a single impeller for off bottom suspension is sufficient. In that case, a stable clear layer might exist at the very top of the batch during operation. If an overflow outlet is required, uniform suspension will be required and thus more than one impeller typically will be required. Many customers will not desire a clear layer during operation and more than one impeller will be required for cases of off bottom as well. Power Level and Impeller Choice: The majority of agitator designs for large slurry storage tanks are based on previous installation experience and laboratory testing of actual slurries. More recently, theoretical models concerning the zone of influence of the impeller in highly viscous shear thinning fluids (n = 0.3 or less) has been applied to hindered settling applications. Cavern theory applies when the viscosity level is driving the impeller selection as this theory was developed for non-Newtonian blending of a fluid. The goal is to determine the minimum impeller speed at a given impeller diameter to tank diameter ratio (D/T) to yield minimum motion at the tank wall. The height of the “cavern” formed will depend somewhat on impeller type selected but typically is in the 0.4-0.6 height to tank diameter range. Thus to achieve full tank motion, it is best to consider multiple impellers if the height to tank diameter ratio exceeds these values. This theory only predicts bulk fluid motion within the vessel and does not address the solids suspension portion of the application (pp. 521-522 Handbook of Industrial Mixing).

Off Bottom UniformOff Bottom Uniform


The more common method of selecting agitators for hindered settling is to use existing installation data. Rules exist for many different applications that include power level required (power per unit volume), impeller style and impeller diameter ratio to tank diameter ratio. These rules are based on not only actual installation feedback but also prior lab testing performed. The majority of applications utilize high efficiency hydrofoil impellers to yield the lowest torque agitator selection (Figure 9).

Figure 9 – High Efficiency Hydrofoil Impeller (A310 – LIGHTNIN)

The LIGHTNIN A310 impeller is one of many high efficiency hydrofoils that exist in the market. The velocity profile below the impeller is purely axial in the turbulent regime (Reynolds numbers above 5000), which is ideal for solids suspension (source - High Efficiency Impeller for Slurry Storage). The Reynolds number (NRe) of an open impeller is defined as the following:

µ

ρ2

Re

NDN =

Where ρ is the slurry density, N is the impeller rotational speed, D is the impeller diameter and

µ is the impeller viscosity, which is determined from the viscosity versus shear rate data

available. All hydrofoil impellers produce an axial flow pattern at low power consumption, which makes them ideal for use in large slurry tank applications (Figure 10).

Figure 10 – Laser Scan A310 (LIGHTNIN)


The impeller diameter to tank diameter (D/T) range is typically 0.3-0.4 for most hindered settling slurry applications. When viscosities exceed 2000 cps at 2.3 s

-1 , large impeller diameter ratios

typically need to be used in order to ensure that motion at the tank wall occurs. The typical impeller diameter to tank diameter ratios in these cases are 0.4-0.5. If off bottom suspension is required, a single impeller may be used provided a stable clear layer is acceptable at the top of the batch. If this is not acceptable and/or uniform suspension is required, more than one impeller should be used. For viscous slurries (viscosities exceeding 2000 cps at 2.3 s

-1), more than one

impeller may be required especially if the slurry is highly non-Newtonian and thus the n value is very low (<0.5). In these cases, the slurry is more than likely a yield stress fluid and in areas far from the impeller (outside the “cavern” zone), the slurry will become too viscous and won’t be able to be easily reincorporated into slurry form by the impeller. Many of these large vessels feed equipment downstream that require uniform solids feeding them in order to operate properly. In cases where the viscosity of the slurry is very high (>10,000 cps at 2.3 s

-1 ), the impeller flow

regime will more than likely be in the transitional range (Nre < 1000). High efficiency hydrofoil impellers do not produce axial flow at such low Reynolds numbers and thus either high solidity impellers or close clearance impellers need to be considered (see Figure 11).

Figure 11 – High Solidity Open Impeller and Close Clearance Impeller (A320 and A620 – LIGHTNIN)

The overall power required to achieve a given process result can be scaled from existing installations by scaling the installed power by the difference in volumes between the two installations. Impeller power draw is defined as the following:

53DNNHP pρ=

Where Np is the power number for the impeller and is a function of not only impeller type but also the Reynolds number and additional impeller proximity effects within the tank (e.g. off bottom distance / impeller diameter ratio). When scaling from one volume to another, the volume ratio is raised to an exponent, which needs to be selected based on experience. The exponent is less than 1.0 and varies based on application. This method of scale up yields a conservative agitator design. The one criteria that must be followed with this method is the geometry of each case must be the same. This means the tank geometric factors (e.g. tank height to tank diameter ratio or Z/T) must remain the same and the impeller style and geometry must remain the same between the two cases (e.g. impeller diameter to tank diameter ratio or D/T, number of impellers, impeller off bottom to diameter ratio (C/D) and impeller spacing to diameter ratio (S/D)).

High ViscosityOpen Impeller

Transitional FlowRegime Impeller


If a reference installation does not exist, laboratory work needs to be completed. A typical lab set up is shown in Figure 12.

Figure 12 – Typical Test Tank / Lab Set Up

One issue with scaling from the full scale to the lab scale is the Reynolds number on the lab scale will be lower than in the full scale. In cases where the slurry is viscous, the Reynolds number in the lab scale might be lower than 1000 and thus the flow regime in the lab scale will be transitional versus turbulent. Therefore the fluid motion observed in the lab cannot be used as an accurate prediction of the fluid motion in the full scale. This is where the scale up exponent and the minimum required power level in the lab must be carefully chosen. Experience shows that testing in vessels that are much smaller than 17.5” diameter will not lead to meaningful results. Low Level Mixing – Process Design Considerations: Many large slurry tanks will operate at various liquid levels. The range of operation needs to be addressed up front during the design stage as it will affect impeller design. In general, an axial flow impeller requires at least half a diameter of coverage in order to pump properly. This becomes especially important when draining a large slurry tank as in most cases, the lower impeller will not provide mixing for the very last amount of slurry in the vessel during draw off. A kicker impeller can be added to the design in order to provide some level of fluid motion near the bottom of the tank. Typically this design, although providing motion, does not provide sufficient solids suspension and thus larger particles may drop out during the draw off process.


Agitator Selection – Mechanical Design Considerations:

As with all agitators, large slurry tank agitators require mechanical design checks with regards to torque, thrust, bending moment and critical speed. There are also additional special design considerations that need to be considered for these applications. These will be discussed below. Torque: Torque is defined as the tendency of a force to cause an object to rotate about an axis. With regards to agitator design, torque is defined as the following:

N

HP=τ

Torque is expressed in either in-lbs. or N-m. The magnitude of torque will determine the gearbox selection. Therefore, it is often desirable to test various impeller diameters and styles in order to find the lowest torque solution, which will yield the lowest cost agitator. If critical speed and bending are not issues and there is not significant viscosity to the slurry, a smaller impeller diameter to tank diameter ratio should be explored in order to reduce torque and optimize gearbox selection. Thrust: Thrust is the upward load imparted on the agitator for a down pumping impeller. Thrust is expressed in lbs. or N. Although typically not a limiting factor on agitator design for open tanks, the gearbox must be able to handle the upward load (i.e. the thrust bearing must be properly sized for this expected load). Bending Moment: A bending moment is imparted on the shaft support bearings and thus gears when fluid forces act upon the impeller. Fluid force acting on a rotating impeller for an agitator can be expressed by the following equation (p. 363 Fluid Mixing Technology):

42DNNF F ρ=

where NF is a force number that is dependent on the impeller type. Pilot lab testing using an agitator on load cells can be used to determine the fluid force equation for any impeller. The diameter of the impeller has a strong influence on the magnitude of the fluid force. Therefore minimizing impeller diameter will always assist in reducing fluid force and may help optimize agitator design. The bending moment will be affected by the magnitude of the fluid force and the distance from the impeller to the bearings. The longer the distance from the impeller to the bearings, the higher the bending moment will be. Therefore, tank geometry does become important, especially for large tanks. Typically high aspect ratio vessels (Z/T > 1.2) will require a more costly agitator design versus a square batch vessel (0.9 < Z/T < 1.1) for the following reasons:

� A high Z/T will typically require more than one impeller if an upper clear layer is not required / acceptable.

� A high Z/T will have smaller diameter impellers and a higher operational speed than a tank with a lower Z/T. This could lead to a critical speed problem.

Agitator suppliers have designed gearboxes to handle agitator fluid force loads. Therefore when considering different options, gearbox design and the size of the low speed bearings (that support the shaft/impeller system) become very important. Agitator vendors design gearboxes to handle a certain level of bending moment during normal operation. It is important to note that many commercial gearboxes available are not designed for external agitator loads and thus are torque transmitters only. In those cases, external bearing members typically need to be utilized in order to prevent the fluid forces from being imparted directly on the gearbox itself. This typically yields


a more costly design versus considering agitator gearboxes that are specifically designed to handle agitator external loads. One way to eliminate a large portion of the bending moment is to utilize in-tank bearings, or steady bearings. The issue with this in large slurry tank design is the slurries are almost always highly abrasive and the life of the steady bearing bushing would be very short. Alternatively, a flush line can be used on the steady bearing components. However, most plants want to avoid this due to the maintenance requirements. Therefore in large slurry tank designs, steady bearings are typically not used. Critical Speed: The critical speed of a component is the speed at which the frequency of the rotating component equals the natural vibration frequency of the component (p. 408 Fluid Mixing Technology). Also known as a natural frequency of a rotating system, the first lateral natural frequency is what we are concerned with in terms of agitator shaft design. The standard vibration equation can be applied to agitator shaft design and is as follows:

)(2

2

tfkxdt

dxc

dt

xdm v =++

where m is the mass, cv is the damping coefficient of the system, k is the effective spring constant and f(t) is an external force (p. 1295 Handbook of Industrial Mixing). In general, it is accepted that agitator actual speed is to be less than the critical speed by ~20% for large agitators. This comparison of the actual operating speed to the first lateral critical speed is called the critical speed ratio:

Critical

ActualCR

N

NN =

Critical speed comes into consideration on all agitator systems. When the NCR is greater than 0.4, there is a need to consider impeller stabilizers. Impeller stabilizers or fins are extensions attached to the bottom of each blade of an impeller in a vertical direction (see Figure 13).

Figure 13 – Stabilizer Fins

The purpose of adding the stabilizers is to dampen the effect of unbalanced fluid forces when an impeller is operating near draw off (p. 1296 Handbook of Industrial Mixing).


Low Level Mixing – Mechanical Design Considerations: When an agitator impeller operates at draw off, fluid forces increase and thus the bending moment imparted on the gearbox will increase. If low coverage can occur for a long period of time (very common for upper impellers on large slurry tanks), then an elevated level of bending moment shall be considered for the design. Typically elevated fluid forces will occur on an axial flow impeller when the coverage is approaching half a diameter of coverage or less. The added bending moment during this condition can lead to gearbox failure if not accounted for up front in the design. Start Up in Settled Solids: At the design phase of a large slurry tank agitator, most customers express the need to consider start up after a power outage and after the slurry has “settled.” The mechanical design needs to be robust enough to prevent damage during an attempt at start up. The added loads at start up can easily cause severe damage to bearing and gear components within the agitator gearbox that may not be apparent for some time afterwards (p. 380-381 Fluid Mixing Technology). Typically there is also a requirement for the agitator to be able to re-suspend the tank’s contents in the event a power outage occurs. The time frame for the power outage will range based on the specification but can be anywhere from a few minutes to a day. The first item to address is how does the slurry behave after it has settled for a period of time. Many gold slurries, for example, do not agglomerate as they settle and thus after a period of time, the impeller(s) do not sand in and thus start up easily. High concentrated Iron Ore and Nickel slurries can agglomerate and thus precautions need to be taken in order to prevent damage to the equipment upon start up. The best way to determine whether start up will be a problem is to simulate start up in the lab scale. If the slurry compacts rather quickly in the lab scale, it will more than likely compact quickly in the full scale as well. If the impeller is sanded in, the following pre-cautions should be taken regarding the mechanical design (pp. 381-382 Fluid Mixing Technology):

� A method of lancing air around the impeller prior to start up to loosen up the solids should be considered.

� Some style of torque limiting coupling should be installed between the motor and gearbox. This will not only cushion the shock loads experienced but will also eliminate the high torque loads. A typical torque-limiting coupling is shown in Figure 14. A fluid style coupling (not shown) can also be considered in addition to a soft starter for the drive motor.

Figure 14 – Torque Limiting Coupling (Falk Corporation)

� The impeller, shaft and gear drive needs to be able to safely deliver the torque required

to break the impeller free upon start up. This means designing the agitator system for more torque than the full load operating torque under normal conditions. This means thicker impeller blades, larger bolts, larger diameter shafts, etc. This will increase the cost of the agitator.


� Include a limit ring underneath the impeller. This is a ring around the shaft that restricts

shaft deflection when the shaft breaks free from the settled solids (see Figure 15). It does not protect the drive and motor from shock overload.

Figure 15 – Limit Ring – Supports not Shown

Designing for start up in settled solids only guarantees that the mixer will not have a severe mechanical failure during start up. Re-suspending the solids may or may not be possible. This will depend on the power level and nature of the slurry. If there is a significant yield stress to overcome and the agitator is not designed properly, full re-suspension may not occur. This again is best simulated on the lab scale. The minimum power level determined in the lab for suspension and blending may not be enough power to re-suspend the slurry after it has had a chance to settle for a period of time.


Tank Design Considerations:

Improper tank design can adversely affect the agitator’s performance even when the agitator process and mechanical design is carried out correctly. Some key items to consider in the design phase are as follows: Tank Geometry: As mentioned earlier, tank geometry will drive the number of impellers required for a given application. Typically the ideal Z/T for a solids suspension for a large slurry tank will be 0.8 to 1.1. At higher Z/T’s, the need for more than one impeller becomes apparent, especially if an upper clear layer is not desired and/or if uniform suspension is required. Multiple impellers, although feasible, typically will lead to higher bending moments and issues meeting critical speed requirements. The tank bottom geometry is also very important. Although not practical for large tanks, ASME dish bottoms are the preferred design for solids suspensions and result in a lower power requirement over flat bottomed tanks, especially for uniform suspensions (p. 577 Handbook of Industrial Mixing). The majority of tanks used are flat bottomed tanks. In large slurry tanks, cone bottomed tanks are rarely used (Figure 16).

Figure 16 – Different Tank Designs

The main issue with a flat bottomed tank is the formation of fillets. A fillet is a stagnant collection of solids, which normally forms in the corners of a flat bottomed tank. Since the power required to eliminate the fillet is normally not practical, it is allowed to form. If the fillet remains stable, there typically is no issue (source - New Concepts in Slurry Storage). When determining minimum power required from either an existing installation or lab testing, it is important to note the fillet size and whether it is a stable fillet or not. A typical fillet ratio is 3-10%, where the fillet ratio is the fillet size divided by the tank diameter.

Cone Flat Dish


Tank Baffles: Baffles are required in large slurry tanks in order to prevent tangential swirl and provide a top to bottom flow pattern when the viscosity of the slurry is low. Figure 17 shows the expected flow pattern with and without baffles utilizing a single axial flow impeller.

Figure 17 – Fluid Flow with and without Baffles

Standard baffles for low viscosity slurries is four vertical baffles spaced 90 degrees apart with a width one twelfth that of the tank diameter (p. 15 Fluid Mixing Technology). One advantage of high efficiency impellers such as the LIGHTNIN A310 is the low torque requirement for a given pumping rate allows for less baffles to be used, in most cases three versus four (source - High Efficiency Impeller For Slurry Storage). The viscosity level of the slurry oftentimes dictates using baffles that are “non-standard.” As viscosity increases, the width of the baffle needs to be reduced in order to reduce the drag at the tank wall. Figure 18 shows the effect of batch viscosity on baffle width.

Figure 18 – Viscosity Effect on Baffle Width

Since most slurries are non-Newtonian, it is best to estimate the viscosity at the tank wall as 0.23 s

-1 and use this value to estimate the baffle width. It is assumed that the highest shear rates

occur near the impeller and the lowest shear rates near the wall (pp. 55-56 Fluid Mixing Technology). If the tank is overbaffled, the slurry can become stagnant near the tank wall even with a properly designed agitator. This stagnant zone can build over time and cause process problems. In addition to baffle width, it is important to maintain an off wall distance and an off bottom distance for the baffles. By maintaining an off wall distance, slurry is allowed to flow freely behind the baffles and not build up in the corner of the baffles. The typical off wall distance should be 1/36 of the tank diameter. Off bottom distance from the tank floor should be one to two feet in order to create some swirl near the bottom of the tank and help reduce fillet formation. As long as the main impeller is fully baffled, top to bottom turnover will occur and thus solids suspension will

w/o Baffles With Baffles


not suffer. Although not required, cutting baffles below the maximum liquid level will help to produce additional surface motion during operation, especially on more viscous slurries. Although this is not required from a pure mixing standpoint, surface motion is often desirable at the plant. Outlet Locations: The process requirements downstream dictate the outlet location for a given vessel. Most storage vessels have bottom outlets and top feed points in order to prevent short circuiting and maximize particle dwell time in the vessel, especially if the process is continuous. Some vessels such as leach vessels in a gold processing plant require overflow outlets. For bottom outlets, the pump location should be located on the higher velocity side of the tank baffle. Typically an agitator designed for off bottom suspension is sufficient for proper system operation. If the slurry is highly concentrated and highly pseudoplastic, it may be necessary to consider an upper impeller to keep the slurry fluidized in order to prevent plugging during pump out when the tank is emptied. For overflow outlets, care must be taken to insure that the agitator design achieves uniform suspension. The best way to handle an overflow outlet is to make use of a riser pipe. The riser pipe is extended from the overflow location down into the vessel. It is designed to yield a pipe velocity that significantly exceeds the highest settling velocity of the largest particles within the tank. This insures the pipe will allow even the heaviest solids to leave the vessel. In this case, the agitator can be designed for off bottom suspension, thus saving valuable power (Figure 19).

Figure 19 – Depiction of Riser Pipe

The riser pipe should be terminated just above the location of the lower impeller.

Outlet

Suspended Solids Level


Mixer Mounting: The majority of installations for large slurry tanks are open top with a bridge agitator mounting or beam mounting. A typical beam mounting arrangement is shown in Figure 20).

Figure 20 – Beam Design

To handle dynamic agitator loads, the beam structure must be stiff in all directions. The two main I beams must be cross-braced to prevent movement during operation. The spacing between the beams should be 15% of the tank diameter or less to achieve a stiff design (p. 1312 Handbook of Industrial Mixing). The agitator vendor must provide the end user / tank vendor with design loads (torque, bending moment, vertical downward load) based on the actual agitator design with a sufficient safety factor. Typically large slurry tanks are open and thus a sealing arrangement is not required.


Conclusions:

A successful agitator installation for a large slurry tank design requires a full understanding of the rheology of the slurry prior to design, most importantly viscosity. The viscosity affects not only the number of impellers and diameter but also the style of impeller chosen. Existing installations and modern theory can be used to arrive at an agitator design that will achieve the desired process results. Lab work is the best way to determine the rheology of the slurry and the optimum impeller configuration and power level required to achieve the desired process result. Agitator mechanical design considerations, including a full understanding of whether start up in settled solids needs to be considered, are very important to insure the agitator design has a long running life with little down time. Even if the agitator is properly designed, a poor tank design, baffle design and mounting structure can lead to run-ability issues. The tank design can have a direct impact on both the process and mechanical selection so both need to be carried out concurrently in order to yield an optimum design that operates with little down time.

References:

V. Atiemo-Obeng, S. Kresta, E. Paul., “Handbook of Industrial Mixing – Science and Practice”, John Wiley & Sons, Inc., Hoboken, New Jersey, 2004. J. Oldsue., “Fluid Mixing Technology”, McGraw-Hill Publications Co., New York, New York, 1983. R. Fox, A. McDonald., “Introduction to Fluid Mechanics – 4

th Edition”, John Wiley & Sons, Inc.,

Hoboken, New Jersey, 1992. A. Olderstein, J. Pharamond., “New Concepts in Slurry Storage – A Technical Reprint”, Presented at The Fifth International Technical Conference on Slurry Transportation, March 1980. C. Coyle, et al., “High Efficiency Impeller for Slurry Storage – A Technical Reprint”, Presented at The Fifth International Technical Conference on Slurry Transportation, March 1983.


THE EVOLUTION OF THICKENERS

By

Ronald Klepper

FLSmidth Minerals

Presented by

Ron Klepper [email protected]

CONTENTS

1. INTRODUCTION 2

2. SIZING AND SELECTION 3

3. CONVENTIONAL THICKENERS 3

4. HIGH RATE THICKENERS 5

5. HIGH DENSITY AND PASTE THICKENERS 6

6. CONCLUSIONS 12

7. REFERENCES 12


2

1. INTRODUCTION

This paper is a review of the evolution of thickeners used in hydrometallurgy and will define and explain differences between the various thickener types discussed in feasibility studies and in sales literature. Conventional, High Capacity, High Rate, High Compression, High Density, Ultra-High Rate, Ultra High Density, Deep Cone and Paste thickeners are names used by more than 20 thickener suppliers worldwide. All names will be classified into four thickener types to simplify the discussion.

Gravity thickeners and clarifiers are two of the major unit operations in hydrometallurgy flowsheets, since dissolution and precipitation are the primary physical chemical characteristics in all hydrometallurgical processes. The solid/liquid separation unit operation is a thickener when suspended solids concentration in the underflow slurry is most important. Or the solid/liquid separation unit operation is a clarifier when suspended solids concentration in the overflow stream is most important. The solid/liquid separation unit operation is a thickener/clarifier when both underflow and overflow qualities are important and by coincidence or due to budget restraints both suspended solids concentration objectives are assigned to one device.

Typical hydrometallurgical thickener applications are:

• Pre-leach thickening to minimize water in leach feed slurry minimizing acid dilution or optimizing energy requirements,

• Counter current decantation (CCD) thickening to separate and wash soluble metals from leach residue producing pregnant liquor solution with the first stage CCD thickener/clarifier required to remove suspended solids to clarify the pregnant liquor solution and the washed tailings in the last CCD thickener prepared for neutralization and storage,

• Impurities removal to separate precipitated impurities or undesired metals for disposal and in some cases for precipitated products to minimize suspended solids in overflow liquor so the product is not contaminated,

• Intermediate product recovery to concentrate mixed sulfides or hydroxides in the underflow slurry for filtration and ensure capture by minimizing product in the overflow,

• Thickened tailings disposal of leach residue and impurities residues.

A question asked frequently is, “What are the differences between the different types of thickeners?” To understand thickener differences requires a general understanding of thickener evolution history.

Arguably, Dr. John Van Nostrand Dorr invented the first continuous process sedimentation thickener in about 1905. Slurry flowed into the top center of the thickener through a pipe and suspended solids settling by gravity, concentrating in the bottom as underflow and clarified water flowed through a peripheral launder near the top as overflow. The thickener consisted of a circular tank with scraper blades or rakes attached to radial arms slowly rotating in the bottom to move coarse solids to the center underflow discharge location.

The Dorr Continuous Thickener 1912 Advertisement Picture Figure 1.


3

A picture of the Dorr Continuous Thickener from the first advertisement in about 1912 is exhibited in Figure 1. This can be considered the beginning of convention thickeners. Numerous scientists and mathematicians studied and modeled the separation of suspended solids from liquids in an attempt to determine the full-scale capacity of thickeners from small “representative” core samples of ground ore.

2. SIZING AND SELECTION

In 1851 George Stokes derived an expression for friction drag of single particles pulled through a liquid by gravity, “Stokes Law” which is the simplest model of the settling rate of individual particles in gravity thickening, but this is flawed as particles interfere with each other.

However, in 1916 Coe & Clevenger defined four zones existing in both batch settling tests and in full-scale continuous conventional thickeners as exhibited in Figure 2. The zones are named;

I. The clear water zone,

II. The constant or free settling zone (particles not in contact),

III. The transition of hindered settling zone (particles in contact), and

IV. The compression zone (particles packed with limited permeability).

Coe & Clevenger proposed that the solids flux-density was a maximum value in the thickener at some concentration between the feed and discharge suspended solids concentration and thickener sizes were to be sized and selected using this rate.

I

II

III

IV

Settling Zones found In Thickeners Defined by Coe & Clevenger 1916

Figure 2.

In 1952 G.J. Kynch proposed that at any point in the suspended solids thickener bed the settling velocity is a function of only the local suspended solids concentration and the thickener size for a desired suspended solids concentration can be selected at a given solids flux. However this theory did not consider compression.

‘Talmage & Fitch’ and ‘Naide & Wilhelm’ proposed methods that determined the thickener underflow suspended solids concentration at a solids flux rate or unit area and added a correction for bed compression effects.

3. CONVENTIONAL THICKENERS

Conventional thickeners are thickeners that are sized using one of the methods described above or a thickener that does not use flocculants to form agglomerates of both fine and coarse particles that has a particle size distribution d80 much greater than that of the feed slurry. Conventional thickeners today are basically the same design conceived by Dr. Dorr more than a century ago.

Conventional thickener design prior to organic polymer or flocculant development was basically larger thickeners sized based on the settling rates of the smallest particles with rakes to move the coarse particles that settle faster into the center of the thickener for discharge as underflow slurry.

Figure 3 exhibits the approximate number of thickeners >90 m installed from 1950 to date. The reason for the reduction of large thickeners up to about 2000 is the settling rate of suspended solids


4

has significantly increased because of the use of flocculants and design changes in thickeners to optimize the flocculation process.

Thickeners >90m Diameter

In The Past 50 Years

0

1

2

3

4

5

6

7

8

9

1950 1960 1970 1980 1990 2000 2010

Year

Nu

mb

er

Ins

tall

ed

Large Thickener Frequency Figure 3

The increase of large thickeners after 2000 is either significant increases in throughput using new High Rate thickeners or retrofitting existing conventional thickeners with feed dilution systems to optimize flocculation and increase throughput or conversion to High Rate thickeners.

1st HRT-Supaflo

1st DCT-Alcan

1st HDT-EIMCO

>141 DCT & HDT

ALCAN + EIMCO

Moa Bay Nickel

1st HRT-Amstar

1st HRT-EIMCO

Thickener Evolutionary Timeline Figure 4

A time line of events affecting thickeners is exhibited in Figure 4. Simply, flocculant development led to High Rate Thickener development that led to Paste & High Density thickener development. Thickeners have evolved from conventional thickeners to predominately High Rate thickeners and arguably another evolution is in progress with changes from High Rate thickeners to High Density or Paste thickeners using current knowledge to produce underflow slurries with greater denser at greater throughput rates in smaller thickeners.


5

4. HIGH RATE THICKENERS

Organic polymers or flocculants began development in the 1950’s and were used to agglomerate fine particles into larger groups that settled faster. These polymers can be very complex with high molecule weights and designed to have surface charges either anionic, cationic or nonionic polarities with varying charge densities. It became apparent in the 1970’s that thickener capacity could be significantly influenced by the type and quantity of flocculent added into conventional thickeners.

The first high capacity thickeners were developed as thickener/clarifiers in the sugar industry to produce a clear sugar solution prior to crystallization. In 1970 the Enviro-Clear Company introduced the first High Capacity thickener patented by Amstar Corporation seen in Figure 5. One of the design and operating concepts was to introduce and mix flocculant into the feed stream in the feedwell and have the flocculated feed slurry exit into the hindered settling zone in an attempt to capture fines. This was significantly different than convention thickener feedwells where the feed slurry exited above the hindered settling zone.

In 1978 Envirotech/EIMCO introduced the “Hi-Cap” Thickener used in uranium CCD circuits and in coal refuse disposal. In 1983 the Supaflo® high rate thickener was invented as seen in the sketch in Figure 7. Delkor and others offered many types of High Rate or High Capacity Thickeners eventually becoming the new thickener state-of-the-art over conventional thickeners.

Enviro-Clear High Rate Thickener Figure 5.

The main consequence of high rate or high capacity thickeners were smaller thickeners designed to use flocculant added directly into the feed stream. The more flocculant used the smaller the thickener. Small thickeners were also easier to control automatically as the lag time created by large quantities of suspended solids within the thickener was reduced.

DELKOR High Rate Thickener Figure 6.

Controlling large conventional thickeners was nearly impossible, because first the inventory of suspended solids was very large creating an enormous lag time following any process change and secondly because sands or coarse particles typically separated from the fines and the fines tended to accumulate in the thickener as slimes the coarse passed through.

Smaller high rate or high capacity thickeners initially required the feed to enter the thickener into the hindered settling zone and is was necessary to maintain a sufficient inventory of suspended solids within the thickener to cover the feedwell outlet. Introducing the flocculated feed slurry into the hindered settling zone is no longer a requirement to obtain high settling rates.


6

Supaflo High Rate Thickener Genesis Figure 7.

5. HIGH DENSITY AND PASTE THICKENERS

In the 1970’s at least three organizations began the development of thickeners designed to use gravity and bed compression to produce as high a suspended solids concentration as possible in the underflow slurry. This was the beginning of the development of the paste thickener. The result of the efforts was the British National Coal Board’s development of the “Deep Cone Paste Thickener” seen in Figure 8. The BNCB installed numerous units at various collieries in an effort to thicken coal tailings to a “paste” consistency for direct deposit on to a conveyor to a coal refuse disposal site. Unfortunately none of these are in operation today.

British National Coal Board “Deep Cone Paste Thickener” Figure 8.

A few years later Alcan and Alcoa independently sought to maximize recovery of alumina and minimize caustic and water loss when disposing of aluminum laterite leach residue or red mud. Alcan identified that the outer portion of conventional washer/settlers were not active and concluded a smaller diameter and taller tank could be as affective. The first deep paste thickeners, seen in Figure 9, were 10m in diameter with a flat bottom and no rakes producing concentrations 70%


7

greater than a conventional thickener. Rakes were later added resulting in an additional 20% increase in suspended solids density of the underflow slurry.

Alcan’s First Deep Cone Paste Thickeners - Jamaica Figure 9.

Alcoa’s efforts resulted in a 90m diameter High Density thickener with a traction drive capable of producing 13,600 kNm torque or a K factor of 115 lbf/ft installed in 1987 at the Pinjarra Refinery in Australia, as seen in Figure 10. A consequence of this large thickener was significant pulp depth producing compression similar to the smaller Alcan deep paste thickeners

Alcoa 90m High Density Thickener – Australia Figure 10.

A key discovery to the success of producing paste was a phenomenon of “optimized flocculation”. Initially almost all High Capacity or High Rate thickener designs added flocculant into the feedwell where blending of the flocculant into the feed occurred simultaneously as floccules were formed and destroyed. The feed slurry exited the feedwells, extended into the hindered settling zone within the thickener, in an effort to repair damaged or destroyed floccules and capture fines.

Figures 11 exhibits examples of two different nickel laterite ores where a maximum settling flux is measured at about 5 wt% suspended solids concentration. However the thickener feed slurry was initially >15 wt%. Dilution of the feed slurry using thickener overflow solution is necessary to obtain optimum flocculation and suspended solids concentration to produce a maximum settling flux.


8

Settling Flux

vs. Suspended Solids Concentration

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0%

Diluted Feed Suspended Solids Concentraion (wt%)

Sett

lin

g F

lux (

t/h

/m2)

50 g/t Dose 60 g/t Dose 70 g/t Dose 80 g/t Dose

Settling Flux

vs. Suspended Solids Concentration

0.0

0.5

1.0

1.5

2.0

2.5

3.0

2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0% 11.0%

Diluted Feed Suspended Solids Concentration wt%

Sett

lin

g F

lux (

t/h

/m2)

20 g/t Dose 40 g/t Dose 60 g/t Dose 80 g/t Dose

Measured Maximum Solids Flux at Optimized Suspended Solids Concentration Figure 11.

The discovery that a maximum solids flux exists at some dilute suspended solids concentration in thickener feed slurry, led to a variety of dilution systems developed utilizing feed slurry kinetic energy, hydrostatic head differences in the feedwell and overflow solution, and simply pumping the required quantity of dilution liquor. Figure 12 exhibits a system using the kinetic energy of the feed slurry to dilute and obtain optimized flocculation and maximum solids flux.

Existing Conventional or High Rate Thickeners not operating with optimized flocculation can have feed dilution systems retro-fitted to either minimize flocculant consumption and/or increase suspended solids throughput.

FLSmidth Minerals E-Duc Feed Dilution System Figure 12.

The impact of optimizing flocculation can be seen in Figure 13 which represents a shift in size of the settling zones defined by Coe & Clevenger. With optimized flocculation the free settling (II) and hindered settling (III) zones are minimized allowing the compression zone (IV) to maximize. In conventional thickeners and in most high rate thickeners the compression zone is minimized by operations personnel to prevent plugging of the underflow. However High Density and Paste thickeners are designed to use the compressive forces created by mass of solids creating greater rates of consolidation. The combination of gravity and significant mass of solids creates a greater combined driving force to release water. This is the same technique used in soil compaction during road construction.


9

Settling Zone Change in High Density and Paste Thickeners Figure 13.

It is time to visually summarize the difference in underflow slurry characteristics produced in the various types of thickeners. Figure 14 exhibits a typical relationship of yield stress and underflow suspended solids concentration with suspended solids consolidation as the limit where all flow characteristics are no longer. The rate of consolidation is different for all slurries depending on physical chemical characteristics such as densities and particle size distribution.

Thickener Definition by Underflow Suspended Solids Concentration vs. Rheology Figure 14.

• Conventional thickeners (CT) typically are sized based on the settling flux rate of the smallest particle and segregation of small and large particles is the norm. The underflow slurries produced are at a suspended solids concentration typically exhibiting Newtonian rheology. Typically the maximum suspended solids concentration in the underflow slurry is defined by operating at less than maximum rake drive torque or at underflow concentrations that are dischargeable to prevent plugging. Typically dischargeable underflow slurry exhibits a yield stress less than 20 Pascal (Pa).

• High Rate or High Capacity thickeners (HRT) are sized using flocculants to produce a underflow slurry with minimal particle segregation. The underflow slurries produced are at suspended solids concentrations typically exhibiting Newtonian rheology. The maximum


10

suspended solids concentration in the underflow slurry is defined by the rake drive torque and the ability to discharge. Excess flocculant must be avoided to prevent forming a donut shaped mass of viscous slurry that adheres to and rotates with the rakes arms. This viscous mass blocks the flow or movement of thickened suspended solids to the center discharge, eventually filling the thickener with suspended solids. Large amounts of structural steel to support the rakes produce resistance and excess flocculant combines to create this operating problem. Typically the underflow slurry yield stress needs to be less than 20 Pa to avoid discharge problems.

• High Density thickeners (HDT) are sized based on optimized flocculation or maximum solids flux rate. HDT’s are designed to produce a slurry that can be pumped by a centrifugal pump and typically have all of the physical design features of a paste thickener but simply are controlled to produce an underflow slurry exhibiting a yield stress of about 100 Pa.

• Paste or Deep Cone Paste thickeners (DCPT) are sized based on optimized flocculation or maximum solids flux rate and the compression residence time required to obtain the desired suspended solids concentration. DCPT are designed to produce and discharge underflow slurries exhibiting yield stresses of less than 500 Pa. The most viscous paste is typically produced for underground mine backfill where cement is added for strength. Paste thickeners are predominately used for thickened tailings disposal with the consistency of the underflow slurry determined and controlled to meet design criteria of the disposal site. The particle size distribution required to produce paste must have a minimum of 20wt% particles less than 20 microns as the fine particles prevent consolidation.

Figure 15 exhibits drawings to scale of different thickener types sized as CCD thickeners at the same throughput for a recent hydromet project. The Conventional and High Rate thickeners were an on ground tank designs producing a underflow slurry suspended solids concentration of ~50 wt% and the High Density thickener was an elevated tank design with positive leak detection producing a underflow slurry suspended solids concentration of >60 wt%.

80 m Diameter

Conventional Thickener (CT)

55 m Diameter

High Capacity Thickener (HRT)

40m Diameter

High Density Thickener (HDT)

Conventional, High Rate & High Density Thickeners Figure 15.

There are numerous design features in High Density and Paste thickeners to produce and discharge paste from a Paste thickener as seen in Figure 16.


11

Figure 16.

• The thickener aspect ratio of height to diameter is much greater than Conventional or High Rate thickeners to produce adequate compressive forces as seen in Figure 16.

Paste Thickeners and Conventional Thickeners in CCD Alumina Red Mud Washing Circuits Figure 16.

• Optimized flocculation by diluting the feed slurry to the optimum suspended solids concentration is mandatory to maximize the solids settling flux rate.

• Feedwell design has been optimized using CFD modeling to totally change the direction of the feed slurry from horizontal to vertical minimizing damage or destruction of floc structure.

• Rakes and support arms are designed to minimize cross-sectional area that produces drag or torque and minimize obstruction at the center of the thickener to allow slurry to flow out of the thickener, yet still have structural strength to rotate at unit torques 10-20 times greater than used in Conventional or High Rate thickeners.

• Rakes are also designed to shear the paste so gravitational force can pull the paste down the 30

o or 45

o sloped bottom. The angle of repose formed by the sloped

bottom of the thickener floor are similar to piles of ore.

• Pickets are also designed to attach to the rotating rakes to travel through the compression zone creating channels allowing the release of liquid from deep within the compression zone that would not have exited the zone due to low permeability.

• Underflow slurry exits through a discharge cylinder rather than a discharge cone which is larger in volume and more open than in other thickener types. The mass of the contents above act as a piston creating adequate net positive suction head for discharge pump with minimal suction pipe. The discharge cylinder can also be


12

used to temporarily reduce the underflow slurry yield stress by shear thinning with a centrifugal pump.

• Automatic control of optimum feed suspended solids concentration, flocculant dose and mass balance of suspended solids inventory is mandatory. A typical control P*ID is exhibited in Figure 18.

Dry Flocculant

Water

Figure 18.

6. CONCLUSIONS

There was a general evolution from Conventional thickeners to High Rate thickeners starting in the early 1970’s as a result of flocculants enhancing the settling rates of suspended solids. The hydrometallurgical industry is in the midst of another evolution from High Rate Thickeners to High Density or Paste thickeners as a consequence of optimizing flocculation and pioneering work by Alcan and the British Coal Board demonstrating the affects of compression and means to overcome limitations of reduced permeability of the concentrating suspended solids. Paste technology is well developed and currently available in sizes of 60m diameter for HDT’s and 40m diameter for DCT’s. The process and financial benefits of using advanced thickener technology in hydrometallurgy are improved soluble metal recovery, using less water with less capital and operating costs.

7. REFERENCES

Caransa, A., 1993, “Experience Has No Substitute: A Sixty-Year History of Dorr-Oliver in the Netherlands”, pp. 8-11.

Coe, H.S., and Clevenger, G.H., 1916, “Methods for Determining the Capacities of Slime Settling Tanks”, Trans, AIME, pp. 55, 356-384.

Delkor Web Site

Enviro-Clear Web Site

Jewell, R.J., and Fourie, A.B., 2006, “Paste and Thickened Tailings – A Guide”, second edition, pp 69-122.

Kynch, G.J., 1952, “A Theory of Sedimentation”, Transactions Faraday Society, pp. 48, 166.

Schoenbrunn, F.R. “A Short History of Deep Cone Thickener Development”, Proceedings of the Tenth International Seminar on Paste and Thickened Tailings, pp. 51-55

Supaflo Web Site

Wilhelm, J.H., and Naide, Y., 1978, “Sizing and Operating Continuous Thickeners”, SME preprint 79-30 SME-AIME Annual Meeting.


1

EXPLORING SYNERGIES IN WESTERN AUSTRALIAN NICKEL PROJECTS

By

H.W. Scriba, S. Spencer and J. Connor

SNC-Lavalin Australia

Presented by

Hermann Scriba

[email protected]


2

Disclaimer

Information contained in this presentation is provided in good faith based on the experiences of the authors. Opinions expressed are those of the

authors’ and not those of SNC-Lavalin. Cost indications have been provided for comparative

purposes only. The information should not be used to evaluate projects or as a basis to make any commercial decisions. SNC-Lavalin disclaims any

liability against the use of the data and information contained in this presentation for whatever purpose.


3

Contents

� SNC-Lavalin and Nickel Projects

� Current Situation in Western Australia

� Synergies, and their potential

� Drivers in Nickel Laterite and Nickel Sulphide Projects

� Synergistic Processing Opportunities

� Methodology for further exploration and realising of

Synergistic Opportunities

� Conclusions


4

Theme

� Western Australia is unique in having both Nickel Sulphide and Nickel Laterite Deposits in close

proximity to each other.

� The current circumstances present opportunities to

explore methods of Synergistic Processing of ores from these deposits to increase the overall viability

of Nickel Projects.


5

SNC-Lavalin and Nickel Projects

� Currently involved in Ambatovy (Madagascar) and Goro (New Caledonia) as EPCM contractor, which will be the largest Nickel Laterite operations in the

world at 50,000 tpa Ni each.

� Rio Tuba (Coral Bay) : 10,000 tpa Ni successfully

ramped up in 1 year (2005)

� Involved in local Nickel Laterite operations (MMO,

Bulong) through EPCM and studies (latest: KNP PFS)

� Studies conducted on Hydrometallurgical Treatment of Nickel Concentrates and Ni Tailings.


6

Goro under Construction


7

Australian Nickel Production

NICKEL QUANTITY

0

40

80

120

160

200

1970 1975 1980 1985 1990 1995 2000 2005

Thousand tonnes

Rest of Australia

Western Australia

Source: DMP and ABARE


8

Situation in Western Australia

� What happened ?


9

Nickel Price, Sulphur and Transport Costs

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

50,000

55,000

Ju

l-9

8

Oct-

98

Fe

b-9

9

Ju

n-9

9

Oct-

99

Fe

b-0

0

Ju

n-0

0

Oct-

00

Fe

b-0

1

Ju

n-0

1

Oct-

01

Fe

b-0

2

Ju

n-0

2

Oct-

02

Fe

b-0

3

Ju

n-0

3

Oct-

03

Ja

n-0

4

Ma

y-0

4

Se

p-0

4

Ja

n-0

5

Ma

y-0

5

Se

p-0

5

Ja

n-0

6

Ma

y-0

6

Se

p-0

6

Ja

n-0

7

Ma

y-0

7

Se

p-0

7

Ja

n-0

8

Ma

y-0

8

Se

p-0

8

Ja

n-0

9

Nic

ke

l P

ric

e (

$U

S /

to

nn

e)

Ba

ltic

Dry

In

de

x

0

200

400

600

800

1,000

Su

lph

ur

($U

S /

to

nn

e)

Ni $US / tonne Ni (cash) [Source - LME]

Baltic Dry Index [Source - Bloomberg]

Sulphur $US / tonne (FOB UAE) [Source - ADNOC]


10

Situation in Western Australia

� Boom : Escalation in all costs !

� Worldwide Credit Crunch and Recession

� Nickel Oversupply – Prices Drop

� Gas supply cut off – Varanus Island Explosion

� Operating costs for Laterites High due to Sº

� Closures of Operations:

– Ravensthorpe : shock, future still unclear

– Norilsk shuts all WA operations indefinitely

– Sinclair, Copernicus, Blair, Beta Hunt, Miitel...

� Thousands lose Jobs (operations and services)


11

Prognosis for Western Australia

� Big companies mothball / sell off marginal assets

� Long Recovery Period – World Economy / Nickel Price may take a while to recover before it becomes profitable to re-open projects. Spare Capacity taken up first

� Worldwide there are several projects lined up to commence production of nickel and cobalt in the next few years: Goro, Ambatovy, Ramu, Koniambo,Onca Puma, Barro Alto – all very large Nickel Laterite Projects

� Big and Small Companies looking for Opportunities for Survival of Operations

� Initially Limited Capital Available for Large Projects � Normalisation of costs: Reagents, Labour, Materials, Energy

and Transport as well as Engineering and Construction

� Opportunities to consider written-down assets and Synergise for sustainable future profitability


12

Synergistic Combinations

� Combinations of the following feed stocks:

– Laterite Ore of Low-Mg and High-Mg type

– Nickel Sulphide concentrates

� Combination of following Unit Processes:

– HPAL

– LT/MT POX leach

– Atmospheric Leach

– Heap Leach

� Single Site and Plant

� Common Downstream Processing

� Potential Vertical Integration, e.g. processing to LME Nickel Grade.


13

What is Synergy?

� Definition of Synergy (also WIN-WIN):

1. The interaction of two or more agents or forces so that their

combined effect is greater than the sum of their individual effects.

2. Cooperative interaction among groups, especially among the

acquired subsidiaries or merged parts of a corporation, that

creates an enhanced combined effect.

(www.freedictionary.com)

� Synergies Applied to Projects:

• Reinvigorate dormant operations by obtaining advantage from

other projects or create new viable projects from synergistic

combination.

• As metallurgists, developing and optimising process flowsheets is

a constant quest to explore synergies between different ores and

processing options. This approach could be applied outside

traditional boundaries.


14

Application of Synergies

Potential Agents of Synergy� Ore bodies /deposits and ore types within a deposit� Primary Processing (Mineral Processing and Hydromet Plants)� Smelting and Refining Operations� Utilities and Reagents� Value adding through vertical integration. (e.g. Final Metal

Production)Participating Parties� Within a company and its operations/prospects – (independence)� Within a specific region, between companies that may benefit –

(interdependence)� Creates Opportunities for acquisitions, take-overs. � Centre for Sustainable Development Studies - conducted

systematic studies for the Kwinana and Gladstone areas. – Triple bottom line (sustainability).


15

Potential Agents of Synergy

Ores /

ResourcesHydromet

ProcessingRefining

Ores /

Resources

Ores /

Resources

Concentrator

ProcessingSmelting

Refining Hydromet

Processing


16

Application of Synergies

Opportunities in realising Synergies:

� Single, larger operation treating several feedstocks

� Increase Available Resources to a Project

� Increase throughput

� Reduce unit labour cost

� Reduce unit reagent costs

� Reduce other unit operating costs

� Reduce unit capital investment

� Use existing capital assets where available


17

Major WA Nickel Projects


18

HPAL Process / Project Drivers

� Resource Grade and Size

– >1.5% Ni in Autoclave Feed for 5 years

– 30-50,000 tpa Ni for >20 years

– Acceptable Acid Consumption < 30 t/t Ni

� Ore types (Some ores are upgradeable)

� Low cost when compared to competitors

� Favourable market conditions foreseen

Ore types % Ni % Fe % Mg HPAL ATM Heap

Limonite 0.6-1.2 40-50 1-2 Yes No No

Nontronite 1.0-1.5 15-25 3-5 Yes Yes Y/N

Saprolite 0.5-1.5 8-10 10-14 No Yes Yes


19

Capital Cost – Ni Laterite HPAL Plant

� Capex = $US 40,000-60,000 / annual tonne of NiMining, 0%

Ore Beneficiation, 8%

Pressure Leaching,

20%

CCD, 6%

Purification, 8%

Product Precipitation,

2%

Process Utilities, 2%

Acid Plant, 15%

Reagents, 1%

Power Supply, 8%

Water Supply, 8%

Other Sevices, 5%

Infrastructure, 16%


20

HPAL Plant – Operating Cost

� Opex = $US 3.50-5.00 / lb Ni ($US 8-11,000 / t Ni)

Mining, 18%

Sulphur, 17%

Limestone, 4%

Lime, 5%

Magnesia, 5%

Flocculant , 3%

Consumables, 1%

Power/Gas, 11%

Product Transport, 2%

General, 4%

Tailings, 4%

Labour, 13%

Maintenance, 14%


21

Heap Leach (Stand-alone) Drivers

� 4 m heap height

� 200-300 days irrigation

� High evaporation rate

� Must be shown to be practically possible

� Very dependent on ore type

� Benefits from Cheap Sulphuric Acid

� Main Acid Consumer is Iron

� High Acid Use means high Limestone Use

� Heap Leach better as integrated operation


22

Capital Cost – Standalone Heap Leach

� Capex = $US 30,000-50,000 / annual tonne of Ni

Mining, 3%

Ore Processing , 15%

Heap Leaching, 13%

PLS Processing, 12%

Acid Plant, 31%

Reagents, 1%

Water , 7%

Power and Utilities, 10%

Infrastructure, 5%


23

Stand-alone Heap Leach – Operating Cost

� Opex = $US 5.50-7.00 lb Ni ($US 12-15,500 / t Ni)

Mining, 11%

Sulphur , 34%

Limestone, 22%

Lime , 1%

Magnesia, 3%

Consumables, 1%

Utilities, 1%

Labour, 11%

Other, 3%Transport, 1%

Maintenance, 9%

Tailings Disposal, 3%


24

Atmospheric Leaching

� Can overcome the high cost and risk of HPAL.

� Dependent on Ore types and Reactivity of Ores

� Relies on Iron Precipitation as Jarosite occurring in the leach and regenerating acid

� Not favourable for Limonite ores

� Combination of Nontronite and Saprolite could possibly work- requires testwork

� Currently being considered at Jump-up Dam


25

Nickel Sulphides - Integration of WA Operations

~ 50 ktpa Ni

35-40 ktpa Ni

~110 ktpa Ni in Matte

35-45 ktpa Ni

35-40 ktpa Ni

550 ktpa

Sulphuric ~ 65 ktpa Ni

Acid

35-40 ktpa Ni

~ 200 ktpa

Acid

Mt Keith

Leinster

Kambalda

Other

Concentrators

Kalgoorlie

Nickel

Smelter

Kwinana

Nickel

Refinery

Overseas

Refineries

Overseas

Smelters

Cawse

HPAL

Murrin Murrin

HPALCSBP

Other Local

Users


26

Sulphide Concentrate Treatment

� Default Route : (KNS) Smelter – Refinery

� KNS Concentrate : ~14% Ni, Fe:MgO > 5, As limits

� Other Operators : Norilsk, Xstrata – Export

� Current Surplus Acid supply - opportunity

� Potential for Hydromet on surplus/non-smeltable(lower grade, high Mg, high As) concentrate.

� Smelter TC & RC

� Yakabindie, Honeymoon Well – Activox

� Other Confidential Studies on Hydromet Options

� Long term – smelter replacement considered

� Transportability an advantage over laterites


27

Sulphide Concentrate - Hydromet

� Choice of Leaching Technology depends on many factors

– Concentrate Mineralogy, Chemistry of Reactions

– Acid Balance

– Water available

– Flexibility of Process to Feed Variability

– Development and Support from Technology Provider

– Availability for Licensing

� Prominent Technologies:

– CESL Process

– Activox® Process

– Albion Process

– Other Medium to High Temperature Pressure Oxidation

� Downstream Processing Similar to HPAL.


28

Capital Cost – Nickel Concentrate Leaching

� Capex =$US 11,000-14,000 / annual tonne of NiSite Preparation,

2%

Hydromet Plant,

43%

Process Utilities,

12%

Services, 29%

Infrastructure, 7% Feed Preparation,

6%


29

Nickel Concentrate Leach - Operating Cost

� Opex = $US 1.00-1.50 lb Ni ( $US 2,200-3,300 / t Ni)

Magnesia, 17%

Other Reagents, 3%

Utilities, 6%

Concentrate

Transport, 2%

Product Transport,

20%

Maintenance, 16%

Labour, 26%

Other, 11%


30

Synergistic Combinations

� Combinations of the following feed stocks:

– Laterite Ore of Low-Mg and High-Mg type

– Nickel Sulphide concentrates

� Combination of following Unit Processes:

– HPAL

– LT/MT POX leach

– Atmospheric Leach

– Heap Leach

� Single Site and Plant

� Common Downstream Processing

� Potential Vertical Integration, e.g. processing to LME Nickel Grade.


31

Heap Leach Integration with HPAL

� Favoured approach is through introduction of PLS into HPAL

(to mill or wash in with 2 stage CCD)

� All Acid associated with Fe and Al becomes available to the

HPAL.

� Effect of integration on Costs (Nontronite ore- Case Study) :

� MMO integration post HPAL on small heap

� Same approach may be possible for Atmospheric – Heap

Leach, but needs reactive ore.

Option Capital Cost Operating Cost

US$/annual tNi US$/lb Ni

Standalone Heap Leach 35,000 5.90

HPAL Integrated Heap Leach 18,000 2.40


32

Heap Leach Integration into HPAL Circuit

CCD

Purifi-

cation

Ore

PrepHPAL TSF

Tailings

Neutral.

Product

Recovery

Product

Handling

Neutra-

lisation


33

Heap Leach Integration into HPAL Circuit -1

Fresh Water

Acid

CCD

Purifi-

cation

Ore

PrepHPAL TSF

Tailings

Neutral.

Product

Recovery

Product

Handling

Neutra-

lisation

Heap

Leach 1


34

Heap Leach Integration into HPAL Circuit -2

CCD

Purifi-

cation

Ore

PrepHPAL TSF

Tailings

Neutral.

Product

Recovery

Product

Handling

Neutra-

lisation

Heap

Leach 2


35

HPAL or AL with Saprolite Neutralisation

� Acid is available after HPAL due to leach solubility requirements.

� This acid can be used to leach reactive Saprolite while extracting more nickel.

� Atmospheric Leach can also be used in this way, but requires high Sap: Lim ratio to cater for Iron

� Results in Reduction in Acid Consumption and Operating Costs.

� Availability of Reactive Saprolite creates the Synergy.

� Reactive Concentrator Tailings may be considered as alternative.


36

HPAL or AL with Saprolite Neutralisation

0

10

20

30

40

50

60

70

80

90

100

Lim PAL Sap PAL PAL-SN Lim-ATM Sap ATM AL-SN 50:50

Acid

(t/t N

i)

Mg

Ni, Co

Cr,Si, Mn

Al

Fe

Free Acid

Acid Economy (t Acid / t Ni)


37

Integration of Saprolite into HPAL/AL Circuit

CCD

Purifi-

cation

Ore

PrepHPAL TSF

Tailings

Neutral.

Product

Recovery

Product

Handling

Neutra-

lisation


38

Integration of Saprolite into HPAL/AL Circuit

Acid

High Mg / Reactive Ore

or Nickel Tailings

CCD

Purifi-

cation

Ore

Prep

HPAL /

ALTSF

Tailings

Neutral.

Product

Recovery

Product

Handling

SAP

Neut.


39

Sulphide Integration into HPAL circuit 1

Approach 1: Sulphides into Autoclave

� Has been used by Cawse to regulate Eh and increase extraction (alternative to Sº)

� Limitations due to Oxidation Potential and Solubility Constraints

� Limit on concentrate addition without Oxygen � “Hybrid Process” using Pyrite described by Curtin Uni Paper

has potential savings in steam and acid – Replace pyrite with LG Ni Con to get energy, acid and Ni credits.

� Risk in using oxygen (Ti-fires) and sulphides (reducing agents) – corrosion

� Manganese Ore can be added instead of oxygen –Manganese extracted downstream - tried before.

� These approaches need proper testing.


40

Integration of Sulphides into HPAL Circuit 1

Benefits in Autoclave:

Energy

Acid

Nickel Credits

Nontronite Ore

Nickel Containing

Sulphide Concentrate

CCD

Purifi-

cation

Ore

PrepHPAL TSF

Tailings

Neutral.

Product

Recovery

Product

Handling

Neutra-

lisation


41

Sulphide Integration into HPAL circuit 2

Approach 2: Sulphides leached post Laterite Leach

� Use HPAL PLS – Acid consuming concentrates e:g: High carbonates – use Albion Process

� Use PLS after Sap Neutralisation for Acid Producing concentrate and MT POX

� Use AL PLS and use LT POX to precipitate Fe

� Increases Solution Ni Tenor and uses same downstream process

� Increased throughput – benefit of scale

� Value adding and benefit of scale

� Lower risk than using HPAL Autoclaves


42


Reactive Ni Sulphide

Concentrate

Nontronite Ore

Less reactive

Nickel Sulphide

Concentrate

CCD

Purifi-

cation

Ore

PrepHPAL TSF

Tailings

Neutral.

Product

Recovery

Product

Handling

Neutra-

lisation

Sulphide

Leach

Sulphide

Leach

Thickener


43


Saprolite Ore

Nontronite Ore

Nickel Containing

Sulphide Concentrate

CCD

Purifi-

cation

Ore

PrepAL TSF

Tailings

Neutral.

Product

Recovery

Product

Handling

MT POX


44

Other Synergistic Opportunities

� Power Availability : Partly Provided by Sulphuric Acid Plant. Other power infrastructure.

� Water availability : Hypersaline mine water has potential benefits for Atmospheric Leaching

� Shut operations : Cheap Capital Assets

� Unused Plant and Equipment : Operating Plant first Prize, but parts can be redeployed as well.

� Shared Infrastructure : e.g. rail line extension to Geraldton

� Resource assets sold off to service debt

� Surplus Acid : not enough for large Laterite project, but could be used with a revived Cawse and Sulphide integration

� Normalisation of costs.

� Environmental Management – Tailings in one area


45

Methodology : Pursuing Synergies

Identification

� Characterise Ore by Mineralogy and Response to

Processing

� Brainstorm

� Use mixture of “youth and experience”

� Gather intelligence

� Identify need for project cost reduction

� Identify Potential Stake-holder

� Communicate need for Synergies to Stake-holders

� Consider Models for Ownership of Asset development

� Define and List Opportunities


46


Evaluation

� Evaluate Potential Benefits

� Rank Opportunities according to value and risk

� Conduct more detailed technical studies� Involve Specialists

� Conduct Metallurgical Testwork

� Simulate Process options using e.g. METSIM

� Involve other parties, if applicable, and talk more Win-Win

� Re-evaluate Financial Viability (Scoping Study)� Narrow down options using Risk Assessment (Synergies

can add substantial Risk, especially if more than 1 party is involved)

� Start talking commercially with other parties

� Involve Financiers in Ownership Models


47


Implementation

� Reach Preliminary Commercial Agreements

� Conduct a Pre-feasibility Study

� Conduct a Definitive Feasibility Study or Engineering Cost Study

� Finalise Commercial Agreements, if applicable.

� Engineer

� Construct

� Commission

� Reap the Benefits


48

Conclusion

� The current circumstances present an opportunity to explore synergies to improve project economics on Nickel projects.

� Using the traditional HPAL Laterite Process plant as a basis consideration can be given to incorporation of treatment of some laterite ores via heap leaching and atmospheric leaching and treatment of sulphide concentrates in the same flowsheet.

� Alternatives that exclude HPAL but include the other options may be considered for the right combination of ore types.



LATERITE HEAP LEACHING


1

NICKEL LATERITE PROCESSING

A SHIFT TOWARDS HEAP LEACHING

NICKEL LATERITE PROCESSINGNICKEL LATERITE PROCESSING

A SHIFT TOWARDS HEAP LEACHINGA SHIFT TOWARDS HEAP LEACHING

By

Bruce Wedderburn

Malachite Consulting, Australia

Presented by

Bruce Wedderburn

[email protected]


2

DisclaimerDisclaimer

The following presentation is based mainly on public

domain information and represents Malachite Consulting’s

best judgment at the time of presentation. The contents

include forward looking statements prepared on the basis

of assumptions which may prove to be incorrect.

This presentation should not be relied upon as a

recommendation or forecast by Malachite Consulting.

No representation or warranty is made as to accuracy,

completeness or reliability of the information.


3

IntroductionIntroduction

Malachite Consulting

• Process engineering consulting company

• Extensive experience with nickel laterites

• Visited over 20 nickel laterite projects & operations

• Evaluated over 30 nickel laterite projects

• Expertise in pressure, atmospheric & heap leaching

• Experience with FeNi and Caron processing

• Not selling or promoting a specific technology


4

AgendaAgenda

1. Current supply demand environment

2. Supply response to nickel price

3. Long term nickel prices

4. Laterite processing technologies

5. Heap leach projects

6. Capital and operating costs


5

LME Nickel PriceLME Nickel Price


6

Stainless Steel ProductionStainless Steel Production


7

Nickel Supply Demand BalanceNickel Supply Demand Balance


8

LME Nickel StocksLME Nickel Stocks


9

Nickel Mine Closures in 2008Nickel Mine Closures in 2008

China – Reduced pig nickel production – Aug 08

Murrin Murrin - Minara to delay a A$300 million expansion – Aug 08

Falcondo - Xstrata suspends operations due to high cost of oil (update - closed indefinitely)

Jinchuan – China, to reduce its nickel output for 2008 by 20,000 tonnes

Blair Mine - Australian Mines suspended operations

Shakespeare Mine – Canada, Ursa Major suspending operations – Sep 08

Hitura Mine – Finland, Belvedere suspending operations – Oct 08

Cawse Mine - Norilsk suspended operations – Oct 08

Lockerby Mine – Canada, First Nickel suspended operations – Oct 08

Lac des Isles Mine – Canada, North American Palladium suspends operations – Oct 08

Levack Mine – Canada, FNX Mining suspends operations

Berong Mine – Philippines, Toledo suspends mining operations – Nov 08


10


Copernicus Mine – Australia, Panoramic suspended operations – Nov 08

Beta Hunt Mine – Australia, Consolidated Minerals to suspend operations by mid Dec 08

Sorowako – Indonesia, PT Inco - Shutting down all the thermal power generators

Ufaleynikel – Russia, Russia's 3rd largest producer suspends operations – Oct 08

Moa Bay – Cuba, Sherritt will suspend expansion of its the Moa nickel operations – Oct 08

McWatters Mine – Canada, Liberty suspending operations – Oct 08

Redstone Mine – Canada, Liberty suspending operations – Oct 08

Mincor – Australia, reducing production

Fenix Project – Guatemala, Hudbay announces delay in construction start-up – Nov 08

Ban Phuc Nickel project – Vietnam, Asian Mineral Resources suspends construction

Taganito – Philippines, SMMC says it may delay the start of its Taganito nickel project

Craig Nickel Mine – Canada, Xstrata ceasing operations by June 09


11


Thayer-Lindsey Mine – Canada, Xstrata ceasing operations by Jan 09

Gag Island Project – Indonesia, BHPB cancels development of $4.5 billion mine

Waterloo – Australia, Norilsk suspended operations – Nov 08

Silver Swan – Australia, Norilsk suspended operations

PT Inco – Indonesia, will reduce production in 2009 by 20% if price of nickel stays low

PT Aneka Tambang – Indonesia, reducing production to 11,500-12,000 tonnes

Activox Demonstration Plant – Botswana, Norilsk will shut down permanently

Trojan Mine – Zimbabwe, Bindura suspending operations – Nov 08

Shangani Mine – Zimbabwe, Bindura suspending operations – Nov 08

Ambatovy – Madagascar, project is on "care and maintenance" for at least 6-8 months

Altai Project – Russia, Russian Nickel Company puts project on hold – Dec 08

Copper Cliff Mine – Canada, Vale suspends operations beginning in January 09


12


Copper Cliff Deep Project – Canada, Vale delays development for 12 months

Voisey's Bay – Canada, Vale will shut down for month of July 09

Talvivaara – Finland, cuts 2009 production target estimates by 2,000 to 5,000 tonnes

Eramet - New Caledonia, Nickel output will be adjusted down to about 50,000 tonnes

Miitel Mine – Australia, Mincor Resources to suspend production end of Dec 08

McCreedy West Mine – Canada, FNX suspend production

Avebury nickel mine – Australia, OZ Minerals Ltd closes new mine due to low price of nickel

Barro Alto – Brazil, Anglo American delays construction by 12 months.

Tati Nickel – Botswana, Norilsk halts production for 12 days after fire in smelter


13


Doniambo - New Caledonia, Eramet announces production cut to 50,000 tonnes

Loma de Niquel – Venezuela, Anglo American halts production 'temporarily’ – Jan 09

Ravensthorpe – Australia, BHPB announced closing indefinitely – Jan 09

Yabulu refinery – Australia, BHPB announced it will stop processing nickel hydroxide

Mount Keith mine – Australia, BHPB announced reducing rate of production

PT Aneka Tambang – Indonesia, to cut its output of ferronickel by 32% in 2009

Isabela mine – Philippines, announced that the mine has been shut down

Craig and Thayer-Lindsley Mines – Canada, both mines on care and maintenance

Berong Nickel Mine – Philippines, Toledo Mining announced all production halted

Black Swan, Lake Johnston – Australia, Norilsk Nickel closes remaining two nickel operations

Kavadarci FeNi smelter – Macedonia, will reduce production by 80% in 2009


14


Larco – Greece, announced that it will cut production by 55% in 2009

Munali – Zambia, Albidon suspends operations, 350 jobs eliminated

Onça Puma Project – Brazil, Vale delays start up to January 2011

Sudbury Operations – Canada, Vale shuts down operations for additional two months

Goro – New Caledonia, Vale delays commissioning till 2H09, following acid spill

Leinster – Australia, BHPB announces that production at Rocky’s Reward has ceased

Kalgoorlie Nickel – Australia, Vale announces withdrawal from the project


15

Long Term Nickel PricingLong Term Nickel Pricing

1. Previous forecast of US$4 to US$6/lb Ni *

2. No reason to change forecast

3. Price capped by Chinese pig iron production

4. Costs of Ni pig iron at US$8 to US$10/lb and falling

5. Challenge is to bring on new projects using a long

term price of US$4 to 6/lb Ni

* Australian Journal of Mining - November/December 2007


16

Laterite Processing OptionsLaterite Processing Options

Ferronickel smelting (FeNi)

Blast / Electric furnace (Pig iron)

Caron process (Reductive roast)

AMAX process (HPAL & AL)

High pressure acid leach (HPAL)

Atmospheric leaching (AL)

Sulphation atmospheric leach (SAL)

Heap leaching (HL)


17

Tropical Tropical LateriteLaterite ProfileProfile


18

FeNi SmeltersFeNi Smelters

Existing Operations• Doniambo (New Caledonia)

• Falcondo (S/D) (Dominican Republic)

• Cerro Matoso (Columbia)

• Larco (Greece)

• Pobuzky (Ukraine)

• Fenimak (Macedonia)

• Ferronikeli (S/D) (Kosovo)

• Indonesia (Pomalaa, Soroako)

• Loma de Niquel (Venezuela)

• Japanese (Various)

Committed Projects• Gwangyang (POSCO) - Completed Nov 08

• Onça-Puma (CVRD Inco)

Well Advanced Projects• Koniambo (Xstrata Nickel)


19

FerronickelFerronickel

Reduced Production• Falcondo ~ 29,000 tpa Ni

• PT Inco ~ 25,000 tpa Ni

• Eramet ~ 12,500 tpa Ni

• Larco ~ 10,000 tpa Ni

• Fenimak ~ 8,000 tpa Ni

• Ferronikeli ~ 8,000 tpa Ni

• ANTAM ~ 6,500 tpa Ni

• Pobuzky ~ 3,500 tpa Ni

Commentary• Falcondo & Feronikeli shut-down

• PT Inco shut-down thermal power generation

• ANTAM moving to coal fired power plants

• Onça Puma completion delayed till January 2011


20

Nickel Pig IronNickel Pig Iron

Existing Operations• China (> 30 producers)

• Zhejiang Huaguang (~1 Mtpa pig iron)

Commentary• Pig iron production 30ktpa Ni in 2006, 85ktpa Ni in 2007 and

90ktpa Ni in 2008

• Production expected to decline to 40ktpa Ni for 2009

• Nickel ore stockpiled in China ~ 8Mt

• EAF’s capable of higher quality product (12-15% Ni)

• Blast furnace cash costs ~ US$11 to 13/lb Ni

• Electric furnace cash costs ~ US$8 to 10/lb Ni

• Not a sustainable industry at current nickel prices

• Swing producer and potentially caps price spikes


21

Caron ProcessCaron Process

Existing Operations• Nicaro (Cuba)

• Punta Gorda (Cuba)

• Yabulu (QLD)

• Tocantins (Brazil)

Closed Operations• Nonoc (Philippines)

• Las Camariocas (Cuba)

Planned Projects• None


22

HPAL ProcessHPAL Process

Existing Operations• Moa Bay (Sherritt)

• Murrin Murrin (Minara)

• Cawse & Bulong (S/D) (Norilsk)

• Rio Tuba (Sumitomo)

• Ravensthorpe (S/D) (BHP Billiton)

Committed Projects• Goro (CVRD Inco)

• Ambatovy (Sherritt)

• Ramu (Metallurgical Construction Corp)

Well Advanced Projects• Vermelho (CVRD Inco)

• Weda Bay (Eramet)


23

AL / SAL ProcessAL / SAL Process

Existing Operations

• None

Committed Projects

• None

Early Development Projects

• Fenix (SAL) – Guatemala (Skye Resources)

• Sechol – Guatemala (Jaguar Nickel) *

* Project changed to ferronickel in 2005


24

Heap Leach ProcessHeap Leach Process

Committed Projects

• Çaldağ - Turkey (European Nickel)

Well Advanced Projects (Demonstration Plants)

• Yuanjiang - China (Yunnan Tin Group)

• Murrin Murrin (Minara)

• Niquel do Piauí (Vale)

• NTUA – Greece (Larco)

Early Development Projects

Nornico – QLD NickelOre – WA

Rusina’s Acoje – Philippines GME – WA


25

Heap Leach HistoryHeap Leach History

• Nickel laterite heap leaching R&D commenced in

early 1990’s at NTUA in Greece

• NTUA undertook column leaches on Greek laterites

• Greek laterites consist mainly of hematite and the

nickel is contained in chlorite

• Amenable to heap leaching with minimal acid

consumption

• BHP, WMC and others undertook R&D commencing

in mid to late 1990’s


26

Heap Leach ChallengesHeap Leach Challenges

Metals Extraction

• Nickel leach efficiency

• Iron chemistry

Heap Leach Design

• Acid addition

• Bed permeability

• Agglomeration

• Downstream processing


27

Iron ChemistryIron Chemistry

HPAL

Fe2(SO4)3 + 3H2O Fe2O3 + 3H2SO4

Atmospheric Leach

Fe2(SO4)3 + ⅓Na2SO4 ⅔NaFe3(SO4)2(OH)6 + 2H2SO4

Heap Leach

Fe2(SO4)3 + 3CaO 2FeO(OH) + 3CaSO4


28

Heap Leach TestworkHeap Leach Testwork

Heap Leach Testwork

• Bottle rolls

• Column tests (1m, 4m and 8m)

Mineralogy

• Impact of agglomeration and leaching

Geotechnical

• Permeability under load

• Triaxial and trafficability

Downstream processing


29

Balkan Laterites (Greece)Balkan Balkan LateritesLaterites (Greece)(Greece)


30

Balkan Laterites (Albania)Balkan Balkan LateritesLaterites (Albania)(Albania)


31

Balkan Laterites (Turkey)Balkan Balkan LateritesLaterites (Turkey)(Turkey)


32

Çaldağ - TurkeyÇaldağ - Turkey

Commercial Operation

• Infrastructure development commenced in 2006

• Forestry permits approved in February 2009

• Heap leaching expected to commence in mid 2010

• Full production of 20 Ktpa Ni & 1.2 Ktpa Co in 2011

• Capital cost of US$428m


33



34



35



36



37


ROM

CRUSHER

PIPE

CONVEYOR

COARSE ORE

STOCKPILE

STACKER

SULPHUR

ACID

PLANT POWER

HEAPS

PONDSLIMESTONE

SODA ASH

IRONREMOVAL

THICKENER

IRON

FILTERCAKE(WASTE)

NICKEL

PRECIPITATIONSTAGE 1

THICKENER

NICKEL

PRECIPITATION

STAGE 2

THICKENER

FIRST NICKELPRODUCT

PROCESS

WATER

SECOND NICKEL

PRODUCT


38

Two products

First Nickel Product (FNP)

34% Ni, 1-1.2% Co, <2% Mn

Second Nickel Product (SNP)

24% Ni, 0.8% Co, <8% Mn

Potential customers

Nickel refineries

Smelters

BHP Billiton exercised option to purchase 100% of MHP output.

Subject to renegotiation to supply 50% to Chinese processors

Nickel product



39


Updated Economics 2006 2009

Production tpa 20,400 Ni in MHP

Construction Cost US$m 238 277

Long-term Ni Price US$/lb 4.25 6.00

Capital Intensity US$/lb 5.57 6.12

Annual Cash Operating Cost US$m 70 103

Cash Operating Cost US$/lb 1.41 1.99

NPV (10% discount) US$m 180 207

IRR % 23 20

Total Capital Cost US$m 300 428

Source: European Nickel


40

Murrin MurrinMurrin Murrin

Demonstration Heap

• 400,000 tpa of scats under leach

• Geotechnical stability of heap – no issues

• Percolation rate – good percolation being

achieved on scats and ore

• Recovery rate – better than expected recovery

curve

• Scats nickel recovery 78% after 180 days –

slightly better than testwork predicted


41

Murrin MurrinMurrin Murrin MurrinMurrin


42

Murrin MurrinMurrin Murrin

Growth Targets

• 400,000 tpa of scats under leach

• Ore is being leached together with scats

• Current pad capacity yields 2,000 tpa Ni

• Heap Leach infrastructure sized for 10,000 tpa Ni

• Expansion of the project has been deferred due to

current market conditions


43



44



45



46



47

Níquel do Piauí - BrazilNíquel do Piauí - Brazil

Demonstration Plant• Vale constructed a demonstration plant in northern state of

Piauí in Brazil

• GRD Minproc awarded contract in 2008 for front end engineering study comprising

• Flowsheet development,

• Engineering for crushing, infrastructure and

• Nickel bearing solution metallurgy,

• Capital and operating cost estimates.

• Completion of basic engineering planned for 2009

• Commercial production estimated for 2012 with a capacity of

36,000 tpa Ni


48



49



50

NTUA - GreeceNTUA - Greece

Demonstration Plant

• Built a small demonstration plant at Larco mines

• Mineralogy is mainly hematite & low acid consumption

• Process consists of:

• Heap leaching of laterite ore at ambient temperature

• Purification of the leach liquor by hydrolytic precipitation

• Removal of Fe, Al and Cr

• Simultaneous SX of Ni and Co using the Cyanex 301

• Stripping of the loaded organic by Ni spent electrolyte

• Co extraction over Ni and Co using the Cyanex 272

• Stripping of the loaded organic by Co spent electrolyte


51

NTUA - GreeceNTUA NTUA -- GreeceGreece


52



53



54



55

Metallica MineralsMetallica Minerals

Nornico Project

• Heap leach testwork completed at HRL and MFC

• Flowsheet changed to incorporate IX to recover Ni and Co after Fe precipitation

• 1.0 Mtpa ore to be processed by heap leaching

• 15 year mine life

• 7,000 tpa Ni & Co as an intermediate product

• Funding feasibility study been deferred due to current market conditions


56

Metallica MineralsMetallica Minerals

Lucky Break Project

• MFC to fund entire A$20m development

• 250 Ktpa ore to be processed by combined heap and vat leach

• Expect 85% Ni extraction after 3 months

• On-off heap leach pads

• 4 to 5 years mine life

• 1,600 tpa Ni contained in carbonate

• Delayed due to acid purchase contract dispute


57

Heap LeachingHeap Leaching

Positives

• Process both limonite & saprolite

• Process far less complex than HPAL or AL

• Lower capital intensity

Negatives

• Heap permeability is critical

• Sensitive to mineralogy

• Iron chemistry

• Acid consumption


58

Capital CostsCapital Costs

FeNi Smelters

• Onça-Puma US$2.3bn => US$18.30/lb Ni

• Koniambo US$3.9bn => US$29.50/lb Ni

HPAL Projects

• Vermelho US$1.9bn => US$18.70/lb Ni

• Ramu * US$1.4bn => US$19.85/lb Ni

• Gladstone US$3.4bn => US$24.50/lb Ni

• Ravensthorpe * US$2.8bn => US$28.20/lb Ni

• Goro US$4.1bn => US$30.90/lb Ni

• Ambatovy US$4.5bn => US$34.20/lb Ni

Source: Public Domain Information

* HPAL capital costs to intermediate product


59

Capital CostsCapital Costs


• Approximately 80% of capital cost of HPAL and similar

operating costs (BHPB - June 2001)

Heap Leach Projects

• Çaldağ US$428m => US$9.70/lb Ni

• GME US$960m => US$12.50/lb Ni

• NickelOre US$650m => US$14.70/lb Ni

• Metallica US$315m => US$19.00/lb Ni

• Heron US$610m => US$27.70/lb Ni

Heap leach capital costs to intermediate product


60

HPAL Capital CostsHPAL Capital Costs

Capital expenditure, US$ million

Source: CRU, Brook Hunt & Public Domain


61

Operating CostsOperating Costs

HPAL Projects (net of Co credits)

• Murrin Murrin US$5.55/lb (Minara Feb 09)

• Moa Bay US$4.12/lb (Sherritt Mar 09)

• Ramu US$2.90/lb (Brook Hunt)

• Ravensthorpe US$2.70/lb (Brook Hunt)

• Goro US$1.15/lb (Inco Feb 05)

• Ambatovy US$0.77/lb (Dynatec Apr 06)


• Similar operating costs to HPAL

Source: Brook Hunt & Public Domain


62

Operating CostsOperating Costs

Heap Leaching (net of Co credits)

• Çaldağ US$1.99/lb Ni (European Nickel)

• NiWest US$2.37/lb Ni (GME)

• Canegrass US$2.67/lb Ni (NickelOre)

• Heron US$6.39/lb Ni (Heron)

Notes:

Heap leach operating costs to intermediate product

Opex = C1 cash costs net of Cobalt credits

Source: Public Domain


63

Operating Costs at Laterite Projects are also much higher than initially presented

Operating Costs at Laterite Projects are also much higher than initially presented

Cash operating costs after by-product credits, US$/lb Ni

Operating cost estimates unusually high due to operation below full capacity, technical problems and high maintenance expenditure – CRU Presentation LME Week 2004


64

By-product CreditsCobalt Price Volatility

ByBy--product Creditsproduct CreditsCobalt Price VolatilityCobalt Price Volatility


65

Operating CostsImpact of Cobalt By-product Credits

Operating CostsOperating CostsImpact of Impact of CobaltCobalt ByBy--product Creditsproduct Credits

Opex Q1 2009 Q1 2008

Mining & Processing US$5.00/lb US$5.63/lb

Third Party Feed US$0.35/lb US$1.32/lb

Cobalt Credits (US$1.48/lb) (US$5.06/lb)

Other US$0.25/lb US$0.06/lb

C1 Cash Costs US$4.12/lb US$1.95/lb

Sherritt Quarterly results for March 2009

Cobalt Price US$16/lb US$46/lb


66

Capital & Operating CostsCapital & Operating CostsCapital & Operating Costs

HPALAtmos

Leach

Heap

Leach

CapexUS$25 – 30

/lb Ni

US$20 - 24

/lb Ni

US$10 – 15

/lb Ni

OpexUS$3.50 –

4.50 / lb Ni

US$3.50 –

4.50 / lb Ni

US$2.50 –

3.50 / lb Ni

Capacity 60,000 tpa 60,000 tpa 30,000 tpa

Opex = C1 cash costs net of Cobalt credits at US$7/lb

Heap leach capex/opex costs to intermediate product


67

Concluding CommentsConcluding Comments

FeNi Smelter

• Low power costs & high grade ores are essential

• Robust process suitable for saprolite ores

HPAL

• Capital and operating costs are high

• Requires high nickel price to support project economics

Heap Leaching

• Mineralogy dependent and not suitable for all deposits

• Lower capital costs are attractive to smaller operators


THE DEVELOPMENT OF NICKEL LATERITE HEAP LEACH

PROJECTS

By

ML Steemson and ME Smith

Vector Engineering Inc., Ausenco Group, Australia

Presented by

ML Steemson

[email protected]

CONTENTS

1. INTRODUCTION 2

2. KEY ISSUES IN A NICKEL LATERITE HEAP LEACH PROJECTS 2

3. NICKEL LATERITE PROJECT DEVELOPMENT ISSUES 3

4. INTEGRATION OF DEVELOPMENT ISSUES 21

5. ACKNOWLEDGEMENTS 22

6. REFERENCES 22


2

1. INTRODUCTION

Owing to the potential for comparatively low capital costs and lower economic cut-off ore grades, heap leaching has been proposed as an alternative processing route for the processing of nickel laterites by a number of recent projects, including European Nickel’s Çaldağ project in Turkey and Acoje project in the Philippines, Vale’s Piauí Project in Brazil and Metallica Minerals’ Nornico project in Australia, among others. Indeed, most new nickel projects are investigating heap leaching as an option for at least a portion of the production.

On the surface, heap leaching of a nickel laterite would not appear to be attractive owing to the low permeability of most nickel laterite ores, coupled with the expected high acid consumption and long leach cycles. The concept of heap leaching of nickel laterites was firstly proposed by Agnatnizi-Leaonardou and Demaki

1, who achieved over 80% nickel recovery during column leach tests

conducted on Greek laterite samples. More extensive development work on heap leaching and metal recovery from solution was undertaken by BHP Billiton, who applied acid agglomeration to a range of nickel laterite ores, including limonites, saprolites and fine clay bearing ores

9. They also

introduced the concept of two-stage heap leaching, using a lead and a lag heap. An important observation was made in column testing in that by controlling the acidity of leachate solutions, nickel could be preferentially leached over other major gangue elements, especially iron-containing minerals. This observation has been reinforced in more recent project investigations including the Çaldağ and Nornico projects. The observation that nickel-bearing minerals often leach at faster kinetics than iron-bearing gangue means that acid usage is generally lower than more omnivorous processes, such as atmospheric leaching in tanks.

Since 2005, Vector Engineering, Inc. (now part of the Ausenco group of companies) has been involved in the development of a number of nickel laterite heap leach projects around the world, initially in terms of geotechnical design and development, and more recently process design, PLS recovery and environmental integration. These projects include current nickel laterite projects under development such as Çaldağ, Acoje, Piauí and a number of other project investigations in Central and South America, Indonesia, the Philippines and the USA. The impetus for the design of heap leaching in these projects has been from experience developed over many years of copper and gold heap leaching; however, nickel laterite projects involve unique issues which need to be considered in project development. The current paper discusses some of these issues, particularly the integration of geotechnical, PLS processing, water management and environmental issues.

2. KEY ISSUES IN A NICKEL LATERITE HEAP LEACH PROJECTS

In consideration of nickel laterite heap leaching there are generally several major questions which need to be addressed, some of which are generic, and some project specific. Based on recent experience some of the main issues to be addressed are:

Geotechnical Issues

• Heap permeability, which affects not only the height of heap, but also the mode of operation and cycle time, and how permeability changes with time.

• Method of heap construction and (for a dynamic heap) off-loading.

• Residue or tailings management – not only of spent leach ore (generally ‘ripios’ in South America and that term has been adopted here), but also process plant residues which can be similar in volume.

• High rainfall effects and the application of “raincoats”.

• Physical stability of the heap & residue or residue storage facility (RSF), including slope stability, erosion, and traffic support.


3

Processing Issues

• Heap leach sizing and scale-up.

• Nickel (and cobalt) recovery from pregnant leach solution (PLS) to achieve high quality product.

• Iron Control.

• Magnesium Control.

• Acid Consumption.

Water Management

• High rainfall management

• Water supply & surplus water issues Environmental Issues

• Liquid effluent control

• Ripios (spent ore) and RSF run-off

• Ripios and RSF post-closure stability

• Containment systems and covers for the leach pad and RSF

Depending on project specifics, each of these issues needs to be addressed in the project development program, with some being more critical for particular project scenarios. For example, high rainfall management is a critical issue for tropical, inland projects where the discharge of sulphate-containing effluent is generally not allowable.

The following section discusses nickel laterite heap leach project development relating to these issues.

3. NICKEL LATERITE PROJECT DEVELOPMENT ISSUES

3.1 GEOTECHNICAL ISSUES

Heap Permeability

Maximum sustainable irrigation rates are directly related to permeability. Thus, a key issue applicable to all laterite and saprolite deposits is the low permeability of the ore and its sensitivity to heap height. With nickel laterites this is further complicated by dissolution of up to 30% of the solids and the related destruction of the permeability-enhancing agglomerates. Therefore, ore permeability testing should include both fresh agglomerate and leached residues or ripios (representing the lowest permeability) over a range of simulated heap or lift heights. In several recent testing programs including column metallurgical tests followed by geotechnical analyses, ripios samples reported permeability values consistent with sustainable irrigation rates of 5 to 10 L/m

2/hr with maximum lift heights of 4 to 8 m. Higher irrigation rates or thicker lifts would have

resulted in fully saturated heaps, surface ponding, slope instability and high susceptibility to liquefaction (both earthquake induced and static). Within these ranges of lift thicknesses and irrigation rates, unsaturated permeability testing generally suggested a high degree of saturation by the end of the leach cycle, which also affects slope stability and liquefaction potential. These data further suggest that multi-stack heap drainage would be significantly impaired, thus indicating the need for thin liners and drainage systems between each lift, as is common in oxide copper heap leaching

15.

For steady-state irrigation under assumed fully saturated conditions (as a limiting condition), the following Table 3.1 indicates the maximum possible irrigation rate as a function of saturated permeability. For practical purposes the allowable irrigation rate would be much lower – one third to one fifth would be a typical range – of these values to provide for variations across the heap and to avoid an unacceptably high degree of saturation.


4

Table 3.1: Maximum irrigation rates as a function of saturated permeability

(fully saturated lift)

Permeability, Saturated

(cm/s)

Maximum Irrigation Rate

(L/m2/hr)

1 x 10-3

36

5 x 10-4

18

3 x 10-4

10

1 x 10-4

5

5 x 10-5

2

A key point in interpreting metallurgical column leach data in terms of permeability has been that small columns generally are optimistic in terms of drainage properties. This is caused by a number of factors, ranging from lower ore densities to bridging along the column walls to solution channelling. In one case a sample was leached in a small (nominally 200 mm) diameter clear plastic column. The laboratory reported sustainable irrigation rates of up to 100 L/m

3/hr though the ore was a limonite. Visual inspection of the column showed that

most of the solution had followed channels along the column-ore interface (Figure 3.1). Larger columns (say around 1,000 mm diameter), properly loaded, tend to report more realistic drainage information and there is good correlation between performance in these columns, geotechnical testing and actual heap performance.

Figure 3.1 Channelling in a small diameter column

Key factors that affect permeability and thus heap drainage are:

• Physical properties of the ore, including: particle size distribution; clay content and clay mineral type; density and porosity; degree of saturation; and segregation of the stacked ore. Most of these change for the worse with time in an acid leach environment.

• Agglomeration: principally influencing drainage and acid distribution early in the leach cycle, as the agglomerates tend to fully degrade during leaching. Also, agglomerated ore tends to segregate less during stacking and thus will channel less even after the agglomerates have degraded.


5

• Method of heap stacking (trucks, stackers working in advance or retreat modes) which affects as-stacked density, segregation and agglomerate quality.

• Segregation in the heap (see Figure 3.2), which can create zones of lower and higher permeability, and the solution will preferentially flow though the high permeability zones, starving the other areas for acid, as well as destabilizing slopes.

Figure 3.2 Segregation and solution channeling in a lateritic heap trigger a large slope failure

• Degree of saturation: the permeability increases as the saturation level approaches 100%. Most geotechnical laboratory data is based on 100% saturation (or very close thereto), but in operations initial conditions will be much lower (thus sometimes limiting the start-up irrigation rate for fresh ore) and some variation across the heap is expected. Further and very importantly, heaps should not be designed to operate fully saturated, but rather saturation should be limited to a thin zone at the base of the heap. Thus, both conventional saturated and unsaturated permeability testing should be conducted and used in the heap modelling.

• Lift thickness: ore compresses with load which reduces the permeability, and the matrix can collapse with sufficient load which can drastically reduce permeability. Figure 3.3 demonstrates the typical permeability response with increasing load for a range of laterite ores.

Figure 3.3 Permeability vs. Lift Thickness by Laterite Ore Quality


6

Heap Construction and Off-Loading

Either multi-stacking (conventional, permanent heap) and dynamic heap (on/off leach pad) are appropriate for nickel laterites, depending on site and ore factors. For several project trade-off studies both multi-stacking and dynamic heaps were found to have similar initial capital costs and similar life-of-project NPVs. Thus, the decision between these two technologies will often be driven by factors other than economics. Generally speaking, the relative advantages of each approach can be summarized as:

• Leach cycles: short, well defined leach cycles tend to favour dynamic heaps while longer or variable cycles favour multi-stacking.

• Land availability: dynamic heaps require more total land because of the ripios disposal, but multi-stacking requires more area for the initial operation and larger “flat” areas.

• Ripios disposal and dump closure costs: a combined dump (with ripios from a dynamic heap and plant residue) can be smaller and thus less expensive to close than two separate piles (a conventional heap separate from a residue disposal area);

• Ability to leach the spent ore (a ripios leach facility) to recover additional metal should be considered. For most nickel laterites, however, the ripios will be of such poor drainage quality that additional leaching will probably not be practical.

• Water balance: because a dynamic heap requires a larger total area (heap plus ripios dump), but the water from the ripios dump can be easier to divert from process than from a larger multi-stack heap.

• Traffic support capacity of leached ore: supporting the stacker over the prior lifts of leached ore is a key limitation for a multi-stack system with lateritic ore.

• Risk factors can vary considerable between the two approaches, and it is often a risk analysis that leads to the final selection

For both multi-stack and dynamic heaps, there are four approaches to stacking the heap:

• Retreat stacking of the leach heap with a conventional radial stacker. Lifts of 2 to 10m can be accommodated with conventional radial stackers (see Figure 3.4).

o Applicable to either multi-stack or dynamic. o Most common method for small and medium size copper operations o In use at: Tintaya (Cu, Peru), Cerro Verde (Cu, Peru), and many others.

• Advance stacking using trucks to dump the ore on the active life.

o Common in gold and silver heap leaching where the quality of the ore is generally very good, and in thick lift, high tonnage copper dump leaching.

o Probably not applicable to nickel laterites due to the susceptibility to compaction, agglomerate degradation, etc.

• Advance stacking of the leach heap with a low-height radial stacker (a hybrid system using the simple equipment of the retreat stacker concept but in advance mode). This offers the advantage of not needing to operate the stacking system over the underlying leached ore lift (and thus reducing traffic support issues).

o Generally only applicable to multi-stacking. o Can damage fresh agglomerates and compact the top of the active lift o In use at Ocampo (Au, Mexico)

• Advance stacking with a self-propelled tripper-stacker (see Figure 3.5).

o Suitable for a very broad range of lift thicknesses. o Same advantages of advance stacking with a radial stacker, but suitable for

higher tonnage rates and generally requires less labour.


7

o In use at: El Tesoro (Cu, Chile), Gaby (Cu, Chile), Florida Canyon (Au, USA)

Figure 3.4 Radial Stacker Operating in Retreat Mode, Chile

Figure 3.5 Tripper Stacker Operating in Advance Mode, Chile

Dynamic heap require that the ripios or leached/spent ore be removed after each leach cycle. Ripios off-loading options can be summarized as:

• Loaders or shovels feeding light trucks.

o Applicable to smaller operations o Lower capital but higher operating costs o In use at: Tintaya (Cu, Peru)

• Loaders feeding conveyors

o Not common but may be well suited to some nickel heap leaching. o Portable conveyors (e.g., “grasshoppers”) can be used on the pad, feeding an

overland conveyor to the ripios dump.

• Reclaimer system (see Figure 3.6)

o Most common system with larger production rates


8

o Economics improve as lift thickness and tonnage rates increase o In use at: El Abra (Cu, Chile), Radomiro Tomic (Cu, Chile), Gaby (Cu, Chile)

Figure 3.6 Reclaimer Bucket Wheel, Chile

Plant Residue Management

Plant residue generally consists of three components: iron precipitates (the largest fraction), sulfates (gypsum, magnesium sulfate), and small quantities of other process sludges, and can be produced in any of these forms:

• Conventional or thickened slurry

• Paste

• Filtering

The disposal or management options thus include:

• Conventional tailings impoundments (slurries or paste)

• In-pit disposal (slurry, paste or filtered tailings)

o Slurry or paste will require containment berms similar to those required for ex-pit disposal

o Filtered residue can be dry stacked in the pit and may require little or no structural containment

A number of authors have suggested that dry stacking of filtered tailings in a Residue Storage Facility (RSF) is the preferred method for all sites and all tailings streams when considering stability and environmental containment. Unfortunately, often the cost of producing filter cake is prohibitive. In most of the nickel heap leach operations studied by the authors, residue filtering was elected because of the process benefits (improved metal and acid recovery) and thus dry stacking of the residues was selected for tailings disposal. Since the residues contain very high levels of sulphates and magnesium, low pH and often elevated manganese, an environmental containment system (base liner and drains and closure cap) will often be required unless good geologic containment and suitable climate factors are present. Because the shear strength of the residues will generally be very low and the material is susceptible to liquefaction, some method of stabilizing the final slopes will usually be needed. Analyses have indicated that combinations of small stabilizing buttresses (which have a height a fraction of the ultimate dump height) and compaction of the exterior shell of the dump (with or without the use of fly ash or Portland cement) can achieve good stability as well as erosion protection (and often become part of the closure capping system). While the residue will usually have a low permeability, leachate will be produced by compression of the residue as well as normal leaching actions (especially in wet climates) and thus a system to collect and remove this water would generally be required. If the process selection does not lead to filtering of the residue, then the tailings management options would broaden to include conventional slurry, thickened slurry and paste, with either all the


9

residue streams comingled or managed separately, and management would be similar to that required for PAL or HPAL residue, with some differences in chemistry and thus environmental containment requirements. However, conventional systems can present a significant long-term risk that should be factored into the analysis (see following section on Physical Stability).

Mine backfilling is often considered in nickel projects and when circumstances allow can be a very good management method. Often either overlooked or underestimated is the complexity of coordinating mine planning with tailings management, and the need for proper containment and closure (either physical or environmental or both) for the tailings whether in- or ex-pit, as well as the complexities brought by high rainfall sites.

Physical Stability

Ripios: Multi-stack Heap or Ripios Dump

The geotechnical behaviour of laterites is significantly different from other ore types, and this is aggravated by strong acid leaching. Freshly agglomerated nickel laterite ores can have relatively good strength and permeability, but both non-agglomerated and leached agglomerates can have very poor geotechnical properties. Leached ore can also experience significant changes in physical and chemical properties with time and exposure to acid, creating complexities in modelling the behaviour as well as sampling and testing. Shear strength will decrease as the agglomerates degrade and a large fraction (20 to 30%) of the mass of the ore is dissolved. Permeability decreases (often by a factor of 100) due to the same factors and can be aggravated by reprecipitation of iron compounds as is common in two-stage leaching. As the permeability decreases the in-leach ore experiences increasing saturation, which increases the vulnerability to liquefaction. By the end of the leach cycle, which can be longer than one year, the permeability may be sufficiently low that the heap effectively does not drain. This makes either working over the leached lift (in a multi-stack heap) or removing the ripios (in a dynamic heap) complicated in that even after weeks of drainage the ripios can be at over 90% saturation. Tests from several sites indicated that CBR (a common measure of traffic support capacity) can be under 5, which is too weak to support even low ground pressure equipment without some form of ground improvement (e.g., treatment and compaction, addition of a layer of waste rock, use of reinforcing grids, or a combination of all of these approaches). When placed in a ripios dump the traffic support capacity can be further reduced since any cohesion or fabric retained in the heap is lost, and the act of dumping the ripios can create excess pore pressures, reducing the shear strength and destabilizing the slopes. Residue or Tailings Storage Facility

Tailings management facilities tends to be one of the higher risk activities in modern mining and a recent survey found 122 modern Tailings Storage Facility (TSF) failures. Of these, 75% were directly related to seepage or poor drainage issues

2, and nickel residue is one of the poorest

draining tailings. Recent significant TSF failures include: Omai (Guyana, 1995), Manila Mining (Philippines, 1995 & 1999), El Porco (Bolivia, 1996), Marcopper (Philippines, 1996), Las Frailes (Spain, 1998), Aurul/Baia Mare (Romania, 2000), Remin (Romania, 2000), and Zhen’an Gold (China, 2006). Nickel leach residues differ from conventional tailings in that they are largely composed of chemical precipitates (principally iron, magnesium and gypsum) rather than ground rock with just traces of chemical residues. Most of the projects studied plan on using filters for the residue to improve metal and acid recovery and thus slurry tailings may not be common in nickel heap leaching. However, even filtered residues can have relatively high degrees of saturation (above 80% has been measured, and for many geotechnical purposes anything above about 85% is effectively fully saturated) and thus have the same problems as the ripios, but because the residue is chemical precipitates, it lacks a granular media (rocks, sand) that creates shear strength and traffic support capacity. These residues also tend to have much lower permeabilities than ripios, with bench scale and pilot plant residues reporting as low as 1x10

-7 cm/s, a value commonly specified for compacted

clay liners. This makes the residue dumps highly prone to liquefaction in that the pore water cannot drain even under relatively slow loading conditions. In addition to having poor slope stability and very poor drainage properties, the residue can be prone to erosion.


10

Laboratory testing and limited field test pads have indicated that the addition of either Portland cement or pozzolanic fly ash can significantly improve the geotechnical characteristics of residues. In one example the CBR was increased from less than 2 (untreated) to 13 with 1% Portland cement (by dry weight). Cohesion will also increase which reduces susceptibility to liquefaction, and by increasing the Proctor optimum moisture content can also make the residue more workable (operators will report, for example, that fly ash will “dry back” the residue but it’s really increasing the optimum moisture content and thus causing the residue to behave as if it were drier). Dry-stack tailings or residue dumps will be limited in the allowable rate of rise (e.g., the maximum crest rise, in meters per year, allowable without inducing static liquefaction) due to creation of excess pore pressure and the potential for liquefaction and flow-slide failures. This will generally be on the order of a few to perhaps 10 or 15 m per year. The allowable rate of rise directly controls the area required for active tailings disposal, which affects capital expenses, water management (larger areas cost more and take in more rainwater) and closure costs.

High Rainfall and Raincoats

Most nickel laterites outside of Australia are found in tropical or sub-tropical climates characterized by both very high annual precipitation and intense peak storm events. This is not unique to nickel laterites, however, and a number of successful heap leach projects have been conducted in high rainfall climates, including: Panama, Myanmar, Ghana, Costa Rica, Peru, Brazil and the Philippines. In such sites the proper management of rain and storm water can be a key driver. One of the techniques almost universally used to manage the high rainfall is the application of temporary geomembrane covers or “raincoats.” More specifically, a raincoat is placed over the heap, ripios or residue dump to shed rainwater from the system before it enters the process circuit. An industry review completed in 2006

10 found 19 heap leach projects that have used or are planning to use

raincoats. Among these are current installations at Pierina (Au, Peru), Philex (Au, Philippines) and four projects in planning (two commercial gold plants in Northern Mexico and nickel pilot plants in the Philippines and South America). Raincoats were first used in heap leaching in the late 1980s on gold ore heaps in Costa Rica

13 to allow continuous wet season heap leaching in a very high-

rainfall climate. The covers provided several wet season improvements including:

• reduced surplus water and reduced water management issues.

• less dilution of process solutions for improved metal recovery.

• reduced reagent consumption in recirculated solutions.

• reduced likelihood of accidental spills due to excessive storm water accumulation or excessive flows in process solution channels or piping.

• reduced damage to the surface of the heap and ore agglomerates caused by falling raindrops (impact damage) and sheet flow (erosion).

Unlike semi-permanent to permanent covers used in other industries such as landfills, raincoats are generally used for short-term wet-season use, often with dry-season removal to aid in ore placement, irrigation network maintenance, and to encourage evaporation.

Interlift Liners for Multi-Stack Heaps

Spent ore will continue to consume acid beyond the economic value of the nickel recovered; thus, for the multi-stack leaching it is important to remove the PLS or ILS from the heap once it has leached the active lift. To accomplish this thin liners are placed over each leached lift of ore. These interlift liners generally consist of thin, non-reinforced geomembrane liner placed over the stabilized top of the prior lift. For the retreat stacking mode, the liner is installed as the radial stacker retreats (approximately 5 to 15 m per day depending on the tonnage rate and geometry of the active stacking area). The interlift liner system is advanced and drainage pipes are installed on very close spacing which is becoming the industry standard for copper oxide heap leaching (also to reduce acid consumption) and has been successfully employed at many operations including the Mantoverde copper project in Chile where this technology was first commercialized at large scale. For advance stacking, the interlift liner can be installed ahead of stacking or concurrently, at the operators’ discretion. An alternative to using a thin geomembranes is to compact the surface of the spent ore, if a sufficiently low permeability can be achieved. This option is currently being studied at one nickel laterite project and the initial data is favourable with compacted spent ore reporting permeabilities of less than 1x10

-6 cm/s. This would reduce operating costs slightly and, more importantly, simplify


11

operations and reduce potential conflicts between the liner installation crew and the conveyor operations. However, compacting the ore in wet periods may not be practical. Temperature Effects on Liner and Drainage Pipes

Most leach pad liners and essentially all drainage pipes are made of polyethylene (PE), which is a thermoplastic. As such, its physical properties can change significantly with temperature. Computer simulations on large copper sulphide heaps predict sustained temperatures of 40 to 50°C at the base of the heap. Anecdotal information from nickel laterite pilot plants suggests temperatures of over 70°C can be achieved. Limited laboratory data is available on the temperature effects on geomembrane puncturing (see Figure 3.7). Better information is available for drainage pipe performance, such as presented in Table 3.2, which presents modelling results calibrated with laboratory data. While laterite lifts tend to be thin, the overall height of a multi-stack heap or the ripios/residue dump can exceed 100 m. Note that PE pipes generally collapse at around 25% deflection

14,15. In both cases the data suggest that temperature should be considered

in design.

Figure 3.7 1.5mm thick LLDPE

1 tested for puncture at 120m simulated heap depth at (1055E)

23°C and (1058H) 60°C (white circles indicate failures)

Table 3.2: Vertical Pipe Deflection for Dual-wall 150 mm Nominal Diameter

Corrugated HDPE2 Pipe (as % of initial diameter)

Heap Height

(m) Pipe Deflection @ 23°C

(%)

Pipe Deflection @ 50°C

(%)

20 5.0 8.0

60 11.6 17.6

100 16.8 24.9

140 21.4 31

1 LLDPE: Linear Low Density Polyethylene; HDPE: High Density Polyethylene

2 HDPE: High Density Polyethylene


12

3.2 PROCESSING ISSUES

General

Processing issues are critical in a nickel laterite heap leach project because of the high gangue uptake per unit of nickel leached. Typically between 20 and 30% of the ore is dissolved to achieve a nickel recovery in the range 65 to 75%, as compared with 1 to 3% in a typical copper heap leaching project. The comparatively high acid consumption in a nickel laterite project has the following consequences:

(1) High acid usage. For a 1.3% Ni head grade, with 65% nickel recovery to product and

an acid consumption of 500 kg/tonne of ore, the acid usage becomes about 60 kgs acid per tonne of recovered nickel.

(2) Leach kinetics become critical. Longer leach times not only increase heap leach size, but can affect the acid usage, as gangue continues consuming acid. There is often an optimum leach time balancing nickel recovery and acid usage.

(3) Heap leach PLS processing needs to be carefully considered owing to the large amount of gangue leached alongside with nickel and cobalt, especially iron and magnesium uptake.

Therefore, although many of the techniques and methods developed for copper heap leaching can be applied to nickel laterites, there are significant differences which must be considered in the process design.

Critical Heap Leach Process Data

The critical heap leach data required for sizing and ramp-up are:

(1) An understanding of acid consumption versus metal (including impurity) uptake.

(2) An understanding of the impact of agglomeration on heap leach performance and the optimum agglomeration method.

(3) Leach kinetics for nickel/cobalt and key impurities.

(4) Development of the most appropriate heap leach cycle, for example, the extent of recycle needed to increase nickel tenors in the PLS.

(5) An understanding of scale-up issues in nickel laterite heap leaching.

(6) An understanding of the variability in heap leaching behaviour for different parts of the orebody, especially if there are significant variations in mineralogy.

These data need to be generated from both column and pilot scale test heaps, as nickel laterite heap leaching is relatively new, and scale-up issues are still being developed. Illustrations of an operating test heap from the Çaldağ Nickel Project are given in Figure 3.8

10.

Figure 3.8: Çaldağ Pilot Test Heap during Operation

Figures 3.9 and 3.10 present the key heap leach process results from a nickel laterite heap leach project under development. The illustrated response is typical of many nickel laterite projects, and


13

indicates why heap leaching has become a potential low cost alternative for many nickel laterite projects. Metal recovery versus acid usage data (Figure 3.9) shows that nickel and cobalt containing minerals are leaching preferentially to the bulk magnesium silicate and iron oxide containing gangue. However, there is an optimum nickel extraction, above which additional acid addition is largely being used to consume more gangue with diminishing nickel recovery. The order of metal recovery is cobalt > nickel > magnesium silicate minerals > iron oxides (largely goethite).

Figure 3.10 presents the leach rate data for test columns and a corresponding test heap (recovery versus flux), which provides the basis for commercial heap leach sizing and leach cycle development. This data is indicating that there was a reasonably good correlation between column and test heap results, particularly when sample variability is considered. Other projects, such as Metallica Minerals pilot scale tests conducted on samples of Bell Creek North ore, have shown similar results. Such scale-up data is considered to be critical for current nickel laterite heap leach projects until a better understanding of process scale-up issues are developed in projects. The columns and test heaps also allow a range of geotechnical data to be determined (as previously discussed). For example, only at test heap scale can trafficability and compaction issues be properly assessed, which are critical for a decision of multi-stacking versus dynamic heap leaching. As heap height and lift thickness are critical design parameters, not only affecting permeability (refer to Table 3.1), but also leaching rates, column and test heap data needs to be generated at heap heights similar to that expected in the commercial plant. For current nickel laterite projects, the expected range is 3 to 8 m. For this reason it is considered that small scale testing (say with 1 m high columns) cannot be employed for design criteria development of projects.

The leach rate data also provides valuable input into the design of the heap leaching circuit. For example, with slow leaching ores, recycle of intermediate leach solution (ILS) can be employed to improve nickel and cobalt tenors in the final PLS. The high levels of impurities present in the PLS limits the extent of recycle possible. Nickel and Cobalt Recovery from PLS

A range of processing options are being developed for nickel and cobalt recovery from heap leach PLS, many of which have been discussed in Willis

20. These options include:

• Mixed Hydroxide Precipitation (MHP)

• Mixed Sulphide Precipitation (MSP)

• Ion Exchange (IX) Options using a range of nickel and cobalt resins (including Lewatit TP207 from Lanxess, Dowex 4195 from Dow Chemicals, WP-2 from PSI, and MRT Superlig

®).

The PLS generated from nickel laterite heap leaching differs from pressure leach discharge liquor (where most current nickel recovery routes have been developed) owing to the high iron and aluminium-to-nickel ratios present in solution. A typical nickel laterite heap leach PLS can contain 2-4 g/L Ni, 15-30 g/L Fe, 2-5 g/L Al and 20-40 g/L Mg. The relatively high impurity levels can result in substantial nickel and cobalt losses during iron precipitation, often 15-20%, which is difficult to recover. This is particularly a problem if a high purity MHP product is required for further processing in a “QNi’ type circuit where the MHP product has tight iron and aluminium specifications. Although ferric iron can be precipitated in the pH range 2.5 to 3.5, aluminium requires a pH of 4-5, and ferrous iron a pH of 4-5.5 (with oxidation to ferric iron). The nickel and cobalt co-precipitation losses at these pH levels are substantial.

Options for reducing nickel co-precipitation losses are:

• Two stage iron precipitation circuits.

• The use of IX based systems for selective nickel extraction, especially from ferrous iron and aluminium.

• Considering an MSP type product where aluminium and ferrous iron removal are not required prior to the production of a nickel/cobalt containing product.

In addition to nickel and cobalt losses during iron precipitation, solid / liquid separation issues also need consideration, especially iron residue thickening and filtration (assuming iron residue is to be handled by dry stacking). The relatively high iron and aluminium levels in solution mean that solid /


14

liquid separation of iron residue needs to be carefully considered to avoid this area being a bottleneck in processing. Some options to be considered to improve iron residue solids handling include:

• Operation of iron precipitation at elevated temperatures to produce goethite or para-goethite.

• The use of seed recycle to improve the degree of compaction of iron residue, and therefore its physical handling properties.

• Careful evaluation of flocculants and coagulants during the evaluation program.

• Selection of the most appropriate iron residue filtration system, particularly minimizing soluble nickel losses to the iron residue. The iron residue washing circuit design can have a significant impact on the overall water balance, therefore needs careful attention.

Figure 3.9: Example of Key Metallurgical Data generated from a Nickel Laterite

Heap Leach Development Program – Extraction versus Acid Usage

0

10

20

30

40

50

60

70

80

90

100

0 100 200 300 400 500 600

Cu

um

ula

tiv

e E

xtr

ac

tio

n (

%)

Acid consumption (kg/dt ore)

Leaching Recovery vs Acid Consumption (Trial Heap)

Nickel

Magnesium

Iron

Cobalt


15

Figure 3.10: Example of Key Metallurgical Data generated from a Nickel

Laterite Heap Leach Development Program – Leach Kinetics

A second area to consider is the type of nickel/cobalt product. As many nickel laterite heap leach projects are being considered at the 10-30 ktpa range of nickel production, full value adding to nickel metal can be difficult to justify, both in terms of circuit complexity and capital costs (i.e. the reason for nickel laterite heap leaching is often to develop a low capital cost, low complexity project). Intermediate products then become the preferred product, but this requires a market for the product. Although more complex to produce, an MSP product may generally be preferable to an MHP as it is more ‘marketable’ as an intermediate product, especially as a ‘QNi type’ MHP is difficult to produce to the required specifications.

Manganese and Magnesium Control

The heap leaching of nickel laterites results in significant amounts of magnesium and manganese sulphate being dissolved into solution, which needs to be handled in the overall process flowsheet. This has two consequences – firstly soluble magnesium and manganese must be disposed of in the process circuit, and secondly high soluble levels of magnesium can impact on heap leach performance, therefore concentration levels must be maintained below levels which would adversely affect heap leaching (such as double salt precipitation in the heap).

Magnesium sulphate control is closely linked to water management. For dry climates, evaporation ponds can be employed, however even here the water balance can be an issue owing to reduced evaporation rates from saline containing solutions. For wet climates (with a net negative evaporation rate), the issue is more serious as evaporation ponds cannot be employed. In projects close to the coast, ocean disposal of neutralised effluent is an option, subject to environmental restrictions. This generally means controlling manganese to discharge levels in the range 1-10 mg/L, and producing an ocean discharge of similar composition to that of the surrounding ocean. For inland projects in wet climates, the problem of magnesium sulphate disposal is substantial, as

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6 7 8 9 10

Cu

mu

lati

ve

Ex

tra

ctio

n (

%)

PLS flux (m3/dt ore)

Cumulative Metal Recovery vs. PLS Flux

Ni (Trial heap)

Ni (Column)

Fe (trial heap)

Fe (column)

Mg (trial

heap)

Mg (column)


16

even fully neutralised liquor will not generally meet World Bank standards for sulphate disposal in fresh water systems. Vector has been involved in several tropical inland projects where sulphate containing effluent disposal is not possible. Some strategies being considered for these projects are:

• Rain covers on the heap leach to minimize rainfall ingress.

• Continuous sealing and capping of residue dumps to minimize sulphate-containing seepage.

• Adequate bleed neutralisation capacity within the process circuit.

• A good understanding of ‘dry’ and ‘wet’ water balance conditions to both ensure adequate water supply (dry conditions) and effluent treatment (wet conditions).

3.3 WATER MANAGEMENT ISSUES

Overall Water Management

Outside of critical geotechnical issues, water management is likely to be the most critical issue in nickel laterite project development. Although water management issues are project specific, for nickel laterite heap leaching the following will generally need to be considered:

• Water supply issues, even in wet climatic conditions.

• High rainfall management.

• Water balance and effluent disposal.

• Emergency pond sizing.

• Water storage pond sizing.

• The interaction between process and emergency ponds to minimize dilution of process solutions.

Figure 3.11 presents an overall schematic of water management within a nickel laterite heap leach project. The key water inputs and outputs are:

Inputs:

• Water associated with feed ore.

• Fresh water requirements for process plant

• Ripios / Residue Storage Facility (RSF) return liquor

• Rainfall feeding heap leach and ponds

• Run-off from heap leach (particularly covered heaps) and process plant

Outputs:

• Water associated with ripios and RSF discharge

• Water in product

• Evaporation losses

• Excess neutralised liquor to either evaporation ponds (dry weather project only) or effluent disposal (if permitted).

Based on these streams, project water balances can be developed based on different scenarios, in particular ‘dry weather’ conditions which will determine the maximum water demand from the project, and ‘wet weather’ conditions which will determine the potential for excess water generation, and therefore what steps are needed to either control excess water production, or to treat this liquor.


17

PROCESS PLANT

WATER

POND

BARREN

POND

RAIN

WATER TO BARREN POND

WATER TO PROCESS PLANT

PLS

POND

PLS

BARREN LIQUOR

MAKE-UP WATER

HEAP LEACH

EXCESS WATER

RUN-OFF TO WATER POND

RIPIOS

RSF

RUN-OFF RETURN

FRESH ORE

Figure 3.11: Nickel Laterite Heap Leach Water Balance


18

High Rainfall Management

High rainfall management is particularly relevant to wet tropical projects, but can be an issue in any weather climates. Tropical laterite projects can experience very high annual rainfall, with average annual precipitations of 2,000 mm to 4,000 mm common. Tropical storms can also be very intense, with peak rainfall of over 200 mm in a few hours. The high annual rainfall presents problems of surplus water, with wet months making some construction activities very difficult or impossible (including operation of the stacking equipment for ore, residue and ripios stacking as well as deploying liners), and the intense storms can damage agglomerates, create zones of channelling within the heap and otherwise impede leaching. To address these various concerns a multi-facetted approach can be integrated into the project, including:

• Raincoats (thin geomembrane liners) to shed rainwater from the heap and dumps, and to protect fresh agglomerates (discussed previously);

• Dual stacker systems for multi-stack heaps, using retreat methods for heap normal operations and advance stacking for the residue dump, all with low ground pressure equipment;

• Heap stacking may need to be discontinued during high rainfall periods (and thus the design should recognize this in the agglomerator and stacker availability);

• Fly-ash or Portland cement stabilization of the external shell of the dumps for both erosion and slope stability (discussed previously);

• Aggressive stabilization of the underlying lift of leached ore in multi-stack operations, using combinations of reinforcing grids, waste rock, and fly-ash or Portland cement stabilization of the ripios.

• Adequate bleed neutralization capacity to prevent effluent disposal.

Water Supply Issues

Because of varying climatic conditions throughout the year, steady state mass balances cannot be employed for determining maximum project water demand, and water storage requirements. Dynamic water balances of ‘dry’ and ‘wet’ project scenarios can be utilized for this purpose, applying the water balance model illustrated in Figure 3.11. The water balances should consider:

• Peak seasonal storage: Wet cycle (usually 100-year return) net cumulative water gain.

• Dry season water supply: Dry cycle (25- to 100- year return) maximum water demand.

For projects employing heap leach raincoats, surplus run-off water generated from the heap leach (and possible the residue dumps) can produce significant fresh water which can be used for water make-up to the water pond.

If inadequate fresh water is available to meet project water requirements, saline or hypersaline water has been considered as an alternative water supply. The use of saline water would have a substantial affect on project development as it can affect process chemistry as well as materials of construction.

Emergency Pond Sizing

Emergency pond sizing is critical in nickel laterite heap leach projects owing to the substantial liquid hold-up within a heap. The moisture level in an operating nickel laterite heap is in the range 30-40% depending on the mineralogy of the ore, which is much higher than most comparable copper heap leach projects. Similarly drain-down volumes are higher during power outage conditions, which must be accommodated in the emergency pond. Typical sizing criteria for the emergency pond is:

• Peak storm: 100-year, 24-hour return event.

• Power outage condition: heap drainage without return irrigation for a period of 12 to 48 hours (depending on conditions such as availability of back-up power and replacement pumps).


19

• Net cumulative water surplus: from a water balance model considering the design wet year (e.g., the 100-year return wet season) which considers both above-average rainfall and below-average evaporation.

3.4 ENVIRONMENTAL ISSUES

Effluent Control

The main issue in effluent control within a nickel laterite heap project is the handling of sulphate-containing effluents. Strategies for handling effluent disposal in different climatic conditions have been previously discussed in Section 3.3 as effluent control is closely associated with magnesium sulphate control. This is because of the substantial magnesium sulphate deportment into solution during heap leaching, which is not removed from solution during nickel recovery, and therefore needs to be treated in the bleed control circuit. Such effluent control strategies need to be considered under the range of weather conditions expected to be experienced in the project including extreme dry cycle (e.g., 100-year dry return period), typical ‘dry cycle’, ‘normal’ or ‘average; operating conditions, ‘wet’ processing conditions (e.g., 100-year wet return period) and peak storm events. To assist in this evaluation the authors have coupled steady state process models of heap leach / metal recovery plants with dynamic water input conditions, particularly rain and evaporation patterns during a simulated ‘wet’ or ‘dry’ years.

The advantage of heap leaching in these climatic conditions is the capacity to recycle solutions within the process circuit. For this reason, heap leaching can be a preferable alternative to pressure acid leaching or atmospheric leaching where sulphate containing effluent disposal is a major challenge. Ripios / RSF Run-off Control

The ripios and RSF run-off control system depends on project environmental conditions, with the most attention being needed for high rainfall conditions where water balance issues are critical. In this case, raincoats provide a method for limiting rain ingress into the ripios and RSF, presupposing the use of dry stacking for tailings disposal.

When raincoats are employed most of the runoff will be clean water and thus the management issue is principally one of managing large peak flow rates. However, part of the active face of the heap and dumps will always be exposed to rainfall regardless of raincoat usage, and these uncovered areas will be vulnerable to producing contaminated runoff. The most common and probably most reliable way to manage this water is to allow it to enter the process circuit and size the conveyance and storage components accordingly. This will also increase the fines entering the circuit, but with sufficiently large ponds the water should be satisfactorily clarified before entering the plant.

Raincoats are also not designed to the same standards as “environmental” containments (such as the leach pad liner) and are prone to both routine leakage and partial failure. For water balance modelling some leakage or bypass should be assumed, and the amount depending upon the quality of the raincoats and the degree of management to be used by the operator. The authors have generally used about 20% bypass for a well designed raincoat system. These bypass rates include allowances for localized failures or loss or raincoats (for example, due to wind damage).

For heaps without raincoats two approaches can be used. The inactive areas of a multi-stack heap can be sealed by compacting the surface and then that runoff directed away from the system with routine testing to verify that the water meets discharge standards. The active area of a multi-stack heap and all of a dynamic heap would take the runoff into the process circuit. This will dilute process solutions but in drier climates the water gain may be a benefit.

If a conventional (i.e. slurry based) TSF is to be employed for residue disposal, then the facility will need to be designed to conventional tailings standards including a robust dam and appropriate provisions for water management. For wet climate sites this will include a large spillway and some sort of management of any spilled water which will probably not meet discharge standards. This could result in serious water balance issues, and may mean that a conventional TSF is


20

inappropriate for the project. Raincoats can be considered, but to date these have never been used on a conventional slurry based TSF and would generally be impractical for all but the smallest facilities.

Containment & Cover Systems

Base liners for the leach pad, ripios dump and residue storage facility (or TSF) will generally need to meet environmental containment standards and thus require a high level of design, good quality construction and a rigorous system of inspection and quality assurance during installation. Conceptually, the liner “system” includes the foundation preparation and bedding layer, sometimes a compacted clay liner component, a geomembrane liner, and a drainage system over the geomembrane. Leak detection and collection may also be included depending on local standards and the level of reliability desired. A number of excellent papers have been published giving design and installation guidance and thus that will not be repeated herein, and interested readers are directed to Breitenbach et al

6,7,8,Thiel el al

18 and Smith et al

12,15,16, among other authors.

Cover or capping systems are not yet common in the mining industry but will likely become more so as environmental standards continue to increase and projects are advanced in increasingly challenging settings. Nickel residues especially are prone to post-closure problems and it is likely that nickel laterite heap leach projects will incorporate some form of closure capping to secure the residues. As with environmental liners, a wealth of information on cap design and construction is available including Breitenbach

4, Smith et al

13,17. One key feature of nickel residues is the

inherently low permeability of the materials, and the ability to enhance this with relatively low cost techniques such as fly ash or cement treatment, allowing the residues or ripios to become part of the capping system. An aspect of geomembrane technology that is currently subject to rapid advancement is construction quality assurance, especially in the area of defect detection post-installation and the use of geoelectrical methods. A number of recent papers have been published on this technology, speaking to both the technology and the economic and environmental advantages. The readers are directed to Beck, et al

2 and Thiel, et al

19 for more information.

Post Closure Stabilization of RSF and Ripios Dump

Most mine closure designs are addressing large quantities of relatively stable waste rock or well-contained tailings. For spent nickel leach ore and plant residues, however, the materials are both relatively weak and degrading with time (especially if any residual acidity remains). This poses a challenge for the designer in predicting the long-term physical and chemical properties of the residues. The fact that nickel leach products age is something not well recognized in geotechnical engineering and some (perhaps most) data produced from geotechnical laboratories is not properly connected to the state of the samples tested (that is, the history of their production, the handling and ageing post-production) and this causes some problems in applying that data in the design. Thus, a key point in nickel project closure – both heap leaching as well as more conventional PAL or HPAL – is to properly address this in the geotechnical program.

Beyond sample ageing some of the key closure issues facing nickel laterite heap leaching are (Smith

13,15,16,17, Ramey, et al

11):

• High rainfall: the need for and long-term performance of closure capping systems including the erosion control components.

• Settlement: heaps, residue dumps and RSFs will be subject to large post-closure settlement which can disrupt drainage courses, affect pipes and other structural components, and rupture the capping system unless conservatively predicted and properly accommodated in the design. Large settlements can also adversely affect slope stability both by changing the slope configuration and by reducing the effectiveness of diversion and drainage systems (and wet slopes are far more likely to fail than dry ones).

• Post-closure maintenance: this will be more important than in an “average” mine due to factors such as climate, the physical and chemical properties of the residues, and often the high seismicity of the sites.

• Effluent management: All of the waste facilities will produce some effluent for some period after closure (e.g. consolidation water from the dumps and heaps) which


21

could continue for several years after cessation of operations and completion of the closure construction. Some facilities, such as conventional slurry TSFs, can produce effluent on a long-term basis.

4 INTEGRATION OF DEVELOPMENT ISSUES

The previous section has discussed the main geotechnical, process development, water management and environmental issues which are critical in the development of a nickel laterite heap leach project. Because many of these issues affect other project areas, an integrated project approach is required. This is not always straight-forward, as often different engineering groups or consultants are responsible for separate areas of the project, and there is a decided culture of compartmentalization in mineral project development. By way of example, the following project interactions are illustrated: Heap Height

Generally the higher the height of a heap, the lower the permeability in the lower zones and therefore the maximum operating irrigation rate. However, there can be substantial process or operational benefits in operating at a higher heap height (such as reduced operating costs in multi-stacking or residue handling, higher specific nickel extraction rates, or lower impurity uptake); therefore, an optimization between these factors can be required. For this reason, it is recommended that testwork be conducted across the expected operating range of heap heights. Applying geotechnical and process data collected at one heap height to other heap heights can lead to problems, especially if there are substantial differences between project decision and available data. For this reason, heap height and ore variability should be investigated early in a project, particularly as column tests may need to be operated for considerable time (6-9 months in typical) therefore it is difficult to generate new data later in a project. Residue Management

Dry stacking of process residue in an RSF is generally the preferred residue disposal option, however, this requires additional capital expenditure in the process plant, and an evaluation of methods for producing handleable iron residue with minimal nickel losses via co-precipitation. Washing of entrained soluble nickel also affects the overall water balance and needs careful design. Similarly a conventional (slurry based) TSF can be considered to reduce processing costs, but may not be possible in high rainfall situations due to its adverse impact on the overall water balance. Type of Heap Leach Operation

The mode of operation of the heap leach can affect the overall water balance. For example a dynamic heap can require additional total area, therefore is subject to additional rain ingress when compared with a multi-stack operation. Additionally, a multi-stack operation can lead to additional surplus run-off water generated from the heap leach when compared with a dynamic heap leach, which could be critical in some projects where water supply is a major factor. Water Management

All aspects of the project impact on the water balance, especially in high rainfall situations, and need to be carefully considered. Many of these aspects have been discussed during the current paper including stacking systems, the use of raincovers, ripios and TSF disposal, containment systems, processing issues, effluent disposal and dynamic water balance considerations for a project. These examples highlight the need for a multi-dimensional approach to a nickel laterite heap leach development taking into account geotechnical, metallurgical, water management and environmental issues. Particularly important is the development of the metallurgical and geotechnical testwork program to provide key design criteria. The long timeframes for column and test heap programs generally mean that data collection and evaluation become the rate limiting step within a project, requiring serious planning as it is difficult to generate new information part way through a project.


22

The current paper particularly highlights the geotechnical and water management issues relevant to high rainfall projects. Strategies are highlighted for the management of high rainfall during heap leaching, ripios and residue disposal and processing of heap leach PLS. The use of dynamic water balance modelling is discussed in overview to evaluate ‘wet’ and ‘dry’ climatic conditions and their impact on a project.

5 ACKNOWLEDGEMENTS

The authors wish to acknowledge the support of BHP Billiton and European Nickel PLC who have been at the fore-front of the development of nickel laterite heap leach processing. We wish to particularly acknowledge Dr. Hou Yuan Liu, formerly of BHP Billiton NTC, who has been instrumental in the development of the technology since the 1990’s and can be considered to be the “father” of the heap leaching of “clay like” nickel laterites. Further thanks are due to the efforts and advice provided by Lourdes Valle of Vector Engineering, Inc. who assisted in the preparation of this paper.

6 REFERENCES

1. Agnatzini-Leonardou, S. and Dimaki, D., “Nickel and Cobalt Recovery from Low Grade Nickel Oxide Ores by the Method of Heap Leaching using Dilute Sulphuric Acid at Ambient Temperature”, Greek Patent No. GR1001555, 1978

2. Beck, A. and M.E. Smith, “Design Considerations for the Use of Geomembranes for Phosphate Tailings Impoundments,” to be presented at and published in proc COVAPHOS3, Marakesh, Morocco, March, 2009. 3. Beck, A., Smith, M.E. and Colmanetti, J., “Métodos Geoelétricos empregados na Minimização de Perdas de Solução em Pilhas de Lixiviação de Minério” Proceedings of the VI Congresso Brasileiro de Geotecnia Ambiental e V Simpósio Brasileiro de Geossintéticos, Recife, PE, Brasil, 18-21 June 2007. 4. Breitenbach, A. J. and Smith M. E., "Geomembrane Raincoat Liners in the Mining Heap Leach Industry", Volume 25, No. 2, pp. 32 to 39, April, 2007. 5. Breitenbach, A.J. and M.E. Smith, “La Historia de las Geomembranes en la Industria Minera,” in Minería & Medio Ambiente magazine, pp 8-11, Dec 2007. 6. Breitenbach, A.J. and R.H. Swan, Jr. “Influence of High Load Deformations on Geomembrane Liner Interface Strengths.” Proceedings of Geosynthetics ’99, Industrial Fabrics Association International, Boston, MA, USA, Vol 1, pp. 517-529, 1999. 7. Breitenbach, A.J. and R. Thiel, “A Tale of Two Conditions: Heap Leach Pad versus Landfill Liner Strengths,” joint proceedings of Geosynthetics Research Institute GRI-19 and NAGS conference, Las Vegas, Nevada, USA, Dec. 2005. 8. Breitenbach, A.J., “Improvement in Slope Stability Performance of Lined Heap Leach Pads from Design to Operations and Closure,” Geotechnical Fabrics Report, January 2004. 9. Duyvestyn, W., Liu, H. And Davis, M., “Heap Leaching of Nickel Containing Ore”, US Patent No. 6312500, 2001 10. Oxley, A., Sirvanci, N. and Purkiss, S., “Caldag Nickel Laterite Atmospheric Heap Leach Project”, Association of Metallurgical Engineers of Serbia AMES, 2006

11. Ramey, T.V., M. E. Orman and M. E. Smith, “Designing Leach Pad Liners for Differential Settlement”, Randol Gold Forum ’96, Squaw Creek, CA, USA, May, 1996. 12. Smith, M. E. and J. P. Giroud, “Influence of the Direction of Ore Placement on the Stability of Ore Heaps on Geomembrane-Lined Pads” published in Slope Stability in Surface Mining, Chapter 49, Society of Mining Engineers, Hustrulid, W.A., M.K. McCarter and D.J.A. van Zyl, eds, 2000.


23

13. Smith, M. E. and M. E. Orman, “Design Considerations for Closure Caps for Waste Piles and Tailings”, Randol Gold Forum ’96, Squaw Creek, CA, USA, May, 1996. 14. Smith, M., Beck, A., Thiel, R., and Metzler, P., “Designing for Vertical Pipe Deflection Under High Loads”, NAGS/GRI-19 Conference Proceedings, Las Vegas, Nevada, USA Dec 2005. 15. Smith, M.E., “Nickel Laterite Heap Leaching,” presented as part of the Emerging Issues in Heap Leaching short at Geoamericas 08, Cancun, Q. Roo., Mexico, March 2, 2008a. 16. Smith, M.E., “Emerging Issues in Heap Leaching Technology.” Paper no. 270, proc EuroGeo04, Edinburgh, Scotland, Sept 2008b. 17. Smith, M. E., “Putting Your Safety Cap On – Flexibility is the Key in Landfill Cap Design” Civil Engineering, January 1997. 18. Thiel, R. and M.E. Smith, “State of the Practice Review of Heap Leach Pad Design Issues,” Proceedings of the bi-annual meeting of the Geosynthetics Research Institute, Las Vegas, Nevada, USA, Dec. 2003. 19. Thiel, R., Beck, A. and Smith, M., “The Value Of Geoelectric Leak Detection Services For The Mining Industry.” Geofrontiers Conference Proceedings, USA, 2005.

20. Willis, B., “Downstream Processing Options for Nickel Laterite Heap Leach Liquors”, Nickel/Cobalt Pressure Leaching and Hydrometallurgy Forum (ALTA), May, 2008, Perth, Western Australia.


DEVELOPMENT OF HEAP LEACHING AND ITS INTEGRATION

INTO THE MURRIN MURRIN OPERATIONS

By

D. J. Readett and J. Fox

Minara Resources

Murrin Murrin Operations Pty Ltd, Australia

Presented by

David Readett

[email protected]

CONTENTS

ABSTRACT 2 1. INTRODUCTION 2 2. SCATS HEAP LEACHING 4 3. ORE HEAP LEACHING 7 4. COMPARATIVE HEAP PERFORMANCE 8 5. OPERATING STRATEGIES 10 6. SOLUTION INTEGRATION 13 7. HEAP LEACH PRODUCTION 13 8. CONCLUSIONS 14 9. ACKNOWLEDGEMENTS 14 10. REFERENCES 14


2

ABSTRACT

The Murrin Murrin Nickel Cobalt Joint Venture Project, located near Leonora, Western Australia, is operated by Murrin Murrin Operations Pty Ltd (MMO) on behalf of the MMJV. The project was commissioned in 1999 and utilises high pressure acid leaching (HPAL) and refining processes to produce Ni and Co briquettes from laterite ores. For a number of years, MMO has been investigating alternate technologies for the recovery of Ni and Co from its laterite resources. MMO’s research in relation to the heap leaching of Ni laterite ores culminated in the decision to build a Demonstration Plant in 2006. The 400,000tpa Heap Leach (HL) Demonstration Plant has been operational since early 2007. The HL Demonstration Plant has proven the technical and economic viability of the Heap Leaching of Scats and Ore. The rates and extent of Ni and Co recovery have been consistent with those predicted from the laboratory research conducted. The HL Demonstration Plant is now considered an ongoing integral component of the Murrin Murrin Operations, however there is still a significant effort being expended to further optimise the heap leach operations.

1. INTRODUCTION

For a number of years Minara has been pursuing alternate leach technologies to High Pressure Acid Leaching (HPAL) with a view to incorporate the alternate leach technology into the existing flowsheet and thereby reduce the overall production risk. The two technologies investigated were agitated atmospheric leaching and heap leaching. Preliminary laboratory based test work indicated that scats were amenable to heap leaching.

Heap Leaching is a mature technology and is applied widely throughout the gold and copper mining industry. The technology however, had yet to be established commercially for Ni laterites.

The majority of the heap leach testwork was centred around “column “ leaching tests (Readett, Meadows and Rodriguez 2006) with a key deliverable being to generate the process data and design criteria necessary to allow for future feasibility studies to establish the technical and economic viability of the heap leach projects

Key elements of the laboratory and site based testwork program were:

• Heap leaching of scats

• Heap leaching of ores

• Integration of heap leach solutions into existing process facilities

The testwork concluded that heap leaching of scats and ore was technically feasible and a basic design criteria was generated.

The primary justification for undertaking the Scats/Ore HL Demonstration Plant Project was to reduce the risk of future investment in heap leaching by addressing key Heap Leach scale-up risks including geotechnical stability, permeability, rate and extent of Ni, Co and Fe extraction, stackable ore depth, leaching cycle time, pad design and solution/stage leaching management. The Plant would also confirm the testwork findings that heap leach solution can be integrated into the existing MMO process facilities. Construction of the HL Demonstration Plant commenced in 2006 and it was commissioned early in 2007. The Plant was completed within the overall capital budget which totalled $49.6 million. The plant has operated for the past two years with no significant design or equipment issues. There has been no additional capital sought or spent to sustain the ongoing operation.

The HL Demonstration Plant consists of scats/ore feed system with agglomeration and radial stacking of the product onto an engineered impermeable leach pad. The stacked material is leached with acidified process solutions that are managed by a series of impermeable lined process pond and a pumping and solution distribution system. Leach solution enriched in Ni and Co are collected and pumped back to the process facilities for conversion through the existing refinery into saleable Ni and Co product.


3

Figure 1 – HL Demonstration Plant Schematic Flowsheet

Figure 2 – HL Demonstration Plant Layout


4

2. SCATS HEAP LEACHING

Pad 1 Cells 1 and 2 were stacked with agglomerated scats to a height of 4 metres. The scats stacked contained 30% Moisture, 1.01% Ni, 0.06% Co and 16.7% Fe. During the stacking process probes were placed within the heap to allow for continuous in-situ monitoring of heap stability, heap saturation/moisture, solution chemical potential and temperature. The metallurgical accounting indicated a recovery of approximately 70% Ni as shown in Figure 3 against a prediction of 74% for the same time period. Figure 4 shows Ni recovery plotted against solution application (which takes into account the period of reduced application rate) and this shows a rate recovery in line with prediction. Figure 5 shows the acid concentration in the leach solution applied and the acid concentration in the solution discharging from the heap. This shows that the heap has been consuming less acid with time and the acid concentrations are converging. The column test work had shown that when the acid concentrations converge then the leach material is nearing its terminal extraction. This is also consistent with a total acid consumption of 350kg/t.

Figure 3 – Pad 1, Cells 1 & 2 Ni Recovery


5

Figure 4 – Pad 1 Ni Recovery vs Solution Application

Pad 1, Cells 1 and 2 - Feed and Effluent Acid

0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

1 31 61 91 121 151 181

Days under Leach

Ac

id C

on

ce

ntr

ati

on

(g

/L) Feed Effluent

Figure 5 - Pad 1 Feed and Effluent Acid Concentrations

To assist in establishing true recovery a drilling program was undertaken. A dedicated fit for purpose drill rig was purchased as part of the Demonstration Plant as it was always considered that drilling would be required to establish a final metallurgical balance on each heap. 42 drill holes were drilled on a 9m x 10m spacing covering Cells 1 and 2. Samples were taken at metre intervals at depth. Each sample was dried, washed, dried and then the residue analysed.


6

The results confirm that to date 74.4% of Ni had been leached. This compares to the metallurgical balance calculation of 70% and a target predicted Ni recovery of 69% at 15kL/t. Figure 6 shows the recovery calculated from drilling compared to predicted recovery and the metallurgical balance. A Co recovery of 75% was also achieved. The drilling results were independently analysed by Arcadis and they verified the recoveries reported above.

Figure 6 – Ni Recovery (Drilled) vs Solution Application

The Ni recovery profile based on the residue also follows a similar trend to that defined in the column testing whereby the average recovery decreases with depth through the heap. This correlation on selected holes shows recoveries in the first metre of 85% trending down to an average of 72% at 3-4m depth (Figure 7). This result also confirms that further leaching of Pad 1 will result in the final average 80% recovery throughout being achieved.

Figure 7 – Pad 1 Ni Residue and Recovery with Depth


7

There was also a correlation with moisture content and Ni recovery with lower moisture content holes showing lower recoveries (Figure 8). This would indicate these locations had not received the same solution application as the heap average.

Figure 8 – Relationship of Moisture and Ni Recovery

The residue grades also confirm the strong relationship between acid consumption and Ni recovery. During the stacking process probes were placed within the heap to allow for continuous in-situ monitoring of heap stability, heap saturation/moisture, solution chemical potential and temperature. These probes have indicated that for the period of the leaching Pad 1 has remained geotechnically stable and has been maintained at below saturation. The probes have also indicated that from both a moisture and chemistry perspective there does not appear to be any significant channelling, stratification or variation either across or at depth on the heap. The tightly bounded moisture and recovery data from the drilling programme confirm the results from the probes. These results indicate that the process of heap leaching scats on a commercial scale has been proven. However, there still remains significant scope for further ongoing optimisation.

3. ORE HEAP LEACHING

The rear half of Pad 3 was stacked with ore. The ore selected was from the existing Murrin Murrin ROM and prior to delivery to heap leach it was screened to remove any +100mm material. The ore was agglomerated and stacked to a height of 4m during September and October 2007. These cells contained ~ 44,000 dry tonnes at 26% moisture containing 1.2% Ni, 0.1% Co and 24.7% Fe. The resultant Ni recovery profile is shown in Figure 9. The design criteria, which was conservatively based on the test work, anticipated 75-80% recovery of Ni in 540 days at an application rate of 40-55 kL/t. As can be seen in Figures 9 and 10 a recovery of 72% Ni was achieved in 180 days and only 12 kL/t of solution.


8

Figure 9 – Pad 3 Ore Ni Dissolution

Acid consumption for the ore was, as anticipated, under 450 kg/t and at 55 kg/kg Ni dissolved. A Co recovery of 59% and Fe recovery of 20% were also achieved. The Co and Fe are slightly lower than anticipated when compared to the Ni recovery. These results are consistent and comparable to the column leach tests. As with the first scats heaps, a series of probes were inserted into the ore heap to assess the degree of saturation, temperature and ORP of the ore heap, and to ensure that the heap remained geotechnically stable. Again the results from this investigation were consistent with previous data from the scats. The ore has proven to be geotechnically stable and it is possible to heap leach the ore at application rates of 5-20 L/hr/m2 without saturating the heap.

4. COMPARATIVE HEAP PERFORMANCE

Having analysed each heap in isolation, data for each heap was summarised, tabulated and then the relative performance of each heap could be assessed. The leaching is acid constrained so it is dependent on the volume and concentration of acid delivered during the leach cycle. The comparative data set indicates that the mode of acid delivery significantly impacts the leach efficiency. Simplistically, those heaps that operated at high average application rates and high average acid concentrations achieved significantly higher metal recoveries in less time than heaps operated at low application rates and acid concentration. This is highlighted in Figure 10 where Pad 1 Cells 1&2 and Pad 3 had high application rates at high acid concentrations whereas Pad 1 (total) and Pad 4 had low application rates at low acid concentrations. This results in leach times of -100 days to achieve 50% recovery versus +200 days and -150 days to achieve 60% recovery versus +350 days (Figure 11). So to fully optimise the capitalised leach pad area available optimum leach conditions must be maintained. Although the time frame for leaching is significantly reduced the actual total volume of solution that is required does not vary dramatically (Figure 12).


9

Figure 10 – Comparison of Rate of Ni Dissolution for a number of Pads

Figure 11 – The effect of acid addition on leach kinetics


10

Figure 12 – Comparison of rate of Ni dissolution vs solution application

For the early heaps the proving up of geotechnical stability was paramount and leach surface area was sacrificed. With the latter heaps where the stability was known a far greater area of heap surface has been utilised. The aim is to maintain high surface coverage (ore under leach) of 90-95%.

5. OPERATING STRATEGIES

A number of operating strategies have been put in place and adjusted over the last 2 years. They have been driven by the following factors:

• Levels of total suspended solids (TSS) in feed liquor (FL) solution

• Acid concentration in FL solution

• Acid utilisation

• Ni solution concentration “building”

• Budget metal production

• Ferric concentration in PLS, ILS and FL

• Downstream limitations, including H2S availability, shutdowns, repairs and plant availability. Figure 13 shows the configuration of the initial start-up. This was established to provide sufficient solution to provide for pad and pond solution and Ni solution inventory. Once the Ni concentration had increased sufficiently, the discharge solution was split between the PLS pond and the main operating plant. Due to the low acid concentrations in the FL, fresh concentrated sulphuric acid was added to the ILS pond to achieve +40g/L.


11

Figure 13 – Initial Start up configuration

In order to maintain as higher recirculating flow as possible, to maximise free acid from the plant and due to reductions in percolation rates and on-flow rates caused by elevated levels of TSS and calcium the feed flow was split over several pads, as below (Figure 14).

Figure 14 – Single Stage Leach

Initial Startup

Pad 1

FL ILS PLS Flushing Water

3400TK

09

Conc Acid

To 3400TK09 and downstream processing

Single Stage Leach

Pad 2


3400TK

09

Conc Acid

Pad 1Pad 3


12

The co-current two-stage leaching (Figure 15) was utilised to enable an increase in Ni concentration with every pass of solution. The TK09 solution was directed to the FL pond and then to the newest pad where the bulk of the acid was consumed, further acid was then added to the ILS pond before this was consumed and the solution discharge to the PLS pond.

Figure 15 – Co-current two stage leach

Two stage leaching and water flushing (Figure 16) is utilised when a fully leached pad needs to be flushed to recover the Ni retained in the pad solution inventory. Limited by pumping capacity, one cell at a time is flushed, the flushing pad off-low is discharged into the ILS pond and utilised as evaporation makeup water.

Figure 16 - Two stage leach and water flushing


Two Stage Leach + Flushing

Pad 2


3400TK

09

Conc Acid

Pad 1Pad 3


Two Stage Leach

Pad 2


3400TK

09

Conc Acid

Pad 1Pad 3


13

The counter current two stage leach (Figure 17) is the configuration currently utilised. It allows two stages of leaching, the ‘dirty’ FL solution (containing high TSS) is fed to the older pads where they are used as a filter, the ‘clean’ solution is then split between the FL pond and the ILS pond where fresh acid is added to increase the acid concentration to > 40g/L. The fresh scats on the new pads are then leached with ILS and the discharge goes to the PLS pond and onto 3400.

Figure 17 - Counter current two stage leach

The overall operating strategy for heap leaching has had to be adjusted to suit both operational constraints (2007 Major Shutdown, suspended solids in feed liquor, HDS project being put on hold), as well as economic constraints (2008 sulphur price, 2009 Budget). In all instances the heap leach plant/process has proven its flexibility and robustness.

6. SOLUTION INTEGRATION

Integration of the product solution into the existing MMO circuit has been achieved. This required an initial change to the H2S addition strategy and has required the implementation of Ni powder addition to 3700 to minimise the downstream acid neutralisation due to the additional Fe load from heap leach.

7. HEAP LEACH PRODUCTION

Stacking on Pad 1 started 1

st Feb 2007, since then 659,567 dry tonnes of Scats and ore has been

stacked over 7 pads (Figure 2). Nickel grade has been 1.054% and Cobalt grade 0.07% resulting in 6950t Ni and 465t Co being stacked. Over 2,300 t Ni has been exported to the MMO plant and the pond inventory has been built to > 400 t Ni. Pad inventory is also sitting at approximately 740t Ni. The amount of Nickel stacked versus Nickel dissolution is sitting at 50.5%, a total of 3508 t Ni and this is in line with expectation based on current modelling. The heap leach demonstration plant has been utilised on a number of occasions to make up production short falls in the MMO circuit. It has demonstrated an ability to produce from inventory between 10-15tpd of Nickel.


Counter-current Two Stage Leach + Flushing

Pad 2


3400TK

09

Conc Acid

Pad 1Pad 3Pad 4


14

The HL ponds have also been used to store excess HPAL solution where the containment ponds have been full and further downstream issues have prevented flow through 3500. The HL plant also has the ability to consume excess sulphuric acid allowing MMO circuit and utilities to maintain optimal production. These are all additional and significant contributions from the demonstration plant and further reduce the production risk at MMO. HL Ni production costs (including downstream refinery process charges) are currently lower than the HPAL production costs. By years end the HL costs are expected to fall below US$3.00/lb.

9. CONCLUSIONS

The Heap Leach Project has proven to be a commercial success for MMO with the key project technical and financial objectives being met. Heap Leaching of Scats and Ore, and subsequent integration of product solution into the MMO plant has been commercially proven. To date over 2,200t Ni and 110t Co product has been generated by heap leach. This is made more significant when considering that the commercialisation period for this novel technology, from initial test work, has only been 5 years. (European Nickel commenced testwork on the Caldag project in 2001 and Metallic commenced on its Nornico project in 2004 and neither have operational facilities). The initial target of 75-80% Ni and Co recovery has been reduced to 70-72% based on findings from the demonstration heaps. However, this is compensated for by increased leaching rates. With a return to historical sulphur pricing the heap leach economics and operating costs remain competitive and <US$3/lb is forecast by year end. The overall operating strategy for heap leaching has had to be adjusted to suit both operational constraints (2007 Major Shutdown, suspended solids in feed liquor, HDS project being put on hold), as well as economic constraints (2008 sulphur price, 2009 Budget). In all instances the heap leach plant/process has proven its flexibility and robustness. On a number of occasions the heap leach operation has also successfully been able to provide additional production, at short notice, to make up production shortfalls within the MMO plant. The success to date has paved a promising future for the project at MMO and when external factors allow, the technical and economic drivers for future heap leach expansion at MMO will be well understood.

9. ACKNOWLEDGEMENTS

The authors would like to thank all of the Minara personnel who have been involved and contributed positively to the success of this project. The ongoing positive attitude and ability of the Heap Leach crew to adapt to ongoing challenges has helped to ensure the success of this project. The early support from operational personnel through the design review construction and commissioning phases was invaluable.

10. REFERENCES

1. Readett, D. J., Meadows, N.E. and Rodriguez, M., “Murrin Murrin Heap Leaching Project”

AusIMM Metallurgical Plant Design and Operating Strategies 2006, Perth, Australia, 2006.



COBALT EXTRACTION &

REFINING SYMPOSIUM


THE USE OF BIOLEACHING FOR COBALT/ARSENIC TAILINGS

REMEDIATION IN ONTARIO CANADA

By

Paul Miller

BacTech Mining Corporation, Canada

Presented by

Paul Miller [email protected])

CONTENTS

ABSTRACT 2

1. INTRODUCTION 2

2. TAILS CHARACTERISTICS 3

3. CONSIDERATIONS FOR BIOLEACHING 3

4. PROPOSED PROCESS 4

5. KEY PROJECT DRIVERS 5

6. CONCLUSION 7

7. REFERENCES 7


2

ABSTRACT

BacTech Mining Corporation is a publicly traded Canadian company based in Toronto. For the last fifteen years the Company has been well recognised for commercial application of its proven bioleach technology for treatment of refractory arsenical gold concentrates. BacTech has now developed a reactor bioleaching process that can be applied to the remediation of polluted mine tailings in an economically beneficial manner. It is the company’s vision to apply bioleaching technology to abatement projects and remove harmful elements such as arsenic from the environment where this can be assisted by a positive cash flow from metal recovery. This short paper highlights a particular project located in the Provence of Ontario in Canada where the company will be applying its technology to the clean- up of silver cobalt arsenic tailings from old mine workings in the region. The paper gives some details of the proposed process and discusses the issues which make bioleaching environmentally and economically favourable for remediation projects.

1. INTRODUCTION

Historic mineral extraction practices in many mining regions of the world, involved virtually unregulated processing of ores with no disposal control systems for the unwanted rock. Many of these resulting tailings contain sulphides, toxic elements, and where previous technology limitations existed, economically recoverable quantities of precious or other valuable metals. The sulphides in these tails, readily react with the atmosphere to create an acidic solution referred to as acid mine drainage. This acidic solution is very efficient at liberating certain metals, resulting in pollution of the surrounding watersheds.

The town and surrounding area known as Cobalt in Ontario, Canada, inherited environmentally damaging mine tailings left behind by a silver rush in the early 1900’s. Silver and cobalt were first discovered in 1903 at the Cobalt Camp and by 1922, cumulative production had reached over 333 million ounces of silver (1). Given the fact that there were very few applications for cobalt metal prior to World War II, most of the cobalt was disposed of along with the tailings on surface or pumped into local lakes. Eighteen million tonnes of tailings were left in lakes, on shorelines and in open areas over a large region. These tailings contain high levels of arsenic that have been leaching into local lakes, streams creating a serious environmental legacy that must now be dealt with as it poses potential risks not only to the environment but also to human health.

The use of bioleaching to enhance metal recovery by controlled oxidation of sulphides as well as stabilisation of arsenic values has been a technique practised in the gold industry for treatment of refractory gold concentrates since the mid 1980’s. The treatment of base metal sulphide concentrates which contain smelter penalty elements is also receiving more serious consideration using similar techniques of bioleach processing. Such concentrates can be relatively low grade compared to those for smelting and contain a variety of toxic elements such as: arsenic; bismuth; antimony; mercury and cadmium.

BacTech has many years of experience in the field of bacterial oxidation or “bioleaching” and established its first plant in 1994 for treating refractory arsenical gold concentrates. Since this time the company has successfully designed, engineered licensed and built bioleach plants for clients in the gold industry. The Company is now poised to transfer its technology to environmental remediation applications that offer favourable economics through metal recovery. BacTech intends to build and operate a demonstration plant capable of treating 200,000 tonnes of mine tailings per annum (“tpa”) in Cobalt, Ontario, that will effectively remove the source of arsenic pollution in the region along with recovery of cobalt, silver, and nickel for sale to market. Successful operation of the demonstration plant will lead to the construction of a plant capable of treating 1,000,000 tpa over 15 years or more.


3

2. TAILS CHARACTERISTICS

As already mentioned, the tails in the Cobalt area were laid down over a period of time beginning in the early 1900,s. There were many different scales of operation and many different mines which resulted in the tailing deposits which are present today. While it is to be expected that both the arsenic and other metal contents varies between different tail deposits, the available information suggests that many of the tails deposits have similar mineralogical attributes and virtually all of them contain significant metal values for extraction.

The major polluting elements in the Cobalt tails are:

• arsenic

• sulphur

• cobalt

• nickel

• copper

• silver

• antimony and mercury

A varied sampling campaign reported in 2007 of various tailing sites implied that, on average, a thousand tonnes of tails may be expected to contain: 4.2t of arsenic, 1.5t of sulphur, 1.2t of cobalt, 0.6t of nickel, 0.4t of copper, and 75kg of silver (2). Clearly, with 18Mt of tails as an inventory, this represents a considerable polluting load to the environment. The three elements of arsenic, cobalt and nickel, have been cited as the most notable for migration in waterways. The presence of small amounts of mercury is to be expected because it is a natural element associated with the local silver mineralogy.

The mineralogy of the original ore type which was mined is quite unusual. Native silver occurs in veins associated with arsenides, sulfarsenides, and sulphides such as nickeline (niccolite), cobaltite, saffrolite, lolingite, rammelsbergite, gersdorfite, skutterudite, arsenopyrite, tetrahedrite, chalcopyrite, bonite, galena, sphalerite, pyrite and marcasites. Gangue minerals include calcite, dolomite, quartz and chlorite. Oxidation of the primary ore minerals produces secondary minerals in particular, erythrite, annabergites, and scorodite. Sulphates (eg gypsum and thenardite), Fe and Mn oxides and oxyhydroxides also occur as well as clay minerals. Much of the arsenic is separate from the sulphur species, with CoAs3 and NiAs being examples of common cobalt and nickel minerals which are present

3. CONSIDERATIONS FOR BIOLEACHING

Very early operations in the Cobalt region used crushing and simple gravity devices to concentrate coarse silver values and the remainder of the material was discarded. The recovery method was improved by the addition of flotation to capture finer silver values, but for many operations the tails still contained considerable metal values by to-days standards. Much of the arsenic which remains forms an intimate part of the cobalt and nickel mineralogy and is also a significant ingredient to the remaining silver values. Therefore a pre-requisite of the process technology to be used, is the ability to successfully manage arsenic both environmentally and economically. A further consideration is that in order to maximise metal and arsenic recovery, only relatively low grade concentrates can often be generated from the tails. The process must also be able to manage a wide range of metal types and be accountable for the fate of all elements in the final processing scheme. High cost oxidation processes, or those unable to manage arsenic effectively with total extraction, or those where the fate of some elements is unclear, cannot be considered. Commercial bioleaching for refractory gold concentrates using agitated reactors has demonstrated that this process is able to meet these types of criteria and readily justifies the development work necessary to adapt the process to remediation projects


4

4. PROPOSED PROCESS

A simplified flow diagram below outlines the proposed process for tailings treatment at Cobalt. Process development work is quite well advanced with the intention of establishing an on–site demonstration plant in 2010 with engineering to begin in late 2009. An important feature of the processing strategy is that over 90% of the arsenic is collected into a small concentrated mass of about 7% of the original tails volume before bioleaching treatment is applied. This results in the tails material being stripped of arsenic very early in the process and effectively over 90% of the material is disposed back to site immediately following the pre-concentration operation. For the proposed demonstration plant, laboratory metallurgical test work has shown that virtually all of the different types of polluting elements will report to a concentrate to be treated by bioleaching. The tails from the concentration operation will be further protected in the environment by a natural increase in the carbonate: sulphide ratio in the proposed operation. This is due to the majority of the carbonates in the feed reporting back to the tails during concentration, whereas sulphide values report to the concentrate for bioleach treatment

The small remaining mass of concentrate containing the arsenic, sulphides and other metals is fed into bioreactors with appropriate reagents where bacterial metabolism liberates arsenic and the other metals into an acidic and oxidized solution. It is a continuous process and after only five to six days residence time of treatment, benign silver solid is removed for further processing and sale to market and the solution is treated with limestone to precipitate ferric arsenate. According to Ontario Regulation 558, ferric arsenate is classified as non-hazardous and can be sent to a landfill site. The pregnant liquor containing the nickel and cobalt is precipitated to produce a mixed precipitate or metal/ precipitate for sale to market.

ARSENIC

TAILS

GRAVITY &

FLOTATION

REACTOR

BIOLEACH

NEUTRALISE PRECIPITATE

Or METAL

Arsenic & Metals in

Concentrate

Arsenic Free Tails

Benign Oxidised Solid to

Silver Recovery

Benign Stable Arsenate

to Disposal

Cobalt Nickel Saleable

Products

Mine

Water

Clean water

Solid

Liquid

SCHEMATIC FOR BIOLEACHING ARSENIC COBALT TAILS


5

The account of toxic elements in the bioleach processing scheme can be summarised as follows:

• Arsenic is captured to the concentrate and extracted into solution in the bioleach process. Neutralisation of the solution by increasing the pH with limestone will result in the binding of the arsenic to soluble iron values to form ferric arsenate. This is then separated as a precipitate for disposal

• Sulphur will also report to the concentrate and is converted to a weak sulphuric acid solution during the bioleach process. This is also neutralised in the same process as arsenic with limestone to form gypsum for disposal.

• The base metals of cobalt, nickel and copper are captured into the concentrate and extracted into solution in the bioleach process, together with the leachable iron present. Selective precipitation is employed to create separate base metal precipitate products for direct sale or for re-leaching and solvent extraction electrowinning routines to be used to produce pure metals.

• Silver report to the residue after bioleaching, which is then treated further to produce a high grade precipitate which is sent off-site for refining.

• When the bioleach liquor has been stripped of all metal values by selective precipitation, the clean water with a neutral pH is suitable for re-use within the process. In current commercial bioleach operations, much of the water is re-used in this way.

5. KEY PROJECT DRIVERS

5.1 SUCCESSFUL CONCENTRATE PRODUCTION

A key ingredient to the success of the remediation process is the ability to produce an economic concentrate for the bioleaching process which will use stirred aerated reactors. For metallurgical testing of the cobalt tails material, a bulk sample was taken from one of the tails deposit with the objective of testing the effects of using both gravity and flotation techniques to upgrade the tailings to give concentrates suitable for bioleach test work

Head assays for the tail sample were: 0.12% cobalt with 9 oz/t silver; minor gold values and 0.5% arsenic

Concentration tests to date have achieved over 92% recovery of arsenic into the concentrate at a 7% mass pull. In addition, gold, silver, and cobalt recoveries have averaged in excess of 76%. The high retention of arsenic in the concentrate, which is one of the main objectives of the test work, is very encouraging from an environmental rehabilitation perspective

These results were obtained using a combination of gravity and flotation methods on as received tails without any regrinding. It is hoped that recoveries may be further improved and mass pull reduced, by incorporating regrind steps where appropriate. For the initial upgrading work, screening was used to remove coarse material, followed by use of a cyclone for de-sliming. Spirals were then used as a low gravity separation technique on the remaining bulk of material to give gravity concentrates which were further upgraded by tabling. The spiral tails were then subjected to flotation to produce a further scavenger concentrate. Two concentrate composites were produced which represent suitable could feed stocks for a bioleach treatment facility. The first composite concentrate (Composite "A") consisted of the course and slimes fractions composited with the table concentrate from spirals and the scavenger flotation concentrate. For the second concentrate composite (Composite "B") tabling of the coarse fraction was conducted to reduce the mass of this material prior to compositing with the table concentrates from spiralling and flotation concentrate and slimes fraction. The recoveries obtained with these two concentrate composites are given below.


6

January 2009 results: COMPOSITE “A” Recovery of elements to concentrate at 10% mass

pulls using gravity and flotation (no grinding):

Ag Au Co Ni As S

78.3% 76.8% 77.6% N/A 93.2% 56.6%

January 2009 results: COMPOSITE “B” Recovery of elements to concentrate at 7% mass pulls

using gravity and flotation (no grinding):

Ag Au Co Ni As S

75.3% 75.8% 76.2% N/A 92.4% 52.9%

Metallurgical test work is continuing with the objective of further improvements being made to arsenic and metal recovery and to decreasing the concentrate mass for bioleach treatment. A discussion on the grades of concentrate produced is given below.

5.2 HIGH ARSENIC EXTRACTION; AND STABILISATION FOR REMEDIATION

Bioleaching is often valued for its ability to manage arsenic effectively creating a benign precipitate of ferric arsenate after its extraction, as demonstrated on existing refractory gold projects treating arsenical pyrites by bioleaching. Toxicity Characteristics Leach Procedure (TCLP) testing meeting US EPA requirements is often chosen as the method for demonstrating the stability of the arsenic in such precipitates. Other workers have conducted quite rigorous research to compliment this type of testing and confirmed the benign nature of such material (3).

Bioleaching test work to be followed by downstream liquor neutralisation routines for arsenic rejection is currently underway. This work is using test quantities of concentrate produced from the preliminary metallurgical test work described above. Bacterial adaption trials to the mineralogy type have been very positive in terms of demonstrating high extractions of arsenic and cobalt into solution using a mixed mesophilic culture. The final concentrate which would be generated for a commercial process is expected to contain between 7% and 10% arsenic which is quite manageable for a conventional bioleaching operation in which arsenic is leached to extinction into the bioleach liquor. Some concentrates for the tails remediation project may be deficient in soluble iron, and the amount of iron present may be insufficient to stabilise the arsenic in the downstream liquor treatment as ferric arsenate. The feed to the bioleach process may therefore be supplemented by other tails which are pyrite rich and carry minor precious metal values worth recovering. This is seen as a cost effective method for increasing the soluble iron content of the concentrate feed and ensure an excess of iron is present in the bioleach liquor for downstream arsenic precipitation.

5.3 HIGH METAL RECOVERY FOR REVENUE GENERATION

The low grade concentrates produced as feed to the bioleach facility are expected to contain between 1.5% to 2% cobalt with up to 1% nickel complimented by silver values of about 35oz/tonne or higher. Even In uncertain times of fluctuating metal prices, current modelling is showing the project to have good internal rates of return, ensuring that the project is sustainable over the time required for remediation of the entire tails inventory. The bioleach process is believed to be particularly cost effective for this project as the ratio of metal values to the quantity of sulphide for oxidation is much higher than for many bioleach facilities currently in operation.


7

6. CONCLUSION

Many of the challenges for transfer of the technology to remediation projects lie in the successful integration of bioleaching into upstream and downstream processing rather than in the bioleach operation itself. Hence much of the development work to-date is orientated towards investigations of concentrate production and also metal recovery routines, the latter of which are not covered within this paper.

The success of the demonstration plant to be established on–site in Ontario in 2010 will confirm that bioleaching can be used as a viable processing option for environmental remediation projects and lead to a commercial installation being established shortly thereafter. This will lead to a much broader scope of application for bioleaching of concentrates beyond its current accepted role in refractory gold concentrate processing. A number of other remediation type projects are also under review by BacTech which could ultimately lead to a new generation of bioleach plants dedicated to environmental clean-up.

7. REFERENCES

1. Dumaresq, Charles;” Historical and environmental legacy of mining in Cobalt, Ontario” dedicated website www.cobaltmininglegacy.ca.

2. Percival, Jeanne B, et al;“ Distribution of As, Ni, and Co in Tailings and Surface Waters in The Cobalt Area Ontario”; Presented at Mining and the Environment IV Conference , Sudbury, Ontario, Canada , October 19-27 2007.

3. Nyombolo, B.M; et al; MINTEK South Africa “Neutralisation of bioleach liquors “Colloquium Bacterial oxidation for the recovery of Metals, Indaba Hotel Sandton South Africa July 4 2004 Paper 2.


THE APPLICATION OF MOLECULAR RECOGNITION TECHNOLOGY

(MRT) IN THE PURIFICATION OF COBALT PROCESS AND

ELECTROWINNING STREAMS

By

Steven R. Izatt, Neil E. Izatt, Ronald L. Bruening, John B. Dale

IBC Advanced Technologies, Inc., USA

Presented by

Steven R. Izatt, [email protected]

CONTENTS

1. INTRODUCTION 2 2. MOLECULAR RECOGNITION TECHNOLOGY 2 3. THE COBALT MARKET 2 4. CLASSICAL COBALT SEPARATION PROCESSES 3 5. CURRENT COBALT REFINING FLOWSHEETS 4 6. MRT CAPABILITIES FOR PURIFICATION OF COBALT STREAMS 5 7. SUMMARY 6 8. REFERENCES 7


2

1. INTRODUCTION

It is extremely important to minimize content of critical impurities in the production of cobalt and cobalt salts for demanding applications such as batteries. This paper provides a summary of MRT capabilities for purification of cobalt streams, and the significance of this technology in the context of current and forecast market requirements for cobalt metal and chemicals.

2. MOLECULAR RECOGNITION TECHNOLOGY

MRT has been used since 1990 in base and precious metals refining. The hydrometallurgical MRT process uses SuperLig® products, composed of highly selective functionalities attached to solid beads such as silica, polystyrene or polyacrylate, that are specific for target ions of interest. The MRT process, and its application to the extraction, recovery and purification of various precious and base metals, has been described previously [1 – 3].

3. THE COBALT MARKET

Total cobalt consumption in 2006 was estimated at approximately 56,000 metric tons [4]. Total estimated consumption on a sector basis is provided in Figure 1 below [4]. Until recently, cobalt consumption has grown at an annual rate of 6.6% from 1996. In early 2008, the Cobalt Development Institute (CDI) forecast the following annual consumption growth rates for the foreseeable future: battery chemicals 12%, superalloys 5%, hardmetal 4 – 5%, magnets 5%, other (including pigments, catalysts, and adhesives 4%. The average annual total consumption growth rate was forecast at 3% per year, and total consumption by 2015 was forecast at approximately 80,000 metric tons per year. Due to the current recession, consumption growth has, at least temporarily, fallen off this trend. However, latest estimates indicate that, in spite of the recession, consumption for battery chemicals will continue to increase dramatically by 40 to 50% by 2013 – 2015 [5]. The indication is that cobalt bearing lithium – ion and nickel hydrate batteries will hold approximately 50% of this market. Please refer to Figure 2 below.

Figure 1: Estimated Cobalt Consumption in 2006 by Sector


3

Figure 2: Estimated Cobalt Consumption in 2015 by Sector

3.1 DRIVERS – RECHARGEABLE BATTERY MARKET

The optimized rechargeable battery has [5]:

• The highest energy density

• The best safety factor

• The longest life in terms of discharge cycles and ease of maintenance

• An environmentally friendly footprint

• High metal purity – material components in the battery are a major determining factor in meeting the above objectives. This highlights the need for efficient and cost-effective refining of cobalt that goes into the rechargeable battery market.

4. CLASSICAL COBALT SEPARATION PROCESSES

Classical cobalt separations processes usually involve the use of multi-stage batch process chemistry involving the following process stages:

• Dissolution

• pH changes

• Chemical additions

• Precipitation

• Filtration

Due to the inherent lack of selectivity of classical cobalt separation methods, multiple separation stages are required, thus resulting in high operational and capital costs. In addition, and most importantly, the efficiency of cobalt recovery using classical methods is usually only 50 – 80%.


4

5. CURRENT COBALT REFINING FLOW SHEETS

In more current cobalt refining processes, cobalt is recovered and refined mainly by hydrometallurgical processes in which the final stage of recovery is usually electrowinning. The most popular electrowinning recovery process for cobalt metal utilizes a sulfate medium. To optimize energy consumption and product quality, the electrolyte must be free of any significant concentration of impurity elements. Many cobalt metal and salt products also require very high purity for use in certain applications such as battery manufacture.

5.1 THE EFFECT OF IMPURITIES IN THE COBALT REFINING SYSTEM

Many cobalt refiners are now utilizing more impure primary and secondary materials for processing. The trend in the industry is to use modern electrolyte purification techniques. Extensive electrolyte purification is necessary to decrease the concentration of harmful impurities to acceptable levels. The impurities exert a deleterious effect on current efficiency [6, 7], as well as on the nature and purity of the cobalt deposit. The major elements of concern are usually: Fe, Cu, Ni, Zn, Mg, Cd, Pb, and Se [6, 7]. Cobalt solution purification can be achieved by such techniques as chemical precipitation, cementation, solvent extraction, ion exchange, and electrowinning, usually performed in a series of steps to achieve the desired results. Specific examples of removal capabilities for these various processes have been reviewed in the literature [7]. Each of these processes, however, has specific disadvantages, typically associated with the multiple steps required to achieve the desired cobalt purity. Such extensive processing results in high capital and operating costs and lower than desired yields.

5.2 COBALT PRODUCT SPECIFICATIONS

Cobalt metal and salts are produced in a wide range of purities. There is no single standard specification, particularly for high purity products. All the major cobalt refiners appear to have their own specifications for cobalt metal and chemical products, most of which are highly proprietary, and not readily publically available. Despite this lack of standardization, the trend is to produce higher purity products, particularly for end use applications such as batteries.

A typical specification for cobalt for battery production, from a major international cobalt refiner, is provided in Table 1 below.

Table 1: Typical Specification for Cobalt Used in Battery Production

Element Target Specification (ppm) Nominal Specification (ppm)

Fe < 5 < 10

Cu < 5 < 10

Zn < 5 < 10

Ni < 5 < 10

Cr < 5 < 10

Pb < 10 < 100

Ca < 100 < 200

Si < 10 < 50

Na < 100 < 300

K < 10 < 50

Cd < 10 < 50


5

6. MRT CAPABILITIES FOR PURIFICATION OF COBALT STREAMS

The elements for which MRT has removal capabilities from sulfuric and nitric acid based cobalt process and electro winning streams include the following: Fe [8], Cd [9], Cu [10], Ni [11, 12], Pb and Zn. Typical plant applications are discussed in the cited references. 6.1 SUMMARY OF KEY TECHNICAL OPERATING BENEFITS OF MRT COBALT PURIFICATION

PROCESSES

The MRT cobalt purification process provides the following key advantages and benefits:

• The SuperLig® product is highly selective for the target ion(s) of interest.

• There is no co-loading of other elements.

• Loading capacity for the target ion(s) is high.

• Due to the high selectivity, lack of co-loading, and high loading capacity for the targeted ion(s), the quantity of SuperLig® product required compared to ion exchange (IX) resins and solvent extraction (SX) solvents, is considerably less.

• The volume of chemicals required for the elution and washes are lower as less SuperLig® product volume is required.

• The process is environmentally friendly, and does not use or generate hazardous materials or solutions.

• The acid elution and washes are fully compatible with the refinery chemistry.

• A pure, concentrated, eluent product solution is generated.

• No resin or solvent regeneration is required.

• The capital and operating costs are comparatively low for the above reasons.

• The modular design is compact and can be readily expanded.

• The system can be fully automated.

• The system can be designed to treat any input concentration. 6.2 TYPICAL MRT SYSTEM LAYOUT FOR IRON REMOVAL FROM A COBALT STREAM

Figure 3 below provides a 3D view of a typical MRT system plant layout for removal of iron from a cobalt process feed stream. This system has been previously described in detail [8].


6

Figure 3: 3D View of a typical MRT System Plant Layout for Iron Removal from a Cobalt Stream

An example of an MRT column design is provided below in Figure 4. The appropriate SuperLig® material is loaded into the column.

Figure 4: Typical MRT Column Design

7. SUMMARY

Table 2 below provides examples of SuperLig® materials available for purification of cobalt streams, and for nickel/cobalt separations.

Table 2: Examples of SuperLig® Separations Used for Purification of Cobalt Streams, and for Nickel/Cobalt Separations

Application

SuperLig® Number

Extraction and Polishing of Iron from a Cobalt Stream in a Sulfuric Acid Matrix

SuperLig® 48

Cadmium Removal from Cobalt Electrolyte SuperLig® 177

Extraction and Purification of Copper from a Cobalt Process Stream

SuperLig® 77

Extraction and Polishing of Nickel from a Cobalt Stream in a Sulfuric Acid Matrix

SuperLig® 241


7

Extraction and Polishing of Nickel from a Cobalt Stream in a Nitric Acid Matrix

SuperLig® 199

Co- Extraction of Copper, Iron, and Nickel from a Cobalt Stream in a Sulfuric Acid Matrix

SuperLig® 176

Co- Extraction of Cobalt and Nickel together from a Nickel Laterite Ore Process Stream With Separate Elutions for the Nickel and the Cobalt

SuperLig® 138

Separate Extractions of Nickel, Copper, and Iron from Concentrated Acidic Cobalt/Base Metal Solution With Separate Elutions Producing Pure Salt Products

Solution pH 1 Cu: SuperLig® 86 Fe:: SuperLig® 14 Ni: SuperLig® 199 Co: SuperLig® 86

Separate Extractions of Copper/Iron, Nickel and Cobalt from Concentrated Acidic Cobalt/Base Metal Solution With Separate Elutions Producing Pure Salt Products

Solution pH 2 Cu/Fe: SuperLig® 145 Ni: SuperLig® 199 Co: SuperLig® 138

8. REFERENCES

1. Steven R. Izatt “Innovative Separations in Precious Metals Mining and Refining,” The Australian Journal of Mining, September 2008.

2. S. Bélanger M. Malone, N.E. Izatt, S.R. Izatt, J.B. Dale, and R.L. Bruening, “Selective Removal of Nickel from Cadmium- and Zinc-Rich Sulphate Electrolyte in the Zinc Industry,” Hydrometallurgy 2008, Phoenix, AZ, August 7 – 8, 2008.

3. N.E. Izatt, R.L. Bruening, K.E. Krakowiak, and S.R. Izatt, “Contributions of Professor Reed M.

Izatt to Molecular Recognition Technology: From Laboratory to Commercial Application,” Ind.

Eng. Chem. Res., 2000, 39, 3405-3411.

4. Cobalt Development Institute, www.thecdi.com/cobaltnews.php, "Outlook for

the Global Cobalt Market", Cobalt News, January, 2008, pp 3 - 7.

5. Reportlinker.com, accessed March 10, 2009. 6. A.E. Elsherief, “Effects of Cobalt, Temperature and Certain Impurities Upon Cobalt Electrowinning

from Sulfate Solutions”, Journal of Applied Electrochemistry 33: 43-49, 2003. 7. Shijie Wang, "Cobalt, Its Recovery, Recycling, and Application", Journal of Metals, Volume 58,

No. 10, October, 2006, pp. 47 - 50 8. Steven R, Izatt, Ronald L. Bruening, Neil E. Izatt, and John B. Dale, “The Application of Molecular

Recognition Technology (MRT) for Removal of Impurities from Cobalt Feed Streams in the Production of High Purity Cobalt and Cobalt Chemicals,” Alta 2008 Nickel/Cobalt Conference, Perth, Australia, June 16 – 21, 2008.

9. van Deventer, J., du Preez, R., Scott, S., and Izatt, S.R., “Cadmium Removal from Cobalt Electrolyte,” The Fourth Southern African Conference on Base Metals, Symposium Series S47, Swakopmund, Namibia, July 23 – 27, 2007.

10. Paul Kiggala, Amos Silungwe, Stanford Saungweme, Moses Mugabi, Steven R. Izatt, John B. Dale, Neil E. Izatt, and Ronald L. Bruening, “Environmentally Friendly Processing and Purification of Cobalt at Kasese Cobalt Company Limited Using Molecular Recognition Technology,” Alta Nickel/Cobalt Conference, Perth, Australia, May 21 – 26, 2007.


8

11. Steven R. Izatt, John B. Dale, Neil E. Izatt and Ronald L. Bruening, “Recent Advances in the Application of MRT to Nickel and Cobalt Separations from Primary and Secondary Process Streams,” Alta Nickel/Cobalt Conference, Perth, Australia, May 15 – 17, 2006.

12. Steven R. Izatt, Neil E. Izatt, Ronald L. Bruening, and John B. Dale, “Separation, Extraction, and Refining of Cobalt and Nickel from Base Metal Feed Streams Using Molecular Recognition

Technology (MRT),” Alta Nickel/Cobalt Conference, Perth, Australia, May 19 – 20, 2003.


DESIGN OF COPPER-COBALT HYDROMETALLURGICAL CIRCUITS

By

Graeme Miller

Miller Metallurgical Services Pty Ltd, Australia

Presented by

Graeme Miller

[email protected]

.

CONTENTS

ABSTRACT 2 1. INTRODUCTION 2 2. COBALT PRODUCT SELECTION 2 3. COBALT LEACHING 3 4. COPPER RECOVERY AND REMOVAL 5 5. SOLUTION PURIFICATION 7 6. SOLVENT EXTRACTION 9 7. PLANT MANAGEMENT 11 8. PLANT EXPERIENCE 11 9. CONCLUSIONS 14 10. ACKNOWLEDGEMENTS 14 11. REFERENCES 14


2

ABSTRACT

Copper-Cobalt hydrometallurgy has seen a resurgence in recent years with the development of

projects in Zambia and the DRC. Many of the flow sheets are unique in their use of

hydrometallurgical techniques to improve performance compared to the older styles in current use.

Many of the cobalt purification by precipitation steps have been replaced with solvent extraction; while

the remaining precipitation stages have been enhanced with use of oxidants to improve impurity

removal. Application of recent developments in precipitation circuit designs has improved solid-liquid

separation performance by better crystal growth. Operating experience from two of these circuits has

shown the benefits of the new technologies and the improved overall recovery of both copper and

cobalt into higher quality products.

The emphasis has moved from direct copper electrowinning / stripping with multiple stages of

precipitation for removal of impurities, and concentration of cobalt; to innovative use of solvent

extraction and ion exchange with higher cobalt recoveries to final product. The poor quality direct

electrowon copper has been supplanted with solvent extraction-electrowinning; with consequent

production of LME “A” quality metal.

1. INTRODUCTION

Copper-Cobalt hydrometallurgy has seen resurgence in recent years with the development of projects

in Zambia and the Democratic Republic of Congo DRC (formerly Zaire). Many of the flow sheets are

unique in their use of hydrometallurgical techniques to improve performance, compared to the older

styles in current use. The emphasis has moved from direct copper electrowinning / stripping followed

by multiple stages of precipitation for removal of impurities and concentration of cobalt; to innovative

use of solvent extraction and ion exchange with higher cobalt recoveries to final product.

Cobalt is recovered as a by/co-product of copper production and is generally treated as a commercial

bonus in the metallurgical copper circuits. As a result it has not been until recent newer projects have

come into the development programme, that a significant focus on cobalt hydrometallurgy has

become more important. For some projects in the DRC the potential cobalt income is now of the

same order of magnitude as the copper income (Anon, 2008). Maximising the recovery and value

addition to the product is a significant driver for hydromet process development.

It is impossible to divorce the selection of the cobalt recovery flowsheet from the specific markets to

be targeted. However there are a number of common unit operations; that are used to remove

specific groups of ions; that have all been advanced in the most recent round of process designs.

Many projects have taken increased development times in order to pilot the cobalt recovery part of the

process plant. The commonly used unit operations and their relative place in the process chain are

discussed in the body of this paper.

2. COBALT PRODUCT SELECTION

The selection of the cobalt product to be produced on site is vitally important in the development of

the required flow sheet. This motherhood statement is often overlooked by the project owners who

assume that the cobalt market is similar to other metal commodities. It is however an extremely

fragmented market, with many different product types, purity requirements and market demands.

One of the more important drivers for recent projects of modest size has been the availability of

excess cobalt metal refining capacity in Europe and China. Both of these markets will take an impure

cobalt intermediate salt and refine to final metal. The available capacity is ultimately limited, and new

large projects will need to address the market directly with finished (or semi finished) products for

direct input to customer processes. The diversification of the BHP Billiton Yabulu refinery into cobalt

chemicals is a good example of this change in product emphasis.


3

The potential hydrometallurgical cobalt products are extremely wide ranging. However it is the major

current markets that most green fields project are targeting. The more significant cobalt products

include:

o Cobalt Salts

• Hydroxide from either lime, magnesia or NaOH precipitation

• Carbonate usually using soda ash for precipitation

o Value Added products

• CoO from calcination of either hydroxide or carbonate

• Sulphate from purified leach liquors and evaporative crystallisation

o Cobalt metal has many grades of product depending on chemical purity and physical

condition. Some potential metal producers are considering the full range of post-production

metal enhancement with:

• Hydrogen de-gassing furnaces

• Size control

• Surface burnishing

• Specific packaging for market.

It is evident that there is no single target product similar to LME “A” grade copper. As a result there

are many possible cobalt flow sheets and processes to generate a marketable product that is

optimum of the specific operation. There are however a number of elements that are specifically

targeted to improve the realised value of the cobalt. The major ones include:

o Copper

o Iron

o Calcium

o Zinc

o Magnesium

o Manganese

o Aluminium

o Silica

o Nickel

Specific products require the removal of these ions to a greater or lesser extent. Particularly with

intermediate salts (for further refining) the level of penalty metals can be in the order of 2% to 4%

without too high a reduction in income. On the other hand production of super high quality electrowon

metal needs the preparation of high purity solutions with controlled maximum amounts of all the ions

mentioned.

3. COBALT LEACHING

Particularly in the DRC and to some extent in Zambia, the oxide ores with copper and cobalt are

leached in atmospheric systems. The dominant mineralisation is malachite/azurite with accessory

chrysocolla and minor other secondary copper minerals. The cobalt is present as heterogenite with

cobalt in both the Co2+

and Co3+

oxidation state. The cobaltic minerals are about 50 per cent of the

total cobalt, and are not direct acid leachable at normal temperatures and pressures. Alternate

methods of enhancing the cobalt leaching are required.


4

The leach trains are designed to leach the copper oxides in mildly oxidizing conditions, to recover the

minor cuprite, chalcocite and native copper. The leach time can be extended if there is significant

chrysocolla present (Miller, 2005). Copper leaching is generally taken close to completion in four to

eight hours. Those projects that do not have significant secondary copper minerals can achieve high

leaching efficiencies in as little as two hours from the rapidly leaching malachite and azurite (Crease,

2006).

About 50 per cent of the cobalt is leached (along with the copper) mainly from the Co2+

minerals. The

other cobaltic minerals need to be reduced to the 2+ state in order for them to leach. This is achieved

with controlled reductive leaching. The main reductant used to date has been sodium meta bi-

sulphite (SMBS: Na2S2O5) (Mwema et al). In an acid solution the SMBS disassociates to form SO2aq

which lowers the Eh in solution and reduces the cobalt oxidation state. SMBS is costly and is only

partially utilised with side reactions producing sulphuric acid – particularly in the presence of

manganese ions in solution.

A more recent development has been the use of gas from a sulphur burning acid plant (SO2 and N2)

as a direct Eh management tool. The gas is injected into the later parts of the leach train to enhance

the cobalt leaching. The SO2 is readily soluble and reduces the cobalt while undergoing conversion to

sulphate with acid production. The large volumes of nitrogen (about 88% v/v) can be an issue with

scrubbing of the SO2 from the solution, if the gas is not dispersed properly. One operation in Zambia

has used a Pressure Gas Disperser (PGD) in order to create ultra fine gas dispersion using a high

velocity pressurised jet of solution to entrain the gas (John, 2006). Another alternative is to disperse

the gas into a recycle stream of the leach slurry via a venturi eductor/mixer. Other newer projects are

considering the use of liquefied SO2 to remove the voluminous nitrogen (Grosse, 2007). This

eliminates the nitrogen scrubbing and makes the gas dispersion much more controlled.

Venturi Eductor/Mixer for Gas Dispersion into Leach Slurry


5

Two projects in Zambia, at Chambishi and Nkana, calcine a copper/cobalt/pyrite concentrate in a

sulphation roast, using fluid bed contactors (Sole et al, 2005). The copper and cobalt are in the

oxide/sulphate form and are leached (along with a lot of the iron) with dilute sulphuric acid. Since the

cobalt has been reduced in the sulphation roast to the Co2+

form it is readily soluble in the acid

solution. Both projects use recycled solution ‘barren’ in copper. This ‘preg builds’ the cobalt solution

levels to +10 g/L for subsequent recovery.

Other projects also undertake ‘preg building’ to increase the cobalt concentration in solution. Most

often this is accomplished by recycling a majority of the copper SX raffinate back to the leach. A

smaller bleed stream is treated for further copper removal and recovery of the cobalt. The technique

has some issues with the control of cobalt soluble loss in the solid/liquid separation steps. Most of the

oxide ores are very weathered and have poor filtration and thickening characteristics. As a result all

newer projects have opted for CCD trains. The wash liquor needs to be fresh water or recycled cobalt

plant effluent, low in cobalt, to enable the recovery of the dissolved cobalt along with the copper.

These CCD trains occupy a significant proportion of the project foot print and contribute a large

proportion of the project capital costs.

4. COPPER RECOVERY AND REMOVAL

4.1 PRIMARY COPPER REMOVAL

The primary metal production from most projects is copper. The existing Zambian and DRC

operations used direct electrowinning and electro-stripping to produce impure metal, that generally

had to be re-refined to produce a saleable product. The electrowin and electro-strip process is also

inefficient with current efficiencies as low as 65%. Both Zambian cobalt projects have committed to

changing the copper recovery to solvent extraction and electrowinning (Miller and Nisbett 2005;

Kordosky, 2008; Sole et al, 2005). T he Nkana project has already been in operation for some time

with excellent results (Mwakila, 2008). The copper is LME Grade ”A” quality and the plant copper

production has increased seventy five per cent. The production increase is directly linked to the

ability to add more copper into the circuit with the removal by SX. The direct EW tankhouse has been

reconfigured from starter sheets to stainless steel cathodes; and the capacity increased from 14,000

tpa to 30 000 tpa in a reduced footprint (Kordosky, 2008).

Rebuilt Nkana 30,000 tpa EW


6

The new generation of projects has gone directly for primary copper recovery from solution by solvent

extraction and subsequent electrowinning on to stainless steel cathodes. In conventional SX plants

this will recover up to 98% of the copper – still leaving 0.10 g/L to 0.20 g/L copper in the final raffinate

that progresses to the cobalt recovery. This level of copper is an issue with cobalt product quality and

needs to be reduced to less than 1 ppm (Kongolo et al, 2005).

4.2 SECONDARY COPPER REMOVAL

Preparation of copper SX plant ‘raffinate’ for downstream cobalt production needs the copper to be

reduced to less than 1 ppm for metal electrowinning. In the past this was achieved by co-precipitation

of the iron and any residual copper at pH +4.5; and disposal of the gypsum/iron/copper solids. The

copper was a direct loss from the process.

Traditional single stage copper SX will produce the raffinate described above. This represents a

significant loss of copper that could be recovered into the primary product. Cost benefit analyses on

three projects to date have shown that a unique SX Split-Circuit™ can be justified (Miller and Nisbett,

2006). This circuit (shown in Figure 1.0) takes the cobalt bleed steam and removes the copper to

consistently less than 10 ppm. The copper is recovered as EW LME Grade “A”. The key to the SX

plant performance is to integrate it with the precipitation removal of the other ions particularly iron.

Figure 1: Copper-Cobalt Split-Circuit™

The iron is precipitated at pH 3.2 with only marginal loss of copper. The iron free high pH solution is

contacted with standard copper SX extractant (in a second SX section – SX2) to remove the copper

quantitatively. The copper in the SX2 raffinate is thought to come mainly from the entrainment of EW

electrolyte in the stripped organic. The copper ‘free’ solution is sent on to further purification. Should

the final route be to metal, the solution quality of < 1 ppm copper can be achieved with a much

smaller IX plant or less reliance placed on purification via cobalt SX.


7

Kabwe Integrated Split-Circuit™ SX Plant – Large Units SX1, Small Units SX2.

Using MMS Side-Feed™ Mixer-Settlers.

5. SOLUTION PURIFICATION

5.1 IRON REMOVAL

Iron removal has been undertaken for many years in many hydrometallurgical process plants. The

classic method is air oxidation to ferric and precipitation with lime and or limestone. All the current

and previous Zambian and DRC project use this basic method. However the older style plants all

suffer from the usual problems of:

o downstream gypsum precipitation and

o fouling of process equipment and pipes.

Many of the new projects are using more recent process developments to minimise the calcium over-

saturation and subsequent gypsum precipitation issues. The major focus has been on:

o Use of air/SO2 as an enhanced oxidant for the iron and multiple stage addition of precipitant

(Papangelakis, 2004; Demopoulos, 2004)

o Higher temperature operation to enhance kinetics of crystal growth and kinetics (ibid)

o Recycle of seed crystals (ibid)

o Use of High Density Sludge HDS™ style techniques to improve precipitant utilisation and

further reduce calcium over-saturation (HGE, 2007).

The air/SO2 system has been the subject of many papers in the recent past (Krause E, 2007; Ho and

Ring 2007; Ferron CJ and Turner D, 1999). The system is reasonably robust and can reduce iron to

< 5 ppm with <1 ppm a common result. The major engineering issues with the system are:

o The very high gas volumes, when using acid plant feed gas.

o The change from reaction rate limited to gas dispersion limited in vessels over ten cubic

meters volume (Van Royeen, Archer and Fox, 2007)


8

o The competing kinetics of gas solubility reduction and crystal growth kinetics with increasing

temperature.

Optimum temperature is quite solution chemistry dependent, but generally falls in the range of 45 0C

to 50 0C (Krause, 2007). The effect of ions in solution is quite marked with reduction in Fe oxidation

rates of between 10% to 25% with H+, Cu

2+, Co

2+, Cl

- (Krause, 2007). As these ions are in most

solutions the rate of Fe oxidation needs to be confirmed from laboratory test work on ‘real’ process

solution.

The gas flow issues in the precipitation tanks are quite important as they are sufficiently high to cause

both SO2 stripping from the solution and to flood impellors – even with appropriately designed ones.

The key control criterion is the SO2/O2 ratio that can seldom be greater than 1:5. These factors are

drivers for eliminating significant gas volume by using liquefied SO2 from the acid plant gas. The

volume reduction will to a large degree overcome the solubility issues and subsequent OH&S issues

with SO2 in the ambient air. At least two new projects are using this technique for enhanced iron

removal.

Multiple stage reagent addition is just good chemical engineering to control the over concentration of

calcium. Papangelakis (2004) has shown the benefits of the method in the control of calcium

concentration in the final solution. He has also recommended that a final stage of solution maturation

without reagent addition be used to further minimise the calcium saturation.

Higher temperature operation has also been adopted in a number of projects where the benefits were

required. The first of these was in Zambia where direct steam injection was used to raise the

temperature. Both Zambian roast-leach projects have enhanced PLS temperature (as a result of

leaching hot calcine) and their subsequent iron removal is generally good without the use of oxidants

other than air.

Crystal seed recycle has been practiced for many years to allow growth of larger crystals and

minimisation of downstream gypsum fouling. The operations at the three Australian Nickel Laterite

plants showed how important this unit operation is to achieving high plant utilisation. None of the

plants utilised all the techniques mentioned here in their original design. Plant shut downs of one to

two days every three to four weeks were common; until better control of the iron precipitation step

was achieved. Seed recycle is an integral part of this control and has been included in all new

projects.

The HDS™ technique (HGE, 2007) and other similar patented processes (RAMS, 2007) were

developed primarily to provide larger crystals from dilute water treatment solutions. The key process

is to have large seed recycles (up to 2000% from dilute solutions) combined with a modified reagent

addition regime. The first reactor receives the recycled crystals and the precipitant. Here the

precipitant is adsorbed on to the crystal seeds. In the second reactor(s) the process solution is added

and reacts directly on the particle surface growing there, in preference to fresh nucleation in the

solution. This technique claims (Gabb et al, 1995) to reduce downstream calcium precipitation and to

improve solid-liquid separation rates. Many variations on this theme are available as technology

packages. Most need to develop specific ‘recipes’ to address the solution chemistry to be used.

However almost all the new projects have incorporated very flexible systems of tankage, reagent

addition, seed recycle and oxidation intensity to allow them to adjust their plant to suit changes in the

actual chemistry presented.

5.2 MANGANESE REMOVAL

Manganese removal is done mainly to achieve levels that are suitable for the product specification. It

is accomplished at a higher pH than iron and has greater potential to co-precipitate cobalt. As a result

it is often done as a separate step to iron, and the resultant solids recycled back to the leach to

recover some of the precipitated cobalt. This is similar to the technique used at Bulong for recovery

of co-precipitated nickel (O’Callaghan, 2003). The process used to date in Zambia and DRC has

been a combined iron and manganese precipitation. This has removed the copper and some zinc;

but also co-precipitated significant cobalt which has been lost. Enhanced methods of manganese

removal have been developed to minimise this loss.


9

The amount of manganese that can be precipitated depends on the level of Mn4+

present. But due to

the reductive leaching for cobalt, this is generally a small portion of the total manganese. The air/SO2

oxidation system has been investigated ( Zhang W, Singh P and Muir D, 2001; Ferron CJ and Turner

D, 1999; Wythe J and Vegter NM, Lunt et al, 2001; Schulze-Messing J, 2006) and included in a

number of circuits to lower the manganese concentrations. The level of reduction needs to be

determined for the specific product. In many cases removal to 1 g/L is sufficient if subsequent

operation is via Co SX-EW or for production of an impure intermediate product. Some manganese is

acceptable in Co EW electrolyte provided that anode maintenance is regular and the deposited MnO2

is removed.

If manganese removal is required to lower levels, then it has been found that the rate of reduction is

enhanced if a two stage process is used (Van Royen J, Archer S and Fox M, 2007). A primary stage

to reach 1 g/L; followed by a solid-liquid separation step. The liquor is then treated in a second

oxidative step to remove manganese to ppm levels. The kinetics are slow and long residence times

are required. The same issues with gypsum management and gas volumes need to be addressed in

the design of this circuit.

Bulong used an alternate method by precipitating the cobalt as a sulphide and leaving the manganese

in solution. The sulphide was re-leached in a small autoclave. This process was complex and

involved more stages of treatment. As a result it was discontinued in favour of producing an

intermediate sulphide product for sale.

5.3 ZINC REMOVAL

At this point it is possible to precipitate a reasonable quality cobalt salt without further purification.

However if zinc levels are elevated in solution they will be elevated in the product, as the zinc will

precipitate with the cobalt. Zinc removal in the traditional circuits has been via lime precipitation at

elevated pH and recycle of the solids to the leach for cobalt recovery. This has proved to be marginal

at best and a circulating load of zinc has built up. Both Chambishi and Nkana use a D2EHPA SX step

to extract the zinc prior to cobalt metal production (Sole et al, 2005). The Nkana plant is a simple

1E+1S that takes out a portion of the zinc to stabilise the circulating concentration to a low enough

level to be acceptable. Chambishi have a more sophisticated SX plant with multiple stages of

extraction, scrubbing, stripping and washing. They also include an HCl regeneration stage where

extracted iron is removed from the D2EHPA.

Other newer projects are also considering the benefits of Zn SX ahead of the cobalt purification plant.

Cyanex 272 has been proposed for this duty (Tinkler et al, 2007) with specific pH and scrubbing

conditions to target zinc rather than cobalt. Both D2EHPA and C272 suffer from issues of calcium

saturation in the strip solution that must be addressed by large volumes of Zn stripping solution to limit

the calcium to below saturation conditions in this waste stream. The solution can not readily be

reused for CCD wash as the zinc content will build up with the circulating load. As a result further

processing may be required to precipitate the zinc so that the water can be reused in the process.

A small D2HEPA Zn SX step was in operation at Bulong that removed this ion before cobalt EW. It

suffered severely from reagent poisoning with ferric whenever the iron and sulphide precipitation

plants were not operated well. No HCl regeneration was carried out.

6. COBALT CONCENTRATION

6.1 SOLVENT EXTRACTION

At this stage the older Co EW plants would direct electrowin from the cobalt solution. Some control of

nickel would be required, generally on a side stream to reduce the cost of the IX plant. Cobalt

precipitation and re-leaching were required to obtain the required water balance, neutralise the acid

from EW and to increase the cobalt concentration in the advance electrolyte. These precipitation

steps also lead to cobalt losses – especially from the less than perfect re-leach step.


10

More recent projects are considering the production of high quality metal by EW from an SX

electrolyte. The preferred reagent is Cyanex 272 or its analogues. The cobalt is removed from the

PLS at elevated pH and stripped in a very dilute acid electrolyte. The EW of cobalt in un-divided cells

is well understood, and the maximum operating concentration of acid is around 8 g/L in the spent

electrolyte. This corresponds to a cell and SX delta cobalt of around 5 g/L.

The engineering issues are around the selection of the pH control chemical – ammonia gas, ammonia

solution, NaOH solution or others. The lack of an ammonia gas infrastructure in Africa is leading

many operations to consider ammonia solution (Tati Nickel) or dilute NaOH (Sole et al 2005). In all

cases there is an issue of residue disposal unless steps are taken to eliminate the common ions:

ammonium or sodium from the effluent.

African Style NaOH Addition System

6.2 PRECIPITATION

Cobalt precipitation is also undertaken at this stage for production of a saleable salt or to reject water.

Precipitation has been conducted traditionally with lime slurry when re-leaching; or sodium carbonate

for a saleable product. However a number of projects have considered the use of MgO instead of

sodium carbonate. This is largely based on the costs of the reagents and the higher utilisation of

MgO than Na2CO3.

The precipitation is also likely to be conducted at elevated temperatures of up to 80°C. This is used

to enhance crystal growth rates and morphology and to achieve extremely high recoveries from the

enhanced kinetics. Crystal recycle is also an integral part of process to achieve coarser product size

distribution that has enhanced dewatering characteristics.

6.3 OTHER IONIC CONTROLS

Control of other ionic species is not generally practiced specifically unless one or another is

introduced as part of the process reagent(s). One operation in Zambia could not use local limestone

as it was running 4% acid soluble oxide zinc in the material. Likewise introduction of ammonia or

sodium carbonate / hydroxide have issues with the disposal of tails having high available nitrogen or

high sulphate (as the sodium salt).


11

Aluminium and silica are generally precipitated with the higher pH reactions. At a pH of 4.5 most of

the aluminium and silica precipitate with the iron and /or manganese removal. This is non specific

removal as both co-precipitate.

Use of MgO for precipitation of cobalt hydroxide introduces Mg into the solution. This can be

removed by lime precipitation of magnesium hydroxide, but at the cost of the lime to force the

reaction. Disposal of the calcium saturated water is now possible once it is re-acidified to pH 6.0 to

8.0.

7. PLANT MANAGEMENT

Many of the ancillary process steps are designed to minimise the effects of one process on the

succeeding process(s). The significant number of plant designs that are using sequential solvent

extraction processes will all require the minimisation of reagent carry over. Bulong suffered from carry

over of Cyanex 272 from the cobalt SX into the Versatic 10 nickel SX (O’Callaghan, 2003).

Insufficient consideration of this interaction has led to other operations also having difficult operations

(Kasese, 2007). Newer plant designs are utilising some or all of the following process steps to reduce

the intermixing of SX reagents:

o Settler designs with lower entrainment losses such as the MMS Side-Feed™ settler

o After-settlers and coalescers for removal of bulk entrained organic.

o Use of a diluent scrub stage for recovery of Cyanex 272. This was combined with a

saponification process at Bulong to recover the C272 in a concentrated stream (O’Callaghan,

2003).

o Dual media filtration to remove organics to < 5 ppm entrained.

o Carbon adsorption to reduce organics to < 1 ppm total entrained and dissolved.

Other process steps are designed to provide higher plant utilisation between shut downs for gypsum

removal:

o Clean-in-place acid circulation systems for dissolving gypsum from heat exchangers and key

process pumps.

o ‘Spare’ mixer-settler unit for on-line clean out – especially for zinc D2EHPA stripping stages.

There is generally a need to address the issue of silica in the primary copper SX plant PLS. The silica

will be removed in the iron precipitation, but only after the primary copper extraction has been

completed. Plant operation in organic continuity may be required, and the control of entrainment in

the loaded organic is necessary. Most new copper SX plants are either including loaded organic

coalescing or making design provision to allow easy retrofitting of the coalescing system.

8. PLANT EXPERIENCE

It is important to look at the experience of existing plants to try to build on the knowledge base that

exists. Many errors of engineering application can be avoided if the lessons can be included in the

new plant designs. Most of the unit operations have been run for long periods at one level of

sophistication or another.

8.1 AUSTRALIAN NICKEL OPERATIONS

The Australian HPAL nickel operations have all suffered to some degree from significant issues of

gypsum precipitation; which has resulted in lower than design utilisation and subsequent cash flow

issues. This is a prime lesson in making sure that all the points regarding process optimisation are

used for the iron / manganese precipitation processes.


12

None of the HPAL operations practice copper recovery specifically; and copper is removed by

precipitation as part of the iron removal process. Bulong had a final copper clean up prior to Co EW

using IX resins.

Cawse

The Cawse plant used an intermediate precipitation of mixed Ni-Co hydroxide with re-leach in

ammonia. Ni and Co were separated in ammonia solution using a LIX reagent. The cobalt was

subsequently precipitated as a salt for sale. Other than the gypsum issues, there are no specific

items that can be taken to new cobalt production facilities.

Bulong

Bulong went the full route to cobalt EW metal production in the initial design and operation. This

proved to be problematic due to the high reagent costs and large number of operators required to

keep the section running. The long train of processes to provide an electrolyte for EW metal

production was such that it created one of the bottle necks in the plant. It was subsequently modified

to produce an intermediate sulphide salt. The main issues to take from this operation are the

simplification of the process with an intermediate product, and the difficulty in operating the long train

of processes continually.

As part of their in-house development to improve plant utilisation, Bulong undertook a six month

commercial trial of an anti-scaling reagent (O’Callaghan, 2003). This was successful in reducing the

build up gypsum in the process vessels and pipes, and increased their run time from three weeks to

over two months. Although expensive, the reagent was cost justified on the basis of the reduced clean

out costs and improved plant utilisation and total production.

8.2 AFRICAN EXPERIENCES

Confidential Client DRC

This project is planning to make a high grade cobalt metal product for market. They currently operate

a facility making an intermediate cobalt basic sulphate. As part of this process the cobalt is reduced

with acid plant SO2 gas stream, and iron is precipitated with air and lime using indirect steam heating.

The operating plant has developed a number of techniques for minimising the downstream

precipitation of gypsum.

A semi commercial pilot plant has been running for more than eighteen months producing 500 kg/day

of high grade cobalt metal. This process uses the sequential zinc SX extraction with D2EHPA and

cobalt concentration with a Cyanex 272 analogue. Nickel, magnesium and other metal rejection is

high from a solution that is essentially iron free. Adjustment of pH in the Co SX is with NaOH in dilute

solution. This pilot plant is being used to develop design criteria for a full scale operation to produce

10 000 tpa cobalt metal.

Tati

LionOre (now Norilsk) have constructed and operated a semi commercial pilot plant for their Activox™

process at Tati Nickel in Botswana. The process involves:

o Initial copper removal by single stage SX-EW,

o Iron / copper / manganese / silica precipitation in two separate stages at different pH,

o Cobalt recovery with Cyanex 272 SX and

o Subsequent Nickel recovery with Versatic Acid 10 (the same circuit as used at Bulong).

Cobalt is precipitated as the carbonate for sale as an intermediate product. Many of the innovations

in iron precipitation and gypsum management are included in the plant design. This, combined with

other proprietary designs, has meant that the pilot plant has run without significant gypsum

precipitation in the cobalt or nickel SX plants.


13

Sable Zinc Kabwe

The SZK operation has the first Split-Circuit™ copper SX to produce very low copper concentrations

in the cobalt plant feed (Kordosky, 2008). It has also been the first to start up with SO2 reductive

leaching and air/SO2 iron and manganese oxidative precipitation. To date the leaching has been

successful while the precipitation is still undergoing parameter optimisation and operational

adjustments. The use of seed recycles and an elevated temperature has meant that there is only

slight gypsum scaling in the secondary copper SX plant.

Kabwe Secondary Copper SX with Minimal Gypsum Precipitation

Nkana

Nkana operations have improved with the conversion of the copper electrowin – electro-strip

operation to a Split-Circuit™ copper SX and stainless steel cathode EW plant (Mwakila et al, 2008).

The copper recovery has improved 5% and the cobalt recovery has also improved by a further +5%.

The copper product is LME Grade “A”. The increased cobalt recovery is attributed to the reduction in

pH required for the iron precipitation stage and the subsequent gain in cobalt no longer co-

precipitated with the iron. Further cobalt production has been possible by adding a bleed of raffinate

from the newly commissioned Nkana leach plant as the make up water to the calcine leach. This has

further increased cobalt production by around ten per cent.


14

Nkana Spilt-Circuit™ with MMS Side-Feed™ Mixer-Settlers

9. CONCLUSIONS

The selection of an appropriate cobalt recovery process is a tortuous path threading between product

specification, leach solution chemistry, cost/benefit analysis and operational complexity. There is no

one process selection that is appropriate to all operations – even those with similar chemistry.

Unit operations for use in cobalt circuits have undergone major development since the last significant

cobalt plant was constructed. As a result the operations of the next generation of cobalt plants are

likely to be more stable with higher recoveries in the long term; but fraught with the usual unforeseen

issues in the period immediately after commissioning.

10. ACKNOWLEDGEMENTS

Miller Metallurgical Services would like to gratefully acknowledge the permission of various clients to

publish this paper. The client bodies are also responsible for most of the pilot plant testing and

engineering proving of the improved unit operations. Without this individual and collective

commitment, none of the innovations now being put into practice would have seen the light of day.

11. REFERENCES

Anon, 2008. Katanga, Nikanor Merger Completed, Engineering and Mining Journal, 290(1):p18.

Crease, R, 2006. Personal Communication. July.

Demopolous, G P, 2004. Aqueous crystallisation and precipitation research at McGill, paper

presented to A J Parker Centre Hydrometallurgy Conference, Perth, April.

Ferron, C J and Turned D, Purification of Cobalt Solutions containing Iron and Manganese with

oxidation mixtures containing SO2 and Oxygen, International Patent WO 00/56943, 2002.


15

Gabb, P J, Howe, D L, Purdie, D J and Woerner, H J, 1995. The Kennecott smelter hydrometallurgical

impurities process, paper presented to Copper 05 – Cobre 95, Vol III Electrorefining and

Hydrometallurgy of Copper.

Grosse, J, 2007. Personal Communication. September.

HGE Corporation, 2007. HDS™ Technology Description.

Ho, E and Ring, B, 2007. Oxidants for Uranium Leaching, paper presented to ALTA Uranium

Conference, Perth.

John, L W, 2006. Personal Communication. August.

Kasese Cobalt Company Limited, World’s First Cobalt Bioleach Plant. Company presentation.

Kongola, K, Mutlae, C T and Kalenga M K, 2005. Contribution of nickel, zinc and sulphur co-

deposition during cobalt electrowinning. Journal SAIMM, October, 2005, pp 599-602.

Krause, E, 2007. Ferrous Oxidation with SO2-air an alternative uranium leaching system? Paper

presented to ALTA Uranium Conference, Perth.

Kordosky, G, 2008. The Copper of Africa – A Renaissance in Copper Hydrometallurgy.

Lunt BJ, Pryor BR, Collette FFJ, Whyte RM, Mann RW, Alexander D, A Method for the Recovery of

Cobalt, International Patent Application WO 03/054238 A1, 2002.

Miller, G, M, 2005. Analysis Of Leaching Data – Methods And Interpretation. First Extractive

Metallurgy Operator’s Conference, AusIMM, Brisbane, Nov.

Miller, G M, Nesbitt, A, 2005. Decreasing Operating Costs and Soluble Loss in Copper

Hydrometallurgy with use of Innovative Solvent Extraction Circuits. First Extractive Metallurgy

Operator’s Conference, AusIMM, Brisbane, Nov.

Mwakila, B, Silwamba, R, Miller, G, and Chisakuta, G C, 2008. Improvement In Copper And Cobalt

Production Due To Inclusion Of Solvent Extraction Of Copper At Mopani Copper Mines Plc’s Nkana

Cobalt RLE Plant, ISEC 2008.

Mwema, M D, Mpoyo, M and Kafumbila, K, Use of sulphur dioxide as reducing agent in cobalt

leaching at Shituru hydrometallurgical plant, J. South African IMM, Vol. 102 Number 1, Jan/Feb. 2002.

O’Callaghan, J, 2003. Process Improvements at Bulong Operations Pty Ltd, paper presented at ALTA

Nickel Conference, Perth.

Papangelakis, J F A, 2004. Aqueous scale formation and control, paper presented to A J Parker

Centre Hydrometallurgy Conference, Perth, April 2004.

RAMS Tech (University of Nevada, Reno), 2007. High Density Sludge Treatment,

www.unr.edu/mines/ramstech/1highdensity.asp [accessed September 2007].

Schulze-Messing J, Alexander DC, Sole KC, Steyl JDT, Nicol MJ and Gaylard P, An Empirical Rate

Equation for the Partial Removal of Manganese from Solution using a gas Mixture of Sulphur Dioxide

and Oxygen, Hydrometallurgy 86 (2007) pp 37-43.

Sole, K C, Feather, A M and Cole P M, 2005. Solvent Extraction in southern Africa: An update of

some recent hydrometallurgical developments, Hydrometallurgy 78 (2005) 52-78.

Tinkler, O, Flett, D, and Bourget, C, 2007. Flowsheet Considerations for Copper-Cobalt Projects in the

DRC, paper presented to ALTA Nickel Conference, Perth, May.

Van Royeen J, Archer S, Fox M, Manganese Removal from Cobalt Solutions with Dilute Sulphur

Dioxide Gas Mixtures, Proc Base Metals 2007, SAIMM.

Wythe, J and Vegter, NM, Manganese removal from Zinc Electrolytes with Sulphur Dioxide – Oxygen,

International Symposium on Lead and Zinc Processing, Japan 2005, MMIJ, pp1077 - 1090

Zhang, W, Sing, P and Muir, D, Oxidative Precipitation of Manganese with SO2/O2 and separation

from Cobalt and Nickel, Hydrometallurgy 63 (2002), pp127-135, Elsiver Science Press.


alta 2009 nickel/cobalt conference - alta ... 2009 nickel/cobalt conference may 25-27, 2009 sheraton...

Documents