5th International

Seminar on Process
Hydrometallurgy

editors

Fernando Valenzuela
Courtney Young

5th International
Seminar on Process
Hydrometallurgy

5th International
Seminar on Process
Hydrometallurgy

July 10-12, 2013

Santiago, Chile

editors

Fernando Valenzuela
Courtney Young

Copyright
Gecamin, Chile. All rights reserved.
No part of this publication may be reproduced,
stored or transmitted in any form or by any
means, electronic, mechanical, by photocopying,
recording or otherwise, without the prior
written permission from Gecamin.
Authors disclaimer
Any views and opinions presented in the articles
published in these proceedings are solely those of
the authors and do not necessarily represent those
of Gecamin. The authors take full and exclusive
responsibility for technical content, style, language
and accuracy of the information published herein.
This information is not intended nor implied
to be a substitute for professional advice. The
editors are not responsible for any damage to
property or persons that may occur as a result of
use of the information contained in this volume.
I.S.B.N. 978-956-8504-88-5

Gecamin
Paseo Bulnes 197, Piso 6
Santiago, Chile
Postcode: 833 0336
Telephone: +56 2 2652 1500
www.gecamin.com

Online proceedings
www.gecaminpublications.com/hydroprocess2013

Contents

11

Organizers

13

Committees

17

Foreword

19

Preface

21

Acknowledgements

chap. 1

Plenary presentations
Impregnated activated carbon for gold extraction from
thiosulfate solutions. C
 ourtney Young and Nick Gow

25
Geotechnical lessons learned from the operation of
Cerro Verdes Crush Leach Pad 4-A. H
 elbert Galdos,
Javier Guevara and Arnaldo Saavedra
35
chap. 2

Base metals hydrometallurgy

Daily process mineralogy: A metallurgical tool for optimized
copper leaching. Wolfgang Baum, Kevin Ausburn and Randy
Zahn

49
Direct production of high-purity cobalt sulfate heptahydrate
from a NICO deposit concentrate. A
 lex Mezei, Michael Johnson,
Cornelia Lupu, Ron Molnar, Keith Lee and Robin Goad
59
Copper extraction from oxide ore using sea water. Cynthia
Torres, Mara Elisa Taboada, Tefilo Graber, Hctor Galleguillos
and Helen Watling
69
Kinetics of chalcopyrite leaching in either ferric sulfate or
cupric sulfate media in the presence of NaCl. T
 cia Veloso,
Johne Peixoto, Michael Rodrigues and
Versiane Albis Leo
79
Mineral QTs in place leaching, standard sector. Hugo Letelier,
Patricio Gimnez, Carola Gonzlez, Luis Zenteno and Carlos
Castillo
87
Study of the steel dust leaching for the recovery of zinc and
cadmium. E
 rnesto de la Torre, Alicia Guevara and
Cynthia Espinoza
93

Zinc recovery from electric-arc furnace dust by hydrochloric

leaching and bi-electrolyte electrolysis. Jos Ricaurte and Ernesto
de la Torre
103
A step closer to the leaching of primary sulfides. G
 uillermo Velarde
111
Controlling the irrigation flow in heap leach piles of El Soldado
mine by canopy system with thermal camera. J avier Ruiz-del-Solar,
Omar Daud, Mauricio Correa and Marco Torres
119
Hydrothermal mineral replacement reactions and their
applications in mining and processing. J ing Zhao, Allan Pring, Jol
Brugger, Fang Xia, Kan Li and Yung Ngothai
127
Study of the recovery of silica from the zircon chemical treatment.
Laura Rocha and Carlos Morais
135
Maboumine process: Example of a promising process for
developing a polymetallic ore deposit. F
 lorent Delvalle, Valrie
Weigel and Antoine Greco
143
chap. 3

Cyanidation, leaching and recovery of gold

A theoretical study of SART precipitate generation: Operational
and safety impacts. Humberto Estay, Pablo Carvajal, Karina
Gonzlez and Vernica Vsquez

153
Development of a new technique to recover cyanide for gold
mining using membranes contactors. H
 umberto Estay, Miguel
Ortiz and Julio Romero
161
Leaching of sulfide gold ores using dithiooxamide, a non-cyanide
solvent of low-toxicity. J uan Pablo Serrano and Ernesto de la Torre
171
Copper cyanocomplexes adsorption on activated carbon: Effects
on the selectivity of gold dicyanide adsorption. C
 lauson Souza,
Virginia Ciminelli and Daniel Majuste
181
Cyanide leaching of copper-gold-silver ores. Humberto Estay, Pablo
Carvajal, Karina Gonzlez, Hctor Yez, Waldo Bustos, Sergio
Castro and Francisco Arriagada
191

Successful application of SART technology for

gold-copper ore deposits. Kresimir Ljubetic, scar Lpez, David
Kratochvil and David Sanguinetti
201
chap. 4

Solvent extraction and ion exchange

Analysis phase separation profiles in copper extraction. P
 atricio
Navarro and Sebastin Jara

211
Mo recovery process from high acidity solutions via solvent
extraction using CYANEX600. M
 auricio Salamanca, Alejandro
Quilodrn and Osvaldo Castro
217
Use of NR reagents in presence of nitrate ion in SX: A revision
of the present moment. R
 odrigo Zambra, Alejandro Quilodrn,
Gonzalo Rivera and Osvaldo Castro
225
Advantages and contributions of modifiers to copper extraction.
Rodrigo Zambra, Alejandro Quilodrn, Osvaldo Castro and Sara
Pascuale
235
Regeneration process for copper solvent extraction reagents.
Piritta Salonen and James Kabugo
245
Metallurgical performance of small and medium size copper SX
plants: Relevance of the SX reagent formulation. P
 hilippe Joly and
Francisco Reyes
253
Removal of some heavy metals from a sulfate-containing residual
mining solution using nano-structured calcium silicates. F
 ernando
Valenzuela, Jaime Sapag, Carlos Basualto and Luis Verdugo
261
chap. 5

Electrometallurgical processes and electrochemistry

Electrochemistry of enargite: Reactivity in alkaline solutions. Nick
Gow, Courtney Young, Hsin-Hsuing Huang and Greg Hope

265
Effects of organic impurities on zinc electrowinning. D
 aniel
Majuste, Virginia Ciminelli, Eder Martins and Adelson de Sousa
275

SELE Technology: An alternative to boosting current efficiency

and cathode quality in EW plants. Pedro Aylwin, Nicols Lagos and
Patricio Melani
285
Tankhouse parameters for transition from lead to alternative
anodes. Scot Sandoval, Casey Clayton, Ephrem Gebrehiwot and
Jason Morgan
295
CoolBar: A new intercell bar for electrolytic processes. G
 erardo
Cifuentes and Rodolfo Mannheim
303
Practical determination of galvanic corrosion of steel induced by
mineral particles. Genny Leinenweber and Luis Cceres
311
MIRS Starter Sheet Robotic Stripping Machine (SSRSM). L
 uis
Felipe Ramrez
321
chap. 6

Bioleaching processes
Bioaccelerant: A new biotechnological product for bioleaching
process optimization. M
 ara de la Luz Osses, Matas Saavedra and
Simn Beard

329
The effect of temperature on column bioleaching of secondary
copper sulfide ores. M
 ichael Rodrigues,
Klinger Lopes, Hamilton Lencio, Tcia Veloso and Versiane Albis
Leo
339
Leaching of copper oxide ore by in situ biological generation of
sulphuric acid. Debora Monteiro de Oliveira, Luis Sobral and Diogo
de Oliveira Padro
347
A real-time method for detecting active microorganisms in
commercial-scale biohydrometallurgical processes. D
 avor Cotoras
and Pabla Viedma
355
Bioleaching of covellite from low grade copper sulfide ore and
tails. V
 ctor Zepeda, Pedro Galleguillos, Cecilia Demergasso, Dina
Cautivo, Jos Soto and Yasna Contador
365

chap. 7

Modeling, and optimizing hydrometallurgical

operations and circuits
Implications of hydrodynamic testing for heap leach design.
Stefan Robertson, Amado Guzmn and
Graeme Miller

375
Dynamic model of copper recovery in heaps, Compaa Minera
Doa Ins de Collahuasi. E
 duardo Flores, Juan Pablo Garcs,
Christian Hu and Jess Casas
389
Dynamic simulator of ore preparation processes and leaching.
F rancisco Reyes, Gabriel Tejeda, Pablo Karelovic, Aldo Cipriano,
Miguel Herrera, Fernando Romero, Solange Rojas and Cristian
Salgado
401
DROP Project: Leaching process optimization for reducing
water consumption in copper mining. U
 lrike Broschek, Jorge
Lobos, Jorge Cornejo, Luis Bravo, Josu Lagos and Karien Volker
409
Rheometallurgical process risk mitigation through engineering
data generation. Alex Mezei and Mike Ashbury
417
chap. 8

Hydrometallurgical processes for producing

salt and non-metals compounds
Some improvements in Caliche heap leaching. Javier Ordez,
Silvia Valdez, Luis Moreno and Luis Cisternas

429
Author's index
439
Editors
443

Organizers

The Hydroprocess 2013 Seminar was organized by Montana Tech of the University of Montana, USA and Gecamin, Chile.

Montana Tech of the University of Montana

With more than 2,800 students on two campuses, Montana Tech is composed
of the School of Mines & Engineering; College of Letters, Sciences & Professional Studies; Highlands College; Graduate School; and the Montana Bureau
of Mines & Geology.
Montana Tech emphasizes teamwork, collaboration, and hands-on learning.
Montana Techs vision is to meet the changing needs of society by supplying
knowledge and education through a strong curriculum augmented by research,
graduate education and service. Its vision is to be a leader for higher education
and research in the Pacific Northwest in engineering, science, energy, health,
information sciences, and technology.
All programs derive a special character and emphasis from the unique
setting and continued tradition of high quality that has characterized Montana
Tech since its founding. Montana Tech has a long standing reputation for
producing outstanding graduates and is committed to research; resulting in
an unprecedented growth in its funded research over the last several years.
Learn more by visiting www.mtech.edu

Gecamin
Powering professional development
for sustainable mining
Gecamin is a Chilean company with 15 years of experience organizing technical
and international conferences for the mining industry. Our conferences aim
to inform and inspire professionals from all over the world, fostering the
exchange of best practices and innovative experiences.
Over 12,000 professionals have attended our events and have been trained
in areas fundamental to the mining industry. These areas include Geology and

11

Mining, Mineral Processing, Hydrometallurgy, Sustainability and Environment,

Water and Energy, Maintenance and Automation, and Human Capital.
Gecamin seeks to contribute to the sustainable development of the mining
industry by openly addressing its most pressing concerns and by offering a
platform for knowledge exchange that aims at identifying the most sustainable
solutions.
In 2012, Gecamin organized 10 conferences and 12 courses, with a total of
501 technical presentations, gathering more than 2600 delegates. A total of
128 mining houses from 37 countries were represented.
Learn more about Gecamin conferences by visiting www.gecamin.com

12

Committees

Executive Committee
chair
Gabriel Meruane, R&D Project Director,
SQM Industrial, Chile

Cleve Lightfoot, Global Practice Leader

Technology, BHP Billiton, Chile
Percy Mayta, Technical Services Manager,
Freeport-McMoran Copper & Gold Inc., Peru
scar Rosas, Plant Manager, El Soldado

co - chairs

Division, Anglo American, Chile

Courtney Young, Department Head and Lewis S.

Gustavo Tapia, Technological Innovation and

Prater Distinguished Professor, Metallurgical and

Process Manager, Antofagasta Minerals, Chile

Materials Engineering, Montana Tech, USA

Advisory Committee

Javier Guevara, Hydrometallurgical Process

Manager, Sociedad Minera Cerro Verde, Peru
2012 chair
Sergio Castro, Senior Processing Consultant,
Arcadis, Chile
executive director
Carlos Barahona, General Manager, Gecamin, Chile
t e c h n i c a l co o r d i nat o r
Fernando Valenzuela, Professor, Universidad

Pablo Amigo, Jacobs Chile S.A.

Corby Anderson, Colorado School of Mines, USA
Francisco Arriagada, Arcadis, Chile
Antonio Ballester, Universidad
Complutense de Madrid, Spain

Directing Members
Alejandro Dagnino, Mining Resources and
Development Manager, Minera Gaby SpA, Chile
Marcelo Jo, General Manager of Technical
Support, Xstrata Copper, Chile

de Chile

Cleve Lightfoot, Global Practice Leader

s e m i na r co o r d i nat o r

Percy Mayta, Technical Services Manager,

Fabiola Bustamante, Gecamin, Chile

Freeport-McMoran Copper & Gold Inc., Peru

a s s i s ta n t co o r d i nat o r

Division, Anglo American, Chile

Rebekah Zale, Gecamin, Chile

Directing Members
Alejandro Dagnino, Mining Resources and
Development Manager, Minera Gaby SpA, Chile
Marcelo Jo, General Manager of Technical
Support, Xstrata Copper, Chile

Technology, BHP Billiton, Chile

scar Rosas, Plant Manager, El Soldado

Gustavo Tapia, Technological Innovation and
Process Manager, Antofagasta Minerals, Chile

Advisory Committee
Pablo Amigo, Jacobs Chile S.A.
Corby Anderson, Colorado School of Mines, USA
Francisco Arriagada, Arcadis, Chile

13

Antonio Ballester, Universidad

Luis Snchez, BHP Billiton Pampa Norte, Chile

Complutense de Madrid, Spain

Nathan Stubina, Barrick Gold, Canada

Enrique Carretero, AMIRA International

Juan Carlos Tapia, AMEC International

Latin America Ltda., Chile

Engineering and Construction, Chile

Dick Celmer, Fluor, Chile

Julio Cesar Tremolada, Iberometex S.A.C., Peru

Virginia Ciminelli, Universidade Federal

Petrus Van Staden, Mintek, South Africa

de Minas Gerais (UFMG), Brazil

Guillermo Velarde, Sociedad Minera Cerro Verde S.A., Peru

Clenilson Da Silva Souza Junior, Instituto

Joan Vials, Universidad de Barcelona, Spain

Federal do Rio de Janeiro, Brazil

Ernesto de la Torre, Escuela Politcnica Nacional, Ecuador
George P. Demopoulos, McGill University, Canada
Claudia Diniz, Vale, Brazil
David Dixon, University of British Columbia, Canada
Steve Dixon, GoldCorp, USA
scar Ferrada, Minera Escondida, Chile
Rick Gilbert, Freeport-McMoran Copper & Gold Inc., USA
Manuel Guzmn, Molibdenos y Metales S.A., Chile
Miguel Herrera, Universidad Adolfo Ibez, Chile
Jorge Ipinza, JRI Engineering, Chile
Joaqun Martnez, Royal Institute of Technology, Sweden
Jorge Menacho, De Re Metallica Ingeniera Ltda., Chile
John Monhemius, Imperial College
London, United Kingdom
Miguel Monroy, Xstrata Copper, Lomas Bayas, Chile
Luis Moreno, Royal Institute of Technology, Sweden
Patricio Navarro, Universidad de Santiago de Chile
Felipe Nez, Minera Florida, Yamana, Chile
Manuel Olivares, ARA Worley Parsons S.A., Chile
Kwadwo Osseo-Asare, Penn State University, USA
Vladimiros Papangelakis, University of Toronto, Canada
Eduardo Patio, BHP Billiton Chile Inc.
Sergio Rivera, El Salvador Division, Codelco, Chile
David Robinson, CSIRO, Australia
Eloy Romn, Hochschild Mining, Peru

14

Technical Committee
Fernando Acevedo, Pontificia Universidad
Catlica de Valparaso, Chile
Jaime Alfaro, Xstrata Copper, Tintaya, Peru
Luis Bergh, Universidad Tcnica Federico Santa Mara, Chile
Germn Cceres, Universidad de Atacama, Chile
Francisco Carranza, Universidad de Sevilla, Spain
Jess Casas, Process Consulting, Chile
Gerardo Cifuentes, Universidad de Santiago, Chile
Luis Cisternas, Universidad de Antofagasta/
CICITEM/CSIRO, Chile
Francisco Cubillos, Universidad de Santiago de Chile
Manuel Chvez, Freeport-McMoran, Minera El Abra
Francisco Daz, The Chilean Commission
for Nuclear Energy, Chile
Jos Fernndez, Mantos Blancos
Division, Anglo American Chile
Juan Pablo Garcs, Ca. Minera Doa
Ins de Collahuasi, Chile
Juan Carlos Gentina, Pontificia Universidad
Catlica de Valparaso, Chile
Tefilo Graber, Universidad de Antofagasta, Chile
Nlida Heresi, JRI Ingeniera, Chile
Christian Hu, Ca. Minera Doa Ins de Collahuasi, Chile
Jos Hernndez, Universidad de Chile

Juan Patricio Ibez, Universidad Tcnica

Jess Casas, Process Consulting, Chile

Federico Santa Mara, Chile

Sergio Castro, Arcadis, Chile

Hugo Letelier, El Teniente Division, Codelco, Chile

Gerardo Cifuentes, Universidad de Santiago, Chile

Hugo Maturana, Universidad de La Serena, Chile

Luis Cisternas, Universidad de Antofagasta/

Gonzalo Montes-Atenas, Universidad de Chile

CICITEM/CSIRO, Chile

Rafael Padilla, Universidad de Concepcin, Chile

Davor Cotoras, Biohidrica, Biotecnologas del Agua, Chile

Carolina Paipa, Universidad de Playa Ancha, Chile

Clenilson Da Silva Souza Junior, Instituto

Nelson Parra, JRI Ingeniera, Chile

Federal do Rio de Janeiro, Brazil

Geysa Pereira, Hydrometallurgy, Vale, Brasil

Ernesto de la Torre, Escuela Politcnica Nacional, Ecuador

Eduardo Robles, Hatch, Chile

Javier Delgado, Novigi Ltda., Chile

Eladio Rojas, Chuquicamata Division, Codelco, Chile

Diana Endara, Escuela Politcnica Nacional, Ecuador

Leonardo Romero, Universidad Catlica del Norte, Chile

Humberto Estay, Arcadis, Chle

Vernica Rueda, SNC Lavalin, Chile

Helbert Galdos, Sociedad Minera Cerro Verde S.A.A., Peru

Andrs Soto, Universidad Mayor, Chile

Nick Gow, FLSmidth, USA

Mara Elisa Taboada, Universidad de Antofagasta, Chile

Tefilo Graber, Universidad de Antofagasta, Chile

Jaime Tapia, Universidad Arturo Prat, Chile

Alicia Guevara, Escuela Politcnica Nacional, Ecuador

Diego Verdejo, Antofagasta Minerals, Chile

Nlida Heresi, JRI Ingeniera, Chile

Miguel Herrera, Universidad Adolfo Ibez, Chile

Editorial Committee

Juan Patricio Ibez, Universidad Tcnica

editors

Federico Santa Mara, Chile

Fernando Valenzuela, Universidad de Chile

Marcelo Jo, Xstrata Copper, Chile

Courtney Young, Montana Tech, USA

Hugo Letelier, El Teniente Division, Codelco, Chile

Daniel Majuste, Federal University of Minas Gerais, Brazil

copy editor

Hctor Mlaga, Sociedad Minera Cerro Verde S.A.A., Peru

Rebekah Zale, Gecamin, Chile

Joaqun Martnez, Royal Institute of Technology, Sweden

reviewers
Versiane Albis Leo, Universidade
Federal de Ouro Preto, Brazil
Jaime Alfaro, Xstrata Tintaya, Peru
Francisco Arriagada, Arcadis, Chile
Pedro Aylwin, New Tech Copper SpA, Chile
Antonio Ballester, Universidad
Complutense de Madrid, Spain
Carlos Basualto, Universidad de Chile

Jorge Menacho, De Re Metallica Ingeniera Ltda., Chile

Gabriel Meruane, R&D Project Director,
SQM Industrial, Chile
Alex Mezei, SGS Mineral Services,
Metallurgical Operations, Canada
John Monhemius, Imperial College
London, United Kingdom
Luis Moreno, Royal Institute of Technology, Sweden
Caroline Muzawasi, University of Cape Town, South Africa

15

Patricio Navarro, Universidad de Santiago de Chile

Rafael Padilla, Universidad de Concepcin, Chile
Vladimiros Papangelakis, University of Toronto, Canada
Jochen Petersen, University of Cape Town, South Africa
Li Qian, School of Minerals Processing and
Bioengineering, Central South University, China
Francisco Reyes, DICTUC S.A., Chile
Stefan Robertson, Mintek, South Africa
Julio Romero,Universidad de Santiago de Chile
Javier Ruiz-del-Solar, Universidad de Chile
Scot Sandoval, Freeport-McMoRan Copper & Gold Inc., USA
Ruberlan Silva, Vale, Brazil
Luis Sobral, Centre for Mineral Technology, CETEM, Brazil
Andrs Soto, Universidad Mayor, Chile
Clauson Souza, Federal University of Minas Gerais, Brazil
Mara Elisa Taboada, Universidad de Antofagasta, Chile
Luis Alberto Texeira, Pontificia Universidade
Catlica do Rio de Janeiro, Perxidos do Brasil
Csar Ugarte, Hatch, Peru
Petrus Van Staden, Mintek, South Africa
Diego Verdejo, Antofagasta Minerals, Chile
Jacques Wiertz, Universidad de Chile
Rodrigo Zambra, Cytec, Chile

16

Foreword

In the last three decades the role of hydrometallurgy has changed. What
started as a secondary process alternative is now recognized as a primary
option with key advantages. Hydro-processes are able to treat lower quality
minerals, reduce the environmental impact of mining and recovery processes
significantly, and require lower initial capital costs thus giving opportunities
to small and medium sized producers.
However, as an established technology further growing faces new challenges.
Complex ore mineralogy, including lower grade materials, forces us to look
for continuous improvement in chemical processes. It is challenging to move
over the standard operation and force continuous innovation cycles to adapt
mineral processing to new conditions.
Another critical limitation is water availability. The intensive use of sea
water shows new challenges and opportunities. The requirement to increase
water reuse and recycling is now not only an environmental responsibility
but also a cost factor. Increasing water cycling produces an increase in yield
but impurities accumulate as a side effect. New process conditions should be
faced and the proper forecast is a key to sustain the operation standard.
Thirdly, we observe that the role of hydrometallurgists has changed in recent
years. Ones initial role preparing flow sheets and balances has been superseded
by new responsibilities. Safety management, environmental responsibility,
operations and human resources management, risk assessment, and finances
appear more often. The traditional academic curriculum for mining, metallurgy
and chemical professionals should be critically analyzed to include new time
requirements. Limitations are observed not only with regards to the availability
of qualified professionals but also in their initial toolkit of abilities.
Global markets have given us a very good opportunity to grow. However,
we are now facing the instability and uncertainty associated with the European
markets. The normal reaction of shareholders is to contain production costs
to be more competitive. It is our challenge to support the innovation cycles
despite limits in resources. Crisis periods are especially fertile for creativity
and innovation.

17

Hydroprocess 2013 is an opportunity to look over these and other important

challenges in the industry for the coming years. Here we will share our positive
results, critical opportunities and foster coming dreams. Certainly we are discussing the state of art in the aqueous processing industry, and we feel proud to be
part of it every year.
Many thanks and welcome to you all.
Gabriel Meruane Naranjo
chair
Hydroprocess 2013|5th International Seminar on
Process Hydrometallurgy

18

Preface

The realization of the 5th International Seminar on Process Hydrometallurgy

is the outcome of great efforts and special dedication from all those involved.
This seminar provided an excellent opportunity to work with many renowned
professionals and researchers from different countries who have selflessly
collaborated with us. Hydroprocess 2013, held from July 10-1 2, 2013 at the
Sheraton Hotel in Santiago, Chile, is effectively the fifth in a series of international seminars on Process Hydrometallurgy initiated in 2006 by Gecamin.
Undoubtedly, this seminar has become the site of an international forum
where all professionals and executives involved in Hydrometallurgy can analyze
and discuss innovations and developments concerning the liquid processes
of ores and materials involving the use of aqueous chemistry and physics. The
papers presented in these proceedings of the fifth version of Hydroprocess
speak of a consolidated seminar given the important assortment of technical
subjects covered and the number of enterprises and countries represented.
As was defined in the initial call, the main objectives of this seminar are
to (i) learn about innovations and developments in the hydrometallurgical
processing of metals including valuable metals, non-metal compounds and
industrial minerals, (ii) identify the best practices and technologies used in
hydrometallurgical plant operation and design, and (iii) fortify an international
network of collaboration and exchange between professionals related to
hydrometallurgy.
This book contains 45 abstracts written by delegates from 11 different
countries. The conference has been organized by area of interest, including:
Plenary presentations (2); Base metals hydrometallurgy (12); Cyanidation,
leaching and recovery of gold (6); Solvent extraction and ion exchange (7);
Electrometallurgical processes and electrochemistry (7); Bioleaching processes (5); Modeling and optimizing hydrometallurgical operations and
circuits (5); and Hydrometallurgical processes for producing salt and nonmetals compounds (1).
I would like to thank all those whose efforts have helped in making Hydroprocess a success. Thank you to the authors and their organizations for
submitting papers; to all the technical experts, for sharing their expertise,

19

dedicating valuable time correcting the articles, and for providing insightful
comments thus enhancing the quality of this publication and the standard of the
seminar. Finally, I would like to thank all the Gecamin personnel for their efforts
in ensuring all the goals proposed for the seminar were successfully achieved.
Fernando Valenzuela
Technical Coordinator a nd Editor
Hydroprocess 2013|5th International Seminar on
Process Hydrometallurgy

20

Acknowledgements

The Organizing Committee acknowledges with gratitude the efforts of all

the authors for contributing a large variety of high quality, detailed and innovative papers to the technical program. We also would like to thank the
reviewers, Montana Tech of the University of Montana, the employees from
Gecamin, and all those involved in the creation of these proceedings for their
assistance. The support of the Organizing, Advisory and Technical Committees
has been greatly appreciated, as has been the support of the Hydroprocess
2013 Chair, Co-Chairs and the Chairs of the technical sessions.
The Organizing Committee also wishes to thank the following sponsors
(as of June 7, 2013 in alphabetical order) for their generous support:
Gold: BASF, GEA Westfalia, Inppamet and SNF FloMin
Silver: BTA, Cytec, Metalex, Outotec, RSR Anodes
and Verne SpA
Social: SQM
Official Material: Arcadis and BASF
Institutional Partners: Consejo Minero, Chile; Instituto de Ingenieros de
Minas del Per; Servicio Nacional de Geologa y Minera (SERNAGEOMIN),
Chile; and Sociedad Nacional de Minera (SONAMI), Chile
Official Media: AreaMinera, Chile
Media Partners: Ecoamrica, Chile; Elsevier, United Kingdom; Infomine,
Canada; and Mining
Engineering, usa.

Finally, we would like to thank all the delegates who attended the seminar
and exchanged their valuable knowledge and expertise, thus contributing to
the great success of this 5th edition of the International Seminar on Process
Hydrometallurgy, Hydroprocess 2013. We are looking forward to seeing you
all again during the next version of Hydroprocess, in the year 2014.
Executive Organizing Committee
Hydroprocess 2013|5th International Seminar on
Process Hydrometallurgy

21

chap. 1
Plenary presentations

Impregnated activated carbon

for gold extraction from
thiosulfate solutions

Hydrometallurgical processing of gold is almost exclusively accomplished with

cyanidation. However, cyanidation has been attacked from environmental,
health and safety aspects due to cyanide toxicity, poor tailings management,

Courtney Young. Montana

Tech, USA
Nick Gow. FLSmidth, USA

and desires to eliminate open pit mining. Although these attacks are predominantly unwarranted, they have led to an increase in studies about cyanide
alternatives, particularly thiosulfate. Thiosulfate leaching of gold is similar to
cyanidation; however, gold recovery from thiosulfate solutions is not possible
with conventional carbon adsorption. This necessitates the use of more expensive resin adsorption/ion exchange processes to recover the gold from thiosulfate solution. In order to make gold recovery cheaper as well as cost-competitive
against cyanidation, a novel non-resin technology is described and characterized. In this technology, activated carbon is impregnated with cyano-cuprous
species which allows for high gold extraction followed by traditional elution.
Elution efficiency depends on how much copper is present with the gold on
the activated carbon surface. Optimal conditions for extracting and eluting
the gold were identified from computational models developed from statistically designed experiments. The impregnated activated carbon technology
has been patented because it makes thiosulfate leaching cost effective compared
to cyanidation by replacing resin adsorption/ion exchange; however, the technology needs to be tested on both a continuous and large scale.

25

INTRODUCTION
During cyanidation, gold (Au) is reacted with cyanide (CN-1) and oxygen (O2) causing its oxidation
and dissolution by forming cyano-aurous complex [Au(CN)2-1] and hydroxide (OH-1) (Marsden &
House, 2006). The process was patented by MacArthur (1916) and was conclusively shown to be
electrochemical in nature by Kudryk & Kellogg (1954) but was not combined with carbon
adsorption for recovering the Au from solution until 1970 (Marsden & House, 2006). If copper (Cu)
is present, it will leach via the same mechanism and form one of four cyano-cuprous species
[CuCN, Cu(CN)2-1, Cu(CN)3-2 and Cu(CN)4-3] depending on the pH (Adams, 1994). Because Cu is
usually present in larger amounts than Au, cyano-cuprous species result in higher concentrations
and compete more with cyano-aurous for the adsorption sites on the activated carbon (Jay, 2000).
Using cyanide toxicity and poor tailings management as excuses (Young, 2001), political groups
have attacked cyanidation to eliminate mining, a foundation of modern society. In response,
researchers in the mining industry have examined alternatives to cyanide leaching including, but
not limited to, chlorine, bromine, ammonia, nitric oxide, thiourea, thiocyanate and thiosulfate
(Young, 2001). Based on its leaching mechanism being similar to that of cyanide including the use
of a neutral pH, thiosulfate leaching is the leading contender even though various chemicals are
needed to stabilize the thiosulfate by preventing polythionate formation. In this regard, Cu has
been found to be the most effective and additionally shown to act as a catalyst (Marsden & House,
2006). Unfortunately, Au recovery from thiosulfate solution can only be accomplished by ion
exchange with resins, which are inherently expensive and suffer from fouling with polythionates.
However, Young et al. (2005) conceived an activated carbon technology that shows great promise
for making thiosulfate leaching cheaper and therefore cost-competitive against cyanidation.
It is well known that activated carbon, by itself, has no affinity for Au thiosulfate (Marsden &
House, 2006) and must be pretreated in order to recover Au from thiosulfate solution. Knowing
that Cu is needed as a catalyst as well as a thiosulfate preservative, Young et al. (2005) suggested
using Cu to impregnate the activated carbon with cyano-cuprous by chemisorption forming
Cu(CN)2-1 ads in order for the following metal exchange reaction to take place:
Au(S2O3)2-3 + Cu(CN)2-1 ads

Au(CN)2-1ads + Cu(S2O3)2-3

[1]

and thereby extract the Au from thiosulfate solution. In this paper, the development of this novel,
non-resin technology is reviewed from its proof of concept (Gow, 2006) at low adsorption density of
Cu(CN)2-1 to confirmation (Melashvili, 2009) at high adsorption density and subsequent elution.
Results have been presented in detail elsewhere (Young et al., 2012) but are reviewed here along
with, for the first time, a brief discussion on cost.

CYANO-CUPROUS CHEMISORPTION
Experiments of cyano-cuprous chemisorption on activated carbon were conducted at room
temperature (20C) and elevated temperature (40C) as well as variable pH (9-12). Adsorption
densities averaged near 15,000 g Cu/g C which equates to approximately 1,500 opt Au assuming
Cu and Au exchange in a 1:1 molar ratio according to Reaction 1. Because carbon adsorption

26

during cyanidation yields Au loadings near 300 opt (Marsden & House, 2006), this would more
than satisfy industrial needs. Furthermore, adsorption densities were found to increase with
increasing temperature and therefore to be endothermic yielding maxima of approximately 8,000
g Cu/g C at room temperature (20C) and 20,000 g Cu/g C at elevated temperature (40C) at pH 9.
However, these values were found to increase moderately by approximately 2,000 g Cu/g C at pH
12. The studies were conducted using caustic (NaOH) to adjust the pH and therefore in the absence
of lime [Ca(OH)2] which eliminates Ca+2 adsorption and therefore the formation of ion pairs
(MacDougall et al., 1980). This was done to keep the chemistry of the system as simple as possible
and thereby enable thermodynamic calculations to verify that cyano-cuprous chemisorbs.
After converting the adsorption densities to surface coverages, fitting the resulting values to
Langmuir Isotherms, and using the isotherms to determine reaction constants according to the
Stern-Langmuir Equation, free energies of adsorption (Gads) were determined and then used to
calculate enthalpies (Hads) and entropies (Sads) of adsorption as well using the ClausiusClapeyron Equation and the fundamental thermodynamic expression, respectively (Young 1994).
Results presented in Table 1 show that Gads at 20C and 40C average -27.4 and -32.6 kJ/mol,
respectively. Because Gads are negative and Hads are positive, adsorption is both favorable and
endothermic. Furthermore, because the free energies are significantly lower than -20 kJ/mol,
adsorption is confirmed to be chemisorption.
In order to optimize chemisorption densities, tests were conducted via full, two-level, factorialdesigned experiments using Stat-Ease software by systematically varying 4 factors between low
and high values, initial Cu concentration (0.001 or 0.1 M); pH (9 or 12); time (1 or 5 hours); and
temperature (20 or 40C), and using mid-point determinations (0.01 M, pH 10.5, 3 hours and 30C).
Resulting models were used to construct 3-dimensional plots to clearly illustrate the conditions that
yield maximum chemisorption. Results were found to be relatively independent of time and were
higher at increased temperature (40C) verifying its endothermic behavior. Taking cross-sections of
the plot at constant pH yields adsorption isotherms similar to those described earlier but as a
function of surface coverage as opposed to adsorption density. Clearly, the optimal conditions for
cyano-cuprous chemisorption on activated carbon occur at shorter time (1 hour), higher
temperature (40C), higher pH (pH 12), and higher cyano-cuprous concentration (0.1M).
Table 1 Thermodynamic Data for Cyano-Cuprous Chemisorption on Activated Carbon
Temperature (C)

Gads (kJ/mol)

20 @ pH 9

-26.8

40 @ pH 9

-31.4

20 @ pH 12

-28.1

40 @ pH 12

-33.8

Hads (kJ/mol)

Sads (J/mol/K)

41.02

223.8

55.65

276.8

GOLD-COPPER ION EXCHANGE

Impregnated carbon prepared under conditions where minimum chemisorption would occur in
anticipation that they would be favored by industry: 1 hour, 20C, 0.001M and pH 9. It was then
contacted with 20 ppm Au thiosulfate solution and characterized with Raman Spectroscopy to see if
the proposed reaction products could be observed (see Reaction 1). Raman spectra of impregnated

27

carbon before and after contact with Au cyanide are shown in Figure 2. Raman shifts near 2100 cm1 in Figure 2a are due to cyano-cuprous with the band at 2118 corresponding to dicyano-cuprous
[Cu(CN)2-1] and those at 2096 and 2127 cm-1 representing tricyano-cuprous [Cu(CN)3-2] (Young et al.,
2008). These peaks are slightly shifted from their aqueous counterparts at 2137, 2094 and 2108 cm-1,
respectively (Lukey et al., 1999). Likewise, Raman band at 2227 cm-1 in Figure 2b are due to cyanoaurous that formed at the activated carbon surface. It is slightly shifted from the 2239 cm-1 band for
solid KAu(CN)2 (Parker et al., 2008). It is understood that, as the pH increases, dicyano cuprous
will convert to tricyano cuprous with excess cyanide present (Marsden & House, 2006).
Consequently, it is reasonable to conclude that increased chemisorption at increased pH is due to
the additional tricyano cuprous available.

Surface Coverage ()

1.000

0.873

0.745

0.617

12.00
11.25

0.490

10.50
0.100

0.075

pH

9.75
0.051

0.026

0.001

9.00

Initial [Cu]

Figure 1 3-dimensional plot showing the effect of cyano-cuprous concentration and pH on surface coverage
() at 40C after 1 hour of chemisorption (Young et al., 2012)

28

a) Cu-CN
Bands Appear

b)
Au-CN Band
Appears

Cu-CN Bands
Disappear

2050

2100

2150

1500

2000

2500

3000

-1

Raman Shift, cm

Figure 2 Raman spectra of activated carbon showing (a) the appearance of cyano-cuprous bands following
impregnation and (b) their disappearance along with the appearance of the cyano-aurous band following Au
extraction (Young et al., 2012)

Because all of the cyano-cuprous bands disappeared following contact of the impregnated carbon
with Au thiosulfate solution, the Au must have ion exchanged with both of the cyano-cuprous
species, resulting in the formation of cyano-aurous species. Clearly, Reaction 1 is confirmed but the
results also suggest that the reaction mechanism should also include tricyano-cuprous:
Au(S2O3)2-3 + Cu(CN)3-2 ads Au(CN)2-1ads + CN- + Cu(S2O3)2-3

[2]

As with Reaction 1, it is assumed in Reaction 2 that the Cu and Au exchange in a 1:1 molar ratio. To
test this, impregnated carbon was prepared under various conditions to yield different cyanocuprous chemisorption densities and then contacted with Au thiosulfate solutions of varying
concentrations (5-50 ppm). Resulting Au concentrations were measured by ICP. Differences
between initial and final Au concentrations were used to calculate the amount extracted. Results
showed that, when the molar ratio of Cu on the impregnated carbon to Au in solution (Cu:Au) was
greater than approximately 1.5:1, Au extraction efficiencies of 100% were observed. Cu
concentrations were also measured but results were indeterminant yielding ratios ranging from 0:1
to 1:1, likely caused by precipitation when excess thiosulfate was not present to keep it solubilized.
Optimization tests of Au extraction via its ion exchange with Cu were also conducted with StatEase software using full two-level, factorial-design. In this case, three factors were systematically
varied and midpoints were examined: time (1 or 4 hrs), pH (10 or 12) and Au concentration (10 or
20 ppm). The temperature was fixed at 20C and the cyano-cuprous impregnation was established
at approximately 20,000 g Cu/g C so that Cu:Au ratios at the surface would not exceed either 2:1
or 1:1 assuming complete Au extraction. All variables were found to be important but Au extraction

29

at high Au concentration was slower. The 3-dimensional plot in Figure 3 shows the effect of time
and pH on Au extraction at the high Au concentration. It indicates that Au extraction is a maximum
at high pH and long times. Projecting the plot onto the base (i.e., pH-Time plane) yields contours of
which five are shown and three are labeled 90, 70 and 50%. For example, the conditions needed to
achieve >90% extraction range from pH 10.5 and 4 hours to pH 12 and 2.5 hours. Clearly, in this
case, the contours allow the conditions to be seen easier.

Au Extraction (%)

109

94.3

79.5

64.8

50

90%
4.0

11.5

50%

2.5

Time

12.0

70%

3.3

11.0

1.8

10.5
1.0

10.0

pH

Figure 3 3-dimensional plot showing the effect of time and pH on Au extraction by impregnated carbon from
Au thiosulfate solution at 20 ppm (Young et al., 2012)

GOLD ELUTION
Au-loaded, impregnated carbon produced in this manner was eluted using the AARL method
(Marsden & House, 2006) and thereby pretreated with 2% NaCN and 1% NaOH solution for an
hour, transferred to a water-jacketed column controlled at 97C, and eluted with distilled water.
Resulting eluant was collected as a function of time, reported in bed volumes (BV), and analyzed
for free cyanide by ISE and Au and Cu concentration by ICP. Example profiles and a Au recovery
curve are shown in Figure 4. Cu is being eluted somewhat selectively which is common in
cyanidation circuits (Marsden & House, 2006). It was not until 5 BVs have passed that the Au
concentration became significant. After approximately 15 BVs, Au concentrations reached a
maximum of 420 mg/L (ppm) which is comparable to cyanidation circuits as well. In this case, Au
recovery reaches a maximum of 83% at 70 BVs but other tests ranged from lows of 5% when Cu
was still present in significant amounts on the surface to 100% when relatively no Cu was on the
surface. Clearly the best results were obtained at low chemisorption densities of cyano-cuprous
such that, after Au extraction, relatively no Cu was left on the surface. However, if some Cu did
remain, it could be selectively eluted or separated by conventional smelting technology.

30

a)

b)

Figure 4 Elution (a) and recovery (b) of Au from impregnated activated carbon as a function of bed volume at
97C using distilled water, 2% NaCN and 1% NaOH according to the AARL method (Young et al., 2012)

FLOWSHEET DESIGN AND COST ANALYSIS

Figure 5 shows a conceptualized flowsheet for thiosulfate heap leaching as determined in this study
and compares it to cyanidation. As can be seen, the two processes are virtually identical. In both
cases, a leachant is added to the heap through which it trickles to leach the Au. The pregnant
solution is then extracted of its Au by adsorption onto activated carbon such that the barren
solution is recycled back to the heap and the loaded carbon is sent to stripping. At this point,
because the carbon contains an adsorbed cyano-aurous species, both processes include stripping
with conventional elution technology followed by electrowinning onto steel wool and eventually
smelting to produce Au dore. Stripped carbon will then be reactivated and recycled back to the
carbon columns, and raffinate from electrowinning will be recycled back to stripping. Steps for
carbon reactivation are not shown in these flowsheets.

31

a)

b)

Figure 5 Flowsheets for heap leaching and Au recovery by (a) cyanide and (b) thiosulfate using the novel,
non-resin, activated carbon process with cyano-cuprous impregnation (Young et al., 2012)

Clearly, the envisioned thiosulfate process uses the same unit operations as cyanidation. The cost
difference between the two heap leaching operations will therefore be predominantly due to the
impregnation step as well as the consumption of chemicals throughout. The impregnation step will
require conditioning tanks and pumps beyond that needed for cyanidation but capital and
operating costs for these items will be minimal. Although lime/caustic consumption will increase
due to the impregnation step, thiosulfate leaching (at pH 8) has less consumption than cyanidation
(at pH 10.5); hence, the overall consumption rates will be similar. Assuming the leachants (cyanide
vs thiosulfate) have the same costs, the only difference will be the cost for cyano-cuprous reagent
which is available for the plating industry at moderate costs near US$6/kg and in large quantities
on the order of MTPD. With chemisorption densities near 15,000 g Cu/g C, cyano-cuprous costs
will also be negligible. It is additionally noted that, because of the impregnation step, the Au dore
product could become contaminated with Cu. This could be prevented by selective elution, as
mentioned earlier, and/or by the proper management of the smelting step. If Cu is present in the ore
either process may be pursued and consequently these costs are not considered as well.

CONCLUSIONS
A novel carbon adsorption technology for extracting Au from thiosulfate solutions has been
developed and shows promise at becoming a cost effective method allowing thiosulfate leaching to
become economically competitive against cyanide leaching. Unless impregnated, activated carbons
will not adsorb Au thiosulfate. However, by first adsorbing cyano-cuprous species onto the carbon,
Au can be extracted from thiosulfate solution. Thermodynamic analysis of the adsorption data

32

verified that the adsorption process was chemisorption and endothermic. Factorially-designed
experiments verified that maximum cyano-cuprous adsorption occurred at high pH and
temperature and verified the tests. Similar tests conducted for Au extraction showed the Au
followed the Cu. Raman spectroscopy revealed that increased adsorption with increasing pH was
due to both di- and tri-cyano-cuprous and the resulting adsorbed Au species was likely cyanoaurous [Au(CN)2-1]. Flowsheets similar to current cyanidation processes were designed and could
be employed immediately. At question is whether Au and Cu separation can be done in the process
or in subsequent smelting and refining processes. Based on these promising results, a patent has
been granted; however, the technology needs to be tested on a continuous basis.

ACKNOWLEDGEMENTS
Many thanks are extended to Newmont Mining Corporation for their support of this project as well
as the patent granted in the U.S.A. and pending in Canada and Australia.

REFERENCES
Adams, M.D. (1994), Removal of Cyanide from Solution using Activated Carbon, Mineral Engineering,
7(9):1165-1177.
Gow, R.N. (2008), Pretreatment of Activated Carbon for Gold Adsorption from Thiosulfate Leach Liquors,
Thesis, Montana Tech, Butte, MT.
Jay, W.H. (2000), Copper Cyanidation Chemistry and the Application of Ion Exchange Resins and Solvent
Extractants in Copper-Gold Cyanide Recovery Systems, In: Proceedings of Alta 2000 Conference,
Adelaide, Australia.
Kudryk, V. and H.H. Kellogg, (1954), Mechanism and RateControlling Factors in the Dissolution of Gold in
Cyanide Solution, Trans. AIME J. of Metals, 541-548.
Lukey, G.C., van Deventer, J.S.J., Huntington, S.T., Chowdhury, R.L., Shallcross, D.C. (1999), Raman Study on
the Speciation of Copper Cyanide Complexes in Highly Saline Solutions, Hydrometallurgy, 53:233-244.
MacArthur, J.S. (1916), Discovery of Cyanidation, Mining & Scientific Press, London.
MacDougall, G.J., Hancock, R.D., Nicol, M.J., Wellington, O.L. and Copperthwaite, R.J. (1980), The Mechanism
of the Adsorption of Gold Cyanide on Activated Carbon, J. S. Afr. Inst. Min. & Metall., 80: 344-356.
Marsden, J. and I. House (2006), The Chemisty of Gold Extraction, Ellis Horwood Publishers, New York.
Melashvilli, M. (2009), Gold Recovery from Thiosulfate Solutions using Activated Carbon Pretreated with
Copper-Cyanide: Mechanism, Quantification and Elution, Thesis, Montana Tech, Butte, MT.
Parker, G.K., Gow, R.N. and Young, C.A. Twidwell, L.G. and Hope, G.A. (2008), Spectroelectrochemical
Investigation of the Reaction between Adsorbed Cuprous Cyanide and Gold Thiosulfate ions at
Activated Carbon Surfaces, In: Hydrometallurgy 2008: Proceedings of the 6th International Symposium
Honoring Robert S. Shoemaker, C.A. Young, P.R. Taylor, C.G. Anderson and Y. Choi (Editors), SME,
Littleton, CO.
Young, C.A. (1994), Characterization of Adsorbed Oleate at Calcite and Flourite Surfaces by Infrared and
Raman Spectroscopy, Dissertation, University of Utah, pp. 287.

33

Young, C.A. (2001), Cyanide: Just the Facts, in: Cyanide: Social, Industrial and Economic Aspects, C.A. Young, L.G.
Twidwell and C.G. Anderson (Editors), TMS, Warrendale, PA.
Young, C.A., Gow, R.N. and Melashvilli, M. (2011), Method for Aqueous Gold Thiosulfate Extraction Using
Copper Cyanide Pretreated Carbon Adsorption, U.S. Patent Publication Number US 20110259148 A1.
Young, C.A., Gow, R.N., Melashvilli, M. and LeVier, M. (2012), Impregnated Activated Carbon for Gold
Extraction from Thiosulfate Solutions, In: Separation Technologies for Minerals, Coal and Earth
Resources, Proceedings of the Roe-Hoan Yoon Symposium, C.A. Young and G.H. Luttrell (Editors),
SME, Littleton, CO.
Young, C.A., Gow, R.N., Parker, G. and Hope, G. (2008), Cu-Cyanide Adsorption on Activated Carbon, In:
Hydrometallurgy 2008: Proceedings of the 6th International Symposium Honoring Robert S. Shoemaker, C.A.
Young, P.R. Taylor, C.G. Anderson and Y. Choi (Editors), SME, Littleton, CO.
Young, C.A., Twidwell, L.G and Hope, G. (2005), Recovery of Gold from Thiosulfate Leach Liquor Using Activated
Carbon, CAST Proposal, Montana Tech, Butte, MT.

34

Geotechnical lessons learned

from the operation of Cerro
Verdes Crush Leach Pad 4-A

The leaching process started in Cerro Verde in the late 1970s. At that time, the
leach pad operating criteria centered around metallurgical and production
parameters only. Little or no consideration was given to slope stability, phreatic

Helbert Galdos, Javier Guevara and

Arnaldo Saavedra. Sociedad Minera
Cerro Verde S.A .A ., Peru

levels or permeability issues. Later on, when new leach pads were started up
in the early 1990s, unusual slope stability issues began to take place; however,
these did not pose risks to the operation because overall pad heights were
small compared to todays permanent pads. In 1996, the Crush Leach Pad 4-A
was engineered and constructed using the best technology available at the
time. The design considered twenty 4-meter lifts, which quickly became 6-meter
and then 8-meter lifts with improvements in ore agglomeration technology,
forced air injection and larger and taller stacking equipment. Recently, Pad
4-A has reached its maximum capacity; twenty lifts have been stacked at an
average of 5.5 meters per lift. The metallurgical optimization, coupled with
changes in ore quality, entailed geotechnical challenges that led to modified
variable irrigation rate schemes, ore stacking, lift rinsing practices, dewatering
and phreatic level controls. Lessons learned at Pad 4-A, described in this paper,
have been taken into consideration in the engineering and construction of
the new Crush Leach Pad 4-B. This paper also describes the geotechnical best
practices and design considerations implemented at this new leach pad.

35

INTRODUCTION
The main purpose of this paper is to describe the experience gained by Sociedad Minera Cerro
Verde S.A.A. (SMCV) in the geotechnical control of permanent leach pads of crushed and
agglomerated mineral. Proper control is critical to the slope stability of leach operations. The
phreatic levels inside the heaps must be controlled as they can seriously affect safety, the
environment, metallurgy and the operation. Nevertheless, they are not always given the attention
which they deserve.
On the other hand, the geotechnical control of leach heaps has been improving. It has become more
complex and elaborate as the leach heaps, platforms or pads, have grown in size. Furthermore,
governmental and internal mining company regulations have been changing as part of more
demanding risk management programs.
In general, it is clear that the geotechnical control of leach heaps has evolved in a simple, almost
always reactive manner, instead of in a more elaborate preventative way. Experience has taught
that it is much more convenient to incorporate geotechnical controls from the engineering stage
than when the leach platforms have already been started and have several layers of mineral being
treated.
The first part of this paper briefly describes the leach operation at SMCV, then the geotechnical and
operational problems encountered at the leach heap of crushed and agglomerated Pad 4-A will be
described, along with the corrective measures implemented. Next, the geotechnical controls
implemented at Pad 4-A as part of a preventative program will be discussed. Finally, the paper will
show the application of the experience obtained at Pad 4-A for the engineering, construction and
operation of the new leach platform, Pad 4-B, which began operations in December 2012. All of
which were intended to guarantee, from the initial design, the stability of the heap, and its
operational continuity to avoid having to implement much more expensive corrective measures
later.

The leach process at Sociedad Minera Cerro Verde S.A.A.

At Sociedad Minera Cerro Verde, secondary sulfide and transitional coppers are leached. The
leaching of low-grade (ROM) minerals is carried out at the leach heaps MegaPad ROM and Pad 1X.
On the other hand, the leaching of high grade mineral, previously crushed and agglomerated with
sulfuric acid, takes place at the heap Pad4-A and, more recently, at Pad 4-B.
Both processes involve multilayer permanent heaps of the valley-leach type, i.e. the mineral is
stacked up until it reaches the storage capacity of the pad. The heaps at Cerro Verde are unique in
that they are built in small ravines, that are filled until they reach the height of the platform and
then an upright pyramid is formed by continuing to stack the mineral until the maximum area of
the platform is reached. Thereafter, the pyramid is formed with ever decreasing stack area. The
total height of Pad 4-A is 110 meters with 20 layers of mineral and its slopes have global gradients
of 2.5H : 1V.

36

Operational and geotechnical problems of leach heaps

The most common operational, metallurgical and geotechnical problems found at the permanent
heaps are as follows:

Operational problems
The collapse of the leach solution collection pipes and/or their connections occurs due to lack of
quality control during installation, or while the protective geo-membrane material (overliner) is
being put in place, or by the uncontrolled movement of heavy equipment such as tractors or
conveyor belts. These collapses significantly reduce the drainage capacity of the system,
producing an increase in the phreatic level inside the pad, reducing the factor of safety and
increasing the risk of liquefaction of the saturated mineral and destabilization and collapse of
the slopes.
Leakages from the pipes near the slopes affect the geometry of the slopes.
Removing leached material from the bases of the slopes, for example, to open/maintain access
routes to the leach heaps, weakens the base of the slope.
Increasing the rate of irrigation without bearing in mind the limit of permeability of the material,
or the settling of the underlying material with each layer placed on top.

Geotechnical problems
Internal erosion and piping due to the concentration of fluid at the base of the slope. In some
cases, a cavern is developed which grows towards the interior of the heap.
Localized instability of a side (bank) due to an increased phreatic level near the slope.
Interruptions to vertical flow caused by the presence of layers of low permeability due to their
high clay contents, badly scarified interfaces (badly carried out drainage and ripping).
The presence of craters (sink holes) in the surface produced by damaged collection pipes
followed by internal erosion.
Surface channels caused by rain, flow from broken pipes, or the overflow of an area with
ponded solution resulting in superficial slipping on sides or banks.
Internal canalization of the solution through high-permeability zones (due to the presence of
highly permeable or segregated material) and lateral discharge over the perimeter.

Metallurgical problems
Excessive increase in clay content (montmorillonite and kaolinite) and high percentage of fines
(below mesh 100) in the agglomerated mineral stacked on the leach heap.
Low quality agglomerated mineral due to insufficient control during agglomeration and mineral
transportation to the leach platforms.
Pooling (ponding) of the irrigation solution due to the presence of mineral with high-clay and
and content and ore fines, which generate areas of very low permeability.

37

Solutions implemented
The operational, metallurgical and geotechnical solutions implemented were the following:

Operational solutions
The localizing of segments of collapsed main drainage collection pipes with the help of infrared
cameras mounted on small remote-control vehicles moving inside the pipes, or by following the
behavior of craters or sink holes on the irrigated surface.
Excavating until the collapsed main pipes are located, replacing the damaged sections and
replacing the removed material, increasing its permeability by mixing it with coarser material.
Blending of minerals in the mine to reduce the concentration of clays and fines being fed to the
crushing and agglomeration plant.
Inspection and periodic replacement of the main and auxiliary high-density polyethylene
(HDPE) pipes in the irrigation system, which show damage due to a reduction in their thickness
caused by handling or transportation, or excessive contraction and expansion due the daily
fluctuations in ambient temperature. The latter involves the use of ultrasound techniques to
evaluate the integrity of the fused unions of the HDPE pipes.

Geotechnical solutions
The installation in the bases of the slopes of banks of horizontal drains of 38 and 102 milllimeters
of exterior diameter, 100 and 150 meter long, respectively. In the right number and length, these
horizontal drains help to reduce the phreatic levels near the slopes of the leach heaps.
Reconstruction of slopes affected by the incorrect removal of material from the toe of the slope.
The latter can be aided by analysing the stability of the slope before and after the reshaping
work to ensure that the minimally acceptable geotechnical safety factors have been maintained
for static and post-seismic conditions.
The construction of small rock buttresses or retaining walls at the base of unstable slopes. This
method stabilizes the slope and canalizes the solution so that it can be drained, which avoids a
localized increase in ore moisture and saturation.
The installation of French drains to capture and evacuate the pools of solutions which can
accumulate on the surface near the base of the slope. These consist of ditches excavated in the
gravel, which are refilled with selected crushed stone (filter), which contains a perforated or
slotted pipe. The pipe must have a slope of at least 1% and operate at full flow. Before the end of
discharge, the slotted pipe will be replaced by a blind pipe and a low-permeable plug put in
place.
The installation of banks or batteries of vertical dewatering wells with exterior diameters of
between 204 and 356 millimeters of perforated PVC pipe to lower the phreatic levels inside the
leach heap. These wells have submergible pumps, pressure sensors (transducers), and low-level
switches to optimize the recovery of the solution retained inside the heap due the loss of
permeability of the mineral.

38

 Stabilization of the craters (sink holes) by refilling with gravel or highly permeable mineral with
particle sizes of from 102 to 204 millimeters; the repaired area is then covered with layers of nonwoven geotextil and geonet to stop the sink hole from affecting underlaying layers.
Repairing of channels in the slopes of the leach heaps, caused by rain or by the spillage of
solution, which result in erosion. The loose material is removed (as it has a high content of fines)
and is replaced by material of an appropriate size distribution, which is wetted and compacted
layer by layer to avoid further erosion.
Stability buttresses are constructed at the base of the principal slopes of the leach heaps, which
due to the high phreatic levels or deterioration in the angle of design of the slope, do not meet
the minimum factors of safety. The main purpose of a buttress is to increase vertical pressure
and to keep the base of the slope in place (only a localized effect). Likewise, as a side-effect, it
strengthens shear resistance when potential faulty zones pass through it. Buttresses are made
from leached material, rubble or waste material from the mine. The material is checked, wetted
and compacted in horizontal layers. Its weight is crucial. These buttresses contain filter zones or
drains of highly permeable gravel to lower or eliminate the phreatic level to ensure their long
term slope stability.

Metallurgical solutions
Optimization of the quality of the agglomerated mineral, ensuring the correct level of humidity,
dosage of sulphuric acid during the agglomeration, transportation and stacking stages of highgrade mineral. For the above, an instrument has been designed, which by measuring the
conductivity of the agglomerated mineral can optimize both the dosage of leach solution and
acid for curing of the mineral. It should be mentioned that an excessive dosage of acid
deteriorates the matrix of the rock of the mineral.
A widening (coarsening) of the size distribution curve of crushed mineral to reduce the level of
fines in the crushing circuit. This method is adopted after ensuring that an increase in ore size
would not reduce copper recovery any more than that caused by the presence of clays and fines
due to damage to or compaction (settling) of the mineral with every new layer stacked on the
permanent pad.

Implementation of geotechnical controls

The geotechnical controls of a leach heap are by nature almost always preventative, focused on
monitoring the phreatic levels, the surface and depth movements of key slopes, the vertical
pressure on the base of the slope and the settling of the mineral. Alternatively, the settling which
takes place between the layers can be monitored which would allow changes to the dry density and
porosity of the material to be controlled.
If the global hydraulic conductivity of the leached material is to be evaluated, a large-scale
pumping test should be carried out, which would obtain much more realistic results than localized
testing, an example is provided by the Lefranc tests which are made inside perforations. Another
alternative to measure hydraulic conductivity, but only locally, are the water addition (slug) tests
which take place inside an open-tube piezometer. These tests can also determine whether the
instrument can detect, over time, losses in permeability of the surrounding material due to
deterioration (decrepitation) or aging of the mineral.

39

If the leach heap contains gravel which is likely to suffer from liquefaction, seismographs should be
installed to register important seismic events and to correlate these to any resulting damage to the
heaps. A fixed seismograph should be installed on solid rock and a mobile one set up on the crest of
the heap to analyze signal amplification and the filtration of frequencies.
The geotechnical controls used at SMCV are as follows:

a) Periodical evaluations of slope stability

The periodical evaluations of stability analyze the stability of the heap slopes using the limit
equilibrium method. This analysis is bi-dimensional and as part of the hypothesis, the surface is
considered to act like a rigid solid. This method does not measure movements, only the resulting
safety factor. For a given surface, which is likely to suffer a fault, this method divides the block
into vertical slices, calculates and compares the resistance forces with the destabilization forces.
If the forces which are available to withstand movement are greater than the forces which
destabilize the slope (FS1), then the slope is considered to be stable.
FS = (Resistance Forces / Destabilizing Forces) 1
In general, the following conditions can be analyzed in heaps: static (short and long-term),
pseudo-static, post-seismic with reduced shear resistance, and post-seismic with liquefaction.
Given the large size of the slope, local, intermediate and global faults are investigated. The
potential fault surfaces are circular, blocks or random. Each type of analysis has its
corresponding minimum recommended factor of safety, established by the industry and by
technical literature. Normally, several control sections are established and for each one, the
factors of safety are calculated, which are compared with the required minimum values. Table 1
shows the required minimum safety factors for the leach platform Pad 4-A.
Table 1 Factors of Safety Leach Heap Slope Stability

Condition

Notes

Short-Term Static

1.3

Only for temporary slope cut work

Long-Term Static

1.5

For all important slopes

Pseudo-Static

1.0

Not applicable as the mineral of Pad 4A below the

phreatic level will undergo liquefaction in the event of
a high magnitude seism.

Post-seismic with
degraded resistance

1.3

Not applicable as the mineral of Pad 4A below the

phreatic level will undergo liquefaction in the event of
a high magnitude seism.

Post-seismic with

1.1

Normal value to control slope stability. This is the most

critical parameter to be complied.

Liquefaction

40

Safety Factor (FOS)

b) Maximum permissible phreatic levels

The maximum permissible phreatic levels are limits or threshold values beyond which it is not
safe to operate the leach heap. Below these thresholds, all the applicable safety factors are met
(static, pseudo-static and/or pseudo-seismic with liquefaction). These are established using slope
stability analysis (by the limit equilibrium method); they depend on the properties of shear
resistance of the terrain, the geometry of the slope and the phreatic level. Figure 1 shows an
example of the control of phreatic levels with respect to the maximum permissible values.

Figure 1 Comparison of Phreatic Levels with Maximum Permissible Values

c) Manual of operations
The Manual of Operations is a field document which attempts to standardize criteria, provide
instructions, procedures and recommendations to monitor and control the phreatic level inside
the leach heap by setting a series of maximum permissible operational levels. The manual
indicates the actions to be taken if these permissible limits are exceeded.
d) Prisms
These are optical instruments which facilitate the monitoring of the surface movements of the
slope in three directions; they are fixed to the terrain by a metallic tube which is embedded in a
concrete base. They can detect settling and/or bulging on the face of the slope, and are read
using electronic surveying equipment with accuracies of a few millimeters.

Geotechnical controls incorporated in the design of Pad 4B leach heap

Based on geotechnical experience acquired during the operation of the leach heap of high-grade
mineral, Pad 4-A; it was decided to include, from the engineering stage, the following elements of

41

geotechnical control in the design of the new high grade ore leach platform, Pad 4-B, which was
successfully started-up in December 2012.

a) Vibrating-wire settlement sensors

This type of sensor allows continuous monitoring of the vertical deformation of the foundations
of the leach heap, due to the vertical pressure of the stacked mineral and the phreatic level. This
sensor is especially useful when the terrain of the foundations has compressible layers or there
are thick refills. Furthermore, they allow indirect measurement of geomembrane stretching or
elongation. They are normally placed under the membrane (underliner layer), either on the mass
or on the structure of the refill. They are accurate to a few millimeters.

b) Inclinometers
Inclinometers are used to measure lateral deformation of the terrain of the foundations of the
heap which are caused by deep breaches (shear surfaces). They are installed outside the heap,
either on the containment berm or on the containment wall. An inclinometer consists of a plastic
tube with special grooves in two perpendicular directions which is installed and connected to
the terrain with grout or with slurry of cement and bentonite. Readings are taken with a biaxis
electronic probe (torpedo) which registers, over 50 centimeter sections, vertical deviations of the
tube in two perpendicular directions. They are accurate to a few millimeters.

c) Vibrating-Wire piezometers
Piezometers measure the pore pressure of the fluid contained in the void spaces of the leached
material (ripios) or in gravel when they are completely full of liquid, that is, saturated (100%
degree of saturation). They are based on the vibrating-wire principle and are very accurate.
Knowing the specific density of the fluid, it is possible to calculate the equivalent height of
column of liquid (piezometric height).

d) Vibrating-wire pressure cells

Vibrating-wire pressure cells measure the total vertical pressure increase at the base of the heap
over its useful lifetime. This pressure is that of the weight of the mineral and the solution.
Knowing the height of the heap, it is possible to estimate the average wet density if required,
which is important operational information. The pressure measurements and those taken from
the foundations allow the plotting of the settling curve of the pad against vertical pressure and
hence calculate the rigidity of the foundation terrain, which is important information for future
designs.

e) The main solution collection system

The system of collection of copper pregnant leach solution (PLS) at Pad 4-B has undergone the
following improvements:

42

A system to collect the main solution has been installed on a cut zone of bedrock material at
the bottom of the ravine to prevent differential settlement (uneven settling), which can
decouple the pipe joints. In other words, by installation of the main collection drains on

bedrock, and not on structural fill, the possibility of future differential settlement is
eliminated.

The principal collection pipes are interconnected which prevents the retention of solution
inside the heap due to obstruction of the collection system.

Together with the system of corrugated and perforated pipes, a system of solid perforated
HDPE collection pipes has been installed in order to ensure continued drainage if a collapse
occurs in the corrugated collection pipes.

Collection pipes have been installed inside trapezoidal ditches (canals) to generate an arc
effect in the mineral located above them and thus transfer the vertical loads of the heap
towards the sides and so reduce deformation.

f) Gravel screens
To prevent phreatic levels from exceeding those considered in the geotechnical design of the
leach, Pad 4-B has permeable gravel screens installed in the space formed by the union of toe of
the slopes of two adjacent cells. These gravel screens are installed in two sectors near the front
slope by the PLS weir upstream the PLS ponds. These gravel screens will also be installed in the
upper levels up to Lift 4, so that, they form a permeable front, perpendicular to the flow of leach
solution on its way to the PLS weir. These screens allow the removal of the leach solution from
the leach zones most adjacent to the front slope, avoiding solution build-up and abating high
phreatic levels affecting the factor of safety of the front slope. This area is the most critical area of
the pad from a geotechnical perspective.

Figure 2 Gravel Screen Pad 4B

g) Large diameter wells for the recovery of solution from the phreatic level
The purpose of these large size wells (nominal diameter of 2.50 meters) is to limit the phreatic
levels inside the heap from the beginning of the operation by providing the opportunity of
installing and operating pumps when the phreatic level starts to increase. The idea is to avoid
costly drilling for dewatering purposes. They have the advantage that they are built in stages
with each new lift being stacked on top of the previous. These well rest on the base of the heap,

43

taking advantage of the entire saturated height of mineral; their size ensures that they have the
required capacity. Solid, perforated pipes of 457 millimeters in diameter have been installed.
They have a gravel band of around 1m thick around them. Due to their large diameter, the pipes
remain upright until agglomerated mineral is stacked around them. Before being put into
operation, the wells should be cleaned and developed to remove fines from the filter and the
surrounding gravel to maximize solution collection and recovery pumping.

h) Wick drains
In the future, to eliminate serious permeability problems due to the presence of perched solution
(pockets of solution), wick drains will be installed; these are vertical drains made from a
corrugated plastic center wrapped in a nonwoven geotextile. They are used to manage complex
hydrogeological processes and to hydraulically connect the pad vertically. They allow the
drainage of perched phreatic levels, the crossing of areas of low hydraulic conductivity and the
reduction of pressure in artesian zones. A metallic lance (pole) is used to sink them into the
terrain. They can reach depths from 30 to 35 meters in leach ripios, with a maximum diameter
size of from 19 to 25 millimeters.

i) Horizontal internal drains

In the future, to eliminate serious permeability issues of the mineral and thanks to the presence
of bedrock banks located to the east and south-east of Pad 4-B, a design for the installation
(while stacking) of horizontal internal drains has been developed. It is much more practical to
proactively install drains than to try to perforate horizontal drains once the heap exhibits poor
permeability; these drains will be made from corrugated perforated pipe, 102 millimeters in
diameter, which will be connected to the PLS collection secondary drain pipes located on top of
the bedrock banks.

CONCLUSIONS
The process metallurgists and leach operators should be aware of the geotechnical concepts
involved in heap leaching practice in order to understand the mechanical, physical and hydraulic
phenomena, which occur inside heaps, their affect on process performance and the physical
stability of the slopes.
In light of the concepts described above, it is possible to avoid future slope stability problems, both
in dynamic and permanent multilayer leach pads by developing solutions that integrate
geotechnical, operational and metallurgical concepts to protect personnel from safety incidents,
meet environmental requirements and achieve production targets.

AKNOWLEDGEMENTS
This paper describes the development of solutions for the geotechnical and operational problems
which occurred at the permanent leach heaps of Cerro Verde. These ideas generally came from all
the members of the Hydrometallurgical Processes Area and the consultancy firms with which they
work. We would like to thank all of the people involved for their creativity and effort.

44

REFERENCES
Memorndum Tcnico: Preliminary Extraction Well Design, North Face, Pad 4A, SMCV, Arequipa, Peru, URS,
Febrero 2009
Evaluacin Geotcnica Integral del Pad 4A, Vector S.A.C., Octubre 2009
Leach Pad N 4A Buttress and Slope Regrading, URS, Noviembre 2009
Memorndum Tcnico: Final Cerro Verde Pad 4A - Liquefaction Analysis, Summary of Runout Estimate, URS,
Julio 2010
Memorndum Tcnico 002 - Anlisis de Estabilidad Adicional Seccin C-C Pad 4A, Ausenco vector,
Septiembre 2011.

45

chap. 2
Base metals hydrometallurgy

Daily process mineralogy:

A metallurgical tool for
optimized copper leaching

Although future copper ores will increasingly require concentrator operations,

heap and stockpile leaching (including bio-leaching of low grade primary

Wolfgang Baum, Kevin Ausburn and

Randy Zahn. FLSmidth, USA

sulfides) will continue to represent a significant production segment in copper

mining. Future copper leach operations will be faced with lower grades, harder,
finer grained and more acid-consuming ores, complex mineralogy and cost
increases related to water, power, reagents and steel wear. It is, therefore,
imperative that existing and new leach operations are well-designed and have
robust data bases and daily control of the ore feed mineralogy. In order to
achieve optimal and consistent crushing, best-practice agglomeration, good
permeability, efficient curing, and lowest acid consumption, a leach operation
will require quantitative routine mineralogy data. Several heap leach operations
in Arizona, Peru and Northern Chile have benefited from long-term mineralogical feed and leach residue analyzes. Typically, modern laboratory technology
such as xrd Rietveld, near infrared, optical microscopy and automated mineral
analyzers were used. The application of the characterization techniques for
daily blast hole analysis in two Arizona mines and select operations in Chile
and Peru has minimized ore routing errors, supported a better p80 and
throughput in crushing, reduced permeability failures, and optimized hydrometallurgical treatment. This paper provides application examples and recommendations for production mineralogy in conventional copper leaching,
bio-leaching and/or hybrid heap-stockpile concepts. The use of automated
small to large mineralogy laboratory modules for mine-site mineralogy will
be illustrated.

49

INTRODUCTION
Future mining of sulfide ores will be faced with lower grades, harder and finer grained rock
matrices, complex mineralogy, more reagent consumption, remote locations, severe staff shortages,
and cost escalations for power, water, reagents and (steel) wear. To exacerbate these challenges, the
cyclical exploration approach may not be able to provide sufficient re-fills in the reserve
pipelines. High throughput heap leach (and stockpile bio-leach) operations of the future will have
to use considerably better ore characterization and production control mineralogy in order to
achieve good copper extraction while maintaining low operating cost.
These better ore characterization approaches were pioneered in the 1980s and started to become
increasingly accepted in the new copper leach operations in Chile in the 1990s.
One case study example is the step-changing process mineralogy work at the El Indio gold-silvercopper operation (then owned by St. Joe Minerals Corporation) which resulted in substantial plant
metallurgy improvements and significant gold, silver and copper recovery increases (Baum et al
1989).
A break-through for process mineralogy for copper leach operations was accomplished with the
extensive use of mineralogical ore characterization by both the El Abra and Radomiro Tomic leach
operations. Specifically at Radomiro Tomic, the first robust, semi-quantitative copper and alteration
mineralogy was carried out on large sample numbers and probably contributed to good startup
and continuously high leach extraction (Cuadra C. and Rojas S. , 2001) and Baum 1998b cited in
Cuadra and Rojas).

Challenges in copper leaching

A ground rule for optimized copper leaching (J.Campbell, 2004, pers. comm.) states maximize
your first cycle leach recovery and control your mineralogy
The copper leach operations of the past had the advantage of relatively good ore grades combined
with oxide copper mineralogy. These ore types are becoming depleted and a variety of ores
including secondary sulfide ores, mixed oxide-sulfide ores, bornite-dominated and, finally, lowgrade chalcopyrite ores will make up the feeds of the future. Also, the future ores may be harder
and will have high(er) acid consumption depending on the host rock type and increase of skarns
and volcanic rocks. Finally, in order to compensate for the declining grades, leach operations
require higher throughput. All of the factors mentioned will place considerable demands on good
ore control which is synonymous with reliable ore characterization.
Operating competitive future heap and bio-stockpile leaching will not be economically attractive
without routine quantitative feed and leach residue mineralogy. Mineralogical analyses, today, are
capable of providing daily/weekly data on:
Mineralogical data
Overall gangue mineralogy
Pyrite content
Copper deportment
Clay content

50

Process parameters optimized by mineralogy

Geo-technical features
Blast indexing & crusher efficiency
Agglomeration
Curing requirements

Acid consumers
Salting potential

Permeability failures
Acid, Ferric, PLS Impurities, Cu extraction

Although diagnostic extraction tests (i.e. flash leach tests), copper assays, bottle roll, column and
mini heap tests remain baseline metallurgical tools, they cannot and will not provide the answers
for low recoveries, high acid consumption, poor agglomeration and low permeability (to name a
few). Mineralogical analyses are required to assess the cause(s) for poor metallurgy.
The following examples (Tables 1 and 2) illustrate the importance of mineralogy for identifying
leach problems:
Table 1 Variance in Acid Consumption Oxide Leach Ore Northern Chile
Parameter
Leach extraction, %
Acid consumption, lbs/t
Matrix fracturing, %
Iron hydroxides, %
Ca-Montmorillonite,%

Ore Type A
74.3
18.9
26.5
16.2
4.3

Ore Type B

58.9
29.6
41.2
25.4
15.8

Modified after Baum, Smith & Sepulveda 1996

Table 2 Diagnostic Quick Leach & Copper Deportment Errors

Total Cu = 0.89% - Cyanide Soluble Cu = 69%
Diagnostic Leach Conclusion: 77% Covellite/Chalcocite Mineralogy
Actual Mineralogy
Copper Mineral
% Distribution
Leach Type
Cu-Fe-hydroxide/Chrysocolla
Covellite/Chalcocite
Chrysocolla/Malachite
Chalcopyrite/Turquoise

36
20
29
15

Semi-Refractory
Ferric Acid Leachable
Acid Soluble
Refractory

Estimated Total Recovery

Mineralogy
Diagnostic Leach
Assays
Bottle Roll Tests
45 Day Column Tests

60%
77%
53%
65%
59%

Modified after Baum 1998a

METHODOLOGY
The currently available mineralogical tools consist of 5 major optimized techniques or innovations
which were introduced during the last 20 years:
1. Automated sample preparation for high throughput and fast turnaround

51

2.
3.
4.
5.

Polarized Light Microscopy

XRD (x-ray diffraction) Rietveld mineralogy (quantitative)
NIR or FT-NIR (near infrared or Fourier transform near infrared) analyses
Automated mineralogy (via QEMSCAN, MLA, TIMA, others*)

*QEMSCAN, QEMSCAN WellSite, MLA, MLA Express, ASPEX: FEI, Hillsboro, OR USA
TIMA: Tescan USA, Cranberry Township/PA USA
RoqSCAN: Fugro Robertson/Carl Zeiss, Houston, TX USA
EVO MA 15 & Particle SCAN VP: Carl Zeiss, Oberkochen, Germany

Further, the availability of larger, automated or central laboratories, permits mineralogical work on
daily blast hole samples and the use of these data for constant ore control, ore blending, ore routing
and, last but not least, for process control (Allen et al. 2007, Baum, 1996a & b, 1998a, 1999, 2007,
2009, 2013).

Automated sample preparation

The introduction and implementation of robotics technology for sample preparation at various
mining companies has been a step change in providing fast and high-throughput laboratory data
(Best et al, 2007). Specifically, for mineralogy, the availability of automated XRD-NIR laboratories
has enabled data feedback for mining and processing on a daily basis (Baum, 2009).
-

Both Freeport-McMo-Ran Copper & Gold laboratories, the Central Analytical Service
Center (Best et al 2007) and the AXN XRD-NIR Mineralogy Lab (Baum 2009) are good
examples of integrating the daily use of mineralogy for leach and concentrator operations.
Other milestone labs include the largest automated laboratory in mining, Anglos Easter
Bushveld Regional Laboratory (EBRL) and the BHP Newman Iron Ore Lab in Australia
(IMP 2012)). The FLS-built automated commercial Labtium Lab in Finland supports this
increasing trend to lab automation, even for smaller lab modules (FLSmidth 2009).

XRD Rietveld mineralogy

The introduction and optimization of Rietveld refinement methods has been the breakthrough for
quantitative XRD mineralogy. For leach operations, this method offers daily mineralogy profiling
on blast holes or any other samples. As such, key minerals which govern hardness and crushing
such as QKP (quartz-K-feldspar-plagioclase), changes of the ore types gangue, sulphides, acid
consuming minerals and minerals impacting geo-technical features can be consistently monitored.

Case study example 1

The Safford Mine (a copper oxide heap leach operation) in Arizona utilizes the XRD and NIR
mineralogy for better blasting and crusher optimization (Brandt et al., 2011).

On the basis of the XRD Mineralogy, i.e. using the QKP = Quartz+K-feldspar+Plagioclase
content as rock hardness indicators which are directly related to rock type and alteration.
The drill index number typically corresponds to geological characteristics of the rock and
has a very strong correlation with QKP.
With this information and other data, linear equations were developed relating to the
tertiary crusher equivalent throughput.

Although many of these applications are ore-specific, may not always exhibit clear relationships
and require continuous refinement, they offer considerably more continuous control and more

52

optimization than the previous conventional approaches which did not include quantitative
mineralogical data.
A key element in any use of new standard operating procedures and better practices in both ore
characterization and hydrometallurgy is stated well by Brandt et al (2011): It has also shown that for
this process to remain successful there must be a team with a champion to keep the process alive.

Case study example 2

In one of the Arizona copper leach operations, Allen et al (2007) illustrated the improvements of
using QEMSCAN and XRD ore profiling over merely visual ore classification using ternary mineral
diagrams.

The visual ore classification of the alteration coding resulted in considerable overlap of key
minerals and no clear discrimination of the ore types.
The use of QEMSCAN and XRD (which apply quantitative measurement techniques)
enabled the mine to reliably profile distinct ore types (which have distinct processing and
recovery features).
Allen et al (2007) indicate that the combined use of XRD and NIR lab technology for ore
control may assist in reducing acid consumption in one specific case by 3%.

Clay models and permeability

The combination of conventional analyses for swelling clay and the capability to generate spectral
models for clays and other hydrated alteration minerals via near infrared (NIR) or FourierTransform near infrared (FT-NIR) analyzers enables mine sites to perform continuous clay and acid
consumption analyses (Allen et al., 2007). These methods, in concert with XRD, permit a mine to
forecast the permeability features of an ore and avoid massive permeability failures (as they have
occurred in the past).

Case study example 3

Allen et al (2007) documented in an excellent study the impact of clay (and lack of good
mineralogical clay control) on permeability and copper leach losses. The assumption was that high
clay contents (normally >10-12% clay note by author) stacked daily with the heap leach ore will
result in an ongoing 10% loss in recovery in the heap. The cost figures are in 2007 US $.

For ores with 0.6% Total Cu content and production of 80,000 st/d
@ 10% leach recovery loss = 1.2 lbs Cu/st = 96,000 lbs Cu/day
Margin of $ 0.25/lb for lost copper
Total loss = $ 24,000 = $ 168,000/week = $ 8.7 M/Year

Case study example 4

Allen et al (2007) also showed a case were daily XRD or NIR mineralogy control could have
prevented serious problems with material handling and equipment shutdown.

A significant increase of muscovite (from 15% to 38%) in concert with elevated clay (10%)
resulted in high moisture additions to maintain good agglomeration. Ultimately, these
uncontrolled moisture additions plugged the conveyor chutes and forced a shutdown.

53

Acid consumption control

One of the most important (i.e. costly) but still neglected issues in copper leaching is the better
forecasting and control of short- and long-term acid consumption. Since acid consumption may
represent 10->25% of the operating cost, this should become a major optimization target. Todays
laboratory technology (XRD and NIR/FT-NIR) can assist companies in improving the acid
consumption issues by establishing acid consumption models. In addition to copper minerals, a
large range of gangue minerals (carbonates, feldspars, biotite, chlorite, clays, talc, Fe-hydroxides,
calc-silicates, tourmaline, serpentine, etc.) consume acid.

Case study example 5

Baum, Smith, & Sepulveda M. (1996) showed that, in a specific case, the acid consumption is
directly related to a combination of copper mineralogy, rock fracturing (=texture), presence of
hydrous iron oxides and swelling clay.

They documented that both the copper extraction differences of 74.3% vs. 58.9% and the
acid consumption differences of 18.9 lb/t and 29.6 lb/t were results of the above
mineralogical parameters.
With a better or control, ore routing and adjustment of crushing, curing, agglomeration and
leaching, the plant could have achieved higher copper extraction.

Case study example 6

The same authors (Baum, Smith & Sepulveda M. 1996) also showed in a case study that
discrepancies between so-called flash leach tests and column/heap extractions to be distinctly
related to a combination of textural copper occurrence.

Ore zones with higher amounts of coarse pyrite (>100 micron) caused lower column and
heap extractions due to the poor accessibility under conventional heap p80 crushing.

Automated mineral analyses (QEMSCAN, TIMA, MLA, Others)

The major innovation breakthrough for mineralogy was the introduction and optimization of the
automated mineral analyzers. Several systems are on the market. However, the authors have
specific and extensive experience with QEMSCAN and TIMA use, and the following observations
and analyses are primarily based on these systems.
The automated mineral analyzers are automated SEMs with special software and hardware to
perform quantitative mineralogy to concentrations of <1%, particle size analysis, liberation-locking
analysis and size-by-size mineralogy on ore and process samples including material balances to
validate the mineralogy. A milestone project was the large-scale heap survey (Fennel, Guevara et
al., 2005) which represents the very first systematic, quantitative mineralogical characterization of a
copper heap leach operation. This work has been refined and optimized in various mining
companies. Today, automated mineral analyses can forecast column leach extractions on the basis
of quantitative gangue/copper mineralogy, particle size, and textural features.

Case study example 7

Fennel, Guevara et al (2005) have conducted (worldwide) the 1st QEMSCAN heap leach survey at
the Cerro Verde operation. The purpose of this survey was to demonstrate the utility of

54

QEMSCAN quantitative mineralogy for monitoring ore response through leach cycles and assist in
refining the leach kinetics for better metallurgical processing.
In addition, they evaluated the impact of mineralogy on leach recovery.

They approached the objective with a very robust sampling program.

The mineralogical data and assay reconciliation with QEMSCAN-calculated recoveries
were excellent and confirmed the importance of the extensive sampling. It also confirmed
the sampling emphasis advocated by Lotter et al 2010.
The overall mineralogical data obtained assisted them in delineating the key opportunities
for additional copper recovery which were mainly related to the textures and mineral
associations (intergrowth) of the unleached chalcocite and covellite.
They also concluded that mineralogy has a direct effect on the leaching process, as
mineralogy changes through the leach cycle have been profiled.
Follow-up work can then be targeted on optimizing metallurgical practices as they related
to the unleached copper.

CHANGES IN BIO-LEACHING
It is recommended that, specifically the copper and gold industry, consider different heap and/or
stockpile leach approaches (J. Campbell, 2009, pers. comm.) which (a) include the use of HPGR as a
secondary or tertiary crusher and (b) apply innovative leach concepts such as deslimed heaps
which was recommended some time ago (Herkenhoff & Dean, 1987, Baum, 1998).
It is viable to conclude that within 2 years, the HPGR use for heap and agitation leaching will make
a measurable breakthrough. Particularly in concert with designed and instrumentalized heaps and
stockpiles, very large future HPGR systems have the potential to facilitate the leaching of very low
grade chalcopyrite ores. This could dramatically improve the extraction of copper (or gold) from
the coarse material fractions via better bacterial and solution access as well as significantly
optimized leach kinetics (Thompson, Baum & Ausburn, 2012).
Hopefully, larger scale pilot testing of hybrid systems including the use of HPGR and desliming
(Baum & Ausburn 2011) will be tested in the near future. Leaching of increasingly important
bornite ores (or mixed chalcopyrite-bornite) is accompanied by challenges of substantial jarosite
salting. Additional development is required for the high-extraction heap leaching of mixed bornite
ores.
Mineralogy with its process-impacting gangue and alteration minerals, acid consumers,
bactericides, textural parameters including liberation/locking or solution access and the myriad of >
20 mineralogy-related leach parameters represents a big impact and/or govern any bacterial
metallurgy. Bio-leach plants, like any other process operation, will not run efficiently without
continuous material characterization chemical/microbiological analyses alone will not suffice.
The inclusion of frequent ore characterization, geo-metallurgical databases and continuous use of
process mineralogy to rapidly trouble shoot and improve their heap, stockpile or agitation leach
systems, will be of substantial importance for plant optimization.

55

CONCLUSIONS
State-of-the-art mineralogical laboratory technologies enable todays and future copper leach
operations to (a) conduct robust forecasts, (b) minimize ore type variance, control and minimize
acid consumption, (c) achieve best practice agglomeration and permeability, (d) minimize geotechnical failures and (e) achieve maximum copper extractions. A considerable amount of the
mineralogical analyses can be performed on-site via small, efficient, automated lab modules. These
on-site labs can work with consulting service labs in order to achieve high quality and/or perform
routine external audits.
The expanded implementation of bio-leaching in future mining operations may benefit from a
combined approach using well-engineered and instrumentalized heaps, stockpiles and hybrid
stockpile operations which may include innovative comminution such as HPGR systems and a
revised approach to leaching using deslimed heap strategies.
In the age of larger capital layouts, higher throughput and de-bottlenecking or process optimization
of existing plants, the use of routine ore characterization and process-accompanying mineralogy
will be a key to a more competitive leach operation.

REFERENCES
Allen, D., Baber, B., Eady, S. & Baum W. (2007) Ore characterization, alteration coding, and lant feed control in
copper operations as directed by routine semi-automated mineralogical analyses, Copper 2007,
Toronto/Canada, Proceedings, Vol. IV, Book 2, pp. 247-258.
Baum, W. (2013) Innovations in process mineralogy and laboratory automation, Innovations in Minerals Processing,
SME-MPD Symposium, 6p., proceedings in print.
Baum, W., & Ausburn, K. (2011) HPGR comminution for optimization of copper leaching, Minerals & Metallurgical
Processing, Vol. 28, No. 2, May 2011, 77-81.
Baum, W. (2009) The new automated XRD-NIR mineralogy lab high capacity tool for short and long-range ore
characterization, SME-2009 Arizona Conference, presentation & abstract.
Baum, W. (2007) Material characterization on the cutting edge. CSIRO Process, June 2007, p. 12.
Baum, W. (1999) The use of a mineralogical data base for production forecasting and troubleshooting in copper leach
operations. Copper 99, Proceedings, Vol. IV, The Minerals, Metals & Materials Society, 1999, pp. 303408.
Baum, W.(1998a) Copper mineralogy and extraction problems. How to reduce your losses, Copper Hydromet
Roundtable, RANDOL, Proceedings, 1998, 41-54.
Baum, W. (1998b) Mineralogical analysis of 97 copper ore samples from Codelco Radomiro Tomic. PMET report 5 M
129, 15 p., unpublished Codelco report.
Baum, W. (1996a) Optimizing copper leaching/SXEW operations with mineralogical data, SME Annual Meeting,
Phoenix, Arizona, 1996, Pre-Print No. 96-84, pp. 1 7
Baum, W., Smith, E.H., and Sepulveda M., C. 1996b) Uso de informaccion mineralogical para optimizer las
operaciones. Mineria Chilena, 16, No. 185, pp. 85-91.

56

Baum, W., Sanhueza, O., Smith, E.H., and Tufar, W. (1989) The use of process mineralogy for plant optimization at
the El Indio gold-silver-copper operation (Chile). Erzmetall 42, Nr. 9, pp. 373-378.
Best, E., Baum, W., Gilbert, R., Hohenstein, B., & Balt, A. (2007) The Phelps Dodge Central Analytical Service
Center: step change technology implementing robotics systems and lab automation in the copper industry,
proceedings, Copper 2007, Toronto, Vol. VII, pp. 119-126.
Brandt, D., Martinez, I., French, T., Slattery, T., & Baker, G. (2011) Better blast fragmentation through teamwork at
the Dos Pobres Mine in Safford, AZ, Mining Engineering, October 2011, pp. 30-35.
Cuadra C., P. and Rojas S., G. (2001) Oxide mineralization at the Radomiro Tomic porphyry copper deposit, Norther
Chile. Economic Geology, Vol. 96, 2001, pp.387-400.
Fennel, M., Guevara, J., Canchaya, S., Velarde, G., Baum, W., & Gottlieb, P. (2005). QEMSCAN Mineral Analysis
for ore characterization and plant support at Cerro Verde, 27th Mining Convention, Arequipa, Peru,
Technical Papers, pp. 1-11.
FLSmidth (2009) Labtium Lab: http://www.flsmidth.com/enus/Services+and+Capabilities/Case+Stories/Increasing+capacity/Labtium+RoboLab
Herkenhoff, E.C, & Dean, J.G. (1987) Heap leaching: Agglomerate or deslime?, Eng. Mining Journ., 1987, June, 3239.
IMP 2012: http://www.impautomation.com/Automation-Projects/Platinum/Polokwane-Smelter-Complex-Laboratory-inPolokwane-(EBRL).html
Lotter, N.O. (2010b). Stratified sampling of drill core. Paper No. 11, Canadian Mineral Processors, Ottawa, p. 163-179.
Thompson, P., Baum, W., & Ausburn, K. (2012) Biohydrometallurgy emerging partner for processing low-grade and
complex sulfide ores and expanding reserves, MEI Conference on Biohydrometallurgy 12. Falmouth/UK,
abstracts.

57

Direct production of high-purity

cobalt sulfate heptahydrate from
a NICO deposit concentrate

Direct production of cobalt sulfate heptahydrate from a cobalt concentrate

originating from Fortune Minerals nico deposit was investigated at a pilot
scale. The process flowsheet tested involved pressure oxidation at 180C, 60
minutes residence time, arsenic removal, purification, cobalt solvent extraction, manganese removal, evaporation and crystallization. The final heptahydrate product averaged 20.75% mass fraction Co, and in g/t, 2.5 Ni, <10 Fe, <13
Mn, <2 Zn and 35Mg. Crystallization discharge predicted the maximum

Alex Mezei, Michael Johnson and

Cornelia Lupu. SGS Mineral
Services, Canada
Ron Molnar. METNETH2O Inc.,
Canada
Keith Lee and Robin Goad. Fortune
Minerals Limited, Canada

vacuum filtration throughput at 0.9bar and 18mm cake thickness of 1197kg/

m2h as well as a discharge cake residual moisture mass fraction of 18.2%.
Cobalt was recovered from the mother liquor bleed stream by a scavenging
carbonate precipitation. This produced a pure cobalt carbonate by-product
assaying 48.5% Co, and in 126g/t Ni, 124g/t Fe, 61g/t Mn, 7g/t Zn and 59g/t
Mg. The final discharge stream was subjected to scavenging precipitation
aiming to remove any residual metals. Based on the dynamic settling response
of the discharge scavenging precipitate stream at ~100g/t anionic flocculant
dosage and 1.2 hours solids residence time, the predicted specific underflow
and hydraulic thickener unit areas were 1.6 and 0.7 m2/tday, respectively. The
thickened underflow displayed a light Bingham plastic rheological behavior
with a critical solids density mass fraction of ~5.9%, showing an unsheared
yield stress of5.5 Pa. A comprehensive process model was established and
validated by the test data. The model output provided the complete process
and equipment design package for the commercial plant.

59

INTRODUCTION
The NICO cobalt-gold-bismuth-copper deposit is located in the Mazenod Lake district near the
Snare hydroelectric complex, approximately 160 km northwest of Yellowknife in Canadas
Northwest Territories. The mine development is planned based on a reserve of 33.1 million tonnes
of ore averaging 0.11% Co, 0.14% Bi, 0.04% Cu and 1.03 g/t Au that will be mined using a
combination of open pit and underground methods within a lower grade overall resource. These
reserves support a minimum 20-year mine life at a production rate of 4,000 tonnes per day. The
economic metals occur throughout the deposit in concordant and discordant sulphide-rich fractures
and disseminations. Cobalt is contained predominantly within a solid solution series between
arsenopyrite and cobaltite, with minor cobaltian loellingite. Bismuth occurs as native bismuth,
bismuthinite, and tellurides, while copper is hosted in chalcopyrite. Gold occurs as microscopic
grains of native gold and as various gold-bismuth-antimony-tellurium alloys.
Cobalt and bismuth are concentrated by bulk flotation followed by selective flotation. Most of the
gold reports to the resulting cobalt and bismuth selective concentrates. Bismuth is recovered from
its concentrate by oxidative hydrochloric acid leaching followed by purification and
electrowinning. The cobalt concentrate is subjected to pressure oxidation which extracts the cobalt
while rejecting most of the arsenic and iron. Residual iron and arsenic are further removed by
normal pressure oxidation. According to the original flowsheet, the iron-arsenic-free stream is
further purified by selective copper precipitation and ion-exchange and directed to electrowinning,
as described in previous papers (Mezei et. al., 2000, 2008). This paper reports the results of the
development work aimed towards establishing the process for the direct production of high purity
cobalt sulphate heptahydrate (CoSO47H2O) along with a cobalt carbonate (CoCO3) by-product.
Direct production means that the process does not use cobalt cathode as feed. Instead, the preferred
purification route was solvent extraction (SX) of cobalt in conjunction with advanced manganese
removal. The process development consisted of conceptual bench-scale testwork validated by an
integrated pilot exercise proving its technical feasibility and generating engineering design criteria
for the NICO processing plant which will be built in Saskatoon, Saskatchewan, Canada. Samples of
the cobalt sulphate heptahydrate produced were sent to potential off-take partners for evaluation.

METHODOLOGY
The testwork was structured around extensive bench-scale SX along with base-line manganese
removal and evaporation-crystallisation testwork. The results were used to establish a real data
process model linking the SX circuit producing the cobalt strip solution to the downstream
manganese removal (MnRe), evaporation (Ev) and crystallisation (Cry) circuits. An integrated pilot
plant was designed using the process model, then built and operated continuously for 120 hours.
After reviewing the results (while updating the process model), the MnRe circuit was operated for
an additional 24 hours, this time being fed with SX feed (not strip) as the goal was to prove that
manganese could also be removed prior to the SX. SGS proprietary standard safety and operating
procedures were applied throughout the bench and pilot scale metallurgical testwork. This paper
describes the results of the pilot testwork.

60

Solvent extraction
The solvent extraction (SX) reagent consisted of Cyanex 272 (Cytec Industries) in Orfom SX80
diluent (Chevron Phillips). Cyanex 272 is bis (2,2,4-trimethylpentyl) phosphinic acid. Orfom
SX80 is a high flash petroleum distillate containing about 80% paraffinic hydrocarbons with the
balance being alkyl aromatics. Organic mixtures from 10 to 15% by volume extractant were tested
in bench-scale experiments and 10% was selected for the piloting.
The solvent extraction pilot circuit was comprised four extraction stages (E1 to E4), two scrub stages
(B1, B2), three cobalt strip stages (CS1 to CS3) and two final acid strip stages (ZS1 and ZS2),
nominally to strip any co-extracted zinc as well as any other elements that were loaded but not
stripped with cobalt. The organic flow was from E1E4B1B2CS1CS3ZS1ZS2 and back
to E1. All the mixer-settler units were of a conventional design with square mixer boxes and
rectangular settlers. Each mixer unit was equipped with a variable speed mixer driving a pump-mix
impeller. The impellers were of a design having a flat disc top shroud over swept-back blades. This
reduced the shear minimizing the formation of micro-droplets and emulsion. All interstage
pumping was by the pump-mix impellers. The circuit was fed continuously at about ~310 mL/min
aqueous feed. Extraction operating temperature 55C and pH was controlled using 150 g/L Na2CO3
solution. Loaded organic was scrubbed at ~54C with a bleed of cobalt strip liquor diluted with
water at a pH of ~4.0-4.5, attained without pH modifying reagents. Cobalt strip operated at
pH ~2.6-2.8, 40C and zinc strip at pH below 1.0 and ambient temperature as there was after the
scrub stages. Sulphuric acid solutions were used to control pH in both strips. Advance volumetric
phase ratios (organic/aqueous) averaged 0.7, 40 and 17 in extraction, scrub and Co strip. No
aqueous was advanced in Zn strip because of the low Zn tenor and relatively short campaign.
Minor phases recycles were pumped in scrub and strip to achieve phase ratios of ~1 in the mixers.

Manganese removal, evaporation and crystallisation

The manganese removal (MnRe) pilot circuit consisted of two baffled agitated tanks equipped with
pH and temperature control and ORP (redox potential) Total residence time across the circuit was
12 h at the 16 mL/min target feed rate. A dual gas sparging system delivered the oxidizing mixture
of 2.5 % by volume SO2 in O2. The pH was maintained at the 3.6 target using 300 g/L NaOH
solution fed through a solenoid valve actuated control loop. The circuit discharge was flocculated
using ~330 g/t equivalent dry anionic flocculant. The settled slurry was filtered and sampled. The
filtrate was analysed and then advanced to the evaporation circuit. The precipitate containing the
rejected manganese was washed, dried and analysed. The wash water was collected and recycled to
the SX circuit to be used for the fresh strip solution make-up.
The manganese removal, evaporation and crystallisation pilot circuits operated simultaneously. The
evaporation circuit consisted of two tanks with a 2.26 h total residence time across the circuit at the
14 mL/min target feedrate. The evaporators were maintained at the target temperature of ~101C
using thermocouple-controlled heating mantles. They operated under 0.474 bar vacuum. The
flowrates were controlled using calibrated metering pumps. Condensate was trapped and
withdrawn at the prescribed rate with any excess recirculated. The evaporation circuit discharge
was pumped directly into the crystallisation circuit consisting of three tanks. Total residence time
was 3 h at the 8.5 mL/min target feedrate. Each crystalliser stage was kept at a relatively constant
equilibrium temperature by indirect cooling using an open-loop thermocouple-controlled water-

61

cooling system. The discharge slurry from the third crystallisation stage was vacuum filtered.
Filtrate and crystals were collected, weighed, sampled and analysed.

Measurement, sampling and analysis

All operating parameter data were collected manually based on pre-set sampling and measurement
timetables. In addition, parameters such as pH, ORP, temperature and flowrates were read by the
in-house data acquisition system. Samples were taken on a regularly scheduled basis to provide
control and full profiles suites. Aqueous and organic samples were prepared and analysed for
cation content by the SGS analytical laboratories either by Atomic Absorption Spectroscopy (AAS)
or by Inductively Coupled Plasma Optical Emission Spectrophotometry (ICP-OES), as appropriate.
Organic samples were also analysed by ICP-OES after dilution with an organic solvent.

Targets definition and real model updates

The pre-pilot update of the real data model was based on bench-test results. It included the
process and equipment design criteria needed for the pilot plant design, ranging from feed
characterization through operating targets definition for all unit operations. The final update
reflected the validation through the pilot results and included rheometallurgical data needed for
scale-up as well as equipment design criteria for the physical separation and materials handling
unit operations. The final output included all the necessary data required for the full feasibility and
engineering design studies.

RESULTS AND DISCUSSION

Feed preparation
The pilot plant feed sample consisted of about 2200 L bulk POX (pressure oxidation at 180C, 60
minutes residence time,) discharge PLS (pregnant leach solution) and wash solution mixture
produced during a prior pilot campaign. Iron and arsenic were removed from the PLS according to
the process parameters previously established, namely 50C, 6.2 h residence time, aeration and a pH
of 3.6, after which copper was taken out at pH 6.9, 50C, 1 h residence time. Lime slurry was the
reagent used throughout the entire purification sequence. The resulting copper-free solution
became the SX extraction feed for the pilot testwork. In TABLE 1, important analytical results on
this stream are summarized as averages of the results for samples taken during pilot plant.
Table 1 Cobalt solvent extraction feed for pilot plant
Co
2275

Mg
976

Mn
66.7

Ni
55.1

Na
675

Ca
495

K
29

Cu
<0.1

Cl
245

Zn
<1

Solvent extraction results

Bench SX data (Figure 1A) predicted good cobalt extraction at and above pH 4.5, with some
possible some Ca and Mg co-extractions, to be minimized by loading the organic well with cobalt,
and by scrubbing the loaded organic at pH 4 to 4.4. Manganese was loaded at a pH just below that
for cobalt, thus precluding its displacement. The isotherm shown in Figure 1B reflects the true

62

capacity of the organic phase in terms of all the extractable elements. The McCabe Thiele diagram
predicted that four stages would yield good extraction and that organic loading capacity would
approach 3.5. g/L Co.

4000

100
90

3500
Co Equivalent in Organic, mg/L

80

Percent Extracted

70
60

Co

50

Ca

Mg

Mn

40
30

Co+Zn+Mn

3000
2500
2000
1500

A/O = 1
1000

20

Ni

10

500
0

[A]

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Equilibrium pH

500

1000
1500
2000
Co Equivalent in Aqueous, mg/L

[B]

2500

3000

Figure 1 [A] Extraction vs. pH; [B] Cobalt-(equivalent) extraction isotherm, pH 4.25

76
5
6|4
5
4
6|3
5
4
6|3
5
4
6|3
5
4
6|3
5
4
3

E1
E2

E3

E4

B1

B2
0

[A]

Mixer pH

Mixer pH

The pilot plant solvent extraction circuit operated continuously for 106 h, a relatively short pilot
campaign, with stable performance (Figures 2 and 3). Organic degradation, stable emulsions and
crud formation were not observed. Excellent (i.e. rapid and clean) phase separation occurred in all
stages. The results demonstrated that cobalt could be recovered with high efficiency with overall
cobalt recovery exceeding 99.9% in four extraction and three stripping stages. The major impurities
in the strip liquor were manganese and magnesium with lesser amounts of calcium and nickel.
Data suggested that the magnesium could be further reduced by maximizing cobalt loading,
adding a third scrub stage and lowering the pH profile in the scrub stages. Nickel and calcium
could also be lowered by more efficient scrubbing.

10

20

30

40

50

60

Elapsed Time, h

70

80

90

100

6
5
4
3
2
5|1
4
3
2
1
5|0
4
3
2
1
5/0
4
3
2
1
5/0
4
3
2
1
0

CS1

CS2

CS3

ZS1

ZS2
0

110

[B]

10

20

30

40

50

60

70

80

90

100

110

Elapsed Time, h

Figure 2 Relevant pH profiles: [A] extraction (E), scrub (B); [B] strip (CS-Co, ZS-Zn)

63

Downstream circuit results

The performance of the manganese removal circuit downstream of SX was robust throughout the
entire 126 continuous operating hours under the bench-work defined targets. Relevant cobalt and
manganese tenor profiles are shown in Figure 3A. Steady state manganese removal efficiency from
the cobalt strip solution during the 12-hours residence time was 96.6% with the cobalt loss at 6.3%.
The manganese removal rate averaged 195, 33 and 227 mg/Lh in tanks 1, 2 and overall,
respectively. Most of the manganese (82.7%) was removed in tank 1, with 13.9% removed in tank 2.
The manganese removal discharge dry cake assayed on average 13% wt. Mn, 45% wt. Co, less than
500 g/t Na, and less than 100 g/t of other elements.
The evaporation and crystallisation circuits ran well in tandem throughout the entire operation.
The crystallisation mass balance indicated that about 37% of the cobalt was recovered in the
heptahydrate form, 53% reported to the mother liquor and ~10% to the wash solution. These two
liquid streams were recycled to solvent extraction upstream so that their cobalt content was
recovered. The heptahydrate product entrapped about 20% of the manganese, 7% of the
magnesium and 1% of the sodium. The cobalt tenor of the stage-3 crystallisation discharge mirrored
quite closely that of the circuit feed suggesting a consistent cobalt delta or bite of 24 g/L. This
enabled the production of crystalline heptahydrate by cooling (Figure 4B). The Co and Na content
profiles (Figure 5) of the crystallizer feed indicated that the sodium brought into the system
increased continuously, building up to levels that could not be controlled by crystallisation despite
its high selectivity, i.e. 99% rejection on average throughout the campaign. The cobalt sulphate
heptahydrate product averaged 20.75% wt. Co and had less than 2.5 g/t Ni, less than 10 g/t Fe, less
than 13 g/t Mn, less than 2 g/t Zn, 35 g/t Mg, 99 g/t Ca and 1500 g/t Na. The results confirmed that
the flowsheet tested allowed for the direct-production of a high-purity cobalt sulphate
heptahydrate product with the caveat that the sodium build-up in the system could lead to product
purity issues. Therefore, the flowsheet was changed by relocating the manganese removal circuit
ahead of the solvent extraction.
`

3
4

Co, mg/L

100

4
3
2
1

10

4
3
2
1

4
3
2
1
4
3
2
1

E4R
2
E3R
3
E2R
1
E1R (Raffinate)
E4O (Loaded)
E3O
2
E2O
E1O
1

1
1

1
1

4
3
4

2
3

2
3

1
2

2
1

2
1

CS1O

600
500
400
300

200

10

20

30

CS2O

40

50

60

70

Operating Time, h

80

90

100

110

[B]

10

20

30

3
40

50

CS3O
60

70

Operating Time, h

Figure 3 Cobalt concentration profiles in: [A] extraction; [B] strip

64

2
2

Loaded
Organic

100

[A]

700

Co in Strip Organics, mg/L

Feed

1000

4
3
4

80

33
90

100

110

Co in Loaded Org, g/L

4
3
4
2
3

800

10000

Co in Feed
Co in Discharge
Mn in Feed
Mn in Discharge

110

180

1350

170

1200

160

1050

150

900
90

750
600

80

Mn, mg/L

Cobalt, g/L

100

1500

Cobalt, g/L

120

Crystallisation
Feed

140
130
120

450

Crystallisation
Stage 3 Discharge

110
300

70

100
150

-0
3

-0
2

ay

ay
M

80
0

-0
1
ay

pr
-3
A

pr
-2
A

pr
-2

pr
-2
A

pr
-2

pr
-2

pr
-2
A

pr
-2
A

90

60

[A]

Date

10

20

30

[B]

40

50

60

70

80

90

100

Elapsed Operating Time, h

170

45

160

40

150

35

140

30

130

25

120

20

110

15

100

10

90

80

0
0

[A]

10

20

30

40

50

60

70

Elapsed Operating Time, h

80

90

100

Manganese, mg/L

50

280

35

240

30

200

25

160

20

120

15

80

10

40

Sodium, g/L

180

Sodium, g/L

Cobalt, g/L

Figure 4 Feed and discharge profiles: [A] Mn Removal circuit; [B] Crystallisation circuit

0
0

10

20

30

40

50

60

70

Elapsed Operating Time, h

80

90

100

[B]

Figure 5 Crystallisation feed profiles: [A] Co and Na; [B] Mn and Na

Final flowsheet
The final flowsheet (Figure 6) was established based on the performance of the pre-SX manganese
removal circuit that defined the optimized conditions for both reactors as follows: pH ~4.2
maintained with lime, ORP ~500 mV (measured vs. Ag|AgCl, satd KCl standard electrode)
realized with 2.3% (volume fraction) SO2 in O2, 50 C and ambient pressure. Under these operating
conditions, the manganese removal rates were 53 mg/Lh in tank 1, 27 mg/Lh in tank 2 and
40 mg/Lh across the circuit. The corresponding manganese removal efficiencies per hour of
residence time were ~65% in tank 1 and ~33% in tank 2, i.e. 49% across the circuit equivalent to 98%
with 2 h residence time. The corresponding overall cobalt loss was 12%. Overall pure reagent
consumptions in kg of reagent per kg manganese removed were 39.4 Ca(OH)2, 2.9 SO2 and 63.8 O2.
The results allowed prediction of an achievable sodium target of less than 5 g/t for the commercially
produced NICO heptahydrate cobalt sulphate product. Furthermore, the results of this work

65

considered with the proposed flowsheet allowed prediction of the key assays of the commercially
produced NICO cobalt carbonate by-product produced by the precipitation from the crystallisation
mother liquor bleed stream with a purity of 48.5% Co, with in g/t: 126 Ni, <124 Fe, 61 Mn, 7 Zn and
59 Mg. All of these targets were attained in the pilot plant product.

Cobalt Concentrate

O2

POX

CIL

180C 1 h
10% solids pH 1.8

O2

Co SX Feed
150 g/L
Na2(CO)3

Ca(OH)2

Na2CO3

Raffinate

Co Extraction
pH 5.3-4.8, 55C
4 stages
Loaded
Organic

Pond

Ni Ppt'n
pH 7

Scrub
Liquor

Gold

L
(acid or
base)

Fe-As Ppt'n
50C, 2 h, pH 4.6

Recycle to Fe-As Removal

H2SO4

Pond

Storage

10% v/v Cyanex 272

in Orfom SX80

Co Strip
pH 3.0-3.3, 30-40C

3 stages
Co-stripped Organic

L
Fe

Scrub
Feed

2 stages
Scrubbed
Organic

from Fe-As Ppte

pH 4.6-4.4, 55C

Cu Releach

S
L

Water

LO Scrub

Tailings

Zn-stripped
organic recycle

Zn Strip

H2SO4

pH 0.9-1, 25-30C

Cu Cementation

to extraction

2 stages
ZnSO4 Strip Liquor

S
Na2CO3

L
Smelter

Co Strip Liquor

Zn Ppt'n

Na2CO3

to evaporator

pH 10

Cu Removal

pH 6.45-7.45

S
Zn Sludge
to Storage

S
L
Cu Removal Discharge

SO2/O2 @
2.5% (v/v)

Ca(OH)2

Evaporation
Crystalliser Feed

Mn Ox/Ppt'n
50C, 500 mV

Ag|AgCl

pH 4.2

Mother
Liquor

Crystallisation

DI water
to wash

Flocculant

Na2CO3

Co Scavenging
Precipitation

DI Water

L
Tailings
Purified
Co SX Feed

Condensate

Wash
Water

Crystal Washing

(Centrifuge)

CoSO47H2O

CoCO3

Figure 6 Final flowsheet

Based on the dynamic settling response of the discharge scavenging precipitate stream at ~100 g/t
anionic flocculant dosage and 1.2 hours solids residence time, the predicted specific underflow and
hydraulic thickener unit areas were 1.6 and 0.7 m2/tday, respectively. The thickened underflow
displayed a light Bingham plastic rheological behaviour with a Critical Solids Density mass fraction
of ~5.9%, showing an unsheared yield stress of 5.5 Pa. Crystallisation discharge predicted the
maximum vacuum filtration throughput at 0.9 bar and 18 mm cake thickness of 1197 kg/m2h as

66

well as a discharge cake residual moisture mass fraction of 18.2%. Therefore, centrifugation could
be considered for the commercial operation. Predicted maximum vacuum filtration throughput for
the cobalt carbonate product under similar vacuum conditions with a 5 mm cake thickness was 127
kg/m2h and a discharge cake residual moisture mass fraction of 47.7%.

CONCLUSION
A process flowsheet for the direct production of high purity cobalt sulphate heptahydrate and
cobalt carbonate by-product was established and validated.

REFERENCES
Mezei, A., Ferron, C.J., Krstic, S., & Goad, R.E. (2000) Recovery of Cobalt, Gold and Bismuth from the NICO
Deposit, NWT, Canada - Part II: Recovery of Bismuth, Proceedings of SME Congress, Minor Elements
2000, Salt Lake City, Utah.
Mezei, A., Fleming, C.A., Canizares, M., Ashbury, M., Brown, J., Todd, I., Imeson, D., Jackman, I., Goad, R.E.,
Samuels, M., Ferron, C.J., Hayden, A., Molnar, R., Konigsmann, K., King, D., Marshall, D., & Ngai, E.
(2008) Recovery of Cobalt, Gold and Bismuth from the Polymetallic Concentrates - NICO Deposit,
NWT, Canada Pilot Plant Results, Proceedings of ALTA 2008 Congress, Nickel and Cobalt, Perth, WA,
May.

67

Copper extraction from oxide

ore using sea water

Because mining activities in Antofagasta Chile, an arid region, have limited

access to water, the use of sea water has been an option for many copper mines
in the region. The Chilean mining industry has extensive knowledge of conventional mining processes using untreated, potable water sources. The
substitution of sea water for freshwater has been used in different copper
processes such as leaching by some mining companies. However, there is very
little information in the public domain about the consequences of substituting
seawater for fresh water. The primary goal of this research is to understand
the mechanisms (from a chemical and physical-chemical point of view) that

Cynthia Torres, Mara Elisa

Taboada, Tefilo Graber and Hctor
Galleguillos. CSIRO Chile
International Centre of Excellence in
Mining and Mineral Processing,
Universidad de Antofagasta, Chile
Helen Watling. CSIRO Chile
International Centre of Excellence in
Mining and Mineral Processing and
CSIRO Minerals Down
Under Flagship, Australian Minerals
Research Centre, Australia

are affected by the use of sea water in leaching and other processes. Focused
studies simulating different unit processes will become the foundation for
optimizing those processes and/or describing the consequences of using
saline water.
For this initial study, the effects of water quality (tap water, seawater, process water) on copper leaching (H2SO4, HCl as leaching agents) were investigated at three temperatures (25C, 35C and 45C) using a low-grade copper
oxide ore (0.36%Cu) pulverized to 100% passing 100mesh (-150m). The
leached residues were characterized using X-ray diffraction and sem-eds. An
empirical equation was used to adjust the experimental data of the leaching
kinetics correlating the extraction (%) with the copper concentration, and
temperature. As preliminary results it was observed that; (i) copper extractions
between 70 and 80% were achieved using sea water, similar to the extractions
obtained using tap water. (ii) Rapid copper dissolution was obtained using a
temperature higher than 45C.

69

INTRODUCTION
Because mining activities near Antofagasta-Chile, an arid region, have limited access to fresh water,
the use of sea water has been an option for many copper mines in the region. The Chile mining
industry has extensive knowledge of conventional mining processes using untreated, potable water
sources. The substitution of sea water for freshwater has been used in different copper processes
such as leaching by some mining companies(Bernal. Minera Esperanza 2009). However, there is
little information in the public domain about the consequences of substituting seawater for fresh
water. (Senanayake 2007) .
Leaching of copper oxide ore has been studied in many investigations. It is well known that copper
oxide ore can be leached using sulphuric acid as leaching agent and, depending on ore mineralogy,
acid consumption ranges from 0.4 to 0.7 ton H2SO4 per ton of copper recovered (Bingl and
Canbazolu 2004). During leaching experiments, the aim is to dissolve the maximum amount of the
oxide ore in the minimum amount of time. At the same time, for the purpose of optimising leaching
or making comparisons, it is important to control those parameters that affect the rate of
dissolution, such as temperature, solution acidity (Park, Mohapatra et al. 2007), the ratio of
solid/liquid (solids loading) and water quality. (Habbache, Alane et al. 2009). The effect of acid
concentration using H2SO4 solution in a concentration range of 0.12 M at 25C has been studied by
(Shayestehfar, Nasab et al. 2008). They reported that the amount of copper extracted increased
when the acid concentration was increased from 0.1 to 0.5 M, but that no further extraction
occurred when the acid concentration was increased from 0.5 to 2 M. (Bingl and Canbazolu 2004)
investigated the effect of solid-liquid ratio on the dissolution of malachite. They reported that
copper recovery increased with decreased solids loading. This result was attributed to the
decreased amount of leaching agent per unit amount of solid with the increase in solids loading.
The primary goal of a broad research strategy at the Laboratorio de Investigacion de Procesos (LIP),
University of Antofagasta (UA) is to understand the mechanisms (from chemical and physicalchemical points of view) that are affected by the use of sea water in leaching and other processes.
Recent studies at LIP have been focused on the effects of seawater on the solubility and physicalchemical properties of acidic copper sulphate solutions (Hernndez, Galleguillos et al. 2012), the
behaviour of sodium nitrate and caliche mineral in seawater (Taboada, Hernndez et al. 2012), and
the recovery of nitrates from leaching solutions using seawater (Torres, Meruane et al. 2013).
Overall, the results confirm that seawater can be substituted successfully for fresh water in different
mining processes. As a result, a secondary goal is to find effective ways of exploiting and managing
the use of seawater in mining processes. Focused studies simulating different unit processes will
become the foundation for optimising those processes and/or describing the consequences of using
saline water.
In the present study, copper extraction from a low-grade copper oxide ore was investigated using
H2SO4 or HCl as leaching agents following example of (Majima, Awakura et al. 1980) in seawater at
three temperatures. In addition, comparison was made of the leaching of the ore in seawater, fresh
water and process water in acidified media. Soluble copper was measured during the leaching
experiments and the leached residues were analysed using digestion and ICP-AES analysis and also
characterised using quantitative X-ray diffraction and SEMEDS.

70

MATERIALS AND METHODS

Materials
Analytical grade reagents: H2SO4, HCl, NaCl, CuSO4.5H2O, FeCl3 and NaNO3 (Merck) were used to
prepare leach media. The seawater used in the experiments was obtained from the Marine
Resources department at UA via an intake collector 200 m off the coast of San Jorge Bay,
Antofagasta, Chile. The seawater was passed through a quartz sand filter (50 micron) and a
mechanical polyethylene filter (1 micron) before use, to remove solid particles and microorganisms.
Synthetic process water was prepared by adding Na+, Cl, Cu2+ and SO42 to seawater. The major
element concentrations in seawater and synthetic process water are shown in Table 1. The copper
oxide ore was obtained from a mining Company located closed to Antofagasta and contained 0.36
% Cu, mainly as tenorite (Table 2) with acid consumption of 37.30 Kg H2SO4 per ton of ore. The ore
was crushed and milled to a fine powder (-100 mesh)as recommended by (Ata, olak et al. 2001),
who showed that the extraction of copper increases with decreasing particle size.
Table 1 Element concentrations in seawater and process water

Description
Seawater
Process water

Na
9480
17255
+

K
374
374
+

Mg
1190
1190
2+

Ca
386
386
2+

mg L1
Cl
SO42
18765
2771
30765
5794

NO3
2.05
2.05

B3+
4.6
4.6

HCO3
142
142

Cu2+
0.072
2072

Table 2 Mineralogical composition of the test ore

Name of mineral

Formula

(%)

Quartz

SiO2

39.62

Albite

NaAlSi3O8

18.58

Illite

(K,H3O)Al2Si3AlO10(OH)2

18.06

Chlorite-1MIIb

(Mg5Al)(Si,Al)4O10(OH)8

1.38

Orthoclase

KAlSi3O8

13.04

Hematite

Fe2O3

1.86

Jarosite

K(Fe3(SO4)2(OH)6)

0.41

Anhydrite

CaSO4

1.90

Biotite

K(Mg,Fe ) (Al,Fe )Si3O10(OH,F)2

4.65

Tenorite

CuO

0.50

+2 3

+3

Copper extraction in flask tests

In a typical experiment, 0.2 g of ore and 100 mL of medium (Table 3) were each added to 250 mL
Erlenmeyer flasks. The flasks were placed on an orbital shaker (200 rpm). Samples of the leach

71

solution were obtained periodically using a membrane filter (pore size 0.45 m) on a syringe.
Copper concentrations in leachates were determined using atomic absorption spectroscopy.
Table 3 Conditions used in flask tests with seawater (see also Figure 3)

Test
A
B
C
D

Ore
Mass (g)
0.2
0.2
0.2
0.2

Mesh
100
100
100
100

Condition
Acid (M)
FeCl3 (M)
HCl (0.5)
H2SO4 (0.5) HCl (0.5)
0.002
H2SO4 (0.5) -

T (oC)
25
25
25
25

NaNO3 (M)
1

Copper extraction in agitated reactors

Agitated leaching experiments were conducted in instrumented reactors. Each reactor had a 2000
cm3 volume capacity, glass top with multi-port lid on which a stirrer was mounted (Figure 1). A
sample mass of 100 g per litre of solution was weighed into each reactor (unless otherwise noted).
As above, solution samples were collected periodically for analysis.

Figure 1 Instrumented reactor

RESULTS AND DISCUSSION

Agitated leaching with shake flasks
Small-scale leach tests in shake flasks provided preliminary copper extraction data (Figure 2). The
results were equivocal but might be interpreted as showing that H2SO4 was a more effective
leaching agent than HCl. The additions of 1 M NaNO3 (85 g L1) to the H2SO4 medium or of 0.002 M
FeCl3 (8 g L1) to the HCl medium had no significant effect on copper extraction. Ferric ions would
only be expected to enhance copper extraction if the copper was present as a sulphide mineral.

72

Overall, it was concluded that the initial quantity of sample used in these tests was not sufficient to
obtain robust extraction data. Therefore, leaching tests were performed using a greater quantity of
sample in an agitated reactor in order to probe more carefully the effect of leaching agent and
additives on copper extraction as well as secondary reaction products in leached residues.

% Cu Recovery

% of Cu Recovery vs.time 25C

45
40
35
30
25
20
15
10
5
0

A
C
D
B

20

40

60

80

100

Time (hour)
Figure 2 Cu recovery (%) versus time in different media (see Table 3, AD)

Agitated leaching with instrumented reactor

A comparison of the two leaching agents, H2SO4 and HCl (0.5 M), under otherwise similar leaching
conditions in seawater, showed no important difference between them (Figure 3). Subsequent
reactor tests were conducted in media acidified with H2SO4.

% of Cu Recovery vs. Time 25C

60

% Cu Recovery

50
40
30
20

HCl + SW
H2SO4 + SW

10
0
0

50

100
Time (Hour)

150

Figure 3 Copper recovery (%) versus time using different leaching agents in seawater media

73

Rapid copper extraction occurred in both seawater and tap water media, with ~70% Cu recovery
within one hour (Figure 4), which is a positive result from an economic point of view. But the
greatest extraction was achieved using synthetic process water. In that test, copper extraction began
slowly, lagging behind the extractions in seawater and tap water in the first eight hours but then
exceeded them. The final copper extraction was >90% at 20 hours. The quantitative XRD and
chemical analyses of the leached residues revealed only subtle differences for the different media.
The key parameter in synthetic process water causing enhanced copper extraction is the subject of
ongoing research.

% Cu Recovery

% of Cu Recovery vs. Time 25C

100
90
80
70
60
50
40
30
20
10
0

H2SO4 0.5M + SW
H2SO4 0.5 M + TW
H2SO4 0.5 M + PW

,00

5,00

10,00

15,00

20,00

25,00

Time (hour)
Figure 4 Cu recovery (%) versus time using seawater (SW), tap water (TW) or process water (PW)

Leaching experiments in 0.5 M H2SO4 medium at temperatures 25, 35 and 45 C yielded between
6570% copper extractions within the first hour. Slightly increased copper extraction was measured
with increased temperature over 20 h (Figure 5) with final extractions in the range 7085%. The
small increase in copper extraction at increased temperature is consistent with what is known about
oxide leaching in heaps, which typically do not become hot unless there is a sulphide component in
the ore to generate heat during oxidation. It is suggested that the process is chemically controlled,
an hypothesis that should be confirmed by kinetic analysis. The equation used to adjust the
experimental data is:
=

(1)

Using

74

(2)

The parameters Ea, ko, n were obtained by minimization of the objective function (2) getting the
following values; Ea= 7102 (cal/mol), k0=4.72*104, n=0.4. . Where E is the extraction percentage of
copper,
= the extraction percentage of copper at infinite time, t is time in minutes, n is an
adjustable parameter, k is the kinetic contacts of the extraction process

% of Cu Recovery vs Time

100,00
90,00

% Cu Recovery

80,00
70,00
60,00
50,00

H2SO4 0.5 M + SW+25C

40,00

H2SO4 0.5M+SW+35C

30,00
H2SO4 0.5 M +SW+45C

20,00
Equation

10,00
,00
0

200

400

600

800

1000

1200

1400

Time (min)

Figure 5 Cu recovery (%) versus time at different temperatures

An increase from 100 g to 300 g L1 of ore did not influence copper extraction (Figure 6). High solids
loading may have suppressed the leaching of an iron-bearing mineral, leaving a greater iron
content in the leached residue. Biotite was the main gangue mineral that dissolved under the leach
conditions, with minor amounts of calcium silicate also being dissolved. The main secondary
reaction product in the leached residues was gypsum.

% Cu Recovery

% of Cu Recovery vs Time S/L change at 25C

80
70
60
50
40
30
20
10
0

H2SO4 0.5 M + SW
H2SO4 0.5 M + SW + 300g
H2SO4 0.5M + SW + 200g

10

15

20

25

Time (hour)
Figure 6 Cu recovery (%) versus time for different solids loadings

75

CONCLUSIONS
Prelmnary tests showed that either H2SO4 or HCl could be used in seawater media with equal
effect; copper extraction was rapid and in the range of 65-70% within an hour and increased slightly
in a 20 hour leach. The addition of NaNO3 to seawater acidified with H2SO4 or the addition of FeCl3
to seawater medium acidified with HCl had little to no effect on copper extraction. In larger scale
tests acidified with H2SO4, copper extraction was initially rapid but similar in seawater and tap
water media. However, while extraction was slow initially in synthetic process water medium, it
increased to >90% in a 20 hour leach, prompting further research to identify the key component that
had caused the enhancement in copper extraction. A small increase in copper extraction occurred
with increased temperature but increased solids loading had minimal effect.

ACKNOWLEDGEMENTS
Assistance in ore preparation by the Mining Engineering and Chemical Engineering Departments
of UA is acknowledged with thanks. The authors are grateful for the financial support provided by
CSIROChile, CONICYT through Fondecyt Project N 1100685, and CICITEM R04I1001.

REFERENCES
Ata, O. N., S. olak, et al. (2001). "Determination of the Optimum Conditions for Leaching of Malachite Ore in
H2SO4 Solutions." Chemical Engineering & Technology 24(4): 409-413.
Bernal., L. (Minera Esperanza 2009). "Taller "Uso eficiente del Recurso Hidrico en Mineria" Universidad
Catolica del Norte."
Bingl, D. and M. Canbazolu (2004). "Dissolution kinetics of malachite in sulphuric acid." Hydrometallurgy
72(12): 159-165.
Habbache, N., N. Alane, et al. (2009). "Leaching of copper oxide with different acid solutions." Chemical
Engineering Journal 152(23): 503-508.
Hernndez, P. C., H. R. Galleguillos, et al. (2012). "Effect of Seawater on the Solubility and Physicochemical
Properties of Acidic Copper Sulfate Solutions." Journal of Chemical & Engineering Data 57(9): 2430-2436.
Majima, H., Y. Awakura, et al. (1980). "Acid dissolution of cupric oxide." Metallurgical Transactions B 11(2): 209214.
Park, K.-H., D. Mohapatra, et al. (2007). "Dissolution behavior of a complex CuNiCoFe matte in CuCl2
NaClHCl leaching medium." Separation and Purification Technology 56(3): 303-310.
Senanayake, G. (2007). "Review of theory and practice of measuring proton activity and pH in concentrated
chloride solutions and application to oxide leaching." Minerals Engineering 20(7): 634-645.
Shayestehfar, M. R., S. K. Nasab, et al. (2008). "Mineralogy, petrology, and chemistry studies to evaluate oxide
copper ores for heap leaching in Sarcheshmeh copper mine, Kerman, Iran." Journal of Hazardous
Materials 154(13): 602-612.
Taboada, M. E., P. C. Hernndez, et al. (2012). "Behavior of sodium nitrate and caliche mineral in seawater:
Solubility and physicochemical properties at different temperatures and concentrations."
Hydrometallurgy 113114(0): 160-166.

76

Torres, M. A., G. E. Meruane, et al. (2013). "Recovery of nitrates from leaching solutions using seawater."
Hydrometallurgy 133(0): 100-105.

77

Kinetics of chalcopyrite leaching in

either ferric sulfate or cupric sulfate
media in the presence of NaCl

The study of leaching kinetics is particularly important for process design,

optimization and control. Despite a considerable volume of published literature
addressing the kinetics of chalcopyrite leaching with Fe(III) and more recently,

Tcia Veloso, Johne Peixoto,

Michael Rodrigues and Versiane
Albis Leo. Universidade Federal de
Ouro Preto, Brazil

Cu(II) solutions, there is no agreement on the rate-determining step, the activation energy and the influence of the solution composition on the process.
Aiming at bringing further insight into these topics, the present work sought
to compare the leaching kinetics of chalcopyrite with either ferric sulfate or
cupric sulfate, in the presence of NaCl. Batch experiments were performed in
a 1L water jacketed reactor, containing 500mL of pulp with 0.1% solids (w/v),
1mol/L H2SO4, at 850min-1. The effects of temperature (70 to 90C) and reagent
[Fe(III), Cu(II) and Cl-] concentrations on chalcopyrite dissolution were determined. The results showed that chalcopyrite leaching is faster with cupric
sulfate than ferric sulfate and that the activation energy was 92,9kJ/mol with
the former and around 122,3kJ/mol with the latter. The addition of sodium
chloride to the leaching solutions increased the dissolution rate, as observed
in the experiments performed with 2mol/L chloride and 1mol/L ferric ion,
in which chalcopyrite dissolution approached 24%, in 180 minutes, whereas
it was only 5% in the absence of anion. When ferric iron was replaced by cupric
ions (in the presence of chloride), chalcopyrite dissolution increased to 46%
in the same period.

79

INTRODUCTION
Chalcopyrite (CuFeS2) is the most abundant copper mineral, comprising roughly 70% of the
worldwide copper reserves and its relevance to the copper industry is evident, accordingly (Kaplun
et al. 2011). Such a mineral is also a very stable copper sulphide and refractory to
hydrometallurgical processing (Aries & Leo, 2007). For this reason, chalcopyrite leaching is slow as
compared to that of other copper sulphides such as chalcocite (Cu2S) and bornite (Cu5FeS4). Thus,
numerous studies have been carried out seeking to determine the various parameters influencing
its leaching kinetics, such as temperature, particle size, stirring rate, solution type, among others
(Yvenes, Miki & Nicol, 2010).
With respect to the leaching solution, hydrometallurgical processes applied to chalcopyrite ores
and/or concentrates can be classified as nitrate, chloride, sulphate or ammonia leaching and all of
them require the presence of an oxidant agent, such as either ferric or cupric ions or oxygen. The
sulphate-based processes have some advantages because sulphate leaching is generally simpler,
better understood and the recovery of copper by solvent extraction - electrolysis is relatively
straightforward. However, there is an critical shortcoming: the leaching kinetics of chalcopyrite is
slow (Hacklet al, 1995). One way of circumventing this problem, which has been extensively
investigated, is the addition of chloride ions to the sulphate medium (Ruiz, Hills & Padilla, 2011).
According Aries & Leo (2007), the presence of chloride ions: (i) reduces chalcopyrite passivation,
whereas the presence of the Cu(I)-chloride complexes adds a second redox couple to the system,
thereby improving leaching kinetics.
Chalcopyrite oxidation by with ferric iron is given by equation (1) (Dutrizac, 1981):
CuFeS

( )

+ 4Fe(

= Cu(

+ 5Fe(

+ 2S( ) (1)

Furthermore, when chloride ions are present, stable complexes with cuprous ions are formed and
chalcopyrite oxidation can also be carried out by the Cu(II)/Cu(I) couple as represented by the
equation (2) (Carneiro & Leo, 2007). In the presence of either oxygen or ferric iron, Cu(I) is readily
converted back to Cu(II).
CuFeS

( )

+ 3Cu(

= 4Cu(

+ Fe(

+ 2S( ) (2)

The reason for the low leaching rate observed with ferric sulphate is the formation of a passivation
layer, as stated, which occurs during the oxidative leaching and that has been extensively studied.
There is still no consensus of the nature of this layer, as some authors claim that it is formed by
elemental sulphur, while others argue that it is a metal polysulfide. Regardless of its nature, such a
layer is seen as a barrier to the diffusion of the oxidant (Munoz, Wadsworth & Miller, 1979). There
is also disagreement with respect to which step controls the leaching kinetics of chalcopyrite,
chemical reaction (Hirato, Majima & Awakura, 1987a; Cordoba et al., 2008; Ruiz Montes & Padilla,
2011; Kaplun, et al., 2011) or diffusion through the elemental sulphur layer (Munoz, Miller &
Wadsworth, 1979; Bonan et al., 1981; Aries & Leo, 2007). Within such a context, the aim of the

80

present work was to determine the effects of temperature, type of oxidant (cupric or ferric ions) and
their initial concentration on the leaching rate of chalcopyrite with both sulphate and chloride ions.

MATERIALS AND METHODS

Sample preparation
The sample utilized in the leaching experiments was a chalcopyrite chunk, which was crushed and
then ground in a orbital mill. This sample was subjected to analysis in a laser particle size analyzer,
model Cilas 1064. The average particle size of the sample was 17,72m, with d10=2.64 m,
d50=13.56m and d90=39.36m. X-ray diffraction (Shimadzu, model 6000) showed that the main
mineral phase in the sample was chalcopyrite, and traces of magnetite, pyrite and calcite. Its
content of copper, iron and sulphur was quantified by acid digestion followed by analysis in an
atomic emission spectrometer (Varian, 725), which revealed the presence of 30.6% Cu, 28.2% Fe and
34.3% S, indicating that chalcopyrite represents nearly 90% of the sample.

Leaching tests
The chemical leaching tests were performed batchwise, using a 1L jacketed glass reactor.
Temperature was controlled by circulating water heated by a thermostatic bath equipped with an
external circulation pump. The glass reactor was magnetically stirred and the value of 850min-1 was
selected to ensure that the film diffusion resistance was negligible. Leaching solutions were
prepared from analytical grade chemicals (ferric sulphate, cupric sulphate, sulphuric acid and
sodium chloride) and distilled water and the effect of chloride ions (0mol/L to 2mol/L) on
chalcopyrite leaching was determined at 80C, using 1mol/L sulphuric acid and 1mol/L ferric
sulphate. The influence of ferric iron (0mol/L to 1mol/L) was also assessed at 80C, in the presence
of 1mol/L sulphuric acid and 2mol/L sodium chloride. Similar conditions were utilized to analyse
the effect of cupric copper (0mol/L to 0.5mol/L) on the leaching rate. Finally, temperature was
varied from 70C to 90C, in tests performed with 1mol/L sulphuric acid, 2mol/L sodium chloride
and either 1mol/L ferric iron or 0.5mol/L cupric sulphate.
Throughout the experiments, a 0.1% (w/v) solids concentration was utilized and 1.5mL samples
were collected regularly for chemical analysis (copper and iron), carried out in an ICP-OES (Varian,
725). In the experiments with ferric iron, copper concentrations were utilized to determine
chalcopyrite dissolution, whereas iron concentrations were used with the same aim in the
experiments with cupric sulphate and sodium chloride.

RESULTS AND DISCUSSION

Effect of chloride ions on chalcopyrite leaching
It is well known that the presence of chloride ions increases the rate of chalcopyrite dissolution in
sulphate media, but there is no agreement on what is the threshold for which copper extraction
stops to be benefited. Lu, Jeffrey & Lawson (2000) and later Ruiz, Hills & Padilla (2011) suggested
that sodium chloride significantly increases the leaching rate up to 0.5mol/L NaCl, while Palmer et
al. (1981) and Lamb & Lion (2007) set such threshold at 1.0mol/L NaCl.

81

Figure 1 shows copper extraction yields as a function of time for concentrations of chloride ions
varying from 0mol/L to 2.0mol/L. It can be observed that, in three hours of testing, copper
dissolution increased from 5% in the absence to 22% with 1.5mol/L NaCl confirming the positive
effect of chloride ions in copper extraction from chalcopyrite. Nevertheless, such improvement on
copper extraction became negligible at high chloride concentrations because when the NaCl
concentration was increased further to 2.0mol/L, the final extraction reached only 24%.
Notwithstanding, 2.0mol/L NaCl solutions were utilized throughout the kinetic study to ensure
that chloride concentrations did not limit copper extraction.
3+

1.0 mol/L H2SO4, 1.0 mol/L Fe and T= 80C

Concentration of NaCl (mol/L):
0.0
1.0
1.5
2.0

Extraction of copper (%)

25

20

15

10

0
0

2000

4000

6000

8000

10000

Time (s)

Figure 1 Effect of concentration of NaCl on chalcopyrite dissolution Error! No hay texto con el estilo
especificado en el documento.

Figure 2 depicts copper extraction from chalcopyrite for different concentrations of either ferric iron
(a) or cupric ions (b). In the absence of either oxidant, 15% copper and 14% iron were extracted,
while copper extraction increased with the ferric iron concentrations up to 0.5mol/L Fe3+. Final
extractions reached 24% and 46% copper dissolution with 1.0mol/L ferric iron and 0.5mol/L cupric
ions, respectively, in three-hour testing. Such results show that in sodium chloride-sulphuric acid
solutions, chalcopyrite oxidation by cupric ions is more effective than that achieved with ferric
sulphate. Such outcome has been ascribed to a faster and more reversible Cu2+ reduction to Cu+ (on
the chalcopyrite surface), as compared to that of the Fe3+/Fe2+ couple (Dutrizac, 1992).
Regarding the initial oxidant concentration, it was noticed that copper extractions increased from
18% to 24%, as the initial Fe3+ content varied from 0.1mol/L to 1.0mol/L, respectively (Figure 2 (a)),
i.e., a relatively small increase. According to Cordoba et al. (2008) the rate of chalcopyrite
dissolution is strongly affected by the initial ferric iron concentration, but only for values below
0.01mol/L. At higher concentrations, such effect is negligible, which is consistent with the findings
of the present study. In those tests carried out utilizing cupric ions as the oxidant (Figure 2(b)),
similar results were observed (Figure 2 (b)). Chalcopyrite dissolution in experiments with 0.125
mol/L and 0.5mol/L CuSO4 were 42% and 46%, respectively, i.e. a small increase. This implies that
the rate of chalcopyrite dissolution is not strongly dependent on the concentration of oxidizing
agent.

82

1.0 mol/L H2SO4, 2.0 mol/L NaCl and T= 80C

a)

2+

1.0 mol/L H2SO4, 2.0 mol/L NaCl e T= 80C

50

Concentrao de ion Fe (mol/L):

0.0
0.02
0.1
0.25
0.5
1.0

Extraction of iron (%)

Extraction of copper (%)

3+

40

30

20

10

0
0

2000

4000

6000

8000

10000

Time (s)

b)

Concentration of ion Cu (mol/L):

0.0
0.125
0.25
0.5

50

40

30

20

10

0
0

2000

4000

6000

8000

10000

Time (s)

Figure 2 Effect of concentration of ion Fe3+ (a) and Cu2+ (b) on chalcopyrite dissolution

Kinetics of chalcopyrite leaching

Temperature is an important parameter in the dissolution of copper sulphides and, as expected, its
increase results in faster leaching kinetics and larger extraction yields. The effect of temperature on
the leaching kinetics of the chalcopyrite sample under study was investigated in the range 70C to
90C, so that the apparent activation energy (Ea) could be assessed for both oxidants in the presence
of sodium chloride. Figure 3(a) presents chalcopyrite leaching as a function of time for a solution
containing 1.0mol/L ferric iron, whereas Figure 3(b) shows the same results for the Cu(II)/Cu(I)
couple (0.5 mol/L cupric ions). Applying ferric ions (Figure 3(a)), copper extraction increased from
17% to 50% as the temperature varied from 70oC to 90oC. In the presence of cupric ions (Figure 3
(b)), the same parameter increased from 35% to 70%, in three hours of the experiment. Such results
emphasise the dependence of chalcopyrite dissolution on temperature. Analysing both figures, it
can be noticed that both leaching rate and extraction yields are higher when leaching is carried out
with cupric ions. Nonetheless, the performance of both oxidants approach each other at higher
temperatures, i.e. up to 80C, the leaching rate determined in tests with cupric ions was
approximately twice as higher as that achieved with ferric ions, but at both 85C and 90C, it is only
1.4 times larger.
From data presented in Figure 3, the kinetics parameters related to chalcopyrite leaching were
determined by applying the shrinking core model. After reducing the film diffusion resistance, the
experimental data were fitted to both chemical reaction and ash layer diffusion controls, but the
latter described better the kinetics of the oxidative leaching of chalcopyrite (Figure 4). According to
this latter mechanism, the fraction reacted (X) and time (t) are related by equation 3, where Kd is the
rate constant (Munoz, Miller & Wadsworth, 1979):

F(X) = 1 3 (1 X) + 2 (1 X) = K t (3)

83

1.0 mol/L H2SO4, 2.0 mol/L NaCl

Temperature:
70 C
75 C
80 C
85 C
90 C

Extraction of copper (%)

60
50

1,0 mol/L H2SO4, 2,0 mol/L NaCl

a)

Temperature:
70 C
80 C
90 C

70

Extraction of iron (%)

70

40
30
20
10

60

b)

75 C
85 C

50
40
30
20
10

2000

4000

6000

8000

10000

2000

4000

Time (s)

6000

8000

10000

Time (s)

Figure 3 Effect of temperature on chalcopyrite dissolution: a) 1.0mol/L Fe3+ and b) 0.5mol/L Cu2+

2/3

0.16

y = 2.27E-5x R = 0.99

y = 1.16E-5x R = 0.99

0.12

y= 8.02E-6x R = 0.98
0.08

y= 5.24E-6x R = 0.94

2/3

1-3*(1-x) +2*(1-x)

0.20

1.0 mol/L Fe
Temperature:
70 C
75 C
80 C
85 C
90 C

0.10

1-3*(1-x) +2*(1-x)

0.24

a)

3+

b)

2+

0.5 mol/L Cu
Temperature:
70 C
75 C
80 C
85 C
90 C

0.08
0.06

y= 9.24E-6x R = 0.99

y= 4.76E-6x R = 0.99

0.04
2

y= 1.88E-6x R = 0.99
0.02

0.04

y= 1.28E-6x R = 0.99
2

y = 3.57E-6x R = 0.93

y= 9.4E-7x R = 0.99

0.00

0.00
0

2000

4000

6000

8000

Time (s)

10000

12000

2000

4000

6000

8000

10000

12000

Time (s)

Figure 4 Fit of shrinking core model for diffusion control a) 1.0mol/L Fe3+ and b) 0.5mol/L Cu2+

The application of equation 3 to data produced at different temperatures, the activation energy was
determined as 122.3kJ/mol for experiments with 1.0 mol/L Fe3+ and 92.9 kJ/mol, for 0.5 mol/L Cu2+
ion. Presented in Table 1 are activation energy values achieved in different studies and values
ranging from 75kJ/mol to 130kJ/mol are reported for ferric iron leaching in the 35C - 100C
temperature range (Cordova et al. 2,008; Dutrizac, 1981). Leaching with both copper and chloride
ions (from 60C to 105C) produced activation energy values in the range 59.5 kJ/mol to 134.7kJ/mol
(Bonan et al. 1981; Hirato, Majima & Awakura, 1987a). High values of the activation energy imply
in the need of high temperatures to achieve large copper extractions in short periods. Dutrizac
(1981) explained such differences by stressing that activation energies are affected by both the (i)
temperature, (ii) reagent concentration and (iii) the particle size range investigated as well as the
(iv) presence of impurities and other copper minerals in the chalcopyrite sample along with (v) its
origin and finally by (vi) the kinetic model selected. The values herein achieved are too high for an
ash layer diffusion-controlled system. Such a result is very likely related to the broad particle size
distribution utilized in the experiments, which results in an erroneously fitting to equation 3, even
though chemical reaction is actually controlling the leaching kinetics as discussed by Gbor & Jia
(2004).

84

Table 1 Some values of activation energy for chalcopyrite dissolution reported by literature
Medium

T (C)

Ea (kJ/mol)

Kinetic control

Reference

Ferric Sulphate

50-94

75

Parabolic, diffusion control

Dutrizac (1981)

Ferric Sulphate

60-90

83,7

Parabolic, diffusion control

Muoz et al. (1979)

Ferric Sulphate

50-78

88

Linear , diffusion control.

Hirato et al. (1987b)

Ferric Sulphate +
NaCl

70 - 90

122,3

Parabolic, diffusion control

This work

Ferric Sulphate

35-68

130,7

Parabolic, Chemical control

Crdoba et al. (2008)

CuCl2 + NaCl

85-104

71

Parabolic, diffusion control

Bonan et al. (1981)

Cupric Chloride

60-85

81,5

Linear, Chemical control

Hirato et al. (1987a)

Cupric Sulphate +
NaCl

70-90

92,9

Parabolic, diffussion control

This work

CONCLUSION
As suggested by the high values of activation energy, chalcopyrite leaching is highly temperature
dependent, with faster kinetics in cupric sulphate-chloride medium as compared to ferric sulphatechloride solutions, i.e. 50% dissolution was achieved using ferric iron and 70% with cupric ion. The
improvement provided by chloride additions is larger up to 1.0mol/L NaCl, reducing when the
concentration is increased further to 2.0mol/L NaCl. A broad particle size distribution very likely
influenced the kinetic analysis and implied in a good data fitting to the equation representing ash
layer diffusion control, despite the very large values of the activation energy achieved i.e.
92.9kJ/mol (with 0.5 mol/L Cu2+) and 122.3kJ/mol, for experiments with 1.0mol/L Fe3+.

ACKNOWLEDGEMENTS
The authors are grateful to Conselho Nacional do Desenvolvimento Cientfico e Tecnolgico
(CNPq), Fundao de Amparo Pesquisa do Estado de Minas Gerais (FAPEMIG), Coordenao de
Aperfeioamento de Pessoal de Nvel Superior (CAPES) and Financiadora de Estudos e Projetos
(FINEP) for financial support. The Conselho Nacional de Pesquisas CNPq scholarship to V. A.
Leo and to Peixoto, J. J. M., the Aperfeioamento de Pessoal de Nvel Superior CAPES
scholarship to T. C. Veloso are especially acknowledged.

85

REFERENCES
Bonan, M., Demarthe, J.M., Renon, H. & Baratin, F. (1981) Chalcopyrite Leaching by CuCl2 in Strong NaCl
solutions, Metall. Trans. B, vol. 12B, pp. 269274.
Carneiro, M.F.C. & Leo, V.A. (2007) The role of sodium chloride on surface properties of chalcopyrite leached
with ferric sulphate, HydrometaIlurgy, vol. 87, pp. 7382.
Crdoba, E.M., Muoz , J.A., Blzquez M.L., Gonzlez, F. & Ballester, A. (2008) Leaching of chalcopyrite with
ferric ion. Part I: General aspects, HydrometaIlurgy, vol. 93, pp. 8187.
Dutrizac, J.E. (1981) The Dissolution of Chalcopyrite in Ferric Sulfate and Ferric Chloride Media, Metall.
Trans. B, vol. 12B, pp. 371378.
Dutrizac, J.E. (1992) The leaching of sulphide minerals in chloride media, HydrometaIlurgy, vol. 29, pp. 145.
Gbor, P.K. & Jia, C.Q. (2004) Critical evaluation of coupling particle size distribution with the shrinking core
model, Chemical Engineering Science, vol. 59, pp. 19791987.
Hackl, R.P., Dreisinger, D.B., Peters, E. & King J.A. (1995) Passivation of chalcopyrite during oxidative
leaching in sulfate media, HydrometaIlurgy, vol. 39, pp. 2548.
Hirato, T., Majima, H. & Awakura, Y. (1987a) The leaching of chalcopyrite with cupric chloride, Metall. Trans.
B, vol. 18B, pp. 3139.
Hirato, T., Majima, H. & Awakura, Y. (1987b) The leaching of chalcopyrite with ferric sulfate, Metall. Trans.
vol. 18B, pp. 489496.
Kaplun, K., Li, J. Kawashima, N. & Gerson, A.R. (2011) Cu e Fe chalcopyrite leach activation energies and the
effect of added Fe3+, Geochimica et Cosmochimica Acta, vol 75, pp. 58655878.
Lu, Z.Y., Jeffrey, M.I. & Lawson, F. (2000) An electrochemical study of the effect of chloride ions on the
dissolution of chalcopyrite in acidic solutions, HydrometaIlurgy, vol. 56, pp. 145155.
Lundstrm, M., Aromaa, J., Forsn, O., Hyvrinen, O. & Barker, M. H. (2005) Leaching of chalcopyrite in
cupric chloride solution, HydrometaIlurgy, vol. 77, pp. 8995.
Majima, H., Awakura, Y., Hirato, T. & Tanaka, T. (1985) the leaching of chalcopyrite in ferric chloride and
ferric sulfate solutions, Canadian Metallurgical Quarterly, vol. 24, pp. 283291.
Munoz, P. B., Miller, J.D. & Wadsworth, M.E. (1979) Reaction Mechanism for the Acid Ferric Sulfate Leaching
of Chalcopyrite, Metallurgical Transactions B, vol. 10B, pp. 149158.
Palmer, B.R., Nebo, C.E., Rau, M.F. & Fuerstenau, M.C. (1981) Rate phenomena involved in the dissolution of
Chalcopyrite in chloride-bearing lixiviants, Metallurgical Transactions B, vol. 12B, pp. 595601.
Ruiz, M.C., Montes, K.S. & Padilla, R. (2011) Chalcopyrite leaching in sulfatechloride media at ambient
pressure, HydrometaIlurgy, vol. 109, pp. 3742.
Yvenes, L.V., Miki, H. & Nicol, M. (2010) The dissolution of chalcopyrite in chloride solutions Part 2: Effect of
various parameters on the rate, HydrometaIlurgy, vol. 103, pp. 8085.

86

Mineral QTs in place leaching,

standard sector

At El Teniente mine there are abandoned productive sectors containing

mineral resources categorized as broken ore, which could not be extracted
by conventional processes profitably. This paper presents a test in Quebrada

Hugo Letelier, Patricio Gimnez,

Carola Gonzlez, Luis Zenteno and
Carlos Castillo. El Teniente Division,
Codelco, Chile

Teniente mine, Standard Sector xc 31 in which an in place leaching was held

to remove the oxidized copper from the broken ore. Studies realized during
2010 show us that this sector possesses hydric resources and presents good
geomechanical stability. The characterization indicates a copper grade of
1.3%, with 33% oxides. These parameters establish that this mineral is ideal
for the industrial proof. This process conducts a 70 meter drill core through
solid. This drill core is coated with a 84mm hdpe pipe having 4 holes at the
top for watering the mineral, with a vertical pump 6L/s of water acidified with
H2SO4 at 8g/L is injected with a static mixer connected to the pipe. This solution
exits through the top of the drill core, leaching the oxides present in the broken
mineral, once drained the pls is collected through pipes and sent to the SX.
The results shown in this test indicate that the 98% of the pls is captured
and in a 6 month period over 50% of the oxides present in the mineral irrigated
were recovered, reaching a leaching rate of 2.2m3/ton.

87

INTRODUCTION
Since the beginning of its underground operation in 1906, El Teniente deposit has stored a mineral
resource known as broken ore in its crater. It is not feasible, neither technically nor economically, to
mine such resources, nor is it to obtain any benefit from it by using any of the existing conventional
mining methods.
In the last few years El Teniente Division has been trying to develop a number of initiatives and
projects on Technology Innovation and Development, to achieve a greater impact on the
feasibility, sustainability and competitiveness of the mining business. This is the background of a
research effort and a series of tests to validate a technology to conduct in-place leaching of this
broken ore.
This test will take place in a sector named Mina Quebrada Teniente inside El Teniente underground
mine which ceased its operations in the last decade, using one of the crosscuts in order to assess the
following parameters: Solution injection method, establishing ore extraction from the broken ore in
the crater, water balance determination, establishing seasonal effects on the operation of solutions
injection and on the hydraulic balance in the area due to natural runoff.

EXPERIMENTS
Characterization of broken material
The test was conducted in crosscut Nr. 31 of Quebrada Teniente sector. Table 1 shows mineralogy
of such crosscut. This characterization shows that the copper grade is 1.3%, with 33% of oxides.
Table 1 Mineralogical makeup of core drilling in crosscut 31
Mineral

Total Cu
Chalcosine

% Weight

1.3
33.1

Coveline

5.5

Bornite

0.4

Chalcopyrite
Tennantite

Oxides

32.2
0.9

33.0

Leaching solution injection system

A 70-meter long core drilling is conducted in the Production Level; through which a leaching
solution is injected with a concentration of 8 g/L of H2SO4, as shown in Figure 1.
The acidified solution flows through the top of the core drilling, thereby leaching the oxides present
in the broken ore, subsequent to drainage the PLS is collected through pipes and sent to the SX
plant.

88

Broken Material

PLS

PLS

XC 31 Production
Level

PLS

XC 31 Ventilation
Sub-level

Injection Pump
inyeccin

Sand filter

5 m3

Storage Tanks

5 m3
Sulfuric Acid
Tanks

Figure 1 Injection System and Solution Collection System

RESULTS AND DISCUSSION

This broken ore was leached with an acidified solution of 8 g/L of H2SO4 for 6 months, with the
objective of assessing the solution injection method, water balance, irrigation surface and copper to
be extracted from the core drilling.
Figure 2 shows the time percentage in which PLS was observed through the drainage zones during
core drilling. The average impregnation surface was 415 m2 according to this figure, with some
zones in which the surface are reached more than 500 m2, which were observed after shutting down
and re-starting the injection pump.

89

1-25%

25-50%
50-75%
75-100%

Figure 2 Percentage of time in which PLS is observed through drainage zones in the core drilling

Figure 3 shows solution balance, during the March-June period a recovery of 98% of solution was
obtained, the subsequent outflow is greater than the supply, which could be due to water
infiltration coming from winter rainfall.

12

flowrate, l/s

10
8
6
4
2
0
mar

abr

may
Caudal de salida

jun

jul

ago

sep

Caudal de entrada

Figure 3 Incoming and outgoing flowrates

Figure 4 shows the copper concentration variation during drilling period. At the beginning there
were high copper concentrations, with a diminishing trend in time, reaching concentrations of 0.30.4 g/L in the last months of operation.

90

Cu concentration, g/L

4
3,5
3
2,5
2
1,5
1
0,5
0
mar

abr

may

jun

jul

ago

sep

Figure 4 Cu concentration in the PLS

Figure 5 shows the kinetic curve, reaching 2.2 m3/ton during a 6 month operation period.
60

Recovery ,%

50
40
30
20
10
0
,000

,500

1,000
1,500
3
Leaching ratio (m /ton)
Fino Soluble

2,000

2,500

Fino total

Figure 5 kinetic curve for recovery

During a 6-month period soluble copper recovery reached 54%, accounting for 80 ton of fine
copper, with an injection rate of 5,8 L/s, and an irrigation rate of 49 L/h/m2.

CONCLUSIONS
From an operating perspective this is a very straightforward process. Moreover, it affords the
possibility of removing broken material from its current location in the mine.

91

Up to 98% of the injected solution could be captured during a period with no evidence of seasonal
effects. A kinetic curve was build indicating a recovery in excess of 50% of the oxides present in the
irrigated volume, during a 6 month period, injecting a solution of 6 L/s with an acidity of 8 g/L.
When working in a sector with a high percentage of secondary sulphides (33%), there are
significant possibilities for this process if in-place Bioleaching is tested. The same existing facilities
can be used for this purpose.

LITERATURE
Dr. Mara Cristina Ruiz (2007), Hidrometalurgia, Universidad de Concepcin, pp 5-7

92

Study of the steel dust leaching for

the recovery of zinc and cadmium

Steel dust, waste obtained from gas streams of smelting and steel refining
furnaces, consists mainly on zinc oxides, iron oxides and small amounts of
heavy metals such as lead and cadmium. These materials, classified as toxic

Ernesto de la Torre, Alicia Guevara

and Cynthia Espinoza. Escuela
Politcnica Nacional, Ecuador

and hazardous waste, can be evaluated by the extraction of zinc.

The characterization of the steel dust was done by atomic absorption spectrophotometry (aa), X-ray diffraction (xrd) and X-ray fluorescence (xrf).
Contents of Zn 33%; Fe 18%; Pb 2.7%; Cd 300g/ton, like zincite (ZnO) 39% and
magnetite (Fe3O4) 33% were found. Steel dust was determined as hazardous
and toxic, because the levels of lead exceeded the maximum permissible values
established by the toxicological characterization epa 1311 (tclp 1311).
Leaching essays were performed using different acids like: Sulfuric, hydrochloric, nitric, acetic, and tartaric. Also, sodium cyanide, caustic soda,
sodium carbonate and ammonia were used for 24hours at 750rpm. The best
results were obtained with sulfuric acid.
The tests with H2SO4 (150[g/L]; 30% solids) reached recoveries of 75.6% Zn,
after 6 hours of leaching at 750rpm. After 30 minutes of processing, 77.2% Zn
was recovered. Then, the solution was treated in an electrolysis process for 10
hours with a current of 400A/m2 to obtain the zinc cathode. Finally, the product was smelted at 450C to recover metallic zinc of 92.3% purity.

93

INTRODUCTION
Nowadays, the recycling of metals is being developed because it saves energy significantly and
reduces the demand of natural minerals. Metals are 100% recyclable and energy savings by using
recycled metals are above 60%. In the case of zinc, energy savings are 70% (Asher, 2000; Pardav,
2006).
During the melting and refining of steel, which is carried out at temperatures above 1600 C in
electric arc furnaces (EAF), fly ash and solid residues are generated. These residues are carried by
the gas stream and accumulated in the filters of the gas treatment system. These residues are called
steel dust and about 8 million tonnes per year are generated worldwide. From these tonnes, 3.7
million come from electric arc furnaces (Altadill, 2009).
Developed countries have strict environmental legislation, as well as, various industrial processes
for recovering nonferrous metals from steel dust. However, in third world countries, which do not
have industrial treatments for waste or steel dust, this residue is arranged in landfills or dumps,
sometimes in poorly controlled conditions, without knowing the destination of leachates. Indeed,
these residues are in direct contact with the soil and organisms, causing negative environmental
impacts due to the mobility of heavy metals in the percolation water (Delvasto, 2010; Parrn, 2009).
The chemical and mineralogical composition of the steel dust varies, depending on the operating
conditions of each steel plant and the chemical composition of the raw materials used in the
process. The steel powders are usually mainly composed of simple or complex metal oxides, iron
oxides and nonferrous metals, such as zincite (ZnO) and franklinite (ZnFe2O4), with a particle size
below 40 m. The major component of steel dust are iron oxides. Also, significant contents of
heavy metals such as lead, zinc and cadmium can be found. For there reasons, this residue is
classified as hazardous waste (Sinclair, 2005; Doronin, 2010).
The steel powder contains 20-55% Fe, 10-35% Zn, 1-8% Pb and 0.05 to 0.20% of Cd; its recovery,
therefore, is positive to the environment and economically profitable (Doronin, 2010).
In Europe, according to the European Waste Catalogue 10 02 01, steel dust is considered as
hazardous waste, also the U.S. EPA considers it as hazardous Class I Code K061. Pb and Cd are
considered to be the most dangerous, while zinc is the most interesting compound (EPA, 1991).
Zinc has interesting commercial importance. The objective of this research is to value the steel dust
by the recovery of zinc by leaching and electrolysis, reducing the environmental impacts caused by
this waste from steel mills.

Methods used in the treatment of the steel powder

For the recovery of steel powders, industrial reductive fusion processes are used in order to obtain
an iron rich slag. Meanwhile, zinc and lead are volatilized and collected as oxides. The most used
method is the "Waelz System", which is carried out in rotary tubular kilns. Solid waste is loaded
into the furnace together with coke powder and is moved through the kiln in the opposite direction
of the combustion gases.
The oxidation products (zinc oxides and volatile elements like lead and cadmium) are withdrawn
from the furnace with the flue gases and deposited on filters. The final product is crude zinc oxides
with high content of impurities (lead, cadmium, chromium and nickel). For this reason, these

94

material has be purified, to be recycled. For this process, the steel powder should contain more than
15% zinc recovery and the residual slurry should not contain more than 4% of zinc. The
disadvantages are the following: difficulty in regulating the temperature, composition and
unrecoverable losses of iron: the iron oxide content reaches 50% of the mass of slag (Doronin, 2010;
Madias , 2009).
Steel powders are also treated by leaching processes that allow direct recovery of electrolytic zinc.
Sulfuric acid, hydrochloric acid, ammonia or caustic leaching is made to obtain zinc of zincite (ZnO)
and franklinite (ZnFe2O4). Furthermore, leaching can also be performed with organic acids like
acetic, tartaric and citric (Altadill, 2009; Delvasto, 2010).
The recovery of zinc from primary metals like zinc blende [(Zn, Fe) S] is carried out by oxidizing
roasting, which transforms it into zincite (ZnO) and franklinite (ZnFe2O4). Later, they are dissolved
in sulfuric acid according to the reaction 1
ZnO + H2SO4 = ZnSO4 + H2O

(1)

The sulfuric acid used in the leaching usually has a concentration of 100 g/L to 150 g/L, and can
recover about 50 g/L of zinc. Howver, these levels vary from one plant to another (Sinclair, 2005;
Docstoc, 2011).
The leaching solution is purified with an organic solvent and, subsequently, the methalic zinc is
recovered by electrolysis in cells with lead anodes and aluminum cathodes, on which zinc is
deposited (Sinclair, 2005).

METHODOLOGY
Materials
Steel dust from gas streams of cleaning systems, which is generated during the smelting and
refining of steel in electric arc furnaces in the recycling industry of Ecuador, is used.

Methods
Mineralogical and chemical characterization of steel dust
Steelwork dusts are characterized mineralogically by X-ray diffraction (XRD) on a Bruker D8
Advance to determine its components.
The chemical characterization was carried out using atomic absorption (AA) with a AAnalist 300
instrument and X-ray fluorescence (XRF) on a Handheld XRF Analyzer, to determine the amount of
zinc, iron, lead and cadmium of the sample.

95

Influence of the type of bleaching agent in pulp dilution

In order to determine the best leaching agent and evaluate the behavior of waste under acidic
conditions and the presence of organic and inorganic anions, the steel powders were leached with
different solutions of 100 g/L concentration and 1% of solid percentage.
Leaching was carried out for 24 hours and a stirring speed of 750 rpm. The used leaching solutions
were: sulfuric, hydrochloric, nitric, acetic, tartaric acid, also sodium cyanide, caustic soda, sodium
carbonate and ammonia.
Different pulps were filtered to obtain a strong solution and a weak solution (product cake
washing). Then, both solutions were analyzed by atomic absorption (AA) in a Perkin Elmar
AAnalist 300 to determine the concentration of metals (zinc, lead and) in solution. Metallurgical
balances were performed to determine the recovery of metals in the different solutions.

Effect of pulp density and the concentration of the sulfuric acid and acetic acid
Steel powders are leached with different concentrations of H2SO4 and acetic acid (50g/L, 100g/L and
150g/L) and different percentages of solids (1%, 10% and 30%).
The steel powders were leached in stirred bed for 24 hours at a stirring rate of 750 rpm. Then, the
pulps were filtered and analyzed by atomic absorption. Metallurgical balances were performed to
determine the recovery of metals of interest.

Kinetics of sulfuric acid leaching

To evaluate the kinetics of leaching with sulfuric acid, the concentration of the metals of interest in
solution was measured at different times between 0.5 and 24 hours, maintaining the concentration
of H2SO4 constant during the process, at 30% solids. The leaching kinetics were determined to find
the time at which there is a maximum recovery of the metal of interest.

Electrolysis of zinc
The electrolysis process is conducted with a pre-purificated solution, adjusting the pH between 4
and 5.
For these tests, an electrolytic cell with lead anode and aluminum cathode where zinc was
deposited, was used. The current density was 400 [A/m2].

Fusion
The zinc, obtained by the electrolysis, was smelted with ammonium chloride and potassium
chloride. The process was conducted at 450 C for 2 hours.

RESULTS AND DISCUSSION

Chemical and mineralogical characterization of steel dust

96

The steel powder contains 30.0% zinc, 18.7% iron, 2.7% lead and 0.03% cadmium and other minor
elements. Zincite (ZnO) is found as the main component, apart from ferrous oxide, iron oxide and
zinc (franklinite).
Table 1 shows the results of the mineralogical characterization of steel dust.
Tabla 1 Mineralogical characterization of steel dust
Mineral

Frmula

Composicin (%)

Zincita

ZnO

39.0

Magnetita

Fe3O4

33.5

Maghemita

Fe2O3

11.0

Halita

NaCl

8.5

Franklinita

ZnFe2O4

5.5

Grupo plagioclasa
(albita, andesita, anortita)

(Na,Ca)Al(Si, Al)Si2O8

2.0

Hematita

Fe2O3

1.0

Influence of the type of leaching agent in pulp diluted

Figure 1 Zinc recovery in solution (1% solids, during 24 hours and 750 rpm)

Due to the fact that zinc is present in 80.3% as zincite (ZnO), Figures 1 and 2 shows that the zinc
dissolution in acids: sulfuric, hydrochloric, nitric, acetic and tartaric, is more than 80%.

97

Figure 2 Zinc recovery in solution (1% solids, during 24 hours and 750 rpm)

From the leaching with sodium carbonate, ammonia and sodium hydroxide very low recoveries of
zinc were obtained: 3.91% with sodium hydroxide, 0.38% with sodium carbonate and 0.01% with
ammonia.

Figure 3 Zinc recovery in solution with sodium cyanide (1% solids, pH=10.5, during 24 hours and 750 rpm)

Figure 3 shows that when working with sodium cyanide 20 g/L, the recovery of zinc is 84%. The
sodium cyanide concentration is high due to the high concentration of zinc present in the steel
powder.

Effects of pulp density and sulfuric acid and acetic acid concentrations
Tests with higher pulp densities were made with H2SO4 and acetic acid, with concentration of 150
g/L, and 1%, 10% and 30% of solids. The results are shown in Figure 4.

98

Figure 4 Zinc recovery in solution (H2SO4 150g/L, during 24 hours and 750 rpm)

The best results were obtained with H2SO4, recovering zinc as a zinc sulfate solution (ZnSO4).
Besides, other undesirable elements have precipitated as insoluble sulphates. Recovery is greater
with dilute suspensions.

Kinetics of leaching with sulfuric acid

With 30% of solids and H2SO4 150 g/L, maintaining the concentration of H2SO4 constant during the
process, aliquots of the solution were taken during 24 hours to determine the reaction kinetics.

Figure 5 Zinc recovery in solution (H2SO4 150 g/L, during 24 hours and 750 rpm)

The maximum recovery of metals was obtained at the sixth hour: 75.6% Zn, 66% Cd, 16.9%Fe and
0.1% Pb. At the first 30 minutes of process 77.2% of the soluble zinc was recovered. The leaching
solution contains 72.5 g/L of zinc, which can be recovered by the electrolysis technique.

99

Electrolysis of zinc and fusion

The pH of the solution was mantained between 4 and 5 with NaOH. The zinc is deposited on the
cathode and removed each hour during 10 hours of operation. A current efficiency of 70% was
achieved.
Fusion was performed to recover the zinc of the electrolysis, with ammonium chloride used as a
filler flux, at 450C for 2 hours; the metallic zinc obtained has a purity of 92.3%.

Figure 6 Cathodic zinc (left), Metallic zinc (right)

CONCLUSION
The steel powder used in this research was composed of 33.0% Zn, 18.7% Fe, 2.7% Pb and 0.03% Cd,
as zincite (ZnO) and franklinite (ZnFe2O4). For this reason, it is feasible to perform a recovery
process by leaching with acids: sulfuric, hydrochloric, nitric, acetic and tartaric acids.
Sodium cyanide with 20 g/L concentration recovered 84.0% of zinc in solution.
The recoveries of zinc with sodium carbonate, ammonia and sodium hydroxide were very low:
5.01%, 0.40% and 0.01% respectively. That is why it was discarded.
The recoveries obtained with H2SO4 150 g/L and 30% solids were: 75.6% Zn, 66% Cd, 16.9% Fe, 0.1%
Pb, at a leaching period of 6 hours, achieving, at the first 30 minutes,77.2% of soluble zinc.
The recovery of zinc by electrolysis, with 68.5% efficiency, and fusion produced metallic zinc of
92.3% purity.
Hydrometallurgy is an innovative recycling technique, and considering the existing resources in
Ecuador the development of this research is viable and a contribution to the conservation of the
country's mineral resources, plus it is economically profitable.

ACKNOWLEDGEMENTS
The authors kindly thank to the Department of Extractive Metallurgy of the National Polytechnic
School, the project PIS09-04 and the company Adelca, for funding this research,.

100

REFERENCES
Altadill, R., Andrs, A., Bruno, A., Bruno, J, Corts, A. Dez G, Elias, X. Ma, F. & Ripoll, E.(2009) Industrial
waste recycling, Daz de Santos Editorial, Espaa, pp. 11381143.
Aser, S.A. (2000) Recycling metals from scrap and industrial waste, Ingenera Qumica, vol. 32, no. 367,p. 211.
Delvasto, P., Ibaez, B. & Sandoval, D. (2010) Initial characterization of steel dust and evaluating their
leachability in dilute organic acids, IIBEROMET XI, Chile.
Docstoc (2011) Chapter III: Metallurgy of zinc, viewed march 2011, <http://www.docstoc. com/docs/21967194/
Captulo-III-ElCinc-y-su-metalurgia>.
Doronin, I.E. & Svyazhin, A.G. (2010) Commercial Methods of Recycling Dust from Steelmaking, Metallurgist,
vol. 54, no. 10, p. 653.
EPA

(1991) Land disposal restrictions for electric arc furnace dust

<http://www.epa.gov/osw/laws-regs/state/revision/frs/fr95.pdf>.

(K061),

viewed

mayo

2011,

Madias, J.(2009) Recycling of duste of arc furnace, Acero Latinoamericano, vol. 23, no. 513, p.38.
Pardav, W. (2006) Industrial metal recycling: an approach, Ecoe Editions, Bogot, Colombia, pp. 2-5.
Parrn, M. A. & Herrero, A. (2009) Design of industrial model to recover metals in PDH Steelworks, viewed march
2011, <http://www.uca.es/dpto/C120/ficheros/dpto/C120/ ficheros/ ponencias/maparron_2>.
Sinclair, R J. (2005) The Extractive Metallurgy of Zinc, The AusIMM, Melbourne, Australia, pp. 57-59.

101

Zinc recovery from electric-arc

furnace dust by hydrochloric leaching
and bi-electrolyte electrolysis

The steel industry generates a lot of waste which is rich in heavy metals for the
melting of scrap from steel manufacturing. The most waste generated is electric
arc furnace dust (eafd) from the gas exhausting system of the furnace whose

Jos Ricaurte and Ernesto de la

Torre. Escuela Politcnica Nacional,
Ecuador

composition is basically zinc oxides (zincite) and iron oxides (maghemite).

This work shows the results of zinc recovery from eafd with a content of
31% zinc and 17% iron by hydrochloric acid leaching, purification of the leaching solution and electrolysis using a bi-electrolytic cell. Purification of the
leaching solution was carried out by pH adjustment and cementation with
zinc powder. The bi-electrolytic cell has two electrolytes, the leached solution
as catolite and a 1M sulfuric acid solution as anolite, both separated by an
impermeable cation-exchange membrane.
The parameters that were varied included; for leaching, HCl concentration
between 50 and 150g/L and solids between 10 and 30%. For electrolysis, current
density varied between 725 and 2000A/m2.
Leaching with 150g/L of HCl and 30% of solids the zinc recovery reached
71% in 2 hours of operation. The process of purification reached eliminations
close to 99% for Fe, Cd and Pb by increasing pH up to 4.5 and cementation
with zinc dust. Electrolysis with a current density of 725A/m2 has a current
efficiency of 53%. The advantage of hydrochloric acid leaching over sulfuric
acid leaching is the reduced time needed to obtain high zinc recoveries and
the possibility of reusing the acid solution obtained after electrolysis for the
same process or other leaching processes. The need to improve the current
efficiency of electrolysis processes and to find uses for the solution obtained
after electrolysis warrant the necessity of further research.

103

INTRODUCTION
The steel industry plays an important role in the world development because steel is one of the
most used materials in construction for its ductility and resistance. Its worldwide demand in 2012
was 1441 million tons, up over last year's demand spurred by economic recovery (Gerdau, 2011).
As with any industrial process, the development of a product generates byproducts and steel
industry is not the exception. Between 12 and 14 kg of waste is generated in form of powder per ton
of steel produced from smelting of scrap in electric arc furnace (Da Silva & Brehm, 2006). These
wastes are called electric arc furnace dust (EAFD) and are classified by EPA as a hazardous waste
because of its high content of heavy metals such as zinc, lead, cadmium and chromium. This type of
waste is named as K061 (EPA, 1991).
To extract zinc from EAFD, hydrochloric acid was used as leaching agent for its selectivity towards
zinc (Nez & Vials, 1984). Generally, hydrochloric acid is not used as a leaching agent in
industrial scale processes because, subsequently, zinc dissolved in solution must be recovered in
metallic form by electrolysis and contact between anode and hydrochloric acid forms chlorine gas.
To avoid the formation of chlorine gas in this study, we worked with a bi-electrolytic cell to
separate the anode and hydrochloric acid.
The presence of metals such as iron, cadmium and lead in the solution resulting from the leaching
of EAFD, presents a big problem in electrolysis, lowering zinc recovery and contaminating the
deposited zinc. Iron presents the greatest problems because, on one hand, it is deposited and
contaminates the zinc deposit and, on the other hand, the presence of this metal decreases the
electrical efficiency of the cell because the ferrous ion (Fe+2) is continuously oxidized at the anode
and the ferric ion (Fe+3) is unceasingly reduced at the cathode (Sinclair, 2005).
Purification processes are used to precipitate the undesirable metals in the leached solutions by the
alteration of their solubility varying pH. The ferric ion (Fe3+) precipitated with pH between 4 and 5,
unlike the ferrous ion (Fe2+) which remains stable in this range of pH. To precipitate ferrous ion, it
must be oxidized to ferric ion and this can be achieved with aeration (Baik & Fray, 2000).
The pH of the leached solution is acid in the order of 0.5 to 3.0. pH of the solution can be increased
by adding a base such as sodium hydroxide or even steel powder, due to its buffering capability, is
between 12 and 13. To remove contaminating metals such as cadmium and lead, zinc powder is
used in a process called cementation. The process consists on putting the leaching solution in
contact with zinc powder so that, by the action of the difference of standard electrode potentials
between zinc and the metals to be removed, the latter are deposited in the zinc powder.
The next step for zinc recovery is the electrolysis but if an electrolyte solution containing HCl is put
in contact with the anode, the acid get decomposed over the anode surface forming chlorine gas
which is toxic and corrosive. The problem is avoided separating the leached solution from the
anode by a cation-exchange membrane (Baik & Fray, 2000). To close the electrical circuit and obtain
hydrogen ions (H1+), a strong acid like sulfuric acid is put in contact with the anode, becoming itself
a hydrogen ions font. Thereby, hydrogen ions generated at the anode (sulfuric acid) travel through
the cation-exchanged membrane to the cationic side (leached solution) and reduce zinc ions (Zn2+)
to metallic zinc (Zn0), depositing it over the cathode surface.

104

METHODOLOGY
Physical, chemical and mineralogical characterization
The electric arc furnace dust (EAFD) was chemically characterized by atomic absorption (AA)
spectrophotometry in a Perkin Elmer AA analysis 300 instrument to determine the Zn, Fe, Cd and
Pb concentrations. The mineralogical characterization was carried out using X-ray diffraction (XRD)
on a Bruker D8 Advance and EVA-TOPAS programs.

Kinetics of leaching, purification of leached solution, electrolysis of zinc

The kinetics of the leaching reaction were determined by varying the reaction time (0,5; 1,0; 1,5; 2,0;
3,0; 4,0; 6,0 y 8,0 hours), the concentration of hydrochloric acid (50, 100 y 150 g/L) and the
percentage of solids (10% y 30%). Stirring was constant at 750 rpm. Also, the HCl concentration was
constant over the time of reaction to prevent zinc recovery to be dependent on the amount of
available acid. The recovery of zinc and iron were determined at each time, analyzing the leached
solution by atomic absorption spectrophotometry.
Air was bubbled into the solution to oxidize the ferrous ion (Fe2+) to ferric ion (Fe3+). To raise the pH
of the leached solution from pH = 0.5 to pH between 4 and 5, EAFD and sodium hydroxide were
added in proportions of 64 g/L and 108 g/L, respectively, to precipitate iron. Thereafter zinc powder
was added (14.8 g/L) to remove cadmium and lead.
Figure 1 schematizes the bi-electrolytic cell used for the electrolysis of zinc. Current densities were
tested in a range between 725 to 2000 A/m2. It used an air pump with diffuser to oxygenate the
cathodic solution, allowing the reduction of zinc and preventing the dendritic growth of the
deposit. The deposits obtained at each current density were analyzed using a scanning electron
microscope (SEM) with X-ray microprobe to determine the composition of the deposits and to
determine the recovery of zinc.

Figure 1 Bi-electrolytic cell diagram

105

RESULTS AND DISCUSSION

Physical, chemical and mineralogical characterization
Tables1, 2 and 3 show the physical, chemical and mineralogical characteristics of electric arc furnace
dust sample. Zinc is present in form of zincite (ZnO) and iron mainly in form of maghemite (Fe2O3).
Table 1 Physical characterization
real (g/ml)
aparent (g/ml)
d80 (um)

2.62
0.92
150.00

Table 2 Chemical characterization

Zn (%)

Fe (%)

Cd (%)

Pb (%)

31,47

17.32

1.48

0.02

Table 3 Mineralogical characterization

Mineral

Formule

Contents (%)

Zincite

ZnO

40

Maghemite

Fe2O3

33

Hematite

Fe2O3

Periclase

MgO

Magnetite

Fe3O4

Quartz

SiO2

Plagioclase

(Na,Ca)Al(Si,Al)Si2O8

Kinetics of leaching
Figure 2 shows zinc recoveries of about 70%, obtained in the first hour with the three tested
concentrations. It was also noticed that the trend of recovery obtained with the three concentrations
of acid were very similar, indicating that a 10% solids and concentration of acid in a range of 50 to
150 g/L is not a factor that defines the zinc recovery.
According to the Figure 3, iron recovery was clearly affected by acid concentration. Significant
recoveries of iron are obtained because, at 10% solids, there is an excess of the amount of acid in
respect to the amount of zinc in the dust and, therefore, that amount of acid reacts with iron.

106

100

Recovery (%)

80
60
150 g/L

40

100 g/L
20
50 g/L
0
0

4
Time(h)

Figure 2 Zinc recovery with HCl leaching (10% solids)

100
150 g/L

Recovery(%)

80

100 g/L
50 g/L

60
40
20
0
0

4
Time(h)

Figure 3 Iron recovery with HCl leaching (10% solids)

In contrast to tests with 10% solids, with 30% of solids, zinc recovery is affected by acid
concentration, being much smaller with 50 g/L of HCl as it can be seen in Figure 4. The greatest
recovery was obtained at 2 hours of operation, leaching with 150 g/L of HCl, with a recovery of
70.9% of zinc. As shown in Figure 5, iron is recovered only at concentrations of 150 g/L of HCl. At
concentrations of 100 g/L of HCl or less, no iron recovery is obtained within the 8 hours of leaching
of the study. It means there is not sufficient acid at concentrations of 100 g/L not even at 50 g/L to
leach iron at 30% solids. These results differ from the findings of recovery of iron at 10% solids. It is
also observed, that when the leaching is done at 150 g/L of HCl, the iron recovery is higher for 10%
solids than for 30% solids during the time range studied, because at 10% solids there is more
amount of acid to react with the dust.

107

100

150 g/L
100 g/L
50 g/L

Recovery (%)

80
60
40
20
0
0

Time (h)

Figure 4 Zinc recovery with HCl leaching (30% solids)

100
150 g/L

Recovery (5)

80

100 g/L
50 g/L

60
40
20
0
0

Time (h)
Figure 5 Iron recovery with HCl leaching (30% solids)

Purification of leached solution

After raising the pH of the leached solution to precipitate iron by adding sodium hydroxide (108
g/L) and EAFD (64 g/L) and, subsequently, cementing cadmium and lead with zinc powder (14.8
g/L), 99.9% Fe, 99.1% Cd and 99.5% Pb were removed. Iron was removed as a precipitate after
raising the pH to 4.5 during 1 hour. The solution was filtrated and then sodium hydroxide and
EAFD was added. The solution was stirred for 15 minutes and then filtrated to obtain a purified
solution.

108

Electrolysis of zinc
Table 4 Electrolysis results
Current density
(A/m2)
725
1150
2000

Operation
time (h)
5.1
4.0
2.2

Deposit
(g)
5.9
5.3
5.1

Deposit composition (%)

Zn
O
66.1
12.3
52.1
22.5
28.4
28.4

Cl
5.8
12.1
18.6

Na
5.7
5.1
10.1

Figure 6 Deposits recovered. (A) 2000 A/m2; (B) 1150 A/m2; (C) 725 A/m2

According to Table 4, working with a low current density gives a deposit with fewer impurities and
higher zinc content. The presence of oxygen is due to dissolved oxygen by aeration in the
purification process. The presence of chlorine is due to cathode solution containing chlorine from
the HCl used for leaching. The sodium comes from the caustic soda used to raise the pH of the
solution in the purification process. The physical difference of deposits regarding the content of
impurities is seen in Figure 6.
Table 5 Power consumption and current efficiency
Current
density
(A/m2)
725

Voltage
(V)

Current
(A)

Energy
kWh/kg Zn

1.1

Current
efficiency
(%)
52.7

4.3

1150

5.1

1.8

39.7

14.0

2000

6.1

3.2

21.1

30.3

6.8

As shown in Table 5, the energy consumption per kilogram of recovered zinc working with 725
A/m2 is less than the consumption obtained working at greater current densities, even though, the
operation time is higher. This is because zinc recovering is greater at this current density.

109

CONCLUSIONS
The steel powder sample used in this study contains 31.4% Zn, 17.3% Fe, 1.4%Cd and 0.02% Pb. The
zinc is present in form of zincite (ZnO) and iron mainly in form of maghemite (Fe2O3).
In two hours of leaching with HCl, working at 30.0% solids, the zinc recoveries reached 70.9%,
52.9% and 32.6% at 150, 100 and 50 g/L HCl, respectively. With 10% solids, zinc was recovered in
86.2%, 77.8% and 75.8% at same concentrations of HCl.
The process of purification reached eliminations close to 99% of Fe, Cd and Pb by increasing pH
until 4.5 and cementation with zinc dust.
Conducting electrolysis process with a bi-electrolytic cell avoided chlorine gas formation and
allowed electrodepositing of zinc at the cathode surface. Furthermore, the leached solution was
clarified reaching a pH of 0.5and can be reused in the same process or in other processes of
leaching.
Working at a current density of 725 A/m2 a deposit composed of 66.1% Zn, 12.3% O, 5.8% Cl and
5.7% Na was obtained. Electrical efficiency at this current density is 52.7% and the energy
consumption is 6.8 kWh/kg of recovered Zn.
All these facts make this process an auspicious technique that should be evaluated at a pilot plant
level to determine its worthiness and in a future, scale it up to the industrial level.

REFERENCES
Baik, D. & Fray, D. (2000) Recovery of zinc from electric-arc furnace dust by leaching with aqueous
hydrochloric acid, plating of zinc and regeneration of electrolyte, Mineral Processing and Extractive
Metallurgy, vol. 109, no. 3, pp.121-128, viewed January 2012, <
http://www.ingentaconnect.com/content/maney/mpem/2000/00000109/00000003/art00002>.
Da Silva, J. & Brehm, F. (2006) Characterization Study of Electric Arc Furnace Dust Phases, Materials Research,
vol. 9, no. 1, pp.41, viewed January 2012,
<http://www.lume.ufrgs.br/bitstream/handle/10183/20331/000592671.pdf?sequence= 1>.
Environmental Protection Agency (1991) Land disposal restrictions for electric arc furnace dust (K061),
Environmental Protection Agency, United States of America, viewed January 2012,
<http://www.epa.gov/osw/laws-regs/state/revision/frs/fr95.pdf>.
Gerdau, A. (2011) Demanda mundial de acero tocar record el 2012, viewed January 2012,
<http://www.gerdau.cl/demanda-mundial-de-acero-tocara-record-el-2012>.
Madas, J. (2009) Reciclado de polvos de horno elctrico, Avances tecnolgicos en la industria siderrgica,
Reprints de artculos de Jorge Madas, Revista Acero Latinoamericano 2008-2012, pp. 26-35, viewed January
2012, <http://www.construccionenacero.com/Boletin%20Tecnico/2_Q_2012/Reprints_Tecnolog%C3%
ADa.pdf>.
Nez C. & Vials J. (1984) Kinetics of leaching of zinc ferrite in aqueous hydrochloric acid solutions, Metall.
Trans., vol. 15B, no. 6, pp.221.
Sinclair, R. (2005) The extractive metallurgy of zinc, Spectrum Series, vol. 13, pp. 93-97.

110

A step closer to the leaching of

primary sulfides

The majority of copper sulfide reserves are found in the form of primary
sulfides, which may be processed by flotation followed by smelting/refining.

Guillermo Velarde. Sociedad Minera

Cerro Verde S.A .A ., Peru

This metallurgical route is almost generally the preferred method to process

this type of ore. However, certain hydrometallurgical process can be less
expensive and potentially more amenable with the environment. Despite the
huge amount of research carried out over several decades on the leaching of
primary sulfides, to date, a viable technique has not been found to adequately
process marginal or below cut-off grades ores for concentrators. Currently,
copper is recovered with relative success from low-grade ores by bioleaching;
however, this depends on the pyrite content of the ore. The Cerro Verde mine
is well aware of the importance of temperature on the metal recovery process
and thus is focusing efforts to develop technology to raise heap temperatures
to well above normal levels using solar energy as a primary source of heat.
This approach can also be supplemented with the heat produced by the biooxidation of pyrite. With the aid of solar energy, it will be possible to treat
minerals with low levels of pyrite and at the same time, obtain a higher and
much more uniform temperature over the leaching cycle. This paper present
several variables in terms of their effect on viability, including the quantitative,
linear influence of temperature on copper recovery, resulting in an increase
of around 2% for each Celsius degree. Pilot work has demostrated that heating
of the heap can be increased by 16C in 40 days, using a solar-heated solution
at an irrigation flow rate of 17l/h/m2.

111

INTRODUCTION
It is widely accepted that temperature greatly affects chemical reactions and the kinetics of other
physical-chemical processes. It increases the velocity and/or levels of efficiency of many processes
and it especially affects copper leaching. According to our experiments, the kinetics and copper
recovery of primary mineral sulphides are much more sensitive to temperature than secondary
sulphides. A higher temperature influences the process chemistry; mass transfer and both the
quantity and type of the bacterial population and activity (Velasquez et al., 2012 & Demergasso et
al., 2012).
Likewise, it is well-known that the leaching of primary sulphide minerals is an extremely slow
process with very low rates of recovery at atmospheric temperature and pressure. Therefore, until
now, leaching is only feasible with low-grade minerals with no size reduction (ROM), or only after
the ore has been processed through primary crushing in a marginal operation, such is the case of
the Toquepala Dump Leach Operation in southern Peru. If the pyrite content of the mineral is high,
this process is more likely to be successful (e.g. at Toquepala, the pyrite content of the low grade
primary sulphide is in the order of 5 to 8%). The heat, generated mainly by the bio-oxidation of
pyrite, raises the temperature of the heap or stockpile, which in turn increases copper recovery and
produces sulphuric acid as well.
Thus, the pyrite content of the mineral has a direct impact on the kinetics of copper recovery.
Nevertheless, it is difficult to control the kinetics of oxidation of the pyrite sufficiently to reach and
maintain the desired temperatures for long enough to significantly impact recovery. Therefore, by
itself, this process is not efficient enough to be used in the leaching of primary sulphides (Velarde,
2013 & Demergasso et al., 2012).
This paper summarises experiments carried out into heap leaching, solar heating and the irrigation
of a leaching heap. The relationship of the recovery of sulphides of copper to temperature is
investigated. An example shows how the temperature of the heap can be increased by almost 20
Celsius degrees using solar heating. Solar heating of leaching solutions has shown being more
effective than the practice of covering heaps with transparent films which only provides a delta in
temperature between 3 and 6 Celsius degrees depending on climatic conditions (Servicios
Asociados, 1999, & Velarde, 2002)
The higher process temperature, obtained with the aid of solar heating, improves pyrite oxidation
kinetics. In turn, the heat generated from oxidation helps off-set deficiencies in solar heating during
cloudy days and low ambient temperature moments of the day or the year. Therefore, the
combination of solar heating and pyrite oxidation can make the leaching of primary sulphides more
viable than ever.

THE EFFECT OF TEMPERATURE ON THE KINETICS OF COPPER RECOVERY:

RESULTS AND DISCUSSION
Secondary sulphides
Figure # 1 shows copper recovery over time curves for secondary sulphide leach column tests
performed at four different temperatures at the same grain size gradation corresponding to a P80 of

112

10 mm. The directly proportional relationship between temperature and copper recovery can be
clearly seen after the first 37 days of leaching. Likewise, Figure # 2 shows that this relationship is
linear over the leaching period. It can also be seen that the higher the temperatures, the lower the
delta copper recoveries are. Furthermore, Figure # 3 shows how delta copper recovery drops with
leach time due to increased temperature (35 C compared to 20 C). The mineral sample tested had
1.04% of total copper, 0.21% acid soluble copper, 0.55% cyanide soluble copper, and a leaching
index or solubility ratio of 0.731. This plot can be used for scale-up purposes, since normally
ambient temperature column leach tests are easier to run than heated column tests.

Figure 1 The effect of temperature on the recovery of secondary sulphide copper minerals

Figure 2 The linear effect on recovery

Figure 3 The fall in delta of recovery achieved

by leaching at 35 C compared to 20 C

Primary sulphides
Figure # 4 shows the recovery curves obtained during leaching tests in columns with low-grade
primary copper sulphide with a grain size at a P80 of 10 mm. The columns were run at an
approximate ambient temperature of 37 C. The ore characteristics were: 0.231% total copper,

113

0.011% acid soluble copper, 0.011% cyanide soluble copper soluble, and a leaching or solubility
index of 0.095; the pyrite content was 1.5% and the proportion of chalcopyrite compared to the total
copper minerals present in the sample was 97.5%.
After 850 leaching days, the higher process temperature resulted in a recovery of 60%, compared to
20% for the ambient temperature process. While the curves continued to be divergent, with the
higher temperature process continuing its upward trend, these results were very different from
those obtained with secondary sulphides, where the impact of temperature is lower and decreasing
over time.
Figure # 5, shows that the relationship of recovery to temperature remained linear after 220 days of
leaching at around 60 C. In this case, two points were obtained from the above mentioned tests run
at Cerro Verde, while the rest of the time-recovery points of the curve were obtained from tests run
at the Phelps Dodge Mining Technology Center (Uhrie, 2004). The increase in delta recovery due to
temperature is shown on Figure # 6.

Figure 4 The effect of temperature on the recovery of primary sulphide copper minerals

Figure 5 Linear effect of temperature over

recovery after 220 days of leaching

114

Figure 6 Delta recovery increase over time

HEATING OF LEACHING HEAPS

State of the art
Some sulphide leach operations, especially those located at high altitude, cover the heap with a
transparent plastic film to achieve temperature gains ranging from 3 C to 6 C, depending on
ambient conditions. In this case the operational strategy is more oriented to reduce the heap cooling
rather than increase the temperature inside the heap.
Other operations, such as Quebrada Blanca in Chile, take advantage of the residual heat of their
electricity generators to warm up the air being blown into the heaps.
There is also a patented process to heat heaps, using Fresnel lenses or parabolic mirrors, to heat a
thermal fluid to 150 C, which in turn heats the leach solutions, some of which can even be stored to
be used for night-time irrigation (Hernandez, J. O & Casas de Prada, J. M., 2012).
The high temperatures achieved inside leach dumps of low copper grade primary ores are basically
the result of heat generated by the oxidation of their relatively high pyrite contents. Nevertheless,
this process is difficult to control, since normally at the beginning, the mass of pyrite quickly reacts
and oxidizes producing large quantities of heat, which are lost to the environment mainly by
convection; then, the temperature inside the heap drops at a time when there are copper quantities
yet to be extracted.
Project concept
The project involves the placing of pipes in a serpentine formation, covered with transparent plastic
sheets, in a permanent installation outside the heap. The system is used to capture solar energy and
heat the leach solutions, which flow through the pipes, just like solar-water heaters. Then, the
heated solution is conveyed to the drip irrigation lines installed at the surface of the heap, which is
also covered with transparent plastic sheets to avoid evaporation, reduce heat loss and allow solar
radiation to act over the heap surface. One of the main objectives was to use an economical material
for the pipes, which would result in a simple, cheap and effective system.
Initially, the pilot plant only supplied hot leach solution during sunny periods, 8 hours per day.
Later on, during the second stage of the tests, storage tanks were used to allow irrigation of the
heap over non-sunny periods. The irrigation flow rate during the heating stage was high, 17 l/h/m2,
to compensate for the lack of irrigation at night and to heat the heap as quickly as possible. This
also meant that the equilibrium temperature could be reached sooner.
The irrigation regime should follow the kinetics of the leaching of copper-sulphides process, which
is initially high and later begins to decrease; that is, there should be a high initial flow rate to meet
the preliminary heating requirements of the heap, followed by a reduced flow rate to continue
leaching and maintain the equilibrium temperature of the heap. Nevertheless, every operation is
unique; for example, in the case of a permanent pad with interconnected layers, the lower layers,
which work at ambient temperature, have to be heated to be able to achieve delta copper recoveries,
especially if high-refractory minerals are present.

115

PILOT PLANT TEST

Description
After promising, initial small-scale testing of irrigation with heated solutions, a higher-scale
definitive pilot test was carried out, the main characteristics and results are shown in Table 1.
Table 1 Main characteristics of pilot test
Magnitude

Item
Heated area

750 m2

Test mineral

5,100 mt

Total length of heating pipe

25,000 m

Irrigation area

441 m2

Flow rate of heated solution

125 l/min (33 gpm)

Irrigation flow rate

17 l/h/m2

Evaluated depth of heap

7.0 m

Average solution temperature during the sunny

period at the heating system outlet in Summer.

60 C

Average solution temperature during the sunny

period at the heating system outlet in Winter.

46 C

The heating of a serpentine pipe

Tests prior to the pilot test allowed the determination of the optimum length of serpentine for a
specific solution flow rate, to be determined. Figure # 7 shows the curves of temperature increase
with length of serpentine at different times of a sunny day. Figure # 8 shows the curve of average
temperature achieved versus length of pipe over said period.
80
70

70
60

60

Temperature (C)

Temperature (C)

50

50
40
30
20

40
30
20

10
10

0
0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

170

Pipe Length (m)

8:00 a 9:00

7:35 a 8:00

11:00 a 12:00

12:00 a 13:00

15:00 a 16:10

10

20

30

40

50

Adjusted curve

60

70

80

90 100 110 120 130 140 150 160 170 180

Real Data

Pipe Length of heating (m)

Figure 7 Temperature increase over the length of

Figure 8 Average temperature increase over

heating pipe for a given solution flow at

the length of heating pipe for a given

different times of day

116

13:00 a 14:00

solution flow rate on a sunny day

Heat loss
The heat loss from the surface of the heap to the environment during a sun-free period, as expected,
falls as depth increases, reaching zero at 50 cm. The external perimeter of the leached section also
loses heat towards the surrounding mineral, which greatly affects its temperature. However, the
test demonstrated that it had almost no effect on the centre of the heap (at 10.5 meters). These heap
boundary losses are, as could be expected, relatively unimportant for the system as a whole,
provided the heated leaching area is large. Figure # 9 shows the curves of temperature at different
depths, determined at around 4:00pm one day, and at 8:30am on the following day:
50
45
45

43
41

40

39

Temperature (C)

40

35
35

34

38

37

36

37

1.09

0.00

50

60

32

30

30
28

25
16.47

12.39

9.00

3.56

5.59

Temperature Delta

20
5

10

20

Temp. Morning following day

30

40

Thermometer Depth (cm)

Temp. Afternoon day 1

Figure 9 Cooling of heap during sun-free period

Thermal balance

45

2400000

40

2100000

35

1800000

30

1500000

25

1200000

20

900000

15

600000

10

300000

Heat (Kcal/day)

Temperature (C)

The final objective of the test was to obtain the thermal balance of the heap being heated. The
results obtained are shown on Figure # 10. The curve of speed of heating is adjusted to fit the real,
actual data. The curve of losses to the ambient is based on calculations made from one day to the
next, over the sun-free period. The other curves were also calculated using experimental data of
balanced system temperatures.

0
0

10

15

20

25

Temperature reached
Net Heat required to achieve a daily increase of temperature
Total heat loss non-sunny period from 0 to 60 cm

30

35

40

45

50

Leaching Time (days)

Gross heat applied to system

Heat not used to increase the temperature of the heap

Figure 10 Thermal balance of heap

117

Economics of project
The total unit cost of heating the solution to 53 C at the heating system outlet to achieve a delta of
32 C, and taking into account winter/summer average temperature and the average temperature of
a sunny day, was 0.195 $/m3 with an additional cost to cover the surface of the heap of 1.37 $/m2.

CONCLUSIONS
The oxidation of chalcopyrite, the most refractory of the leaching minerals, is heavily influenced by
the process temperature. Thus, primary sulphides are much more sensitive to temperature than
secondary sulphides. Therefore, additional heating combined with the heat provided by the
oxidation of pyrite, could improve the leaching of primary sulphides in the dump or heap
The success of leaching at above normal temperatures using solar energy, depends on the
application of new irrigation concepts, the level of solar radiation, the climate, and the
characteristics of the mineral to be processed.
The heating system tested increased the temperature of the heap by an average of up to 37 C,
with a temperature delta of 16 C during 40 summer days. The average depth of the heap was 7 m.
The total solar-heated irrigation system unit cost was 0.195 US$/m3 of heated solution plus 1.37
US$/m2 of heap area subjected to irrigation.

ACKNOWLEDGEMENTS
The author wishes to thank Javier E. Guevara, Manager of Hydrometallurgical Processes of
Sociedad Minera Cerro Verde, for supporting this project and its presentation at Hydroprocess
2013.

REFERENCES
Uhrie, L. John, (2004) Effect of Temperature and Bacterial Population on Pyrite and Chalcopyrite Leaching,
Process Technology Center of Phelps Dodge Mining Company, Arizona, USA.
Hernandez, J. O & Casas de Prada, J. M. (2012) Process and System for Solar Leaching Cells Using Solar
Concentrating Collectors Based on Fresnel Lens with Solar tracking.
http://ip.com/patfam/es/46639000#desc.
Demergasso, C., Iturriaga, V., Galleguillos, P., Acosta, M., & Soto, P.. Espoz Ch., Velasquez A. & Pinilla C.
(2012) A Shift from Mesophilic to Moderate Thermophilic Populations in an Industrial Operation of a
Heap Bioleaching Process, Minera Escondida, Universidad Catlica del Norte and CICITEM, paper
published at Hydroprocess 2012, Chile.
Velasquez, A., Sutherland, F., Espoz Ch. & Pinilla C. (2012) Evolution of the Microbial Consortium through
Leaching Cycles in Minera Escondida, paper published at Hydroprocess 2012, Chile.
Servicios Asociados (1999) Aumento de temperatura en las pilas de lixiviacin y el efecto del film Therm-2,
Chile.

118

Controlling the irrigation flow in

heap leach piles of El Soldado
mine by canopy system with
thermal camera

This work describes the development of a mechatronic system which is capable

of developing a top/down view over on/off heap leach piles of the El Soldado
mine, in order to recognize and detect (through a thermal analysis) where the
piles may become locally saturated from leach solution irrigation. Such situations can trigger the potential of liquefaction and should be considered, con-

Javier Ruiz-del-Solar, Omar Daud

and Mauricio Correa. Advanced
Mining Technology Center,
Universidad de Chile
Marco Torres. El Soldado Division,
Anglo American Chile

trolled, and avoided. In order to reduce the possibility of liquefaction, and to

achieve full stability in the pile, the irrigation system must be properly controlled.
For instance, in potentially saturated zones the irrigation flow can be reduced
or eliminated until the saturation has disappeared.
This mechatronic system consists of a canopy system, equipped by a thermal imaging camera which is mounted below its structure, pointing down to
the ground. A canopy system consists of a pulley system attached between
two opposite ends of a heap pile. The mounted thermal infrared camera is
used to obtain the temperature information of the heat transfer from the pile
to the environment. Such information is very useful, as the level of saturated
zones can be traced.
The communication between the operator of the irrigation system and the
canopy system is based on a radio-frequency link. The operator is in charge
of the Programmable Logic Controller, while the information from the canopy
system is transmitted by geo-referenced images.

119

INTRODUCTION
Heap leaching is a mineral process in which crushed or run-of-mine rock are leached with chemical
solutions in order to extract valuable minerals. Heap leaching technology is largely applied to
copper-containing minerals in which these are irrigated with a weak sulphuric acid solution. This
solution dissolves the copper from the mineral and the pregnant leach solution (PLS) passes
down through the ore pile and is recovered at the bottom. The bottom of the leach pile is usually
constructed with a geo-membrane liner, and a permeable crushed rock drainage system along with
a drainage pipe network. Copper is extracted from the PLS using electro-winning processes and the
acidic solution is recycled back onto the leach pile. Leach pads can be divided into four categories:
conventional or flat pads, dump leach pads, valley fills and on/off pads (Thiel & Smith, 2004).
Conventional leach pads are relatively flat, and the ore is stacked in relatively thin lifts (5 to 15 m
typically). Dump leach systems are similar or can include rolling terrain; lifts can be much thicker
(up to 50m). Valley fill systems are leach pads properly designed in natural valleys. On/off pads
(also known as dynamic heaps) are hybrid systems. A relatively flat pad is built using a robust liner
and over liner system (Thiel & Smith, 2004). Then a single lift of ore, usually from 2 to 8 meters
thick, is loaded and leached. At the end of the leach cycle the spent ore is removed for disposal and
the pad recharged with fresh ore. Usually the piles are loaded automatically, using conveyors and
stackers; in some cases bucket wheel excavators and conveyors are used in the unloading cycle, in
others loaders and trucks are used. For on/off pads the critical design loads for the containment
system come from the ore handling equipment, not the weight of the ore (Thiel & Smith, 2004).
The critical aspects of heap leach piles can be summarized as the depth of the ore, and for on/off
pads the equipment loads, presence of water and local terrain (Thiel & Smith, 2004). All these
variables can dramatically affect stability. Since the piles may become locally saturated from leach
solution irrigation, the potential for liquefaction also exists and is often considered; moreover, if the
pile is located in a seismically active zone. Liquefaction typically occurs when saturated, loose
granular material contracts, or collapses, under some triggering event. A classic triggering event is
seismic shaking. Seismically-induced liquefaction is typically limited to approximately 20 m in
depth, as the confining pressures at greater depths reduce susceptibility to this type of failure.
Liquefaction flow slide failures have also been known to occur under static conditions; these are of
particular interest because of their catastrophic consequences.
Modern heap leaching practices represent an expanding opportunity to develop new technologies
dealing with new operating systems in order to improve the overall process. Therefore, this paper
addresses the problem of the irregular irrigation in heap leach piles, in order to control irrigation
and to avoid liquefaction, and herein the correlated collapses, along with their catastrophic
consequences. Throughout an easy and simple canopy system in which is installed a thermal
camera, thermography maps can be obtained, and through image processing, these thermal
information can be used for control irrigation.

METHODOLOGY
The main scope of this approach is to control the irrigation of the pile, in order to avoid saturation
zones, which comprise the leaching process and stabilize the pile. In order to address the early

120

detection of the saturation zones, it is necessary to survey and monitoring the pile, especially
during irrigation. However, with this system, it is also intended to recognize the zones that are not
irrigated, in order to improve the leaching process as well. This monitoring action was conducted
by a thermal camera attached to a canopy system (see Figure 1). The idea is to detect the saturation
zones and not irrigated zones by thermal analysis, which can be conducted by image enhancement
and processing. Therefore, as the saturation zones are detected, suitable control actions in the
irrigation system can be performed. In such case, in those zones in which there is an excess of
solution, the corresponding valves should be closed.

Figure 1 Setup of the system

The heap leach pile under analysis corresponds to the one of the Cancha 4 from El Soldado Mine.
The dimensions of the pile are L 48.4 x W 24.4 x H 2.5 m. The distance between two rows of
irrigation line is three meters. The thermal images were acquired using a FLIR 320 TAU Thermal
Camera, 9mm-f1.25, with sensitivity in the range 7.5 -13.5 m, and a resolution of 324x256 pixels.

121

Figure 2 Example of the thermal image. Deep blue zones are the saturated areas, while white zones
corresponds to dry areas inside the heap leach pile

Image analysis
In order to solve the problem explained above, three classes are defined using the different zones of
the irrigation inside the heap leach pile: saturated zone, irrigated zone and dry zone. Given the
images that were obtained from the thermal camera (see Figure 2), two sets of images are
generated. The first set is for training, while the second set is used for evaluation. From those
images, corresponding to the training set, several crops for each class were selected. It was used the
value of each pixel as the data (Gonzalez & Woods, 1992). With these data, three histograms were
calculated (see Figure 3). Three Gaussian models are estimated; given the data distribution in the
histograms (see Figure 4).

300
200
100
0

50

100

150

200

250

50

100

150

200

250

50

100

150

200

250

4000

2000

0
2000

1000

Figure 3 Histogram of each class

122

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2

Class1
Class2
Class3

0.1
0

50

100

150

200

250

300

Figure 4 Gaussian models estimated given the distribution of the data in the histograms

Naive Bayes is a robust classifier based upon the principle of Maximum A Posteriori (MAP), and,
for this reason, Bayes is used in order to determine the class of each pixel (Rish, 2001). Given a
problem with K classes { , . . , } with so-called prior probabilities
,,
(Zhang, 2004), it
can be assigned the class label, c, to an unknown example with features =
,,
such that:
=

,,

which chooses the class with the maximum a posterior probability given the observed data (Mardia
& Kanji, 1994). This a posterior probability can be formulated, using Bayes theorem, as follows:
=

,,

,,
,,

As the denominator is the same for all classes, it can be dropped from the comparison. Now, it
should be computed the so-called class conditional probabilities of the features given the available
classes. This can be quite difficult taking into account the dependencies between features. The naive
bayes approach assume class conditional independence i.e. , ,
which are independent given
the class. This simplifies the numerator to be:
=

and then choosing the class c that maximizes this value over all the classes = 1, , . Clearly this
approach is naturally extensible to the case of having more than two classes, and was shown to
perform well inspire of the underlying simplifying assumption of conditional independence. In this
case, we have three classes that are independent. It is supposed that any class c have the save
probability, then
=1 =
=2 =
= 3 = . Again the denominator is the same for all
classes; therefore, it can be dropped from the comparison. Finally, for each pixel the class is given
by =
, = .

123

RESULTS AND DISCUSSION

Results
Based on the image analysis, explained above, the thermal images were segmented as shown in
Figure 5. These images where segmented under three classes in a grayscale (0 to 255):

Black (0): for saturated zone.

Gray (127): for optimal zone.
White (255): for dry zone.

Discussion
Based on the results shown above, it is clear that the segmentation of the thermal images was
successfully conducted, giving solid information about the classified zones. Through image
processing, it is possible to detect specifically where the saturated or the dry zones are, as well as
the zones that are irrigated in a proper way.
Indeed, such information could be used in order to recognize specifically the zones in which the
pile is saturated. Moreover, as the irrigation system follows an irrigation line, it can be easily
understood what line can be controlled automatically, as this image processing system can fed with
this kind of information the PLC that controls irrigation. Herein, if a saturation zone is detected
though thermography, the PLC is fed with this information in order to compute a correcting action,
i.e. the corresponding valves should be closed, until the saturation is gone. Taking into account the
opposite case, in which some zones of the pile are dry, i.e. the irrigation control system must send a
higher flow and/or to more open the valve in order to increase the flux.

a1

124

a2

b1

b2

c1

c2

d1

d2

Figure 5 Some examples of the segmentations. a1, b1, c1, d1: the thermal images. a2, b2, c2, d2: the segmented
images

As the excess of humidity can seriously put in danger the stability in the heap leach pile, raising
exponentially the risk of collapse because of liquefaction, such monitoring system helps
maintaining the stability of the pile. This leads to an increase of safety in the leaching process, as no
human operator needs to be in the immediate surroundings, where these collapses can damage or
put in danger the life of the operator.

125

CONCLUSIONS
The thermal images are efficient for controlling irrigation in heap leach piles as they give valuable
information about the zones, if they are correctly being irrigated.
As this algorithm permits the classification in order to successfully separate and segment different
zones of the pile under irrigation, the loop can be closed as the irrigation system can be controlled
automatically.
Such irrigation control can give the chance to monitor the pile in order to avoid collapse because of
liquefaction.
As this approach is non-invasive, it is not necessary that an operator goes and check inside the heap
leach pile this information that is taken from the thermal camera with the canopy system. This kind
of approach dramatically reduces the risk for the operator, as he/she is neither exposed directly to
the solution nor above the pile.

ACKNOWLEDGEMENTS
The authors would like to thank the miners of the El Soldado Mine for their assistance during the
surveys and the engineers of the AMTC for their precious support during the execution of the
experimental evaluations.

REFERENCES
Thiel, R. and Smith, M. (2004) State of the practice review of heap leach pad design issues, Geotextiles and
Geomembranes, vol. 22, no. 6, pp. 555568.
Gonzalez, R. C. and Woods, R (1992) Digital Image Processing, 2nd edn, Addison-Wesley Longman Publishing
Co., Inc., Boston, MA, USA.
Rish, I. (2001) An empirical study of the Naive Bayes classifier, IJCAI Workshop on Empirical Methods in
Artificial Intelligence, pp. 41-46.
Zhang, H. (2004) The Optimality of Naive Bayes, Proceedings of the Seventeenth International Florida Artificial
Intelligence Research Society Conference (FLAIRS 2004), pp. 562-567.
Mardia, K. V. and Kanji, G. K. (1994) Statistics and Images: 2. Advances in Applied Statistics, Carfax Publishing Co.
Ltd., Abingdon, Oxfordshire.

126

Hydrothermal mineral replacement

reactions and their applications in
mining and processing

Over the past decade, many mineral replacement reactions in nature have
been recognized to occur via the coupled dissolution re-precipitation (cdr)
mechanism, i.e. they proceed via dissolution of the parent phase into a fluid,
coupled with the precipitation of the product, that might occur at the reaction
front, or some distance away from the reaction front. The product is porous,

Jing Zhao, Allan Pring, Jol Brugger,

Fang Xia and Kan Li. University of
Adelaide and South Australian
Museum, Australia
Yung Ngothai. University of
Adelaide, Australia

enabling reaction to run to completion. Recent research has dramatically

improved our understanding of the mechanisms and kinetics of cdr reactions,
with applications to understanding mineral formation processes in nature,
but also in materials synthesis and it is the process central to mineral leaching.
We review existing studies relevant to mineral processing, including Fe-Ni-S,
Au-Ag-Te, and Cu-Fe-S (+U) ores, and explore the potential of the new concepts
for in situ mining and leaching. (i)Pentlandite [(Fe,Ni)9S8] and violarite
[(Ni,Fe)3S4] are the main Ni ores in massive sulfide deposits; the nature of
these phases is important for flotation and smelting. Using crd reactions,
pentlandite can be quickly transformed to violarite under mild hydrothermal
conditions. (ii)Cdr processes enable the transform of refractory Au-Ag-tellurides into gold in ~24hours under hydrothermal conditions (200C), providing
an alternative to roasting. (iii)Studies of the reactions among hematite, chalcopyrite and bornite provide new insights into heap leaching of chalcopyrite
and bornite, and predict the distribution of uranium in iocg deposits.

127

INTRODUCTION
Aqueous fluids in the crust are the most important medium for hydrothermal mineral replacement
reactions over a wide range of hydrothermal conditions: fluids with varying amounts of salts (up
to > 50 wt%) and volatiles (e.g., CO4, CH4, H2S), at pressure up to many kbar and temperatures up to
600 C or higher. The understanding of dissolution, transport, and deposition of metals in aqueous
hydrothermal fluids is central to understanding hydrothermal ore deposits genesis. In the
replacement process, the existing set of minerals reacts with the ubiquitous aqueous fluids in the
crust and results in the formation of a different minerals assemblage. The nature of fluid-mediated
replacement reactions can vary over different length and time scales. In principle they are interface
coupled dissolution-reprecipitation reactions (ICDR reactions), where the dissolution of parent
mineral and the product precipitation are coupled both in space and time, allowing the external
dimension and in some cases even the crystallographic orientation to be persevered (Putnis & Putnis,
2007; Xia et al., 2009). But in the more general case the dissolution takes place at the reaction interface
and the products, or part of the product can be deposited some distance away from the reaction
front. In both cases, some elements may remain in solution; such situations are the essence of the
leaching process. In leaching of, for example Cu ores, Cu does not diffuse out of the solid directly
into the solution. At temperatures below 200 C the rate of solid-state diffusion in ore minerals, even
in copper sulphides, is negligible. Neither is it an ion exchange processes, with the Cu ions being
pulled from the solid into the solution as ions from solution are sucked in to maintain charge
balance like an electro-chemical cell; ore minerals are not good ionic conductors even at high
temperatures. What actually happens in heap leaching is that the mineral is dissolved into solution;
this dissolution increases the concentrations for some ions in the solution near the surface of the
dissolving grain, causing the solution to become supersaturated with another mineral, which will
then precipitate. In the case of the dissolution of chalcopyrite, the precipitated phase might be
goethite (FeOOH) or a jarosite-group mineral.
Our recent research into the nature of CDR reactions on Fe-Ni-S, Au-Ag-Te, and Cu-Fe-S (+U) ores
has led us to believe that these reactions can be employed in mineral processing for in-situ mining
and leaching. Here, we review the mineral replacement reaction of pentlandite ((Fe,Ni)9S8) by
violarite ((NiFe)3S4), the transformation of gold silver tellurides to Au-Ag alloys and the formation
reactions of copper iron sulphides from iron oxides, and emphasize how the unique properties of
CDR reaction mechanism can be used in the mining and processing industry.

The mechanism of coupled dissolution reprecipitation (CDR) reactions

Mineral replacement reactions in nature have been recognized to occur via the coupled dissolution
re-precipitation (CDR) mechanism. In contrast to solid-state reactions, which are controlled by
diffusion, CDR reactions proceed via the dissolution of the starting material into a fluid. In
interfaced couple dissolution re-precipitation (ICDR) reactions, this is followed by the immediate

128

precipitation of the product mineral at the reaction front (Figure 1) (Brugger et al., 2010; Putnis,
2009; Putnis & Putnis, 2007). The dissolution front and the site of precipitation are coupled in
variable length scale ranging from ~1 nm to up to 10s of m or more. This length scale is controlled
by the interplay between the kinetics of dissolution and precipitation, and these are in turn
controlled by the relative solubilities of the parent and product phases. Therefore, the overall rate
depends on the reaction pressure, temperature, solution chemistry (e.g. pH, redox, ligand
concentrations) and the kinetics of nucleation and growth of the crystals. ICDR reactions usually
lead to pseudomorphic replacements. The external morphology and, in case where the first and
second phases have a close crystallographic relationship, even the crystallographic orientation of
the parent phase can be preserved by the product. In the more general case when the dissolution
and the precipitation steps are not tightly spatially coupled over the duration of the reaction, the
local chemical gradients near the reaction front, which is different from that of the bulk solution,
will play a role in controlling the rates of dissolution and precipitation. In all CDR reactions the
product is porous, enabling mass and fluid transport between the reaction front and the bulk
solution until the completion of the reaction. In the cases where the product is not porous, or the
porosity decays during the reaction, then the parent grain becomes armored and the reaction stalls.
Studies on model reaction replacement systems the (e.g., formation of KBr from KCl) show that the
porosity can heal, or be annealed out over time, even at room-temperature.

Figure 1 Schematic diagram of an interface coupled dissolution-reprecipitation reaction

Recent work on mineral replacement reactions in aqueous fluids has dramatically improved our
understanding of the mechanisms and kinetics of these reactions. At a more fundamental level,
understanding CDR processes improves our understanding on the interplay between the physical
chemistry of mineral dissolution and crystallization, solution chemistry, and fluid transport along
grain boundaries and through the porosity within minerals. The mechanism could help us to
understand the formation of hydrothermal ore deposits, develop methodologies for porosity
materials synthesis and improve the mining and processing industries.

129

IRON NICKEL SULFIDES

Pentlandite ((Fe,Ni)9S8) and violarite ((Ni,Fe)3S4) are primary nickel sources in massive sulphide ore
deposits. Violarite replaces primary nickel sulphide minerals such as pentlandite in the supergene
alteration zones of nickel sulphide deposits during the natural weathering and oxidation processes
(Misra & Fleet, 1974). The formation of violarite from pentlandite was believed to occur in the solid
state, and hence to be driven by metal diffusion (Thornber, 1975). Later studies however indicate
that the replacement reaction is actually a CDR reaction (Tenailleau et al., 2006; Xia et al., 2007;
2009).
Xia et al. (2009) studied the transformation process of pentlandite to violarite under mild
hydrothermal conditions using both polytetrafluoroethylene-lined stainless steel autoclaves and in
situ using stainless steel flow-through hydrothermal cells. Reaction kinetics shows a complex
behaviour depending on various physical and chemical parameters, but the mechanism had a first
order dependence on solution chemistry. The reaction rate increases with temperature from 80 to
125C, and then decreases as temperature further increases to 210C. The effects of pH shows that
the reaction only occurs in the pH 1 to 6 range, with maximum reaction extent at pH 4. Above pH 2
the reaction is an ICDR reaction, which is pseudomorphic. The rate of the reaction is controlled by
the rate of dissolution of the pentlandite, i.e., violarite precipitation is faster than pentlandite
dissolution (Figure 2a). At lower pH the rate of dissolution of pentlandite is faster than the rate of
precipitation of violarite, so the interface coupling is broken and hollow casts of violarite are
formed (Figure 2b).

Figure 2 Backscattered electron images of partially reacted grains at 125 C and (a) pH~3.9, (b) pH~1

These studies on the transformation of pentlandite to violarite provide us with a fundamental

understanding of reaction mechanism and the thermodynamics and kinetics of the formation of
violarite in nature. This can be used to predict the synthesis conditions of nickel iron sulphides

130

under hydrothermal conditions (Xia et al., 2008). If it is possible to transform violarite back to
pentlandite, the hydrothermal method could be used as a pre-treatment to eliminate fine-grained
and relatively porous supergene violarite, which has a poor response in many systems to ores
separation by floatation. The method would improve the nickel recovery rate in sulphides
processing industry.

GOLD SILVER TELLURIDES

Gold-bearing tellurides, such as sylvanite, krennerite, calaverite, and petzite, are important sources
of gold in some deposits around the world (e.g., Shackleton & Spry, 2003), but are the most
refractory gold ores in extraction process. Tellurides are normally pretreated by roasting, chemical
oxidation, high pressure acid leaching, or biooxidation (Kongolo & Mwema, 1998; Climo, Watling
& Van Bronswijk, 2000; Spry, Chryssoulis & Ryan, 2004). Zhao and coworkers (Zhao et al., 2009;
2010; 2013a) found that Te could be effectively leached from Au-Ag tellurides over a wide range of
pH (212), oxidant nature and concentration, and temperature (140 to 220C) under hydrothermal
conditions. The pseudomorphic transformation proceeds via an ICDR mechanism, with the
dissolution of gold silver tellurides being rate-limiting relative to Au-Ag alloy precipitation. The
Au-Ag alloy precipitates as a spongy mass of fine filaments similar in microstructure to mustard
gold. Tellurium is lost to the bulk solution as Te(IV) complexes, and precipitates away from the
dissolution site (e.g. autoclave walls) as TeO2(s). The reaction increases linearly with the surface
area of the starting minerals. For example, at 200C, solid-state replacement of calaverite by gold is
very slow (months), but calaverite grains (~100 m in size) are fully replaced within 48 hours under
hydrothermal conditions. Results also show that basic buffer solutions favour the replacement
reaction, due to the high solubility of Te(IV) in basic solutions. Studies on other gold silver
tellurides, including sylvanite (Zhao et al., 2013a), krennerite (Xu et al., 2013) and petzite all show
the loss of tellurium to solution and the precipitation of Au-Ag alloy. Thus, the ICDR of gold silver
tellurides provides an efficient and less toxic alternative pretreatment for the industrial treatment of
gold-bearing telluride minerals.

COPPER IRON SULFIDE (+U)

Iron oxide-copper-gold (IOCG) deposits, such as Olympic Dam and Prominent Hill in South
Australia, are important sources of copper, gold and uranium, and have received considerable
attention from researchers in the last few decades. Studies of the reactions among hematite,
chalcopyrite and bornite provide new insights into heap leaching of chalcopyrite and bornite. Based
on the studies of iron or iron nickel sulphides (Xia et al., 2007; 2008; 2009; Qian et al., 2010; 2011;
2013), chalcopyrite has been successfully synthesized by the replacement of hematite (Zhao et al.,
2013b) under hydrothermal conditions, and the newly formed chalcopyrite could be further
replaced by bornite. The replacement of natural chalcopyrite by bornite was carried out under

131

hydrothermal conditions (Zhao et al., 2013c). The reaction kinetics shows that high temperature (up
to 300 C), high pressure and basic buffer solution (pH 10) favour the reaction rate. This study shed
some light on the formation of IOCG deposits and may lead to improved mineral exploration,
mining, and processing of complex IOCG ores. At the same time, uranium has been successfully
incorporated in sulphide minerals via ICDR reactions using a flow-through hydrothermal reactor,
which allows for the first time a direct insight into the processes that led to the formation of the
chalcopyrite-bornite-uraninite association. Further experiments using synchrotron technique would
help us to predict the distribution of uranium in IOCG deposits. Ultimately, this work will lead to
improved methods for prospecting and for processing complex U ores.

CONCLUSIONS
CDR reactions are often extremely complex, since they result from a dynamic interplay between
dissolution and precipitation, controlled critically by phenomena such as surface nucleation,
porosity creation/destruction, and transport of aqueous species to and from the reaction front
through an evolving porous medium. Understanding the molecular level mechanisms at work in
these reactions is the key to controlling the textures and compositions of the products, and requires
the concerted efforts of physical chemists and mineralogists, but may have many benefits in terms
of mining and processing industry.

ACKNOWLEDGEMENTS
This work is funded by the Australian Research Council Project (DP0878903; DP0772299;
DP1095069) and the South Australian Museum. We also thank the staff at Adelaide Microscopy for
their help with SEM imaging.

REFERENCE
Brugger, J., McFadden, A., Lenehan, C.E., Etschmann, B., Xia, F., Zhao, J. & Pring, A. (2010) A Novel Route for
the Synthesis of Mesoporous and Low-Thermal Stability Materials by Coupled
Dissolution-Reprecipitation Reactions: Mimicking Hydrothermal Mineral Formation, CHIMIA
International Journal for Chemistry, vol. 64, pp. 693-698.
Climo, M., Watling, H.R. & Van Bronswijk, W. (2000) Biooxidation as pre-treatment for a telluride-rich
refractory gold concentrate, Minerals Engineering, vol. 13, no. 12, pp. 1219-1229.
Kongolo, K. & Mwema, M.D. (1998) The extractive metallurgy of gold, Hyperfine Interactions, vol. 111, pp.
281-289.
Misra, K.C. & Fleet, M.E. (1974) Chemical composition and stability of violarite, Economic Geology, vol. 69, pp.
391-403.

132

Putnis, A. (2009) Mineral Replacement Reactions, Reviews in Mineralogy and Geochemistry, vol. 70, no. 1, pp.
87-124.
Putnis, A. & Putnis, C.V. (2007) The mechanism of reequilibration of solids in the presence of a fluid phase,
Journal of Solid State Chemistry, vol. 180, pp. 1783-1786.
Qian, G., Brugger, J., Skinner, W.M., Chen, G. & Pring, A. (2010) An experimental study of the mechanism of
the replacement of magnetite by pyrite up to 300 C, Geochimica et Cosmochimica Acta, vol. 74, no. 19,
pp. 5610-5630.
Qian, G., Xia, F., Brugger, J., Skinner, W.M., Bei, J., Chen, G. & Pring, A. (2011) Replacement of pyrrhotite by
pyrite and marcasite under hydrothermal conditions up to 220 C: An experimental study of reaction
textures and mechanisms, American Mineralogist, vol. 96, no. 11-12, pp. 1878-1893.
Qian, G., Brugger, J., Testemale, D., Skinner, W. & Pring, A. (2013) Formation of As(II)-pyrite during
experimental replacement of magnetite under hydrothermal conditions, Geochimica et Cosmochimica
Acta, vol. 100, pp. 1-10.
Shackleton, J.M. & Spry, P.G. (2003) Telluride mineralogy of the Golden Mile deposit, Kalgoorlie, Western
Australia, Canadian Mineralogist, vol. 41, pp. 1503-1524.
Spry, P.G., Chryssoulis, S. & Ryan, C.G. (2004) Process mineralogy of gold: Gold from telluride-bearing ores,
Journal of the Minerals, Metals and Materials Society, vol. 56, pp. 60-62.
Tenailleau, C., Pring, A., Etschmann, B., Brugger, J., Grguric, B. & Putnis, A. (2006) Transformation of
pentlandite to violarite under mild hydrothermal conditions, American Mineralogist, vol. 91, no. 4, pp.
706-709.
Thornber, M.R. (1975) Supergene alteration of sulphides, I: A chemical model based on massive nickel
sulphide deposits at Kambalda, Western Australia, Chemical Geology, vol. 15, pp. 1-14.
Xia, F., Zhou, J., Pring, A., Ngothai, Y., ONeill, B., Brugger, J., Chen, G. & Colby, C. (2007) The role of
pyrrhotite (Fe7S8) and the sample texture in the hydrothermal transformation of pentlandite
((Fe,Ni)9S8) to violarite ((Ni,Fe)3S4), Reaction Kinetics and Catalysis Letters, vol. 92, no. 2, pp. 257-266.
Xia, F., Zhou, J., Brugger, J., Ngothai, Y., ONeill, B., Chen, G. & Pring, A. (2008) Novel Route To Synthesize
Complex Metal Sulphides: Hydrothermal Coupled Dissolution-Reprecipitation Replacement
Reactions, Chemistry of Materials, vol. 20, no. 8, pp. 2809-2817.
Xia, F., Brugger, J., Chen, G., Ngothai, Y., O'Neill, B., Putnis, A. & Pring, A. (2009) Mechanism and kinetics of
pseudomorphic mineral replacement reactions: A case study of the replacement of pentlandite by
violarite, Geochimica et Cosmochimica Acta, vol. 73, no. 7, pp. 1945-1969.
Xu, W., Zhao, J., Brugger, J., Chen, G. & Pring, A. (2013) Mechanism of mineral transformations in krennerite,
Au3AgTe8, under hydrothermal conditions, American Mineralogist (Accepted).

133

Zhao, J., Brugger, J., Grundler, P.V., Xia, F., Chen, G. & Pring, A. (2009) Mechanism and kinetics of a mineral
transformation under hydrothermal conditions: Calaverite to metallic gold, American Mineralogist, vol.
94, no. 11-12, pp. 1541-1555.
Zhao, J., Xia, F., Pring, A., Brugger, J., Grundler, P.V. & Chen, G. (2010) A novel pre-treatment of calaverite by
hydrothermal mineral replacement reactions, Minerals Engineering, vol. 23, no. 5, pp. 451-453.
Zhao, J., Brugger, J., Xia, F., Ngothai, Y , Chen, G. & Pring, A. (2013a) Dissolution-reprecipitation versus
solid-state diffusion: mechanism of mineral transformations in sylvanite, (AuAg)2Te4, under
hydrothermal conditions, American Mineralogist, vol. 98, pp. 19-32.
Zhao, J., Brugger, J., Ngothai, Y. & Pring, A. (2013b) The formation of chalcopyrite and bornite under
hydrothermal conditions: an experimental approach (Submitted to American Mineralogist).
Zhao, J., Brugger, J., Ngothai, Y. & Pring, A. (2013c) Mechanism and kinetics of a mineral transformation
under hydrothermal conditions: chalcopyrite to bornite (draft).

134

Study of the recovery of silica from

the zircon chemical treatment

Zirconium silicate (ZrSiO4), commonly known as zircon, is the most abundant

mineral of zirconium and the main source for the production of metallic zirconium, as well as its compounds and alloys. The present work aims at studying
the recovery of silica present in the product of the fusion of zircon with sodium

Laura Rocha and Carlos Morais.

Centro de Desenvolvimento da
Tecnologia Nuclear, Comisso
Nacional de Energia Nuclear (CDTN/
CNEN), Brazil

hydroxide. During the alkaline fusion, the zirconium silicate (ZrSiO4) is converted into sodium zirconate and sodium silicate. The study of silicon recovery
was performed by leaching the silicate from the product of the fusion with
water (alkaline leaching) followed by the precipitation of silica and by leaching
the fusion product with nitric acid (acid leaching) for the dissolution of zirconium. After the leaching process, the solid-liquid separation was carried out
through vacuum filtration, the residue was dried and the chemical analysis of
the residue and the liquor was performed. From an alkaline molten product
of zircon containing 31.8% ZrO2, 14.1% SiO2 and 47.8% Na2O and 5.7% humidity,
in the alkaline leaching, a dissolution of approximately 73% of silica, a liquor
with about 15g/L of silica and a residue with 89% of ZrO2 and 6.7% of SiO2 was
obtained. In the acid leaching carried out with HNO3, the dissolution of ZrO2
was approximately 92%, obtaining a nitric liquor containing 20g/L of Zr (27g/L
ZrO2). The content of SiO2 in the acid leaching residue was 92% with a mass
reduction of approximately 80%.

135

INTRODUCTION
Zircon is a tetragonal zirconium orthosilicate (ZrSiO4). The chemical composition of the mineral
mass contains approximately 65% of ZrO2, 32% of SiO2, and other elements at low concentrations,
such as Hf, Fe and Ti (Habash, 1997; Manhque et al, 2003; Bertolno et al, 2008). By presenting high
thermal and chemical stability, the decomposition of zircon, requires drastic conditions. Different
techniques are used in the decomposition of the mineral and may occur by thermal or chemical
decomposition at high temperatures. The alkaline fusion with sodium hydroxide is a method for
chemical decomposition of zircon and is the most used process in obtaining metallic zirconium.
This fusion occurs at temperatures between 550-700 C. The variables that affect the characteristics
of the product of the fusion are: mass ratio, temperature and reaction time (Abdel-Rehim, 2005;
Zolfonoun et al 2010).
The decomposition of zircon generates large ammounts of sodium silicate, along with zirconium
oxide, according to Equation 1 (Abdel-Rehim, 2005).

ZrSiO 4 + 4 NaOH Na 2 ZrO 3 + Na 2SiO 3 + H 2 O

(1)

According to Bswas et al, 2010, considering the reaction presented in Equation 1, it is possible to
form other products, besides zirconate and sodium silicate, such as Na4SiO4, Na2Si2O6, Na6Si2O7,
Na4ZrSi3O12.
The recovery of the silica present in zircon is fundamental as the environmental and economic
issues are concerned, since silica is an inorganic material widely used in a variety of applications
such as resins, molecular sieves, catalyst carriers, as raw material in the glass, and electronic
industry, amongst other applications. This removal is generally accomplished by alkaline leaching
using water as a leaching agent. In this step there is a very high consumption of water and a highly
pollutant effluent having a high content of silica is produced. An alternative method evaluated in
this work was the leaching of sodium zirconate, keeping the silica insoluble. This new route can
reduce the cost of the process with satisfactory results. The process variables investigated were:
reaction time, temperature, percentage of solids and nitric acid concentration (HNO3).

METHODOLOGY
Reagents and solutions
The sample of zircon concentrate used in this work was provided by the Nuclear Industries of
Brazil - INB. Prior to this study, the decomposition of the zircon concentrate by alkaline fusion with
sodium hydroxide was performed. In the study of the separation of Zr/Si; 100 kg of zircon
concentrate was melted, generating approximately 200 kg of molten mass (alkaline melt product).
The sample was homogenized and reduced to fractions of different masses in order to enable the
laboratory experiments and the chemical characterization.

136

The sodium hydroxide used was of commercial grade (99% w/w) and the nitric acid (HNO3) was of
analytical grade. Distilled water was used for alkaline leaching and also in the preparation of the
solutions.

Methodology
The leaching studies were performed in a 250 mL beaker using mechanical agitation at controlled
temperature. After the reaction time stated, the pulp was vacuum filtered using buchner funnil and
kitasato. After filtration, the solid was transferred for drying in an oven at 110C. Then the residue
was weighted and the volume of the filtrate was measured and sent to chemical analysis. The
metals present in the residue and in the leaching liquor were analyzed by Energy Dispersive X-ray
Fluorescence Spectrometry - EDXRF - using an X-ray spectrometer of dispersive energy brand
SHIMADZU, model EDX-720. The nitrate content in the liquor was determined by molecular
absorption spectrophotometry in the ultraviolet region (302 nm) using an UV visible
spectrometer, brand VARIAN model Cary 50 conc.

RESULTS AND DISCUSSION

Chemical characterization
The chemical composition of the main components of the zircon concentrate and of the molten mass
(fusion product) of zircon used in the experiments of recovery of silica is shown in Table 1, below.
Table 1 Chemical composition of zircon before and after the alkaline fusion.

Sample
Zircon
Fusion
product

Species / Content (%)

ZrO2

SiO2

HfO2

Na2O

H2 O

Er2O2

Y2O3

Al2O3

TiO2

Fe2O3

BaO

65.0

29.0

1.33

0.58

0.18

2.22

0.94

0.53

0.58

5.70

31.8

14.1

0.57

47.8

Alkaline leaching
The chemical opening of zircon performed through fusion with sodium hydroxide produces a
strongly alkaline molten mass. This allows to use water as the leaching agent for the the dissolution
of the sodium silicate formed during melting. In the study of silica dissolution with water (alkaline
leaching) the following process parameters were investigated: percentage of solids, temperature
and reaction time.

Reaction time
The reaction time was investigated ranging from 1 to 6 hours, within a percentage of solids of 10%
and at room temperature. The results showed that there was only a slight variation in the
percentage of the silicate dissolution 73 to 76%. Considering the small diference in the silicate

137

dissolution obtained in this step of the study, the reaction may be conducted with 1 hour of stirring
at room temperature.

Variation of the solids percentage and the temperature

The influence of the solids percentage and temperature on the sodium silicate dissolution was
investigated for percentages of 10%, 15%, 20% and 30% at temperatures of 25C and 70C with a
reaction time of two hours. The results are presented below in Table 2.
Table 2 Effect of the percentage of solids in the dissolution of sodium silicate generated in the
alkaline fusion of zircon
Percentage of solids
10
10
15
15
20
20
30
30

Temperature (oC)
25
70
25
70
25
70
25
70

Dissolution (%)
73
76
75
73
72
72
71
72

As can be seen in Table 2, the influence of the percentage of solids and the temperature on the
dissolution of sodium silicate was not significant. However, one can observe a slight tendency
towards higher dissolution in solids percentages between 10 and 15%. In the intervals investigated,
the dissolution of the silicate ranged from 71 to 76%, considering an experimental error of about 4%,
it can be said that the dissolution of the silicate present in the fusion product is not influenced by
temperature.

Final products of the alkaline leaching

At the end of the experiments presented in this paper, an alkaline liquor containing approximately
15 g/L of Si and a residue containing 89% of ZrO2 and 6.7% of SiO2 was obtained. The silicon
present in the liquor was precipitated through acidification with hydrochloric acid 5.0 mol/L up to
pH 1.0. The chemical composition of the alkaline solution, the residue of leaching and silica
obtained in the neutralization liquor is given in Table 3.
The mass reduction in the alkaline leaching was approximately 45%. The dissolution of the residue
of the alkaline leaching with HNO3 generated a liquor containing around 50g/L Zr, 2.5 g/L Si, 3.5
g/L Na, 1.0 g/L Hf and 0.2 g/L Fe. The yield of the dissolution of Zr and Hf from the residue of the
alkaline leaching was 97%.
Table 3 Chemical characterization of the products of the alkaline leaching
Sample
Alkaline Liquor (g/L)
Residue of leaching (%)
Silica (%)
*NaOH = 86.6 g/L

138

SiO2
31.1
6.73
95.3

ZrO2
0.09
89.,4
2.30

Species / Content (%)

HfO2
Na2O
Fe2O3
_
0.002
67.1*
0.96
0.98
1.14
0.09
0.57
0.04

TiO2
_

0.54
_

CaO
0.08
0.26
0.20

Acid leaching
As shown in Table 3, the residue of the alkaline leaching had a SiO2 content of 6.7%, which is not
desirable in the process. Therefore, another process route was investigated; the dissolution of Zr
without previously removing the silicon. The dissolution of Zr was made by leaching with HNO3 in
a beaker using mechanical stirring at controlled temperature. The process variables investigated
were: acid concentration, temperature, reaction time and percentage of solids.

Variation of the concentration of HNO3 and temperature

Zr recovery (%)

The influence of the concentration of HNO3 in the dissolution of zirconium from the fusion product
was investigated at the range of 4 mol/L to 12 mol/L at three different temperatures: 25C, 50C and
70C. The experiments were performed in triplicates and the results are shown in Figure 1. As seen
in Figure 1, there is an increased dissolution of Zr as the temperature increases. The influence of
temperature is more significant for lower concentrations of HNO3 (4.0 mol/L). The dissolution of
silica was below 2% for all the concentrations of HNO3 investigated. The subsequent experiments
were carried out in acidity of 12 mol/L. The high acidity of HNO3 is justified by the fact that in the
next stage of the study, which comprises the separation of Zr/Hf by solvent extraction, is performed
at high acidity. The behavior of Hf was similar to that of Zr and therefore was not reported in
Figure 1.

100

25 degree Celcius

95

50 degree Celcius

90

70 degree Celcius

85
80
75
70
4

10

12

Nitric acid conc. (mol/L)

Figure 1 Influence of temperature and HNO3 concentration in the dissolution of zirconium.

139

Reaction time
The reaction time was varied from 0.5 to 4 h, using a HNO3 concentration of 12.0 mol/L (ratio
acid/water of 3:1) at a temperature of 70C and using a percentage of solids of 10%. The
experiments indicated a progressive increase in the dissolution of the zirconium with an increase of
time, however negligible. For the reaction time of 30 min. (0.5 h), the dissolution of Zr was
approximately 89% and 95% for the time of 4 hours. The results are shown in Table 4 below. The
subsequent tests were performed at a reaction time of 3 h. The dissolution of Si was lower than 2%
in all experiments.
Table 4 Effect of reaction time on the dissolution of SiO2 and Zr
Time (h)
0.5
1.0
1.5
2.0
2.5
3.0
4.0

Dissolution of Zr (%)
88.7
91.5
92.0
92.7
94.0
92.9
94.9

Dissolution of Hf (%)
85.6
89.4
91.2
92.7
92.0
91.6
93.5

Purity of Si (%)
84.2
85.5
86.5
86.7
88.7
87.0
89.3

Percentage of solids
The influence of the percentage of solids, in the dissolution of Zr, was investigated at intervals
between 5% and 30%. The experiments were conducted with 3 hours stirring at 70C and 2.0 mol/L
HNO3. The results are shown in Table 5. As can be seen there is a reduction in the dissolution of Zr
as well as the purity of the silica when the solids content is higher than 15%. Under the
experimental conditions, the dissolution of Si was below 2%.
Table 5 Influence of the percentage of solids in the dissolution of Zr and the in the purity of the
silica.
Percentage of solids (%)
5
10
15
20
25
30

Dissolution of Zr (%)
94.7
93.4
91.7
90.4
90.3
86.2

Dissolution of Hf (%)
93.3
91.0
89.8
87.0
85.8
82.9

Purity of Si (%)
91.3
90.0
87.0
85.8
85.0
81.2

Final products of the acid leach

The best results obtained in the zirconium leaching without the prior removal of silica were: (i) a Zr
liquor containing about 20g/L of Zr, 1.2 g/L Si, 0.4 g/L Hf and 20 g/L Na; the Fe content in the liquor
was below 0.05 g/L, (ii) a residue containing about 43% Si (92% SiO2), 6% Zr and 0.2% Hf, (iii) the Zr
recovery was of approximately 95% and the mass reduction in the leaching was around 80%.

140

CONCLUSION
The previous dissolution of the sodium silicate by alkaline leaching before the dissolution of the
zirconium, is already being used industrially, however the drawback of this route is the high water
consumption, as water plays an important role in the sustainability of the environment. The route
of zirconium leaching without the prior leaching of sodium silicate has the advantage of
demanding lower water consumption besides saving one step in the purification of zirconium.
However, the quality of the product obtained by acid leaching was lower when compared with the
product obtained from the alkaline leaching. The liquor obtained fron the alkaline leaching residue
was richer in Zr when compared to that obtained from direct acid leaching. For a more noble
application, the silica obtained from acid leaching needs to go through a purification step.

ACKNOWLEDGEMENTS
The Authors are grateful to FAPEMIG, FINEP, CNPq and CAPES for the financial support and also
thankful to the CDTNs technicians involved in this study, mainly Luiz C. da Silva, Walmir M.
Garcia and Liliane P. Tavares.

REFERENCES
Abdel-rehim, A.M. (2005). A new technique for extracting zirconium from Egyptian zircon concentrate.,
International Journal of Mineral Processing, Elsevier, pp. 234243.
Bertolino, L.C.; Palermo, N.; Sampaio, J.A.; Frana, S.C.A. (2009) Zirconita, Rochas & Minerais Industriais: Usos e
especificaes, Cetem/MCT, pp. 917930.
Biswas, R.K.; Habib, M.A.; Karmakar, A.K.; Islam, M.R. (2010) A novel method for processing of Bangladeshi
zircon: Part I: Baking and fusion with NaOH., Hydrometallurgy, Elsevier, vol. 103, pp. 124129.
Habashi, F. (1997) Handbook of Extractive Metallurgy, Wiley-VCH, vol. 3, pp. 16491684.
Manique, A.; Kwela, S.; Focke, W.W. (2003) The wet process for the beneficiation of zircon: optimization of
alkali fusion step., Industrial & Engineering Chemistry Research., American Chemical Society, vol. 42,
pp. 777783.
Zolfonoun, E.; Monji, A. B.; Taghizadeh, M.; Ahmadi, S. J. (2010) Selective and direct sorption of zirconium
from acidic leach liquor of zircon concentrate by rice bran., Minerals Engineering, Elsevier, vol. 23, pp
755756.

141

Maboumine process: Example of a

promising process for developing a
polymetallic ore deposit

As a growing mining company, ERAMET is looking for new positions on the

strategic metals market, as ores are becoming increasingly complex to handle.
In this context, the Gabonese Maboumine pyrochlore deposit is currently

Florent Delvalle, Valrie Weigel

and Antoine Greco. ER A MET
Research, France

being studied for the recovery of niobium, tantalum, rare earths and uranium.
This ore shows mineralogical and chemical characteristics that make it challenging to process by already known practices, from both technical and economical points of view. Pyrochlores are in an advanced weathered state, as
well as very fine and entrapped into a goethite and crandalite matrix. Nevertheless recent improvements in niobium, tantalum and rare earth chemistries
in high ionic strength and hot aqueous media have been made. It has allowed
ERAMET Research to develop an innovative hydrometallurgical process, which
has been successfully operated at a pilot plant scale. The upstream part of the
process which aims to dissolve valuable elements consists in several steps. A
first one is dedicated to dissolve the ore matrix, leaving the pyrochlore unleached. A second step consists of chemically breaking down pyrochlores in
order to form soluble species of the valuable elements. Niobium and tantalum
on the one side, and rare earths and uranium on the other, are separated into
different flows. Finally each one feeds a specific section of the downstream
part of the process, which is dedicated to obtain commercial grade products,
such as ferro-niobium, rare earth and uranium concentrates.

143

INTRODUCTION
ERAMET is a world leader in alloying metals and in high-quality metallurgy. The group employs
about 15,000 people in 20 countries, through their subsidiaries, like, for example COMILOG in
Gabon for manganese, Socit Le Nickel in New-Caledonia for nickel and Aubert & Duval for
alloys in France. As a growing company, ERAMET has major research and development projects in
new business lines with high growth potential, such as titanium dioxide, zircon, lithium, as well as
recycling.
Niobium, and rare-earth elements have been listed by European Union and United States
authorities as critical metals from the point of view of supply and production. Indeed, almost 85%
of the world production of niobium is ensured by Brazil, and 95% of the rare-earths by China.
Niobium is mainly used in microalloyed steel production and rare-earth have a wide diversity of
applications as for example magnets for wind-turbines or hybrid cars, or optical applications for
LCD television.
In this context, the Gabonese Maboumine polymetallic deposit is currently being studied by
ERAMET. An innovative hydrometallurgical process has been developed for the recovery of
niobium, tantalum, rare earths and uranium.

METHODOLOGY
Maboumine process development involves around a hundred people worldwide, with laboratories
in France, Australia, Canada, the United States, South-Africa and Finland. Research for process
development is organized around ERAMET Research with a forty people team and follows a
several steps methodology, as described hereafter:

Mineralogical study of the ore

Laboratory test work to define the process steps and the related chemical and physical
conditions

Pilot plant scale operation in order to validate the process in a continuous configuration

Demonstration plant in order to validate technology options, and to provide data for the
scale up to the commercial plant

Today, the process has been entirely operated at pilot plant scale in France since the end of 2012. A
demonstration plant is planned for 2015. 4 patents have been written on the process.

MABOUMINE ORE
The deposit
The Maboumine deposit is located near the Mabouni River in the Gabonese rainforest. Geology of
the Maboumine ore body has been extensively studied by ERAMET recently and Bureau de
Recherches Gologiques et Minires (BRGM) in the past after several drilling campaigns. The
vertical profile of the ore body can be represented by 3 layers. The top layer is the superficial and
ribonned one that is treated for the recovery of niobium, tantalum, rare-earth and uranium.

144

Underneath lays a phosphate ore that can be mined out, for example to produce phosphoric acid.
These two upper layers result from the weathering of the carbonatite bottom layer.

Figure 1 Vertical profile of the deposit and map of Maboumine site with delimitation of the drilled area

The mean chemical composition of the deposit for the superficial and ribonned layer is given
hereunder in Table 1:
Table 1 Mean chemical composition of the superficial and ribonned layer of the deposit
Fe

Al

Ti

Mn

Si

Nb

RE

Th

Ta

35%

6,1%

2,7%

2,5%

2,4%

1,6%

1,2%

1,4%

0,10%

0,03%

0,03%

Valuable elements are rather diluted in the Maboumine ore, considering 1.2% niobium, 1.4% rareearth, 0.03% uranium and 0.03% tantalum. Among the on-going project of rare-earth mines
industry outside of China, Maboumine ore deposit can be considered as rich with heavy rare earth.
It is to mention that uranium and thorium are radioactive elements.

Mineralogy
Samples of the superficial and ribonned layer have been characterised over the past several years
using SEM, MLA or QEMSCAN, XRD and chemical analysis. It has shown that the ore consists in a
complex mixture of minerals, with iron bearing goethite and magnetite as well as calcium,
phosphate and aluminium bearing crandalite matrix. The ore contains a wide diversity of other
minerals such as the titanium bearing ilmenite and rutile or manganese bearing pyrolusite, as well
as the silica bearing quartz, etc. Most of the valuable elements, typically 70% to 80%, are hold by
pyrochlore minerals. This mineral, which is in an advanced weathered state, exists in several forms
in the Maboumine ore, with multiple texture and compositions. Indeed, pyrochlore can be

145

represented as an A2B2O6(OH,F) formula with the A site as U, RE, Na, Ca or Th and the B site as Nb,
Ta, Ti or Fe.

Figure 2 SEM views of Pyrochlore grains from Mabounine ore with varying texture on the left hand-side, and
varying composition on the right-hand side

Particle size distribution study coupled with chemical analysis on samples of the ribonned and
superficial ore showed that it is very fine. Typical values obtained on a given sample exhibit that
the inferior to 10 m size fraction contained 31% of the pyrochlore within 58% of the total mass
distribution, whereas the superior to 50 m size fraction contained 36% of the pyrochlore within
17% of the total mass distribution. Even if liberated pyrochlore can be easily obtained by fine
grinding, traditional beneficiation techniques may be applicable only to the coarser material, but
there will be a need for novel beneficiation techniques to deal with the ultra-fine (e.g.
approximately <10 m material). In any case, the diversity of pyrochlores compositions and surface
properties due to their advanced weathered state is making floatation very difficult.

PROCESS DEVELOPMENT
Classical routes for niobium production and application to the Maboumine ore
Pyrometallurgy process
Classical route for ferro-niobium production requires pyrochlore to be concentrated by mineral
processing involving floatation, prior to pyrometallurgical treatment (Gupta, 1994). Pyrochlore
concentrate requires then a dephophorization step which leads to obtain a phosphorous pig iron
and a niobium rich slag. The former is introduced to an aluminothermy step for the production of
ferro-niobium.

146

Ore

Pyrochlore
floatation

Pyrochlore
concentrate

Dephosphorization

Ferro-niobium
Niobium
rich slag

Phosphorous
pig iron

Tail

Aluminothermy

Al2O3, CaO
slag

Figure 3 Pyrometallurgical route for production of ferro-niobium

Since only a small fraction of pyrochlores of the Maboumine ore can be concentrated by mineral
processing, classical route for ferro-niobium production is not economical for ERAMET. Moreover,
pyrometallurgical process does not allow recovering rare-earth and uranium which are lost in the
Al2O3-CaO slag.

Hydrometallurgy process
Niobium species are known for having poor solubilities in most of mineral acids, except
hydrofluoric acid, but forms soluble species with complexing agent such as oxalic acid or in very
basic media such as potash (R.J.H. Clark, D. Brown, 1973). The only known hydrometallurgical ore
treatment process for the recovery of niobium that has been industrialized so far is hydrofluoric
acid route (Polak). Actually it is rather a tantalum recovery route. It is illustrated by the scheme
below.

Nb stripping
solution
HF/H2SO4

Ore

Leaching

TaF72-(aq)

MIBK
NbOF52-(aq) solvent extraction
Ta stripping
solution

Nb
precipitation
Spent
solution

Ta
precipitation

Figure 4 Hydrofluoric route for niobium-tantalum ore treatment

Hydrofluoric route consists in leaching niobium-tantalum ore by a mixture of hydrofluoric and

sulphuric acid. Niobium and tantalum are separated by a ketone based solvent extraction, and then
selectively striped one after the other from the loaded organic solution. Niobium and tantalum are
then precipitated form their respective streams and subsequently calcined in order to obtain high
purity oxides. The main disadvantage of this route is the use of hydrofluoric acid which is major
health safety and environment issue. Besides, it would not be an economical route to treat
Maboumine ore.

Maboumine process
Considering that classical pyrometallurgical or hydrometallurgical routes can not be applied to
Maboumine ore, alternative process has been developed by ERAMET Research. For informations

147

protection reason, the process description will not be explicit. Nevertheless, it can be mentioned
that it is hydrofluoric acid free. The upstream part of the Maboumine process is dedicated to
dissolve valuable elements and consists in several steps that are explained downunder and
illustrated by Figure 5 below :

Ore dressing and magnetic separation: After grinding and scrubbing steps, magnetic
separation enables to remove around 30% of the ore mass, mainly as magnetite. This
fraction does not contain valuable elements. The remaining fraction is called the non-mag
ore.

Matrix leaching: the main purpose of this step is to leach goethite and crandalite matrix.
Pyrochlores remain unleached and are concentrated into solid that feeds the Pyrochlore
breakdown step. Since part of the valuable elements are not holds by pyrochlores, they
dissolved to a certain extend at this leaching step. This is the reason why, rare earth and
uranium solution from downstream step is recycled to the leaching. The solution exiting
this step is the rare-earth and uranium pregnant leach solution

Pyrochlore Breakdown: Pyrochlores are among the least leachable minerals of the ore.
Therefore, they undergo a specific treatment in order to breakdown their structure and
form soluble species of the valuable elements.

Valuable elements dissolution: This step allow all valuable elements to be dissolved and
produces a PLS that contains most of the niobium, tantalum and rare-earth. The solid
residue is mainly silica containing species.

Nb/Ta precipitation: niobium and tantalum are recovered from PLS by precipitation as
impure concentrate.
Ore

Ore dressing
Magnetic separation

Magnetic
fraction

Non-mag
ore

Pyrochlore
Breakdown

Matrix
Leaching
RE & U PLS

Nb/Ta precipitation

Spent
solution

Nb/Ta
precipitate

Nb, Ta,
RE & U
PLS

Valuable Elements
dissolution
Final residue

Figure 5 Upstream part of the Maboumine

Niobium, rare-earth, tantalum and uranium dissolution yields in the upstream part of the process
reach about 80% respect to the ore at pilot plant scale. One particulartiy of the process is that
niobium and tantalum on the one side, and rare earths and uranium on the other side, are separated
into different streams. Each one of them feeds a specific section of the downstream part of the
process, which is dedicated to obtain commercial grade products, such as ferro-niobium, rare earth
concentrates and uranium concentrate.

148

Challenges related to processing polymetallic ores

Modelling is an important aspect of process design for which thermodynamics and species
speciation are essentials. On the one side, existing models for metal chemistry usually describe
chemical behaviour of a limited amount of species in a rather diluted media with low ionic
strength, and quite commonly at ambient temperature. On the other side, aqueous solutions
obtained by dissolution of complex ore bodies are likely to have the following properties:
Concentrated media, with very high ionic strength
Valuable elements are in minor or trace concentration compared to impurities from trace
concentration to major concentration
High temperature
Therefore thermodynamic modelling can be hardly applied to many of actual cases of
Hydrometallurgical ore treatment. Besides, extrapolating existing models to real cases, as far as
they could be appropriate, faces a lack of data. This lack of data and of the absence appropriate
models has at least 2 explanations:
The complexity of the actual solutions makes modelling extremely difficult
Measurement methods for determining the species in which metals are involved are to be
developed.
The problem of measurement method is also critical for other reasons besides thermodynamics
modelling. Speciation and chemical behaviour understanding in such complex media are not
always described in scientific literature. Phenomenon that can be observed and even controlled
experimentally can remain unexplained, because classical measurement methods may not be able
to give answers. Indeed measurement in complex and concentrated media is a challenge
considering that measuring trace metals in a concentrated matrix is difficult. The response of the
observed metal can be totally hidden by the noise or other components response.

CONCLUSIONS
The Gabonese Maboumine pyrochlore deposit is currently being studied by ERAMET for the
recovery of niobium, tantalum, rare earths and uranium. This ore shows mineralogical and
chemical characteristics that make it challenging to process by already known practices, from both
technical and economical points of view. Pyrochlores are in an advanced weathered state, as well as
very fine and entrapped into a goethite and crandalite matrix. Nevertheless recent improvements
in niobium, tantalum and rare earth chemistries in high ionic strength and hot aqueous media have
been made. It has allowed ERAMET Research to develop an innovative hydrometallurgical process,
which has been successfully operated at a pilot plant scale.

REFERENCES
C. K. Gupta, A. K. Suri (1994), Extractive Metallurgy of Niobium, CRC Press,
C. Polak, Mtallurgie et Recyclage du niobium et du tantale, Dossier Techniques de lingnieur, m2365
R.J.H. Clark, D. Brown (1973), The Chemistry of Vanadium, Niobium and Tantalum, chapter 34 & 35,
Comprehensive Inorganic Chemistry, Pergamon Press.

149

chap. 3
Cyanidation, leaching and recovery of gold

A theoretical study of SART

precipitate generation:
Operational and safety impacts

SART (Sulphidization, Acidification, Recycling and Thickening) process is a

cyanide recovery technology currently installed in different gold cyanidation
plants around the world. This technology reduces the operational costs by

Humberto Estay, Pablo Carvajal,

Karina Gonzlez and Vernica
Vsquez. Arcadis, Chile

recovering cyanide and increases the income through the production of a

copper precipitate. The literature has reported that the precipitate is mainly
Cu2S and, consequently, the engineering design of SART plants have been
based on this fact. However, the experimental and operational results available, such as the required dosage of reagents and the observed copper grade
in the precipitates, introduce doubts about the correct characterization of the
copper precipitate. These results suggest the formation of other species
different from Cu2S, which then bias the true estimation of reagents dosage
and the selection of the safety criteria needed in the drying and storage
facilities. The present work applies thermodynamic modeling to identify
theoretically formation of the species CuS in the precipitate, as a function
of the operational conditions. This paper proposes the use of a 120% NaHS
stoichiometric addition based on Cu2S in order to obtain copper recoveries
around 90%; however, NaHS dosages above 120% stoichiometry increases the
NaHS consumption without increasing copper recovery. Furthermore, the
presence of oxidation reactions in the stored precipitate promotes the ability
to undergo spontaneous combustion under certain conditions. Therefore,
the inclusion of safety criteria in the storage facilities must be incorporated
in the design and operation of sart plants in order to minimize hazardous
situations in the facilities.

153

INTRODUCTION
Description of SART Process
The SART process (Sulphidization, Acidification, Recycling, and Thickening), illustrated below in
Figure 1, is a technology currently incorporated in the cyanidation circuit by some gold mining
operations in order to reduce the cyanide consumption associated to the processing of ores with
high cyanide-soluble copper content.
The process begins with the Acidification and Sulphidization of the feed solution according to
equation (1) and (2) by adding sulphuric acid (H2SO4) and sulphide (in the form of NaHS or Na2S),
producing precipitation of copper sulphide (Cu2S) in the precipitation reactor, among that other
cations present in the solution (Estay et al., 2012). The hydrogen cyanide generated remains
dissolved in the solution (MacPhail, Fleming & Sarbutt, 1998).

Figure 1 Schematic block diagram of the SART process

2CN + H 2 SO4 2 HCN (ac) + SO42

(1)

2Cu (CN ) 32 + S 2 + 6 H + Cu 2 S ( s ) + 6 HCN (ac )

(2)

The solids formed in the precipitation reactor are removed from the process by sequential stages of
thickening, filtration, and drying, while the treated solution is neutralized by adding milk of lime
which produces the conversion of dissolved HCN into calcium cyanide (Ca(CN)2) and precipitation

154

of gypsum (CaSO4*2H2O) (Barter et al., 2001). Finally, the solution produced by the SART process is
put back into the leaching circuit, having all of its cyanide content as soluble Ca(CN)2, which is
equivalent to free cyanide for the purposes of gold dissolution in the cyanidation process
(MacPhail, Fleming and Sarbutt, 1998).

Precipitate characterization
The literature has reported that Cu2S is the main species formed in the precipitation stage of the
SART process (MacPhail, Fleming & Sarbutt, 1998; Barter et al., 2001; Estay, Arriagada & Bustos,
2010). However, the experimental and operational data suggest that other species may be the
reaction product. There are three issues which indicate the possible formation of others species, as
follows:
1. The typical copper grade in the SART precipitate is around 65% (Ford, Fleming & Henderson,
2008; Estay et al., 2012), almost 15% lower than the stoichiometric relation of Cu2S (79.9%). In
fact, the precipitation of other cations contained in solution such as zinc, nickel, and silver can
not explain the difference of copper grade between operational results and the expected value
given by the stoichiometric relation.
2. The operational data indicate that the copper recovery in the SART process varies between 90%
and 95% when the NaHS addition is 120% stoichiometric based on copper (Estay et al., 2012;
Estay, Arriagada & Bustos, 2010). Precipitation of Cu2S (equation 2) is very fast and determined
by thermodynamic equilibrium, therefore the excess of NaHS addition ought to be explained by
the generation of other species.
3. Finally, current SART operating plants have encountered exothermic reactions (T. Van Zeller,
2012, pers. comm., 04 November) that could promote spontaneous combustion of precipitate
bags, after a period of storage. The characterization of the precipitate in this connection is very
relevant since Cu2S has a self-ignition point higher than 400C.
These results are taken to indicate that in the precipitate there are species not considered to date,
which could be formed in the precipitation stage, such as CuS or S. This study analysed the
thermodynamic equilibrium in the system to assess the generation of CuS and S species, in addition
to the main reactions occurring in the SART precipitation stage. Furthermore, the results of the
theoretical model are compared with the literature in order to identify the nature of these
spontaneous combustion events.

METHODOLOGY
The theoretical study for characterization of the precipitate in the SART process involved the
following steps:
1. Review of experimental and operational data to establish the facts about the true
characterization of the SART precipitates.
2. Establish the main chemical reactions that could be involved in the precipitation reactor of the
SART process.
3. Conduct a literature survey of equilibrium constants to build the thermodynamics model.
4. Construction and testing of the thermodynamics model.

155

5. Conduct a literature study regarding the spontaneous combustion of sulphide concentrates.

6. Analyse the thermodynamics model results.

RESULTS AND DISCUSSION

Theoretical model
The theoretical model developed in this work is based on the following reactions, which could
occur during the precipitation stage:
+

(3)

( ),

( )+

) ,

(4)

) +

) ,

(5)

) ,

(6)

( )+

( )+

( )+

+2

(7)

(8)

(9)

(10)

(11)

Thus, the solid species which could form in the precipitation stage are CuCN, CuS, Cu2S and S. The
feed solution could contain cations in low concentration (<10 mg/L) such as zinc, cadmium, nickel,
and silver, which can precipitate. These cations however have not been considered in this present
model.
The model relates the thermodynamic equilibrium equations of each chemical reaction and the
mass balance for cyanide, copper and sulphide to obtain a mathematical model describing the

156

speciation in the solid-liquid equilibrium, as a function of pH. The model has been constructed in
Matlab. The equilibrium constants of equations (3-11) are presented in Table 1.
Table 1 Equilibrium constants values for Equations 3-11 at 25C
logKi

Value

Source

logKCu1

20

Lu, Dreisinger & Cooper, 2002

logKCu2

3.94

Izzat et al., 1967

logKCu3

5.3

Izzat et al., 1967

logKCu4

1.5

Izzat et al., 1967

logKCuS

19.52

Zhang, Fang & Muhammed, 1997

logKCu2S

34.6

Zhang, Fang & Muhammed, 1997

logKS

2.2

Zhang, Fang & Muhammed, 1997

logKH2S

7.02

Zhang, Fang & Muhammed, 1997

logKHCN

9.21

Izzat et al., 1962

Simulations were conducted by considering a hypothetical cyanide solution containing 500 mg/L
Cu and 800 mg/L CN-, treated with NaHS dosages of 80, 100 and 120% stoichiometric based on
Cu2S during the precipitation stage, at different pH values. The copper species distribution results
obtained by the model are shown in Figures 2, 3 and 4 below.
The SART process operates in the pH range of 4 to 5. According to the results of the model, CuS
may be present in the precipitate at pH 5, which has 2, 15 and 36% of copper for 80, 100 and 120%
stoichiometric NaHS addition, respectively. That is, the increase in the sulphide/copper ratio
increases the stability range of CuS in the precipitate. On the other hand, a limited sulphide content
produces CuCN at pH lower than 4, increasing the stability of Cu(CN)2- species at pH 5, which
reduces copper recovery in the SART process. Elemental sulphur is not present for pH lower than 6,
and H2S is present for 120% stoichiometric NaHS addition at pH lower than 4. Hence, from the
thermodynamic standpoint, CuS could be present in the precipitate depending on the process
conditions (solution characterization, sulphide addition and pH).
100

90

Cu distribution, %

80
%Cu Cu+

70

%Cu CuCN

60

%Cu Cu(CN)2

50

%Cu Cu(CN)3
%Cu Cu(CN)4

40

%Cu CuS

30

%Cu Cu2S

20
10
0
0

pH

10

12

14

Figure 2 Copper species distribution for cyanide solution containing 500 mg/L Cu, 800 mg/L CN- and 80% of
NaHS stoichiometric addition (176.4 mg/L NaHS) based on Cu2S

157

100

Cu distribution, %

90
80

%Cu Cu+

70

%Cu CuCN

60

%Cu Cu(CN)2
%Cu Cu(CN)3

50

%Cu Cu(CN)4

40

%Cu CuS

30

%Cu Cu2S

20
10
0
0

pH

10

12

14

Figure 3 Copper species distribution for cyanide solution containing 500 mg/L Cu, 800 mg/L CN- and 100% of
NaHS stoichiometric addition (220.5 mg/L NaHS) based on Cu2S

100

Cu distribution, %

90
80

%Cu Cu+

70

%Cu CuCN

60

%Cu Cu(CN)2
%Cu Cu(CN)3

50

%Cu Cu(CN)4

40

%Cu CuS

30

%Cu Cu2S

20
10
0
0

pH

10

12

14

Figure 4 Copper species distribution for cyanide solution containing 500 mg/L Cu, 800 mg/L CN- and 120% of
NaHS stoichiometric addition (264.6 mg/L NaHS) based on Cu2S

The formation of CuS in the precipitate explains in part (0.5 to 5.5 points of % from the typical
difference of about 15%) the copper grade reported in the The remaining 10 points of % can be
explained by co-precipitation of zinc, nickel, silver o cadmium. It also explains the excess of NaHS
addition required to obtain copper recoveries higher than 90%. Actually, the effective stoichiometric
addition depends on the ratio of CuS and Cu2S generated; thus, when CuS formation is increased
the sulphide consumption increases. Hence, a 120% NaHS stoichiometric addition based on Cu2S
ensures copper recoveries higher than 90%, while NaHS dosages above 120% stoichiometric are not
recommended since the additional sulphide will be consumed to form CuS without increasing
copper recovery. This trend was confirmed by additional simulations conducted above 120% NaHS
stoichiometric (results not shown).

158

The model can predict the copper recovery obtained in the precipitation stage, according to the
solution speciation and pH. Therefore, it can be a useful tool to predict the copper recovery as a
function of NaHS addition.

Spontaneous combustion risk

It is well known that spontaneous combustion of sulphide concentrate is a common phenomenon
under favourable conditions (Yang et al., 2011). The ability of sulphide concentrates to undergo
spontaneous combustion has been investigated (Yang et al., 2011; Yang, Wu & Li, 2011; Payant et
al., 2012; Yang & Wu, 2013). These authors studied the activation energy of spontaneous
combustion in storage yards for different sulphide concentrates, suggesting that the mechanisms
that can promote self-heating may be oxidation reactions, galvanic interactions, mechanical
activation, among others.
According to the results of the thermodynamics model, the characterization of the SART precipitate
is basically Cu2S with the presence of CuS, depending on the process conditions. Even when the
results of some researches indicate that copper sulphide concentrates are less susceptible to
spontaneous combustion at room temperature than iron sulphide concentrates or sulphur-rich
concentrates (Yang et al., 2011), there are experiences in current operating SART plants reporting
self-heating of precipitates in the storage area.
The occurrence of oxidation reactions can be expected when the sulphide concentrate is stored for a
period of time, releasing heat. If the rate of heat generation in the sulphide concentrate storage area
exceeds that of heat removal from the boundaries, the accumulated heat rises the temperature
producing the oxidizing reaction to accelerate up to the self-ignition point (Zarrouk & OSullivan,
2006).
Therefore, future studies should be focused on confirming the causes of spontaneous combustion,
i.e. CuS effect, energy activation, contact area of species, oxidation reactions, among others.
In summary, the SART precipitate as a sulphide concentrate can, under certain conditions, suffer
spontaneous combustion. This fact forces to include important safety criteria in the storage
facilities, such as to avoid large quantities of precipitate stored for extended periods, minimize air
contact, avoid the natural dried and to include the necessary installations to ensure a safety area,
such as fire protection system and a proper handling of the bags.

CONCLUSION
The precipitate formed in the SART process was characterized from a theoretical point of view. The
thermodynamics solid-liquid equilibrium suggested the formation of CuS in addition to Cu2S, its
content being dependent on the operational conditions. It is concluded that copper recoveries of
90% can be reached using 120% NaHS stoichiometric addition based on Cu2S; however, NaHS
dosages above 120% stoichiometric could only consume sulphide to form CuS without increasing
copper recovery.
There is no evidence in the literature about the effect of CuS on promoting self-heating. In this
context, future studies should be focused on confirming the causes of spontaneous combustion.
However, the fact of self-heating of SART precipitate forces the inclusion of safety criteria in the

159

storage facilities, which will be incorporated in the design and operation of SART plants, in order to
minimize the risk in the facilities as well as in personnel.

REFERENCES
Barter, J., Lane, G., Mitchell, D., Kelson, R., Dunne, R., Trang, C. & Dreisinger, D. (2001) Cyanide management
by SART, Cyanide: Social, Industrial and Economical Aspects, Young, C. (Ed.), The Minerals, Metals
& Materials Society, pp. 549562.
Estay, H., Arriagada, F. & Bustos, S. (2010) Design, Development and Challenges of the SART Process,
Proceedings of the 3rd International Workshop on Process Hydrometallurgy, Santiago, Chile, 11-13 August,
Gecamin, Chile, pp. 3031.
Estay, H., Carvajal, P., Hedjazi, F. & Van Zeller, T. (2012) The SART process experience in the Gedabek plant,
Proceedings of the 4rd International Workshop on Process Hydrometallurgy, Santiago, Chile, 12-13 July,
Gecamin, Chile, pp. 144155.
Ford, K., Fleming, C. & Henderson, R. (2008) Application of the SART Process to Heap Leaching of GoldCopper Ores at Maricunga, Chile, paper presented to the 40th Annual Meeting of the Canadian Mineral
Processors, Ottawa, 22-24 January.
Izzat, R., Christensen, J., Pack, R. & Bench, R. (1962) Thermodynamics of metal-cyanide coordination. I. pK,
H0, S0 values as a function of temperature for hydrocyanic acid dissociation in aqueous solution.
Inorganic Chemistry, vol. 1, pp. 828-831.
Izzat, R., Johnston, H., Watt, G. & Christensen, J. (1967) Thermodynamics of metal-cyanide coordination. VI.
Copper(I)-and Silver(II)-Cyanide systems. Inorganic Chemistry, vol. 6, pp. 132-135.
Lu, J., Dreisinger, D. & Cooper, W. (2002) Thermodynamics of the aqueous copper-cyanide system,
Hydrometallurgy, vol. 66, pp. 23-36.
MacPhail, P.K., Fleming, C. & Sarbutt, K. (1998) Cyanide Recovery by the SART Process for the Lobo-Marte
Project, Chile, paper presented to the Randol Gold and Silver Forum, Denver, 26-29 April.
Payant, R., Rosenblum, F., Nesset, J. & Finch, J. (2012) The self-heating of sulfides: Galvanic effects, Minerals
Engineering, vol. 26, pp. 57-63.
Yang, F., Wu, C., Cui, Y. & Lu, G. (2011) Apparent activation energy for spontaneous combustin of sulfide
concentrates in storage yard, Transactions of Nonferrous Metals Society of China, vol. 21, pp. 395-401.
Yang, F., Wu, C. & Li, Z. (2011) Investigation of the propensity of sulfide concentrates to spontaneous
combustion in storage, Journal of Loss Prevention in the Process Industries, vol. 24, pp. 131-137.
Yang, F. & Wu, C. (2013) Mechanism of mechanical activation for spontaneous combustion of sulfide
minerals, Transactions of Nonferrous Metals Society of China, vol. 23, pp. 276-282.
Zarrouk, S., & OSullivan, M. (2006). Self-heating of coal: the diminishing reaction rate. Chemical Engineering
Journal, vol. 119, pp. 83-92.
Zhang, Y., Fang, Z. & Muhammed, M. (1997) On the solution chemistry of cyanidation of gold and silver
bearing sulphide ores. A critical evaluation of thermodynamic calculations, Hydrometallurgy, vol. 46,
pp. 251-269.

160

Development of a new technique to

recover cyanide for gold mining
using membranes contactors

In the last 10 years, operational costs in gold cyanidation processes have increased, fundamentally, due to the treatment of ores which contain high cyanide soluble metals, such as zinc, copper, nickel, among others, increasing
the cyanide consumption along with the increase in cyanide price. Moreover,
the high cyanide consumptions in gold operations have increased the cyanide
contents in leach tailings, forcing the inclusion of processes to recover or
eliminate cyanide. In this way, a membrane contactor operation to recover

Humberto Estay. Laboratory of

Membrane Separation Processes,
Universidad de Santiago de Chile and
Arcadis, Chile
Miguel Ortiz and Julio Romero.
Laboratory of Membrane Separation
Processes, Universidad de Santiago
de Chile

cyanide from gold mining using gas filled membrane absorption is proposed
in this study, performing laboratory tests in order to determine the throughput
of this process and the most relevant parameters, which have an effect on the
cyanide recovery. The Gas Filled Membrane Absorption (gfma) process can
extract cyanide, using a hydrophobic membrane with its pores filled with air
separating an aqueous stream which contains cyanide and an absorption
solution of NaOH. These aqueous solutions cannot penetrate into the pores
of the hydrophobic membrane, promoting the cyanide transfer from the cyanidation solution to the receiving solution. The experimental results show
cyanide recoveries higher than 90% in extraction times close to 10 minutes,
achieving average cyanide transfer rate of 0.01kg m-2 h-1 and identifying the
pH, feed flow rate and copper concentration as the most influential parameters
on the process performance. Additionally, a scoping study comparing this
process with the current cyanide recovery processes used in gold mining has
been performed, based on the experimental results. The NPV results [3.65kUS$/
(ton NaCN/y)] indicate that the GFMA process is an interesting alternative to
recover cyanide in gold cyanidation processes.

161

INTRODUCTION
The recovery of gold and silver by using cyanidation has increased the operational costs, due to the
increase in cyanide price and the high cyanide consumptions, promoted by the treatment of ores
containing cyanide-soluble metals, such as zinc, copper, nickel, among others. These cyanidesoluble metals generate cyanide-complex in plant solutions, which increase gradually the cyanide
concentration in the leach tails (Fleming, 2005). In order to reduce the operational costs of the
cyanidation processes and cyanide contents in leach tails, many cyanide recovery process have been
developed (Fleming, 2005), and currently the gold mining operations have incorporated two
cyanide recovery processes in the cyanidation circuits: the SART (sulphidisation, acidification,
recycling and thickening) process (MacPhail, Fleming & Sarbutt, 1998) and the AVR (acidification,
volatilization and regeneration) process (Smith & Mudder, 1991). The SART process is applied in
operations with high copper contents, precipitating Cu2S and recovering cyanide in the treated
solution. The SART process uses big equipment in solid-liquid separation processes, in order to
recover the copper precipitate and gypsum formed in the process (Estay, Arriagada & Bustos, 2010).
In turn, the AVR process is applied to recover cyanide from tailings and solutions, using big
desorption and absorption towers, and a high air-solution flow ratio to volatilize the hydrogen
cyanide generated in the process.
The Gas filled membrane absorption process (GFMA) has been applied to recover cyanide in
plating (Kenfield et al., 1988) and pesticide (Shen, Han & Wickramasinghe, 2006) solutions in
laboratory and pilot scale, but no experience in gold cyanidation solutions has been reported.
GFMA process operates using hydrophobic hollow fiber membrane contactors, contacting the
cyanidation solution with the absorption phase. The hydrophobic character of the membrane
avoids the penetration of solutions into the pores, which are filled with air. The driving force of this
process is the partial pressure gradient of HCN through the membrane, as shown the Figure 1.

Figure 1 Outline of the gas filled membrane absorption process (Estay, Hasanoglu & Romero, 2011)

162

The HCN transfer through the membrane is achieved by means of the following sequence of steps,
which are represented in Figure 1:
1. Transfer through the boundary layer of aqueous feed solution to be treated.
2. Phase equilibrium between the feed solution and the gas phase (air) retained in the membrane
pores.
3. HCN transfer through the air gap that fills the pores.
4. Phase equilibrium between the gas and the receiving solution. The formation of an absorbed
compound at the interface can be considered in this step.
5. Mass transport of absorbed cyanide into the bulk receiving solution.
The feed stream of the GFMA process must be acidic (pH 4-5) in order to maximize the presence of
cyanide as HCN in the solution (pKa=9.21). Thus, the cyanide contained in the cyanide solution is
acidified to pH 4-5, in order to obtain HCN dissolved, which is transferred through the membranes
pores into the absorption solution (10 wt% of NaOH). HCN transfer in the GFMA process can be
described by the Equation 1.

N HCN = KAC HCN

(1)

where, NHCN is the flux of HCN transferred from cyanidation solution to absorption solution, K is
the overall mass transfer coefficient, A is the mass transfer area and CHCN is the difference of HCN
concentration between the cyanidation solution and the absorption solution. The overall mass
transfer coefficient can be estimated by the following equation:

1
1
1
=
+
K k L m HCN k m

(2)

Where, kL is the local mass transfer coefficient of the feed solution, km is the local mass transfer
coefficient in the pores and mHCN is the partition constant, which depends on the Henrys law
constant (HHCN). The Henrys law constant represents the phase equilibrium of HCN in the interface
of cyanidation solution and the air gap of the pores, and can be described by Equation 3.

H HCN =

PHCN
C HCN

(3)

In equation 3, PHCN is the partial pressure of HCN in the boundary layer of air gap in the
membranes pores and CHCN is the concentration of HCN in the boundary layer of the cyanidation
solution. A third term on the right side of Equation 2, referred to the transfer resistance at the

163

receiving phase boundary layer, can be neglected if the conversion of HCN in cyanide in the
receiving solution (Equation 4) is considered instantaneous and complete.

HCN + NaOH
NaCN + H 2 O

(4)

This study reports the experimental and theoretical analysis of a GFMA process, using a synthetic
solution, which simulates a typical gold cyanidation solution.

METHODOLOGY
Experimental
A GFMA process was implemented at laboratory scale, treating a synthetic solution under different
operating conditions in order to obtain the effect of the main parameters in the cyanide recovery. A
GFMA process laboratory scale was mounted (Figure 2), using a membrane contactor described in
Table 1. Synthetic solutions were prepared using NaCN 98% purity, copper and zinc cyanide
solutions were prepared from CuCN and ZnSO4*7H2O powder by dissolving in NaCN/NaOH
solution to obtain solutions containing 2,000 mg/L of total CN-, pH 11, 0-500 mg/L zinc and 0-100
mg/L copper. Cyanide analyses were performed by titration with AgNO3 and KI as indicator.
Chemical analysis of cyanide was performed in the absorption solution. A final cyanide analysis in
the feed solution was performed, in order to make an accountability of cyanide in the system.
Table 1 Experimental conditions of GFMA process
Description

Value

Unit

Feed volume tank

0.5

Absorption volume tank

0.5

L/min

pH of cyanidation solution

4-7

Temperature

5-25

polypropylene

Number of fibers

7400

Surface contact area

0.58

m2

Fiber OD

330

Fiber, ID

220

Length

0.12

Feed solution flow

Membrane material

164

Figure 2 Experimental setup for GFMA process tests implemented at laboratory scale (Estay, Ortiz & Romero,
2013)

Scoping study
Economical comparison between GFMA process and the current cyanide recovery processes (SART
and AVR) was performed, based on the following steps:
1. Theoretical model of the GFMA process was developed based on thermodynamics and mass
transfer aspects.
2. Scale up of the GFMA process for industrial capacity (150-300 m3/h).
3. Design of SART and AVR processes for industrial capacity (150-300 m3/h).
4. Estimation of capital and operational costs of GFMA, SART and AVR processes.
5. Estimation of comparative economical evaluation between GFMA, SART and AVR processes
(40% accuracy).

RESULTS AND DISCUSSION

Effect of the metals in cyanide recovery
The effect of zinc and copper concentration on the HCN removal capacity of GFMA process was
measured. Cyanide feed solution containing zinc concentration between 0 and 500 mg/L and
copper concentration ranging between 0 and 100 mg/L (in presence of 500 mg/L Zn), were treated
in the membrane contactors at pH 5 and T=15C. The results of the effect of metal concentration on
the HCN removal capacity are shown in Figures 3 and 4 for zinc and copper, respectively.

165

100

Cyanide recovery (%)

80

60

Zn concentration

40

0 mg/L
200 mg/L
500 mg/L
20

0
0

10

15

20

25

Time (min)

Figure 3 Effect of Zn content in the cyanide recovery for GFMA process

100

Cyanide recovery (%)

80

60

Cu concentration

40

0 mg/L
50 mg/L
100 mg/L
20

0
0

10

15

20

25

Time (min)

Figure 4 Effect of Cu content in the cyanide recovery for GFMA process ([Zn]=500 mg/L)

HCN removal tests with metal presence show cyanide recovery values higher than 90% in 10
minutes of extraction. The effect of zinc content was no relevant, since the zinc-cyanide complex is
dissociated below pH 5.5. Thus, the total cyanide content can be considered as HCN dissolved in
the feed solution.
On the other hand, the copper content affects the cyanide recovery because CuCN precipitates
during the process, reducing the total cyanide recovery. The cyanide extraction runs were finished
after 10 min of processing, in order to avoid fouling on the membrane contactor.

166

Effect of pH in cyanide recovery

The effect of pH in cyanide recovery was measured. GFMA process at 15C, [Cu]=100 mg/L and
[Zn]=500 mg/L was carried out ranging pH between 4 and 7. The results are shown in Figure 5.

100

Cyanide recovery (%)

80

60

40
pH = 4
pH = 5
pH = 7
20

0
0

10

12

14

Time (min)

Figure 5 Effect of pH in the cyanide recovery for GFMA process

Cyanide recovery is affected by the pH value, obtaining the best results with pH=4.0, due to the
effect of pH on the dissociation of copper-cyanide complexes. Cyanide recovery values higher than
90% in 10 minutes could be reached operating at pH ranged between 4.0 and 5.0.
Effect of temperature in cyanide recovery
The effect of temperature on the extraction capacity of the GFMA process was assessed, varying the
temperature of solutions between 5 and 25C, at pH 5, [Cu]=100 mg/L and [Zn]=500 mg/L. These
results are reported in Figure 6.

100

Cyanide recovery (%)

80

60

40
T = 5 C
T = 15 C
T = 25 C
20

0
0

10

15

20

25

Time (min)

Figure 6 Effect of temperature on the cyanide recovery for GFMA process

167

Cyanide recovery increases when temperature increases, since the diffusion coefficient of HCN in
water becomes higher, which involves a decreasing of the overall mass transfer resistance in the
process.
The mean value of cyanide transfer flux in the above results was close to 0.01 kg m-2 h-1 at 10 min of
operating time.
Scoping study
The economical evaluation developed in this study is based on the design of each cyanide recovery
process to estimate capital and operational costs. The unitary costs of reagents, labour,
maintenance, construction and equipment acquisition have been obtained from recent data of
Chilean projects. The NaCN price was US$2200/ton NaCN, the discount rate was 5% and the life of
the project was 10 years. The summary of results is shown in Table 2.
Table 2 Results of comparative economical evaluation
Description

GFMA

SART

AVR

1.32

1.22

1.45

Capital Cost, kUS$/(m /h)

56.5

90.0

69.8

NPV, kUS$/(ton NaCN/y)

3.65

4.24

1.82

Operational Cost, kUS$/ton NaCN

3

The GFMA process is an attractive alternative to the current cyanide recovery processes, obtaining
a net present value (NPV) better than AVR process. The SART process shows the highest NPV,
because of the incomes by sulphide precipitates and the lowest operational costs. Thus, the metal
contents in the cyanide solution and the solutions circuit of the plant will be relevant to determine
the best option to install a cyanide recovery process.

CONCLUSION
A new technique to recover cyanide in gold mining has been developed in this work, using gas
filled membrane absorption process. This new process reaches cyanide recoveries higher than 90%
at 10 minutes of operation. The effect of copper content in solution and pH are relevant on the
performance of the process. The NPV of GFMA process for a plant ranging 150-300 m3/h capacity is
3.65 kUS$/(ton NaCN/y), a 50% higher than AVR process and 16% lower than SART process.
Hence, this new process is an attractive alternative to recover cyanide in gold mining. Moreover,
the GFMA process is more safety, because the HCN is retained in the membranes pores, avoiding
leaks of gaseous HCN to the environment.

ACKNOWLEDGEMENTS
This work was supported by the projects FONDECYT 1100305 (CONICYT Chile) and ECOSCONICYT C10E05. The support of ARCADIS Chile to Mr. Humberto Estay is gratefully
acknowledged.

168

REFERENCES
Estay, H., Arriagada, F. & Bustos, S. (2010) Design, Development and Challenges of the SART Process,
Proceedings of the 3rd International Workshop on Process Hydrometallurgy, Santiago, Chile, 11-13 August,
Gecamin, Chile, pp. 3031.
Estay, H., Hasanoglu, A. & Romero, J. (2011) Gas-Filled Membrane Absorption: A Novel Process to Recover
Cyanide in Gold Cyanidation, Proceedings of the XXV Interamerican Congress of Chemical Engineering,
Santiago, Chile, 14-17 November, Chile.
Estay, H., Ortiz, M. & Romero J. (2013). A novel process based on gas filled membrane absorption to recover
cyanide in gold mining. Hydrometallurgy, vol. 134-135, pp. 166-176.
Fleming, C. (2005) Cyanide Recovery, Developments in Minerals Processing, vol. 15, pp. 703-727.
Kenfield, C., Qin, R., Semmens, M. & Cussler, E. (1988) Cyanide Recovery across Hollow Fiber Gas
Membranes, Environmental Science Technology, vol. 22, No. 10, pp. 1151-1155.
MacPhail, P.K., Fleming, C. & Sarbutt, K. (1998) Cyanide Recovery by the SART Process for the Lobo-Marte
Project, Chile, paper presented to the Randol Gold and Silver Forum, Denver, 26-29 April.
Shen, Z., Han, B. & Wickramasinghe, S. (2006) Cyanide removal from industrial praziquantel wastewater
using integrated coagulation-gas-filled membrane absorption, Desalination, vol. 195, pp. 40-50.
Smith, A. & Mudder, T. (1991). The Chemistry and Treatment of Cyanidation Wastes, Mining Journal Books
Limited, London.

169

Leaching of sulfide gold ores using

dithiooxamide, a non-cyanide
solvent of low-toxicity

Cyanide is the most used leaching product for gold recovery due to its effectiveness and low cost. Nevertheless, the environmental problems generated
by this material have motivated the study of new non-cyanide leaching agents

Juan Pablo Serrano and Ernesto de

la Torre. DEMEX, Escuela
Politcnica Nacional, Ecuador

that could form stable complexes with gold and that were less toxic. In basic
medium, dithiooxamide (DTO), C2S2N2H4, forms complexes with gold of the
form A [Au (C2S2N2H2)2], where A is Li, Na or K. In acid medium, it forms the
structure [Au (C2S2N2H3) X2] where X is Cl or Br. This research presents the
results of dithiooxamide (DTO) leaching of Au, Ag, Cu and Fe made with a
sulfide ore from the Ecuadorian mine named Agro-corazn, which contains
10.9g/t Au, 87% quartz, 12% mineral clay, 1% sulfides (pyrite). The post-leaching
process, reactions formulation and the comparison between the toxicity of the
dithiooxamide and sodium cyanide (LC50 in g/L) are also shown. The influence
of solids percentage, dithiooxamide concentration, particle size and oxidizing
agents such as air, potassium permanganate and hydrogen peroxide were
studied as well. Leaching with 10% solids, 0.9g/L DTO and a potassium permanganate of 74m obtained gold recovery of 85%, which is similar to the
one obtained by the cyanidation of this mineral. The low environmental toxicity
of dithiooxamide, despite its current high cost, allows wide perspectives for
the industrial application of this process to be considered.

171

INTRODUCTION
The processing of gold bearing ores with cyanide leaching is a widely used technique because
cyanide forms stable complexes with gold in alkaline mediums and its cost is low. Nevertheless, the
environmental problems associated with cyanide toxicity have motivated the research of new noncyanide leaching agents with lower toxicity (Senanayake, 2004).
Gold is a stable precious metal that does not react with most chemicals. It is only reactive with
chlorine and acid mixture HCl and HNO3. As shown in table 1, the dissolution of gold requires a
complexing agent and an oxidizing medium.
Table 1 Non-cyanide oxidants and leaching for gold recovery processes (Sparrow, Woodcock, 1995)
Oxidants
NH+4
H2O o OHH2SO3
O2 o H2O2
X2, X3 o OXFe(III), Cu(II)

Complexing agents
NH3
OHSO2-3
Cl-, OH-, S2O2-3, HSO-3, NH3
X- (X=Cl, Br, I, SCN)
Cl-, S2O2-3, SCN-, SC(NH2)2, NH2(CH2)COO-, NH2CH(CH3)COO-

The Au (III) complexes in basic medium form the [Au(C2S2N2H2)2]- anion, due to the strong
attractions between sulfur and gold and nitrogen and gold. Figure 1 shows the most probable
complex that the dithiooxamide with gold would form using NaOH as a pH regulator. The positive
charge ion (in this case Na+) is located very close to the nitrogen atoms where it destabilizes its
polar resonance form. This destabilization is larger when the positive charge ion decreases in size,
favoring a form of non-polar resonance that has a weaker C-N bonding force (Slootmaekers, 1996).

Figure 1 The most probable gold-dithiooxamide complex structure in alkaline medium with NaOH.

The movement in the electronic distribution of the thioamide function for the counter ion is
confirmed by the opposite movement of the bond C-S, whose position is linked with the position of
the bond C-N due to the resonance structure of the thioamide function, as shown in Figure 2.
Due to the "symbiosis" effect, the anion [Au(C2S2N2H2)2]- is more stable because of the larger counter
ion (positive charge). The stabilization of the complex causes the Au-S bond to become stronger,
resulting in a weakening of the C-S bond and, due to resonance structure of the thioamide function;
in an increase of the C-N bond strength, so that an Au-S bond is dative to the sulfur.

172

Figure 2 The resonance forms of the thioamide function

The most probable reaction that describes dissolution of gold in basic medium, with the golddithiooxamide complex data proposed by Slootmaekers, would be:

+4

+2

+5

(1)

The form of the gold-dithiooxamide complex in acid medium differs from the form of such complex
in basic media, due to the strong bonds between gold and halogenated compounds such as bromine
and chlorine. The dithiooxamide gold complex is formed only with one dithiooxamide molecule
which has a S-cis conformation with an intra-molecular hydrogen bond NH-N-. This effect is given
for the NH NH position, present in all complexes, which is independent from the position of
halogen ions. This means that the hydrogen of the thioamide is not exactly located but it can bond
with a nitrogen atom, resulting that this atom overpasses the S-Au bond forming a dative bond.
Figure 4 shows the tautomeric forms of the single dithiooxamide deprotonated. This molecule is
characterized by the union of the gold atom with two halogen atoms and two sulfur atoms. Only
one hydrogen atom of the amide group is deprotonated and joined by covalent bond to the
nitrogen and the subsequent dative bond. The complex formed using an acid medium with HCl as
pH adjuster is shown in Figure 5.

Figure 3 The tautomeric forms of the simple deprotonated dithiooxamide

173

Figure 4 Structure of dithiooxamide-gold complex in acid medium, adjusting the pH with HCl

The most probable reaction that describes dissolution of gold in acid medium, with the golddithiooxamide complex data proposed by Slootmaekers, would be:

+2

(2)

The information about the toxicity of dithiooxamide is limited. Its LD50 in rats is 500 mg/kg, which
makes the poisoning by accidental ingestion unlikely. The dithiooxamide, used to control
nematodes in crops, modifies plant growth (it was an herbicide that reduced the width of the
plants) and combats Lepidoptera insect species, these facts lead us to think that it is a manageable
toxic (Luckenbaugh, 1961, Gerhard, 1972). The limited solubility of the dithiooxamide in water and
the fact that it does not form gaseous complexes in environmental conditions restricts the potential
of accidents that cause health damage (Burger, 2011).
The dithiooxamide is not soluble in water, so it has been considered as unsuitable for leaching
precious metals. However, the dithiooxamide is deprotonated at pH>12.5, increasing significantly
its solubility, achieving gold dissolutions above 90% for an ore of Noligwa that has 7.3 g/ton gold
essentially free with a small amount of silver (Burger, 2011.).
In this context, the objective of this study is to evaluate the efficiency of the dissolution of gold with
dithiooxamide in the sulfide mineral.

METHODOLOGY
Ore characterization Agro-corazn
The ore characterization was performed by fire essays and acid disintegration, analyzing the
samples by atomic absorption. Also, the mineralogical compounds that form the ore were
determined by X-ray diffraction. The Agro-corazn ore with d100 = 2mm, was reduced by a ball mill,
working with 62.5% water for 37 minutes to obtain a d80 of 74 m.
Agro-corazn ore leaching with dithiooxamide as leached
The gold recovery with dithiooxamide as leaching is very similar to the one performed with
cyanide at environmental temperature (19 C), at a pressure of 0.71 atm (Quito environmental
conditions). In a stirred reactor, sodium hydroxide (4M NaOH) was added to adjust the pH of the

174

solution between 12.5 and 13.0. An air flow rate of 3 NL/min was injected to keep the solution
saturated with air.
The dithiooxamide concentration was determined by titration with silver nitrate, compound which
forms a black precipitate with dithiooxamide, easy to identify. For a silver nitrate concentration of
0.1 g/L, a relation of 1 cm3 AgNO3 corresponds to 200 g DTO/m3.
To obtain the most favorable conditions for the gold recovery, various parameters such as solids
percentage (between 10-20 %) and dithiooxamide concentrations (0.7-1.1 g/L) were tested, while
maintaining the size of mineral particle constant. To obtain higher gold recoveries, oxidants such as
sodium permanganate and hydrogen peroxide were added in different tests.
Due to the fact that the dithiooxamide does not form gases at any pH, leaching tests in acid medium
(pH = 1.5) with hydrochloric acid were carried out. (Burger, 2012).
Also, tests for gold cementation with zinc powder at different zinc concentrations between 3-6 g/L
and different pHs (11-12) with lead acetate 0.2 g/L were done.
The toxicity bioassay with lettuce seed is a static test of acute toxicity (120 h exposure) on which
you can evaluate the phytotoxic effects of cyanide and ditioxamide at different concentrations in the
process of germination of seeds that serves us to determine the LC50 of both compounds. Seeds
must have a germination of at least 90% and should know their sensitivity against a toxic standard
(in this case ZnSO4). For determine graphically the LC50 was should perform the lethality curve vs
toxic concentration of which was must prepare at least 5 concentrations of each toxic by diluting
each of them with reconstituted hard water (APHA, 1992) that are between 10 to 90% lethality of
the content of lettuce seeds (Castillo, 2004).

RESULTS AND DISCUSSION

The concentration of the metals that constitute the Agro-corazn ore are the following: gold 10.9
g/ton, silver 25.98 g/ton, magnesium 55 g/ton, iron 0,91%, sodium 2,58%, potassium 0,95%, copper
0,05%, cobalt 0,01.
The Agro-corazn ore has minerals with 87% quartz, 12% clay and 1% sulphides (pyrite). There are
3 kinds of mineral clays: poligoskita 5%, muscovite with 5% and clinochlore 2%.
Gold recovery from the Agro-corazn sulphide ore using ditioxamide was done in both alkaline
and acid mediums. These results are presented in figure 6, where it can be observed that gold
recovery in acid medium was 30% after 75 hours of leaching with ditioxamide, while, in basic
medium, the gold recovery was 82%.
The recovery in an acid medium is low even knowing that, in this conditions, dithiooxamide with
gold form complexes with the form [Au(C2S2N2H3)X2], using only one molecule of dithiooxamide,
in comparison to the complex formed in alkaline medium where 2 dithiooxamide molecules link
one molecule of gold.

175

Figure 5 Kinetic of gold recovery with dithiooxamide (10% solids, 0,9 g/l, pH 13, with air as oxidant) in
alkaline medium (with NaOH 4M as pH adjuster) and acid medium (with HCl as pH adjuster) against time

The formation of the gold-dithiooxamide complex is favored with the presence of a strong oxidant
in leaching because gold recovery is strongly bound to electrochemical potential (Burger, 2011). As
it can be seen in Figure 7, the gold recovery comes to 85% with the potassium permanganate (which
is the strongest oxidant) in just 50 hours, while leaching in the same conditions without any oxidant
reached recoveries of just 72% in 75 hours.
Air was also used as oxidant, obtaining gold recoveries of 82% in 75 hours, 13% more than not
using air. With hydrogen peroxide, the obtained results were similar to the ones found leaching the
ore just with agitation, probably because the H2O2 decomposes dithiooxamide.
Therefore, at 50 hours, gold recovery of 85% was obtained with potassium permanganate, 68% with
air, 58% with hydrogen peroxide and 58% with just agitation.

Figure 6 Kinetic of gold recovery with dithiooxamide, 10% solids, 0,9 g/L, pH 13, with air 3 NL/min,
potassium permanganate 0,38 [N], just agitation and H2O2 al 10%

Figure 9 shows the gold, silver, copper and iron recovery as a function of the leaching time. It can
be observed that gold has a dissolution rate higher than copper, while iron and silver have a small
dissolution. At 50 hours of leaching, gold recovery was over 90%, while silver recovery was less
than 2% because the dithiooxamide-silver complex is formed at lower pH. The copper recovery was

176

about 68%, very important percentage due to the concentration in this mineral (0.05%). The copperdithiooxamide form very stable complexes. The iron recovery is <0.5%, despite the high
concentration of this metal (0.09%).

Figure 7 Kinetics of gold, silver, copper and iron recovery with dithiooxamide (15% solids, 0.9 g/L, pH 13)

The sodium cyanide leaching was performed at the best conditions (Jurado, 2008) for this ore
obtaining gold recoveries of 85%. The dithiooxamide leaching, in the first 7 hours with the best
conditions, reached recoveries of only 22%, compared to the 62% recovery in the cyanidation due to
the large amount of complexes that copper forms with the dithiooxamide, as showed in Figure 10.
However, both leaching reached the 85% of gold recovery.

Figure 8 Kinetics of gold recovery with cyanide (CN-) 33% solids, 1 g/L, pH 10,5; with air as oxidant and with
dithiooxamide (DTO) 10% solids, 0.9 g/L, pH 13, with KMnO4 as oxidant

For to study of the toxicity, lettuce seeds with a germination of 90% were used (18 seeds lived of 20
tested) and a sensitivity to toxic standard (zinc sulfate) LC50 of 3.4 g/L was found. Figure 11 shows
that, the LC50 of lettuce seeds is 0.08 g/L with sodium cyanide which is more than 12.5 times lower
than the amount used in the process of gold recovery. In contrast, the LC50 of this kind of lettuce

177

seed for dithiooxamide is 3.29 g/L, which is 3.5 times less than the LC50, which leads to observe that
is acceptable option for the environment.
The cementation with zinc powder was performed with a dithiooxamide-gold complex enriched
solution, 3 g/L of zinc powder, pH range between 11.5 to 12 and 0.2 g/L acetate lead as operating
conditions. Larger recoveries than 3% could not be obtained. For this reason, the amount of zinc
was increased to 5 and 6 g/L, obtaining recoveries between 10 to 12% gold.

Figure 9 Dithiooxamide and cyanide toxicity (pH=12, NaOH 2M adjuster)

CONCLUSIONS
The gold recovery of the Agro-corazn sulphide ore by leaching with dithiooxamide is possible in
alkaline medium obtaining 82% of gold dissolution. In acid medium, just a 30% recovery was
reached. The gold recovery increased in alkaline medium using strong oxidants such as potassium
permanganate obtaining 85% in 50 hours, while with air, the same recovery was accomplished in 75
hours of leaching. With hydrogen peroxide, only 70% of gold was recovered due to the
decomposition of the dithiooxamide.
The gold recovery with dithiooxamide in alkaline medium was larger than other metals above 90%.
Only copper achieved a recovery above 60%. Silver and irons recoveries were below 2%.
Although the gold recovery with cyanidation and dithiooxamide were similar, the cyanidation has
higher kinetics and it can operate at higher solids percentages. The dithiooxamide is 41 times less
toxic than cyanide and can work through all the pH ranges.
The recovery of the gold-dithiooxamide complex in zinc powder should be studied more carefully
in further researches.

REFERENCES
APHA, 1992
Burger, M., Kunene, M., Padayachee, D. & Phala, N. (2011) Aqueous leaching process for recovery of precious metals
with addition of dithiooxamide ligand, anglo operations limited., viewed 01 april 2012,

178

<http://www.ipaustralia.com.au/applicant/anglo-operationslimited/patents/AU2010274628/#.UCiERqEf6xU>
Burger, M., Smith, J. & Schmuhl, J. (2011) Effective leaching of gold ores using a novel low-toxicity lixiviante,
paper presented to 9no. International Conference on Clean Technologies for the Mining Industry, Santiago,
pp. 3-10.
Castillo, G. (2004) Ensayos toxicolgicos y mtodos de evaluacin de calidad de agua: Estandarizacin,
intercalibracin, resultados y aplicaciones. Instituto Mexicano de Tecnologa del Agua, Cuernavaca,
Mxico.
Fisher, J. & Fresno, C. (1978) Method for modifying the growth characteristic of plants, Patent US2837417.
Gerhard, H. & St. Louis, M. (1972) Method of combating chewing insect larvae species of the order
Lepidoptera using N,N-Dicyclopropyl dithiooxamide, Patent US3639627.
Senanayake, G. (2004) Gold leaching in non-cyanide lixiviant systems: critical issues on fundamentals and
applications, Proceeding in the department of Extractive Metallurgy and Mineral Science, Murdoch
University, Georgia, pp. 1297-1305
Slootmaekers, B., Manessi-Zoupa, E., Perlepes, S. & Desseyn, H. (1996) The infrared spectra of complexes with
planar dithiooxamides, Department of chemistry. RUCA, Groenenborgerlaan 171.
Sparrow, G. & Woodcock, J. (1995) Cyanide and other lixiviante leaching systems for gold with some practical
applications. Mineral Processing and Extractive Metallurgy Reviews 14, 193-247.

179

Copper cyanocomplexes adsorption

on activated carbon: Effects on the
selectivity of gold dicyanide
adsorption

The treatment of cyanide solutions with a high copper concentration may

cause negative impacts on the gold adsorption on the activated carbon, since
copper cyanocomplexes have been recognized as important competitors for
the reactive sites of this material. While the adsorption of Au(CN)2- species
has been quite investigated, copper cyanocomplex adsorption has received

Clauson Souza, Virginia Ciminelli

and Daniel Majuste. Universidade
Federal de Minas Gerais and the
National Institute of Science and
Technology on Mineral Resources,
Water and Biodiversity
(INCT Acqua), Brazil

much less attention. Therefore, the purpose of the present work was to investigate the effects of typical parameters, such as aeration, excess free cyanide
and calcium ion concentration, on the copper speciation and selectivity of Au
and Cu adsorption on sample used industrially. The adsorption experiments
were carried out with cyanide solutions at pH 10.5, for 24 hours. The Raman
analysis indicated that under practical conditions of Au adsorption, the main
competitors for the reactive sites of this material are likely the Cu(CN)32- and
Cu(CN)43- species. The Cu(CN)2- species may compete with the Au(CN)2- species
at eventual low concentration of free cyanide in the solution. The point of zero
charge (PZC) value of the activated carbon is 7.1, thus, at pH 10.5, the surface
of this sample has a net negative charge, according to the deprotonation of
acidic and basic groups identified on the sample. The adsorption experiments
revealed that (i) the addition of excess free cyanide to the main solution inhibited
Au adsorption and it caused a large negative impact on Cu adsorption, thereby
increasing selectivity; (ii) the air bubbling into the main solution caused a
positive effect on Cu adsorption and negative effect on Au adsorption; and (iii)
the addition of calcium ion to the main solution caused a positive effect on Au
adsorption and negative effect on Cu adsorption. These experimental trends
were discussed on the basis of copper speciation, surface charge of the adsorbent
material, and electrostatic phenomena.

181

INTRODUCTION
The treatment of cyanide solutions with a high copper concentration may cause negative impacts
on the gold adsorption on the activated carbon (Fleming & Nicol, 1984; Muir et al., 1989). The
copper cyanocomplexes have been recognized as competitors for the reactive sites of this adsorbent
material and, in addition, the presence of these complexes in the solution may affect the loadingelution cycles, by increasing reagent consumption during elution. Thus, a successful application of
the conventional methods for treating low-grade, complexes gold ores, which include the Carbon in
Pulp (CIP) and Carbon in Leaching (CIL) processes, requires a high selectivity for gold adsorption
from solutions containing copper.
While the adsorption of dicyanoaurate (I) (Au(CN)2-) species on the activated carbon has been quite
investigated, the adsorption of copper cyanocomplexes (i.e., Cu(CN)2-, Cu(CN)32-, and Cu(CN)43-),
has received much less attention, and the effects of practical, operating parameters on the selectivity
of Au and Cu adsorption have been overlooked. It has been demonstrated that the adsorption of Cu
complexes on the activated carbon is strongly affected by metal speciation, which depends on the
solution pH and free cyanide concentration. The following adsorption order, Cu(CN)2- > Cu(CN)32> Cu(CN)43-, has been proposed (Fleming & Nicol, 1984). Therefore, the Cu(CN)2- species has been
claimed as the main competitor for the reactive sites of the activated carbon, but this conclusion
possibly took into account the similarities between this species and the Au(CN)2-. Thus, in practice,
the cyanidation conditions are adjusted in order to favor the formation of the Cu(CN)32- and
Cu(CN)43- complexes. It has been found that these ions are more stable at high pH and free cyanide
concentration, but under these conditions Au adsorption becomes unfavorable, as the CN- ions also
compete for the reactive sites.
In spite of the contributions that the previous studies provided, it is important to highlight that the
interaction mechanisms among the copper complexes and the activated carbon remain unclear. A
detailed evaluation of the effects of operating parameters and physicochemical properties of this
material has not been considered in these publications. In this context, the purpose of this work was
to investigate the effects of typical parameters on copper speciation and selectivity of Au and Cu
adsorption on the activated carbon. The parameters of particular interest are aeration, excess free
cyanide and calcium ion concentration. The role of Ca2+ ion in the preferential Au adsorption has
practical implication; lime (CaO) is commonly added to the industrial solution for pH control. Also,
there are evidences that Cu adsorption is enhanced from solutions containing Ca2+ ion (Souza et al.,
2013). The findings of the present paper may be helpful for developing best adsorption practices,
with minimal constraints caused by the presence of copper complexes in the industrial solution.

EXPERIMENTAL
Raman analysis
The analysis of gold and copper cyanocomplexes in the solutions was carried out by using Raman
spectroscopy, using a Horiba Jobin Yvon (Labram HR 800) spectrograph, equipped with a 633 nm
He-Ne laser (60 mW total power, 10 mW incident beam power). The Raman signal was collected by
an Olympus (BHX) Microscope equipped with lenses of 10, 50 and 100X. The spectral resolution

182

was about 2 cm-1, and a minimum of 10 scans with 30 s integration time were recorded. The grating
angle was calibrated using the 520 cm-1 silicon band.

Characterization of the activated carbon samples

In this work, only the activated carbon sample produced from coconut shell (CS) was investigated.
This sample is used for industrial adsorption of gold from cyanide solutions. Prior to the adsorption
experiments, the sample was thoroughly rinsed with 0.1 mol.L-1 sodium hydroxide (NaOH, Synth
99%) solution in order to remove humic acid residues. Following, the sample was thoroughly
rinsed with Milli-Q water (Millipore) and, then, dried in a furnace at 60 C for 24 h. The sample CS
was characterized by the specific surface area (SSA); ash content (A); density of surface functional
groups (DSFG); and point of zero charge (PZC). The SSA was determined by the Brunauer, Emmett,
and Teller (BET) method (Quantachrome, NovaWin2), while A was calculated by the weight loss of
the material, after heating in a porcelain crucible at 900 C for 1 h. The DSFG of the sample CS, as
well the identification of the corresponding groups, was determined by the Boehm method (Boehm,
1994). The PZC was obtained according to the procedure described by Strelko & Malik (2002).

Adsorption experiments
The adsorption data were obtained at 251 C from solutions prepared by dissolving potassium
dicyanoaurate (I) (KAu(CN)2, Aldrich 99%), copper (I) cyanide (CuCN, Aldrich 99%) and potassium
cyanide (KCN, Aldrich 96%) in Milli-Q water. The initial metal concentrations were 250 mg.L-1 Au
and 250 mg.L-1 Cu. The amount of KCN added to the solution depended on the CN/Cu molar ratio
to be investigated. The pH value was adjusted to 10.5, which characterizes practical conditions, by
adding 0.1 N potassium hydroxide (KOH, Synth 99%) solution. The effect of the ionic strength was
investigated by adding Ca2+ ion in the cyanide solution, as calcium chloride dihydrate (CaCl2.2H2O,
Synth 99%). Regarding the experimental procedure, 1 g of solid was added to 100 mL of solution.
This heterogeneous mixture was kept under stirred conditions (200 rpm) in a shaker for 24 h. Air
was added to the solution by using a pump (Whisper 30, flow capacity of 567 L.h-1) connected to a
glass tube (1.5 cm diameter x 3.2 cm height). Afterward, the mixture was vacuum filtered, and the
metal concentrations were analyzed by means of inductively coupled plasma optical emission
spectrometry (ICP-OES) using a Perkin Elmer (Optima 7300 DV) spectrophotometer. Au and Cu
adsorption densities (AD), in mmol.g-1, were calculated by mass balance in the aqueous phase, as
given by:
AD = ([M]i - [M]f).V/m

(1)

where [M]i represents the Au and Cu concentrations in the feed solution (mmol.L-1), [M]f the Au and
Cu concentrations in the filtrated solution (mmol.L-1), V the solution volume (L), and m the mass of
the solid (g). The selectivity of Au and Cu adsorption on the activated carbon sample was evaluated
by the separation factor (), as given by:
= ([Au]ac.[Cu]aq)/([Au]aq.[Cu]ac)

(2)

183

where [ ] represents the metal concentrations (mmol.L-1 in the aqueous phase, mmol.g-1 in the solid
phase), and the subscripts ac and aq mean activated carbon and aqueous phase, respectively. The
variable correlates the amount of metals adsorbed on the material with the remaining amount of
metals in the solution.

RESULTS AND DISCUSSION

Raman findings
The Raman spectrum of the KAu(CN)2 solution indicated only one intense band at 2162 cm-1, which
corresponds to the vibration mode of the Au(CN)2- species (Jia et al., 1998). The spectra obtained
from solutions at varied CN/Cu molar ratios are shown in Figure 1. At CN/Cu = 2.5, the analysis
indicated bands at 2106 cm-1 and 2135 cm-1, and shoulders at lower shift. The band at 2106 cm-1
represents a symmetric vibration of the Cu(CN)32- species, while that at 2135 cm-1 is diagnostic of
the Cu(CN)2- species. At CN/Cu = 3, the analysis indicated an intense band at 2108 cm-1, and band
with relative low intensity at 2094 cm-1. The band at 2094 cm-1 is assigned to a symmetric vibration
of the Cu(CN)43- species and asymmetric vibration of the Cu(CN)32- species (Lukey et al., 1999). At
CN/Cu = 4, bands at 2078, 2094 and 2108 cm-1 were detected. The band at 2078 cm-1 was assigned to
free CN-, which is stable at pH 10.5, and to an asymmetric vibration of the Cu(CN)43- species (Lukey
et al., 1999). At CN/Cu = 7, it can be noticed the same bands detected at CN/Cu = 4, but the relative
intensity of the band at 2108 cm-1 decreases, while a slight increase in the intensity of the band at
2078 cm-1 was observed. Therefore, the stability of the Cu(CN)43- species progressively increases
when the CN/Cu molar ratio increases, while the stability of the Cu(CN)32- species decreases, as
evidenced by speciation curves reported elsewhere (Souza et al., 2013). From solutions containing
Ca2+ ion, it can be seen that the formation of the Cu(CN)43- species was favored (Fig. 1). In
comparison with the spectrum obtained in the absence of this cation, a decrease in the relative
intensity of the band at 2108 cm-1 and a slight increase in the intensity of the bands at 2078 and 2094
cm-1 were observed. The spectrum obtained from aerated solution revealed in turn a slight increase
in the intensity of the band at 2108 cm-1 (Fig. 1), indicating a preferential formation of the Cu(CN)32species. In summary, the Raman analysis indicate that under practical conditions of gold
adsorption using activated carbon (i.e., pH 10.5 and CN/Cu about 4), the main competitors for the
reactive sites of this material are likely the Cu(CN)32- and Cu(CN)43- species. The Cu(CN)2- species
may compete with Au at eventual low concentration of free cyanide in the process solution.

184

Figure 1 Raman analyses: Effects of CN/Cu molar ratio, calcium ion and air on copper speciation
([Cu] = 0.004 mol.L-1; pH 10.5; 25 C)

Characterization of the activated carbon samples

The physicochemical properties (SSA, A, total and specific DSFG, and PZC) of the sample CS are
shown in Table 1. It can be seen that the sample has a large SSA value, which confers to the material
a high adsorption capacity per unit mass, and a small ash content, A value (4%). A low total density
of surface functional groups (1.2 eq.m-2) was also determined. This low number of heteroatoms on
the surface of this material results in a more homogeneous and hydrophobic surface, mainly
dominated by basal planes. The Boehm method revealed a relatively larger density for basic groups
than acid groups (i.e., carboxyl, lactones and phenol). The DSFG values reported here are likely
associated with the decomposition temperature of each functional group to carbon monoxide (CO)
and dioxide (CO2) during the activation process.
Table 1 Physicochemical properties of the investigated activated carbon sample.
SSA (m2.g-1)
931
Carboxyl (eq.m-2)
0.3

A (%)
4.0
Lactones (eq.m-2)
0.2

PZC
7.1
Phenol (eq.m-2)
-

DSFG (eq.m-2)
1.2
Basic groups (eq.m-2)
0.7

The PZC value of the activated carbon sample is 7.1 (Tab. 1). Thus, at pH below 7.1, the surface of
the material is positively charged, since the organic groups are protonated. At pH above 7.1, the
surface is negatively charged; the groups are deprotonated. It has been reported that the carboxyl
deprotonate in pH about 5 and lactones in pH about 7. The phenol and amine groups deprotonate,
respectively, in pH about 10 and 9-11, this value depending on the type of amine (Dean, 1999). On
the basis of the PZC value, at pH 10.5, the surface of the sample CS has a net negative charge,
according to the deprotonation of the carboxyl group, the neutral condition of the lactone group,
and the circum neutral character of the amine groups available on the surface of this material. Such
net negative charge may adversely affect the electrostatic interaction among the sample CS and the
negatively charged metal cyanocomplexes.

185

Effects of the process parameters on the adsorption densities of Au and Cu

Table 2 summarizes the AD values for Au and Cu after 24h. It includes results obtained from the
main solution (250 mg.L-1 Au, 250 mg.L-1 Cu, CN/Cu = 4, pH 10.5), main solution containing excess
of free cyanide (200 mg.L-1 CN-) or calcium ion (0.25 mol.L-1 CaCl2), and aerated main solution (567
L.h-1 Air). It can be seen that the excess of free cyanide in the main solution, which causes increase
in the CN/Cu molar ratio from 4 to 6, decreased the AD values for both metals. The AD value for
Au decreased about 16%, while the AD value for Cu about 72%. Regarding the addition of Ca2+ ion
to the main solution, the AD value for Au slightly increased (about 11%), while the AD value for Cu
decreased about 58%. From the aerated solution (567 L.h-1 Air), the adsorption of Au decreased, but
the adsorption of Cu largely increased. The AD value for Au decreased about 39%, while the AD
value for Cu increased about 4x.

Selectivity of Au and Cu adsorption

Figure 2 shows the variation of log as a function of time for adsorption experiments from the
main solution, main solution containing an excess of free cyanide or calcium ion, and aerated main
solution. A selective Au adsorption was found under all the investigated conditions, since > 1.
Table 2 Adsorption densities (mmol.g-1) of Au and Cu using the activated carbon sample CS

Main Solution
Au

Cu

0.05500.0004

0.00790.0001

Main Solution + CN-

Main Solution + Ca 2+

Main Solution + Air

Au

Au

Au

Cu

Cu

Cu

0.0460 0.0003 0.0022 0.0001 0.0608 0.0005 0.0033 0.0002 0.0336 + 0.0008 0.0323 0.0006

3,00
2,500
Ca2+

log

2,00
1,500

Free CN-

1,00

Main Solution

,500
Air

,00
0

200

400

600

800

1000

1200

1400

1600

Time (min)
Figure 2 Effects of the process parameters on the selectivity of Au and Cu adsorption (Sample CS)

186

From solution containing excess of free cyanide

As already reported, the addition of excess free cyanide to the main solution caused a significant
negative effect on Cu adsorption. It resulted in an increase in the value (Fig. 2) after 24 h, log
increased from 1.57 to 2.06. However, in spite of this higher selectivity, excess of free cyanide
inhibited the Au adsorption. Previous results indicated that in the main solution containing excess
of free cyanide (CN/Cu = 6), the Cu(CN)43- species predominates (Fig. 1), and that Cu adsorption on
the sample CS is favoured when the Cu(CN)32- species is more stable than the Cu(CN)43- species, a
condition typically found at relatively low CN/Cu molar ratios. The higher stability of the more
negatively charged complex favors the electrostatic repulsion among such species and the negative
charges available on the surface of the activated carbon. In addition to the larger charge of the
tetrahedral species, its geometry may also affect Cu adsorption. Regarding the inhibition of Au
adsorption when free CN- was added to the main solution, there are evidences that a competition
among the Au(CN)2- species and free CN- to the reactive sites of the adsorbent material takes place
(Nicol & Fleming, 1984). It can also be seen from Figure 2 that log decreased as a function of time,
which is probably related to this competition, as Cu adsorption is quite low (Tab. 2).

From aerated solution

The air bubbling into the main solution caused a positive effect on Cu adsorption, while a negative
impact on Au adsorption was demonstrated (Tab. 2). These behaviors affected directly the value
after 24 h, log decreased from 1.57 to 0.75 (Fig. 2). According to the Raman analysis carried out
here, aeration affects Cu speciation (Fig. 1). The spectrum obtained from the aerated main solution
indicated an increase in the relative intensity of the band associated with the Cu(CN)32- species. It
may be explained by the possible oxidation of coordinated cyanide in the Cu(CN)43- species (Eq. 3),
but the oxidation of the remaining coordinated cyanides were not confirmed by Raman analysis. In
addition, the activated carbon may catalyze the oxidation of free cyanide (Adams, 1990), causing a
diminution in the CN/Cu molar ratio, thereby favoring the stability of the Cu(CN)32- species.
Cu(CN)43- + O2 = Cu(CN)32- + CNO-

log K = 46.0

(3)

Thus, under aerated conditions, the predominance of the Cu(CN)32- species in the process solution,
in comparison to the Cu(CN)43- species, is beneficial to Cu adsorption, as earlier discussed here, and
due to the competition with the Au(CN)2- species the Au adsorption decreases (Tab. 2).

From solution containing calcium ion

The addition of 0.25 mol.L-1 Ca2+ ion to the main solution caused a positive effect on Au adsorption,
but a negative impact on Cu adsorption was determined (Tab. 2). These behaviors affected directly
the values after 24 h, log increased from 1.57 to 2.10 (Fig. 2). At pH 10.5, carboxyl groups
available on the surface of the sample CS are deprotonated and, therefore, the negative charges are
neutralized by positive charged Ca2+ ions. As already reported, the addition of Ca2+ ion to the main
solution caused a slight increase in the AD value for Au (about 11%) and decrease in the AD value
for Cu (about 58%) (Tab. 2). The increment in the amount of Au adsorbed is not very large, because
the Au(CN)2- species interacts preferentially with the graphene layers of the activated carbon by

187

van der Waals force (Jones et al., 1989), and its adsorption on the positive sites on the hydrophilic
surface of the material (i.e., R-COO- Ca2+) is probably secondary. Regarding the negative impact
on Cu adsorption, it could be explained by (i) a faster adsorption of the Au(CN)2- species, in
comparison with the adsorption of the Cu(CN)32- and mainly the Cu(CN)43- species initially, a
higher log value was observed (Fig. 2), which declines as a function of time due to an increasing,
slow Cu adsorption; and (ii) steric effects involved in the adsorption of the Cu(CN)43- species. The
Raman analysis demonstrated that Cu(CN)43- species predominates over Cu(CN)32- species if 0.25
mol.L-1 Ca2+ is added to the solution (Fig. 1).

CONCLUSIONS
In the present paper, the adsorption of the copper cyanocomplexes on the activated carbon and the
selectivity of gold adsorption were discussed by an approach that took into account the nature of
the interaction among stable copper species and reactive sites of the adsorbent material. The effects
of relevant parameters such as aeration, excess of free cyanide, and calcium ion concentration, on
metal speciation, properties of the solid phase, and selectivity of Au and Cu adsorption were
considered. Some findings that are useful to improve Au adsorption from Cu concentrated cyanide
solutions are:
(i) The flow rate of air into the cyanide solution may be optimized in order to minimize the
predominance of the Cu(CN)32- species, which adsorbs well on the activated carbon, and
competes with the Au(CN)2- species for the reactive sites;
(ii) The addition of Ca2+ ion to the cyanide solution (with CN/Cu molar ratio > 4) may be optimized
in order to maximize selectivity since it favours the stability of the Cu(CN)43- species, which is
poorly adsorbed on the activated carbon;
The results obtained from this investigation also confirmed that an excess of free cyanide in the
solution may increase the selectivity of Au adsorption, however, at relatively high CN/Cu molar
ratios, the excess of free cyanide may decrease the total amount of Au adsorbed on the activated
carbon.

ACKNOWLEDGEMENTS
The authors are grateful to CAPES, CNPq, FAPEMIG, and INCT-Acqua for financial support.

REFERENCES
Adams, M.D. (1990) The Chemical behaviour of cyanide in the extraction of gold, I. Kinetics of cyanide loss in
the presence and absence of activated carbon. Journal of South African Institute of Mining and
Metallurgy, Vol. 90, no. 2, pp. 37-44.
Boehm, H. P. (1994) Some apects of surface chemistry of carbon blacks and other carbon. Carbon, Vol. 32, pp.
759769.
Dean, J.A. (1999) Hanbook of Chemistry. MacGraw Hill. pp. 8.26- 8.57.

188

Fleming, C.A., & Nicol, M.J. (1984) The absorption of gold cyanide onto activated carbon. III -. Factors
influencing the rate of loading and the equilibrium capacity. Journal of South African Institute of Mining
and Metallurgy, Vol. 84, no. 4, pp. 8593.
Jia, Y.F., Steele, A.C., Hayward, I.P., & Thomas, K.M. (1998) Mechanism of adsorption of gold and silver
species on activated carbons. Carbon, Vol. 36, no. 9, pp. 12991308.
Jones, W.D., Klauber, C., & Linge, H.G. (1989) Fundamental aspects of gold cyanide adsorption on activated
carbon. World Gold '89, Littleton, SME, pp. 278-281.
Lukey, G.C., van Deventer, J.S.J., & Shallcross, D.C. (1999) Raman Study on the Speciation of Copper Cyanide
Complexes in Highly Saline Solutions. Hydrometallugy, Vol. 53, pp. 233244.
Muir, D.M., LaBrooy, S.R., & Cao, C. (1989) Recovery of gold from copper-bearing ores. Gold Forum on
Technology and Practices World Gold 89. SME, Littleton, pp. 363374.
Souza, C., Majuste, D., & Ciminelli, V.S.T. (2013) Effects of surface properties of activated carbon on the
adsorption mechanism of copper cyanocomplexes. Hydrometallurgy. (in press).
Strelko, V., & Malik, D.J. (2002) Characterization and metal sorptive properties of oxidized active carbon. J.
Colloid Interface Sci, Vol. 25, pp. 213220.

189

Cyanide leaching of
copper-gold-silver ores

Copper leaching is currently carried out in sulphuric acid media, where the
process is highly known and where metallurgical and economic performances
are mostly reliable. However, acid leaching is not able to recover the gold and

Humberto Estay, Pablo Carvajal,

Karina Gonzlez, Hctor Yaez,
Waldo Bustos, Sergio Castro and
Francisco Arriagada. Arcadis, Chile

silver contained in some copper ores and thus unable to capture their economic
benefits. In this work, the feasibility of using alkaline cyanide solutions for
leaching copper, gold, and silver from Cu/Au/Ag ores is analyzed and compared
with acid leaching. A key operation in the cyanide leaching circuit proposed
is a SART process (Sulphidization, Acidification, Recycling and Thickening),
which is being used presently at many gold plants to recover cyanide and to
remove and recover CN-soluble copper contained in the ore, thus reducing
cyanide consumption. Additionally, operational evidence exists in gold cyanidation plants showing that very low grade copper can be dissolved, which
supports the thesis that under certain conditions the cyanide process might
be able to dissolve lower copper contents, achieving lowers residual copper
in tails than in the acid leaching process. A technical and economical tradeoff analysis of the cyanide and acid leaching processes is presented in order
to establish under what conditions cyanide copper leaching is a competitive
alternative against copper acid leaching. The main inputs parameters used
in the evaluation are presented, seeking to identify which ore type (function
of copper mineralization and contents of Cu, Au, Ag) could be suitably treated
by this process. Final considerations about the cyanide process for coppergold ores are presented.

191

INTRODUCTION
This work addresses the option of alkaline, cyanide heap leaching for copper ores containing gold,
presenting a trade-off analysis between the conventional acidic route and the cyanidation process.
The main basis of this approach are a) the high solubility of a number of copper minerals in cyanide
solutions, b) more than 100 years of industrial experience in gold cyanidation processes, and c) the
current industrial experience of the SART (Sulphidization, Acidification, Recycling and Thickening)
process (MacPhail, Fleming & Sarbutt, 1998), which is able to recover copper and cyanide from
gold-copper cyanide solutions. The objective of the work is to establish under what ore conditions
cyanide copper leaching is a competitive alternative.
The feasibility of using alkaline cyanide solutions for leaching copper, gold, and silver from ores
has been well known at laboratory scale. It is known (Hedley & Tabachnik, 1968; Parkison &
Bhappu, 1995) that native copper, chalcocite, and most copper oxide, carbonate and hydrochloride
minerals dissolve readily in cyanide solutions. For example, leaching 24 h at 23oC and -150 m
results in copper dissolution from 85.5% (cuprite) to 94.5% (azurite). No report was found on
commercial extraction of copper from cyanide solutions, although different cyanide processes have
been proposed but none tested at any reasonable scale. For example, the Treadwell process
involving the precipitation of cuprous cyanide from an acidic cupric sulphate solution, or a further
development where the ore is leached with cyanide and copper is precipitated as copper sulphide
by acidification, similarly to the SART process (Shantz, 1976). The high solubility of these copper
minerals in cyanide solution is indeed observed in the industrial practice, as shown in Table 1 for
different heap leaching operations. Copper dissolution is in the range of 14 to 39% of the total
copper content (due to variable portions of CN-soluble copper), producing leaching tails with
residual copper as low as 0,028%. In these processes, the leaching conditions are selected to enhance
gold dissolution rather than copper.
Table 1 Au and Cu recovery in Heap Leach Projects
Parameter

Project 1

Project 2

Project 3

Project 4

80%-1

80%-3/4

100%-3/8

80%-1/2

Au content, g/t

0.4

0.67

1.51

0.61

Cu content, %

0.05

0.07

0.07

0.04

Au Recovery, %

65

55-70

65

81

Cu Recovery, %

20

14

39

30

0.040

0.058

0.042

0.028

Particle size

Cu tails, %

The copper-cyanide complexes formed will tend to inhibit gold dissolution, unless a given freeCN/Cu ratio can be maintained in the cyanide solution. Also, copper-cyanide complexes are
adsorbed in the ADR stage, which reduces gold adsorption efficiency and needs a separate cold
elusion stage, before gold can be efficiently eluted. Finally, the complexation of copper (or that of
other base metals) with cyanide increases the cyanide consumption in the leaching operation.
Current practice in heap leaching-ADR operations is to keep copper concentration in solution
below 300-400 mg/L (even below 100 mg/L) to avoid gold recovery inefficiencies and high

192

operational costs (Coderre & Dixon, 1999; Dai, Jeffrey & Breuer, 2010; Guzman, 2009; Marsden &
House, 2006; Sceresini, 2005; Sevilla, 2013).
There are several methods of copper removal from cyanide solutions, of which the SART
technology has received increasing interest from the industry. In processing copper-bearing gold
ores, the key application of this process is to recover copper and cyanide from the heap leaching
solutions, removing copper (and other cations) as a sulphide precipitate. The Telfer plant in
Australia pioneered the application of this technology (Barter et al., 2001) by running a 40 m3/h
SART plant. Nowadays, there are five new SART plants: 1400 m3/h in Yanacocha, Peru (Botz and
Acar, 2008; Guzman, 2009); 350 m3/h in Lluvia de Oro, Mexico; 140 m3/h in Gedabek, Azerbaijan
(Estay et al., 2012); 120 m3/h in Mastra, Turkey. Since December 2012 Kinross Gold Corporation is
operating a 750 m3/h SART plant at Maricunga in Chile.

METHODOLOGY
The trade-off analysis of the acid and cyanide leaching routes was conducted in this work at a
scoping level of engineering. The design criteria and cost data were taken from the Data Base of
Arcadis from gold, copper, and gold-copper ores leaching projects. The Case Study was defined,
and the flow diagram and process design criteria were selected for each process. Copper species
dissolution in acid or cyanide solutions was estimated from literature and/or Arcadis experience.
Equilibrium mass balances were completed to determine sizing and capacity of the main unit
operations in each circuit and to establish gold, silver, and copper productions, flows of irrigation
and PLS solutions, concentrations of Cu, Au, Ag, and reagent consumptions. Operational costs and
capital costs were estimated from the mass balance results and equipment sizing, by factoring
capacity on the basis of similar heap leaching projects. Finally, a sensitivity economic analysis was
conducted by varying Au grade and copper mineralogy in the Case Study.

RESULTS AND DISCUSSION

Processes description
Acid leaching is conducted in an on-off heap followed by a solvent extraction-electrowinning plant
(SX/EW) to obtain copper cathodes. Figure 1 shows a schematic block diagram of the process.
H2SO4
Evap

Ore

Crushing Plant
(1 , 2 & 3 )

Water Evap

Rf Pond

Heap Leach
Evap

PLS Pond

SX-EW Plant

Cathode

Figure 1 Schematic block diagram of acid leaching

193

The cyanide leaching uses an on-off heap, similar to that of acid leaching. On the other hand, the
circuit includes a SART plant to recover copper and an ADR (adsorption, desorption and recovery)
plant to recover gold, which includes a CIC (carbon in column) process. A schematic block diagram
of the cyanide leaching plant is shown in Figure 2.
NaCN

Water

Evap

Evap
Lime

Ore

Crushing Plant
(1 , 2 & 3 )

Rf Pond
Heap Leach
Dore
ADR Plant

Evap

Copper

PLS Pond
SART Plant

Gypsum
H2SO4
NaSH
Milk of Lime
NaOH

Figure 2 Schematic block diagram of cyanide leaching

Design criteria
The parameters used in the Case Study presented in Table 2 below are based on typical design
criteria of acid leaching operation and of SART plants, part of cyanide leaching circuits. The Case
Study selected considers that the copper minerals are distributed as 10% Atacamite, 35% Malachite,
45% Chalcocite and 10% Bornite.
Table 2 Parameters used for acid and cyanide leaching

194

Parameter

Acid Leaching

Cyanide Leaching

Processing rate, Mt/y

10

10

Copper grade,%

0.3

0.3

Gold grade, g/t

0.35

0.35

Silver grade, g/t

10

10

Copper recovery,%

70

80

Gold recovery, %

65

Silver recovery, %

40

P80, mm

13

13

Leach ratio, m3/t

1.5

1.5

Irrigation rate, L/(m2h)

10

10

Lift height

10

10

Acid consumption in leaching, kg/t

15

NaCN consumption, kg/t

1.0

CaO consumption in heap leaching, kg/t

3.0

Copper extraction in SX, %

90

Copper recovery in SART, %

90

Gold extraction in adsorption, %

95

Acid consumption in SART, kg/m3

4.88

NaHS stoichiometric addition in SART, %

120

CaO consumption in SART, kg/m3

2.79

Mass balance results

Steady state mass balances have been developed in this work for both circuits (acid and cyanide).
The main results are shown in Table 3 below. Copper is produced in acid leaching in the form of
cathodes, while the copper produced by cyanide leaching is in form of a sulphide precipitate
having 65% copper. Cyanide leaching produces 72,870 Oz/y of gold and involves high reagents
consumption (68 kt/y acid, 14 kt/y NaHS, 68.8 kt/y CaO and 15.8 kt/y NaCN). On the other hand,
the acid consumption of acid leaching is 145% higher than cyanide leaching. In addition, the
cyanide leaching includes SART and ADR plants of 1,590 m3/h capacity, while the acid leaching
circuit includes a 1,590 m3/h SX plant and a EW plant of 21 kt/a.
Table 3 Mass balance results for acid and cyanide leaching
Parameter

Acid Leaching

Copper production, kt/y

Cyanide Leaching

20.83

23.81

72,870

PLS flow, m3/h

1,587

1,587

[Cu] in PLS, g/L

1.67

1.90

[Au] in PLS, mg/L

0.172

[H2SO4] in PLS, g/L

2.38

4,068

Gold production, Oz/y

[NaCN] in PLS, mg/L

Gypsum production, t/y

125,943

167,311

67,859

NaCN consumption, t/y

15,751

Total CaO consumption, t/y

68,777

NaHS consumption, t/y

13,976

NaOH consumption, t/y

828

SART plant capacity, m /h

1,587

ADR plant capacity, m3/h

1,587

1,587

Total acid consumption, t/y

Solvent extraction plant capacity, m3/h

Operational and capital costs results

The operational and capital costs estimation has been developed on the basis of the mass balance
results and the sizing of unit operations. Also, reagents prices, installations and construction costs

195

are based on recent values for similar Chilean projects. The estimation has been performed using
1,500 US$/Oz gold, 2.7 US$/lb Cu and 25 US$/Oz Ag. In the case of the copper sulphide precipitate
formed in the SART process, the prices of copper and silver are taken as 70% of the LME grade
copper price and 80% of the pure silver price. Other costs used were 130 US$/t acid, 1,500 US$/t
NaHS, 150 US$/t CaO and 2,200 US$/t NaCN. Capital cost includes only plant costs as depicted in
the flow diagrams (mine and infrastructure costs are not included). The estimation of the total
operational and capital costs for the Case Study is presented in Table 4.
Table 4 Operational and capital costs results for acid and cyanide leaching
Parameter

Acid Leaching

Cyanide Leaching

Incomes, kUS$/y

123,992

234,222

Operational cost, kUS$/y

85,527

148,057

186

282

264,795

369,896

Operational cost, cUS$/lb Cu

Capital cost, kUS$

The cyanide leaching process produces higher income (234 MUS$) than the acid leaching process
(124 MUS$), due to the additional gold and silver production. On the other hand, the operational
costs of cyanide leaching (282 cUS$/lb) are higher than those of acid leaching (186 cUS$/lb), due to
the high consumption of reagents not used in acid leaching. Furthermore, the capital costs of
cyanide leaching are higher than acid leaching due to the SART plant and filtering and drying
stages. Also, the capital cost of the SART and ADR plants is higher than the capital cost of the SXEW plant to produce cathodes.

Economic evaluation
A project life of 10 years was assumed to evaluate the Case Study. The results obtained by using an
8% discount rate are shown in Table 5.
Table 5 Economic evaluation results
Parameter

Acid Leaching

Cyanide Leaching

NPV, kUS$

-6,696

208,277

The results indicate that cyanide leaching presents a NPV 215 MUS$ higher than that of acid
leaching, for the ore and processing conditions defined for the Case Study. A sensitivity analysis
was performed by keeping constant the total Cu grade and by varying the relative content of
copper sulphide in the ore (from 0% to 70%). Also, the gold grade was increased (0.15 g/t to 0.55
g/t), in order to further explore the economical feasible range for cyanide leaching. In the case of
variable copper sulphide content, the copper recovery was changed according to the mineralogical
distribution chosen. Figure 3 below presents shows the sensitivity results based on NPV differential
between cyanide and acid leaching.

196

80%

0.6
0.5

60%
0.4

50%
40%

0.3

30%

0.2

20%

Sulphide Content

10%

Gold Grade

0%
-200,000

Gold grade, g/t

Sulphide content, %

70%

0.1
0.0

200,000
400,000
Differential NPV, kUS$

600,000

Figure 3 Sensitivity analysis results (NPV, kUS$) for sulphide content and gold grade

According to the results shown in Figure 3, the NPV is mainly affected by gold grade, increasing
from -200 MUS$ to more than 600 MUS$ when gold grade increases from 0.15 g/t to 0.55 g/t. The
gold grade break-even is around 0.25 g/t for 0.3% CuT. On the other hand, the NPV rises from 30
MUS$ to about 230 MUS$ when the copper sulphide content increases from 0% to 70%,
demonstrating that the mineralogy is very relevant on the feasibility of cyanide leaching. The
increase in the oxidized copper content in the ore will favour acid leaching, since copper recovery
will increase in this media. On the other hand, the increase in the total copper grade (above about
0.40%) favours also the acid leaching, due to the high cost of reagents (acid and lime) and capital
cost of filters and dryers in the SART plant under these conditions of high copper load (results not
shown).

Final considerations
The cyanidation process for gold ores having copper contents is practiced today in industry. In
general terms, it can be stated that the proposed cyanidation process for copper-gold ores is not
conceptually different, although it will operate under certain conditions which are different from
the usual processing conditions and parameters of the former operations. In particular, it is
envisaged that the main concerns that the development and application of this process may face are
the following, as compared to a conventional acid leaching process:
1. High cyanide concentrations may have deleterious effects on carbon adsorption efficiency.
Also, gold and silver dissolution may be inhibited by the high copper concentration in the
solution. In addition, the high cyanide concentration may promote higher cyanide
consumptions.
2. High solid load in the SART plant may impact solid-liquid separation efficiency and
design. It may require large number of filtration and dewatering equipment, involving a
large foot-print area.
3. Difficult environmental approval from the Authority, and from the Community. High
cyanide concentrations in the circuit solutions may promote higher risks of infiltration in
liners and solution ponds.

197

4. The copper mineralogy may vary over the mine life to less CN-soluble species, which will
reduce copper recovery.
5. High capital costs are involved.

CONCLUSIONS
A trade-off study of heap acid leaching versus heap cyanidation leaching for copper-gold-silver
ores was conducted at a scoping level. The main conclusions are the following:
1.

For the ore of the Case Study containing 0.35 g/t Au, 10 g/t Ag, and 0.30% CuT (of mixed
oxides-sulphides mineralogy), the cyanidation process presents a differential NPV larger
than 215 MUS$ with respect to the acid leaching case, given that the metal recovery levels
are 65% Au, 40% Ag, and 80% Cu, the latter being 10 points of % above the expected copper
recovery in the acid leaching.

2.

Under these conditions, the expected copper content in the cyanide leaching tails would be
about 0.06%, a level of residual copper which has been readily achieved in gold-copper ores
leaching operations. For the majority of copper minerals, the cyanide leaching is most
effective in dissolving copper than the acid leaching (one exception is Chrysocolla).

3.

The results obtained suggest that there is an opportunity for the cyanidation of copper-gold
ores, which would deserve to be explored in a case by case basis. Mineralogical
characterization of copper species as well as gold and silver grades will be relevant.
Leaching testing will be mandatory to establish the suitable leaching process performance
in terms of metal recovery and reagents consumption.

REFERENCES
Barter, J., Lane, G., Mitchell, D., Kelson, R., Dunne, R., Trang, C. & Dreisinger, D. (2001) Cyanide management
by SART, Cyanide: Social, Industrial and Economical Aspects, Young, C. (Ed.), The Minerals, Metals &
Materials Society, pp. 549562.
Botz, M. & Acar, S. (2008) Copper Precipitation and Cyanide Recovery Pilot Testing for the Newmont
Yanacocha Project, paper presented to. SME Annual Meeting, Denver, February 25-28.
Coderre, F. & Dixon, D. (1999) Modeling the cyanide heap leaching of cupriferous gold ores. Part I:
Introduction and interpretation of laboratory column leaching data. Hydrometallurgy, vol. 52, pp. 151175.
Dai, X., Jeffrey, M. & Breuer, P. (2010) A mechanistic model of the equilibrium adsorption of copper cyanide
species onto activated carbon. Hydrometallurgy, vol. 101, pp. 99-107.
Estay, H., Carvajal, P., Hedjazi, F. & Van Zeller, T. (2012) The SART process experience in the Gedabek plant,
Proceedings of the 4rd International Workshop on Process Hydrometallurgy, Santiago, Chile, 12-13 July,
Gecamin, Chile, pp. 144155.
Guzman, G. (2009) Proceso SART y su influencia en el proceso CIC Minera Yanacocha, paper presented to the
S.R.L. 29 Convencin Minera, Perumin 2009, Arequipa, Per, September 1418.
Hedley, N., Tabachnik, H. (1968) Chemistry of Cyanidation, Mineral Dressing Notes No23, American
Cyanamid Company, New York.

198

MacPhail, P.K., Fleming, C. & Sarbutt, K. (1998) Cyanide Recovery by the SART Process for the Lobo-Marte
Project, Chile, paper presented to the Randol Gold and Silver Forum, Denver, 26-29 April.
Marsden, J. & House, C. (2006) The Chemistry of Gold Extraction, Second edition. Society for Mining,
Metallurgy & Exploration SME, Colorado, USA.
Parkison, G.A. & Bhappu, R.B. (1995) SME Annual Meeting, 1995: The Sequential Copper Analysis MethodGeological Mineralogical and Metallurgical Implication. Denver, Colorado, SME, March 6-9,.
Sceresini, B. (2005) Goldcopper ores. Devolepment of Minerals Processing, vol. 15, pp. 789824.
Sevilla, L. et al (2013) Yanacocha Gold Mill y el Proceso SART 2013-2017, paper presented to the 1st Arcadis
Workshop de Oro, Santiago, 20 March.
Shantz, R. (1976) Cyanide Leaching of Chalcosite. E&MJ Publishing.

199

Successful application of SART

technology for gold-copper
ore deposits

Cyanide is a widely used inexpensive and efficient chemical to extract gold

from ores, leaching the gold from crushed or run of mine ore. Impacting this
process is the presence of cyanide-soluble metals such as copper in the gold
deposit. Copper creates a metallurgical challenge by consuming the cyanide,
reducing the availability of cyanide for the leaching process. The result is

Kresimir Ljubetic and scar Lpez.

BioteQ Water Chile SpA
David Kratochvil and David
Sanguinetti. BioteQ Environmental
Technologies, Canada

increased total cyanide consumption that can render gold recovery uneconomic
and high concentrations of copper-cyanide complexes that represent a challenge for cyanide destruction.
The challenges presented by copper-complexed gold ores can be addressed
with SART (Sulphidization, Acidification, Recycling and Thickening) process
technology. When applied successfully, SART treats the leach solution to break
the copper-cyanide weak acid dissociable (WAD) complexes and in the process,
recover the copper as a high-grade sulfide concentrate and regenerate the
cyanide as free cyanide for recycle to the gold extraction process, eliminating
the downstream environmental liabilities and resulting in improved project
economics. This paper provides an overview of the operational results of a
SART plant BioteQ designed and reviews the key design elements required
for successful SART process control and performance.

201

INTRODUCTION
As easily mined gold ore bodies become depleted, focus is now turning to complex gold deposits
containing relatively large amounts of cyanide-soluble base metals such as copper. The presence of
cyanide-soluble base metals can cause metallurgical and environmental challenges. These base
metals complex with cyanide to form weak acid dissociable (WAD) complexes that increase cyanide
consumption and impact cyanide destruction, both of which increase project operating costs. Borne
out of a study to find an economic and environmentally acceptable solution to remove cyanidesoluble copper from leach solution, SGS Lakefield and Teck Corporation developed SART
(sulphidizationacidificationrecyclethickening) process technology in 1997 (Ford, Fleming &
Henderson, 2008). Since its development, several commercial SART plants have been built around
the world (Kratochvil, Chan & Hall, 2013). Some are small, some are sited in remote locations and
many have not operated on a continuous basis. As a result, very little has been published about
SART operations. BioteQ has been involved in the design, construction, commissioning and
operation of numerous commercial scale SART and sulphide precipitation plants. The knowledge
gained from the latter has contributed to BioteQs expertise in SART of which the sulphidation
circuit is a component.

SART technology
In the SART process, the cyanide solution containing copper-cyanide complexes is treated with
sulphuric acid to yield a pH of 4-5 to break the WAD cyanide-copper bond and convert cyanide to
HCN gas. A sulphide reagent is then added to precipitate the copper as a sulphide product (Cu2S).
The copper sulphide precipitate is thickened and recovered as a valuable, high-grade product,
which can grade up to 70% Cu. The solution is then neutralized to pH 10 or greater with lime,
converting the HCN to free cyanide for recycle back to the leach. Cyanide recycling allows the leach
circuit to operate at higher cyanide levels, maximizing leach efficiency and minimizing copper
deportment to gold electrowinning. Gypsum is formed in the lime addition (neutralization) stage
and recovered by thickening.
The sulphidation and acidification reaction is shown below:
2Na2Cu(CN)3(aq) + NaHS + 2.5H2SO4 Cu2S(s) + 6HCN + 2.5Na2SO4 (aq)
The recycling and thickening reaction is shown below:
6HCN + 2.5Na2SO4 (aq) + 3Ca(OH)2 2.5CaSO4.2H2O(s) + 5NaCN(aq) + 0.5Ca(CN)2(aq) + H2O
Application of SART technology allows complex gold ore bodies to be leached aggressively with
cyanide, thereby maximizing gold recovery without undue concern for copper cyanide formation.
A schematic of the process is provided in Figure 1.

202

Figure 1 The SART Process

Mastra mine SART plant

Located near Gmhane, Turkey, the Koza Gold owned Mastra Mine is an active gold mine with
both open pit and underground operations. Due to high concentrations of cyanide-soluble copper
in the deposit, the mine was stockpiling some of its ore. The copper interfered with the efficient
recovery of gold and resulted in copper-cyanide complexes in tailings that were at risk for
exceeding the 10 mg/L WAD cyanide discharge limit when copper levels in tailings were greater
than 150 mg/L (Lawerence & Lopez, 2011). Tailings copper concentrations were typically more than
800 mg/L.
In 2010, a SART plant with a 2,800 m3/day hydraulic capacity designed and commissioned by
BioteQ went into operation to treat barren leach solution (BLS) containing up to 1,500 mg/L of
dissolved copper. Pulp from the leach tanks is washed in a counter current decantation (CCD)
circuit, the overflow from the CCD goes to carbon-in-column (CIC) for gold extraction and then to
the SART plant for treatment. A large portion of the SART plant effluent containing regenerated
cyanide is recycled back to the leach while a small portion of the SART effluent is directed to
cyanide destruction and discharge to the tailings pond. Washed solids from the CCD go to carbonin-pulp (CIP) for further gold recovery. Tailings from this circuit are blended with the SART
effluent for cyanide detox and discharge to the tailings pond. A flow diagram of the leach solution
circuit is provided in Figure 2.

Figure 2 The Mastra Mine Leach Solution Circuit

203

The SART plant successfully achieves the objective of avoiding the accumulation of copper-cyanide
complexes in the circulating leach solution, allowing for the efficient recovery of gold and
compliance with cyanide limits in tailings. Additionally, the process removes and recovers the
copper as a saleable copper concentrate and regenerates the cyanide for recycle to the gold
operation.

SART process control at Mastra

The objectives of the SART plant at the Mastra Mine were achieved with an overall control strategy
centred on maintaining pH and oxidation-reduction potential (ORP) in all reactors within ranges to
control SART process reactions while ensuring safe operating conditions. Operational data for the
time period of May 2011 to September 2012 will be reviewed to determine the effectiveness of the
control strategy.

Copper concentration and recovery

Copper concentration in the feed has ranged from 100 to 350 mg/L as illustrated in Figure 3. While
some variability in the copper in the SART plant discharge has been observed, it has been
consistently low with a median value of 17 mg/L. Operational data for this period also reports a
fairly consistently high copper recovery average of 89%, exceeding the design specification of 80%.
100
500

90

80
400

60
300
50

40
200

Copper Recovery (%)

Copper Concentration (mg/l)

70

30

20

100

10

Copper Concentration in Feed

Copper Concentration in Product

Copper Recovery

Figure 3 Copper Concentration in Feed, Discharge and Copper Recovery from May 2012 to September 2012

Despite widely fluctuating copper concentration in the feed, constant and consistent behaviour in
both copper concentration in the discharge and copper recovery were achieved. These results
confirm that the pH and ORP controls utilized on the copper contactor were effective and
sufficiently robust to control the process reactions.

204

Cyanide Recovery
In Figure 4, the concentration of free cyanide in the feed is observed to be fairly consistent, ranging
from 408 to 590 mg/L with an average concentration of 484 mg/L. In contrast, free cyanide in the
discharge was observed to trend downwards from 1,154 to 588 mg/L over the reporting period. The
average concentration of free cyanide returned to the leach was 825 mg/L.
1400

Free Cyanide Concentration in

Feed
1200

Free Cyanide Concentration in

Discharge

Free Cyanide Concentration (mg/L)

1000

800

600

400

200

Figure 4 Free Cyanide Concentration in Feed, Discharge and Free Cyanide Recovery from May, 2011 to
September, 2012

The concentration of free cyanide in the SART plant feed reflects operating conditions in the leach
reactors whereby free cyanide levels are adjusted to maximize gold recovery. And the decline in the
concentration of free cyande in the SART plant discharge indicates a concurrent decline in cyanidesoluble copper reporting to SART. These results clearly indicate the ability of SART to recover the
cyanide associated with copper cyanide complexes and make it available for re-use in the leach
process.

Key SART design elements for successful process performance

Based on the experience BioteQ has gained by designing the Mastra Mine SART plant, other SART
and sulphide precipitation plants, and the provision of technical services for existing plants, BioteQ
has observed the following key SART design elements are crucial for successful process control and
performance.

Safety
When designing any plant that involves hazardous gasses such as HCN and H2S, it is imperative
that safety is the first consideration. One of the key elements of safety related to these gasses is the

205

installation of stationary gas monitors. An experienced designer will conduct a detailed analysis of
the plant to determine the correct number and location of monitors required to provide accurate
readings of gas levels. Important considerations for sensor location include the density of the gasses
(note that H2S is heavier than air but HCN is not) and an understanding of where in the plant the
potential exists for gasses to be released. This latter point cannot be understated. A poorly located
gas monitor is considerably more dangerous than not having any gas monitors because the poorly
located gas monitor may indicate that it is safe to enter a plant, even if gas has been released. For
this reason it is critical that the number and location of stationary gas monitors in a SART plant be
determined by a designer with sufficient experience in SART and/or sulphide precipitation to
ensure that it is done correctly.
Another critical safety consideration in the design of a SART plant is the back-up power. The plant
must be designed so that in the event of a power failure there is no risk to the safety of the operators
and site personnel. There are many aspects to be considered, two of which will be discussed here.
The first is that when a loss of power occurs, there is an increased risk of a gas release. To prevent
this, the gas scrubber system must have a continuous power supply so that vessels in the plant
where HCN or H2S are likely to occur are always under negative pressure.
Another consideration for back-up power is the ability to re-start the plant once the main power is
restored. The plant needs to be analysed to determine which elements cannot be allowed to stop
functioning for risk of them being unable to re-start. In addition to the operational concern this also
presents a safety concern, as certain elements of a SART plant that may have problems re-starting
after a power failure contain hazardous gasses. As a result, if these parts of the plant are not
provided with back-up power there will be an increased safety risk to the operators and site
personnel during the re-start. For this reason, the requirements for back-up power in a SART plant
are best determined by a designer with experience in the operation of this type of plant.

Process
In addition to safety concerns that are specific to SART and sulphide precipitation plants, there are
numerous process design considerations that designers who are not familiar with SART may not
consider. One of the most important of these is the selection, location and mounting of instruments.
Although most designers will design a SART plant with pH and ORP monitors to control the
addition of the acid and NaHS, there are many details in the monitoring of these key parameters
that may be overlooked. In missing these details, the performance of the plant will suffer noticeably
in terms of copper and cyanide recovery, reagent consumption and also process availability.
Another key detail that can be missed is the importance of the location and mounting of probes. If
poorly located, a probe may become fouled or may take a reading in a poorly mixed region of a
tank that is not representative of the bulk of the tanks contents. An incorrectly mounted probe will
also be difficult to maintain, which will result in poor readings and consequently poor plant
performance.
Similarly, the selection of the type of probe is also critical. One of the primary causes of failure of
pH and ORP probes in SART is sulphide poisoning. While many probes are advertised as being
resistant to sulphide poisoning, BioteQs operating experience has shown that performace can vary
dramatically. In fact, BioteQ has found that the average lifespan of particular brands of probes can
vary from 1-2 months to up to 6 months. Experience has shown that proper selection, location and
mounting of the instruments in a SART plant will directly impact the efficiency and profitability of
both the SART plant and the overall mine itself.

206

Another key aspect in the process design of a SART plant is the design of the neutralization circuit.
If designed properly, the neutralization circuit will safely convert the HCN back to soluble NaCN
and Ca(CN)2, while producing an easy to settle and dewater gypsum cake. If not designed properly,
there is the risk of significant problems in the operation of the neutralization circuit. The majority of
the problems relate to the formation of lime and gypsum scale due to poorly controlled supersaturation. This scale can rapidly coat vessels and equipment, causing lines to plug and agitators
and pumps to fail. Cleaning the scale alone is a difficult process involving a lot of expensive downtime and repairing the plant after the fact in order to prevent a recurrence of the problem is even
more costly. The solution is to ensure that the initial plant design is done by a designer with
experience in the design and operation of this style of plant.
Perhaps the most important component of the plant is the gas scrubbers, which are designed to
neutralize the acid gasses (HCN and H2S) released during the process. A proper scrubber design
ensures that all vessels in the plant which contain acid gasses are at a slightly negative pressure so
that all of those gasses are collected. Inside the scrubber the gasses are absorbed in a solution of
NaOH (caustic soda) thus recovering them in a form that can be returned to the process via a bleed
of spent caustic solution. The design of gas scrubbers is unlike the design of the rest of the
equipment in a SART plant and requires a different set of expertise. If designed properly, the
exhaust gasses from the scrubber will be below threshold levels for the gasses (4.7 ppm for HCN
and 10 ppm for H2S) and the caustic bleed will contain a relatively low concentration of NaOH, thus
minimising reagent consumption.
Critical design issues to also consider include the gas flow required to keep the vessels evacuated,
the height of packing used in the scrubber and the concentration of NaOH used. The gas flow
should be calculated based on an understanding of how the vessels will be operated and what
openings are possible into the system. The height of packing in the scrubber is calculated based on a
mass transfer calculation taking into account all of the elements present in the gas to be scrubbed.
Generic coefficients for these calculations can be found in literature, but the actual values used by
the equipment designers are proprietary. The concentration of caustic used in the scrubber is also a
critical parameter which requires input from an experienced designer. If too high a concentration is
used, solids will form inside the scrubber while too low a concentration will result in poor
scrubbing efficiency. Either of these problems can result in undesired gas emmissions, waste of
reagents or both.

CONCLUSIONS
Cyanide-soluble metals can create costly intereference in gold cyanidation. SART technology can
remove these metals, recovering them as a saleable, high-grade concentrate and regenerate the
cyanide for recycle to the leach. When applied successfully, SART can generate a revenue stream
from the recovered metal, lower reagent cost and reduce residual cyanide levels in tailings. BioteQ
has utilized its experience to design SART plants that provide these deliverables and in the process,
amassed knowledge of the key design elements necessary for successful SART plant process control
and performance. SART can be safe, easily controlled and efficient through proper design that
incorporates careful gas sensor location, back-up power requirements, instrumentation selection
and positioning, neutralization circuit specifications and systematic scrubber design.

207

REFERENCES
Ford, K, Fleming, C, and Henderson, R, 2008. Application of the SART Process to Heap Leaching of GoldCopper Ores at Maricunga, Chile. Proceedings of the 40th Annual Meeting of the Canadian Mineral
Processors, Ottawa, Canada, pp 591-613.
Kratochvil, D, Chan, N, and Hall, A, 2013. Integrating SART into Metallurgical Flowsheets for Cyanide
Recovery. In Proceedings of ALTA Au 2013, Perth, Australia. In press.
Lawrence, R. W., & Lopez, O. (2011). Improving the Economics of Gold-Copper Ore Projects Using SART
Technology. World Gold Conference, Montreal, Canada.

208

chap. 4
Solvent extraction and ion exchange

Analysis phase separation

profiles in copper extraction

This work studies the progress of binary mixtures interphases, constituted

by an aqueous phase that simulated physics and chemical properties of a
leaching solution of copper mineral and an organic phase formed by copper

Patricio Navarro and Sebastin

Jara. Universidad de Santiago de
Chile

extractant (LIX 984N) diluted in a commercial solvent.

The phenomenon was studied in aqueous continuity, inside a glass reactor
of cylindrical shape of 1 litre capacity used as mixer-settler. The profile was
obtained through an image analysis, and during the experiments, the copper
concentration in the leaching solution was changed (1-6g/L) as well as the
extractant concentration of the organic phase (10-30% v/v), keeping the agitation time and the stirring velocity constant.
Through the analysis of the profiles, three ways of phase distribution dispersed between fronts are proposed, each one of these ways would be developed between inflection points of each curve. Alongside the sedimentation
profile and separate by the inflection point, there will be two types of sedimentations. Before the change of sedimentation velocity curve, the dispersed
phase would sediment without interaction between drops (free sedimentation). After the change in the sedimentation velocity, there would be an interaction between drops, increasing the sedimentation profiles velocity (mixed
sedimentation). Under the front height of coalescence, the dispersed phase
would be set forming a zone of drops packaging, which would gain control of
the separation of phases when the sedimentation profile would be fully
developed.

211

INTRODUCTION
Hydrometallurgy of copper has grown exponentially thanks to the invention of the hidroxime type
organic extractants (approximately 1960), which over the years have evolved in selectivity and
physical properties for the different conditions that present processing plants of copper minerals
that includes leaching, solvent extraction and electrowinning steps (Castillo, Biela & Navarro, 2012).
In the copper extraction process, the most common form of mixing the organic and aqueous phase,
to carry out the transfer reaction, is to disperse as droplets one phase inside the other by mechanical
agitation, and to make the process useful, it is necessary that the dispersion be of a temporary
nature (Castillo, Biela & Navarro, 2012).
When dispersed liquids, the most influent factors in the rate of phase separation are: the size of the
droplets, the viscosity of the continuous phase and the density difference between the phases
(Hartland & Jeelani, 1988), (Nadiv & Semiat, 1995).
The droplet size influences the mass transfer and the speed at which the phases are separated. A
small drop increases the mass transfer, because there is more surface area available for the reaction
of chelatin molecules and copper molecules (Cu+2), however, a very small droplet diameter
increases the time of phase separation, producing an important effect on the size of the equipment,
the performance of the plant and the formation of a stable dispersion (emulsion) (Kordosky, Virnig
& Boley, 2006).
Increasing or decreasing the viscosity of the continuous phase, influences the droplet freedom to
move towards its interface, also the greater the density difference, the shorter the time of phase
separation. The effects of viscosity and density are confused when there is a chemical reaction
between the phases, for example, when the mass transfer occurred from the aqueous to the organic
phase would cause that these phases change their physical properties (Kordosky, Virnig & Boley,
2006).

EXPERIMENTAL METHODS
Equipment
To measure progress of the coalescence and sedimentation profiles, we used a jacketed reactor of 1
liter of capacity, 14 cm high and 10 cm of internal diameter. The reactor is equipped with baffles of
14 cm long by 1 cm wide and a graduated scale to indicate the height of the advance front. To
disperse phases it was used a circular base stirrer with six equidistantly spaced blades under the
surface and two blades over the circumferential surface of the agitator. The agitator was connected
to a variable speed motor (0-2.000 rpm).
The profiles of each experiment were captured by a digital camera that processes 15 frames per
second. The camera is connected to a computer which performs the analysis of each front height
versus time (Castillo, Biela & Navarro, 2012).

212

Procedure
To perform the experiments, we used a synthetic leaching solution with 1-3-6 g/L of copper
concentration and concentration of extractant (LIX 984N) of 10-20-30% v/v diluted in commercial
solvent (Shellsol 24 AR). Equal volumes of each phase were mixed, with initial heights of 50 mm.
Each experiment was performed with the conditions shown in Table 1. Conductivity of this
solution was monitored to ensure the aqueous continuity of the system.
Table 1 Operation conditions
Parameter
Stirring time (s)
Stirring rate (rpm)
Continous phase
Initial pH
Temperature (C)
Sulphate concentration (mol/l)

Value
180
800
Aqueous
2
20
0,25

RESULTS
Figure 1 shows the sedimentation and coalescence profiles obtained by varying the extractant
concentration. The sedimentation rate practically was not affected, however, coalescence rate
decreased with increasing extractant concentration.

60

Perfiles de Coalescencia

Altura del Frente de Avance, mm

40

20

-20

Perfil de Sedimentacin (1 g/L [Cu+2] - 10% v/v extracatante)

Perfil de Colescencia (1 g/L [Cu+2] - 10% v/v extracatante)
Perfil de Sedimentacin (1 g/L [Cu+2] - 20% v/v extracatante)
Perfil de Coalescencia (1 g/L [Cu+2] - 20% v/v extracatante)

-40

Perfil de Sedimentacin (1 g/L [Cu+2] - 30% v/v extracatante)

Perfiles de Sedimentacin

-60

10

20

Perfil de Coalescencia (1 g/L [Cu+2] - 30% v/v extracatante)

30

40

50

60

Tiempo, s

Figure 1 Profiles obtained when extractant concentration is varied

Figure 2 shows the sedimentation and coalescence profiles obtained by varying the concentration of
copper (Cu+2). For low concentrations of copper, lower coalescence speeds are detected.

213

60

Perfiles de Coalescencia

Altura del Frente de Avance, mm

40

20

-20

Perfil de Sedimentacin ( 1g/L [Cu+2] - 30% v/v extractante)

Perfil de Coalescencia ( 1 g/L [Cu+2] - 30% v/v extractante)
Perfil de Sedimentacin ( 3g/L [Cu+2] - 30% v/v extractante)
Perfil de Coalescencia ( 3 g/L [Cu+2] - 30% v/v extractante)

-40

Perfil de Sedimentacin ( 6g/L [Cu+2] - 30% v/v extractante)

Perfiles de Sedimentacin
-60

10

20

Perfil de Coalescencia ( 6 g/L [Cu+2] - 30% v/v extractante)

30

40

50

60

Tiempo, s

Figura 2 Profiles obtained when copper concentration is varied (Cu+2)

Table 2 presents the turning time of the settling curve (t*s), the turning time of coalescence curve
(t*c) and the phase separation time (tf) of each experiment.
Table 2 Inflection and phase separation times of each experience
Experience
1 g/L Cu+2 10% v/v of extractant
1 g/L Cu+2 20% v/v of extractant
1 g/L Cu+2 30% v/v of extractant
3 g/L Cu+2 10% v/v of extractant
3 g/L Cu+2 20% v/v of extractant
3 g/L Cu+2 30% v/v of extractant
6 g/L Cu+2 10% v/v of extractant
6 g/L Cu+2 20% v/v of extractant
6 g/L Cu+2 30% v/v of extractant

t*s (s)
12
13
14
10
12
13
12
12
13

t*c (s)
37
44
44
35
37
38
35
34
38

tf (s)
47
56
57
48
50
51
47
47
51

The experiences developed, showed that on average, the first ten seconds, the coalescence profile
remains motionless and there are only advance of the sedimentation profile. When the coalescence
profile starts to develop, there is a turning point in the sedimentation profile. The sedimentation
front accelerates after the first ten seconds elapsed separation, and runs until finish approximately
in the time there was an inflection point in the profile of coalescence.
Based on the results obtained, the dispersed phase would be ordered in three ways between the
sedimentation and coalescence profile.
Beside the sedimentation profile until the inflection point (approximately in ten seconds), the
droplets of the dispersed phase would rise without interaction with other drops, to this way of find
the dispersed phase was called free sedimentation zone. After the inflection point, the

214

sedimentation profile is accelerated, so it is believed that the droplets sediment together increasing
their speed in the continuous phase, to this area or way to find the dispersed phase was called
mixed sedimentation zone.
The third way to find the disperse phase between the profiles is next to the coalescing profile,
forming a zone of packing of droplets. The rate of the coalescence profile increase when detecting
an inflection point of the curve which coincides approximately with the end of the sedimentation
profile and exist only where the packaging area between the profiles (see Figure 3). The continuous
phase volume that is retained in the packing area when the sedimentation profile is over, it is small
due to the high degree of packing of the droplets.

60

Altura del Frente de Avance, mm

40

Zona de Empaquetamiento
20

Sedimentacin Mixta
0

-20

Sedimentacin Libre
Perfil de Sedimentacin

-40

Perfil de Coalescencia

-60

10

20

30

40

50

60

Tiempo, s

Figure 3 Methods to find the dispersed phase between the coalescence and sedimentation profiles.

CONCLUSIONS
The speed at which develops the sedimentation and coalescence profile is strongly linked with the
phases physical properties. Within the system under study, it was found that result of the chemical
reaction that takes place at the dispersed phase interface, is favored the movement of the droplets
through the continuous phase, because of this, the sedimentation finishes before the coalescence
phenomenon.
The analysis of the phase separation profiles allow, through the turning points of the curves,
elucidate how the dispersed phase is distributed as it develops coalescence and sedimentation of
the liquid-liquid system under study.

ACKNOWLEDGMENT
We acknowledge the research funding to the Universidad de Santiago de Chile through 051114ND
DICYT project.

215

REFERENCES
Castillo J., Biela F., Navarro P. (2012). Estudio de separacin de fases en sistemas lquido-lquido usando LIX
984N en fase orgnica. Revista de Metalurgia. 48, 107-117.
Hartland S., Jeelani S. (1988). Prediction of Sedimentation and Coalescence Profiles in Decaying Batch
Dispersion. Chemical Engineering Science. 43, 2421-2429.
Kordosky G., Virnig M., Boley B. (2006). Solvent Extraction in Hydrometallurgy: Present and Future. Tsinghua
Science & Technology. 11, 160-164.
Nadiv C., Semiat R. (1995). Batch Settling of Liquid-Liquid Dispersion. Industrial & Engineering Chemistry
Research. 34, 2427-2435.

216

Mo recovery process from high

acidity solutions via solvent
extraction using CYANEX600

Miners are increasingly challenged with higher operating costs, lower grades
and hard to process minerals. To make the mining industry more profitable
and sustainable in the long term, there have been efforts to develop processes

Mauricio Salamanca, Alejandro

Quilodrn and Osvaldo Castro.
Cytec Chile Limitada

and reagents that can address the challenges outlined above. In some operations
the molybdenum concentrate is discarded because of its low grade, however
this material could be a good source of molybdenum if leached and processed
via solvent extraction. Technologies are available to leach these concentrates
using sulfuric acid. The resulting pregnant leach solution (PLS) will usually
exceed 2g/L Mo concentration.
Cytec has developed a technology based on CYANEX600, capable of recovering, purifying and concentrating molybdenum from acidic aqueous
solutions. CYANEX600 is a highly selective stable reagent based on phosphinic acid. The characteristics of the PLS coming from the leach of these molybdenite concentrates are suitable for processing with CYANEX600. Cytec has
trialed the technology in several pilotings and these have demonstrated that
it is possible to obtain molybdenum recoveries exceeding 90%. Additionally
the technology is highly selective and can generate a high purity rich solution.
Typically this concentrate would exceed 60g/L and at that strength it is then
crystallized to make a high purity molybdate salt.

217

INTRODUCTION
At present more mining companies are trying to expand and diversify the metals that they recover.
There has been important progress in the hydrometallurgical processes of materials which had been
previously disregarded (Cruz & Reghezza, 2007; An et al., 2009; Valenzuela et al. 1995, Zeng &
Cheng, 2009). One of these is the leaching of molybdenite concentrates. However, these low grade
concentrates can now be processed via acid leaching.
Cytecs new solvent extraction process aims to recover Mo from solutions obtained from the
leaching of copper oxides as well as from acid leach of other types of materials, such as foundry
dust, slag and molybdenite (Soderstrom et al., 2010; Cocalia et al., 2011)

Extraction, %

This type of extractant is capable of loading different metals depending on the pH of the reaction.
At a low pH (closer to or lower than 1), CYANEX600 is highly selective in the recovery of Mo and
is very stable. Figure shows the selectivity curves of this reagent for different metals (Fe, Al, and
Cu). It can be seen that at a low pH, CYANEX600 has very good selectivity for Mo. In some
operations, especially when smelter dust is leached, the content of arsenic is high (over 10 g/L).
Using CYANEX600, the selectivity factor Mo:As is > 2000.
100
90
80
70
60
50
40
30
20
10
0
0.0

0.2

0.4

Mo

0.6
0.8
Equilibrium pH

Fe

Al

1.0

1.2

Cu

Figure 1 Selectivity of Reagent Cyanex600 at Low Ph

The proposed mechanism for Mo extraction is shown below:

(

(1)

This reaction is not reversible due to the stability of the organometallic complex. The extraction
kinetics are very fast and are not the limiting stage of the process.

218

Due to the strength of the formed complex acid is not suitable stripping. However, using a basic
media, recovery is very high. The base allows breaking the organometallic compound and
releasing Mo to the aqueous solution. The mechanism is shown below:

(2)

Molybdenum remains in the aqueous solution called rich solution or RS- as ammonium
molybdate, which may be crystallized using known methods such as evaporation or acidification
yielding various molybdate salts which can be sold to existing marketing or used as a feed to
thermal processing to make pure metal or metal oxides. Examples of the ammonium molybdate
products include ammonium tetramolybdate, ammonium heptamolybdate, and others.

MATERIALS AND METHODS

A solution coming from an acid leach of molybdenite concentrates, whose Mo concentration was
2.76 g/L and an acid concentration of 517 g/L, was made available for this study. The ICP analysis
was performed by SGS.

Piloting
A pilot plant was operated with a pumping capacity of 25 mL/min in a 2E+ W1+2S configuration.
(See Figure 2). The purpose of using two extraction and two stripping stages was to maximize Mo
recovery. The organic wash stage was included in order to minimize physical contamination of the
rich solution.
PLS
E1

TK

E2

TK
Water

Ammonia Make
up

pH
Meter

S1

SR

S2

Organic

Figure 2 Pilot Plant Flow Diagram

The plant operation lasted 34 hours.

Acidulated water organic washer.

219

RESULTS
Characterization
The solution used in the piloting operation was characterized. Table 1 shows the results of such
analysis.
Table 1 characterization of PLS used
[Mo]
[Acid]
Density
Viscosity
Superficial Tension

2.76
517.6
1.31
2.85
68.5

g/L
g/L
g/mL
cP
D/cm

The Mo content measured in the PLS was 2.76 g/L and the Cu content was 0.9 g/L. Physical
properties (viscosity, density and interfacial tension) maintain values within the range
recommended for piloting operations.
An ICP was performed on a PLS sample. Table 2 shows the results of such analysis. The
contaminants of this solution are: Cu (1.35 g/L) and Fe (1.04 g/L).
Table 2 ICP of SGS for a PLS sample
Element
Unit
PLS

Element
Unit
PLS

Element
Unit
PLS

Ag

Al

As

Ba

Be

Bi

Ca

Cd

Co

Cr

Cu

Fe

MG/L

MG/L

MG/L

MG/L

MG/L

MG/L

G/L

MG/L

MG/L

MG/L

G/L

G/L

34.2

3.5

0.064

400

90.6

1.1

1.1

Ga

La

Li

Mg

Mn

Na

Nb

Ni

Pb

Sb

MG/L

MG/L

MG/L

MG/L

G/L

MG/L

G/L

MG/L

MG/L

G/L

MG/L

MG/L

10.6

0.08

7.7

0.11

50.1

0.02

Sc

Se

Sn

Sr

Te

Ti

Zn

Zr

MG/L

MG/L

MG/L

MG/L

MG/L

G/L

MG/L

MG/L

MG/L

0.04

80

Results of the piloting operation

Physical properties
One of the most important aspects in an SX operation is the control of physical variables. This
piloting operation showed the following:
Organic band: the ideal aim is to maintain a stable organic band that allows the minimization of
entrainment. Any unexpected change may result in some effects on the organic band. During the
piloting operation, the height of the organic was intended to be maintained around 3.2 cm.
Figure 3 shows the data of the organic bands along the experiment for the 5 stages used. As
shown, the height of all stages remained around the targeted value, which allowed the balance

220

of all the system. Also as shown in Table 3, the average bands remained close to the objective
with a deviation near 10%.
Dispersion band: corresponds to the measurement of dispersion. The ideal operation does not
have a dispersion band thus assuring that the phases are transferred with minimal entrainment.
During Cytecs piloting there was no appearance of a dispersion band in the phase transfer zone
and thus entrainment must have been low.
Crud band: crud is a solid emulsion that may alter the stability of the plant and increase
entrainments. This will strongly depend on the amount of solids that are present in the plant.
Crud formation is a phenomenon inherent to every plant that must be controlled both in a
preventative and palliative manner. The presence of crud in the system was not detected during
the piloting operation. The absence of crud shows the stability of the operation.
Phase disengagement time: high phase disengagement times could increase entrainments.
Figure 4 shows the progress of Phase disengagement in the piloting operation. It is important to
mention that an SX operation may tolerate up to 150 s measured in standard conditions. In this
piloting operation the disengagement time varied between 80 and 120 s. The only stage that
presented higher values was S1.

4,5

3,5
3

Banda, cm

2,5
2

1,5
1
0,5
0

10

15
E1

20
Time , hr

E2

S1

S2

25

30

35

40

Figure 3 Organic bands

Table 3 Organic bands

E1
E2
S1
S2
W

Average
Band, cm
3.18
3.26
3.50
3.42
3.30

Percentage
deviation, %
11
10
6
12
14

221

200

180
160
140

120

Time, s

100
80
60

40
20
0

10

15
E1

E2

20
Time, hr
S1

S2

25

30

35

40

Figure 4 Phase disengagement times

Metallurgical results
Figure 5 shows the evolution of the recovery and the Mo concentration in PLS and Raffinate. The
average recovery obtained in this piloting was 92.35%.
100
90
2

80
70

1,5

60
50

40

Recovery, %

Mo Concentration (PLS and Raffinate), gpl

2,5

30

0,5

20
10

10

15
PLS

20
Time, hr
Raff

25

30

35

40

Recovery

Figure 5 Evolution of the MO recovery and MO concentration in PLS and Raffinate

The rich solution was generated after the organic was loaded with up to 2 g/L of Mo. The
composition of this rich solution is shown in Table 6. As it can be seen, Mo could be concentrated in
the rich solution until it reached 66.2 g/L in a period of almost 34 hours. During this period the
inventory of organic circulated three times through the pilot plant.

222

70

Mo Concentration, g/L

60
50
40
30
20
10
0

10

15

20
Time, hr.

25

30

35

40

Figure 6 Enrichment of rich solution

Selectivity
The selectivity of CYANEX 600 is important because as Mo is concentrated in the rich solution, the
impurities are not built up which allows the development of a high purity final product without
additional purification steps. Table 4 shows a comparison of contaminants that are present in the
PLS and the rich solution during the piloting. Because of the excellent selectivity, Mo concentrated
almost 30 times while the presence of other metals in the rich solution was very low. In average, the
estimated Mo selectivity against these elements was 40.000:1.
Table 4 Comparison of the chemical composition of PLS and the rich solution
PLS
Rich Liquor

Al

As

Ca

Cd

Cr

Cu

Fe

Mg

Mn

Mo

Na

Ni

Pb

Ti

Zn

34,2
<0.1

3,5
<0.1

63,9
<0.1

0,4
<0.1

90,6
<0.1

1117,8
<0.1

1151,7
<0.1

10,6
12,7

82,5
<0.1

7,7
<0.1

2102,8
66200,0

114,8
5,9

50,1
<0.1

1,0
<0.1

37,6
0,3

79,9
<0.1

CONCLUSIONS
The Mo recovery and concentration through the technology developed by Cytec using
CYANEX600 was successful. The piloting operation demonstrated the robustness of the
technology. The process was carried out smoothly, with an absence of crud, dispersion or emulsion
and confirmed the high stability of the reagent at low pH. The average recovery was 92%,
selectivity was high, and there were no signs of contamination whether by chemical co-extraction or
physical entrainment.
Using the PLS, mentioned above it was possible to create a rich solution of 66 g/L of Mo in a 34hour period.

REFERENCES
An, J.W., Lee, Y.H., Kim, S.J., Tran, T., Lee, S.O. & Kim, M.J. (2009a) Production of high-purity molybdenum
compounds from a Cu-Mo acid-washed liquor using solvent extraction. Part 1: Laboratory studies,
Minerals Engineering, vol. 22, pp. 10201025.

223

An, J.W., Jung, B.H., Lee, Y.H., Tran, T., Kim, S.J. & Kim, M.J. (2009b) Production of high-purity molybdenum
compounds from a Cu-Mo acid-washed liquor using solvent extraction. Part 2: Pilot and plant
operations, Minerals Engineering, vol. 22, pp. 10261031.
Cocalia, V., Bednarski, T., Harris, D.J., Yaez, H., Soto, A., Soderstrom, M. & Kamenetzky, E. (2011) Recovery
of molybdenum from copper leach solutions by solvent extraction, ISEC 2011, Santiago, Chile,
Gecamin, Santiago.
Cruz, A. & Reghezza, A. (2007) Molybdenum: New byproduct at hydrometallurgical operations in CODELCO
North?, HydroCopper 2007, Proceedings of the IV international copper hydrometallurgy workshop, Via del
Mar, Chile, Gecamin, Santiago, pp. 349355.
Soderstrom, M.D., Bednarski, T., Yaez, H., Soto, A., Kamenetzky, E. & Cocalia, V. (2010) Getting more out of
your Copper SX Plant, Proceedings of Copper, 2010, Hamburg, Germany, vol. 5, pp. 20592070.
Soderstrom, M.D., Bednarski, T., Yaez, H., Soto, A., Kamenetzky, E. & Cocalia, V. (2010) CYANEX600 for
molybdenum recovery from acidic leach solutions, Cytec Solutions, Hydrometallurgy and Mineral
Processing, vol. 15.
Valenzuela, F.R., Andrade, J.P., Sapag, J., Tapia, C. & Basualto, C. (1995) The solvent extraction of
molybdenum and copper from acid leach residual solution of Chilean molybdenite concentrate,
Minerals Engineering, vol. 8, pp. 893904.
Zeng, L. & Cheng, Y. (2009a) A literature review of the recovery of molybdenum and vanadium from spent
hydrodesulphurization catalyst, part I: Metallurgical processes, Hydrometallurgy, vol. 98, pp. 19.
Zeng, L. & Cheng, Y. (2009b) A literature review of the recovery of molybdenum and vanadium from spent
hydrodesulphurization catalyst, part II: Separation and purification, Hydrometallurgy, vol. 98, pp. 1020.
Trademark Notice: The indicates a Registered Trademark in the United States and the or *
indicates a Trademark in the United States. The mark may also be registered, the subject of an
application for registration or a trademark in other countries.
Disclaimer: Cytec Industries Inc. in its own name and on behalf of its affiliated companies
(collectively, "Cytec") decline any liability with respect to the use made by anyone of the information
contained herein. The information contained herein represents Cytec's best knowledge thereon
without constituting any express or implied guarantee or warranty of any kind (including, but not
limited to, regarding the accuracy, the completeness or relevance of the data set out herein). Nothing
contained herein shall be construed as conferring any license or right under any patent or other
intellectual property rights of Cytec or of any third party. The information relating to the products is
given for information purposes only. No guarantee or warranty is provided that the product and/or
information is adapted for any specific use, performance or result and that product and/or
information do not infringe any Cytec and/or third party intellectual property rights. The user should
perform its own tests to determine the suitability for a particular purpose. The final choice of use of a
product and/or information as well as the investigation of any possible violation of intellectual
property rights of Cytec and/or third parties remains the sole responsibility of the user.
2013 Cytec Industries Inc. All Rights Reserved.

224

Use of NR reagents in presence

of nitrate ion in SX: A revision of
the present moment

Since the emergence of extractants based on modified aldoximes with better

resistance against the nitration phenomena, there has been growing interest
in the performance of pilot and industrial studies to demonstrate the effective-

Rodrigo Zambra, Alejandro

Quilodrn, Gonzalo Rivera and
Osvaldo Castro. Cytec Chile
Limitada

ness of their protection against nitration without sacrificing copper transfer

or Cu:Fe selectivity.
At present, there are four Cu plants in Chile that show levels of nitrate ions
indicating that they are at risk of nitration. In this paper, these plants are
identified as Plants A, B, C and D. Three plants chose the NR extractant
(modified aldoxime protected against nitration).
Herein we present the results of different studies on the use of NR extractants carried out at laboratory, pilot plant and industrial scales. The goal was
to determine the advantages of the modified aldoxime reagent with respect to
its metallurgic benefits, its greater Cu extraction efficiency and Cu transfer,
and its ability to protect against nitration.
The results shown in this paper correspond to measurements of the chemical and physical properties made at industrial plants in which ACORGA
NR was added as a substitute for LIX84I. The most important results evidence
a greater selectivity, an increase in the Cu transfer and a greater protection
against nitration.

225

INTRODUCTION
This work shows the results of different solvent extraction operations in the north of Chile at plants
where nitrates are present in the PLS feed. Currently, there are four plants producing copper via
hydrometallurgical processes that contain nitrate ions in their leach solutions, which may present
some level of nitration of the organic phase.
Of these four plants, two began operations with extractants based on ketoxime, favouring
protection against this nitration over maximizing the capacity of copper transfer and Cu:Fe
selectivity. As a result, Cytec developed an extractant with better resistance to nitration based on a
modified aldoxime that allows solvent extraction plants to maintain or improve protection against
nitration while simultaneously improving the level of Cu transfer and production due to the better
selectivity associated with aldoximes. At present, three of these four plants are using the extractant
ACORGA NR developed by Cytec Industries Inc. (modified aldoxime with protection against
nitration). The fourth plant is in the evaluation stage at the pilot level. The results obtained are
shown below.

BACKGROUND INFORMATION
Nitration is a phenomenon that attracted the interest of the copper mining industry in the late 90s
due to the experience at Lomas Bayas. Since then, the majority of projects have used extractants
based on ketoximes, thus generating the paradigm in which the the nitrate ion in the PLS must be
extracted from the SX.
Cytec therefore developed an extractant that maintains the benefits of extraction and selectivity
associated with aldoximes yet provides equal or improved levels of protection against nitration as
those obtained with ketoximes.

What is Nitration?
Nitration is a substitution reaction involving the oxime and nitronium ions (NO2+) present in PLS
solutions or electrolytes. This reaction results in the pseudo stabilization of organometallic
complexes of copper, which prevents the stripping and removal of Cu ions from the electrolyte
solution, resulting in a circulating load that avoids the transfer of copper (Yez, Soto, Soderstrom,
Bednarski, 2009). The formation of the nitronium ion results from:
NO3- + H2SO4
HNO3 + H2SO4

HNO3 + HSO4NO2+ + H2O + HSO4-

The proposed mechanism for the nitration reaction of the oxime is:

226

(1)
(2)

OH
OH

OH

R1

OH

OH

NO2+

O 2N
+

R2

OH

R1

O 2N

R1

R2

R2

R1 H or CH3 (aldoxime or ketoxime)

R2 C9H19 or C12H25 (nonylaldoxime or dodecylaldoxime)
The nitration reaction directly depends on the nitrate concentration present in the aqueous
solutions of the plant and the acidity level, but is also affected by temperature, redox potential,
interfacial tension and the reactivity of the aqueous and organic phases (Hurtado-Guzmn, C. &
Menacho, J., 2003). Depending on each of these variables, the nitration conditions will be more or less
aggressive.
Oxime nitration may cause several problems in the normal operation of the SX plant, the most
important being the following:

Loss of copper transfer capacity.

Increased phase disengagement times.
Reduction of interfacial tension.
Increased entrainments from one phase into the other.
Increased consumption of extractant.
Increased speed of hydrolytic degradation.

Operations where nitrate ions are present under the above-mentioned conditions should establish
routine controls in order to avoid or reduce the possibility of nitration of the oxime, namely:
Determination of the residual copper. The non-nitrated organics must be completely discharged
under standard laboratory conditions. The copper detected after this discharge is known as the
residual copper and is directly proportional to the amount of nitrated oxime.
Analytical measurement of the content of nitrated products (gas chromatography).
Below are shown the results of different studies with the extractant ACORGA NR that were
carried out at different copper solvent extraction plants known to have nitrate ions in their leach
solutions.

RESULTS AND DISCUSSION

Case 1, Plant A
The conditions at the plant prior to substitution of the extractant are listed in Table 1.
Table 1 Characterization of solutions at Plant A
Elements
Cu
pH/H2SO4
NO3ORP

Unit
g/L
-/g/L
ppm
mV

PLS
4.60
1.60
1,890
470

Spent
36
180
63
500

227

Simulations
In order to determine the extraction efficiency of the reagents Lix 84I and ACORGA NR10, the
extraction and stripping isotherms were measured in the laboratory using real plant solutions. The
results are presented in Table 2.
Table 2 Results of Simulations with plant solutions (23% extractant).
Extractant

Efficiency Train A
%

Efficiency Train B
%

LIX 84I

89.23

77.51

NR 10

95.21

87.33

The extraction efficiency of the ACORGA NR10 extractant was 6 percentage points higher than
that of the LIX 84I extractant, which was used in the plant. The basic explanation for this difference
is that the ACORGA NR10 extractant has a greater extraction power than ketoxime extractants at
pH levels below 1.8. At the same time, the aldoxime has higher extraction kinetics, which implies a
higher efficiency near the equilibrium point (EAE). Based on these results, a decision was made to
use the ACORGA NR10 extractant in the plant due to its great metallurgical advantages, including
its selectivity and the avoidance of nitration due to the low presence of nitrates in the PLS and
electrolyte.

Accelerated nitration tests

Several

tests were carried out in order to evaluate the behaviour of the extractant in a possible
nitration scenario as defined according to the mining plan. The properties of the evaluated PLS feed
are shown in Table 3. Untreated PLS feed and feeds with freshly added LIX 84I (ketoxime), LIX 860
(pure aldoxime) and ACORGA NR10 (modified aldoxime) extractants were evaluated.
Table 3 PLS conditions for the accelerated nitration tests.
Properties
Cu
NO3
FeT
Cl
P. Redox
T
pH

Value
2.50
62.0
4.70
10.30
752
40
1.10

Units
g/L
g/L
g/L
g/L
mV
C
-

The tests were carried out under standard conditions for accelerated nitration (see Table 3). The
results obtained clearly show, as can be seen in the two graphs presented in Figure 1, that there was
a stronger resistance to nitration when using the ACORGA NR10 reagent (approx. 50%) than that
observed for the untreated feed or the feeds with freshly added LIX 84I and LIX 860.

228

Figure 1 Results of accelerated nitration tests based on residual copper and nitroxime.

Transfer and selectivity results

The plants transfer and selectivity results are shown in Figures 2 and 3. As can be seen in the
graphs in Figure 2, there was a reduction in the efficiency of the extraction due to an increase in the
Cu concentration in the PLS (4,6 to 7 g/L), and a decision was made to increase the percentage of
extractant by 2% (25% in all). Later, it was decided to add the ACORGA NR10extractant, which
slightly increased the extraction efficiency in Train A, serial circuit 2E + 2S, but strongly increased
the efficiency in Train B with a 2E + 1E// + 1S circuit, resulting in an increase of approximately 10
percentage points.
On the other hand, the addition of ACORGA NR Acorga NR resulted in an increase in the plant
selectivity from 200 to 900 (seen figure 3).

ComienzoagregadoAcorg
a NR 10

ComienzoagregadoAcorg
a NR 10

Figure 2 Train A and B extraction efficiency.

229

Figure 3 Plant selectivity, Train A and B.

Case 2, Plant B
A comparison of the behaviour of the ACORGA extractant and the reagent currently in use at the
plant was conducted with a focus on their resistance to nitration and metallurgical response.

Simulations
Simulation tests were carried out by generating isotherms using real plant solutions (PLS and
Spent) in order determine the extraction efficiency of the two reagents. The evaluated reagents
were fresh LIX 84I and ACORGA NR20.The results are presented in Table 4.
Table 4 Evaluation of extractants for plant PLS and Spent feeds.

/v
Extractant

HEAP
Extraction
Efficiency
(%)

ROM
Extraction
Efficiency
(%)

Organic
Load (%)

LIX 84 I

24.9

87.60%

67.70%

86.20%

Acorga NR 20

24.9

89.40%

72.10%

98.50%

Extractant
Type

As can be seen in Table 4, for the ACORGA extractant, the global HEAP extraction efficiency was
1.80 percentage points higher and the PLS ROM extraction efficiency was 4.4 percentage points
higher than that of the extractant currently used in the plant (LIX 84I).

Accelerated nitration tests

Next, accelerated nitration tests enabled the evaluation of the LIX 84I and Cytec-developed
ACORGA NR20 extractants under aggressive nitrating conditions in the laboratory. Both
extractants were compared to a conventional aldoxime.
The evaluation took place over a period of 150 days. The PLS solution used in this study was the
same as that used in the HEAP extraction efficiency test, but modified with added impurities
(chloride, iron, and nitrate) and a lower acidity (pH: 1.0) (Table 5). The extractants were mixed in a

230

1:1 ratio, and the solution was submerged in a thermostatic bath at a temperature of 40 C with
constant agitation.
Table 5 Characterization of the PLS HEAP Solutions
COMPOSITION

INITIAL HEAP
PLS

MODIFIED
HEAP PLS

Acidity

g/L

0.9

5.7

NO3-

g/L

11.07

58.8

FeT

g/L

2.24

3.08

g/L

8.8

10.38

Cl

It can be clearly seen in Figure 4 below that both the ACORGA NR20 and Lix 84I extractants had
an appropriate resistance to nitration, as verified against the nitration performance using the
aldoxime standard (ACORGAPT5050), which was nitrated.

Figure 4 Results of accelerated nitration tests based on residual copper and nitroxime.

Pilot plant evaluation

Based on these positive laboratory results, the ACORGA NR20 extractant was evaluated in a 100
cm3/min (deep cell) pilot plant under the same conditions as those used for the accelerated nitration
tests, and its performance was compared to that of the LIX 84I extractant. The initial conditions for
the pilot study are presented in Table 6.
The configuration of the pilot plant corresponded to that of an industrial plant, and the extractant
was added at 26.30 % for Lix84I and 24.94% for ACORGA NR20. The results of the tests are shown
below in Figure 5.
Table 6 Pilot Plant Test Initial Conditions

Cu PLS
pH / H2SO4
O/A E
O/A S

g/L
- / g/L

HEAP
1.97
2.08

ROM
1.63
1.81

0.95
1.24

0.95
1.24

Spent
42
175

231

Figure 5 Pilot plant results.

It can be clearly seen in the figure that both the HEAP and ROM extraction efficiencies increased
with the use of ACORGA NR20 compared to the efficiencies obtained with the extractant that was
currently in use at the plant (LIX 84I). The average values obtained are presented in Table 7.
Table 7 Extraction Efficiency, Pilot Plant results.

Extractant

HEAP Extraction
Efficiency
(%)

ROM Extraction
Efficiency
(%)

LIX 84 IC

89.44

67.24

ACORGA NR 20

95.28

76.05

The extraction efficiency results clearly show a better metallurgical response for the ACORGA
NR20 extractant under the conditions used in the plant, with a better performance for the
ACORGA reagent in the SX of 5.8 percentage points for the HEAP efficiency and 8.8 percentage
points for the ROM efficiency.
Both extraction efficiencies are enhanced using the ACORGA NR20 extractant because this
reagent is based on modified aldoxime chemistry, which has favorable kinetics for mass transfer as
compared to those of extractants based on ketoxime chemistry. As a result, a better mixing
efficiency near the equilibrium point is achieved. At the same time, the ACORGA extractant also
tolerates a wide range of pH, maintaining good chemical and metallurgical performance from pH
1.0 to 2.5.
In addition, the results for the selectivity of the ACORGA NR20 extractant conclusively confirm
that the new reagent improves the plant selectivity by approximately 50%. The obtained Fe
co/extraction isotherms for both the HEAP and ROM PLS streams are shown in Figure 6 below.

232

Figure 6 Fe Co-extraction isotherms in organic and PLs Heap and ROM solutions.

CONCLUSIONS
Based on the results of the studies in the laboratory, pilot plant and actual plant described above,
the following conclusions can be made:
There is a great increase in the extraction efficiency and transfer of copper when using the
ACORGA NR extractant, mainly because it provides better performance at low pH and
enhanced extraction kinetics, which helps improve the efficiency near the equilibrium point.
In all of the cases studied, the ACORGA NR reagent performed better by at least two
percentage points, with a maximum difference of 8 percentage units.
Cu/Fe selectivity increased significantly, increasing by 50%.
The ACORGA NR extractant offers protection for the plant organic inventory under nitration
conditions, ensuring a similar or better response than the LIX 84I extractant.

REFERENCES
YezH , Soto A., Soderstrom, M., Bednarski T., Nitration in copper SX? Cytec ACORGA provides a new
reagent, Hydrocopper 2009, Antofagasta, Chile 332-341.
Hurtado-Guzmn, C. & Menacho, J. (2003).Oxime degradation chemistry in copper solvent extraction plants,
Copper 2003. The Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Quebec,
Canada, 719 734.
Bednarski T., Soderstrom, M., Tinkler O. Reagent Development: New ACORGA Formulations with enhanced
Stability, Alta 2010, Auckland, New Zealand.
Trademark Notice: The indicates a Registered Trademark in the United States and the or * indicates a
Trademark in the United States. The mark may also be registered, the subject of an application for registration
or a trademark in other countries.
Disclaimer: Cytec Industries Inc. in its own name and on behalf of its affiliated companies (collectively,
"Cytec") decline any liability with respect to the use made by anyone of the information contained herein. The
information contained herein represents Cytec's best knowledge thereon without constituting any express or
implied guarantee or warranty of any kind (including, but not limited to, regarding the accuracy, the
completeness or relevance of the data set out herein). Nothing contained herein shall be construed as
conferring any license or right under any patent or other intellectual property rights of Cytec or of any third
party. The information relating to the products is given for information purposes only. No guarantee or

233

warranty is provided that the product and/or information is adapted for any specific use, performance or
result and that product and/or information do not infringe any Cytec and/or third party intellectual property
rights. The user should perform its own tests to determine the suitability for a particular purpose. The final
choice of use of a product and/or information as well as the investigation of any possible violation of
intellectual property rights of Cytec and/or third parties remains the sole responsibility of the user.
2013 Cytec Industries Inc. All Rights Reserved.

234

Advantages and contributions of

modifiers to copper extraction

Since the first application of solvent extraction as part of the hydrometallurgic

process for Cu, several studies have been carried out to optimize extractant
formulations. The introduction of modifiers mixed with aldoximes has con-

Rodrigo Zambra, Alejandro

Quilodrn, Osvaldo Castro and Sara
Pascuale. Cytec Chile Limitada

tributed significantly to the improvement of extraction efficiency (Cu transfer),

Cu:Fe selectivity, degradation stabiltiy and the ability to withstand diverse
conditions for Cu leaching (concentration, pH, acids in the electrolyte, etc.)
At present, reagents based on modified aldoximes are leaders in copper
extraction plants due to their physicochemical properties, which allow for
good metallurgic performance. Therefore, to gain more insight into the
performance of various formulation types, an analysis of different extractant
formulations is presented herein along with a comparison of their chemical
stability, physical performance and metallurgic response.
The results show that modified aldoximes have greater stability with respect to hydrolytic degradation, provide a higher Cu extraction efficiency and
provide high Cu:Fe selectivity with respect to the reagents based on ketoxime
and mixing aldoxime/ketoxime pure. It was also determined that the physical
performance of the plant (crud levels, entrainments, consumption, etc.) is
independent of the type of extractant and modifier used.
At present about 65% of the copper hydrometallurgic market is operating
with modified aldoximes, basically due to the high extraction capacity for Cu
concentrations at all ranges, low pH, high concentrations of acid in the electrolyte, different plant configurations and high Cu/Fe selectivity.

235

INTRODUCTION
The first industrial SX-EW plant began operating in Ranchers Bluebird, USA in 1968. It processed
dilute copper solutions, thus avoiding cementation. In 1979, nine plants were operating that used a
reagent based on a weak ketoxime (2-hydroxy-5-nonylbenzophenone), with a configuration that
required three to four extraction stages and two or three stripping stages. At this time, the first
aldoxime-based reagents were introduced, which are stronger extractants with better kinetics and
Cu/Fe selectivity.
Initially, improved reagents based on mixtures of ketoximes and aldoximes were used to optimize
their performance in stripping. Ultimately, new extractant formulations were created for the
industry based on modified aldoximes without the presence of ketoximes. Nonylphenol was the
first modifier to be used, but it was later substituted for tridecanol and ester.

BACKGROUND INFORMATION

OH
OH

The

R1

chemistry of copper solvent extraction is based on two basic oxime

molecules:
+

R2

R1 H or CH3 (aldoxime or ketoxime)

R2 C9H19 or C12H25 (nonylaldoxime or dodecylaldoxime)
The characteristics of the two base oximes may be summarized as follows:

Aldoxime-based extractants are:

Strong extractants at low pH
Require modification for effective stripping
Show very high Cu/Fe selectivity
Ketoxime-based extractants are:
Weak extractants
Provide good performance in stripping

The extraction capacity of oximes based on the pH is shown below (Figure 1).

Figure 1 Oxime extraction capacity based on pH

236

At present there are four general types of formulations for extractants. They include:

Modified aldoxime
Ketoxime
Modified aldoxime/Ketoxime
Ketoxime/Aldoxime

The addition of a modifier to the aldoxime causes a weakening of the extraction power and
improves stripping, depending on the oxime/modifier ratio; that is, the lower the ratio, the lower
the extraction power of the aldoxime (see Figure 2),with this helping to operate optimally for
example in circuit with one stripping stage.
Nonylphenol, tridecanol and esters are among the modifiers used today. It should be noted that,
because the type of modifier will determine the level of modification of the oxime, the
oxime/modifier ratio is not comparable for different modifiers. The effect of the addition of the
modifier is shown in Figure 2 for an ester-modified aldoxime. In this case, the aldoxime used was
compound name, and the components in the mixed formulations included compound1 and
compound 2 in ratios of 1,65:1 (Acorga M5640), 1,24:1 (Acorga M5774) and 1:1 (Acorga M5910).

Figure 2 Effect of the addition of a modifier (ester) to an aldoxime

Solvent extraction plants have recognized that modified aldoximes are useful as extractants with
good chemical and physical properties for the Cu production process because of their high
selectivity for Cu extraction in relation to Fe, their high chemical stability (degradation level), their
extraction power over a wide range of Cu concentrations and pH levels and because they do not
interf with the physical performance of the plant. The latter has been exhaustively shown during
the last several decades and is in opposition to the paradigms spread in the industry about the
relationship between modifiers and the formation of crud.

RESULTS AND ANALYSIS

Analysis of the Cu extraction efficiency
In order to carry out a comparative study, four relevant variables were considered for the extraction
efficiency: the Cu PLS concentration, the pH of the PLS, the spent electrolyte acid content and the

237

percentage of extractant. Simulations were carried out for a 2E + 1S traditional series circuit. The
evaluated extractants included a modified aldoxime, an aldoxime-ketoxime mix at a 1:1 ratio and a
ketoxime. Other variables affecting the extraction efficiency were set as follows: O/A extraction 1:1,
O/A stripping 2,5:1, spent electrolyte Cu 38 g/L and stage efficiency (EAE) 95%. The latter was
considered constant for the different formulations, despite the fact that there is a lower EAE when
there is a higher presence of ketoxime in the extractant. (See Figure 2). The results are presented in
Figure 3.
It can clearly be seen in the figure that the modified aldoxime presents an advantage for all ranges
of the different simulated variables. Specifically, when the PLS Cu concentration increased, the
reduction in the extraction efficiency was least for the modified aldoxime. This effect is mainly due
to the greater load capacity of the modified aldoxime compared to that of the mix of pure
compounds and the ketoxime.

Figure 3 Extraction efficiency based on Cu concentration in and pH of the PLS, acid content in the Cu Spent
and the % of extractant for different formulations of extractants.

In addition, when the pH in the PLS decreased, the ketoxime was more affected and had a limited
operation range above pH 1.80. However, even at low pH, the efficiency of the modified aldoxime
was 0,80 percentage points higher than the aldoxime-ketoxime mix and 5,0 percentage points
higher than the ketoxime.
Furthermore, when the acid concentration in the spent electrolyte was below 175 g/L, the ketoxime
performed slightly better than the modified aldoxime and aldoxime-ketoxime mix. However, as the
acid concentration increased, the performance of both the modified aldoxime and aldoximeketoxime mix surpassed that of the ketoxime. This variable is of great importance for the daily

238

operation of a solvent extraction plant due to the fact that this it may be manipulated to reduce the
possible increase in the Cu concentration in the PLS and thus the efficiency of the plant.

Cu/Fe selectivity analysis

Cu:Fe selectivity is of vital importance for the electrowinning process due to it effect on the current
efficiency. Bleeding is performed in order to control the level of this impurity in the electrolyte, but
this method leads to losses and inefficiencies in the process that increase the costs due to higher
consumption of cathodic and anodic reagents, losses of copper and acid (approximately 1% of the
Bleed) and higher consumption of the extractant. Therefore, it is of critical importance to maintain a
high selectivity in the SX, which depends on operational conditions (Cu PLS, pH stability, etc.) and
the selective nature of the extractant (Yaez, 2007).
To carry out the selectivity analysis, isotherms for Fe co-extraction of a PLS sample with high Fe
content taken from a plant were measured for an extractant based on 100% ketoxime, and then
different doses of modified aldoxime were added in order to observe the effect on selectivity. The
characteristics of the PLS are listed in Table 1.
Table 1 Characteristics of the PLS used for the Fe co-extraction tests.
Cu PLS

4.04

pH PLS

2.,0

g/L

Cu Spent

32

g/L

H+ Spent

184

g/L

% extractant

23.5

O/A Ex.

1.,0

O/A Str.

1.80

Fe PLS

20

g/L

The various isotherms generated for Fe co-extraction with various levels of modified aldoxime in
the extractant are show in Figure 4. It can be seen in the figure that there is a significant increase in
the selectivity when the modified aldoxime is introduced to the organic inventory. The Fe coextraction was reduced by a facor of six with respect to that of the ketoxime extractant when the
organic load was between 70% and 80%.

Figure 4 Fe co-extraction isotherms in organic

239

Chemical stability
The chemical stability of an extractant depends on several factors that are independent from the
nature of the oxime. The most important of these factors include the level of acidity, the redox
potential, the temperature and the make-up of the extractant (Hurtado - Guzmn, C. & Menacho, J,
2003).
To analyze the chemical stability of various extractans, accelerated hydrolytic degradation studies
were conducted. Pure ketoxime, pure aldoxime, a 70:30 mixture of ketoxime and aldoxime (used
plant), and a modified aldoxime were added to PLS samples and heated for 80 days with agitation
and the physical variables of the organic phase were monitored. The trial conditions are listed in
Table 2, and the results are presented in Figure 5.
As can be seen in Figure 5, the physical parameters were not significantly affected during the trial.
There was, however, a slight reduction in the viscosity and the interfacial tension when the
extractant was composed of the pure aldoxime. The chemical stability of the extractants was also
evaluated based on the evolution of the maximum load and the percentage of hydrolytic
degradation (generation of aldehydes and ketones). The results are shown in Figure 7.
Table 2 Conditions for the accelerated hydrolytic degradation studies
Cu

25

g/L

H2SO4

300

g/L

% extractant

15

O/A Str.

2.0

Fe T
ClT
Redox

240

4,7
10.,3
40
653

g/L
g/L
C
mv

Figure 5 Evolution of organic phase physical properties accelerated hydrolytic degradation test

Figure 6 Evolution of maximum load and hydrolytic degradation

The evolution of the maximum load clearly corresponds with the degradation of the aldoxime and
ketoxime to aldehydes and ketones, which reduce the load capacity of an extractant for a system
with no make-up. Notably, the pure aldoxime has the lowest stability of all of the extractact
formulations, followed by the 70:30 keto:aldo mix, which is an expected result due to the fact that
the pure aldoxime is less chemically stable. The 100% ketoxime and modified aldoxime reagents
were found to be more chemically stable, which is due to their less reactive nature (ketoxime) and
the capacity of the modifier (ester) to stabilize the pure aldoxime.

Physical performance analysis

There are many external factors that contribute to the formation of crud (solids content in the PLS,
extractant consumption, flows, etc.) which have no relationship with the nature of the oxime
reagents and the modifiers (Tinkler, O. et al, 2008).
In order to study the affinity of silica to the oximes and different modifiers to determine if there is
any external effect, fine thin layer chromatography (TLC) plates were prepared. TLC is based on the
difference in the affinity of a compound to a stationary (absorbent, silica) and a mobile (eluent,
solvent) phase. The silica affinity was compared for the oximes and different modifiers in
commercial use nonylphenol, ester, alcohol and the NR anti-nitration protector. While this affinity
comparison is of a qualitative nature, the ratio of the distances covered by the eluent provides a

241

relative value for the affinity. The higher this number, the lower the affinity between the organic
compound and the silica. The diluents used in the TLC analysis were hexane and toluene, which
faithfully represent the aliphatic and aromatic mixes that comprise the commercial diluents used in
the solvent extraction process.
As can be seen in Figure 7, both in the aromatic (toluene) and aliphatic(hexane) media, the ester, the
nonylphenol and the NR protector showed a lower affinity with silica than the oximes themselves
(greater relative mobility). In fact, due to its non-aromatic nature, the ester had a higher mobility in
both media. Only the alcohol has a higher polarity and, consequently, a greater affinity with silica
than the commercial oximes.

Mobilidad Relativa (%)

100
80
60
40
20
0
Aldoxime

Ketoxime

Ester

Hexano

Nonylphenol

NR Protector

Alcohol

Tolueno

Figure 7 Relative mobility of organic compounds in relation to hexane and toluene diluents

Additionally, batch tests were carried out in order to measure the solid distribution between the
aqueous and organic phases, the phase disengagement times, the interfacial tension, the density
and the viscosity of the organic phase. The test conditions were those standard for the measurement
of the phase disengagement time at a mixture temperature of 20C, an agitation of 1,200 rpm, an
O/A phase ratio of 1:1 and 100 mg/L solids in the PLS.

Organic phase solids

(ppm)

As can be seen in Figure 8, the results of the batch tests confirm that the ester modifier did not have
an effect on the formation of crud, while a 50:50 mixture of pure aldoxime:ketoxime had the
greatest effect Studies are ongoing to determine if there are any possible synergic effects occurring.
50

43,5
36

40

37
25,5

30

26

20
10
0
Ester

Noylphenol Aldoxima Pura Diluyente

Aldoxima modificada
Mezcla Aldo:Ceto 50:50
Cetoxima

Figure 8 Distribution of solids in organic phase

242

25

19

CONCLUSIONS
The following conclusions about the use of modifiers for copper extraction can be made based on
the results of this study:
The modified aldoxime provides a greater efficiency over a wider range of Cu PLS and pH
levels and acid levels in the electrolyte, which allows a better response in the operation of
hydrometallurgical plants.
The high performance of extractants based on modified aldoximes with respect to Cu:Fe
selectivity has been confirmed and enables avoidance of high concentrations of Cu:Fe in the Fe
electrolyte that are detrimental to the current efficiency and increase bleed costs.
The modified aldoxime exhibits good chemical stability and resistance to hydrolytic
degradation, even under aggressive conditions and without any effects on the physical
parameters.
There is no relationship between the formation of crud and the type of extractant; therefore, the
affinity of esters with the solids that are present in the PLS is the lowest of all of the organic
compounds in the organic phase in a plant.

REFERENCES
Hurtado - Guzmn, C. & Menacho, J. (2003).Oxime degradation chemistry in copper solvent extraction plants,
Copper 2003. The Canadian Institute of Mining, Montreal, Quebec, Canad.
Yez, H. Consumo de extractante en plantas de SX, Seminario Cytec (2007), Iquique, Chile.
Tinkler, O. et al, The ACORGA OPT extractante series: comparative studies against aldoxima: cetoxima
reagents. Alta (2008) Copper Session, Perth, Australia.
Trademark Notice: The indicates a Registered Trademark in the United States and the or * indicates a
Trademark in the United States. The mark may also be registered, the subject of an application for registration
or a trademark in other countries.
Disclaimer: Cytec Industries Inc. in its own name and on behalf of its affiliated companies (collectively,
"Cytec") decline any liability with respect to the use made by anyone of the information contained herein. The
information contained herein represents Cytec's best knowledge thereon without constituting any express or
implied guarantee or warranty of any kind (including, but not limited to, regarding the accuracy, the
completeness or relevance of the data set out herein). Nothing contained herein shall be construed as
conferring any license or right under any patent or other intellectual property rights of Cytec or of any third
party. The information relating to the products is given for information purposes only. No guarantee or
warranty is provided that the product and/or information is adapted for any specific use, performance or
result and that product and/or information do not infringe any Cytec and/or third party intellectual property
rights. The user should perform its own tests to determine the suitability for a particular purpose. The final
choice of use of a product and/or information as well as the investigation of any possible violation of
intellectual property rights of Cytec and/or third parties remains the sole responsibility of the user.
2013 Cytec Industries Inc. All Rights Reserved.

243

Regeneration process for copper

solvent extraction reagents

Hydroxyoxime based reagents are used widely in copper solvent extraction

processes to recover copper from sulphuric acid leach solutions. The hydroxy-

Piritta Salonen and James Kabugo.

Outotec Oy, Finland

oxime extractant may degrade due to acid catalyzed hydrolyses and/or manganese and cobalt catalyzed oxidation. Degradation has major negative impact
on plant operation and performance. It decreases copper loading capacity,
and shifts of the extraction isotherms and changes metal selectivity of extractant. Degradation extends the phase disengagement time in settlers making
operation of the plant challenging. Degradation can be verified by gas chromatographic analysis of organic phase samples.
A regeneration process has been developed to re-oximate degraded extractant with a hydroxylamine salt solution in a liquid-liquid reactor in normal copper solvent extraction environment. The process is patented and is
now called Outotec EMT (Extractant Maintenance Technology). The mobile
unit can be operated in continuous operation at the SX plant. Regeneration
restores the extractant, loading capacity and physical performance in the SX
process. Technology has been industrially operated since 2008 and since
then further developed to present modification. The paper reports performance comparison with degraded and fresh reagent and experimental results
of regeneration of degraded reagents from several process operations in
Chile, Zambia and Europe.
The performance of re-oximated organic solutions shows major improvement in copper loading, and the phase disengagement rate has reduced to its
original level. Process operation has been reported to be improved after the
regeneration campaign.

245

INTRODUCTION
The regeneration process enables treatment of degraded reagents with continuous flows at the
plant site. Hydroxyoxime commercial reagents used in recovery of copper values from sulphric
acid aqueous solutions are well known to degrade under normal copper solvent extraction (SX)
operational conditions. (Merigold, 1996) In some instances, reagent degradation can be too high
which results into increased extractant consumption and difficulty in SX process operation due to
adverse effects caused by presence of degradation compounds. Extractant degradation mainly takes
place via two major routes: acid catalyzed hydrolysis and metal catalyzed oxidation of the oxime
group. (Menacho et al., 2003; Cheng et al., 2000) In both cases the eventual primary degradation
products are corresponding aldehyde and ketone compounds according to reaction 1. Also it has
been reported that breakdown of such oxime extractants can also be accelerated under nitrate and
chloride conditions (Kordosky et al., 2003; Carlo, 2001).

(1)

Re-oximation of degraded hydroxyoxime extractant is normally done using hydroxylamine which

reacts with the aldehyde or ketone to form an oxime. (Virning et al., 1999) Hydroxylamine salts and
mainly hydroxylamine sulphate are preferred over the free base. This is because of the hazards
associated with handling, use and storage of free base hydroxylamine. It is known to decompose
even under ambient conditions but its acid salts are more stable and safer to handle (Cisneros et al.,
2004).
The re-oximation reaction has been studied (Ferguson et al. 1991; Virning et al., 1999) and there are
few cases reported where it has been practiced industrially (Mackenzie et al., 1998; Crane et al.,
2009). However in most cases the reaction was slow and also the organic phase after re oximation
showed poor phase separation characteristics which are undesirable in SX operation, but phase
separation properties of re-oximated extractant can be improved to a certain extent with clay
treatment (Crane et al., 2009). Regeneration does not have an effect to nitrated extractant molecules.
This paper reports a continuous flow process for regeneration of degraded hydroxyoxime copper
extractants. In this process a bleed stream or previously stored batch of barren organic from Cu SX
can be fed directly to the regeneration process without further pre-treatment. The re oximation
reaction takes place in liquid-liquid stirred reactors where hydroxylamine sulphate is used to
convert the aldehyde or ketone in the organic phase back to oxime. Soda ash is used for pH
adjustment of aqueous hydroxylamine solution. Since barren organic from Cu SX is often not
completely stripped, presence of copper loaded extractant do not affect the re-oximation reaction
(Ferguson et al. 1991). Clay treatment of the organic phase after re-oximation removes undesired
reaction by-products and completes the regeneration process.

246

METHODOLOGY
Analysis and chemicals
Organic samples from Cu SX and regeneration circuits were first pre-treated with 2 M H2SO4 to
remove metals. The oxime and its hydrolysis degradates were analyzed by an Agilent 7890A 5975C
Gas Chromatograph-Mass Spectrometer (GC-MS) instrument using 30-m long capillary column, He
carrier gas and a non-derivatization analytical method. Hydroxylamine in aqueous solutions was
determined by a direct titration method and analysis done with the use of Mettler Toledo T50
titrator. Oxime strength or copper loading capacity was determined based on the standard
procedure for such commercial hydroxyoxime reagents (Cytec, 2005). Densities of solutions were
measured using Anton Paar DMA 5400M density meter and at room temperature. All chemicals
used in the analysis were analytical grade.

Laboratory batch tests

Several process organic solutions from different Cu SX plants were tested in laboratory and the
degraded extractant regenerated as previously reported (Paatero et al., 2010). The extent of the re
oximation reaction was similarly monitored by GC-MS analysis of oxime and extractant hydrolytic
degradation products. Effectiveness of regenerated reagent was also verified using copper loading
and extraction phase separation tests.

Extraction phase disengagement

The tests were conducted based on the standard method for copper extractants (Cytec, 2005).The
main parameters were: 6.0 g/L Cu in aqueous feed solution, pH 2.0, T=25C, O/A=1 with total
volume of 800 mL, stirring rate 600 rpm and organic phase continuous.

PROCESS DESCRIPTION
The regeneration process has been developed to treat degraded reagents with continuous flows at
the plant site. Process is called Outotec EMT (Extractant Maintenance Technology) and it is
patented by Outotec (Paatero et al., 2009). The process equipment consists of two separate unit;
regeneration and adsorption. Equipments have been installed in two mobile 40ft containers that can
be transported to different site locations. Process has been designed for constant organic flow of 10
m3/h. EMT internal process areas are connected via buffer tank providing flexibility for operation.
The process uses automation system similar to normal process plant and all process steps are
controlled with automation utilizing on-line chemical analysers. All the process equipments have
been selected according to area classification of ATEX (European Union directives 94/9/EY and
1991/92/EY) (Figure 1).

247

Figure 1 Full-size regeneration unit

Figure 2 shows process connections to copper plant. Regeneration unit includes a liquid liquid
reactor for the re-oximation. The adsorption stage includes a mixed adsorption reactor and a
decanter centrifuge for solid liquid separation. Process operates in normal plant conditions.

Regeneration plant

Regeneration

Chemicals

Chemical solution

preparation

Cu SX PLANT

Buffer tank

Adsorption

Adsorbent

Solid residue

Figure 2 Basic schematic process connection

The process containers can be located close to industrial solvent extraction process. Bleed flow of
barren organic is pumped through the re-oximation containers and returned to back to original
process area. The process operation period length depends on the degradation degree and organic
inventory of current plant. Pre-testing requires four litres of plant organic to make laboratory
regeneration tests and to determine the extraction performance of the treated reagent. The most

248

important tests are determination of copper loading capacity and the phase disengagement rate in
copper extraction conditions.

RESULTS AND DISCUSSION

Continuous flow extractant regeneration

1,0

50

0,9

45

0,8

40

0,7

35

0,6

30

0,5

25

0,4

20

0,3

15

0,2

Aldehyde conversion
(fraction)

0,1

Temperature, C

0,0
1

10

Temperature, C

Aldehyde conversion

In this current case the active extractant in the organic reagent was 5-nonyl salicylaldoxime
modified with 2,2,4-trimethyl-1,2-pentanediol diisobutyrate (TXIB). According to GC-MS analysis
the aldoxime extractant had degraded into aldehyde product. Reagent degradation was above 20%
of the initial active extractant. Re-oximation of aldehyde degradate was successfully done using
Outotec EMT process. The re-oximation circuit was operated in aqueous continuity mode and
reaction conversion as high
as 95% was achieved (Figure 3) under steady state operational
conditions. The reaction was fast and even with relatively low hydroxylamine concentration in the
aqueous phase, good recoveries were observed. It also showed that, hydroxylamine can be
consumed in the process to very low concentrations which is economical and environmentally
attractive.

5
0

1 32 32 44 44 44 46 45 45 48 49 51 51 51 53 53
Day

Figure 3 Effect of temperature on the re-oximation reaction during continuous plant operation

It is also important to note that in this case, reaction temperatures as low as 6C were experienced
during plant operation (Figure 3). However, re-oximation reaction was not seriously affected since
relatively high extractant recoveries were observed. Such low temperatures as expected lead to
slow phase separation after re-oximation.
Adsorption purification stage performed very well in regard to restoring the settling rate of the
organic solution. Regenerated organic solution exhibited faster phase disengagement times than the
original process organic (Table 1). Density measurements did not show any unusual behavior after
organic regeneration. Solid breakthrough in the organic phase exiting adsorption purification stage
was monitored and typical assays are given in Table 1. Although the values maybe relatively on the

249

higher side of SX process requirement, solid breakthrough can be optimized before returning
regenerated organic to Cu SX.
Table 1 Results summary of the organic extractant regeneration process
Organic details

Feed

Re-oximated only

Re-oximated with
adsorption
purification

Phase
disengagement
time, s

Conversion,
%

Increase in Cu
loading, %

Density,
kg/L

Solids,
ppm

400

0.8608

1200

1076

84

44

0.8672

1918

120

87

55

0.8633

48

159

95

60

0.8648

67

128

91

66

0.8638

80

80

95

66

0.8641

151

93

72

0.8633

64

Comparison of various plant samples

Authentic samples have been tested and regenerated to compare the performance and behaviour of
organic solutions. The reagent concentrations varied between about 5 30 vol %. In some cases due
to degradation losses extractant make-up had been added. With degradation samples a significant
increase in viscosity was observed in every case.
In each laboratory batch experiment, 1 L of organic was treated for 2 h. The re-oximation reaction
was monitored at different time intervals. The extractant and degradation products (i.e. aldehyde
and/or ketone) in the organic solutions were similarly analyzed by a gas chromatographic method.
According to Figure 4, under the same reaction conditions, different reaction rates were observed
for particular extractant solution. Currently there are no exact reasons for such behavior but slow
reaction rates can be possibly attributed to low concentration of hydrolysis degradation products,
type of degraded extractant (aldoxime or ketone) and presence other organic compounds in the
process organic. Therefore it is important to test a particular process organic in laboratory in order
to optimize extractant regeneration process parameters before the Outotec EMT process can be
considered.

250

1,0

0,8
Conversion

0,6
A
B
C
D
E

0,4
0,2
0,0
0

10 20 30 40 50 60 70 80 90 100 110 120

Time, min

Figure 4 Kinetics of re-oximation time with different plant samples

A summary of laboratory tests results representing different industrial organic solutions is given in
Table 2. It is clear that degraded oxime extractant can be re-oximated and relatively high recoveries
obtained at least within 2 h. Also Adsorption purification method used in the Outotec EMT
process restores and in many cases greatly improves the phase separation characteristics of the
organic phase. Organic physical properties, density and viscosity are not negatively affected after
extractant regeneration.
Table 2 Regeneration results and physical properties of industrial organic solutions
SX
Plant

Re-oximation
conversion, %

Increase in
Cu
loading, %

Phase
Disengagement
Time, s

Density, kg/L

Viscosity, mPas

Feed

Treated

Feed

Treated

Feed

Treated

95

22

134

55

0.849

0.850

3.53

3.73

76

15

116

70

0.842

0.843

3.22

3.33

83

192

125

0.834

0.834

2.92

2.97

94

14

172

123

0.852

0.853

3.97

4.19

95

22

200

153

0.810

0.810

2.07

2.18

91

32

216

82

0.825

0.825

2.80

3.00

CONCLUSIONS
Degraded hydroxyoxime extractants can be regenerated continuously using the Outotec EMT
process under conditions comparable to that of Cu SX. One removable unit has been built for Chile
area. The regeneration of degraded reagents lowers the operational costs and facilitates process
operation. Performed test campaigns showed that both degraded aldoxime and ketoxime based
extractants can be re-oximated however they require different conditions case by case. Since
particular process solutions may behave quite differently towards extractant regeneration, it is

251

important to first conduct laboratory tests in order to optimize the required process parameters.
Economical comparison of operation and benefit needs also to be done prior regeneration campaign.
From the environmental perspective, the process does not generate hazardous waste substances
which would require expensive treatment. However, it is very important to follow safety
procedures while handling, use, storage and disposal of hydroxylamine and its wastes.
Hydroxylamine in the regeneration process can be consumed to low levels and aqueous effluent
treated together with other industrial waste solutions.

REFERENCES
Merigold, C.R. (1996) Copper extractants: Modified and Unmodified Oximes, a Comparison, MID,CNNMIECYunnan Company, BGRIMM, Annual Technical Seminar, Kumming, China, 112. [1]
Menacho, J.M. & Hurtado-Guzman, C. (2003) Oxime Degradation Chemistry in Copper Solvent Extraction Plants,
in Copper 2003-Cobre 2003, Volume VI - Hydrometallurgy of Copper (Book 2), MetSoc, Canada, pp.
719-761. [2]
Cheng, C.Y., Hughes, C.A., Barnard, K.R. & Larcombe, K. (2000) Manganese in Copper Solvent Extraction and
Electrowinning. Hydrometallurgy, vol. 58, pp. 135150. [3]
Kordosky, G. (2003) Copper SX circuit design and operation-current advances and future possibilities, Alta 2003.
Carlo, C.D. (2001) Use of Anti-degradation Agents for Oxime Extractants in the Solvent Extraction of Metals, Patent
Application, WO 01/14604 A1. [4]
Virnig, M.J., Krbechek, L., Casey, M. (1999) Reoximation of Metal Extraction Circuit Organics, U.S. Patent No.
5993757, Nov. 30, 1999. [5]
Cisneros, L.O., Rogers, W.J. & Sam Mannan, M. (2002) Comparison of the Thermal Decomposition Behaviour
for Members of the Hydroxylamine Family, Thermochimica Acta 414, pp. 177-183. [6]
Ferguson, E.S., Price, M.J., Reid, J.G. (1991), Two Phase Oxime Synthesis, AU Patent No.612528. [7]
Mackenzie, J.M.W., Virnig, M.J., Boley, B.D., Wolfe, G.A. (1998) Extraction of Nickel from Ammoniacal Leach
Solutions: Extractant and Solution Chemistry Issues, ALTA 1998 Ni/Co Pressure Leaching &
Hydrometallurgy Forum, Perth,Western Australia, May 1998. [8]
Crane, P., Urbani, M., Dudley, K., Horner, A. & Virnig, M. (2009) Solvent Extraction Reagent Selection for High
Temperature, Acid, Chloride and Cu PLS at Port Pirie and its Impact on Electrowinning,
Hydrometallurgy Conference 2009, The South Afr. Inst. of Mining and Metallurgy, pp. 283-294. [9]
Cytec Technology Note (2005), ACORGA Extraction Reagents: Standard Test Methods. [10]
Paatero,E., Virolainen, S. Kortelainen, J., Jyrkk, K.(2010) Regeneration of degraded Hydroxyoxime reagents, ISEC
2010, Santiago de Chile, Chile, October 2010. [11]
Paatero,E., Jyrkk, K., Virolainen, S. (2009) Methods and Apparatus for Regeneration of Extraction Solution in Metal
Extraction Processes, FI Patent No. 122939, PCT Pat.appl. WO2010/142841. [12]

252

Metallurgical performance of
small and medium size copper SX
plants: Relevance of the SX
reagent formulation

More than twenty small and medium size copper solvent extraction (SX) plants
are operating in Chile. Production of copper as copper cathodes or as copper

Philippe Joly and Francisco Reyes.

BA SF Chile S.A .

sulfate at these plants varies between two and thirty thousand tons per year.
Almost all of these plants produce an excellent quality copper product at
a competitive cost. While there is a difference in scale relative to larger SX
plants, many operating parameters are similar or equivalent; such as mixing
retention times, specific flows in settlers, electrolyte bleed, SX reagent make
up, diluent addition rate, electrical power usage and others.
Some of these plants operate at high chloride level in leaching. Others
apply non-traditional SX configurations and some of them use one or more
wash stages. The operating conditions are diverse and may change rapidly
which often requires adapting the choice of SX extractant to each plant.
The objective of the paper is to compare several small and medium size
SX plants in terms of operation and design. Emphasis is on the importance
of the SX reagent choice and the flexibility to adapt the extractant formulation
capability when the chemical conditions of the plant are changing, or when
there is no more room for improvement from the design or operational
parameters.

253

INTRODUCTION
Two countries have developed a significant amount of small and medium size copper solvent
extraction plants in the last 15 years: Chile and China (Li, 2010).
Currently, twenty six small and medium size copper SX plants are operating in Chile. The
production capacity and SX configurations of each plant are shown on Table 1 (Hein & Joly, 2011).
Cathode or copper sulphate production from these plants varies between two and thirty thousand
tons of copper per year. The geographical location of these plants covers an area in northern Chile
from Iquique to Rancagua, 1800 km north-south and 100 km east-west.
BASFs Global Mining Solutions unit has been providing LIX reagents with technical service to
most of these plants. The present paper will review specific topics related to SX configurations, PLS
characteristics, production capacity, extractant selection and others. Emphasis is done on the
importance of the SX reagent choice and the flexibility to adapt the extractant when the operating
parameters of the plant are changing.

SX CONFIGURATION
Number of strip stages
The current trend in SX design is to use a single strip stage. Nineteen of the SX plants use a single
strip stage while treating a pregnant leach solution (PLS) containing up to 8 g/L Cu with a pH range
between 1.2 and 2.5.
Seven plants were designed with two strip stages. Only one SX plant really needed a second strip
stage, as all others operate with a low acid PLS, having a pH above 1.8 and a copper tenor below 4
g/L Cu.

Number of wash stages

The main purpose of a wash stage is to reduce the electrolyte bleed caused by non-extractable
species such as chloride, manganese and/or iron transferring by entrainment with the loaded
organic from leaching to electrowinning and by ferric iron chemically loaded onto the organic.
Eighteen plants operate with a PLS containing high concentrations of anions (sulphate, chloride)
and cations (iron, manganese, aluminium, magnesium). High ionic content is due to the
composition of the ore and/or industrial water. Among them four plants operate with chloride
concentrations in the range of 35 to 80 g/L and some plants located in the Second Region of Chile
report nitrate presence in PLS.
Thirteen plants have at least one wash stage, particularly in the north of Chile, where the salt
content of the PLS is typically high.
Two plants have two wash stages in a series arrangement, which proved to be very effective. In one
of these plants, the PLS contains 50 g/L chloride and often more than 100 ppm total suspended
solids (TSS), nevertheless the chloride concentration in electrolyte is maintained at 15-20 ppm.

254

Another plant has a single wash stage combined with a coalescer. The plant operates well up to 80
g/L chloride in PLS with the occasional difficulty in maintaining a low chloride concentration in
electrolyte.
Table 1 List of small and medium size copper solvent extraction plants located in Chile (2013)
Plant

Region

Capacity

Configuration

Note

TPA
ENAMI / Taltal

II

3,000

Series + S + 2 W

Seawater

Trinidad / Nora Plant

III

3,000

Optimum series parallel or

triple series +S+W

Water from tailings dam

San Geronimo / S. Lorenzo

IV

3,000

Series + crystallisation

Copper sulphate

Molymet / Copper Plant

RM

3,000

Triple series + 2S + W

SX Cu downstream of SX Mo

ENAMI / Vallenar

III

5,000

Opt. series par. +S+W

Outstanding Cu quality

ENAMI / Minera Cruz

IV

5,000

Flexible 4 E + S

Cerro Dominador /
Sta.Margarita

II

10,000

Series + S +W

CENIZAS / Oxide plant

II

10,000

Flexible 4 E + S+ 2 W

Seawater, nitrate

ENAMI / El Salado

III

10,000

Series + S + W

Salt water

Pucobre / Biocobre

III

10,000

Series + S

Low reagent consumption

ENAMI / M.A. Matta

III

10,000

Triple series +S+W or

optim. series parallel

CEMIN / Dos Amigos

III

10,000

Opt. series par. + 2 S

Cerro Negro / Oxide

10,000

6 E Parallel + S

Anglo / El Soldado

10,000

Semi parallel + 2 S

Sierra Miranda

II

15,000

Opt. series parallel or

CEMIN / Amalia plant

15,000

4E+S+W

Anglo / Los Bronces

RM

15,000

Series + 2 S

High altitude

Codelco / El Teniente

VI

15,000

2E+S

AMD + smelting dust

HMC / Haldeman Sagasca

II

20,000

Series + 2 S + W

Codelco / Sulfuros BL

II

20,000

Series + S

TECK / CD. Andacollo

IV

20,000

Semi parallel + S

Low Cu grade PLS

VALE / Tres Valles

IV

20,000

4 E Double series par. + S

Two different PLS

Mantos de la Luna

II

30,000

Series + 2 S + W

Seawater

Franke / KGHM

II

30,000

Opt. series par. +S+W

CODELCO / Salvador

III

30,000

Series + S + W

MLCC / Caserones

III

30,000

Triple series +S+W or.

series parallel

or triple parallel

First Optimum S/P

configuration

triple parallel +2S+W

Flexible interlaced

High Altitude, 4000 m

255

Extraction schemes
Extraction staging is chosen based on PLS characteristics, capital investment, organic loss and
copper recovery. Sometimes leaching and solvent extraction have to be optimised jointly. In other
cases, SX configuration is selected based on space availability and/or plant expansion decisions.
Series
(example of complete circuit)
PLS

Triple series

SE

Semi parallel
PLS

PLS

SO

E1

LO

E1

E2
RF

LO

Conventional series parallel

PLS

E1
LO

PLS

E2

E3

RF1

RF2

LO

SO

E1

LO

PLS

E2

E3

RF1

RF2

SO

LO

Double series parallel

PLS

PLS

E2

E3

E4

RF1

RF2

E1
SO

LO

PLS

RF1

E3

PLS

PLS

E1

E2

E3

RF1

RF2

RF3

SO

Optimum triple parallel (4E)

PLS

E2

SO

RF2

Triple parallel

PLS

E1

Interlaced
PLS

LO

SO

Optimum series parallel

PLS

E2

E1

E4

RF2

SO

LO

PLS

PLS

E2

E3

E4

RF1

RF2

RF3

SO

Note: PLS = Pregnant Leach Solution, RF = Raffinate, LO = Loaded Organic, SO = Stripped Organic, SE = Spent Electrolyte, PE =
Pregnant Electrolyte, E=??

Figure 1 Main configurations applied in extraction (Hein, 2003)

In reference to PLS copper concentration, fourteen plants operate with a PLS copper tenor in the
range of 4 to 9 g/L and 12 plants operate with less than 4 g/L Cu (Table 2). One of these plants
operates at less than 1 g/L Cu.
Figure 1 summarizes the most common extraction schemes. Some of these configurations can
change over time based on operating conditions and plant flexibility.
Series and triple series are widely used when copper concentrations in the PLS are above 4 g/L Cu.
Series configuration with two extraction stages is a low cost option and generally leads to extraction
efficiencies of 85 to 93%. Adding a third extraction stage in series can improve efficiencies up to
98%.
Five plants use the optimum series parallel configuration, decreasing plant investment cost. The
PLS flow is divided into two parallel flows; consequently, the mixer settler size is reduced. Series

256

parallel extraction schemes are conveniently designed with flexibility, such that triple series or
triple parallel configurations may be applied whenever the PLS tenor moves up or down.
The semi parallel configuration is useful for maintaining the production level of a series circuit,
whenever PLS copper concentration drops and additional extraction stages become too expensive.
Finally, four plants use four extraction stages. Such SX configuration design should be highly
flexible, allowing for a switch between double series parallel (2 PLS flows), interlaced (2 PLS flows)
and optimum triple parallel (3 PLS flows) configurations.
Table 2 Number of plants classified by SX configuration and PLS copper tenor
Extraction scheme

PLS Cu conc.

Total

< 4 g/L

> 4 g/L

Series

11

Optimum series
parallel

4 E flexible

Semi parallel

Triple series

Other
TOTAL

12

14

26

Table 3 Number of plants classified by SX configuration and type of reagent

Main SX
configuration

Ketoxime /
aldoxime blend

Series

Optimum series
parallel

4 E flexible

Semi parallel

Triple series

Pure
ketoxime

Pure
aldoxime

Modified
aldoxime

Other
TOTAL

1
18

SOLVENT EXTRACTION REAGENT

SX reagent selection
The selection of the extractants is important during the development of any SX/EW plant, especially
with regard to the overall economic performance of small and medium size SX plants. Reagent
selection is based on extractant characteristics, looking for high net transfer and recovery, low

257

reagent concentration, minimum reagent loss and suitable crud generation. Technical service
provided by reagent suppliers is another important factor where reagent formulation adjustment
capability is desired.
Most of the plants (75%) use a non-modified ketoxime/aldoxime blend, mainly LIX 9790N and
LIX 984N (Table 3). Bigger plants may optimize their reagent formulation by varying the
proportions of pure ketoxime (LIX 84-IC) and aldoxime (LIX 860N-IC) added to the circuit.
A few plants operate with modified aldoxime like LIX 612N-LV or Acorga M5774, M5640

Figure 2 Reagent formulations based on Ketoxime, Nonyl Aldoxime, Dodecyl Aldoxime and various kinds of
modifiers sourced from BASF Global Mining Solutions

Examples of active SX reagent formulations

The examples hereafter show how an effective reagent formulation may help the SX plants to adapt
to changes and optimize their operation. The formulation of the organic inventory does not
normally change over short periods of time. In fact, it may take up to several years to dilute
compounds from a previously used reagent.

Case A: Increase in PLS sulphate concentration

The increase in PLS sulphate concentration well above 250 g/l increased the PLS viscosity and
organic losses by entrainment. But it also increased the pH buffer effect of the sulphate, which
allowed the use of a reagent with better stripping characteristics, even at a pH as low as 1,3 in PLS.
As the plant purchases separately pure components LIX 84-IC and LIX 860N-IC, the ratio of
Ketoxime/Aldoxime was easily changed from 1:1 to 3:1, improving the stripping reaction and
lowering reagent concentration and consumption by more than 20 %.

258

Case B: PLS pH flexibility

A new mineral had a very low acid consumption, which raised the concern that the pH of the PLS
could go down to 1,0 from an initially expected 1,8. The plant opted for the flexibility of purchasing
pure components LIX 84-IC and LIX 860N-IC. The ratio Ketoxime/Aldoxime was adjusted
depending on the effective PLS pH, allowing the plant to maintain the technically economical
optimum conditions.

Case C: High chloride concentration in electrolyte

With a chloride concentration in the PLS above 60 g/L, the plant had seen an increase in chloride
entrainment in the electrolyte above 60 ppm. The design of the plant has been improved and the
previous modified aldoxime reagent has been replaced by a lower viscosity non modified ketoxime
/ aldoxime blend, LIX 9790N, lowering the chloride entrainment in the electrolyte down to 15 ppm
(Rojas et al., 2010).

Case D: Extremely low copper concentration in PLS

The Cu concentration decreased to less than 0,5 g/L in the PLS and pH increased above 2 due to
lower mine production. The plant adjusted the reagent concentration and the PLS pH to help
maintain the Cu/Fe selectivity. LIX 684N-LV, a low viscosity modified aldoxime, was proposed to
be part of the organic formulation. This formulation has also been proposed for other plants
(Peralta &Araya, 2005).

Case E: High crud generation and acidity in PLS (>20 g/l)

Years ago, the previous modified aldoxime reagent has been replaced by a pure non modified
aldoxime, resulting in lower crud generation and extraction efficiencies higher than 90 % even with
an extremely high acidity in PLS (> 20 g/L acid). Now that the PLS acidity has lowered, a nonmodified ketoxime / aldoxime blend was proposed to improve copper transfer and lower reagent
use.

Case F: Improving cathode quality

A plant needed to increase the copper and lower the acid concentration of the electrolyte
maintaining the strip eficiency and the viscosity of the organic. A proposal has been made to add
ketoxime to the existing modified aldoxime, in order to lower the strip point of the formulated
organic.

Case G: Plant expansion

The SX configuration of the plant has been converted from series to optimum series parallel to
increase the production, adding an extraction mixer settler. At the same time, the organic
formulation was switched from LIX 984N to LIX 9790N, bringing a higher net transfer with this
new configuration due to the higher ketoxime content (Ibacetta & Garrido, 2003).

259

SX reagent consumption
SX reagent consumption usually varies between 1.0 and 3.5 kg/ton Cu, depending on PLS viscosity,
suspended solids in PLS, reagent characteristics, plant design and operational parameters. Most of
the plants are able to maintain SX reagent consumption below 2.5 kg/ton Cu, which is remarkable.
While suspended solids above 50 ppm in PLS typically affect operation by increasing crud
generation and entrainments, some plants operate at levels above 500 ppm with acceptable results.
In recent years, the PLS viscosity at four SX plants increased to 5 cP and up to 10 cP (at 25C). This
resulted in high organic entrainments until viscosities were brought under control.
The incidence of the SX reagent make up on the C1 production cash cost is generally less than 1 %.

CONCLUSIONS
The LeachSXEW technology can be easily and economically scaled down to as low as 2,000 TPA
copper cathode production depending on the ore resource and its location.
In Chile, existing engineering capability, technical supply support and skilled operators lead to
efficient plants with outstanding copper quality.
The great variety of SX designs and operating parameters at the 26 small and medium Chilean SXEW plants provide a broad range of experience. Of particular interest is the successful operation of
plants using seawater, high suspended solids in PLS, different types of reagents, novel SX
configurations and occasionally two wash stages connected in series.
The flexibility provided by the formulation of the organic composition, based on the various kinds
of SX reagents, is a must for optimization of plant operation.

REFERENCES
Forest, L. The application of Copper SX technology in China Gold smelter, LIX Users Conference 2010,
Puerto Varas, Chile.
Hein, H. & Joly, PH. Metallurgical performance and characteristics of small and medium size copper SX
plants in Chile, Proceedings of the 19th International Solvent Extraction Conference, Santiago, Chile, 3-7
October 2011.
Hein, H. Flexibility trends in copper solvent extraction, Hydrometallurgy 2003, 5th International Symposium,
Vancouver, B.C., Canada, 24-27 August.
Rojas, E., Olgun, D., Valds, R. & Gutirrez, R. Synergy Between Operational Practices And Mix Of
Extractants In Copper Industry History Of Case, 8 LIX Users Conference, Puerto Varas, September
5-9, 2010.
Ibaceta, D., Garrido, J., Hein, H., Araya, G., SX Development at Dos Amigos in Northern Chile, ALTA 2003
Conference, Perth, Australia, May 19-24, 2003
Peralta, R. & Araya, G. Cambio de Configuracin en SX y mezcla optima de extractante, Segundo encuentro
internacional de hidrometalurgia, UCN, Antofagasta, Chile, 04-05-2005

260

Removal of some heavy metals from a

sulfate-containing residual mining
solution using nano-structured
calcium silicates

The aim of this research is to achieve the removal of Cu(II), Zn(II) and Cd(II)
from a waste mining solution with a high sulfate content using nano-structured
calcium silicate. The sorbent was easily synthesized by a chemical reaction

Fernando Valenzuela, Jaime Sapag,

Carlos Basualto and Luis Verdugo.
Universidad de Chile

between Ca(OH)2 and a solution of Na2SiO3 being obtained as insoluble and

amorphous precipitate. A higher stirring intensity during the synthesis generated a finer structure of the prepared solids. The calcium silicate was characterized using different methodologies confirming its amorphous character. It
was shown to have particles whose mean size was 0.5 to 1.0m having an average
surface area of 333.0m2/g. The pores on its surface were very fine with a mean
pore diameter that varied between 15.8-23.6nm and a pore volume between
0.7853-1.5441cm3/g. X-ray diffraction analysis confirmed the amorphous or
polycrystalline character of prepared compounds; patterns associated with
wollastonite, CaSiO3, and larnyta-syn, Ca2SiO4 were detected.
In uptake studies Cu(II), Zn(II) and Cd(II) ions were adsorbed readily and
completely from an acidic mine waste water which contained over 5g/L of
sulfate, probably due to the formation of insoluble hydroxides with the hydroxil and silanol groups associated with the calcium silicates. The amount
of metals adsorbed was much higher than those taken up by many other
natural and synthetic sorbents, especially when compared with the forced
precipitation with lime. The adsorption equilibrium of the metal ions on
nano-structured calcium silicate was examined. The Redlich-Peterson adsorption isotherm matched the experimental results for all metal ions very
well. A high adsorption rate of metals onto nano-structured calcium silicate
was observed. The experimental results were satisfactorily well explained by
applying a pseudo-second-order kinetics model.

This book, both digital and hard-copy, only includes

the abstract of this work.

261

chap. 5
Electrometallurgical processes
and electrochemistry

Electrochemistry of enargite:
Reactivity in alkaline solutions

The presence of enargite (Cu3AsS4) is problematic in gold processing because

it is environmentally hazardous due to its arsenic content but also because it
is double refractory causing both cyanide and oxygen consumption to increase.
Two prevalent categories of hydrometallurgical processes for enargite-bearing
materials exist: The co-dissolution of arsenic and copper under acidic condi-

Nick Gow. FLSmidth, USA

Courtney Young and Hsin-Hsuing
Huang. Montana Tech, USA
Greg Hope. Griffith University,
Australia

tions and the selective leaching of arsenic under alkaline conditions. Selective
leaching and treatment of the arsenic would prove advantageous prior to
applying conventional gold leaching methods. To this purpose, the reactivity
of enargite samples from Butte, Montana and Quiruvilca, Peru were spectroelectrochemically studied under alkaline conditions, pH range of 8-13, using
Raman spectroscopy and cyclic voltammetry. Raman spectra of the surface
were collected during and after voltammograms were potentially cycled in
order to determine and compare surface species against theoretical Eh-pH
diagrams determined from mass-balanced thermodynamic calculations using
StabCal software. Under slightly oxidizing conditions, covellite (CuS) peaks
were found in a short matter of time, above ~100mV vs. SHE for pH 9-13,
suggesting arsenic leaching specifically occurred. At longer conditioning
times, elemental sulfur was observed and tended to passivate the surface.
The presence of sulfur suggests that copper was eventually leached. By operating above pH 12, under slightly reducing conditions near -300mV vs. SHE,
arsenic can be selectively leached as thioarsenate (AsS43-) or arsenate (AsO43-)
without sulfur formation.

265

INTRODUCTION
Environmental limitations on pyrometallurgical processing of ores and concentrates bearing arsenic
and other hazardous elements often outweigh those in hydrometallurgical processing, particularly
when minerals like enargite (Cu3AsS4) are present. Hydrometallurgical processing of enargitebearing materials can be broken down into two categories: the co-dissolution of arsenic and copper
under acidic conditions and the selective leaching of arsenic under alkaline conditions. The
selective leaching of arsenic would preferably leave behind copper-rich sulfide phases, typically
chalcocite (Cu2S) and covellite (CuS), which then become more amenable to conventional copper
(and gold) processing. Furthermore, selective leaching often employs sulfide (S2-), bisulfide (HS-) or
polysulfides (Sx2-) to first extract the arsenic as AsS43- and AsS33- (Anderson & Twidwell, 2008;
Tongamp et al., 2010).
In this study, Raman spectroscopy and cyclic voltammetry were used simultaneously to look at
surface reactions of enargite electrodes under alkaline conditions. Specifically, for this study, a
spectroelectrochemical technique was used at pH 8-13 with potentials cycled between
approximately -1000 and +1000 mV vs. SHE. Precise control of these conditions along with various
species concentrations will help in understanding how enargite behaves in solution as well as
developing a hydrometallurgical separation process.

Thermodynamics
A database of free energies of species in the Cu-As-S-H2O system has been compiled primarily
using the MINTEQA2 (Allison, 1991) database but partially supplemented from other sources
including the U.S. Geological Survey (Robie, 1984). Those values taken not listed in MINTEQA2
and USGS are listed in Table 1. According to Craig and Barton (1973), there are five Cu-As-S
sulfosalts; however, only enargite, sinnerite (Cu6As4S9) and tennantite (Cu12As4S13) have listed free
energies. Because enargite and luzonite have the same chemical formula, they have different crystal
structures and are referred to as polymorphs.
Table 1 Free energy values for Cu-As-S species not found in the MINTEQA2/USGS databases.
Species
Cu3AsS4 Enargite (Craig, 1973)
Cu12As4S13 Tennantite (Craig, 1973)
Cu6As4S9 Sinnerite (Craig, 1973)

G (kcal)
-49.808
-180.908
-110.256

Species
Cu(AsO2)2 (Nishimura, 1993)
Cu3As Domeykite (Barin, 1995)
H3AsSO2 (Helz, 2008)

G (kcal)
-169.2
-3.184
-103.412

StabCal, a thermodynamic stability calculation program (Huang, 2011), was used to create massbalanced EH-pH diagrams with a Cu-As-S molecular ratio for enargite of 0.75:0.25:1 (see Figure 1).
Mass balancing constrains thermodynamic calculations making them more accurate which is
critical for ternary and higher order systems.
There are a few differences in the diagrams when compared to those previously published by
Padilla et al. (2008), namely the inclusions of the minerals olivenite (Cu2AsO4OH) and domeykite
(Cu3As). There are also differences between the two diagrams and are attributed to sulfur
oxidation. Figure 1A was created with consideration of sulfate species whereas Figure 1B was
created without. In this regard, Figure 1A predicts enargite oxidation to sulfate and Figure 1B

266

predicts oxidation to elemental sulfur. The shaded regions in each diagram are areas where solid
arsenic species do not exist and would therefore determine conditions where arsenic could be
selectively leached. Theoretically, according to these thermodynamic calculations, there is a region
where this would occur and, for the most part, it is independent of the sulfur oxidation state and
occurs under basic conditions (~pH > 9) but slightly oxidizing (~EH > -500 mV).

Figure 1 Mass-balanced EH-pH Diagrams for the Cu-As-S system for As-containing species only
A. Oxidation to Sulfate and B. Oxidation to Sulfur

267

Raman spectroscopy
Enargite samples were obtained from Butte, MT and Quiruvilca, Peru. Both samples were
characterized with Scanning Electron Microscopy/Mineral Liberation Analysis (SEM/MLA). The
Peru sample is ~98% pure with 78% minor antimony (Sb) substitution. By comparison, the Butte
sample was ~83% enargite but with only 55% of minor Sb-substitution; however, the sample did
contain significant amounts of chalcopyrite (14.5%).
Once the samples were characterized, sections measuring approximately 1xx were cut
from each. After a copper wire was attached to the back side of the enargite samples using
Electrodag 502 and carbon tape, the assemblies were mounted into polytetraflouroethylene (PTFE)
tubing and then sealed in the tubing using a non-conductive epoxy resin. Conductivity
measurements showed the mineral electrodes were ready for the spectroelectrochemical studies.
Raman spectra and cyclic voltammograms collected from the two electrodes yielded the same
results and were virtually indistinguishable. Consequently, it did not need to be reported which
results were obtained from which sample/electrode. In between these spectroelectrochemical
measurements, the exposed x face was polished with 200-grit SiC sandpaper and cleaned
with de-ionized water.
Raman studies were conducted at Griffith University in Brisbane, Australia. In-situ studies were
conducted using a Renishaw Raman 100 InVia spectrophotometer. The system has a 100-mW HeNe laser as the excitation source (632.8 nm), a MultiPhysics 1000-line grating, 2 Rayleigh-scatter
filters, and a 256x512 CCD array as the detector. Raman spectra were calibrated using the 520 cm-1
primary band of pure silicon. The system programming was controlled by the GRAMS
spectrometer software suite. Because of the selected optics, detection of peaks below 200 cm-1 was
not entirely reliable and, as a result, peak characterization was presumably conducted on the less
intense peaks at higher wavenumbers. The pH was controlled by buffers that were created
according to the CRC Handbook (Haynes, 2011).
Ex-situ spectra were collected using a Renishaw Raman RM2000 InVia spectrophotometer, also
with a 632.8 nm HeNe laser, using a Peltier-cooled CCD detector. The attached Leica LMDM
microscope allows for a rapid visual inspection of the electrode surface with up to 600x
magnification. Because the enargite electrodes and surface species were easily burned under full
laser intensity, the power was reduced to 10% or less. In addition, the number of scans was
increased to 5 with >10 second exposure to maximize the Raman signal. The system was controlled
by the Renishaw WiRE 3.2 software package.

Cyclic voltammetry
Liquid samples were contained in a 5-port electrochemical cell. The cell was clipped into a holding
bracket, which allowed horizontal and vertical adjustment to center the focal point of the laser. The
port closest to the laser was used for the enargite electrode while the others were used for the
Ag/AgCl saturated KCl (+0.197mV vs. SHE) reference electrode and a platinum counter electrode.

Static tests
Initially, the pH and potential conditions were held constant for various defined points of interest
on the Cu-As-S EH-pH diagram while spectra were collected for up to 24 hrs. These tests were used
to identify Raman peak positions of the major mineral phases. The electrodes were held under

268

potentiostatic control until removed from the solution and then dried under nitrogen gas. Spectra
of the surface were collected with the RM2000 at 10% laser power.
The reference peaks for enargite appear at 265, 338, 384, 679 and 725 cm-1 (see Figure 2). For
consistency, these spectra and others were compared against the RRUFF Raman mineral
spectroscopic database (Downs, 2006). Tennantite (Cu12As4S13), chalcocite (Cu2S), covellite (CuS),
olivenite (CuAsO4OH), elemental copper, elemental sulfur, and pyrite (FeS) were identified in this
manner. Copper (II) hydroxide formation was identified via precipitation during testing. Tennanite
was identified by the reversal of peak intensities for peaks at 338 and 384 cm-1. There are no known
spectra for sinnerite or lautite, and none of the experimental results could be attributed to them
based on common bonds.
Sulfur on enargite was identified first at pH 9 and 0 mV vs. SHE as shown in Figure 3. The
reference peaks for elemental sulfur at 150, 220 and 470 cm-1 are included for comparison. The
surface usually indicated a CuS phase, often with a characteristic peak at 470 cm-1 before the sulfur
peaks would build in intensity. The presence of sulfur was followed by a drop in current density
which then indicated no further reaction occurring and likely passivation of the surface.

Figure 2 Raman spectra for the enargite samples used in this study

Figure 3 Sulfur growth on enargite surface (pH 9, 0 mV, 60 min)

269

Multifiles
The multifile capability combine a continuously scanning program for the collection of Raman
spectra which occurs every five seconds and are linked to the cyclic voltammograms. The
electrodes were cycled three times, scanning initially in either the negative or positive directions
but typically at a scan rate of 5 mV/sec between -1000 and +1000 mV. There is a time delay between
changes in the current density slope and when the Raman detected composition changes in order to
maximize Raman signals. However, it is noted that the Raman spectra are therefore collected over a
range of potentials. In this regard, slow scan rates or constant potentials are preferred so that the
Raman spectra can be linked to the voltammograms, and therefore the EH-pH diagrams, as best as
possible.
Cycling the electrode gave relatively the same pattern of voltammograms, independent of cycle
number. A typical example is shown in Figure 4 for voltammograms obtained at pH 8 at a scan rate
of 5 mV/sec with an initial potential of -200 mV and lower and upper potentials of -1200 and + 800
mV vs SHE, respectively.

Cycle

A3
A2

A1

1
2
3

(A5)

(A4)

C1
C2
C3

C4

Figure 4 Peru enargite sample at pH 8, scanning negatively from -200 mV, 5 mV/s scan rate

In this example, resulting peaks on the cathodic scan suggest enargite reduces to tennantite then
chalcocite and finally elemental copper as depicted in the following reduction reactions and
observed from the Raman spectra collected simultaneously:
4Cu3AsS4 + 3H+ + 6e- = Cu12As4S13 + 3HS-

C1

Cu12As4S13 + 7H+ + 6e- = 6Cu2S + 4As + 7HS-

C2

Cu2S + H+ + 2e- = 2Cuo + HS-

C3

It is noted that neither sinnerite (Cu6As4S9) or domeykite (Cu3As) were observed as predicted
thermodynamically in the EH-pH diagrams presented earlier in Figures 1A and B. In this regard, an
EH-pH diagram will be redetermined without conside ring these species as presented and discussed

270

later. Unfortunately, Raman spectra could not confirm arsenic metal, realgar (AsS) or orpiment
(As2S3) as reaction products but Reaction C2 is the only thermodynamically acceptable prediction.
After reversing the scan direction, Figure 4 also reveals anodic peaks which correspond to copper
oxidation to chalcocite, then covellite, and finally elemental sulfur as depicted in the following
oxidation reactions and observed from the Raman spectra collected simultaneously:
2Cuo + HS- =

Cu2S + H+ + 2e-

A1

Cu2S + HS- =

2CuS + H+ + 2e-

A2

CuS + 2H2O =

Cu(OH)2 + So + 2H+ + 2e-

A3a

Interestingly, not only were the two sulfides spectroelectrochemically observed but so was cupric
hydroxide (Cu(OH)2) precipitate. Olivenite was not observed until potentials reached ~500 mV vs
SHE. In this regard, it is suggested that anodic reaction A3 also be represented as:
2CuS + AsO43- + H2O = Cu2AsO4OH + 2So + H+ + 4e-

A3b

such that the arsenate (AsO43-) likely comes from the oxidation of arsenic metal
As + 4H2O = AsO43- + 8H+ + 5e-

A4

where the realgar is presumed to be present as a result of cathodic reaction C2. However, it is also
conceivable that enargite itself oxidizes to arsenate and precipitates out as olivenite and covellite:
Cu3AsS4 + 5H2O = Cu2AsO4OH + CuS + 3So + 9H+ + 9e-

A5

Because Raman spectra showed elemental sulfur (So) formed at each of the pH values investigated
in this study, Reactions A3-A5 are all written with it as a product. This shows that, at least in the
time frames examined in this study, enargite oxidizes to element sulfur and not sulfate.
It is important to note that these results strongly show that arsenic can be leached from enargite to
form covellite and elemental sulfur. Periodically, the sulfur could not be observed until more than
1 cycle was conducted allowing for enough to accumulate. When that happened, the surface was
observed to passivate. More importantly, the sulfur tended to accumulate more at pHs above 10
where enargite is not stable. Under these conditions, enargite could not be regenerated during the
voltametric studies. However, when the pH was below 10, the enargite surface tended to
regenerate as evidenced by the cathode peak C4 in Figure 4 but more species than just enargite was
observed.

Updated thermodynamics
Based on the spectroelectrochemical results just presented, some modifications have been made to
the EH-pH diagrams in Figure 1. Consequently, olivenite will continue to be considered and
domeykite (Cu3As), cuprite (Cu2O) and tenorite (CuO) will not. Furthermore, in regards to kinetics,
only sulfide oxidation to elemental sulfur was allowed such that no sulfoxy ions such as sulfate
were considered, Figure 5. In this regard, leaching conditions should be kept > pH 12 with EH
between approximately -700 mV where thioarsenate (AsS43-) is stable and +100 mV where arsenate
(AsO43-) is stable.
All voltammograms were then re-examined for inflection points in which increased current flow
was observed, indicating an electrochemical reaction was starting to occur. The potentials where

271

this occurred were plotted on the updated EH-pH diagram and results are presented in Figure 6.
Clearly, the potentials line up well with thermodynamic expectations and much better than would
occur in Figure 1B. This suggests the diagram is accurate giving high confidence for the results
presented in this study. It is noted that a series of points occur near 0 mV vs SHE at pH 10-13 and
are attributed to the formation of nonstoichiometric copper sulfides as chalcocite oxidizes to
covellite. These points are not observed at lower pH values because chalcocite is not stable under
these conditions (in the presence of arsenic). As noted earlier, the nonstoichiometric copper
sulfides were stable according to the thermodynamic calculations but were not included in the EHpH diagrams in Figures 1A, 1B, 5 and 6 for clarity purposes.

Figure 5 Updated EH-pH Diagram for the Cu-As-S system with region of selective searching

272

Figure 6 Updated EH-pH Diagram for the Cu-As-S system overlaid with CV inflection points

CONCLUSIONS
Enargite (Cu3AsS4) electrodes were spectroelectrochemically studied under alkaline conditions at a
pH range of 8-13 using Raman spectroscopy and cyclic voltammetry simultaneously in order to
determine and compare surface species against theoretical EH-pH diagrams determined from massbalanced thermodynamic calculations. Under slightly oxidizing conditions, covellite (CuS) peaks
were found in a short matter of time, above ~100mV vs SHE for pH 9-13, suggesting arsenic
leaching specifically occurred. At higher potentials and longer conditioning times, elemental sulfur
was observed and tended to passivate the surface. The presence of sulfur suggests that copper was
eventually leached. By operating above pH 12, under slightly reducing conditions, elemental sulfur
formation would be avoided between approximately -700 mV vs SHE where thioarsenate (AsS43-)
forms and +100 mV vs SHE where arsenate (AsO43-) forms. An average potential near -300 mV vs
SHE is therefore suggested to selectively leach arsenic from enargite and form covellite or
chalcocite. Spectroelectrochemical results match up very well with a newly calculated massbalanced EH-pH diagram.

REFERENCES
Allison, J., Brown, D. & Novo-Gradac, K. (1990) 'MINTEQA2/PRODEFA2 A Geochemical Assessment Model
for Environmental System', V. 3, Environmental Research Laboratory, EPA.
Anderson, C. & Twidwell, L. (2008) 'Hydrometallurgical Processing of Gold-Bearing Copper Enargite
Concentrates', Canadian Metallurgical Quarterly, vol 47., no 3., pp. 337-345.
Barin, I. (1995) Thermochemical Data of Pure Substances, New York, NY VCH Publishers, Inc.
Craig, J. & Barton, P. (1973) 'Thermochemical Approximations for Sulfosalts, Economic Geology, vol. 68, pp.
493-506.
Downs, R. (2006) The RRUFF Project: an integrated study of the chemistry, crystallography, Raman and
infrared spectroscopy of minerals, Program and Abstracts of the 19th General Meeting of the
International Mineralogical Association in Kobe, Japan.
Gow, R., Young, C., Huang, H., Hope, G., & Takasaki, Y., (2013a) 'Spectroelectrochemistry of Enargite II:
Reactivity in Acidic Solutions', submitted to Min. Met. Proc.
Gow, R., Young, C., Huang, H. & Hope, G., (2013b) 'Spectroelectrochemistry of Enargite III: Reactivity in
Alkaline Sulfide Solutions', submitted to Min. Met. Proc.
Haynes, W. (2011) CRC Handbook of Chemistry and Physics, 92nd Edition, CRC Press.
Helz, G. & Tossell, J. (2008) 'Thermodynamic Model of Arsenic Speciation in Sulfidic Waters: A Novel Use of
Ab Initio Computations', Geochemica et Cosmochimica Acta, vol. 72, pp. 4457-4468.
Huang, H. (2011) 'STABCAL', Metallurgical & Materials Engineering, Montana Tech.
Nishimura, T., Itoh, C. & Tozawa, K. (1993) 'Equilibria of Cu(II)-As(III,V) H2O System at 25C', Bull. Res. Inst.
Min. Dress. Metall., vol 49, pp. 61-70.
Padilla, R., Rivas, C. & Ruiz, M. (2008) 'Kinetics of Pressure Dissolution of Enargite in Sulfate-Oxygen Media',
Metallurgical and Materials Transactions B, vol 39B, pp. 399-407.

273

Tongamp, W., Takasaki Y. & Shibayama, A. (2010) 'Selective Leaching of Arsenic from Enargite in NaHSNaOH Media', Hydromet., vol 101, pp. 64-68.

274

Effects of organic impurities on

zinc electrowinning

The effects of organic compounds used in the production of zinc from silicate
ores on (i) process parameters of the electrowinning step, (ii) morphology and
crystalline structure of the deposit, and (iii) electrochemistry of the relevant
cathodic reactions (zinc deposition and hydrogen evolution) are discussed. A
flocculant and a flotation collector, which may be present at residual amounts
in the feed solution of industrial tankhouse, were selected for investigation.
Our approach included electrochemical methods such as electrowinning,

Daniel Majuste and Virginia

Ciminelli. Universidade Federal de
Minas Gerais and the National
Institute of Science and Technology
on Mineral Resources, Water and
Biodiversity (INCT Acqua), Brazil
Eder Martins and Adelson de Sousa.
Votorantim Metais Zinc, Brazil

cyclic voltammetry and linear sweep voltammetry; organic analysis by Fourier

transform infrared spectroscopy; and zinc analysis by scanning electron
microscopy and X-ray diffraction. The analysis of the organics indicated polar
groups in the corresponding molecules, which can form hydrogen bonds with
water and interact with electroactive species (H+ and Zn2+ ions). The electrowinning experiments revealed that current efficiency (CE) drops significantly
when 50 and 100mg/L of the collector are added to the solution. The analysis
of the zinc deposits indicated modifications in the nucleation and growth
patterns of hexagonal platelets and reduction in crystal size, when both organics were added to the solution. Regarding the effects of these impurities on
the cathodic processes, the addition of the flocculant caused a minor increase
in the nucleation overpotential (NOP) for zinc deposition and slight decrease
in the corresponding cathodic current only at 100 mg/L. When the collector
was added, it was difficult to establish the NOP value, but a drastic negative
effect on the cathodic current was observed. With regards to the effects of these
impurities on the hydrogen evolution rate, a slightly beneficial effect was observed. The findings obtained here led to a better understanding regarding the
activity of organic molecules in zinc concentrated, acid solution, which may
be helpful for developing best practices and new additives.

275

INTRODUCTION
During zinc electrowinning, the presence of impurities in the sulphate electrolyte may affect the
quality of the metal electrodeposited on the aluminium cathode and current efficiency, particularly
when the impurities are present at high concentrations. While the effects of metal impurities have
been the theme of many previous investigations, the effects of organic impurities have received less
attention. Some investigations revealed that organic impurities may be harmful to zinc deposition.
The negative effects include the diminution in the current efficiency and modifications in the crystal
structure of the deposits (Mackinnon et al., 1980; Mackinnon et al., 1988; Hosny, 1993; Mackinnon,
1994; Alfantazi & Dreisinger, 2003; Dhak et al., 2011), which may affect its mechanical properties
and, as a result, the effectiveness of the stripping stage.
The activity of organic compounds in electrolyte solutions has been generally described in terms of
adsorption. The adsorption of organic molecules on the cathode surface may affect the kinetics of
electron transfer due to the blocking of a fraction of the active sites available for metal nucleation or
interactions among the electroactive species H+ and Zn2+ ions and the adsorbed molecules. When
the adsorption degree on the metal surface increases, a large blocking of the active sites occurs,
thereby affecting the kinetics of zinc deposition (Eq. 1) and hydrogen evolution (Eq. 2). A higher
driving force is then required for deposition, which may be evidenced by the polarization of the
cathode (Oniciu & Murean, 1991), which often result in an increase of the cell voltage.
Zn2+(aq) + 2 e-

Zn(s)

(1)

2 H+(aq) + 2 e-

H2(g)

(2)

Winand (1992) has discussed the effects of organic compounds during electrowinning on the basis
of their affinity for water. The author has claimed that when the organic is physically or chemically
adsorbed on the metal surface and has no affinity for water, it may cause strong inhibition in the
deposition reaction. In contrast, if the organic has affinity for water, it may cause small inhibition
or, eventually, it may act as an additive. However, it is essential to assume that atoms with partial
negative charges in organic molecules (typically F, O, N, and S) may also interact with positively
charged species, such as H+ and Zn2+ ions, thus enhancing their migration to the cathode (negative
terminal of the electrowinning cell). It seems obvious that the magnitude of the interaction among
heteroatoms and Zn2+ ion may has serious impacts on the deposition rate.
This research is then motivated by the recognition that the behaviour of ions and organic molecules
in electrolyte solutions is complex and difficult to describe in details, but a better understanding on
the activity of specific organic groups in the zinc concentrated, acid solutions may be helpful for
developing best electrowinning practices or even additives with better performance. Due to the
upcoming trend of low-grade, complexes ores, it is expected a gradual rise of unreacted residual
amounts of organic molecules in the feed solution of electrowinning tankhouses, mostly when zinc
silicate ores are treated. The hydrometallurgical route for zinc silicate ores does not include a high
temperature operation such as roasting, in which organic compounds added in prior stages, such as
flotation, are normally degraded. In this context, this paper discusses the effects of selected organics
used in the production of zinc from silicate ores on (i) current efficiency (CE), cell voltage (CV), and

276

specific energy consumption (SEC) during electrowinning; (ii) morphology and crystalline structure
of the deposit; and (iii) electrochemistry of the cathodic reactions.

METHODOLOGY
Organic compounds
Only the effects of two organic compounds are discussed in this paper. Table 1 provides the role of
such organics in the industrial process and characteristic functional groups, which were detected by
Fourier transform infrared spectroscopy (FTIR). It can be seen that Organic 1 (flocculant) exhibits
polar primary amide (RCONH2) and non-polar alkyl (RCH3) groups in its molecule. Organic 2
(collector) exhibits polar ether (ROR) and primary amine (RNH2) groups, and non-polar
alkyl group. It is known that the amide and amine groups can form hydrogen bonds with water
molecules. The partial negative charges available on the O and N atoms for amide, and N atom for
amine, may also attract the H+ and Zn2+ ions, assuming no steric hindrance effects. It is important to
observe that the solubility of organic compounds containing the RCONH2 and RNH2 groups
typically decreases as the size of the hydrophobic carbon chain increases. In addition, the extension
of the interaction heteroatoms-ions also depends on the size and type of carbon chain. With regards
to ether group, slightly polar, the intermolecular interactions are low, as well its affinity for water.
Table 1 Role of the organic compounds in the industrial process and detected functional groups.
Organic
1

Role
Flocculation of solids
(Thickening stage)

Collector of zinc
(Flotation stage)

Alkyl

Functional groups (by FTIR)

Amide

Alkyl

Amine

Ether

Electrochemical experiments
Electrolyte solutions
The electrolyte solutions consisting of 60 g/L zinc (Zn2+) and 160 g/L sulphuric acid (H2SO4) were
prepared with deionized water (Millipore Milli-Q), analytical grade H2SO4 96% (FMaia) and zinc
sulphate heptahydrate >99% (ZnSO4.7H2O) (Sigma-Aldrich). The temperature of the solutions was

277

kept constant at (38 0.5) C by a temperature-controlled hot plate (IKA RET basic). The organics
were added to the solution in the following concentrations: 0, 50, and 100 mg/L.

Electrowinning
Electrowinning experiments were conducted in a four-electrode cell: three Al (99.5% pure) cathodes
and central Pt (99.99% pure) anode, in fixed cell geometry. The cell consisted of a 2 L Pyrex reactor.
Prior to each experiment, fresh cathode surfaces were prepared by wet mechanical polishing with
SiC paper (#1200) and rinsing with deionized water in ultrasonic bath for 15 min. The experiments
were carried out at constant current density (500 A/m2) for 6 h, under stirred conditions (350 rpm).
During electrowinning, CV measurements were undertaken using a Minipa (ET 2615A) multimeter.
The average SEC, in kwh/t Zn, was calculated from CV measurements at 2, 4, and 6 h, as given by:

nFCV
SEC =

3.6MCE

(3)

where n is the number of electrons in the overall deposition reaction (2), F the Faraday constant
(96485 C/mol), M the zinc atomic weight (65.39 g/mol), and CE the current efficiency. After that, the
deposits were manually stripped from the cathodes, rinsed with deionized water in an ultrasonic
bath for 15 min and then dried in a furnace at 60 C for 30 min. The % CE was determined by:

nFm
CE =
x100%
ItM

(4)

where m represents the weight of the deposit obtained after 6 h (g), n, F and M as described above, I
the total applied current (C/s), and t the deposition time (s).

Cyclic voltammetry and linear sweep voltammetry

Electrochemical measurements were conducted in a three-electrode cell (Al (99.5% pure) working
electrode, Pt (99.99% pure) counter-electrode and Ag/AgCl/KCl 3 mol/L reference electrode, in fixed
cell geometry) using Autolab (PGSTAT 20) potentiostat. The cell consisted of a 1 L Pyrex reactor.
Prior to each measurement, fresh working electrode surfaces were prepared by wet mechanical
polishing (SiC paper #1200) and rinsing with deionized water in ultrasonic bath for 15 min. All the
potentials were converted to the Standard Hydrogen Electrode (SHE) scale (i.e., +0.197 V at 38 C).
By potentiometry, the open circuit potential (OCP) on the working electrode was measured during
10 min. The nucleation overpotential (NOP) of zinc cathodic reduction (i.e., the difference between
the initial potential of nucleation and the reversible potential of the Zn2+/Zn couple) was obtained
by cyclic voltammetry. The measurements were started at -0.600 V vs. SHE, scanned in the negative
direction, and reversed to the positive direction at -0.905 V vs. SHE. The cathodic current associated
with the hydrogen evolution rate was obtained by linear sweep voltammetry. The measurements
from acid solution without Zn2+ ions were started at -0.600 V vs. SHE and scanned to -0.905 V vs.
SHE. All the measurements were undertaken at 1 mV/s under stirred conditions (350 rpm).

278

Characterization of the zinc deposits

The morphology of the deposits was analyzed by scanning electron microscopy (SEM) with energy
dispersive spectrometry (EDS), using a JEOL (JSM 6360 LV) microscope and Thermo Noran (Quest)
spectrometer. The images were obtained at 15 kV. The structure of the deposits (i.e., the preferred
crystal orientations) was examined by X-ray diffraction (XRD), using a PANalytical (Empyrean) Xray diffractometer, with Cu K1 ( = 1.5406 ) incident radiation. The XRD patterns were measured
at a 2 range of 10-120 with step size of 0.06, and identified by comparing the d-spacings obtained
with standard reference patterns, supplied by the International Centre for Diffraction Data (ICDD).

RESULTS AND DISCUSSION

Effects of the organic impurities on the process parameters: CE, CV, and SEC values
The average CE and SEC values obtained from the organic-containing solutions are summarized in
Table 2. The average CE value obtained in the absence of organics was 93.67%, while the average
SEC value was 2644 kwh/t Zn. In the presence of Organic 2, significant negative impacts may be
observed. When 50 mg/L was added, CE dropped from 93.67% to 21.58%, and SEC increased from
2644 to 11672 kwh/t Zn. For Organic 1, CE was 0.47 (50 mg/L) and 1.19 (100 mg/L) higher than that
obtained for the organic-free solution. This compound also caused slightly beneficial effects on SEC.
Table 2 Average CE and SEC values obtained from solutions containing Organics 1 and 2.
CE (%)
Organic
1
2

50 mg/L
94.14 0.04
21.58 0.08

100 mg/L
94.86 0.04
21.09 0.09

SEC (kwh/t Zn)

50 mg/L
100 mg/L
2636 8
2594 5
11672 9
11996 31

Obs.: In the organic-free solutions, CE = (93.67 0.05) % and SCE = (2644 3) kwh/t Zn.

SEC is proportional to CV (Eq. 3), which is a function of the equilibrium potentials for the anodic
(Eanode) and cathodic (Ecathode) reactions, their associated overpotentials (anode and cathode) and the
ohmic potential drop through the cell (Eq. 5).
CV = (Eanode + anode ) (Ecathode cathode) + IL/A

(5)

where I is the current (A), L the interelectrode distance (m), A the electrode reaction area (m2), and
the specific conductivity of the electrolytic solution (S/m). Voltage losses in the electrical contacts of
the cell were neglected. Assuming that Eanode, Ecathode, I, L and A are constants, then CV only depends
on , anode and cathode values.
Figure 1 exhibits the average CV values measured at 0, 2, 4 and 6 h. It can be seen that the addition
of Organic 2 to the solution increased CV at both concentrations. An electrostatic interaction was
assumed to occur among the negatively charged Al cathode and the positively charged molecules
of Organic 2. In the solution, the RNH2 group interacts not only with H3O+ ion, but also with Zn2+

279

ion, producing milky agglomerates. The steadiness of higher CV values along the experiment may
be related to the activity of these agglomerates at the interface. With regards to the effect of Organic
1, its addition slightly decreased CV at t = 0 h, and such drop persisted only at 100 mg/L, an effect
which remains to be better investigated. Further observation made from CV measurements (from t
= 2 h) is a gradual, very slight decrease in CV over time. This trend was found to be unaffected by
the type of organic. According to Eq. 5, a diminution in CV during electrowinning could be related
to a slight increase in the value due to a stoichiometric balance (consumption and production) of
ionic species, and a slight decrease in the anodic and cathodic overpotentials. It has been reported
that reduction of Zn2+ ions is faster on Zn substrate than on Al substrate (Recndiz et al., 2007).

Figure 1 Effects of the organics on the CV values. Legend:

0 mg/L;

50 mg/L;

100 mg/L.

Effects of the organic impurities on the product quality

The effects of the impurities on the product quality were evaluated by morphological and structural
analysis of the Zn deposits. Figure 2 shows typical micrographs of samples obtained from organicfree solutions and solutions with 50 mg/L of Organics 1 or 2. For organic-free solutions, hexagonal
platelets randomly oriented were observed (Fig. 2a). When Organic 1 was added, nodular deposits
were observed. The platelets grew randomly oriented in a tower-type structure (Fig. 2b). This effect
is likely related to the competition between the vertical growth (two-dimensional nucleation) and
the lateral growth of the initial layers (Winand, 1992). In the presence of Organic 2 in the solution,
platelets perpendicularly aligned to the surface were observed (Figs. 2c). SEM/EDS analysis also
provided evidences on product contamination. Residual amounts of zinc sulfate (ZnSO4), as needlelike structures, were detected on deposits obtained from solutions with Organic 1. However, after
thoroughly rinsing, most of the salt has been removed. Furthermore, addition of both organics to
the solution caused grain refinement; the highest effect was observed when Organic 2 was added.
All the factors that promote nucleation or disorder in the growing crystal contribute to reduce the
grain size of metal deposit. Although the adsorption of organic molecules on the Al surface may
decrease the density of nucleation sites due to blocking, adsorbed foreign molecules inhibit surface
diffusion of ad-atoms towards growing centers, and create disorder during the incorporation of adatoms into the crystal lattice (Oniciu and Murean, 1991).

280

(a)

(b)

(c)

Figure 2 Typical SEM micrographs of zinc deposits obtained from organic-free solutions (a); solutions
containing 50 mg/L of Organic 1 (b) and Organic 2 (c).

Regarding the structural analysis by XRD, only hexagonal zinc (ICDD 04-0831) was detected. The
preferred crystal planes developed from organic-free solutions are (1 1 4), (1 0 3), (1 1 2), (1 0 2), and
(1 0 1). When Organic 1 was added, the order of preferred planes changed to (1 0 1), (0 0 2), (1 0 0),
(2 0 1), and (1 0 2). The addition of the flocculant suppressed largely crystal growth in the (1 1 4), (1
0 3), (1 1 2) and (1 0 2) directions, and it promoted the development of pyramidal (1 0 1) and basal (0
0 2) planes. Also, an intense growth of crystals in the perpendicular (1 0 0) direction was evidenced,
which supports the tower-type structure reported by SEM (Fig. 2b). When Organic 2 was added to
the solution, the order of preferred planes changed to (1 1 0), (1 0 1), (1 1 2), (1 0 0), and (1 0 2). The
presence of this impurity in the solution fully suppressed the growth of (1 1 4) and (1 0 3) planes.
The detection of the perpendicular (1 1 0) and (1 0 0) planes among the preferred ones support the
findings obtained by SEM (Figs. 2c).
Therefore, the adsorption of organic molecules on the cathode surface affects not only the process
parameters but also the morphology/structure of the product. The adsorbed molecules may prevent
the growth of Zn platelets in one specific direction, inducing different growth rates for the various
crystal planes.

Effects of the organic impurities on the electrochemistry of the cathodic reactions

The measurements by the potentiodynamic methods were conducted after determining the OCP
values in organic-free and organic-containing solutions. If compared to the value obtained from
organic-free solutions (i.e., -0.563 0.001 V vs. SHE), the OCP slightly increased in the presence of
Organic 1 and slightly decreased when Organic 2 was added. The results obtained by linear sweep
voltammetry indicated from organic-free, acid solution, that cathodic current increases gradually as
the overpotential (i.e., difference between applied potential and OCP value) increases. The average
cathodic current recorded at the cathodic limit was (1.06 0.00)x10-2 A. Addition of the flocculant to
the solution slight increased hydrogen reduction rate. The average current recorded at the cathodic
limit increased about 17% (50 mg/L) and 19% (100 mg/L). With regards to addition of Organic 2, it
favored hydrogen evolution only at high overpotentials, which represent more practical conditions.
The average current increased about 31% (50 mg/L) and 42% (100 mg/L) at the cathodic limit.
The main hypotheses for explaining the slight positive effects of organic molecules on the kinetics
of the hydrogen evolution are: (i) effects on the hydrogen discharge mechanism it is known that
desorption of adsorbed hydrogen molecules as bubbles is affected by the surface tension of the

281

electrolyte (Gabe, 1997); and (ii) minor interval by which the hydrogen bubbles remain attached to
the electrode surface. When organic compounds are added to acid solution, its surface tension may
be reduced and, thus, desorption of adsorbed hydrogen bubbles from the metal surface would
occur readily. In this case, the hydrogen bubbles would cause less impact on the electron transfer
rate and, then, the current associated with hydrogen evolution increases. Such hypotheses along
with the formation of milky agglomerates in the solution, which have effects on the interfacial
viscosity, could explain the effects observed when Organic 2 was added to the solution.
The results obtained by cyclic voltammetry are shown in Figure 3. From organic-free solutions, it
can be seen that polarization goes through a region of low cathodic current until zinc deposition
commences. The cathodic current, then, increases until the potential in which the scan is reversed.
In the positive direction the current decreases and reaches zero at the reversible potential. The NOP
value obtained from organic-free solutions was (74 1) mV. The cathodic peak current (due to the
simultaneous reduction of Zn2+ and H+ ions on the Al surface) was (3.18 0.03)x10-2 A. The addition
of Organic 1 (Fig. 3a) slightly increased the average NOP value from (74 1) mV to (79 3) mV at
100 mg/L. This minor shifting of the deposition potential indicates a slight polarization of the Al
cathode. It means that the presence of flocculant in the solution has a very low inhibiting effect on
zinc nucleation, which may be related to a low adsorption degree of molecules on the electrode. At
100 mg/L of Organic 1, Zn2+ reduction occurs at unblocked active sites and, thus, current decreases
with respect to that recorded from organic-free solution; the cathodic peak current decreased about
21%. At 50 mg/L, no shifting of the deposition potential was observed, and a very slight decrease of
the peak current (about 2%) was demonstrated. Such effect and the slight increase in the hydrogen
evolution rate when the flocculant was added to the solution do not support the increase of CE
during electrowinning (Tab. 2). These effects are expected to be balanced by faster reduction of Zn2+
ions on Zn substrate in comparison to Al, and slower evolution of H2(g) on Zn in comparison to Al.
0.04

0.04

(b)

0.03

-Cathodic current (A)

-Cathodic current (A)

(a)

0.02

0.01

0.00
-0.76

-0.78

-0.80

-0.82

-0.84

-0.86

Potential (V / SHE)

-0.88

-0.90

-0.92

0.03

0.02

0.01

0.00
-0.72

-0.76

-0.80

-0.84

-0.88

-0.92

Potential (V / SHE)

Figure 3 Voltammograms recorded from solutions containing (a) Organic 1 and (b) Organic 2.
Legend: 0 mg/L; 50 mg/L; 100 mg/L.

The voltammograms obtained from solutions containing Organic 2 are shown in Fig. 3b. It can be
observed that this impurity caused a large decrease in the cathodic peak current: about 98% and
97% at 50 mg/L and 100 mg/L, respectively. Such negative impacts explain the drastic effect of this

282

compound on the CE value (Tab. 2). In addition, it was complex to establish the NOP value, even
minimizing largely the Y-axis scale. Thus, this impurity has a strong inhibiting effect on the kinetics
and mechanism of Zn2+ reduction, and from the reported increase in the hydrogen reduction rate, it
can be concluded that the drastic effect of Organic 2 is predominantly driven by the negative effects
on zinc reaction (Eq. 1). This distinguished effect could be related to a strong interaction among the
negatively charged N atoms of the RNH2 group and the Zn2+ species.

CONCLUSIONS
This work discussed the effects of organic compounds used in the production of zinc from silicate
ores on the main parameters of the electrowinning stage, as well on the morphology and structure
of the metal deposit, and on the mechanism of the cathodic reactions. A flocculant and a flotation
collector, which may be present at residual amounts in the feed solution of industrial tankhouse,
were selected for investigation. The results indicated that a special care should be taken to avoid
contamination of the feed solution with the collector. An important drop in the CE value along with
a significant increase in the CV and SEC values were observed when this impurity is present in the
solution. Contamination with the flocculant appeared to be unimportant from the CE, CV and SEC
viewpoints. This impurity may in fact act as an additive, enhancing production. Nevertheless, the
addition of both organics to the solution caused modification in the growth pattern of zinc crystals.
The findings obtained by using potentiodynamic methods coupled with organic analysis by FTIR
technique led to a better understanding on the role of these impurities in the electrochemistry of the
cathodic processes. Under practical conditions, the activity of the organic molecules at the cathodeelectrolyte interface may cause a slight beneficial effect on the kinetics of hydrogen evolution. In the
presence of the flotation collector, zinc deposition was largely affected, while in the presence of the
flocculant, the reduction reaction is slight inhibited only at high concentrations. Advances in the
knowledge on the activity of polar organic molecules in zinc electrolyte solutions may be helpful
for developing best practices and new additives.

ACKNOWLEDGEMENTS
The present work has been developed in the scope of the AMIRA Project P705B Improved anode
and cathode processes in the electrowinning of base metals. The first authors would like to thank
the AMIRA International Ltd. and all the sponsors for agreeing with this publication. The authors
are also grateful to CAPES, CNPq, FAPEMIG, and INCT-Acqua for financial support.

REFERENCES
Alfantazi, A.M., & Dreisinger, D.B. (2003) An investigation on the effects of orthophenylene diamine and
sodium lignin sulfonate on zinc electrowinning from industrial electrolyte, Hydrometallurgy, vol. 69,
pp. 99107.
Dhak, D., Mahon, M., Asselin, E., & Alfantazi, A. (2011) The effects of mixtures of acid mist suppression
reagents on zinc electrowinning from spent electrolyte solutions, Hydrometallurgy, vol. 108, pp. 110.
Gabe, D.R. (1997) The role of hydrogen in metal electrodeposition processes, Journal of Applied
Electrochemistry, vol. 27 (issue 8), pp. 908-915.

283

Hosny, A.Y. (1993) Electrowinning of zinc from electrolytes containing anti-acid mist surfactant,
Hydrometallurgy, vol. 32, pp. 261-269.
Mackinnon, D.J. (1994) The effects of foaming agents, and their interaction with antimony, manganese and
magnesium, on zinc electrowinning from synthetic acid sulphate electrolyte, Hydrometallurgy, vol. 35,
pp. 11-26.
Mackinnon, D.J., & Brannen, J.M. (1977) Zinc deposit structures obtained from high purity synthetic and
industrial acid sulphate electrolytes with and without antimony and glue additions, Journal of Applied
Electrochemistry, vol. 7, pp. 451-459.
Mackinnon, D.J., Brannen, J.M., & Lakshmanan, V.I. (1980) The effects of chloride ion and organic extractants
on electrowon zinc deposits, Journal of Applied Electrochemistry, vol. 10, pp. 321-334.
Mackinnon, D.J., Brannen, J.M., & Morrison, R.M. (1988) The effect of thiourea on zinc electrowinning from
industrial acid sulphate electrolyte, Journal of Applied Electrochemistry, vol. 18, pp. 252-256.
Mackinnon, D.J., Morrison, R.M., Mouland, J.E., & Warren, P.E. (1990) The effects of saponin, antimony and
glue on zinc electrowinning from Kidd Creek electrolyte, Journal of Applied Electrochemistry, vol. 20,
pp. 955-963.
Oniciu, L., & Murean, L. (1991) Some fundamental aspects of leveling and brightening in metal
electrodeposition, Journal of Applied Electrochemistry, vol. 21, pp. 565-574.
Recndiz, A., Gonzlez, I., & Nava, J.L. (2007) Current efficiency studies of the zinc electrowinning process on
aluminum rotating cylinder electrode (RCE) in sulfuric acid medium: Influence of different additives,
Electrochimica Acta, vol. 52, pp. 68806887.
Robinson, D.J., & OKeefe, T.J. (1976) On the effects of antimony and glue on zinc electrocrystallization
behavior, Journal of Applied Electrochemistry, vol. 6, pp. 1-7.
Tripathy, B.C., Das, S.C., Hefter, G.T., & Singh, P. (1997) Zinc electrowinning from acidic sulfate solutions.
Part I: Effects of sodium lauryl sulfate, Journal of Applied Electrochemistry, vol. 27, pp. 673-678.
Tripathy, B.C., Das, S.C., Hefter, G.T., & Singh, P. (1998) Zinc electrowinning from acidic sulphate solutions.
Part II: Effects of triethylbenzylammonium chloride, Journal of Applied Electrochemistry, vol. 28, pp.
915-920.
Winand, R. (1992) Electrocrystallization theory and applications, Hydrometallurgy, vol. 29, pp. 567-598.

284

SELE Technology: An alternative

to boosting current efficiency and
cathode quality in EW plants

Selective Electrodeposition Enhancer (SELE) technology is used in the electrodeposition process. At the core of this solution is a custom-made, fibreglass
and modified-plastics device that is inserted into copper electrolytic cells at

Pedro Aylwin, Nicols Lagos and

Patricio Melani. New Tech Copper
SpA, Chile

electrowinning (EW) plants in order to keep electrodes confined in a fixed, parallel

position, while also preventing short circuits and boosting current efficiency.
Additionally, SELE technology improves cathode quality by yielding a
finer-grain, homogeneous deposit while also doing away with the need for
edge strips and anode spacers.
SELE technology makes it possible for EW plants to operate at a high
current density without impairing current efficiency or cathode quality. In
the future, corporate decisions concerning the expansion and/or construction
of new EW plants will take this feature under advisement.
Five years of operating SELE technology in EW plants at current densities
as high as 450A/m2 have yielded excellent operating results, as a direct consequence of reducing the current density variability from one electrode to the
next and within electrodes. These results are directly aligned with the continuous improvement path chartered by the mining industry, making SELE
technology an ideal tool to take the EW process down that path.
This article outlines some of the main features of SELE technology while
focusing on how the SELE solution contributes to reducing the distribution
dispersion of two key factors involved in the EW process, i.e., current efficiency
and cathode weight. Finally, also included herein is a brief analysis illustrating
how much a greenfield project is able to reduce its CAPEX by preferring SELE
technology over conventional alternatives.

285

INTRODUCTION
In the electrowinning (EW) process, copper ions derived from heap leaching and concentrating
copper ore through solvent extraction, are turned into metallic copper on the surface of the cathode
when a current passes through the electrolyte between two electrodes (cathodes and anodes).
Copper deposits obtained through EW have a purity of 99.999% or greater.
The EW process is carried out in large tankhouses containing several electrolytic cells, which for
copper deposition purposes are typically equipped with 60 pairs of electrodes plus an extra anode
and are usually operated at a current density of 320 A/m2 with a voltage potential of two volts
between electrodes. Current efficiency ranges from 82% to 92%, and energy efficiency losses are
primarily associated with short circuits and undesired products resulting from side reactions (Bard,
Stratmann, McDonald, & Schmuki, 2007; Robinson, Jenkins, Rasmussen, King, & Davenport, 2003)
Cathode quality is a key factor in copper production. Copper cathodes can be rejected for a variety
of different reasons; shorts, nodulation and extremely thin deposits are perhaps three of the most
recurring causes. The latter two are directly related to a great deal of variability in cathode weight,
as observed in conventional processes.
Indepth studies aiming to develop more efficient and less-polluting electrowinning processes have
been conducted by Wiechmann et al. (Wiechmann, Morales, & Aqueveque, 2010), Fornari et al.
(Fornari & Abbruzzese, 1999) and Gopala et al. (Gopala & Das, 1992).
By spatially confining the electrodes, SELE technology reduces the variability of the key
parameters involved in the process, such as cathode weight and current efficiency, thereby
improving cathode quality and energy efficiency.
This article examines how variability of current efficiency and cathode weight has improved at an
EW plant equipped with the state-of-the-art Selective ELectrodeposition Enhancer (SELE)
technology versus the commonly-known conventional technology in which an electrowinning cell
is equipped with anode spacers, cathodes edge strips, wishbone busbars, and a cell-attached
electrolyte manifold. Technical data and specific measurements described in this paper were
gathered at an actual tankhouse.

Figure 1 Left, Electrode Loading; Center, Cathode Harvest; Right, Cathode Pallet with SELE Technology

286

SELE Technology
SELE technology is a modular system consisting of a core element and multiple modules capable
of improving different aspects of the electrodeposition process. Each SELE devicecomprised of
its core elements plus modulesis expressly designed, manufactured and tailor-made according to
each EW plants individual specifications and parameters. The device is installed in EW cells and is
intended to secure electrodes in a fixed, parallel position, thereby preventing short circuits,
increasing current efficiency, enhancing cathode quality and improving process performance.

Figure 2 Installation of SELE Technology

SELE technology and its modular system add value by:

Reducing specific energy consumption
Improving cathode quality up to 98%
Increasing current efficiency
Operating efficiently at a high current density
Reducing operational costs; increasing asset life span
Doing away with the need for edge strips and anode spacers
Increasing cathode weight uniformity

Figure 3 Left, Cathodes Obtained with Conventional Technology; Right, Cathodes Obtained with SELE
Technology

This technology also makes it possible to expand production capacity with minimal
implementation time.

287

METHODOLOGY
Taguchi analysis
The Taguchi method was developed by Genichi Taguchi to improve product quality. It is based on
the premise that no amount of inspection can improve a product; therefore, quality must be
designed in throughout the different stages of the process. Robust design is an engineering
methodology for improving productivity through research and development; thus, high-quality
products can be produced rapidly and at low cost (Phadke, 1989).
The traditional approach follows a step function to guarantee performance is kept within upper and
lower specification limits. Moreover, Taguchiss loss function states that the existence of an ideal
performance value and any deviation from such value represents a cost which can be expressed in
terms of a quadratic relationship:

L = k ym

(1)

Accordingly, Taguchis loss function penalizes the deviation of a parameter from a nominal or
target value. This analysis does not contain operation ranges, since there is only one single
operation point where there is no loss. With the above in mind the process aims to work with that
parameter value. There are clearly a number of expressions referring to cases where higher
parameter values are preferable as well as expressions for cases where lower parameter values are
preferable (Sharma, Cudney, Ragsdell, & Paryami, 2007).
When considering a large number of experiments, the average economic loss for each case
described above can be expressed as:

L = k + m

(2)

L=k + 1

(3)

RESULTS AND DISCUSSION

A statistical analysis was performed on current efficiency and electric current data collected from
cells equipped with SELE Technology and Conventional Technology at an EW plant over a
three-month period. Results revealed that the parameters studied come from normal distributions
with a confidence level greater than 90% for both processes.

288

100

SELE Technology
Conventional Technology

Current Efficiency %

96
92
88
84
80
Fe

21

3M

ar

13

Ma

r
23

Ma

2A

pr

12

Ap

r
22

Ap

Figure 4 Current Efficiency as a Function of Time

350

Current Variance [A]

300

Current variance, SELE Technology

Current variance, Conventional Technology

250
200
150
100
50
0
F
21

eb

ar
3M

13

Ma

r
23

r
Ma

2A

pr

12

Ap

r
22

Ap

Figure 5 Current Variance with Conventional-Technology Cells and SELE Technology Cells

Faradays Law depicted below was used to convert data from the electric current domain into the
copper cathode weight domain:
I=

nFM
t

(4)

Current efficiency
The Taguchi study on the variables under analysis employs the statistical parameters obtained from
the previous data. In terms of current efficiency, higher values than target values are preferable.
Reaching 93% efficiency would be the nominal value for plant operations. Current efficiency data
obtained for both technologies can be plotted in a graph such as the one illustrated in Figure 6,
showing the results of current efficiency reached by both technologies during the EW process,

(5)

!", $%%

&''

"

= k S) + ) 1 )

(5)

289

(5)

L*&+&

!", $%%

&''

"

= k S + 1

(6)

Then
(5)

!", $%%

L*&+&

!", $%%

&''

&''

"

= 9.5

(7)

"

The economic loss experienced with conventional technology is 9.5 times greater than the economic
loss recorded under SELE Technology, essentially ensuring benefits associated with one of the
most critical performance parameters in EW. Likewise, when using SELE Technology the process
becomes more robust which, in turn, renders it more stable in the likelihood of potential changes in
other uncontrollable variables.
0,5

250
Conventional Technology
SELE Technology

0,3

150

0,2

100

0,1

50

0,0
80

84

88

92

Process Economical Loss

Taguchi's Economical Loss Function

0,4

Frecuency

Nominal
Parameter
93%
200

0
96

Current Efficiency %

Figure 6 Comparison of Current Efficiency in the EW process using conventional technology and SELE
technology

Cathode weight
Cathode weight is another key variable in the EW process. This parameter must be analysed
differently than current efficiency, since it is important to hit the target value. As such, loss occurs
when operations are above or below this nominal value.

(8)

L*&+&

!",
!",

/ 0 !

= 5.83

(8)

/ 0 !

The Taguchi loss function calculated for the cathode weight distribution obtained through
conventional technology is almost six times greater than the loss function resulting from SELE
Technology, as shown in Figure 7.
According to a variability analysis conducted on these two performance parameters (current
efficiency and cathode weight), SELE Technology proved to be better than conventional

290

technology. By decreasing the variability on these two key parameters, SELE Technology adds
robustness to the EW process and reduces economic costs, making it more efficient and sustainable.
0,15

2000
Conventional Technology
SELE Technology

Frecuency

1500

Nominal
Parameter
75 [kg]

0,09

1000
0,06
500
0,03

0,00
30

40

50

60

70

80

90

100

Process Economical Loss

Taguchi's Economical Loss Function

0,12

0
110

Cathode Weight [kg]

Figure 7 Comparison of Cathode Weights in the EW process using conventional technology and SELE
Technology

SELE technology and its impact on the future of EW Plants

Both mining and engineering companies are considering working at a high current density in
future greenfield EW projects since SELE Technology makes it possible to operate under these
conditions without impairing current efficiency and cathode quality. This consideration is based on
the fact thatwith the same copper cathode production capacityplant size can be reduced when
working at a high current density, thus considerably minimizing project CAPEX. When compared
to a conventional project, CAPEX for a project using SELE Technology are 20% to 25% less as a
direct result of building a plant with a smaller footprint, fewer cells and electrodes, and less
peripheral equipment.
An EW plant with a copper cathode production capacity of 40,000 tonnes per year operating at 300
A/m2 will require 148 cells with an approximate investment of US$31,500,000, while an EW Plant
with an equivalent production capacity equipped with SELE Technology and operating at 450 A/m2
will only require 80 cells and an investment of US$24,514,000, as illustrated in Table 1.
Table 1 Process Parameters and CAPEX Main Indicators, EW Plant. Conventional Technology v/s
SELE Technology

Cells
Current (A/m2)
Equipment Cost (US$)
Building Cost (US$)
Total (US$)

Conventional Technology
148
300
22 636 000
8 864 000
31 500 000

SELE Technology
80
450
19 043 000
5 470 000
24 514 000

%
-68
150
16 %
38%
22 %

Finally, a greenfield project using SELE Technology has lower operating costs.

291

CONCLUSIONS
Results of the statistical analysis support and explain the benefits attained when using SELE
Technology as compared to the use of conventional technology. The SELE technological
innovation reduces the variability of two key parameters in the EW production process, namely,
current efficiency and cathode weight distribution. It is therefore a source of robustness for mining
processes and an effective means to reducing production-related costs.
Working at high current densities reconfigures the paradigm of having to expand the EW
tankhouse to increase cathode production. Implementing SELE technology and increasing current
density will increase tankhouse production capacity, thereby generating savings in costs related to
investments, project, engineering, construction, infrastructure and human capital, for a 25%
reduction in CAPEX, prompting some mining and engineering companies to consider this
alternative in future greenfield EW plant construction projects.
In summary, SELE technology makes it possible to operate an EW plant at high current densities,
with greater current efficiency, improved cathode quality and lower CAPEX.

NOMENCLATURE
F

Faraday constant, C mol-1

Current, A

Constant dependent on costs within specification limits

Economic loss associated to parameter

Mass of deposited copper, gr

Nominal value

Electrochemical equivalent, mol gr-1

Deposition time, s

Experimental value

Nominal deviation factor

Standard deviation

Mean

REFERENCES
Bard, A., Stratmann, M., McDonald, D., & Schmuki, P. (2007). Encyclopedia of Electrochemistry.
Electrochemical Engineering, 5, 161-223.
Beukes, N., & Badenhorst, J. (2009). Copper electrowinning: theoretical and practical design. The Journal of
The Southern African Institute of Mining and Metallurgy, 109, 343-356.
Fornari, P., & Abbruzzese, C. (1999). Copper and nickel selective recovery by electrowinning from electronic
and galvanic industrial solutions. Hydrometallurgy, 209-222.

292

Gopala, P., & Das, S. (1992). Enhancement of operating current density in a copper electrowinning cell.
Hydrometallurgy, 243-255.
Phadke, M. (1989). Quality engineering using robust design. New Jersey: Prentice Hall.
Robinson, T., Jenkins, J., Rasmussen, S., King, M., & Davenport, W. (2003). Copper electrowinning - 2003 world
tankhouse operating data. Proc. Copper - Cobre, 5, 421-472.
Sharma, N., Cudney, E., Ragsdell, K., & Paryami, K. (2007). Quality Loss Function - A Common Methodology
for Three Cases. Journal of Industrial and Systems Engineering 1, 218-234.
Wiechmann, E., Morales, A., & Aqueveque, P. (2010). Improving Productivity and Energy Efficiency in Copper
Electrowinning Plants. IEEE Transactions on Industry Applications, 1264-1270.

293

Tankhouse parameters for

transition from lead to
alternative anodes

Freeport-McMoRan Copper & Gold Inc. (FCX) is commercially operating alternative anodes at three of its electrowinning tankhouses, including Chino,
Bagdad, and El Abra. Alternative anodes (AA) replace lead anodes and decrease

Scot Sandoval, Casey Clayton,

Ephrem Gebrehiwot and Jason
Morgan. Freeport-McMoRan Copper
& Gold Inc., USA

energy consumption in copper electrowinning by lowering cell voltage. Because

A A are composed of titanium with a metal oxide coating, lead is removed from
the circuit, which improves cathode quality. Also cobalt additions to electrolyte
to help stabilize lead anodes are no longer required.
Under specific tankhouse conditions, compounds of Mn and Pb can deposit from electrolyte onto the surfaces of A A during operation, which can
hinder performance. Control of Pb deposition is particularly important
during the transition phase in which the tankhouse is converting from lead
to A A. Means to control Mn and Pb deposition onto A A surfaces have been
determined and are reported in this paper.
During transition from lead to A A the cobalt concentration must be maintained at a steady level. Results show that a significant decrease in cobalt
concentration releases lead into solution from lead anode surfaces which can
then deposit onto the AA surfaces. Increasing the cobalt concentration returns
the lead deposits into solution, but if the lead deposits are allowed to remain
and grow for a period of time the catalytic coating of A A can be damaged
which shortens lifetime.
Controlling the Fe2+/Mn ratio of electrolyte at 8 or above minimizes Mn
deposition. Polyacrylamide addition to electrolyte is a useful tool to prevent
Mn deposition onto alternative anode surfaces even at electrolyte concentrations of 300ppm Mn, but caution must be used with lead anodes still present
in the circuit.

295

INTRODUCTION
In 2006 Freeport-McMoRan Copper & Gold Inc. (FCX) Technology Center began operating an
anode development lab at its SXEW Test Facility located in Morenci, Arizona. The purpose of the
lab was to explore opportunities for replacing conventional Pb-Ca-Sn anodes in copper
electrowinning with an alternative anode (AA) that would decrease energy consumption and
remove lead contamination from the copper electrowinning circuit. Successful lab development led
to pilot testing of commercial scale anodes composed of a titanium structure with a catalytic coating
consisting of crystalline and amorphous Ir/Ta oxide layers. A 15% voltage reduction versus lead
anodes was achieved. Cobalt addition to electrolyte for stabilizing lead anodes was eliminated as
well as lead contamination of copper cathode and lead sludge in copper electrowinning cells.
Alternative anodes differ from lead anodes in several ways. Typical lead anodes are made using a
copper hanger bar and a Pb-Ca-Sn alloy blade that is not flow-through with respect to electrolyte.
Lead anodes typically weigh 80 kg and are robust towards rough handling and towards short
circuit damage. Operating costs are higher because of a higher operating voltage, but lead alloy is a
lower cost material of construction. Cobalt addition to electrolyte is used to decrease surface
corrosion rate of lead anodes during electrowinning. However lead contamination still makes its
way into copper cathode and electrowinning cells must be regularly cleaned of lead sludge.
Alternative anodes are constructed of a copper hanger bar with titanium-clad copper conductor
rods fitted into the hanger bar. Expanded titanium mesh substrate is attached to each side of the
anode structure. This makes the anode flow-through with respect to electrolyte. A catalytic coating
is applied to the expanded titanium mesh substrate. It is the catalytic coating that performs the
work of evolving oxygen in the electrowinning process. Alternative anodes weigh 20 kg and,
although the structure is rigid and straight, the structure is less robust towards physical damage
due to rough handling and towards short circuit damage. Personnel normally devoted to cleaning
lead sludge from electrowinning cells can be reassigned to preventing short circuits, which itself
has a payoff in terms of improved current efficiency for the tankhouse. Cobalt addition to
electrolyte is not required with AA.
Commercialation of AA began at FCXs Chino SXEW (Silver City, NM) in 2008 and followed at
FCXs Bagdad SXEW (Bagdad, AZ) in 2009. Conversion of FCXs El Abra SXEW (Calama, Chile)
began in 2010 and is nearly 100% complete at this time. During tankhouse conversion and operation
of AA at these sites hand-held XRF (x-ray fluoresence) units were used to monitor catalytic coating
thickness and the presence of surface deposits eminating from electrolyte during copper
production. Compounds of Pb and Mn were measured under varying conditions and
circumstances. To facilitate efficient operation of AA the means to control Pb and Mn deposits were
studied and developed and are reported here.

METHODOLOGY
Preparation for installation of AA at a particular tankhouse involved several steps. Training
sessions were held with operators to introduce them to the benefits of AA and to the inherent
differences from lead anodes. Simultaneously several demonstation cells were installed in the
tankhouse to allow the operators to begin hands-on training. Operation of demonstration cells in

296

the tankhouse typically lasted for up to a year before full conversion began. This allowed the
tankhouse to optimize the site-specific design of AA before commercial implementation.
Commercial implementation required less than one year at Chino and Bagdad, but has required
over two years at El Abra because of its size. After the demonstration phase an economic analysis
was conducted which evaluated the measured energy savings due to operation with alternative
anodes. To the energy savings was added the benefit of eliminating cost of cobalt addition to
electrolyte (for lead anode stabilization), the benefit of eliminating copper cathode downgrading
due to lead contamination, and the benefit of eliminating lead sludge cleaning requirements from
electrowinning cells. Positive economics led to full commercialization of AA at the plant.
Continuos electrowinning with AA may require several changes in operating practice. Short circuits
can produce holes in the expanded titanium screen that serves as the substrate for the catalytic
coating. It is therefore desirable to minimize short circuits, which provides an added benefit of
increased current efficiency to the plant. Holes of small to moderate size do not prevent AA from
continuing to plate copper. If the holes become too large so that copper plating is impacted the
holes can be patched on-site by tack-welding coated pieces of screen to cover the holes. Monitoring
of AA performance during operation is chiefly accomplished by regularly scanning the catalytic
coating of sample anodes using a hand-held XRF (x-ray fluoresence) unit (such as Oxford model XMET7000 eXpress). The XRF scan will show catalytic coating thickness that can then be monitored
over time in order to project the coating lifetime. In addition the XRF will reveal elements that may
deposit on the AA surfaces from electrolyte during operation. Two elements, Pb and Mn, have
periodically been detected on AA surfaces during operation at FCX tankhouses. Means to control
these deposits are discussed below.

RESULTS AND DISCUSSION

Control of Pb desposits
Introduction of AA to the copper electrowinning circuit to replace lead anodes will allow the
presence of Pb to dissipate from the circuit over time. Deposits of Pb on AA surfaces have occurred
only during a transition phase when both AA and lead anodes were operating together in the same
electrolyte circuit, although in separate cells. In all cases when lead anodes are removed from the
circuit Pb deposits on AA have not occurred. In the example noted here the Pb deposits were
initiated by a drop in electrolyte cobalt concentration accompanied by a simultaneous rise in cell
current density as shown in Figure 1. Cobalt concentration in electrolyte was decreased from 180 to
130 ppm while amperage increased from 35 to 45 Kamps (i.e. from 257 to 330 A/m2). XRF
measurements showed that Pb desposits appeared on AA surfaces after these events.
Cobalt concentration was then increased to 150 ppm but the Pb deposits remained. It was not until
cobalt concentration was returned to its original 180 ppm that the Pb deposits returned into
solution leaving the AA surfaces as shown in Figure 2.

297

Figure 1 Effect of amperage and cobalt concentration on Pb deposits on AA

Cobalt
returned to
180 ppm

Figure 2 Effect of returning cobalt concentration to 180 ppm on Pb deposits on AA

298

The demonstration phase at this plant had operated the AA demonstration cells at 180 ppm cobalt
and typical amperage of 277 A/m2 for about one year without any Pb deposits.
Figure 3 shows that Pb concentration in electrolyte increased from about 1 to 5 ppm as the cobalt
concentration decreased, returning to its original level of about 1 ppm when the cobalt was
returned to 180 ppm.

Figure 3 Effect of cobalt concentration on Pb concentration in electrolyte

From these results we see that during the transition phase from lead anodes to AA cobalt
concentration should be maintained at a steady level, preferably in the range evaluated during the
demonstration phase. If amperage is significantly increased from demonstration levels during the
transition phase then AA surfaces should be regularly scanned using XRF and cobalt concentration
increased if necessary to prevent Pb deposition. In this example five months passed before cobalt
concentration was returned to 180 ppm after which the Pb deposits returned into solution. Figure 4
shows the effect of extended exposure to Pb deposits on the AA catalytic coating. The surface of the
coating was disrupted by the Pb deposits which shortened the projected lifetime of the coating.

Figure 4 Effect of extended exposure of AA catalytic coating to Pb deposits

299

Control of Mn desposits
Accelerated coating failure testing was conducted at 5,400 A/m2 in a 180 g/L sulfuric acid solution at
ambient temperature. Comparative tests were conducted in which the effect of 250 ppm Mn in the
sulfuric acid solution was measured on coating lifetime. An 18% reduction in lifetime was
measured as a result of the Mn presence. This indicates that there is a lifetime benefit derived from
controlling Mn deposition on AA surfaces.
Figure 5 shows the effect of Fe2+/Mn ratio in electrolyte on Mn deposits on AA as measured by XRF.

Figure 5 Effect of Fe2+/Mn ratio in electrolyte on Mn deposits on AA

It can be seen in Figure 5 that a Fe2+/Mn ratio of 8 or above in electrolyte minimizes the presence of
Mn on AA surfaces. The presence of Fe2+ helps to keep manganese in the Mn2+ valence state in
solution, which does not deposit on anode surfaces.

Figure 6 Chino AA acts immune to Mn desposits even at 300 ppm Mn in electrolyte

300

It is interesting to note that FCXs Chino SXEW operates as low as 3 with its Fe2+/Mn ratio and yet
maintains near zero presence of Mn on AA, even at Mn concentrations of 300 ppm in electrolyte
(Figure 6). This unexpected result was due to the polyacrylamide cathode smoothing agent used at
Chino, Cyquest N-900. Polyacrylamide acts to prevent Mn deposits on AA surfaces. The
mechanism for this phenomenon is being investigated. A pilot plant trial was conducted at FCXs
Test Facility located in Morenci, AZ to show the polyacrylamide effect. In this trial Mn
concentration in electrolyte was increased steadily to 100 ppm. Polyacrylamide (Cyquest N-900)
was utilized as the cathode smoothing agent at an addition rate of 220 grams/ton of plated copper.
For a period of time addition of polyacrylamide was discontinued and the AA surfaces were
scanned using XRF. A significant increase in Mn deposits occurred during this period. When
addition of polyacrylamide was started again the Mn deposits on AA disappeared (Figure 7).
Caution must be used when applying polyacrylamide in the presence of lead anodes. Pilot testing
conducted at the Test Facility in 2001 showed that polyacrylamide when first introduced to
electrolyte caused a change in the lead anode surfaces that released both Pb and Mn as a slime into
the electrolyte. This resulted in a lead upset that contaminated the copper deposit with Pb. Only
several quality shutdowns of the pilot plant allowed the slimes to settle so that lead contamination
diminished in the plated copper. It was shown that the lead anode surfaces will eventually
equilibrate to the presence of polyacrylamide.
Polyacrylamide is also a flocculant and in the solvent extraction plant it will cause the crud layer to
become more dense and compacted. Adjustments to the plant crud management system may have
to be made.
It is interesting to note that in bench-scale electrowinning testing polyacrylamide was shown to
remove Pb deposits as well as Mn deposits from AA surfaces.

Figure 7 Effect of polyacrymide (Cyquest N-900) on Mn deposits on AA

301

CONCLUSIONS
Freeport-McMoRans alternative anode (AA) has been installed at FCXs Chino and Bagdad
tankhouses, and is nearly installed at the El Abra tankhouse. Preparation for conversion from lead
anodes to AA involves a demonstration phase and a transition phase. During the demonstation
phase the key objectives are to train operators on differences between lead anodes and AA and to
operate demonstration cells in the tankhouse to allow the operators a hands-on experience. Results
from the demonstration phase are used to conduct an economic analysis of the benefits of AA to the
tankhouse.
During transition from lead anodes to AA it is important to maintain tankhouse parameters as
much as possible within the range evaluated during the demonstration phase of the project until
lead anodes are completely removed. Once lead anodes are removed cobalt addition to electrolyte
can cease and Pb will diminish from the electrowinning circuit over time.
During transition from lead to AA cobalt concentration in electrolyte must be maintained at a
steady level, preferably in the range evaluated during the demonstration phase. If amperage must
be significantly increased during the transition phase then AA surfaces should be regularly scanned
using XRF and cobalt concentration increased if necessary to prevent Pb deposition. Control of
cobalt concentration in electrolyte as a function of amperage is the primary means to control Pb
deposits on AA until the lead anodes are removed from the circuit.
For control of Mn deposits the Fe2+/Mn ratio in electrolyte can be controlled at 8 or above to
minimize Mn deposition. Polyacrylamide addition to electrolyte acts to prevent Mn deposition, but
caution must be used with lead anodes still present in the circuit.

302

CoolBar: A new intercell bar for

electrolytic processes

When an electric current flow is present in an electrolysis cell, it follows that

the cell potential, Ecell, achieved is equivalent to the voltage difference of the
theoretical thermodynamic equilibrium resulting from the anode and cathode

Gerardo Cifuentes and Rodolfo

Mannheim. G2M Ingeniera Ltda.,
Chile

reactions, Eeq, plus the algebraic sum of the terms that reflects the inertia
of the reactions involved, normally called electrode overpotential, a+c, plus
a resistive term in solution, ReI, and finally plus the resistive term due to the
external electric circuit, Systemlosses.
Losses in the external electric circuit refer mainly to the Joule effect due
to the flow of current through the conductors. Indeed, intercell bars, also
called intercell busbars, work at an average temperature of 70 to 90C, and in
short circuit the temperature can go up to 200C and higher.
Our CoolBar (WIPO International Patent pending), which has a higher
current conductivity compared to presently used regular intercell bars, will
decrease the cell potential by 2 to 7%, decreasing the specific energy consumption (SEC) by an equivalent amount. In addition, the Coolbar optimizes the
use of thermal energy in the electrolytic plant by allowing better current flow
distribution in the cell, causing a significant drop in the carbon footprint of
the electrolytic process.
Finally, an existing intercell bar can be easily replaced in situ by a Coolbar
by just placing the latter over the cell capping board and making a few small
adjustments.

303

INTRODUCTION
Getting a high quality copper cathode implies, among other things, low energy consumption per
ton of copper cathode produced. In the copper Electrorefining process, which seeks to remove the
impurities from the copper anode (99.5% purity) through a process of dissolution in an electrolyte
by chemical and electrochemical reactions, as well as in the copper Electrowinning process to
recover the copper contained in the electrolyte solution, the aim is to get a cathode with 99-99%
purity as the product of these processes [W.G. Davenport, M. King, M. Schlesinder and A.K.
Biswas; Noguchi, Yano, Nakamura and Ueda, (1994); Noguchi, Iida, Nakamura and Ueda, (1992)].
When a direct electric current I flows in an electrolysis cell, the diagram of Figure 1 shows that Eth
is the resultant minimum theoretical thermodynamic voltage difference if the systems are
considered to be fast. The diagram does not consider the resistive terms [Cifuentes, Gerardo;
Simpson, Jaime; Lobos, Francisco; Briones, Leoncio and Morales, Alejandro; (2009 and 2010)]:

Figure 1 Voltage decomposition diagram of the electrode kinetics curves in an electrolysis cell

E cell = E th + a + c + IR e + Systemloss
where

Ecell

: Cell potential, (volts)

: Anodic overpotential, (volts)

| c |

: Cathodic overpotential, (volts)

I Re

: Electrolyte potential drop, (volts)

Re

: Electric resistance of the electrolyte, (ohms)

System losses

: Electric losses in connections, bars, etc., (volts)

304

(1)

Of all the terms in Equation (1), the Systemloss term reflects, among other things, the externalities
such as conductivity by electrons, which can be represented by Ohm's law for the case of copper
bars used as conductors.
Table 1 shows a typical distribution of the cell potential in a copper electrowinning cell. It also
shows the percentage of the voltage consumption at each component described in Equation (1).
The Ecell value corresponds to the measured value between the cell's anode-cathode contacts.
The contact and conduction losses are mainly due to the Joule effect due to the current flow through
the conductors. Indeed, intercell bars, also called busbars, work at an average temperature of 70 to
90 C, which is caused by the electric current going through the bar.
Table 1 Cell voltage distribution in copper electrowinning at the industrial scale.
Parameter

Volt

Eeq = Equilibrium voltage difference

0.9

40.9

|c|= Cathodic overpotential

0.05

2.2

a = Anodic overpotential

0.5

22.7
27.2

IRe Electrolyte potential drop at 200 A/m2

0.6

System losses due to contacts and electric conductors

0.15

6.8

Ecell = Total cell potential

2.2

100

Figures 2A and 2B show a classical arrangement of the cathodes in a cell, and a scheme of a copper
busbar in a copper electrowinning cell, respectively.

(A)

(B)

Figure 2 Cathode layout on the intercell copper busbar (A) and of a copper busbar in a copper electrolysis cell
(B)

Table 2 provides a comparison of some of the properties of copper compared to aluminum, thus
justifying the use of copper for busbars. It is seen that the mechanical strength and conductivity of
copper are better than those of aluminum. The only disadvantage of copper is its density, and for a
particular purpose an aluminum conductor would be lighter, despite its greater cross section. In
closed systems, however, space considerations are more important than weight [Mario Snchez M.,

305

(1987)]. Even in outdoor systems the weight of the bars, which are supported at intevals, is not
necessarily a deciding factor.
Table 2 Typical properties of copper and aluminum
Copper
(CW004A)

Aluminum
(1350)

Units

101
17.5
0.0039

61
28
0.004

% IACS
mmm2/m
1/C

397
1710-6
200 a 250
260 a 300
385
8.91
1.083

230
24 10-6
50 a 60
85 a 100
900
2.70
660

W / mK
1/C
N/mm
N/mm
J/kg K
g/cm
C

Electric conductivity (annealed)

Electric resistivity (annealed)
Coefficient of temperature
resistance (annealed)
Thermal conductivity at 20 C
Coefficient of thermal expansion
Tensile strength (annealed)
Tensile strength (medium hardness)
Specific heat
Density
Melting point

In particular, the high electric current used in the electrolytic processes requires the use of large
conductors, and space considerations can be important. It should be noted that the use of copper at
higher operating temperatures than would be acceptable for aluminum, allows smaller cross
sections and lighter weights to be used than those that would be necessary at low temperatures
[Copper Development Association (1984)].
As the temperature of metals increases, conductivity decreases, with a corresponding increase of
resistivity according to the expression

T = T (1 + T T (T2 T1 ))
2

1 2

(4)

where

= resistivity at temperature T1, (cm)

T T

1 2

= resistivity at temperature T2, (cm)

= resistivity coefficient in the T1 to T2 range, in C

Furthermore, resistance is related with resistivity by

R=

306

l
A

(6)

where R = resistance in ,

resistivity in cm2/cm, l = conductor length in cm, and A =

conductor's cross section in cm.

It follows that the electrical resistance of a metallic conductor also increases with temperature.

METHODOLOGY
Figure 3 shows a graphic summary of the results obtained in the measurement of the relative
conductivity (% IACS) of eight samples (M1 to M8) of commercial busbars at different
temperatures. The results show that increasing the temperature has greatly affected the
conductivity of the conductive material, in this case copper.
105
M1

% IACS

100

M2
M3

95

M4
M5

90

M6
M7

85

M8

80
10

20

30

40

50

60

70

80

90

Temperature, C

Figure 3 Experimental curves, relative conductivity vs. temperature

We see an average decrease of up to 20% in the relative conductivity of the conductive bars at 80 C
compared to ambient temperature (20 C). Therefore, if the electrically conductive bars have a
cavity through which a cooling fluid can flow, they would make it possible to control the
conductor's temperature at sufficiently low levels to allow energy savings of at least 2% because
high conductivity would be achieved at low temperatures.
As an illustration, since the existing geometries of industrial bars is quite varied, Figure 4 shows
two different examples of solid copper bars {1} which have one or more holes bored along them
depending on their geometry {2}, through which a cooling fluid circulates in order to keep a low
temperature along the bar. These bars are then placed as shown in Figure 2, but now with the
possibility of passing a fluid through them to keep them cool during the electrolysis.

307

{1}

{2}

Figure 4 Schematic drawings of two busbars, one rectangular and one triangular, either conductive bars or
busbars supports {1} conductive bars, which have either rectangular or cylindrical holes {2} through them

RESULTS AND DISCUSSION

Copper electrowinning experiments were performed at the laboratory scale to measure the voltage
drops in the system working at 200 A/m2, and comparing them with reported literature values [
W.G. Davenport, M. King, M. Schlesinder and A.K. Biswas; Noguchi, Yano, Nakamura and Ueda,
(1994); Noguchi, Iida, Nakamura and Ueda, (1992)].
Figure 5 shows the cells used in the laboratory tests. They consist of a rectangular cell of acrylic
material capable of withstanding the acidity and temperature conditions required for this type of
test. The conductive "bars" on both sides of the cell from which the electrodes (anode and cathode)
are suspended are also seen; because of the process conditions the cathodes are electrically
connected only to the negative pole and the anodes only to the positive pole. In this particular case,
the "bars" are actually hollow copper tubes through which it is possible to run a fluid (water in this
test) and in this way control and reduce the temperature of the electrical conductor. The points of
measurement of the cell potential (Ecell) by means of a tester are also shown.

Cooled busbars

Cathode

Anode
Cathode

Ecell

Electrolysis cell of
acrylic material

Figure 5 Electrolysis cell used in the laboratory tests

308

Table 3 presents the potential distribution obtained experimentally by testing the conductor bars at
60 and 20 C, showing that these percentage distributions are fairly close to those reported in the
literature.
Table 3 Experimental electric potential distribution with the conducting bars at 60 and 20 C.
Parameters

Volts
at 60 C
0.90

%
at 60 C
40.0

Volts
at 20 C
0.90

%
at 20 C
40.7

0.05

2.2

0.05

2.3

0.50

22.2

0.50

22.6

0.60

26.7

0.60

27.1

0.20

8.9

0.16

7.2

2.40

100

2.21

100

Eeq = Voltage difference at equilibrium

c|=

Cathodic overpotential

a = Anodic overpotential
IRe Electrolyte potential drop at 200 A/m2
System losses at contacts and in electric conductors
Ecell = Cell potential

CONCLUSIONS

The CoolBar system allows direct energy savings in the process of electrolysis. For example, for
a plant that produces 100,000 tons/year and consumes 2,000 kWh/ton, and considering direct
savings of 2% by using the CoolBar system, this would be reflected in approximately US$
600,000 per year.

If we consider the above electrolytic plant with a copper price of 400 US/lb, and now the 2%
energy savings is reflected in increased production, which would amount to 2,000 extra
tons/year of copper produced, equivalent to approximately US$ 17 million per year.

Related improvements not yet considered in the costs:

- Using the heat extracted from the bars.
- Increased power efficiency by better distribution of the current in the bar.
- Decrease by better distribution of the current in the bar.
- Increased durability of the busbars.

Currently tests are being scheduled in Chilean industrial electrolytic plants with the CoolBar
system.

REFERENCES
W.G. Davenport, M. King, M. Schlesinder and A.K. Biswas, Elsevier Science Ltd, ISBN 0-08-044029-0.
Noguchi, Yano, Nakamura and Ueda, (1994) Metallurgical Review of MMIJ, 11 (2): 39-41.
Noguchi, Iida, Nakamura and Ueda, (1992) Metallurgical Review of MMIJ, 8 (2): 83-97.

309

Cifuentes, Gerardo; Simpson, Jaime; Lobos, Francisco; Briones, Leoncio and Morales, Alejandro; (2010) An
alternative copper electrowinning process based on reactive electrodialysis (RED), CISAP4
International Conference, 14-17 March, Florence, Italy.
Cifuentes, Gerardo; Simpson, Jaime; Lobos, Francisco; Briones, Leoncio and Morales, Alejandro; (2009)
Copper electrowinning based on reactive electrodialysis, J. Chil. Chem. Soc., 54 (4).
Mario Snchez M., (1987) Efecto de Impurezas en el Comportamiento de Cobres Comerciales,IIT, Facultad
de Ingeniera, Universidad de Concepcin.
Copper Development Association (1984), Copper for Busbars, publication N22.

310

Practical determination of
galvanic corrosion of steel
induced by mineral particles

The objective of this work was to study the galvanic corrosion of AISI 1020
carbon steel induced by the presence of copper sulfide mineral by electrochemical techniques using a home-made electrode with copper concentrate
and a carbon steel working electrode. The concentrate electrode as a film
attached on a supportive media was prepared by rubbing the surface tip of a

Genny Leinenweber and Luis

Cceres. CSIRO Chile International
Centre of Excellence in Mining and
Mineral Processing, Department of
Chemical Engineering, Universidad
de Antofagasta, Chile

graphite/wax electrode against dried powdered mineral. The graphite/wax

electrode was a plastic cylinder with a cavity filled with a supportive media
and metal fittings for electrical connections to the potentiostat (bare electrode).
The supportive media was a 1.3:1 mixture of candle wax and pencil graphite
respectively, which is a good electrical conductor. The carbon steel working
electrode was prepared from a carbon steel disk embedded in a Teflon cylinder.
A polarization curve from -1100 a 1600mVshe using a graphite/wax electrode
without film mineral indicated the occurrence of the main electrochemical
reactions of hydrogen and oxygen evolution. In between these two reactions
no significant redox activity was observed. This feature is very important to
restrict the effect of interfering possible redox reactions of the particle support.
The galvanic corrosion current density of steel induced by concentrate was
achieved by superimposing individual concentrate and carbon steel polarization curves, both in contact with the same working aqueous solution.
The main findings of this work are: (i) This technique allows the determination of electrochemical features of ground mineral. This is important where
massive samples are not available. (ii) The support for ground mineral exhibit
ideal electrochemical features, and (iii) concentrate induced galvanic corrosion on carbon steel in chloride solutions at ambient conditions was
corroborated.

311

INTRODUCTION
Galvanic corrosion takes place when two different metals or minerals are electrically connected
through an electrolyte generating an electric potential. This potential difference provides a driving
force for the dissolution of the more electrically negative material (Schweitzer, 2007). This is the
case of iron and copper; corrosion of the less resistant metal which is iron increases, and its surface
becomes anodic, while corrosion of the more resistant metal which is copper decreases, and its
surface becomes cathodic.
A particular situation of copper particles suspended in conductive electrolytes such as chloride
solution, takes place at copper mining processes in Northern Chile. Such slurries in contact with
steel surfaces create a condition for galvanic corrosion of steel which is triggered by solid particle
impingement or gravity deposition, combined with attachment mechanism to metal surfaces. For
such situation the direct measurement of corrosion currents is impractical, because a corroding
metal surface consists of many short-circuited corrosion cells with corresponding currents.
Galvanic corrosion rates can be determined from mixed-potential theory by which the sum of all
oxidation currents must equal the sum of all reduction currents. Thus, the mixed potential (or the
galvanic potential, Ecouple) of the galvanic couple and the resulting galvanic current can be uniquely
determined as:
iaAa = icAc at

Ecouple

(1)

where ia and ic are the anodic and cathodic current densities (A/cm2), respectively, and Aa and Ac
are the anodic and cathodic areas (cm2). The ia and ic values may be experimentally determined
from individual polarization curves for copper concentrates and carbon steel respectively using a
common electrolyte.
When massive sample is not available for electrochemical studies, as is the case of copper
concentrate, the preparation of an electrode using an adequate support for solid particles is an
essential preliminary step (Bard, Faulkner & Hu, 2001, Sun & Wang, 2009). This material support
must be both a good electrical conductor and an effective adhesive for particles.
In this work the suitability of a home-made graphite/wax electrode to determine the occurrence of
galvanic corrosion of AISI 1020 carbon steel induced by copper sulphide particles was examined.
Such determination is based on the superimposition of the experimental polarization curves for
concentrate and metal electrode both using the same electrolyte solution.

MATERIAL AND METHODS

A Voltalab electrochemical interface with a rotating electrode system was employed for
voltammetry and impedance measurements. A conventional three electrode cell was used with a
platinum wire as a counter electrode, a Ag/AgCl 1 M KCl cell as a reference electrode and two
rotating working electrodes made of graphite/wax and carbon steel, respectively.

312

Graphite-wax working electrode (GWE)

The GWE material is a mixture of 1.0 g graphite and 1.3 g of candle wax. The graphite is
commercial pencil graphite which is a mix of finely ground graphite, clay powders with a small
amount of wax (Cain et al. 1978). A graphite 8B type was chosen because of its very low clay
content which is 4% w/w. The graphite was hand ground with an agate mortar and pestle.
For the construction of the GWE, a teflon rod with dimensions of 8 mm diameter, 30 mm length
was drilled along its axis to insert a 4 mm diameter brass cylinder in one end, and the shaft of a
rotating accessory of the potentiostat in the other (Fig. 1). The cavity left by the inserted brass was
filled with molten graphite/wax mix at 80 C, which was solidified at room temperature.
GWE-concentrate electrode
The concentrate powder was hand ground with an agate mortar and pestle. This material was
sieved to isolate a homogeneous particle size fraction. The concentrate electrode as a film attached
on the supportive media of the GWE which is the graphite/wax phase. This was prepared rubbing
the surface tip of a GWE against dried powdered mineral.
Carbon steel working electrode (CSE)
The CSE was a 4-mm diameter X 5-mm length cylinder made of carbon steel AISI 1010 with
chemical content (wt%) 98.5 Fe, 0.2 C, 0.6 Mn and traces of P, S, Si, Sn, Cu, Ni, Cr and Mo. An 8 mm
diameter 20 mm length teflon rod was drilled along its axis to insert the specimen in one end and
the rotating shaft in the other.

Figure 1 Graphite/wax working electrode

Electrochemical tests
Electrochemical tests carried out in this work were linear sweep voltammetry and impedance
spectroscopy. Polarization curves were obtained from linear potential sweep at a rate of 3 mV s-1,
(Cceres, Vargas & Herrera, 2007). Potential sweeps were conducted in the anodic direction from
-1000 to 1400 mVSHE with the working electrode rotating at 1200 rpm under air bubbling rate of 6

313

cc/min in 0.5 M NaCl at ambient conditions. Electrochemical corrosion parameters were determined
using mathematical software described in former investigations (Cceres, Vargas & Herrera, 2007).
Impedance measurements for GWE were made in the frequency range 105 to 10-2 Hz at potentials of
-900, -600, -300, 300, 600, 900 and 1200 mVSHE under similar electrolyte conditions.

RESULTS AND DISCUSSION

GWE characterization
A polarization curve obtained at a rotation rate of 1200 rpm for bare GWE is shown in Fig. 2. Its
relevant features are: a) A cathodic region between -1200 to 0 mVSHE corresponding to a process of
hydrogen evolution, b) Intermediate zone between 0 and 900 mVSHE where the redox activity is
negligible, and c) The anodic region between 900 to 1700 mVSHE corresponding to oxygen evolution.
The average Tafel slope value of 130 Vdec-1 for hydrogen evolution in GWE is in agreement with
bibliographical data. In contrast the Tafel slope for oxygen evolution with an average value of 350
Vdec-1 is larger than that reported for most metals with the exception of diamond electrodes for
which there is a very good agreement (kapalka, Foty & Comninellis, 2008). The very low
electrochemical activity in the intermediate zone which is corroborated by very low current density
values indicates a wide electrochemical window. This is a necessary condition for a supportive
media to ensure no interference with the electrochemical interaction between particles and solution.
More information concerning the redox activity originated from the EIS study.
10

I (A/m2)

0
-1200

A1

A2

-600

A3

A4

A5

600
E (mVSHE)

A6

A7

1200

1800

-5

-10
Figure 2 Polarization curve of GWE in seawater. Black dots denote potentials where impedance spectra were
obtained

Impedance spectra for GWE recorded at different potentials in the frequency range of 104 - 10-1 Hz
are shown in Fig. 3. Nysquit and Bode plots at -600 mVSHE are shown in Fig. 4. Each curve was
modeled using an equivalent circuit depicted in Fig. 5. This circuit denotes a single redox reaction
without product precipitation/deposition on the surface. The meaning of the model parameters are
(Bard & Faulkner, 2001): Rs correspond to the electrolyte resistance, R1 is associated to the charge

314

transfer resistance, C to the double layer capacitance; Zw is the impedance associated to a semiinfinite linear diffusion of reactants within the solution. The parameter account for the Cole-Cole
dispersion time constant (Bard, Stratmann & Unwin, 2002). For this model the overall impedance
as a function of frequency w and circuit elements of Fig. 5 is given in the following expression:

=R +

R1 Cw

R1
2
+ 1 1 + R1Zw

where Zw, the Warburg impedance, is defined as:

Zw = Zr 1

and Zr is the Warburg coefficient and j is the imaginary unit in complex number representation.

60000

Zim (ohms cm2)

50000
A5

40000
30000
A3, A6, A7

20000

A1
A4

10000
A2
0
0

10000

20000

30000

40000

50000

60000

Zreal (ohms cm2)

Figure 3 Electrochemical Impedance Spectroscopy of GWE

315

-200
-300
0

Z (ohms cm2)

Zim (ohms cm2)

0
-100

50

100

150

200

250

300

350

400

Zreal (ohms cm2)

600
400
200
0
-1

log(w)
Angle ( )

-20

-40
-1

log(w)

Figure 4 Nysquit (upper) and Bode (Middle and bottom) plots at -600 mVSHE. Continuous lines are
experimental data. Dot marks are fitted data

Table 1 Best fitting parameters of the impedance data of GWE at different potentials

316

Exp.

E
mV SHE

Rs
2
cm

R1
2
cm

A1
A2
A3
A4
A5
A6
A7

-900
-600
-300
300
600
900
1200

5.3
4.6
5.5
4.0
8.2
6.2
4.9

23650
33400
1490
2
361
94700
5006

C
F
4.890
0.005600
0.111
0.00002
0.00003
0.004
0.004

Zw
2
cm

0
0
0
28140
9152
0
0

0.43
0.45
0.44
0.47
0.66
0.63
0.63

C
Rs

R1

Zw

Figure 5 Equivalent circuit used to fit the experimental EIS data for GWE/sea water environment interface

The impedance parameter values shown in Table 1 should be compatible with redox reactions that
occur at given potentials. At A1, A2 and A3 potential values the impedance parameters are
attributed for hydrogen evolution, A4 and A5 are in the intermediate zone (Fig. 2) with very high
impedance values (Fig. 3) which are attributed to very low redox activity (low current density);
parameter values for A6 can be attributed to a situation where oxygen evolution interferes with low
redox activity of the intermediate zone and finally parameters for A7 can be attributed to oxygen
evolution. As expected the electrolyte resistance values Rs are of the same order of magnitude at
different potentials and its values are significantly lower than the interface impedance which is
represented by the combined elements R1, C and Zw of Fig. 5. This means that the electrolyte
resistance do not play any significant role in the electrochemical behaviour of the GWE surface. The
lowest R1 values that are coinciding with the emergence of the Zw at intermediate potentials
suggest that a resistance associated to mass transport of reactive species toward the interface could
takes place from the wax/graphite phase.
Galvanic corrosion
Polarization curves for GWE-concentrate and CSE are shown in Fig. 6. Both clearly intersect such
that the carbon steel is in the anodic branch while the concentrate curve is in the cathodic branch.
According to previous investigations a minimum difference of 100 to 130 mm between the
corrosion potential of the cathode and the anode of the pair (

) are necessary to consider

the galvanic effect significant (Sanchez et al., 2013). According to Fig. 6 the corresponding
!! and
!! values are -269 and 168 mV for concentrate and steel respectively which gives a
difference of 437 mV.
Another evidence to confirm the occurrence of a galvanic couple between carbon steel and copper
concentrate comes from the mixed potential theory. Assuming balanced values of cathodic and
anodic areas the value of the corrosion potential is equal to -188 mVSHE with a galvanic corrosion
current density equals to 2.8 A/m2 (127.68 milli-inch per year) In considering the corrosion current
density of the carbon steel in chloride solution which is 1.2 A/m2 (54.72 milli-inch per year) the
galvanic corrosion is relevant and deserves further attention in order to devise alternatives for
corrosion prevention.

317

Figure 6 Polarization curves for GCE-concentrate and CSE in seawater. Galvanic corrosion potential is
indicated at the point marked as black dot (absolute values of cathodic and anodic current are equivalent)

CONCLUSIONS
In this work the galvanic corrosion of AISI 1020 carbon steel induced by the presence of copper
sulfide mineral using a home-made electrode were studied.
The main findings of this work were:
This technique allows the determination of electrochemical features of ground mineral. This is
important when massive samples are not available.
The support for ground mineral exhibit ideal electrochemical features. This is corroborated from
the fact that no significant redox reactions on the bare wax graphite/wax electrode surface other
than hydrogen and oxygen evolution were observed. The analysis of the impedance spectra
gives a good account for the nature of the redox reactions that occur on the GWE. This evidence
indicates that no product deposition and/or adsorption take place on the GWE surface.
Concentrate induced galvanic corrosion on carbon steel in chloride solutions at ambient
conditions was corroborated.

ACKNOWLEDGMENTS
The financial support from CONYCIT project number 21120803 developed in the Chemical
Engineering Department at the Universidad de Antofagasta, Antofagasta, Chile, is greatly
appreciated.

318

REFERENCES
Cceres, L., Vargas, T. & Herrera, L. (2007) Determination of electrochemical parameters and corrosion rate for
carbon steel in un-buffered sodium chloride solutions using a superposition model, Corrosion Science
49 31683184.
Cain, S.,Cantu, A. A., Brunnelle R & Lyter, A. (1978) Scientific study of lead pencil components, Journal of
Forensic Science 23643-661.
Bard, A. J, & Faulkner, L. R (2001) Thin-Layer Electrochemistry in: Electrochemical methods fundamentals and
applications, John Wiley & Sons, Inc., pp. 459.
Bard, A. J. &. Faulkner L. R. (2001) Techniques based on concepts of Impedance in: Electrochemical methods
fundamentals and applications, John Wiley & Sons, Inc., pp. 368.
Bard, A. J., Stratmann, M. & Unwin P. R. (2002) Impedance methods in: Encyclopedia of Electrochemistry
Vol 3, Instrumentation and Electroanalytical Chemistry, University of Warwick, Coventry, United
Kingdom, Wiley, pp. 196-229.
Hu, Y., Sun W. & Wang D. (2009) Electrochemistry of Flotation of Sulphide Mineral, Springer Ed., pp.197.
kapalka, A., Foty, G. & Comninellis, Ch. (2008) Determination of the Tafel slope for oxygen evolution on
boron-doped diamond electrodes, Electrochemistry Communications, 10607610.
Sanchez-Tobar, R., Montaes, M., Garca-Antn, J., (2013) Contribution of the flowing conditions to the
galvanic corrosion of the copper/AISI 316L coupling in highly concentrated LiBr solutions , Corrosion
Science 68, 91-100.
Schweitzer, P.A. (2007) Fundamentals of metallic corrosion: atmospheric and media corrosion of metals, CRC
Press Taylor & Francis group Eds., N.Y., pp 8-10.

319

MIRS Starter Sheet Robotic

Stripping Machine (SSRSM)

Most SX-EW and Refineries currently use permanent stainless steel cathodes.
However, there are refineries and SX-EW operations that use starter sheets/
traditional cathodes technology: In Chile, Ventanas and Salvador Refineries;

Luis Felipe Ramrez. Mining

Industry Robotic Solutions (MIRS),
Chile

in Peru, Ilo Refinery and Cerro Verdes SX-EW plant; several in the USA and
Mexico, ASARCO s Amarillo Refinery among them. They require copper starter
sheets for their final cathode harvesting.
Stripping starter sheets manually is boring, tiring, unsafe and may produce
lower quality sheets which directly impacts the quality of commercial cathodes
and therefore operator revenue.
MIRS designed and developed a Starter Sheet Robotic Stripping Machine
(SSRSM) focusing on improving safety and occupational health and reducing
hazards in the process. The patented method and separation device also
improves the quality of starter sheets.
In a standard layout, four robotic arms interact in the system: The input
robot picks each cathode from a conveyor and places it in the stripping station;
two stripping robots strip each sheet and the output robot picks the empty
plate from the stripping station and places it in the output conveyor. This robot
also handles rejects and blank reposition.
One MIRS SSRSM was commissioned at Ventanas during 2011. An industrial test was carried out whose objective was to demonstrate functionality to
strip in typical operating conditions. Test results revealed a capacity of 160
plates per hour and stripping performance higher than 98%.
Now we are incorporating a washing station and two laser scanners; one
for sheets and another for base plate quality and a starting sheet weighing
station to sort them into different qualities.
Because the SSRSM is robotic, it is more compact, flexible and configurable
so it adapts to existing layout and equipment. It also is capable of using different stripping strategies according to cathode quality.

321

INTRODUCTION

Figure 1 Starter Sheets in Electrolytic cell

Electrorefining is the last step in copper production, be it in a refinery after smelting concentrates
from copper sulphides or in an SX-EW tankhouse after leaching and solvent extraction from oxides
or lixiviable sulphides.
Electrorefining may be performed in one cycle using stainless steel permanent cathodes or in two
cycles using starter sheets.
Operating and Cost advantages of each method are beyond the scope of this document, but it is
worth mentioning how efficient a refinery must be to transform at a profit- 98.5%-99% copper
anodes to to 99.99% copper cathodes spending a huge Capex and Opex for that small increase in
product grade, which is sine qua non for copper to conduct electricity.
For a Lix/SX-EW operation processing oxides or lixiviable sulphides from the mine, the technology
selection drivers are just the same. These operations face very demanding operating conditions and
thus it never proves feasible to switch fron one technology o the other.

A market niche
Be it as it may, most refineries and SX-EW plants currently use permanent cathodes and need
Cathode Stripping Machines. The balance of them produce traditional cathodes from starter sheets
and need Starter Sheet Stripping Machines.
Traditional cathode producers are:
Operation
Ventanas
Salvador
Ilo
Toquepala
Amarillo
Morenci
Bagdad
Twin Buttes
Miami
Safford

322

Owner
Codelco
Codelco
SPCC
SPCC
Asarco
Freeport McMoran
Freeport McMoran
Freeport McMoran
Freeport McMoran
Freeport McMoran

Location
Chile
Chile
Peru
Peru
USA
USA
USA
USA
USA
USA

Tyrone
Bingham Canyon

Freeport McMoran
Rio Tinto

USA
USA

The last attempt to supply a starter sheet stripping machine was around 1990 by Mitsubishi. MIRS
identified this market niche and developed its Starter Sheet Robotic Stripping Machine (SSRSM)
capable of stripping a cathode in 13.5 seconds.

SSRSM main components

The robotic machine is composed of

1 input robot that feeds incoming cathodes in the stripping station

2 stripping robots which strip the starting sheets

1 output robot that removes empty plates from the stripping station and removes rejects
replacing blanks

1 stripping station

An array of sensors in charge of safety and component statuses

Racks, conveyors

Perimeter Fencing

Power and Control systems

An operation and control station including image cameras which continuously monitor
system variables and allow operator control.

The MIRS SSRSM Core

MIRS SSRSM core is the stripping tool, developed and patented by MIRS.

Figure 2 SSRSMs Stripping Tool

323

The stripping tool performs 3 duties:

Prestrips

Strips

Picks each stripped sheet and leaves it on the rolling conveyor underneath

The stripping method improves sheet quality

History
Long before the first piece of steel of the SSRSM was cut, bent and polished, it went through an
extensive set of virtual design alternatives and tests by MIRS product development team. Safety,
performance, efficiency, process continuity and tolerance to external factors were all scrutinized.
Then the SSRSM was built and pilot tested at MIRS laboratory, and in 2nd half 2011 it was tested at
Ventanas Refinery. Cycle time was 16 seconds/plate with 98% stripping efficiency. The test also
showed an improvement in quality of stripped copper sheets compared to manual stripping.
After successfully completing the industrial test at Ventanas we submitted them a proposal to
modify the layout to improve SSRSMs performance to 13.5 seconds/plate, together with a second
SSRSM unit which would fulfill 100% of Ventanas starter sheet stripping needs.

Figure 3 SSRSM layout at Ventanas, as tested

Current Status
SSRSM now includes an optional washing station/dryer; two laser scanners: one for sheets and
another for base plate quality and a starting sheet weighing station to sort them into different
qualities.
The standard input conveyor now incorporates a transfer station so system operation is not
interrupted when a new batch is being fed by the overhead crane; plus a weighing station to
preclassify cathodes by weight and reject early or adjust stripping method accordingly.

324

Flexibility
Being robotic -therefore not being restricted by standard electromechanical design constraints- our SSRSM
may be setup in different layouts adapting itself to available space, cathode feeding method,
conveyor location, position and length, etc.

Figure 4 SSRSM layout improvements, as proposed to Ventanas after Industrial Test

Figure 5 Standard SSRSM General Arrangement Plan showing optional washing station

325

CONCLUSION
We strongly believe MIRS SSRSM proves to be an affordable, feasible, efficient, safe and overall
better alternative to manual stripping for starter sheets. Layout and component flexibility as well as
options add up to its advantages.

ACKNOWLEDGEMENTS
Rodrigo Abel, Senior Engineer at Ventanas Refinery
Carlos Moya, MIRS development team, who designed the mechanical aspects of the stripping tool
Leonardo Basaul, MIRS development team, who implemented all SSRSM control system
MIRS development team responsible of initial design and the industrial test at Ventanas
Luis Baeza, leader of SSRSM development

NOMENCLATURE
SSSM
CSM
SSRSM

Starter Sheet Stripping Machine

Cathode Stripping Machine
Starter Sheet Robotic Stripping Machine

SxEw

Solvent Extraction and Electrowinning

326

chap. 6
Bioleaching processes

Bioaccelerant: A new
biotechnological product for
bioleaching process optimization

BioAccelerant is a new bioproduct intended for copper sulfide bioleaching

process optimization, which allows cycle time minimization and ore recovery
maximization through the addition of native, metabolically active, iron- and

Mara de la Luz Osses, Matas

Saavedra and Simn Beard.
Biotecnologas Antofagasta S.A .,
Chile

sulphur-oxidizing microorganisms at an industrial scale, without the need

for complex in situ biomass plants construction and operation.
BioAccelerant contains highly concentrated freeze-dried bioleaching
bacteria in powder format. Microorganisms in the product are native, that is,
isolated directly from each client environmental samples, a condition that
ensures a higher degree of adaptation and effectiveness. BioAccelerant can
be applied easily as a biological curing (homologous to acid curing) during
heaps mineral loading or directly through the irrigation system.
Product formulation is defined on a case by case basis, looking for the optimum microorganism combination able to maximize results for each mine site.
BioAccelerant was subjected to laboratory scale validation through unitary column tests. Application of the product on a refractory mineral with
high chalcopyrite content generated a 25% decrease in cycle time and a 20%
increase in copper recovery.

329

INTRODUCTION
Bioleaching of sulphide minerals is a biochemical process involving native microorganisms
naturally present in each mine ecosystem, with the capability to dissolve copper contained in ores
by means of it metabolic activity in acidic environment.
Mining companies are aware of the impact that signifies the presence or absence of active leaching
microorganisms in their extraction processes. Several enterprises have invested in the construction
and operation of complex biomass production and continuous supply systems in their mine fields,
with high associated costs in materials and human resources. However, difficulties associated to the
operation of biotechnological plants in mine fields and the lack of experience in this matter have
prevented a massive adoption of biologic control processes through exogenous inoculation or
bioaugmentation, despite its widely recognized value among the copper mining industry.
Cycle times of bioleaching processes at industrial scale in heaps or dumps are particularly
extensive. The time needed to extract expected amounts of copper from ores normally exceeds 300
days. A reduction, even of a discreet extent, in process cycle time has the potential to produce large
economical profits. A mine can process around 150.000 tons of mineral daily. Based on the latter, a
30 days reduction in processing time (10% considering a cycle time of 300 days) would allow the
additional incorporation of 3.5 million tons of mineral into the productive chain.
In this context, BTA S.A. designed and developed BioAccelerant, a biotechnological product for the
bioleaching industry that allows shortening conventional cycle times in copper mining
biohydrometallurgical processes through exogenous inoculation of freeze dried, metabolically
active, native bioleaching microorganisms at production scale.
In this report, we make a detailed description of product BioAccelerant and present the results of
validation essays at laboratory pilot scale, performed in collaboration with important mining
companies based in Chile using bioleaching unitary columns.

METHODOLOGY
Analytical, microbiological and instrumentation procedures together with unitary column tests
implementation and monitoring were performed according to BTA S.A. internal protocols, based
on standard methods. As these experimental manipulations belong to BTA S.A. specific know-how,
they cannot be fully described in the context of this article.

Cell viability
The proportion of viable cells in liquid samples was determined through epifluorescence
microscope counts using commercial kit BacLigth Live/Dead (Life Technologies). This kit is based
on the use of two fluorescent chemical agents capable of interacting with DNA. One agent,
SYTO9, is able to penetrate living and dead cells, whereas the other agent, propidium iodide, can
only enter dead cells with a damaged membrane. As the two dyes coexist within dead cells, and
living cells are stained only with SYTO9, it is possible to distinguish live (green stained) and dead
(red stained) bacteria using appropriate filters in the epifluorescence microscope.

330

Unitary columns tests

Unitary column tests were realized according to Table 1 general conditions.

RESULTS AND DISCUSSION

BioAccelerant description
BioAccelerant is a bioproduct composed of native leaching microorganisms extracted directly from
the mining field or hydrometallurgical process of each client, produced at industrial level,
concentrated and dehidrated to a powder, what allows industrial scale heap inoculation without
the need for significant investments in infrastructure and operational costs related to the
construction and operation of complex biomass plants. Bioaugmentation by heap inoculation in
turn offers the possibility to optimize metallurgical results, specifically lowering process cycle time
and/or increasing copper recovery yields.
Table 1 Unitary columns operational parameters
Parameter

Value

Ope rational conditions:

Ae ration

No

Te mperature control

No

Duplicate s

Ye s

Tipo de column:
Diame te r (m)

0,162

He ight (m)

Ore mass (Kg)

50 Kg pe r column aprox.

Particle size

P80-18 mm

Le aching solution

Synthetic raffinate pH 1,8

leaching ratio

2,5 m /Ton

Acid curing

No

Cycle time

120 days

BioAcce le rant dose

1,0x10 ce lls/g of ore

Control

uninoculate d columns

PLS control fre que ncy

Daily until day 15, e ach 4 days until day 30 and e ach 7 days until the e nd

Raffinate control freque ncy

Each 15 days

Analitical calculations

Base d on solution and ore determinations

Product formulation
At given environmental conditions, ores can be more or less susceptible to bioleaching depending
on their mineralogy (including gangue) and the type of microorganisms at which they are exposed.
An ore with the potential to liberate toxic chemical species into the leaching medium will probably
be highly detrimental for biological activity, inducing selective inhibition or destruction of

331

beneficial microbial cells. Therefore, native biomass tolerance to potential inhibitors dissolved from
the ores is a determinant aspect for a successful bioleaching process.
On the other hand, the impact of diverse specific metabolic capabilities among the microbial
communities involved in bioleaching processes over metallurgical results can be highly unequal.
Indeed, from a metallurgical point of view, influences of sulfur-oxidizing and iron-oxidizing
bacteria on a given piece of mineral can be very different. As sulfur-oxidizing bacteria could be very
effective in promoting leaching, iron-oxidizing might even have the opposite effect, and vice-versa.
As a consequence, BioAccelerant is tailored to customer and its formulation is designed on a case
by case basis.
Several different proportions of sulfur-oxidizing and iron-oxidizing consortia are included in
various BioAccelerant formulations and tested at laboratory scale for bioleaching effectiveness.
Only the specific formulation that shows the best metallurgical results is considered for production
and supply to a given mining Company.
Parallely, if environmental conditions in the objective industrial bioleaching process are or could be
inhibitory for the native biomass, isolated microbial consortia intended for BioAccelerant
production are submitted to an adaptation procedure in order to increase their tolerance threshold.

Industrial application and management

BioAccelerant is distributed in vaccum 1 m3 maxibags (approximately 100 kg of product) that can
be stored at room temperature for 2 months or at a constant temperature of -20C for up to six
months. Before application, the product has to be resuspended in acid water and incubated at
ambient temperature for at least 5 days, preferably with aeration. The volume of water needed for
this purpose will depend on each mining Company process design and the considered amount of
product. Typically, to treat ten thousand Tons of ore, four kilograms of BioAccelerant will have to
be diluted in one thousand to ten thousand liters of acid water giving rise to solutions with 1,0 x 109
to 1,24 x 1010 cells/mL (see Table 2 for calculation parameters). The goal of resuspension is to
rehydrate and reactivate the microorganisms from the dry product. Figure 1 shows a standard
reactivation curve for BioAccelerant: from powder to metabolically active microorganisms.

100

Live cells (%)

80
60
40
20
0
0

6
Time (days)

332

10

Figure 1 BioAccelerant reactivation curve. The product was resuspended in acid water pH 1,8 and incubated
at room temperarure with aeration.

After five days of incubation, cells viability begins to increase significantly to a high value of 70% at
day eight. This viability score is comparable to a standard laboratory culture at exponential fase.
Figure 2 shows Epifluorescence microscope pictures with live/dead kit results corresponding to
reactivation days cero and seven.
Once reactivated, the aqueous cell suspension can be applied to ROM and heap operations in
several ways (Figure 3). In a ROM dump, cell suspension should be delivered through the raffinate
irrigation system, as a side input line. In the case of a crushed bioleaching heap, the cell suspension
can be administrated either in the ore feeding conveyor belt as a biological curing (homologous to
acid curing), or in the raffinate irrigation system as with ROM dumps.
As a supply, BioAccelerant permits a flexible, low cost and straight forward bioleaching process
inoculation management.
A

Figure 2 BioAccelerant reactivation. Epifluorescence microscope picture with resuspended BioAccelerant in

acid water pH 1,8 at day cero (A) and day 7 (B). Live cells appear in green and dead cells show red
fluorescence

Refine +
BioAccelerant

Product
management

Dump ROM

BioAccelerant
storage (managed as a
supply)

Crushing

BioAccelerant
product reactivation
in water

Mine

Acid curing
Water + BioAccelerant
(Biological curing of ore)

Refine +
BioAccelerant

Heaps

333

Figure 3 BioAccelerant management and application diagram

Experimental validation
Unitary columns bioleaching tests
Figure 4 shows the effect of BioAccelerant application at a 1,0 x 106 cells/g of ore dose in unitary
columns bioleaching tests with a highly chalcopyritic ore (70% chalcopyrite). BioAccelerant is able
to shorten cycle time by 25% and increase total copper recovery by 20% (Figure 4A). One
particularly attractive feature of the product is its capacity to decrease the lag fase and increase the
rate of iron oxidation, establishing as a consequence a higher redox potential for the process at
shorter cycle times (Figure 4B). These effects are proper to biomass inoculation and hence they do
not constitute an innovation by themselves. BioAccelerant main contribution relies on significant
process management advances (see conclusion section).

16%

Cu Recovery (%)

BTA S.A. is currently performing additional validation tests at laboratory scale in collaboration
with several Chilean mining Companies with the intention to begin industrial scale testing and
validation in the short term.

12%

Blank
BioAccelerant

8%

4%

0%
0

20

40

60

80

100

120

140

Cycle time (days)

Eh (mV, vs SHE)

900

Blank (1)

850

Blank (2)

800

BioAccelerant (1)
BioAccelerant (2)

750
700
650
600
0

20

40

60

80

100

120

Cycle Time (days)

Figure 4 Unitary columns bioleaching tests with a highly chalcopyritic ore. A, Copper extraction. B, PLS
Redox potential

334

BioAccelerant production: a cutting edge technological platform

BioAccelerant production process includes three main operations: fermentation, biomass liquid fase
concentration and lyophilization (freeze drying). These operations are widely used in other
industrial fields such as foods and pharmaceuticals, but there are no previous experiences
involving bioleaching microorganisms at the scale required to fulfill the copper sulfide mining
industry potential demand for a product like BioAcccelerant. In this context, Lyophilization
constitutes a particularly complex and rare technique of prime relevance in BioAccelerant
production flowsheet. BTA S.A. had to adapt lyophilization equipment and procedures to the
specific conditions imposed by the BioAccelerant manufacturing technological platform. Next
section covers the basics of lyophilization and its role in BTA S.A. exclusive technological platform.
Table 2 BioAccelerant main parameters
Parameter
Appearance
Microorganisms load
Storage temperature
Shelf life
Application dose

Quantity
Dry powder
2,57E+12
2,57E+15
Ambient or -20
2 or 6
1,00E+06
1,00E+09
1,00E+12
390

Units
cells/g of product
cells/kg of product
C
months
cells/g of ore
cells/kg of ore
cells/Ton of ore
mg/Ton of ore

Lyophilization (freeze drying)

Stability and viability of the active component (living microorganisms) are determinant aspects in
BioAccelerant technology. Biochemical capacities of sulphur-oxidizing and iron-oxidizing
consortia incorporated in the product formulation have to remain unchanged during transport and
storage periods prior to application in order to meet predefined performance and reproducibility
standards. At the end of growing phase during culture, microorganisms undergo significant
changes in their physiology (Nystrm, 2004). These changes, induced by several factors such as
starvation (lack of carbon and energy source) or the accumulation of toxic compounds in the
surrounding medium, trigger and/or facilitate degenerative processes that affect cell viability
(Nystrm, 2004; Ballesteros et al., 2001). To counteract these processes and preserve
microorganisms in good conditions during long periods of time, methodologies involving freezing
(Pegg, 2007) and / or drying (Morgan et al., 2006) of the sample have been used for decades.
Particularly, one of the most accepted procedures for transport and long-term storage of large
microorganisms collections is lyophilization (freeze drying). This technique involves the removal of
the water contained in a frozen sample through sublimation in vacuum. Water, at sufficiently low
pressure (lower than the triple point of water) evaporates directly from the solid state without a
liquid transition. The temperature both in the freezing phase as during lyophilization should ensure
that all components in the processed sample are in solid phase and remain below the Eutectic point
(Adams, 2007) as partial localized melting could generate zones with potentially harmful high
solute concentrations that would eventually cause heavy cell damage and loss of viability (Adams,
2007).

335

In order to put into perspective and to evaluate the scope of lyophilization as a biomass
preservation method, it is appropriate to analyze the advantages and disadvantages related to its
application.
Advantages (Adams, 2007):
Water sublimation from a frozen sample generates a dry, porous structure in which solutes
are spatially disposed in the same way as in the original solution or suspension. This
prevents the occurrence of unwanted phenomena like excessive solute concentration.
It prevents thermal damage and immobilizes the sample components.
Elimination of water supposes a significant weight loss that directly benefits transport and
storage management.
High cell survival rate trough long-term storage
Overcomed Disadvantages:
High sample processing time
High energy consumption
As a result of intensive and systematic research and development activities carried out over several
years, BTA S.A. successfully applies lyophilization to preserve, transport and store native
bioleaching biomass at industrial scale.

CONCLUSION
BioAccelerant value proposition for sulfide bioleaching processes
Lower costs:
Economy of scale through production centralization
No required investments in biomass plants construction and operation
Low CAPEX and OPEX:
- Low investments related to BioAccelerant implementation and application
- Low operation costs
Optimized management:
BioAccelerant can be managed as an industrial supply:
- Safe storage
- Operational flexibility (inoculum at any time and any cell concentration; possibility
to easily stop and resume inoculation processes as a response to operational
contingencies.
- Lower operational complexity

REFERENCES
Nystrm, T. (2004) Stationary-phase physiology, Annu. Rev. Microbiol., vol. 58, pp. 16181.
Ballesteros, M., Fredriksson, A., Henriksson, J. & Nystrm, T. (2001) Bacterial senescence: protein oxidation in
non-proliferating cells is dictated by the accuracy of the ribosomes, EMBO J., vol. 20, pp. 528089.

336

Pegg, D. E. (2007) Principles of cryopreservation, Methods Mol. Biol., vol. 368, pp. 1538.
Morgan, C. A., Herman, N., White, P. A. & Vesey, G. (2006) Preservation of micro-organisms by drying; A
review, J. Microbiol. Methods, vol. 66, pp. 183193.
Adams, G. (2007) The principles of freeze-drying, Methods Mol. Biol., vol. 368, pp. 1538.

337

The effect of temperature on

column bioleaching of secondary
copper sulfide ores

Bioleaching is a mature technology, which can recover metallic content present

in low-grade ores and widely commercially employed for heap and dump bioleaching of secondary copper sulfide ores. In this work, the bioleaching

Michael Rodrigues, Klinger Lopes ,

Hamilton Lencio, Tcia Veloso and
Versiane Albis Leo. Universidade
Federal de Ouro Preto, Brazil

potential of a low-grade copper ore bearing chalcocite (64%) and bornite (36%)
was investigated. Small columns (10cm 1m height) were inoculated with
either mesophiles (Acidithiobacillus ferrooxidans) or moderate thermophiles
(Sulfobacillus thermosulfidooxidans), at 30 and 50C, respectively, aiming at
determining the effects of temperature on both copper extraction and acid
consumption. These microbes have the potencial for applied use in the recovery
of metal values from copper sulfide ores through the process of bacterial
leaching. The effect of the temperature on both copper extraction and acid
consumption was investigated and enabled a comparison between the different
microbial strains and their utilization in bio-heap-leaching operations. Each
column was loaded with agglomerated ore (10kg/ton acid, top size), then
aerated (2.16m3m-2h-1) from the bottom and irrigated with an acid solution
(8.6Lm-2h-1) from the top. After 255 days, copper extraction reached 88%
with mesophiles and 95% with moderate termophiles. It could be observed
that bioleaching at higher temperatures resulted in both higher metal extraction andfaster leaching kinetics.

339

INTRODUCTION
Heap-bioleaching is a mature technology and an alternative to conventional methods for the
processing of secondary copper sulphide ores and mine tailings as it can cost-effectively process
low-grade and complex ores. From a microbiological viewpoint, a heap is an complex habitat
containing microorganisms capable of promoting the oxidation of both ferrous iron and sulphide
minerals (Watling, 2006). Such microorganisms have remarkable physiological characteristics that
enable them to be used in bioleaching operations, such as growth under conditions of high acidity
and high ionic strength. Temperature plays a major role in the process, as the heap is dominated by
mesophiles, (mostly Acidithiobacillus sp.), at temperatures below 40oC, and pH values between 1.5
and 2.5 (Norris, 2007; Paulino et al., 2001; Schippers, 2007). Moderate thermophiles can also be
important to heap bioleaching, as long as sulphide oxidation and heat generation are fast enough to
promote an increase in the heap temperature to values around 50oC, with benefits to the reaction
kinetics (Sicupira, 2010; Sicupira, 2011; Watling, 2006). Microorganisms of the genus Sulfobacillus,
especially Sulfobacillus thermosulfidooxidans, which grow in the pH range of 1.1 to 2.4 are the most
common (Norris, 2007; Schippers, 2007). Both of them are important to bioleaching as they catalyse
the oxidation of Fe2+ to Fe3+, an important and efficient oxidizing agent of secondary copper
sulphide minerals.
Bioleaching experiments were accomplished by Acar et al. (2005) who inoculated two different 20cm diameter columns, loaded with a copper ore assaying 80% covellite and 10% chalcopyrite, with
mesophiles (A. ferroxidans and L. ferroxidans) and moderate thermophiles (Sulfobacillus). Iron
solubility in the columns operated at 23oC and 65oC reached 30 g/L and 15 g/L, respectively. After
345 days, copper extraction was higher in the column operated at 60-65oC (79%) than in the column
at 23oC (12-20% extraction). The studies performed by Witne & Phillips (2001) on the bioleaching of
a concentrate containing chalcopyrite, bornite, chalcocite, covellite, pyrite, pyrrhotite, galena and
others, with mesophiles (Acidithiobacillus), moderate thermophiles (Sulfobacillus) and extreme
thermophiles (Sulfolobus), revealed a slight increase on copper extraction with temperature, i.e., 76%
extraction at 30oC, 78%, with Sulfobacillus (50oC) and 85% extraction at 70oC.
The present investigation compared bioleaching of a secondary sulphide ore bearing mostly bornite
and chalcocite with mesophilic microorganisms (Acidithiobacillus ferrooxidans) and moderate
thermophiles (Sulfobacillus thermosulfidooxidans) in short columns (1m-high). Such study is justified
because it is an important step in assessing the impacts of high temperatures on both copper
bioleaching and acid consumption.

METHODOLOGY
Bioleaching experiments
Bioleaching experiments were carried out in 100 cm-high x 10 cm-diameter columns divided into
two different sets of 3 columns each. One set was kept at 30oC in a temperature controlled room,
while the second, operated at 50oC, by means of water jacketed fibber glass columns. Each column
was loaded with copper ore (~10Kg) crushed to a top size of '' and previously agglomerated with
sulphuric acid (10Kg/ton). All columns were fed by the top from 10-liter containers at a flow rate of

340

1mL/min (8.6L.m-2.h-1), which was controlled by peristaltic pumps. The oxygen supply (air)
required in the bioleaching process was provided to each column by oil-free compressors at a flow
rate of 250 mL-air.min-1 (2.16 m3-air.m-2.h-1) and fed by the column base.
The biotic experiments were run for a period of nearly 250 days and all columns (except the noninoculated ones) were inoculated with bacteria pre-adapted to the ore. For this purpose, the Norris
growth medium ((NH4)2SO4: 0.4 g/L; MgSO4.7H2O: 0.8 g/L and K2HPO4: 0.4 g/L) along with a 1.0 g/L
de Fe2+ solution were utilized for both the mesophilic (Acidithiobacillus ferrooxidans) and moderate
thermophiles (Sulfobacillus thermosulfidooxidans) strains; while the growth medium was
supplemented with 0.1 g/L of yeast extract for the latter. During the column experiments, the
pregnant leaching solution (PLS) was collected at the base of each column from which liquor
aliquots where withdrawn for subsequent analyzes (pH, Eh and metal concentration). The PLS
volume was then recorded and used in copper extraction calculations, whereas evaporation losses
were compensated with distilled water. After the pH had been adjusted with concentrated H2SO4,
the solution was then recalculated (closed-loop tests). The pH of leach solution inside each column
was corrected in weekly-base to 1.70.1 by varying the entry pH of each column, independently.
Solution bleeding was carried out whenever copper concentration in solution exceeded 4 g/L.

Ore characterization
The copper ore used in the experiments was a secondary sulphide ore bearing 1.80% of copper.
Mineralogical analysis performed by optical microscopy and SEM-EDS indicated biotite (42.3%),
magnetite (21.5%) and silicates, especially amphibole (18.9%) and garnet (6.9%). The copper-bearing
minerals comprised bornite (36%) as well as chalcocite (64%).

Analytical techniques
The pregnant leach solution was accumulated into weekly samples and the volume recorded. A
subsample was taken for analysis of redox potencial, pH and metal concentrations. Measurements
of pH and Eh were carried out in a digital pH/mV meter (DIGIMED, DM-20 model) with a DMEtype flow-electrode CV1 (for Eh) and combined platinum electrode DMR-CP1 (for pH). Metal
concentrations in the PLS were assessed by ICP-OES, Varian 725.

RESULTS AND DISCUSSION

Effects of pH on leaching performance and acid consumption
One of the key parameters in bioleaching operations is pH because each microorganism has an
optimum pH range to grow. The lower limit of this range is around 1.5 for A. ferrooxidans and 1.2
for S. thermosulfidooxidans. If the microorganism rapidly oxidizes Fe2+, jarosite precipitation at pH
values above pH 2.0-2.2 defines the maximum working pH. Therefore, pH 1.70.1 was selected for
the experiments with both mesophiles and moderate thermophiles. The present study showed that
there was precipitation of iron as jarosite at temperatures of 30 C and 50 C which, however, does
not seem to be a limiting factor in the oxidation of sulphide minerals. Authors such as Watling
(2006) and Pina (2006) reported that the precipitation of jarosite is not a limiting factor for oxidation
of sulfide minerals. Aiming at achieving the selected value in the columns, the pH of the inlet
solution was varied from 1.0 to 1.6 to avoid that the pH inside each column increased to values

341

above mentioned range due to the presence of acid consuming minerals in the ore. With such an
approach, pH in the column remained at the set point particularly in the later part of testing.
However, in the first week of experiment, pH values in the PLS were next to 4 in all columns
indicating, as expected, the acid attack on readily soluble minerals from the ore. Values of pH above
3.0 continued to be recorded in the following 30 days of testing. After this period, the dissolution of
acid-consuming minerals slowed, resulting in less scatter in pH readings (PLS), which varied
between 1.6 and 1.8 until the end of the experiments.
The presence of acid-consuming minerals resulted in sulphuric acid consumption values ranging
from 133.0 kg/ton (closed loop, biotic test, 50C) to 230.3 kg/ton (closed loop, non-inoculated test,
30C), i.e., experiments with moderate thermophiles produced the lower acid consumption (Table
1). The highest acid consumption rate occurred within approximately the first 75 days, which was
independent of temperature, reducing in the subsequent period. This high rate corresponds, as
expected, to the period in which the input pH was maintained at its lower value (1.0) and that also
represents the period of maximum metal extraction.
Table 1 Sulphuric acid consumption by secondary sulphide ore
Acid consumption
(kg/ton)

Column

T
(C)

Type of column

1
2

30
30

Control
Open loop

30

Closed loop

50

Control

50

Open loop

206.1
164.4

50

Closed loop

133.0

230.3
191.2
143.9

High values of acid consumption are related to the dissolution of Mg-, Ca- and Al-bearing gangue
minerals; although, in this study, magnesium dissolution was not high. Such dissolution was a
purely chemical process, which was not affected by the presence of microorganisms. The final
extractions observed for aluminium were approximately 17% in the experiments performed at 30C
and 50C (Figure 1), which was coupled to peak concentrations of approximately 1.5 g/L and 2.5
g/L, respectively. On the other hand, calcium concentrations in the PLS were controlled by gypsum
solubility (Tariq,Antti, & Olli, 2012); therefore, values ranging from 0.5 g/L to 0.8 g/L were observed
in the PLS of all columns.

342

a)

Closed loop
Open loop
Abiotic

0
0

50

100

150

200

Aluminium concentrations (g/L)

Aluminium concentrations (g/L)

b)

Closed loop
Open loop
Abiotic

5
4
3
2
1
0

250

50

100

150

200

250

Time (days)

Time (days)

Figure 1 Aluminium solubilisation at mesophilic (a) and moderate thermophilic (b) bioleach conditions; the
vertical lines indicate the bleeding operations

Iron effects on the bioleaching performance

Although the focus of copper bioleaching experiments is the metal extraction, it is also important to
monitor the leaching of iron and its effects on Eh values as ferrous ion is a substrate for bacterial
growth, while the produced Fe3+ is used in the oxidation of sulphide minerals. Furthermore, ferric
ion concentrations are affected by pH and values between 1.5 and 2.0 are considered optimal to
prevent its precipitation as jarosita (Bestamin et al., 2007); although, the precipitation in this pH
range is catalysed at higher temperatures (Halinen et al., 2009a, b). Maximum concentrations of up
to 20 g/L Fe (figure 1) were observed in all temperatures studied, which corresponded,
approximately, to 12% and 10% final extractions for the temperatures of 30C and 50C,
respectively, for the open loop columns (Figure 2).

20

15

10

b)

Closed loop
Open loop
Abiotic

15

Iron extraction (%)

Iron extraction (%)

a)

Closed loop
Open loop
Abiotic

20

10

0
0

50

100

150

Time (days)

200

250

50

100

150

200

250

Time (days)

Figure 2 Percent iron extraction at mesophilic (a) and moderate thermophilic (b) bioleaching conditions

343

Iron concentrations in the PLS were affected by the presence of microorganisms with the noninoculated columns producing by far the largest iron concentrations in the PLS, suggesting that
jarosite formation controlled iron solubility in the inoculated experiments.
In those columns inoculated with Acidithiobacillus ferrooxidans, an increase in the recorded Eh to
values around 700 mV vs. Ag/AgCl was observed after approximately 20 days, regardless of the
experimental condition applied. Conversely, the columns inoculated with S. thermosulfidooxidans
depicted longer lag phases, which lasted 54 days. This was particularly observed in the closed loop
condition and implies that mesophilic microorganisms showed a higher adaptability to the ore than
the moderate thermophiles.

Copper extraction
Table 2 presents copper extraction yields both at 30C and 50C and, as expected, those tests
performed in the presence of microorganisms exhibited copper extractions, which exceeds that
observed in the columns without microorganisms (non-inoculated). However, in the first 12 days,
approximately, (data not shown) metal extractions in both tests were equivalent, indicating that
during this period of time, the copper was solubilised via chemical leaching, possibly as a
consequence of the presence of acid in the agglomeration step. It was also observed that final
copper extractions did not vary regardless the conditions applied, which reached 85.88 % and 88.04
% under conditions of open and closed loop, respectively.
Table 2 Chemical analysis of the bioleaching residues copper assay
Copper extraction (%)

Column

T
(C)

Operating
conditions

Cu (%)
residue

1
2

30
30

Non-inoculated
Open loop

30

4
5

50

50

Open loop

0.18

90.96

90.93

50

Closed loop

0.15

92.54

92.90

Residue

PLS

0.36
0.29

83.22
85.88

83.91
88.84

Closed loop

0.24

88.04

84.04

Non-inoculated

0.56

73.28

79.01

Similarly, in columns inoculated with Sulfobacillus thermosulfidooxidans, the final copper extractions
were not significantly affected by the different conditions adopted. However, the use of a higher
temperature (50C) increased both the reaction kinetics and final copper extraction as compared to
the columns kept at 30C. It was observed 90.96 % and 92.54 % copper extractions, respectively, in
the open and closed loop columns. Analyzing the open loop column at 50C, it was observed
slower reaction kinetics than that of the remaining columns, after 120 days of testing, which took
more than 220 days, approximately, to exceed the same extraction (90%).
Chemical analysis of the bioleaching residue revealed differences in copper extraction, as compared
to the values determined from metal concentrations in the PLS (Table 2), as also observed by Acar
et al., (2005). Likely, these small discrepancies were due to the difficulties related to the errors in
chemical analysis as well as the volume of solution used in extraction determinations.

344

CONCLUSIONS
The columns inoculated with Acidithiobacillus ferroxidans (30C) and moderate thermophiles
(Sulfobacillus thermosulfidooxidans, 50C) reached maximum copper extractions of 88.0 % and 92.5 %,
respectively, and faster leaching kinetics at higher temperatures. The fast Fe2+ bio-oxidation by both
bacteria allowed Eh values between 600 mV to 715 mV vs. Ag/AgCl to be achieved, regardless of
the condition adopted (closed or open loops). The dissolution of gangue minerals promoted lower
H2SO4 consumptions in the columns maintained at 50C (132.96 kg/ton - 206.12 kg/ton). Such
consumption was also affected by the conditions prevailing within the same adopted temperature,
as follows: control> open loop> closed loop. It was also observed that most acid has been spent
during the first 75 days, despite the experimental setup. Such behaviour can be explained by the
higher dissolution of gangue minerals during this period, which is supported by relatively higher
calcium and aluminium concentrations in solution. In general, the biotic systems promoted both
faster extraction kinetics and lower H2SO4 consumptions, than those observed in the non-inoculated
columns; therefore, demonstrating the effectiveness of copper bioleaching by both bacteria.

ACKNOWLEDGMENTS
The financial support from the funding agencies FINEP, FAPEMIG, CNPq and CAPES as well as
Vale are greatly appreciated. The CAPES and CNPq scholarships to M. L. M. Rodrigues and V. A.
Leo, respectively, are particularly acknowledged.

REFERENCES
Acar, S., Brierley, J. A., & Wan, R. Y. (2005). Conditions for bioleaching a covellite-bearing ore. Hydrometallurgy,
77(3-4): 239-246.
Bestamin, O., Erkan, S., Pauliina, N., & Anna, H. K. (2007). Iron oxidation and precipitation in a simulated
heap leaching solution in a Leptospirilum ferriphilum, dominated biofilm reactor. Hydrometallurgy, 88:
67-74.
Halinen, A.-K., Rahunen, N., Kaksonen, A. H., & Puhakka, J. A. (2009a). Heap bioleaching of a complex sulfide
ore: Part I: Effect of pH on metal extraction and microbial composition in pH controlled columns.
Hydrometallurgy, 98(1-2): 92-100.
Halinen, A.-K., Rahunen, N., Kaksonen, A. H., & Puhakka, J. A. (2009b). Heap bioleaching of a complex sulfide
ore: Part II. Effect of temperature on base metal extraction and bacterial compositions.
Hydrometallurgy, 98(1-2): 101-107.
Norris, P. R. (2007). Acidophile diversity in mineral sulfide oxidation. In R. D. E., & D. B. Johnson (Eds.),
Biomining: 199-216. Berlin: Springer.
Paulino, L. C., Bergamo, R. F., Maricilda, M., Garcia-Jr, O., Manfio, G., & Ottoboni, L. (2001). Molecular
characterization of Acidithiobacillus ferrooxidans and A. thiooxidans strains isolated from mine
wastes in Brazil. Antonie van Leeuwenhoek, 80(1): 65-75.
Schippers, A. (2007). Microorganisms involved in bioleaching and nucleic acid-based molecular methods for
their identification and quantification. In E. R. Donati, & W. Sand (Eds.), Microbial Processing of Metal
Sulfides: 3-33. Dordrecht: Springer.

345

Sicupira, L. C. (2011). Biolixiviao de Sulfetos Secundrios de Cobre por Sulfobacillus thermosulfidooxidans.

Unpublished Master, Federal University of Ouro Preto, Ouro Preto, Brazil.
Tariq, M. B., Antti, V., & Olli, H. T. (2012). Dissolution of non-sulfide phases during the chemical and bacterial
leaching of a sulfidic black schist. Hydrometallurgy, 117: 32-35.
Watling, H. R. (2006). The bioleaching of sulphide minerals with emphasis on copper sulphides - A review.
Hydrometallurgy, 84(1-2): 81-108.
Witne, J. Y., & Phillips, C. V. (2001). Bioleaching of Ok Tedi copper concentrate in oxygen- and carbon dioxideenriched air. Minerals Engineering, 14(1): 25-48.
Acar, S., Brierley, J. A., & Wan, R. Y. (2005). Conditions for bioleaching a covellite-bearing ore. Hydrometallurgy,
77(3-4): 239-246.
Bestamin, O., Erkan, S., Pauliina, N., & Anna, H. K. (2007). Iron oxidation and precipitation in a simulated
heap leaching solution in a Leptospirilum ferriphilum, dominated biofilm reactor. Hydrometallurgy, 88:
67-74.
Halinen, A.-K., Rahunen, N., Kaksonen, A. H., & Puhakka, J. A. (2009a). Heap bioleaching of a complex sulfide
ore: Part I: Effect of pH on metal extraction and microbial composition in pH controlled columns.
Hydrometallurgy, 98(1-2): 92-100.
Halinen, A.-K., Rahunen, N., Kaksonen, A. H., & Puhakka, J. A. (2009b). Heap bioleaching of a complex sulfide
ore: Part II. Effect of temperature on base metal extraction and bacterial compositions.
Hydrometallurgy, 98(1-2): 101-107.
Norris, P. R. (2007). Acidophile diversity in mineral sulfide oxidation. In R. D. E., & D. B. Johnson (Eds.),
Biomining: 199-216. Berlin: Springer.
Paulino, L. C., Bergamo, R. F., Maricilda, M., Garcia-Jr, O., Manfio, G., & Ottoboni, L. (2001). Molecular
characterization of Acidithiobacillus ferrooxidans and A. thiooxidans strains isolated from mine
wastes in Brazil. Antonie van Leeuwenhoek, 80(1): 65-75.
Pina, P. S. (2006). Estudo da biolixiviao e a da lixiviao qumica de um concentrado sulfetado de zinco. Unpublished
Estudo da biolixiviao e a da lixiviao qumica de um concentrado sulfetado de zinco, Universidade
Federal de Ouro Preto, Ouro Preto, MG, Brazil.
Schippers, A. (2007). Microorganisms involved in bioleaching and nucleic acid-based molecular methods for
their identification and quantification. In E. R. Donati, & W. Sand (Eds.), Microbial Processing of Metal
Sulfides: 3-33. Dordrecht: Springer.
Sicupira, L., C., Veloso, T., C., Rodrigues, I., C., B., Oliveira,V., A., Leo, V., A. (2010). Fluoride ion effects on the
kinetics of ferrous iron oxidation by Sulfobacillus thermosulfidooxidans. Paper presented at the 5th
International Symposium on Bio-&Hydrometallurgy (BioHydromet '10), Cape Town, South Africa.
Tariq, M. B., Antti, V., & Olli, H. T. (2012). Dissolution of non-sulfide phases during the chemical and bacterial
leaching of a sulfidic black schist. Hydrometallurgy, 117: 32-35.
Watling, H. R. (2006). The bioleaching of sulphide minerals with emphasis on copper sulphides - A review.
Hydrometallurgy, 84(1-2): 81-108.
Witne, J. Y., & Phillips, C. V. (2001). Bioleaching of Ok Tedi copper concentrate in oxygen- and carbon dioxideenriched air. Minerals Engineering, 14(1): 25-48.

346

Leaching of copper oxide ore by

in situ biological generation of
sulphuric acid

In the extractive metallurgy sector sulphuric acid is used for extracting

metals from different mineral species. From sulfide minerals, the metals
can be extracted by using acidophilic microorganisms that need low pH for

Debora Monteiro de Oliveira, Luis

Sobral and Diogo de Oliveira
Padro. Centre for Mineral
Technology, Brazil

the oxidation of such species to take place releasing the metal of interest
into a solution to be further recovered through a particular downstream
process. In the case of a weathered ore, where mineral species derived from
previous sulfide species weathering processes in nature are found, such
oxides, carbonates, sulfates etc., one simply needs to react these species
with sulphuric acid solution. However, operations such as metal leaching
often have mines and plants located in remote areas and in countries lacking
the infrastructure necessary to handle the extremely dangerous concentrated
sulphuric acid. Most leaching operations use aqueous solutions of such acid
that contain less than 20g/L-1 of sulphuric acid. Therefore, this technical
contribution goes a little bit further and is aimed at generating sulphuric
acid, biologically, by mixing, in particular, a copper oxide ore with elemental
sulphur, in a bench scale, in a orbital shaker, where the ground copper oxide
ore and elemental sulphur were suspended in an acid solution bearing acidophilic microorganism and nutrients (MKM medium) with initial pH adjusted
to 1.8. During the leaching process the pH and Eh were monitored until the
acid concentration reached nearly 4.17g/L, which corresponds to pH 1.37, as
a result of the bio-oxidation of elemental sulphur. The copper-bearing drained
solution, apart from being sufficiently concentrated in copper, to be further
used in the downstream process, was sufficiently concentrated in sulphuric
acid to be detrimental to the microorganisms metabolism. Finally, it was
realized that when running the leaching process for four days nearly 90% of
the copper was extracted.

347

INTRODUCTION
In times of economical down turn and, therefore, a period of lower copper prices, it is increasingly
important to improve current copper extraction processes regarding their efficiency. So far as we
can tell, approximately 20% of the world copper production are reached by hydrometallurgical
leaching processes, in particular by leaching oxide copper ores with sulphuric acid and subsequent
downstream processing of the leachate through solvent extraction followed by the electrowinning
of copper (SX-EW).
At present, heap leaching is a well-known industrial process for the treatment of minerals with
oxidized species and secondary sulphides from base metals, usually supported by bacterial
techniques. The latter, are now also being widely explored and tested in the treatment of primary
sulphides.
Sulphuric acid is used in a wide variety of commercial settings. In mining operations, such acid is
used in heap or run-of-mine stock-pile leaching of ore materials and recovery of desired metal
values through SX-EW. This acid supply for use in heap leaching of copper, and other base metals
can be obtained from different sources [United States Patent (2005), no. no. US 2004/0023350 and
2004/0022698]. However, there are significant costs associated with the production, purchase,
transfer and transportation of acid that is generated by any of these processes. In addition to the
cost, the amount of acid required for heap leaching operations varies with time depending on the
availability of feed material for raising that heap. In general, this demand has been increasing
worldwide in recent years.
Conventionally, copper oxide ores have been leached in a heap using sulphuric acid solution and
leachate being further sent through downstream processes for getting pure metallic copper
[BioMineWiki; G.A. Kordosky, (2002)]. The Figure 1, as follows, shows the whole process for
extracting copper from oxide ore.

Figure 1 Conventional leach/solvent extraction/electrowinning flow sheet

While various biological oxidation methodologies are known, such methodologies have uniformly
been used, so far, in connection with metal recovery processing techniques [R. Garca de la Fuente
et al., 2007]. Accordingly, a method to produce low cost acid in an environmentally acceptable
manner, such as through the use of biological oxidation processing, would be advantageous. In this

348

study, the ground ore mixed with elemental sulphur powder were suspended in an acid solution
bearing nutrients and mesophiles microorganisms so as to produce most of the acid enough to
solubilise the copper bearing minerals.

METHOLOGY
Mineral sample
In this study a copper weathered ore was used. After crushing, grinding, and homogenization
procedures, a particle size varying from 0.105 to 0.149 mm was classified for being used in a
simplified technological characterization and in leaching experiments. The analyses carried out by
atomic absorption spectrometry, after an appropriate acid digestion of a representative sample of
the ore, revealed a copper content of 0.64%.
The copper-bearing mineral species comprising the copper ore are: Brochantite (Cu4(SO4)(OH)6,
Malachite (Cu2(CO3)(OH)2), Yakhontovite (Na2.25K0.75Mg4Al4.5Mg2.5Si21.5Al2.5O60), Posnjakite
(Cu4(SO4)(OH)6), Woodwardite (Cu4.Al2(SO4)(OH)12), Lavendulan (NaCaCu5.(AsO4)4.Cl.5(H2O),
Byite (Cu.2Pb.4(SO4)(OH)8), Phosphofibrite (KCuFe3+.15(PO4)12.(OH)12.12(H2O)), Tsumebite
(Pb.2Cu(PO4)(SO4)(OH)), as shown in the difratogram of Figure 2.

Figure 2 Difratogram of the copper ore.

349

Micro-organism
The bacterium Acidithiobacillus thiooxidans (Strain FG01) used in that study was cultivated at 30oC, in
orbital shaker at 150 rpm, using the modified 9K culture medium (MALIK et al., 2001), (pH 2.8),
with the following composition: (NH4)2SO4 (1.0 g.L-1), MgSO4.7H2O (0.5 g.L-1), K2HPO4 (0.5 g.L-1), S
(10 g.L-1)

Leaching experiment
The experiments were accomplished in 500 mL Erlenmeyer flasks containing 200 ml of MKM
culture medium (OLSON & CLARK, 2008). in a 1:5 dilution and 10 g of copper oxide ore (pulp
density of 10% w/v). The culture medium had its pH adjusted to 1.5 with 5M H2SO4 solution with
the following composition (Table 2):
Table 2 Composition of the culture medium used in the bioleaching tests.
Reagent

[g.L-1]

(NH4)2SO4

0.4

MgSO4.7H2O

0.4

K2HPO4

0.04

The microbial inoculum was added at a ratio of 10% v/v and 10 g.L-1 of elemental sulphur were also
added. Simultaneously, leaching tests were accomplished without adding elemental sulphur. In
this case, it was necessary to adjust the pH through the addition of sulphuric acid to maintain pH
around 1.5. The Erlenmeyer flasks were incubated in an orbital shaker under at 150 rpm, at a
temperature of 30o 1oC for 15 days. No sterilized samples were used looking for, thus, establishing
conditions close to reality before industrial application (OLIVEIRA, 2009). The water losses by
evaporation were estimated by the mass losses and compensated by adding water before each
sampling.

RESULTS AND DISCUSSION

After 4 days of process 90%, approximately, of copper extraction were reached. The copper
extraction while running the leaching process adding elemental sulphur was similar to that
achieved in conventional chemical leaching as can be observed in Figure 3. The experiment was
carried out for fifteen days; however, it was not possible to extract the whole copper from the ore,
as some of the mineral species bearing copper are not ready soluble in the acid solution used.

350

Figure 3 Copper extraction in the leaching tests

The Figure 4 shows the variation of the redox potential throughout the experiments. It was realized,
in the test where elemental sulphur was added, such potential values remained higher when
compared to conventional acid leaching where the raise of the redox potential is just due to the
available ion species in solution during the acid dissolution of minerals.

Figure 4 Variation of redox potential in the leaching tests

351

Regarding the pH variation (Figure 5a), it was observed the same trend in both tests. The big
difference between these tests is that, in the test called "Conventional acid leaching", it was
necessary to add sulphuric acid for keeping the pH around 1.5. In the test where elemental sulphur
was added the generation of sulphuric acid took place according to equation 1.
A.thiooxidans
S 0 + 1.5O2 + H 2 O

H 2 SO4

(1)

The oxidation of sulphur by the bacteria present in the reaction system has ensured the supply of
sulphuric acid required for dissolving the mineral species bearing copper.

(A)

(B)

Figure 5 pH variation (A) and acid consumption (B) in the leaching tests

To extract 88.71 % of copper using conventional chemical leaching it was necessary to use 35.53
kilograms of sulphuric acid per tonne of ore. In the test with in situ acid production nearly 90% of
copper were extracted in the same period of time (4 days).

CONCLUSIONS
After running the leaching test for four days, in bench scale, it was possible to observe that:
1. Nearly 90% of the copper content in the ore were extracted by in situ biological generation of
sulphuric acid;
2. In the test where elemental sulphur was mixed with the copper ore, the continuous generation
of sulphuric acid took place, which ensured the suitable pH and Eh for leaching most of the
copper-bearing mineral species due to the action of A. thiooxidans in the operation conditions
used;
3. The copper extraction results wave for the possibility of leaching base metals oxide ores, scaling
up such technology so as to get the right figures for raising a demonstration heap with in situ
generation of sulphuric acid.

352

ACKNOWLEDGEMENTS
We would like express our gratitude to Caraba Mining Co. for providing the ore sample for this
study to take place.

REFERENCES
United States Patent (2005), no. US 2004/0023350 A1, February, 4, 2004.
United States Patent (2005), no. US 2004/0022698 A1, February, 4, 2004.
BioMineWiki: www.wiki.biomine.skelleftea.se, status 19.09.09.
Kordosky G.A., (2002) The Journal of The South African Institute of Mining and Metallurgy, Copper recovery using
leach/solvent extraction/electrowinning technology: Forty years of innovation, 2.2 million tonnes of
copper annually, November/December, pages 445-450.
Malik, A.; Dastidar, M. G.; Roychoudhury, P. K. Biodesulfurization of coal: effect of pulse feeding and leachate
recycle. Enzime and Microbial Technology, v. 28, p. 40 56, 2001;
Garca R. de la Fuente, C. Carrin, S. Botella, F. Fornes, V. Noguera, M. Abad, Bioresource Technology, (2007),
Biological oxidation of elemental sulphur added to three composts from different feedstocks to reduce
their pH for horticultural purposes, 98, 3561-3569.
Oliveira, D. M., Influncia do tensoativo (ramnolipdio) comercial na biolixiviao de minrio primrio de
cobre. Rio de Janeiro: UFRJ/EQ, 2009.
Olson, G.J.; Clark, T. R, Bioleaching of molybdenite, Hydrometallurgy v. 93, p. 1015, 2008;

353

A real-time method for detecting active

microorganisms in commercial-scale
biohydrometallurgical processes

Due to the fact that the entire bioleaching process depends strictly on the
presence and viability of microorganisms, it is capital to have methodologies
that allow rapid, reliable and effective monitoring of the biological activity

Davor Cotoras and Pabla Viedma.

Biohidrica, Biotecnologas del Agua
Ltda., Chile

in commercial-scale biohydrometallurgical processes. The existing biological

activity must be known at different sampling points in the leaching heap to
ensure efficient metal recovery and to take corrective measures whenever
necessary. However, there are, currently, no useful methodologies that can
effectively replace classic techniques for determining total viable bacteria,
which have a very long analysis times (between 7 and 14 days). In a former
presentation, we described the development of a rapid and simple method
for detecting the presence of acidophilic microorganisms in aqueous samples.
This method is based on concentration and washing of the microbial cells,
and detection of active microorganisms by means of its cellular ATP using
a bioluminescent reaction. Unlike conventional methods, this technology
enables the determination of active microorganisms in just 10 minutes. To
facilitate implementation of this method a simple kit was designed. This kit,
called LixKit, contains all materials needed to make such determination.
This work aimed at demonstrating the application of LixKit as a real-time
method for detecting the active microorganisms in commercial-scale biohydrometallurgical processes. LixKit has been applied to 85 samples of 6
mining companies and 8 samples of acid mine drainage. The determination
of microbial activity by LixKit was performed at the mining site. In parallel,
the same samples were transported to the laboratory for analysis with conventional methods of total count and viable count by the Most Probable
Number technique. These results showed a significant correlation between
this real-time method and conventional methods. The simplicity of LixKit
allows it to be used without problem by an operator at the mining site.

355

INTRODUCTION
Due to the fact that the entire bioleaching process depends strictly on the presence and viability of
microorganisms, it is capital to have methodologies that allow rapid, reliable and effective
monitoring of the biological activity in commercial-scale biohydrometallurgical processes.
Therefore, during monitoring of sulphide leaching heaps, the existing biological activity must be
known at different sampling points in the leaching heap to ensure efficient metal recovery and to
take corrective measures whenever necessary. However, currently there are no modern
methodologies that can effectively replace classic techniques for determining total viable bacteria,
which have a very long analysis time (between 7 and 14 days) and therefore do not allow carrying
out the required corrective actions in time.
Microorganism detection in bioleaching heaps are currently carried out by using different classic
methods for determining total and viable leaching bacteria. The total microbial count can be
determinate by direct count using a phase contrast microscope and a Petroff-Hausser chamber. The
culture-dependent viable count of microorganisms is based on determining microbial growth in
different adequate culture media (Harrison, 1984; Johnson, 1995; Lafleur et al., 1993).
In the last years, the use of molecular biology techniques has become a very powerful tool to
identify leaching microorganisms without microorganism culturing (Pizarro et al., 1996; Vsquez,
Moore and Espejo, 1999; Bond, Smriga and Banfield, 2000 Coram and Rawling, 2002; Demergasso et
al., 2005 Remonsellez et al., 2009).
Despite these techniques give valuable qualitative and quantitative information about the diversity
of microbial communities that take part in leaching processes or are responsible for acid mine
drainage formation, such techniques are not able to give an account of the activity or viability of
such microorganisms.
An interesting alternative is the detection of intracellular microbial ATP by means of an enzyme
reaction (luciferin-luciferase) according to Reaction 1.
Luciferase
D-Luciferin + ATP + O2

Oxyluciferin + AMP + PPi + CO2 + light

(1)

The amount of light is proportional to the concentration of ATP in the original sample. Due to all
metabolic active microorganisms show practically constant concentration of ATP, the reaction has
been proven to be proportional to the amount of microorganisms present in solution (Hammes et
al., 2010; He et al., 2009).
Despite the relevant development of the bioluminescence-based methodology, none of these
publications describe a technique that could be applied to detect the presence of active acidophilic
microorganisms that are important in bioleaching processes or responsible for the production of
acid mine drainage. Measurement of active acidophilic microorganisms present in these samples
has always had significant difficulties. The main problems for the analysis using this technology are
the presence of inorganic compounds (toxic metals, high ionic strength etc.), the extreme conditions

356

in which these microorganisms proliferate (particularly dominated by very low pH values) and the
low cell density attained by these microorganisms in their growth. This has made their detection by
means of regular bioluminescence techniques very difficult.
In a former presentation, we described the development of a rapid and simple method for detecting
the presence of acidophilic microorganisms in aqueous samples (Cotoras and Viedma, 2011). This
method is based on concentration and washing of the microbial cells, and detection of active
microorganisms by means of its cellular ATP using a bioluminescent reaction. Unlike conventional
methods, this technology enables the determination of active microorganisms in just 10 minutes. To
facilitate implementation of this method a simple kit was designed. This kit, called LixKit, contains
all materials needed to make the determination. This work aimed at demonstrating the application
of LixKit as a real-time method for detecting the active microorganisms in commercial-scale
biohydrometallurgical processes.

METHODOLOGY
Microorganisms and culture conditions
The microorganisms used in this study were Acidithiobacillus ferrooxidans strain ATCC 23270 and
Acidithiobacillus thiooxidans strain DSMZ 14887. Acidithiobacillus ferrooxidans strain was grown in
ferrous iron-containing modified 9K medium as described previously (Viedma, 2010),
Acidithiobacillus thiooxidans strain was cultured in culture medium 71 in accordance with German
Collection of Microorganisms and Cell Cultures (DSMZ) recommendations. 250 mL Erlenmeyer
flasks containing 100 mL of the culture medium were used. Each flask was inoculated with 0.5 mL
of a culture. Subsequently, the flasks were incubated in an orbital shaker at 30C and 120 rpm.
Sampling of bioleaching solutions
Bioleaching solutions were obtained from the following mines: Minera Los Bronces, Minera
Escondida, Minera Carmen de Andacollo, Minera Barrick Zaldivar, Minera Spence and Minera
Quebrada Blanca. Each sample batch consisted of 4-8 samples each time. Samples were collected in
500 mL polypropylene clean jars. The luminescence assays were performed in the field site. Later
on the samples were brought back to the laboratory for the following tests: total count of
microorganisms, viable bacterial counting by the most probable number count (MPN) (Lafleur et al.,
1993).
Bioluminescence-based bacterial counting
The quantification of active bioleaching microorganisms was performed using the LixKit
(Biohidrica, Chile), according to the manufacturer's instructions. A volume of 10 mL of the
microorganism culture to be assayed was passed through a filter holder with a 0.22 m membrane
using a syringe. The second stage of this assay consists in the removal of the agents that are
inhibitory for the bioluminescence reaction. It was carried out by sequentially washing of the
previously concentrated acidophilic microorganisms. In order to do that, 20 mL of solution 1 were
passed through the filter and after that the filtrate was discarded. Then the membrane was rinsed
with 20 mL of solution 2 and the filtrate was discarded again. Finally, using moderate pressure over
a syringe plunger air is allowed to pass through the membrane to remove the remaining solution).

357

A swab was wetted by immersion in solution 2, and was rubbed against the surface of the
membrane, keeping a constant pressure. The third stage of the assay corresponds to the extraction
of intracellular adenosine-triphosphate (ATP) from the acidophilic microorganisms and the
performance of the bioluminescence reaction. Finally, emitted light was immediately measured in a
luminometer (Kikkoman Lumitester PD-20, Japan). The values obtained are expressed in relative
light units (RLU). Figure 1 shows a schematic representation of the procedure applied for the
quantification of active bioleaching microorganisms using the LixKit

Figure 1 Steps of the quantification of active bioleaching microorganisms using the LixKit

RESULTS AND DISCUSSION

Detection of active microorganisms in samples of commercial-scale heap and dump
bioleaching processes using the LixKit method
The goal of this work was to demonstrate the application of LixKit as a real-time method for
detecting the active microorganisms in heap or dump bioleaching samples. The main factors that
could interfere with the analysis of these samples are low microbial concentration and presence of
high concentrations of toxic metals and low pH. A total of 85 bioleaching solution samples were
collected from different heap and dump bioleaching of six mining companies (Minera Los Bronces,
Minera Escondida, Minera Carmen de Andacollo, Minera Barrick Zaldivar, Minera Spence and
Minera Quebrada Blanca). These samples include 40 heap leaching solutions, 14 dump leaching

358

solutions, 10 raffinate solutions, 7 pregnant leaching solutions, 5 intermediate leaching solutions, 6

bioreactor solutions and 3 bioleaching column solutions. The purpose of this is to demonstrate the
applicability of the method in different types of samples. The analysis by LixKit was conducted
immediately at the sampling place. This operation was done by a researcher. However, in some
cases, the challenge for this new kit was even greater, since due to safety regulations, the operation
had to be done only by professionals or technicians of the mining company. Subsequently, the same
samples were analyzed in the laboratory by conventional methods of total count and viable count.

Bioluminescence (RLU)

1000000
100000
10000
1000
100
10
5,000E+03 5,000E+04 5,000E+05 5,000E+06 5,000E+07 5,000E+08
Total count (microrganisms/mL)
Figure 2 Correlation between the LixKit bioluminescence method and total cell concentrations of heap and
dump bioleaching solution samples and pure cultures of A. ferrooxidans and A. thiooxidans (R2=0.774)

Bioluminescence (RLU)

1000000
100000
10000
1000
100
10
1,000E+00 1,000E+02 1,000E+04 1,000E+06 1,000E+08
Viable count (microrganisms/mL)
Figure 3 Correlation between the LixKit bioluminescence method and viable cell concentrations (iron- and
sulphur-oxidizing bacteria) of heap bioleaching solution samples and pure cultures of A. ferrooxidans and A.
thiooxidans (R2=0.565)

359

Figure 2 shows that although occurs dispersion in analyzing actual samples of different origin,
there is a linear relationship between the LixKit method based on bioluminescence and total count
of microorganisms. The data show a correlation coefficient of 0.774, which is normally considered
adequate when validating a new analysis technique. In this case, it is important to realize that they
are being compared two methods of analysis based on completely different principles.
In a second comparison between the LixKit method based on bioluminescence and the viable
counting of microorganisms it is also possible to find a linear relationship (Figure 3). To obtain an
estimate of total viable microorganisms, the viable counting iron- and sulphur-oxidizing bacteria
microorganisms by MPN were added. The data show a correlation coefficient of 0.565. This
correlation may be considered appropriate, since the comparison is between two methods, which
are based on totally different principles of analysis. Furthermore, it should be noted that the
conventional methodology for analysing viable microorganisms based on MPN has a high
experimental error of around 30% (Lafleur et al., 1993), significantly increases the deviation of the
experimental data.
These results showed a significant correlation between this real time method and conventional
methods even though the samples had very different origin, such as bioleaching solutions from
heaps of different operating times, PLS or raffinate. The experience gained in the various mining
companies also showed that LixKit can be used without problem by an operator at the mining site.

Detection of the effect of environmental temperature on the microbial activity in

samples of commercial-scale dump bioleaching processes and acid mine drainages
using the LixKit method
Detecting microbial activity in ROM dump leaching solutions operations represents a particular
challenge because of the low concentration of microorganisms commonly found in these samples. A
similar situation occurs when trying to determine the microbial activity in acid mine drainage. It is
important to note that in Chile dump bioleaching processes and generation of acid mine drainage
occurs mainly in the high mountains. They are open systems that are directly affected by
environmental temperature changes.
In this work, we sampled solutions that percolate from two ROM dump leaching operations and
two solutions of the acid drainage in Minera Los Bronces. It was found possible to determine
microbial activity of these samples with the LixKit. Figure 4 shows the seasonal change of the
monthly average temperature and microbial activity. All determinations of microbial activity
followed the same trend than the average environmental temperatures. In July (monthly average
temperature: -6 C) it was found the lowest bioluminescence, which clearly shows an effect of
environmental temperature on the microbial metabolic activity. This shows the good performance
of the kit in extreme conditions where the microbial detection is difficult due to low population
densities.

360

Bioluminescence (RLU)

1400

1200
0

1000
800

-2

600

-4

400

Average temperature (oC)

1600

1
2
3
4

-6

200

0
-8
02/05/yy 21/06/yy 10/08/yy 29/09/yy 18/11/yy

Autumn

Winter

Spring

Sample date
Figure 4 Seasonal change of the monthly average temperature and microbial activity, measured as
bioluminescence (LixKit method) in two samples of ROM dumps leaching solutions and two samples of acid
mine drainage. (1) Acid mine drainage 1, (1) acid mine drainage 2, (3) dump leaching solution 1, (4) dump
leaching solution 2

CONCLUSIONS
These results showed a significant correlation between this real time method and conventional
methods. The simplicity of LixKit allows it to be used without problem by an operator at the
mining site.
In conclusion, the following are the advantages of LixKit for monitoring the activity of microbial
populations in a bioleaching plant:
Rapid determination (10 minutes)
Easy operation (kit ready for use)
High reproducibility
Allows measurements in the field
Determines the metabolic activity of leaching microorganisms -"health" of the microflora
Facilitates the metallurgical and operational decisions

ACKNOWLEDGEMENTS
This project was funded by Innova Chile. We thank the following mining companies for their
cooperation and supply of samples: Minera Los Bronces, Minera Escondida, Minera Carmen de
Andacollo, Minera Barrick Zaldivar, Minera Spence and Minera Quebrada Blanca. We thank the
company Valtek S.A. by the excellent participation in the production of the kit.

361

NOMENCLATURE
AMP
ATCC
ATP
MPN
PCR
PPi
RLU
ROM

Adenosine monophosphate
American Type Culture Collection
Adenosine triphosphate
Most Probable Number
Polymerase chain reaction
Inorganic pyrophosphate
Relative light units
Run of Mine

REFERENCES
Bond, P. L., Smriga, S. P. and Banfield, J. F. (2000) Phylogeny of microorganisms populating a thick, subaerial,
predominantly lithotrophic biofilm at an extreme acid mine drainage site. Appl. Environ. Microbiol.,
vol. 66,
Coram, N.J. and Rawling, D.E. (2002) Molecular relationship between two groups of the genus Leptospirillum
and the finding that Leptospirillum ferriphilum sp. nov. dominates South African commercial
biooxidation tanks that operate at 40C. Appl. Environ. Microbiol., vol. 68(2), pp. 838-845.
Cotoras, D. and Viedma, P. (2011) A rapid and simple method for detecting active acidophilic microorganisms
in copper bioleaching processes Proceedings of the 6th International Seminar on Copper Hydrometallurgy,
Via del Mar, Chile.
Demergasso, C. S., P. A. Galleguillos, L. V. Escudero, Zepeda V. J., Castillo D., and Casamayor E. O. (2005)
Molecular characterization of microbial populations in a low-grade copper ore bioleaching test heap.
Hydrometallurgy 80:241-253.
Hammes, F., Goldschmidt, F., Vital, M., Wang, Y., Egli, T. (2010) Measurement and interpretation of microbial
adenosine tri-phosphate (ATP) in aquatic environments. Water Research, vol. 44, pp. 3915-3923.
Harrison, A. P., JR. (1984) The acidophic thiobacilli and other acidophilic bacteria that share their habitat. Ann.
Rev. Microbiol., vol. 38, pp. 265-292.
He, B., Liu, X., Yue, W., Zhou, A., Luo, J., & Cai, X. (2009). Rapid detection of bacteria without cultivation with
a portable bioluminescence sensor system. African Journal of Microbiology Research , vol. 3, pp. 575-580.
Johnson, D.B. (1995) Selective solid media for isolating and enumerating acidophilic bacteria. Journal of
Microbiological Methods, vol. 23, pp. 205-218.
Lafleur, R., D.-Roy, E., Couillard,D. and Guay, R. (1993) Determination of iron oxidizing bacteria numbers by a
modified MPN procedure, Biohydrometallurgical Technologies (Ed. A.E. Torma, M.L. Apel and C.L.
Brierley). The Minerals, Metals and Materials Society, pp. 433-441.
Pizarro, J., Jedlicki, E., Orellana, O., Romero, J and Espejo, R. T. (1996) Bacterial populations in samples of
bioleached copper ore as revealed by analysis of DNA obtained before and after cultivation. Appl.
Environ. Microbiol., vol. 62, pp. 1323-1328.
Remonsellez, F., Galleguillos, F., Moreno-Paz, M., Parro, V., Acosta M., and Demergasso, C. (2009) Dynamic of
active microorganisms inhabiting a bioleaching industrial heap of low-grade copper sulfide ore
monitored by real-time PCR and oligonucleotide prokaryotic acidophile microarray. Microbial
Biotechnology vol. 2, pp. 613624.

362

Vsquez, M, Moore ERB, Espejo RT. (1999) Detection by polymerase chain reaction-amplification sequencing
of an archaeon in a commercial-scale copper bioleaching plant. FEMS Microbiol. Lett., vol. 173, pp.
18387.
Viedma, P. (2010) Method for detecting presence of acidophilic microorganisms in bioleaching solution. US Patent
7,851,177 (also patented in Australia: AU 2007203563, Chile CL 48457, Peru PE 6450, Germany
102007035588.4-09 and South Africa N 2007/06363).

363

Bioleaching of covellite from low

grade copper sulfide ore and tails

The economic importance of copper production in Chile has prompted the

study of new strategies for the optimal management of mineral resources,
including processing of low grade ore and even processing tails (processed
ore). Currently, bioleaching is the most convenient technology to treat low
grade copper sulfide ores bearing chalcopyrite, chalcocite and covellite. Covellite is a secondary copper sulfide occurring in copper sulfide deposits and it
is also formed by acidic dissolution of chalcocite. In leaching processes, the

Vctor Zepeda, Pedro Galleguillos

and Cecilia Demergasso.
Universidad Catlica del Norte and
the Center for Scientific and
Technological Research for the
Mining Industry, Chile 
Dina Cautivo, Jos Soto and Yasna
Contador. Universidad Catlica del
Norte, Chile

main factors influencing the dissolution of covellite are the electrochemical

potential (Eh) and temperature.
This work aimed at assessing the effect of different concentrations of total
iron (3.0 and 6.0 g/L) and temperatures (30 and 50C) in the bioleaching of
covellite contained in a low grade copper ore and the same ore after a previous
process of bioleaching at ambient temperature. Initial composition of low-grade ore was chalcocite (0.58%), chalcopyrite (0.56%), covellite (0.30%) and pyrite
(2.73%), while the processed ore contained chalcopyrite (0.47%), covellite
(0.28%), chalcocite (0.10%) and pyrite (3.32%). Covellite was mostly dissolved
(>90%) in both unprocessed and processed ore, independently of the temperature and total iron concentrations. However, a higher total copper recovery
was achieved in tests performed at 50C due to higher dissolution of primary
ores, especially in unprocessed ore. The microbial activity was confirmed
by Most Probable Number (MPN) analysis of iron and sulphur oxidizing
microorganisms. The microbial population was dominated by Acidithiobacillus spp. in tests operated at 30C, while in tests operated at 50C Sulfobacillus spp. were the most abundant organisms.

365

INTRODUCTION
Nowadays, copper bioleaching is considered as the most convenient technology to treat low grade
copper sulphide ores. The increasing global demand of copper combined with bioleaching
technology have expanded the sources of this metal cosidered before as waste, such as low grade
ore and leaching tails materials. In this regard, Chile has been a pioneer country using the
technology of copper bioleaching at industrial level and over 10% of the total copper production is
obtained by bioleaching (Demergasso, 2010). At industrial level, depending on the type and grade
of the sulphide ore, copper bioleaching process is performed either in heaps or dumps. The
mineralogical composition of the ore is an important factor affecting the level of copper recovery in
the bioleaching process. Drew et al., described the preferential order of mineral bioleaching using a
mesophile culture. The dissolution sequence was Chalcocite> Bornite >Cubanite >Covellite >Pyrite
>Enargite >Carrolite >Chacopyrite and the copper recovery levels were among the different ore
species ranged between 40-100% (Drew et al., 2010). Previous bioleaching studies using low grade
copper ores from several mining companies were performed at Centro de Biotecnolgia UCN and
showed that even when suitable levels of copper recoveries are achieved, the mineral tails still
contained copper as chalcopyrite and covellite (demonstrated by QEMSCAN analysis).
Chalcopyrite is a refractory sulphide mineral and generally is poorly dissolved in heaps
bioprocesses operated at ambient temperature (20-30C). By the other hand, covellite is a copper
mineral that occurs in native ore or it is generated after the dissolution of chalcocite by sulphuric
acid (ec. 1).

+2

+2

+2

(1)

The main factors affecting the dissolution of copper sulphide ores are the temperature and Fe III/Fe
II ratio, the latter expressed as redox potential (Boon et al., 1998). Only few studies have been
focused to enhance the copper recovery from covellite in industrial processes. The bioleaching of
covellite using a thermophile culture showed copper recoveries higher than 70%, at 60 C (Sevket et
al., 2005). However, at industrial scale (heap or dump) the system is heterogeneous and
temperature increases slowly favouring the growth of moderate thermophiles, rather than
thermophiles (Demergasso, et al., 2005). On the other hand, the external addition of iron in
bioleaching tests would not have an important effect in the dissolution of covellite when culture of
A. ferrooxidans (Falco et al., 2003) was used. In tests performed using A. thiooxidans, no differences
were found in the dissolution rates compared to abiotic controls, suggesting the absence of a direct
leaching mechanism. The oxidative dissolution of sulphide ores requires oxygen, which is involved
either in chemical or biological reactions. At industrial scale the oxygen gas is supplied by air
blowers located in the base of the heap. However due to the presence of fine material, blockage of
internal stringers or height of the heap, the oxygen would not reach certains zones of the heap, thus
generating anaerobic conditions. Several mesophile and moderate thermophile microorganisms are
able to thrive in anaerobic conditions such as A. ferrooxidans. In this regard Donati et al., found a
significative dissolution of covellite in anaerobic conditions using the mesophile A. ferrooxidans
(Donati et al., 1997). The main goal of this investigation was to assess the effect of different
concentrations of total iron (3.0 and 6.0 g/L) and temperature (30 and 50 C) in the bioleaching of

366

covellite contained in a low grade sulphide ore after being processed and unprocessed in column
tests.

METHODOLOGY
Five column tests were accomplished by duplicate to access the bioleaching of a low grade copper
sulphide ore. Tests were carried out for ores previously processed and original unprocessed. Each
column was 0.5 m height and 0.09 m diameter. Processed and unprocessed ore were sampled and
homogeneously distributed in fourteen loads of 3.0 kg each. The loads of processed and
unprossseced ore were used in the column tests and for other analysis (mineralogical, chemical and
granulometric). The leaching period was 120 days and synthetic raffinate containing 0.1 g/L Cu and
2.0 g/L H2SO4 was used as irrigation solution. Table 1 shows the operational design of each column.
Iron was added to raffinate solution in order to obtain tests conditions (3.0 and 6.0 g/L). Table 2
shows the mineralogical analyses of processed and unprocessed ore.
Tabla 1 Tests column conditions
Column

Type of ore

Fe II (g/L)

Temperature

MC 1

processed

50C

MC 2

processed

30C

MC 3

processed

50C

MC 4

processed

30C

MC 5

unprocessed

50C

Tabla 2 Mineralogical analyses of ore used in column tests

Mineral

Formula

Unprocessed
ore (%)

Processed
ore (%)

Chalcocite

Cu2S - Cu1.4S

0.58

0.10

Covellite

Cu1.4S - CuS

0.30

0.28

Chalcopyrite

CuFeS2

0.56

0.47

Brochantite

Cu4(SO4)(OH)6

0.02

0.01

Chrysocolla

(Cu,Al)2H2Si2O5(OH)4n(H2O)

0.05

0.02

Cu-bearing Clays

Cux.(K,Fe,Mg)x.Alx.Six.(OH)x

0.02

0.00

Pyrite

FeS2

2.73

3.32

Other Sulphides

PbS / ZnS / MoS2

0.11

0.13

Fe Oxides

FeO.OH / Fe2O3 / Fe3O4

1.14

1.14

Quartz

SiO2

23.73

25.51

Chlorite

(Mg,Fe2+)5Al[(OH)8AlSi3O10]

6.92

6.41

Jarosite

(K,Na,H3O)Fe 3(SO4)2(OH)6

0.03

0.03

3+

367

The inocula were composed by microorganisms enriched at 30 and 50C in the presence of iron and
sulphur as energy sources at the Centro de Biotecnologia UCN. The characteristics of the inocula in
terms of oxidizing activity and composition are shown in Tables 3 and 4, respectively.
Tabla 3 Microbial activity characterization of the inoculums used in this study
Inoculum

DAPI
(cells/mL)

NMP-Fe
(cells/ml)

NMP-S
(cells/ml)

Actividad
Ox-Fe (days)

Actividad OxSo (g/L H2SO4)

MC 30oC

1.18E+07

2.30E+04

3.28E+05

3.93

MC 50 C

2.17E+07

1.27E+04

2.16E+03

4.3

Tabla 4 Specific quantification of microorganisms present in the inocula

Inoculum

Bacterias

Acidithiobacillus Acidithiobacillus Leptospirillum Acidiphillum Acidithiobacill Sulfobacillus

Sulfobacillus

Archaea

Totales

thiooxidans

ferrooxidans

Ferriphillum II

sp.

us caldus

MAD

thermosufidoxidans

Total

MC - 30C

2.83E+07

2.12E+06

1.17E+06

7.65E+06

2.78E+06

6.98E+04

3.41E+04

1.95E+04

1.54E+03

MC - 50C

8.43E+07

4.28E+04

3.85E+04

6.68E+04

4.10E+04

3.70E+06

8.51E+06

4.90E+07

4.48E+03

Chemical analysis of the solutions for total iron, copper and aluminum were performed by atomic
absorption spectrophotometry; ferrous iron and proton concentration were determined by titration
and sulphate concentration was determined by gravimetric analysis. pH and Eh were measured
using a pH meter with respective electrodes. Biological analyses including total cell counting by
epifluorescence microscopy (DAPI stain), Most Probable Number (MPN) of iron and sulphur
oxidizing microorganisms and specific counting by real time PCR were performed in liquid and
mineral samples from column tests. The mineralogy of solid samples (including ore and tails) were
performed by QEMSCAN analysis (Andersen et al., 2009).

RESULTS AND DISCUSSION

Copper recovery
Copper recovery was calculated based on the copper concentration determined in irrigation and
percolate solutions, chemical analysis of the processed and unprocessed ore and tails obtained after
120 days of bioleaching. Figure 1 shows the kinetic curves for copper recovery of the column tests.
The copper extracted ranged from 56 to 74%, the highest recoveries were obtained in column tests
MC5 (unprocessed ore, 6.0 g/L iron added and operated at 50 C) and MC3 (processed ore, 6.0 g/L
iron added and operated at 50 C). The curves of Figure 1 show a frist stage including the first 25
days of operation when copper was dissolved at higher rates than other periods. This rapid
dissolution during the first 25 days can be attributed to chemical leaching with no biological effect.
Sequential copper analysis indicated 80.4% of dissolution of oxide ore (Brochantite and chrysocolla)
plus part of the chalcocite leached by effect of the acid (Table 5). In a second stage the dissolution
rate decreased compared to the first 25 days of operation and copper leaching was driven by
microbiological activities. The effect of temperature was also observed in sequential copper analysis
as higher dissolution of primary copper sulphides was achieved in columns operated at 50 C than
30 C (Table 5).

368

Tabla 5 Copper recovery determined in column tests by sequential copper analysis method

Column

Copper
recovered from
oxide ore (%)

Copper recovered
from secondary
sulphide ore (%)

MC1
MC2
MC3
MC4
MC5

76.4 1.3
77.4 5,3
84.9 1.4
81.1 1.5
85.9 1.4

78.6 0.3
72.4 1.7
84.8 1.8
75.2 2.3
82.7 3.1

Copper
recovered from
primary
sulphide ore (%)
43.3 4.7
29.7 3.3
46.7 2.1
29.3 0.9
44.3 1.5

Total copper
recovery (%)

Standard Error
in test (%)

63.1 0.8
57.6 2.8
71.4 1.3
57.9 2.9
74.0 2.8

4.12
4.32
4.11
4.21
4.19

Figure 1 Total copper extraction determined in percolate solutions

The Figure 2 shows the acid consumption in the tests. The highest acid consumption was observed
in column test CM5 (22.05 kg/ton); in addition, column tests operated at 50C had a higher acid
consumption than those operated at 30C. Column test MC4 consumed acid until operation day 23
and then the acid was generated in this test. It can be attributed to higher rate of ferric iron
precipitation as jarosite in the test. The mineralogical analysis of tails of the column test CM6
showed the highest percentage of jarosite (4.26%) compared to tails of tests CM1 and CM5 (0.66 and
0.79%, respectively).

Figure 2 Total acid consumption in percolate solution of column tests

369

The pH profile determined in percolate solution showed higher values at the beginning of the tests
(ranging between 2.5 to 3.0); but, as the process time progressed the pH was maintained around 2.0.
The highest values for the electrochemical potential were observed in column tests performed at
30C. Thus, 600 and 800 mV vs. SHE were achieved in column tests MC2 and MC5, respectively
(data not shown). On the other hand, column tests operated at 50 C maintained the Eh between 550
and 650 mV vs. SHE, during the leaching period (data not shown). The temperature measured in
the base of the columns during the leaching period did not have important variations (Figure 3).

Figure 3 Temperature measured in the base of the columns

Analysis of the microbial population

Total cells counting performed using DAPI stain in percolate solutions showed concentrations
around 106 cells/mL during the leaching period. The MPN of iron and sulphur oxidizers were
higher in column tests MC2 and MC4 observing maximum concentrations of active cells around 104
cells/mL in day 120 (data not shown). Specific counts determined by real time PCR showed
expected compositions of the microbial populations in MC1 and MC4 according to the operation
temperatures (Figure 4).

100%

Relative abundance

Ferroplasma sp.
Sulfobacillus sp. STO

80%

Sulfobacillus sp. MAD

60%

Acidithiobacillus caldus
Acidiphilum sp.

40%

Leptospirillum ferriphilum
Acidithiobacillus thioooxidans

20%
Acidithiobacillus ferrooxidans

0%
PLS
PLS
tail
day 60 day 120 50C
50C 50C

PLS
PLS
tail
day 60 day 120 30C
30C 30C

Figure 4 Structure and dynamics of the microbial population in percolate solutions and tails of tests MC1 and
MC4, determined by real time PCR

370

Mineralogical analyses
The dissolution of covellite observed in tests using processed and unprocessed ores was higher
than 90% and no significant differences were observed by adding different concentrations of iron
(3.0 and 6.0 g/L) or different temperature of operation (30 and 50 C). Previous studies have shown
that maximum dissolution of covellite occurred at 65 C (Sevket et al., 2005 and Lee et al., 2011). In
this study a higher dissolution of chalcopyrite was observed in unprocessed ore compared to
processed ore. This can be attributed the passivation effect of previous bioleaching treatment of the
processed ore (Figures 5 and 6).

Figure 5 Mineralogical analysis of unprocessed ore and tail of MC5

Figure 6 Mineralogical analysis of processed ore and tails of MC1, MC3 and MC4

371

CONCLUSION
The dissolution of covellite in column bioleaching tests using a low grade copper sulphide ore,
previously treated and unprocessed were higher than 90%. The complete dissolution of chalcocite
was achieved at 30 and 50 C and no significative differences were observed. However the effect of
temperature was clearly observed in the dissolution of primary sulphide ores, producing higher
total copper recoveries in column tests operated at 50 C. In tails and percolate solutions of column
tests operated at 30C, the microbial population was maily formed by Acidithiobacillus ferrooxindans
and Acidithiobacillus thiooxindans, while for column tests operated at 50 C Sulfobacillus spp and
Acidithiobacillus caldus were the dominant species.

ACKNOWLEDGEMENTS
Marina Vargas and Mario Avendao, Antofagasta Mineralogy Center Leader BHP Billiton Centro
de Mineraloga Antofagasta.

REFERENCES
A. Sevket, A. James and Y. Rong. Conditions for bioleaching a covellite-bearing ore. Hydrometallrgy 77, pp
239-246.
C. Demergasso, P. Galleguillos, L. Escudero, V. Zepeda, D. Castillo and E. Casamayor. 2005. Molecular
characterization of microbial populations in a low-grade copper ore bioleaching test heap.
Hydrometallurgy 80, pp 241-253.
C. Demergasso, F. Galleguillos, P. Soto, M. Sern, V. Iturriaga. 2012. Microbial succession during a heap
bioleaching cycle of low grade copper sulphides Does this knowledge mean a real input for industrial
process design and control? Hydrometalurgy 104, pp 382-390.
E. Donati, C. Pogliani and J. Boiardi. 1997. Anaerobic leaching of cevellite by Thiobacillus ferrooxidans. Appl.
Microbiol Biotechnol 47, pp 636-639.
J. Andersena, G Rollinsona, B Snooka, R Herringtonb, R. Fairhurst 2009, Use of QEMSCAN for the
characterization of Ni-rich and Ni-poor goethite in laterite ores. Minerals Engineering, 22, pp 11191129.
J. Lee, S. Acar, D. Doerr, J. Brierley. 201. Comparative bioleaching and mineralogy of composited sulphide ores
containing enargite, covellite and chalcocite by mesophilic and thermophilic microorganisms.
Hydrometallrgy 105, pp 213-221.
L. Falco, C. Pogliani, G. Curutchet and E. Donati. 2003. A comparison of bioleaching of covellite using pure
cultures of A. ferrooxidans and A. thiooxidans or a mixed culture of Leptosporillum ferrooxidans and A.
thiooxidans. Hydrometallurgy 71, pp 31-36.
M. Boon, J. Heijnen and G. Hansford. 1998. The mechanism and kinetics of bioleaching sulphide minerals. Min.
P D. Dew, C. Van Buuren, K. McEean and C. Bowker. 2000. Bioleaching of base metal sulphide
concentrates: a comparision of high and low temperature bioleaching. The Jounal of the South African
Institute of Mining and metallurgy, pp 409-414ro. Ext. Rev 19, pp 107-115.

372

chap. 7
Modeling, and optimizing hydrometallurgical
operations and circuits

Implications of hydrodynamic
testing for heap leach design

The design of a heap leach operation incorporates aspects from various engineering disciplines, such as knowledge of the geology, mining plan and
metallurgical performance. The physical and hydraulic (hydrodynamic) properties of ore for heap leaching are often poorly understood and can be the cause
of heap failure. Hydrodynamic performance can become the limiting factor
which controls operational conditions, metallurgical performance as well as

Stefan Robertson. Mintek,

South Africa
Amado Guzmn. HydroGeoSense,
USA
Graeme Miller. Miller Metallurgical
Services, Australia

geotechnical performance of the ore mass. Laboratory procedures have been

developed to characterize ores with respect to hydrodynamic properties. These
include stacking (compression) tests, whereby the bulk density and hydraulic
conductivity profiles as a function of stacking height are determined. Samples
that satisfy the requirements of the stacking test (a minimum porosity of 30%
and a saturated hydraulic conductivity of 100 times the target application rate)
for the selected heap height can be safely advanced to more involved testing
procedures (hydrodynamic or metallurgical column tests). Experience shows
that for copper oxide leaching the maximum liquid saturation should be kept
below 85% and ideally under 65% to accommodate the natural variability
associated with the heap stacking process. For copper sulfide leaching, where
oxygen is required to promote bacterial activity, the maximum liquid degree
of saturation should be less than 65% and ideally below 60% to ensure that
proper air percolation capacity is maintained. Results from the hydrodynamic
characterization together with basic metallurgical data can be used as inputs
for an integrated flow and mineral recovery model. This is based on a solution
of Richards equation coupled with a sorption/desorption approach in order
to calibrate the model on the basis of column data and predict hydraulic and
metal recovery behavior in a heap. This paper presents experimental data for
a mixed oxide-sulfide copper ore and the analysis of the data using a novel
procedure to scale up this information for the purpose of designing a heap
leach process.

375

INTRODUCTION
Although the mining industry invests significant effort on geological and metallurgical
characterization of an ore mined for leaching, little attention has been given to the hydrodynamic
(physical and hydraulic) properties of the ore (Guzman et al, 2008). A proper percolation leaching
process requires that solution flow occurs under partially saturated conditions such that the pore
space is occupied by both solution and air. Poor performance of percolation leach processes can
often be attributed to a lack of understanding of the hydrodynamic properties of the ore governing
the flow of solution and air through the heap profile and the effect of these parameters on the
chemical reactions controlling the dissolution and mobilization of solutes. Guzman et al (2008)
demonstrate the importance of proper agglomeration and its effect on the bulk density of the ore.
The laboratory procedures described by these authors allow determination of the hydrodynamic
parameters necessary for the characterization of solution and air flow through a heap. The
parameters defined by these lab procedures include; a) the density profile the relationship
between bulk density and heap height, b) the conductivity profile the relationship between
saturated hydraulic conductivity and heap height, c) the hydraulic conductivity curve the
relationship between hydraulic conductivity and liquid degree of saturation, d) the moisture
retention curve the relationship between moisture content (or degree of saturation) and the
capillary pressure (suction), and e) the air conductivity curve the relationship between air
conductivity and degree of liquid saturation.
This paper presents hydrodynamic characterization information for a mixed oxide-sulphide copper
ore and the application of the data for the modelling of the expected leaching performance at an
industrial scale. An integrated approach comprising both metallurgical and hydrodynamic
laboratory testing was adopted, with geomechanical testing performed on both feed and residues
from metallurgical columns, at two different testing scales to evaluate the hydraulic properties that
control the leaching performance of the samples. The experimental results are interpreted with a
numerical procedure which facilitates scale up of the data and provides a calibrated parameter set
for the evaluation of the hydrodynamic and metallurgical performance of a heap under alternative
operational conditions.

EXPERIMENTAL PROCEDURES
Stacking tests
The stacking test (ST) is an indicator procedure which determines whether a particular ore sample
will support percolation leaching. Stacking test requirements for a dynamic heap are a minimum
porosity of 30% and a saturated hydraulic conductivity of 100 times the target application rate. The
stacking test is performed by mechanically increasing the confining load in a test cell to determine
the effect of heap height on the bulk density and the hydraulic conductivity of the ore. The stacking
test is conducted under partially saturated conditions, containing only the moisture of
agglomeration, as opposed to fully saturated conditions typically employed by the civil engineering
community. After compression to maximum load, the saturated hydraulic conductivity is
determined by passing solution through the bed from a constant head reservoir. The compressed

376

sample represents a unit volume of ore located at the bottom of the heap. Therefore, the saturated
hydraulic conductivity value at the end of the load-sequence indicates whether the ore is
sufficiently permeable at the corresponding heap height to allow free drainage of the pregnant
leach solution (PLS) to the collection system. Besides qualifying the ability of a sample to support a
percolation leaching process, the ST helps to identify the maximum lift/heap height that the sample
might support. The load exerted by the ram is converted to an equivalent stacking height by taking
into account the bulk density of the ore and area of the test cell. The ram therefore simulates the
weight exerted by a mass of material on a unit are of the heap.

Hydrodynamic column test

The hydrodynamic column (HCT) relies on the density profile derived from the stacking test to
select the bulk density value required to properly represent the desired lift/heap height. The
column (15-cm or 20-cm diameter, 75 cm or 100 cm tall) is loaded with agglomerates at a uniform
target bulk density and is irrigated at incremental solution application rates. The experiment is run
such that an accurate solution balance is maintained throughout the duration of the test. The
parameters obtained from the HCT test include the hydraulic conductivity, air conductivity,
solution content (degree of saturation), pore pressure as a function of solution application rate, the
drain down curve, and total, micro- and macro-porosity. In other words, the HCT provides all the
necessary hydrodynamics required to simulate the solution and air movement through a heap.
From the more practical point of view, the HCT answers the key question for the design of a heap
the degree of saturation resulting from a given application rate. Experience shows that for copper
oxide leaching the maximum liquid saturation should be kept below 85% and ideally under 65% to
accommodate the natural variability associated with the heap stacking process. Leaching of a
sulphide ore requires that the degree of saturation remains below 60% to ensure that forced (or
natural) aeration can be properly implemented. By design a sample subjected to a HCT represents a
unit volume of ore located at the bottom of the lift the portion of the heap which experiences the
most stringent conditions since it is exposed to the maximum lithostatic load and the maximum
degree of liquid saturation along the heap profile. If the hydraulic performance of this portion of
the heap/lift is adequate, it is reasonable to conclude that the rest of the heap/lift profile will
perform just as well or better.

Metallurgical column tests

The results from the hydrodynamic characterization (ST and HCT) allow selection of the bulk
density and irrigation schedule that would likely result in the most favorable conditions for the
leaching process. In this particular case, the metallurgical column tests performed on the oxide ore
types were conducted at two lift heights (2 m and 4 m) to provide calibration and validation data
sets. The results from the metallurgical tests are used within the context of a numerical flow and
transport model to represent the mechanisms responsible for the observed copper recovery.

CHARACTERIZATION RESULTS
The stacking test results are summarized in terms of density and conductivity profiles (Figures 1
and 2) which indicate that all the head samples can be candidates for heap leaching over the lift
heights examined. The steeper slope of the density profiles from the leached oxide residues

377

compared with those from the head samples suggests that leaching of the oxide ore types has
noticeably reduced the physical integrity of these samples. The fresh rock and transition ore
samples do not show a marked increase in the slope of the bulk density curve, and a load up to
18 m was tested, at which point these samples had not yet reached the threshold value of
1.87 ton/m3 (corresponding to 30% porosity). The conductivity profiles of the leached oxide residues
(Figure 2) display a curvilinear behaviour with a steep downward trend, as opposed to the head
samples which are relatively linear. This downward trend suggests that the destruction of the
porous structure resulting from both decrepitation and concomitant compaction is such that it
would render the saturated hydraulic conductivity of most of the leached oxide residues smaller
than the recommended threshold conductivity value (1,000 L/h/m2) at heap heights of 4 m.
The shape of the density and conductivity profiles of the oxide, fresh rock and transition ores (fresh
and leached) show the variability in the expected physical and hydraulic behaviour of the various
ore types. Because a ST can be conducted in a matter of hours, this test procedure is an effective tool
in the selection of ore preparation techniques and alternative levels of crushing.

Figure 1 Bulk Density Profiles

Inspection of the particle size distributions of the metallurgical column feeds and residues
(Figure 3), indicate a substantial increase in fines during leaching of the oxide ore types. The
transition and fresh rock ore types, on the other hand, experienced a moderate increase in fines. The
change in particle size distribution during leaching helps to explain the changes observed in the
bulk density and conductivity profiles. An increase in the content of fines during leaching is
indicative of decrepitation, which induces a transient behaviour during the leach cycle resulting in
destruction of the porous structure, increase in bulk density and reduction of percolation capacity.

378

Figure 2 Saturated Hydraulic Conductivity Profiles

Figure 3 Particle Size Distributions

The results from selected hydrodynamic column tests are presented in Figure 4 in terms of the
degree of liquid saturation versus irrigation rate. It is noted that the resulting degree of saturation
for the target irrigation rate is the key design parameter for a percolation leaching process. Figure 4
indicates that leaching of a 20:80 blend of upper to lower bulk oxides with a bulk density of
1.65 ton/m3 (lift height of ~4 m) under a percolation leaching process on a free-standing heap at an
irrigation rate of 10 L/h/m2 results on marginal conditions (degree of saturation of 77%). Although

379

this saturation is still smaller than the recommended saturation threshold (85%) required for
geotechnical stability of the slopes of a heap; it is too close to accommodate the inherent
heterogeneity resulting from ore changes and process variability. Leaching of this ore blend would
have to be limited to extremely shallow lifts (< 2 m) and relatively small solution application rates
(<5 L/h/m2). Based on the results of the STs and this HCT, it is expected that leaching of upper bulk
oxide at larger blend proportions of 20% would result in a non-performing material. The lower bulk
oxide at a nominal application rate of 10 L/h/m2 resulted in about 62% saturation, indicating that
this material can support percolation leaching under a free-standing heap design. Fresh rock and
transition ore types at a bulk density equivalent to a 17 m lift operate at 50% saturation at an
application rate of 10 L/m2/h, indicating that this lift height can be properly supported.

Figure 4 Degree of saturation versus application rate

As indicated by the data summarized on Figures (1, 2 and 4), the ST and HCT provide practical
information that can be used to make informed decisions during the design of a leaching process.
Although not shown in this paper due to space limitations, the HCT procedure generates all the
parameters necessary to simulate solution and gas percolation through a heap. The next section
illustrates the use of a numerical flow code to develop an integrated analysis procedure for the
design of a heap leach operation.

380

NUMERICAL CONCEPTUALIZATION
For computational purposes the leaching of metals by percolation leaching can be regarded as a
special case of flow and reactive-transport under multi-phase (solution and air) flow conditions.
This process can be described by partial differential equations representing the movement of
solution and solutes through the heap profile, exchange across solid boundaries and diffusion
through water filled pores (at the macro and micro scale). Metal extraction has been simulated over
the last few decades with a variety of approaches from the shrinking core model (Cathles and
Schlitt, 1980) to fully integrated flow and geochemical model (Cooper and Dixon, 2006)
implemented in the context of computational fluid dynamics (CFD). Although a full geochemical
model attempts to represent the overall chemical reactions, experience shows that the biggest
hindrance to optimal leaching arises from natural complexity of solution flow and the lack of
understanding of this process at the relevant (laboratory and heap) scales.
Our work on design, trouble shooting and optimization of heap leach processes at a variety of
scales (from columns, to cribs, to test heap and at industrial heaps) shows that flow heterogeneity
is, in many cases, the reason for impaired performance of the leaching process. The work by Orr
(2002) clearly illustrates the effect of heterogeneity in the hydraulic response of the heap. Our work
on the chemical profiling of a leaching process (Guzman et al, 2007) shows that flow heterogeneity
strongly influences the metallurgical performance of a heap mainly due to non-uniform flow
velocity fields.
Further, physical characterization of ore-for-leach samples shows that as a minimum the effect of
bulk density on percolation capacity of solution and air along the heap profile should be
considered. At the core of the process of design of a leaching process is the issue of how to scale up
the results from the various laboratory-scale characterization activities. Our early work on
numerical modelling of heap hydrodynamics (Guzman et al, 2000) showed that to address this
issue three items are necessary:

A representative sample of the ore body prepared in a consistent manner for testing at
multiple scales. The issue of sample representativeness is a complex topic which could fill
several volumes. Suffice to say that additional development is required in this area.

A method to determine the key hydrodynamic parameters controlling solution and gas
flow through the heap material. As described on the previous paragraphs, the Stacking Test
and the Hydrodynamic Test are specifically designed to provide this information. In fact,
these tests are the only tools available to generate this information for a mine-for-leach
sample.

A numerical tool capable of incorporating these hydrodynamic parameters and assessing

the effect of alternative ore preparation practices (crush size, degree or quality of
agglomeration, density profile, etc.) and/or operational conditions (heap height, irrigation
schedule, whether conditions, etc.).

Early attempts to develop the necessary modelling capabilities included a modified version of the
mobile-immobile concept initially developed by Brusseau et al (1989). To our knowledge, the first
application of this tool in the context of heap leaching was used to develop an improved design of
the leaching of the Ripios Mina Sur Aglomerados (RMSA) for CODELCO during 1999. The

381

results of this effort showed that this concept provides a powerful approach to address several of
the key design questions for a heap leach operation. Srivastava and Brusseau (1996) further
extended this concept to incorporate the multi-process non-equilibrium (MNPE) transport. For the
sake of printing space, the theory behind the implementation of the MPNE concept is not presented
in this document. The following discussion focuses on the conceptual representation of metal
extraction using this method as well as the integration of this numerical tool with the
measurements from the hydrodynamic characterization.
Application of the MPNE concept to the multi-dimensional variable saturated flow model allows
full incorporation of all the hydrodynamic parameters measured for a given sample and provides
an approximation of the chemical reactions via a sorption-desorption model which relies on firstorder kinetics. The MPNE numerical analysis illustrated below represents non-equilibrium physical
and chemical solute transport.
In simple terms, physical non-equilibrium arises from differences in capillary forces between the
mobile (macro) and immobile (micro) solution domains within the porous medium. Chemically, the
non-equilibrium arises from chemical reactions, sorption and desorption processes which occur at
different rates on these two porosity domains. The current version of the model assumes isothermal
conditions in a two-phase (liquid and air) flow system. The MPNE concept is illustrated in Figure 5,
where the relevant transport parameters for the mobile and immobile domains are identified. The
tranport parameters are as follows:

Km and Kim represent the equilibirum (instantaneous) sorption between the mobile liquid
and solids and between the immobile liquid and the solids, respectively;

km and kim represent the sorption rate constant between the mobile liquid and solids and
between the immobile liquid and the solids, respectively; and

m is the mass transfer coefficient between the mobile and immobile water.

382

The fraction of the solid phase which is subject to instantaneous or rate limited sorption/desorption
is selected by the user and adds to the flexibility of the model.
As shown by the example presented below, this modelling approach is well suited to represent
leaching of oxide ore but will not be appropriate to represent the leaching of sulphide ores where
non-isothermal conditions are critical. Development directed to identify the best modelling tool to
represent such process is on-going.
The following paragraphs present an integrated procedure to incorporate the laboratory scale data
(hydrodynamics and metallurgical column results), and facilitate scale-up of these measurements to
predict the behaviour of a leaching process at an industrial scale.

HEAP LEACH DESIGN

As indicated above, fundamental to the design of a heap leach process is the scale-up from smallscale (laboratory) data to represent the leaching mechanisms at the industrial scale (heap). Our
work on this subject shows that in order to successfully complete this task, the laboratory
experiments need to meet some similarity criteria; as a minimum, the sample should be a
representative sample of the ore to be placed on the heap, the particle size distribution of the
samples should be as close as possible, and the physical conditions (level of agglomeration, bulk
density, and saturated hydraulic conductivity) should be selected to correspond to a particular
heap height.
The following procedure has been devised to ensure proper scale-up from laboratory data:
1.

Use the results from the Stacking Tests to select two heap heights and corresponding bulk
density values for the metallurgical tests.

2.

Conduct the metallurgical characterization at the two scales making sure that the
samples/tests satisfy the similarity criteria stated above.

3.

Using the results from the Stacking Tests and Hydrodynamic Column Tests, select the
necessary hydrodynamic properties to represent the conditions of the metallurgical tests.

4.

Interpret the results of the smaller metallurgical column using the MPNE (or equivalent)
model. This interpretation will produce a set of transport parameters for the metallurgical
response of the ore under the flow conditions imposed during the test. In other words, this
step is used to calibrate the model for the particular leaching application. If the
metallurgical tests were conducted such that the similarity criteria were satisfied, this
parameter set would directly apply to the larger column without further adjustment.

5.

Model the metallurgical results from the larger column using the proper hydraulic
parameters and the transport parameter set developed in Step 4.

6.

Use the resulting parameter set to simulate the hydrodynamic and metallurgical response
at the heap scale.

This procedure relies on our observation that the flow and transport of a percolation leaching
process are self-similar. In other words, as long as the experimental set-up satisfy the similarity
requirements, the hydrodynamic characterization and metallurgical response should be scalable.
The scale-up rules are controlled by the density and conductivity profiles as determined by the ST.

383

It is important to note that, a properly designed metallurgical program will need to consider other
factors in the scale up from column to the commercial scale operation: (a) The different grades will
have different leach times with higher grade giving longer times for the same recovery; (b) The
effect that changes in the heap bulk density have in changing the underlying diffusion controlled
transport rates; (c) The effect of heap height and irrigation rate on required leach cycle;
(d) Fundamental issues of higher heterogeneity in the commercial heap and the method of allowing
for these with estimation of longer resulting leach times. The current work provides guidance on
the hydrodynamic effects; which can now be fully accounted for; whilst the other parameters are
still subject to ongoing research to quantify their potential effects.
The process described in Steps 1 to 6 was applied using the hydrodynamic and metallurgical data
for the Lower Oxide samples. Figure 6 shows the results of the calibration step on the copper
extraction data generated from a metallurgical column test of the Lower Bulk Oxide ore at a 2-m
scale. The left frame in Figure 6 shows the measured and simulated PLS flux while the right frame
shows the laboratory measurements of PLS concentration and copper recovery as well as the
corresponding value calculated by the model. The hydraulic parameters used to simulate the flow
on this column are scaled-up using data generated from the corresponding HCT and the density
and conductivity profiles generated from the ST. Overall, the model is able to reproduce both the
flow behaviour and the metallurgical response for the 2-m test relatively well. The parameters
obtained from the calibration shown in Figure 6 together with the corresponding scaled-up
hydrodynamic parameters were used to calculate the metallurgical results of the 4-m column. The
results of this scale-up process are presented in Figure 7. Overall, the model calculated values fit
the experimental data relatively well. Future application of this procedure at larger scales would
provide the necessary basis to consolidate industrial application of the proposed scaling up
procedure.

Figure 6 Calibration run on Lower Bulk Oxide 2 m column

384

Figure 7 Scaled-up simulation on Lower Bulk Oxide 4 m column

CONCLUSIONS

A combination of metallurgical and hydrodynamic testing of ores for heap leaching, in

conjunction with modelling to simulate solution movement and mineral recovery, has been
applied to a mixed copper oxide/sulphide ore to derive design parameters such as stacking
height, and to characterise various ore types within the resource. Hydrodynamic test
procedures such as Stacking Tests and the Hydrodynamic Tests have been designed to
determine the key physical and hydraulic parameters controlling solution and gas flow through
the heap.

Scale-up from laboratory data to represent the leaching mechanisms at the industrial scale
requires the ore to be a representative sample of the ore to be placed on the heap, the particle
size distribution of the samples should be as close as possible, and the physical conditions (level
of agglomeration, bulk density, and saturated hydraulic conductivity) should be selected to
correspond to a particular heap height.

Four ore types (upper oxide, lower oxide, transition, fresh rock) showed varying degrees of
decrepitation during leaching, resulting in an increase in fines content and bulk density, and a
reduction in conductivity during the leach cycle. Especially the oxide ore types experienced a
marked increase in fines during leaching. Stacking test and hydrodynamic column test results
suggest that these ore types should be limited to a 4 m stacking height. The upper bulk oxide
fraction should only be leached in conjunction with other ore types, blended in proportions of
less than 20%. Transition and fresh rock ore types were less prone to physical degradation, and
can accommodate stacking heights of at least 20 m.

Results from the Stacking Tests were used to select two heap heights and corresponding bulk
density values for the metallurgical tests. Metallurgical characterization was performed at the
two scales, and metallurgical results from 2 m column test on the lower bulk oxide ore was
interpreted using the MPNE model producing transport parameters for the metallurgical
response of the ore under the flow conditions imposed during the test. The proposed scale-up

385

procedure was used to reproduce the copper drainage tenor and extraction curve for 2 m and
then to predict the recovery of the 4 m tall metallurgical columns. It is noted that the data
derived from the STs and HCTs eliminated the need for calibration of the hydrodynamic
parameters.

The resulting transport parameter set derived from the 2 m metallurgical test was used in
conjunction with the hydrodynamic parameters derived from the HCT and ST to validate the
scale-up approach by reproducing the PLS flow and copper concentration as well as the overall
copper recovery of the 4 m column. Overall, the fit obtained by the model with the calibrated
parameter set adequately reproduces the experimental data obtained for the 4 m column. It is
noted that modelling of the 4 m metallurgical column was completed without further
adjustment of the parameters derived from the smaller scale test, thereby demonstrating the
ability of the proposed procedure to scale-up between columns of different height.

Future work would look at integrating data from more disparate scales (2 m, 6 m and large
diameter columns cribs) to further validate our approach. Nonetheless, given that the
proposed approach is based on a solid theoretical basis to derive the hydrodynamic
parameters, it is our contention that the scale-up approach presented on this document could
be used in conjunction with other transport models (e.g., full geochemical speciation model) to
represent more complex leaching conditions as those observed in the context of leaching
primary sulphide ore. Application of the approach as presented here to oxide (copper and gold)
should be straight-forward.

ACKNOWLEDGEMENTS
This paper is published with the permission of Mintek, South Africa. Also, we appreciate the
comments from the anonymous reviewers and Randy Scheffel and Petrus van Staden for their
constructive comments and editorial suggestions.

REFERENCES
Brusseau, M.L., Jessup, R.E., Rao, P.S.C.: Modeling the transport of solutes influenced by multiprocess
nonequilibrium. Water Resour. Res. 25(9), 19711988 (1989)
Cathles, L.M., Schlitt, W., J., (1980) A model of the dump leaching process that incorporates oxygen balance, heat
balance, and two dimensional air convection, In Leaching and recovery of copper from as mined
materials, ed W.J. Schlitt, SME.AIME 9-27
Cooper, G. and Dixon, S. Modeling Chalcocite Leaching, Society of Mining Engineers Annual Meeting, Mar.27Mar29, 2006, St. Louis, MO.
Guzman, A., C.Fillippone, R.Srivastava, J.Chahbandour, I.Carrera (2000) Hydrodynamic Characterization and
Optimization of the Chuquicamata Ripios Mina Sur Aglomerados Heap Leach Project Part 2. Hydromet
2000, September 5-8, Tucson, Arizona, USA.
Guzman, A., R.E. Scheffel and S. Flaherty 2006. Geochemical Profiling of a sulfide leaching operation: A case study.
SME 2006 Spring Meeting. March 2006 St. Louis, Mo.
Guzman, A., Scheffel, R., Flaherty, S., (2008), The fundamentals of physical characterisation of ore for leach, R.S.
Shoemaker Symposium, Phoenix, Arizona

386

Guzman, A. Scheffel R.E., Ahlborn, G., Ramos, S., and Flaherty, S. (2007). Geochemical profiling of a sulfide
leaching operation: The rest of the story. Technical presentation at Copper 2007, Toronto, CA. Aug. 28
Van Genuchten, M.Th., (1980), A closed form equation for predicting the hydraulic conductivity of
unsaturated soils, Soil Sci. Soc. Am. J., 44, 892-898.
Books, R,H, and Corey, A.T., (1964), Hydraulic properties of porous media. Hydr. Paper 3, Colorado St. Univ., Fort,
Collins.
S. Orr and V. Vesselinov, (2002), Enhanced Heap Leaching Part 2 Applications, Mining Engineering, Oct, 3338.

387

Dynamic model of copper recovery

in heaps, Compaa Minera Doa
Ins de Collahuasi

Leachable ores from 2011 have experienced a decrease in soluble copper and
increase in residual copper (chalcopyrite) in Compaa Minera Doa Ins de
Collahuasi (CMDIC). This mineralogical change has motivated the development
of a new model to predict the copper cathode production for the 5 year period
of 2013-2017. Copper minerals in the feed to the leach pads are divided into

Eduardo Flores, Juan Pablo Garcs

and Christian Hu. Compaa Minera
Doa Ins de Collahuasi, Chile
Jess Casas. Process Consulting,
Chile

three classes based on leach kinetics: Slow kinetics representing chalcopyrite,

intermediate kinetics for chalcocite, and covellite and fast kinetics representative of chalcanthite and copper-oxides. Kinetic parameters for each individual
heap of the CMDIC plant were adjusted during the period of January 2011 to
May 2012, generating correlations of these parameters with sequential copper
grades and copper recovery results of individual heaps. A global model of the
plant was obtained using the kinetic parameters correlated by statistics. The
global model was used to estimate cathode production of the CMDIC plant,
by simulating the stacking, wetting, irrigation with intermediate solution
(ILS), irrigation with raffinate solution and decommissioning of the heaps,
thereby improving copper accounting. To estimate future copper production
a 5 year plan was generated using the current mining plan and its constitutive
geo-metallurgical units (GMUs). The model was calibrated for 2 GMU s ; one
corresponding to mixed ores from mine stockpiles and the other to oxide
ores from the mine. The results indicate that the plant model has reproduced
cathode productions with around 5% accuracy, yielding a cathode production
in the years 2011 and 2012 of 36,057 and 36,650t/y, respectively.

389

INTRODUCTION
The Mining Company Doa Ines de Collahuasi is conducting a review of its oxide plant operations.
The plant became operational in 1998 and was designed for processing resources of oxidized
copper ores, using crushing, agglomeration, leaching in permanent heaps, solvent extraction and
electrowinning to produce copper cathodes. (Castro, S., 2000, 2)
Currently, mixed ores are processed in a circuit designed to treat oxide ores as a result of gradual
depletion of the oxide ore resources. The stacking and leaching area has been declining after 15
years of operation and reaching the higher lifts of two current heaps. This has resulted in a decrease
in copper recovery and cathode production compared with the original plant design. (Garca, C.,
Garcs, J.P., Rojas, C., and Zrate, G. 2007)
In order to identify opportunities for process improvement and geometallurgical planning for the
next five years 2013-2017 CMDIC targeted the analysis of the feed to the current oxide plant and
developed a phenomenological model to simulate leach kinetics, this was used as inputs for a plant
production model comprising various heaps at different stages of the leach cycle
As a result of this work the predictive model of CMDIC for copper production was improved
including the leaching kinetics of operating and projected heaps according the medium-term mine
plan. The model has been developed in MS-Excel for easy manipulation and updating according to
requirements of planning area and hydrometallurgy operations of CMDIC.
The next steps relating to this development is to incorporate the kinetics results of metallurgical
testing running with low grade ore resources, which will be processed in the future by the oxides
plant, so to calibrate and improve the model for resources with low content of soluble copper in
acid solutions.

METHODOLOGY
Model of unitary heaps
Oxidized copper ores generally exhibit a high dissolution rate in acidic solutions, and both particle
size and sulphation degree achieved in the curing step the most influential variables in the copper
leaching rate.
The mechanism of leaching of copper ore mineral can be represented through the unreacted core
model (Levenspiel, O., 1972). According to this mixed kinetic model, first the reaction occurs on the
outer surface of the mineral particles and then moves inwards in the step to be copper leaching,
generating a porous region which corresponds to the unreacted zone and an inner core being
mineralogical species unreacted. Figure 1 is a schematic graph showing a particle leaching of
copper ore. (Bartlett, R.W. 1997; Casas, J.M. 1996; 7; Madsen, B.W., M.E. Wadsworth, M.E.& R.D.
Groves, 1975).

390

Figure 1 Schematic picture of copper leaching according to the unreacted core model

Below the formulation of the physical-mathematical model is presented (Casas, J.M., 1996). The
leaching of a copper ore may be represented by the following overall stoichiometry:
Ore(s) + a Reactant(aq) -----> b Cu2+(aq) + c Byproduct(aq) + d Residual-Solid
Where a, b, c and d are stoichiometric coefficients.
The ore leaching kinetics is represented according to the unreacted core model (Eq. 1), this model
can generate an equation to calculate the degree of copper conversion or recovery () implicitly,
depending on operating time (t) and reaction (TR) and diffusion (TD) characteristic times of leaching
process.
1

t = T R [ 1 - (1 - )3 ) + T D [ 3 - 2 - 3(1 - ) 3 ]

(1)

Following are the stoichiometries for the heap leaching of main copper species present in an
agglomerated mixed ore.
1) Dissolution of the oxide with fast kinetics representing the acid soluble copper generated in the
agglomeration and acid curing period and during heap irrigation:

Chalcanthite: CuSO *5H O

4

Tenorite: CuO

(S)

+ 2H

2+

(s)

+ H O ---> Cu
2

2+

(aq)

------> Cu

(aq)

(aq)

+ SO

+HO
2

2-

4 (aq)

+ 6H O
2

(Q1)

(Q2)

391

2) Dissolution of secondary sulphides (chalcocite and covellite) with medium kinetics, which
represents the cyanide soluble copper or the copper dissolved by ferric leaching of the mixed
ore throughout the irrigation cycle of the heap:

Chalcocite: Cu2S

Covellite: CuS

(S)

+ 2Fe

3+

2+

(aq)

(S)

(S)

2+

3+

+ 2Fe

------> CuS + Cu

(aq)

------> Cu

(aq)

+ 2Fe

+ 2Fe

(aq)

2+
(aq)

2+

(Q3)

(aq)

+ S

(Q4)

3) Dissolution of primary sulphides with slow kinetics, representative of residual copper,

generated by ferric leaching of chalcopyrite throughout the end period of the irrigation cycle of
the heap:

Chalcopyrite: CuFeS2

(S)

+ 4Fe

3+
(aq)

------> Cu

2+

2+

(aq)

+ 5Fe

(aq)

+ 2S

(Q5)

To estimate the kinetics of copper extraction in the heaps the kinetic model indicated by equation 1
was used and given three main species carrying copper, which are determined through the
sequential copper assay: acid soluble copper (chalcanthite/oxides), cyanide soluble copper
(chalcocite/covellite) and residual copper (chalcopyrite). The stoichiometric combination of the
individual kinetic of mineral species represents the total copper leaching kinetics in each individual
heap, which is calculated through the following Eq. 2. (Casas, J.M., 1996)
R i(t),(%)

[ 1(t )* F1 + 2(t) * F2 + 3(t) * F3 ]

(2)

were i(t) represents the extraction of copper mineral i-th and Fi represents the fraction of copper
mineral i-th in the mixed ore, and F1 + F2 + F3 = 1.
Figure 2 shows the leaching kinetics for copper minerals with different reactivities and for a mixed
copper ore processed in the conditions of CMDIC oxides plant.

392

100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0

Copper Recovery, (%)

Chalcanthite/Oxide

Ore, Heap

Chalcocite/Covellite

Chalcopyrite

25

50

75

100

125

150

175

200

Time, (days)
Figure 2 Graphical representation of the heap kinetic model considering 3 copper species according copper
sequential assay

Leaching model parameters of copper minerals were obtained by fitting the leaching model
represented by eq. 1 to experimental data of copper recovery for a series of individual heaps. A
macro programmed in MS-Excel was developed obtaining calculated recoveries from the implicit
eq. 1 and kinetics parameters of each mineral. Table 1 presents time constants obtained for all
heaps in operation within the period August 2011 to May 2012.
Table 1 Time constants of heap kinetic model (eq. 1)

TR, (days)

TD, (days)

Chalcanthite/Cu-oxide

50

Chalcocite/covellite

400

800

Chalcopyrite

10.000

10.000

Species

Model of heap leaching plant

The model of CMDIC oxides plant aims to calculate the cathode production from individual heap
productions. This model is constructed from metallurgical balance of copper fine, considering a
series of operating heaps. Figure 3 shows a schematic drawing of the leaching model developed for
the oxides plant of CMDIC.

393

Figure 3 Schematic Leaching Model of Copper Oxides Plant

The plant model was calibrated for the period August 2011 to May 2012 based on fine copper
generated daily by the series of heaps in operation. The model was calibrated by using correlated
individual factors F1, F2 and F3 with grades of total copper, acid soluble copper, residual copper and
leach ratio of each heap. From this balance, the monthly copper production was calculated and
compared to the amount of platted copper on a monthly basis, knowing the electrical power
consumption and current efficiency of EW plant, within the defined period of time. Plant model
parameters were adjusted to obtain results within 5% of error.
The model allowed the leaching kinetic projection of heaps to be stacked in 2012 and 2013, with the
inputs of copper grades and mineralogy.

RESULTS AND DISCUSSION

Adjustment of model factors from individual heaps
Values F1, F2 and F3 are obtained by manual adjustment of a set of plant kinetics. Table 2 shows a set
of values for 18 operating heaps.

394

Table 2 Constants F1, F2 and F3 of individual heap model with adjusted kinetics.

Parameter

Unit

Heaps
308

3 ant

19-7

18-7

20-7

21-7

22-7

25-7

24-7

Tonnage

kt

883.5

344.5

294.8

251.3

324.5

330.6

345.6

375.6

363.3

Copper

3056

2387

2212

1648

1989

2393

2284

2482

2584

Cu(tot.)

0.63

0.69

0.75

0.66

0.61

0.72

0.66

0.66

0.71

Cu(ac. sol.)

0.47

0.49

0.48

0.48

0.41

0.46

0.36

0.38

0.4

Cu(CN sol.)

0.11

0.13

0.16

0.11

0.12

0.16

0.18

0.17

0.2

Cu(res.)

0.05

0.08

0.11

0.07

0.08

0.1

0.12

0.11

0.11

kg/t

Acid in agglom.

15.8

14.9

12.1

11.7

11.2

9.6

7.9

8.0

7.6

F1

0.40

0.45

0.32

0.35

0.40

0.35

0.21

0.21

0.27

F2

0.60

0.25

0.56

0.53

0.57

0.53

0.66

0.61

0.62

F3

0.00

0.30

0.12

0.12

0.03

0.12

0.13

0.18

0.11

Table 2 Constants F1, F2 and F3 of individual heap model with adjusted kinetics (continue).

Parameter

Unit

Heaps

23-7

401

402

403

404

405

406

407

408

Tonnage

kt

258.1

335.7

423.4

458.7

445.4

528.3

508.3

230.0

460.3

Copper

1840

2423

3198

3611

3599

4597

4303

1623

4220

Cu(tot.)

0.71

0.72

0.76

0.79

0.81

0.87

0.85

0.71

0.92

Cu(ac. sol.)

0.45

0.48

0.49

0.45

0.46

0.49

0.47

0.41

0.45

Cu(CN sol.)

0.17

0.15

0.16

0.21

0.2

0.23

0.24

0.18

0.28

Cu(res.)

0.09

0.09

0.11

0.13

0.15

0.15

0.14

0.12

0.19

kg/t

6.8

9.3

13.3

13.8

14.5

14.0

12.6

12.8

12.3

F1

0.20

0.26

0.30

0.30

0.18

0.30

0.21

0.22

0.25

F2

0.69

0.68

0.68

0.50

0.70

0.35

0.60

0.50

0.55

F3

0.11

0.06

0.02

0.20

0.12

0.35

0.19

0.28

0.20

Acid in Agglom.

395

60
55
50

Copper Recovery, (%)

45
40
35
30
25
20
15
10
5
0
0

25

50

75

100

125

150

Alpha 3

Model

175

200

Time, (days)
Real

Alpha 1

Alpha 2

Figure 4 Model of Heap 22-P7 with fitted parameters F1, F2 and F3 adjusted to the real kinetics

Correlations of heap factor with copper ore grades

Below are the correlations obtained for factors or ratios of solubility F1 and F3. The factor F2 is
determined as the difference F2 = 1 - F1 - F3. These kinetic factors were used for the heaps in
operation within the period August 2011 to May 2012 and the plant model represented by
Equations 3 to 5.

F1 Factor, (fraction)

0,6
0,5
0,4
0,3
0,2
0,1
0,0
0,45

0,50

0,55

0,60

0,65

0,70

Solubility Ratio, Cu(ac. sol)/Cu(tot), (%)

Figure 5 Correlation for the Factor F1 (ratio of acid soluble copper)

396

0,75

0,50

F3 Factor, (fraction)

0,45
0,40
0,35
0,30
0,25
0,20
0,15
0,10
0,05
0,00
0,05

0,10

0,15

0,20

0,25

Solubility Ratio, Cu(res)/Cu(tot), (%)

Figure 6 Correlation for Factor F3 (ratio of residual copper).

General plant model

The general expression of the plant model is:
P t =

(3)

M L R t

were:
P(t) is the daily production of the plant.
Mi is the tonnage stacked in each individual i-th heap.
Li is the copper ore grade in each individual i-th heap.
R(t)i is the daily copper recovery of i-th heap.
The copper recovery of each individual heap for calculating the production according eq. 3 is given
by the contribution of each geometallurgical unit in the mining plan (see Table 3).
Table 3 Summary of Mining Plan for CMDIC Oxides Plant (GMU mixed-ore)
Parameter

Unit

Jul-12

Aug-12

Sep-12

Oct-12

Nov-12

Dic-12

Tonnage

t/month

26,733

26,449

26,458

26,449

26,458

26,449

Cu(tot.)

0.78

0.90

0.87

0.78

0.77

0.70

Cu(ac. sol.)

0.41

0.45

0.41

0.43

0.43

0.37

Cu(CN sol.)

0.21

0.20

0.16

0.21

0.17

0.19

Cu(res.)

0.16

0.25

0.30

0.14

0.16

0.13

397

Cathode projection applying the leaching model

Figures 7 and 8 present the results of model predictions compared with real copper cathodes
productions of CMDIC oxides plant, on a daily and monthly basis respectively.
The plant model was first calibrated using the plant data with a series of heaps in operation within
the period August 2011 to May 2012 (see comparisons in Fig. 7), and then the set of obtained kinetic
parameters were used to project the copper production in the second term of 2012.
These results indicate that the plant model has reproduced cathode productions with around 5%
accuracy, yielding a cathode production in the years 2011 and 2012 of 36,057 t/y and 36,650 t/y,
respectively.

Figure 7 Copper Production in Year 2012, plant and calculated values.

Main recommendations of this study are:

Incorporate this modeling development as a tool for planning and management in the
medium and short term in CMDIC.

Continue with permanent model improvements and refinement of kinetics parameters.

Increase the kinetics parameters and correlations database for different ores to be
processed.

Develop metallurgical testworks at different scales with future low grade ore resources to
be processed in the oxides plant, so to have the information of the kinetics response of each
ore to be processed in prior to their response on the industrial plant. This will allow better
calibrating the model of plant balance and anticipates changes in operating conditions such
as irrigation rate, leaching cycle (irrigation area), acid dosages and crushing particle sizes.

398

4000

Copper Cathodes, (t/month)

3500
3000
2500
2000
1500
1000
500
0
Jan-12

Feb-12

Real

3543

3269

Mar-12 Apr-12 May-12 Jun-12

3023

2660

2840

3040

Jul-12
2844

Aug-12 Sep-12
3122

3039

Oct-12
3340

Nov-12 Dec-12
2850

3080

Model

3514

3215

2747

2734

2825

2764

2988

3282

3280

3541

3003

2901

Figure 8 Copper Production in year 2012, real and calculated values.

CONCLUSIONS
The copper leaching kinetics in individual heaps of the CMDIC oxides plant can be described
through a model that weight over time (irrigation cycle) the leaching of three species, which can be
identified through sequential copper assay, corresponding to copper sulphate content/oxides
(Cu(ac. sol.)), chalcocite/covellite (Cu (CN sol.)) and chalcopyrite Cu (res.).
Daily copper production of the CMDIC oxides plant can be described through a model that solves
the copper balance from the sequence of operating heaps with their respective kinetic individual
leaching.
Model projections for copper cathodes production in 2012 was 36,794 t, which resulted close to real
production of 36,650 t in the CMDIC oxides plant for this year.
To estimate future copper plant production can use the balance of plant proposed and
incorporating the operationalized mining plan in order to have the sequencing of heaps in
operation, with their mixed ore loads (agglomerated tonnage stacked and mixing ratios for various
types of ores) and the representative composition of each heap based on a sequential copper assay.
From the models developed for individual heaps and for the plant (series of heaps) can generate
strategic information management of CMDIC oxides plant, so to optimize the future production
plan of copper cathodes and define the best conditions of copper leaching kinetics in each heap
under operation.

REFERENCES
Castro, S. (2000) Proyectos Hidrometalrgicos de Plantas de Oxides de Cobre. 51 Convencin Anual del
Instituto de Ingenieros de Minas de Chile. La Minera Contribuyendo al Desarrollo Social. 27-30
Septiembre, Santiago, Chile

399

Un Buen Comienzo para la Planta de xidos de Collahuasi, Segn diseo producir 50.000 ton/ao. Revista
Minera Chilena, N 208, Octubre 1998, pp. 41-45
Garca, C., Garcs, J.P., Rojas, C., y Zrate, G. (2007) Efecto sinrgico del tratamiento de mezclas de minerales
conteniendo copper wad y sulfuros secundarios Compaa Minera Doa Ins de Collahuasi, Iquique,
CHILE. In: HydroCopper 2007, 16-18 Mayo, 2007, Via del Mar, Chile
Levenspiel, O. (1972) Ingeniera de las Reacciones Qumica". 2 ed. 1981, Reverte S.A., Barcelona, Espaa, pp.
385-415
Bartlett, R.W. (1997) Metal extraction from ores by heap leaching, Metallurgical and Materials Transactions BProcess Metallurgy and Materials Processing Science, 28 (4), 529-545
Casas, J.M. (1996) Modelacin de Procesos de (Bio)Lixiviacin de Minerales de Cobre en Pilas y Botaderos.
Tesis de Doctorado en Ciencias de la Ingeniera Qumica. Universidad de Chile
Madsen, B.W., M.E. Wadsworth, M.E.& R.D. Groves (1975) Application of a Mixed Kinetics Model to the
Leaching of Low Grade Copper Sulfide Ores. Trans. Soc. Min. Metall. Petrol. Eng. AIME, Vol. 258, pp.
69-74.

400

Dynamic simulator of
ore preparation processes
and leaching

As one of the copper recovery processes, hydrometallurgy has gained momentum and significant impact due to its ability to process ore with average low
grades at competitive prices compared to other metallurgical processes.
Among hydrometallurgical processes the leaching step is emphasized, a
process that is characterized by its significant temporal and spatial scale of
operation, preventing proper monitoring, control and therefore a stable operation. This requires the use of computational tools to simulate the heap leaching

Francisco Reyes, Gabriel Tejeda,

Pablo Karelovic and Aldo Cipriano.
DICT UC S.A ., Chile
Miguel Herrera. Universidad Adolfo
Ibez, Chile
Fernando Romero, Solange Rojas
and Cristian Salgado. Anglo
American Chile

process to test in less time, smaller spaces and without the need to interfere
in plant operation.
Simulation tools have excelled in different areas due to the possibility of
modeling processes and predicting their behavior considering various situations. Several applications of heap leach simulation can be found. However,
few simulations include interaction with the upstream processes (crushing
and agglomeration). This paper presents the development and implementation
of an integrated dynamic simulator for the heap leach process, considering
also the crushing, transportation and agglomeration process, including the
assembly and disassembly of the heap. The dynamic simulator uses two-dimensional models of fluid transport, transport of solutes and dissolution of
copper in the leaching heap to analyze the effects produced in the variables
of interest, such as the consumption of acid in the copper concentration in
the PLS and leaching times. It also allows the study of various control strategies, including advanced process control, which was the original intention of
developing the model.

401

INTRODUCTION
As an alternative to pyrometallurgy, hydrometallurgy is being used in an increasing number of
copper mining operations extracting copper from average low grades ores. The ability to extract
mineral at low cost mainly lies in the use of the leaching process, in which the previously crushed
and agagglomerated ore is stacked to be irrigated with an acid solution to remove the copper.
Leaching has a high temporal and spatial extent, making it difficult to study and operate.
This paper presents the development of an integrated dynamic simulator of the crushing process,
transport, agglomeration and leaching. We describe the models used, calibration process and an
illustrative example. The dynamic simulator for hydrometallurgy (SDH, for its acronym in Spanish)
has been delivered as an intermediate result of a joint project with Anglo American Copper
Mantoverde Operation. This Operation is an open pit copper mine, crushing plant, a mechanized
transport and stacking of material, heap leaching (dynamic pad and static dumps), solvent
extraction and electrowinning processes for oxidized ores. The simulator is delivered as a prototype
to Mantoverde staff, adjusting its parameters and flowsheets to the existing process plant.
Dynamic simulators are tools, to facilitate the study of complex processes. This allows the study of
the interaction between process variables and the assessment of their impact on the final product.
Examples of simulators are listed in Nikkhah & Anderson (2001), which shows how each simulator
is useful in various stages of completion of the projects. In general the ore preparation process and
leaching have been studied separately, omitting the interactions that occur between processes.
Crushing and classification models are mainly based on the equations proposed by Hatch (1973),
see Equation 1, where F is the feed distribution, P is the product distribution, and B and C are
crushing and classification matrix. On the other hand, models of conveyor belts are based on first
principles, considering its operating and construction features
( ) = ( )(

( )

(1)

Following the scheme proposed by Wang et al. (2007), agglomeration process is modelled through
dynamic population balance equations, similar to Equation 2, where ni represent the number of
agglomerates of size i, N and Nin are the total number of agglomerates in and out of the drum, G,
Bi and Di represent the growth, birth and death rate of agglomerates, respectively. Fin and Fout
represents the total number flow rate in and out the drum, respectively.

= G n +

(2)

The leaching process has temporal and spatial scales of great extension. This involves complexities
in modelling and interconnection with the previous processes. The spatial scale necessitates the use
of partial equations models (Cariaga, Concha & Sepulveda, 2005), similar to Equations 3 and 4,
where is the nabla operator, which is use as a shorthand expression for the divergence and the
gradient. In Equation 2, represents the water content of the heap, D is the diffusion tensor and K
is the hydraulic conductivity and q is the Darcys velocity of the solution. In Equation 3 Ci stands

402

for the concentration of solute (copper or acid) in the leaching solution and represents the rate of
solute added or removed from the solution
= ! + "#$%
= (

& ) + '

(3)
(4)

MODEL DESCRIPTION
The simulator considers static and dynamic empirical and first principles models to describe in
general terms the behaviour of a real plant. Thus, its main objective is the evaluation of different
real time control strategies, rather than the determination of metallurgical parameters. The
following section briefly indicates the simulated variables in each process and how they are
connected.

Crushing and transport

In the crushing process the key variables are the granulometric distribution and the power
consumption, Table 1 shows the simulated variables. Models included in SDH can predict the
granulometric distribution in different points of the crushing process, allowing the study of the
effect of changes in operational variables, such as CSS, screen apertures, conveyor belts velocity, in
the final product that affects agglomeration.
Table 1 Input and output variables in crushing, screening and transport simulation
Variable

Unit

Input/Output

Feed

T/h

Input

Feed size distribution

mm

Input

Throughput

T/h

Output

Product size distribution

mm

Output

Power consumption

kW

Output

Crushing ratio

Output

Screen efficiency

Output

Stock levels

Output

Agglomeration
Table 2 show the input and output variables in the agglomeration process, key variables are the
granulometric distribution, the agglomerate humidity and the acid consumption. These variables
set the initial condition of the leaching process.

403

Table 2 Input and output variables in agglomeration process

Variable

Unit

Input/Output

Feed

T/h

Input

Feed size distribution

mm

Input

Ore humidity

Input

Acid feed

T/h

Input

Water feed

T/h

Input

Drum angular velocity

RPM

Input

Throughput

T/h

Output

Product P80

mm

Output

Power consumption

kW

Output

Product humidity

Output

Internal load

Output

Acid consumption

Kg/T

Output

Heap leaching
An unsolved problem is the interconnection between the processes of ore preparation and
operation of the heap leaching. The simulator uses assembling and dismantling algorithms for the
operation of dynamic heap leaching of oxide coppers, which allows the output variables of the
agglomeration process to take effect on the construction and operation of the heaps. The leaching
process simulation modelling allows the following operation variables, as shown in Table 3.
Table 3 Input and output variables in leaching process

404

Variable

Unit

Input/Output

Feed size distribution

mm

Input

ILS irrigation rate

L/h/m2

Input

ILS copper concentration

gr/L

Input & Output

ILS acid concentration

gr/L

Input
2

Raffinate irrigation rate

L/h/m

Input

Raffinate copper concentration

gr/L

Input

Raffinate acid concentration

gr/L

Output

Mean humidity in heap

Output

PLS flow

m /h

Output

PLS copper concentration

gr/L

Output

PLS acid concentration

gr/L

Output

Copper recovery

Output

Leaching rate

m /T
3

Output

SIMULATOR PROGRAMMING AND IMPLEMENTATION

SDH is presented as an executable compiled program for Windows 32 and 64 bits, allowing its
use in a broad spectrum of computers. Figure 1 shows the splash screen of the simulator, which
shows the work of DICTUC and the financial support from CORFO program InnovaChile.

Figure 1 Splash Screen of the simulator

Calibration
As a prototype designed and adjusted specifically for Mantoverde Operation the simulator must be
calibrated. This process is carried out by measuring operational data delivered by the staff of Anglo
American and its subsequent processing by DICTUC. Procedures, which in the case of the leaching
process are similar to those described in Munoz & Vauclin Rengifo (1997). Table 4 shows the
parameters that must be set for operation of SDH.
Table 4 Parameter to be calibrated
Parameter

Unit

Measure/Calculated

Crusher CSS

mm

Measure

Screen aperture

mm

Measure

Coalescence kernel

1/s

Calculated

Ore density

kg/m3

Calculated

Ore humidity

Calculated

Agglomerate porosity

m3/m3

Calculated

van Genuchen parameters

Calculated

Hydraulic saturation

m/h

Calculated

Copper extraction

m3/kg/s

Calculated

Acid consumption

1/s

Calculated

405

Features and possible uses

SDH is simple and easy to use, as been calibrated and designed to Mantoverde operation; the user
interface is clean with no excess of calibration parameters entries. Any configuration occurs by the
HMI shown in Figure 2 a). SDH also allows individual or integrated simulation of crushing,
agglomeration and leaching process. For this, it has a connection to MS Excel that allows the user
to enter input variables into the simulation and get the simulation results for further processing. It
also has a specialized interface, Figure 2 b), to visualize simulation results, allowing the user to
analyse cross effects between variables.

a)

b)

Figure 2 Main window of SDH. a) Navigation is made by process buttons (1), which changes the process show
in (2). Parameters are entered by pressing button (3), and simulation is controlled by buttons Start and
Stop (4). b) Plotting window of SDH. When simulation is done a new window appears so the user can plot
variables against time. Two plots can be displayed at the same time so relations between variables of different
processes can be seen

SDH, by having dynamic simulation models can be used to assess performance of different
operation strategies, for example, different irrigation cycle patterns in leaching. Also can be used as
a basis for the development of control strategies or operator training.

EXAMPLE OF USE
This section delivers an illustrative example of one of the uses of the simulator. We present the
study of the leaching cycle and how SDH can help to predict the behaviour of the leaching process.
First we simulate a base cycle of 160 days, including an ILS cycle and a Raffinate irrigation cycle.
The recovery of this cycle is plotted in Figure 3 a), where it can be seen that the recovery shows a
steady behaviour between day 40 and 120. The final recovery is 84.49 [%] and the leaching rate is

406

4.0 [m3/T]. A modified cycle is tested under simulation, also including an ILS and Raffinate
irrigation, but with a reduction of 20 days in the ILS irrigation. Bringing forward the Raffinate
irrigation allows energy efficiency, avoiding the energy consumption of the irrigation bombs in the
steady behaviour of the recovery. Figure 3 b) shows that the recovery of the new cycle, with a
similar final recovery (84.50 [%]), and the leaching rate is reduced to 3.3 [m3/T].
a)

b)

Figure 3 Comparison between two different leaching cycles. a) Recovery and leaching rate of the base cycle,
recovery shows a steady behaviour between the 1000 and 2500 hours of operation. b) Recovery and leaching
rate of the new cycle, leaching rate is reduced 25% and final recovery has not been reduced

CONCLUSION AND FUTURE RESEARCH

In this paper we have presented the development of an integrated simulator for the ore preparation
process and further leaching. We describe the simulated variables and parameters adjusted. This
simulator consists of an intermediate result of a project between Anglo American and DICTUC
which seeks to optimize the preparation process ore from the energy standpoint.
The use of simulation tools allows the study of the relationships between process variables,
generating a better understanding of the process and behaviour. Simulation example presented
shows that simulation can be used as a predictive tool for the study of new leaching cycle that
achieve greater copper recoveries and lower energy consumption.
Future work includes the use of SDH and its algorithms as a basis for a decision support system
and a global control system that seeks to maximize the copper production minimizing energy
consumption. We also are looking to expand the scope of the simulator processes, including solvent
extraction and electrowinning.

ACKNOWLEDGEMENTS
The authors wish to thank CORFO for support through Project 11IDL2-10443 Control system for
energy optimization in hydrometallurgy. DICTUC also thanks Anglo American for it support and
authorization to publish this paper.

407

REFERENCES
Cariaga, E., Concha, F., Saplveda, M. (2005) Flow through porous media with applications to heap leaching
of copper ores, Chemical Engineering Science, vol. 111, pp. 151165.
Hatch, C. (1973). Digital Simulation of a Crushing Plant. MSc. Thesis, University of British Columbia, Canada.
Muoz, J.F., Rengifo, P., Vauclin. M., (1997) Acid leaching of copper in a saturated porous material: Parameter
identification and experimental validation of a two-dimensional transport model, Journal of
Contaminant Hydrology, vol 27, pp. 124.
Nikkhah, K., Anderson, A. (2001) Role of simulation software in design and operation of metallurgical plants:
a case study, Proceedings SME Annual Meeting, Denver Colorado.
Wang, F.Y., Cameron, I.T. (2007) A multi-form modelling approach to the dynamics and control of drum
granulation processes, Journal of Powder Technology, vol. 179, pp. 211.

408

DROP Project: Leaching process

optimization for reducing water
consumption in copper mining

The growing water shortage in the north of Chile and the projected growth
of the mining sector has forced the initiation of efforts to optimize the use
of water resources and to minimize water losses. The DROP Initiative, devel-

Ulrike Broschek, Jorge Lobos, Jorge

Cornejo, Luis Bravo, Josu Lagos
and Karien Volker. Fundacin Chile

oped by Fundacin Chile at a mining company, aimed to reduce fresh water

consumption in the leaching process in the framework of the World Class
Cluster Suppliers Program for the global mining industry. The initiative
sought the optimization of the sprinkler irrigation process employed in the
project considering that water losses occurred through: (i) Evaporation during
irrigation; (ii) wind drag of water and; (iii) evaporation of water from the cell
surface during the irrigation resting stages. Optimization of water usage and
the management of water loss have a number of variables and complexities
that must be carefully considered when designing improvements or changes
in the production line. This may involve incremental operational costs, and
may increase complexity in the operation and affect efficiency in the recovery
of minerals. Project DROP was developed to generate an adaptive design to
suit the requirements regarding the intrinsic characteristics of the operation.
A theoretical approach was carried out to evaluate operational changes and
its impacts on water losses. During the second stage, field work was carried
out using these results. The results of the field tests performed on the leach
pads showed that the solution generated can meet the goal of reducing water
losses defined in the beginning of the project of 5L/s. Reductions above 8L/s
where achieved through changes in the sprinkler configuration, the droplet
size and/ or shape of the irrigation. Currently, Fundacin Chile has a set of
variables that can be implemented to reduce important water losses in other
mining sites that also use sprinkler irrigation.

409

INTRODUCTION
Copper mining takes place mainly on the macro zone of northern Chile, where the availability of
water resources is limited mainly due to low rainfall. This availability of water is limited due to
first, the growing increase in demand from mining, and second, the changes in the water cycle
caused by global climate change, particularly evident in recent years (Garredeau 2011 ). As a result
of this situation, the mining projects in the area, both old and new, are affected by this shortage
which has led to the urgent need to develop optimization processes in current water consumption
by mining (Cochilco, 2010 ). This is the vision of project DROP, an initiative led by the Water and
Environment Department of Fundacion Chile alongside a mining company of northern Chile. This
initiative is part of the Cluster Programme World Class Provider for the global mining industry,
which seeks to encourage suppliers to develop greater technology and management in order to
address the challenges that the mining industry has throughout the country and the world
(Fundacin Chile, 2012).
Copper mining in Chile operates in two modes, holding constant the extraction stages, depending
on whether the copper ore is sulfide or oxidized state. If this goes through a process, sulphide
milling, flotation, smelting and electrorefining. If oxidized, leaching, solvent extraction and
electrowinning are the way. In the mining company, through this latter process operates under
dynamic heaps, formed by mixtures of oxide-type materials and copper sulfides in a proportion
between forty to sixty percent of oxides. The project began by identifying the losses in the leaching
process, specifically in the wetting of the heaps under nebulization so established the goal to
develop and evaluate a nebulization system that would reduce water consumption by 5 [L / s] in
the leaching process. This improvement should be carried out keeping current copper extraction
rates.

Operational and environmental critical factors in the process of nebulization

For the leaching process to develop properly, one must consider a number of operational and
environmental critical factors in the process of nebulization. These can be grouped into three
categories: operational factors of process, design factors specific to the nebulization step, and
environmental factors or also known as external factors.
The first category includes the permeability of the material to be leached and the efficiency of
contact between the material and the leaching solution. The permeability of the material to be
leached, and refining the solution must pass through the entire bed of material and takes contact
with it. The effectiveness of contact between the material and the leaching liquid depends on the
surface area, porosity, particle size, roughness of the surfaces (Mellado, Galvez & Cisternas 2011)
amongst other factors. During the spraying the solution, refinement is applied as spray. The correct
operation prevents water loss by crusting on the surface, avoiding surficial ponding and formation
of preferential channel by promoting efficient leaching of ore on the heap. (Avendao, 2004).
The second group of operational factors, relating to the nebulization design, can be named to
irrigated frame (nebulizers configuration), the height of the nozzle, pressure and nozzle diameter.
The frame affects the correct overlap of nebulizers. If nebulizeres are more spaced, the dispersion of
the cloud caused by the wind is more influential, avoiding a well wetting process. The height of the
nozzle determines the residence time of the drop in the air, being higher the losses by drag and

410

evaporation if the residence time is bigger. The pressure and the nozzle diameter determining the
droplet size; smaller diameter, more evaporation.
Finally, in the third category, the environmental factors that promote the loss of water can be
mentioned: a) the temperature, b) air humidity, c) solar radiation, and d) wind.
In this research, efforts have focused on optimizing the irrigation matrix of leaching process. To
meet this goal, we studied the water loss in the current configuration, and furthermore each design
factors was studied (eg nozzle height, irrigation frame, drop size and use of cover, among others)
individually. Then we proceeded to optimize these design factors in order to reduce water losses
and finally quantified the effect on water loss at the entire process and also at the level of each
design parameter.

METHODOLOGY
Experimental design
Tests for water loss were performed in-situ in the leach pads on mining company. This was
achieved on a quadrant of twenty meters long and eighty meters wide on a leach pad. The quadrant
was divided into 8-cell study of approximately ten by ten meters. Each unit contained seven
nebulizers mounted on triangular array distribution as used by the mining, see Figure 1. One of
them was coated by a cover and another only for side walls.

Figure 1 Configuring nebulizers in study cells at heap pad and use of cover

The evaporation and precipitation rate measurement was made using forty-eight catch cans around
center nebulizer. Pressure and flow rate was monitored by a pressure gauge and a flowmeter,
respectively, both located in the input line to each cell. The evaporation loss value was calculated as
the difference between the rate of applied irrigation and water collected at the set of catch cans.
Tests were conducted in three different time zones: 9:30-11:00 am, 11:30 am-13: 00 pm and 15:00
pm-16: 30 pm each day in order to evaluate the effect of environmental variables water loss. For
statistical purposes, each test was repeated 3 times
In addition to the sampling of environmental variables a weather station was mounted on field at a
fixed point of the leaching plant to generate wind roses to determinate wind behavior and its
influence on the losses and water savings achieved in developed tests.

411

Optimizing process design parameters nebulization

According to the operational and environmental variables a test matrix was used for evaluating its
effects on potential water savings. See Table 1.
Table 1 Potential measurements to reduce water consumption
Potential solutions

Unit

Test range

Nozzle height modification

Cm

40 - 20

Frame modification

1,4 - 1,8

Cover system

--

Use/No use

Drop size

190-250

Note that the current system configuration nebulization is 1.8 m separation between nebulizer, 40
cm nozzle height and no use cover. It was also considered a mixture of measures to verify synergy
between each other by means of an experiments factorial design 24.
Finally, to study the droplet size a work station was mounted inside the cell with cover. It
consisted of a nebulizer located at forty centimeters height from the surface and water sensitive
cards along the wet radius in a line. These cards can capture various droplet sizes by changing
from yellow to blue upon contact with the water, registering droplet size with the fingerprint that
is marked on the card. After the test, the cards were scanned and processed in the Steinmaster
software that translates every mark on droplet size. Droplet size was measured for the current
nebulizer and another alternative nebulizer, similar on operational parameters, to quantify the
difference in droplet sizes.

RESULTS AND DISCUSSION

Base analysis
The results presented below are for the period of time between 15:30 to 17:00 hrs, Zone 3, where it
was expected the most favorable condition for high evaporation.

Framework irrigation
The effect of frame did not demonstrate influence on the reduction of the losses by evaporation and
drag. Its important to say that although a reduction was not shown in evaporation losses, there
was a contribution of this measure to the global reduction, considering that if the frame is reduced
more nebulizers are needed in the same area, which means a consequent reduction of the pressure
and flow rate to keep irrigation rate. Taking into account this effect, the overall reduction is
positive, which quantifies potential savings. Those savings compared to the rest of the measures are
discussed in the following section.

412

Nozzle Height

Evaporation + drag losses (%)

The nozzle height did not show a favorable trend to reduce evaporation losses and drag. The
increase may be explained by wind turbulence on the areas close to the surface, the drops would
remain longer in the air.

40%
30%
20%
10%
0%
Height

Frame
irrigation

Nozzle

Current configuration

Cover
Modification

Figure 2 Effect of the proposed measures in the total losses

Covers
The cover isolates in some way the nebulization process from the environment, which is an
unfavorable element in terms of evaporation. This can be seen in Figure 2. Wind speeds
measurements within the cell which showed that the wind is reduced by 75% with respect to the
wind outside, confirming the importance of isolating the wind on nebulization and its influence on
the process.

Droplet Size
In Figure 1, the droplet size shows a difference of almost 10% between the nebulizer and the
alternating current due to the effect of the drop in evaporation. The small drop of alternative
nebulizer is more likely to drag by their weight and evaporation by high surface area. Although the
differences in size between drops is 60 micrometers, the trend is strong, showing a significant
sensitivity to this parameter.

Environmental variables
The weather station worked during most of the study, recording data every 1 minute. This data was
then processed and the values were reported per hour. This allowed this information to be checked
with time zones and see how this variable influenced in terms of water losses. The main parameters
are shown in Table 2.

413

Table 2 Environmental variables and ranges measures on field

Topic

Unit

Zone 1

Zone 2

Zone 3

Wind

m/s

1,3

2,5

3,6

Relative Humidity

20

14

14

Radiation

W/m2

20

530

830

As seen in Table 2, the zone 3 is the most unfavorable for nebulization. There is a low relative
humidity and wind speed and radiation are in highest point, which is consistent with previous
data. Is important to say the results of evaporation losses from the other two zones were not
favorable, because the measures did not show a positive trend in terms of reducing water
consumption.

Integrated optimization variables

After basis analysis, a new analysis was introduced that integrated operational variables with heap
nebulizers results. This analysis was essential to relate the loss by evaporation and drag at a flow
rate that would allow the comparison with the assigned goal flow rate. This analysis was integrated
with the total amount of nebulizers, the flow emitted by each of them, the total nebulization area,
the wetting ramp established, among other variables. Considering this analysis in reducing total
water consumption for the four measures described above, the order taking into account the
greatest reduction is:

frame irrigation > cover > height > drop size

Considering the reduction goal of 5 L / s, four measures were able to overcome it and to reach 8 L/s.
Of these, 3 are a combination of measures:
cover+frame irrigation > frame irrigation+cover+height > frame irrigation > Frame irrigation+height
If the criteria are the costs associated with each measure, taking into account the higher cost is:

Cover+height+frame irrigation > walls+height+frame irrigation > frame irrigation+height > frame irrigation

Taking account the two previous criteria, frame irrigation and height are seen as cost-efficient
options to reduce water losses.

CONCLUSIONS
From this analysis, the alternatives to change frame, with and without the use of sidewalls, are the
most interesting, with significant potential water savings.

414

With the end of this project, Fundacin Chile articulated a methodology for nebulization irrigation
study considering the factors in this process comprehensively. This provides insights for
appropriate solutions to several operational realities, aiming at reducing water losses.

REFERENCES
Avendano, C. (2004). Revision de la Lixiviacion en Pilas de Minerales de Cobre. Paper presented on LX Users
Conference Chile, La Serena, Chile, Junio 2004
Cochilco (2010), Consumo de agua en la minera del cobre, Direccin de estudios y polticas pblicas viewed 4
march 2013, <http://www.cochilco.cl/productos/pdf/2011/INFORME_AGUA_2010.pdf>
Fundacin Chile (2012). Gua programa proveedores de clase mundial, Innovum Centro de innovacin en capital
humano, viewed 1 march 2013 <
http://www.fundacionchile.com/archivos/Gu_a_Programa_Proveedores_de_Clase_Mundial.pdf>
Garreaud, R. ( 2011) Cambio Climtico: Bases fsicas e impactos en Chile. Revista Tierra Adento INIA-Chile, no.
93, pp 13-19
Hooghart, J. C. (1987). Evaporation and weather, 44th edn, Technical Meeting, Netherlands.
Mellado, M., Galvez, E., and Cisternas, L. (2011).On the optimization of flow rates on copper
heapleachingoperations. International Journal of Mineral Processing. vol. 101 no. 14, pp 75-80.
Nin, R.A. (2008). Tecnologia del riego por aspersion estacionario. Calibracin y validacin de un modelo de
simulacin. Universidad de Castilla-La Mancha, Espana
Tarjuelo, J.M. (2005). El riego por aspersion y su tecnologa, 3a edn, Mundiprensa,Espana.
Valiantzas, J. D. (2006). Simplified versions for the Penman evaporation equation using routine weather data.
Journal of Hydrology, vol. 331, pp 690 702

415

Rheometallurgical process risk

mitigation through engineering
data generation

The concept of rheometallurgy is relatively novel as it addresses process risk

mitigation through generation of engineering data from the incipient stages

Alex Mezei and Mike Ashbury. SGS

Mineral Services, Canada

of testwork and process development throughout the completion of the full

feasibility study. At the same time, it is a well proven approach, being based
on a core methodology that quantifies the interaction between the metallurgical process parameters and the flow behavior of slurries and pulps, which is
determinative of true commercial performance of key unit operations. This
paper presents selected data from several commercial projects, with an
emphasis on hydrometallurgical applications involving difficult ores that
generally exhibit handling, flow and separation problems. It was found that
the main root cause of the process risk primarily resides in disconnects
between the optimized metallurgical parameters and the physical conditions
created as a result. The process of actual generation of key engineering data,
whilst monitoring and mitigating technical process risk is presented in detail,
including relevant case studies. Typical examples include scenarios whereby
significant errors of equipment sizing, ranging from +67 to -44% were flagged
and corrected. Furthermore, an example of ionic strength effect is presented,
illustrating the rheometallurgical optimization of an autoclave pressure oxidation process within the 77 to 95% extraction range as a function of key rheometallurgical parameters, including the discharge underflow Critical Solids
Density (CSD), ranging from 60 wt to 28% wt solids content, respectively for
that extraction range. Accordingly, the risk-minimized, optimum target extraction of about 92% was determined as predictable of a continuously operating
commercial plant.

417

INTRODUCTION AND KEY DEFINITONS

Key rheometallurgical concepts and definitions were presented in a previous paper (Mezei et. al,
2011). Accordingly, rheometallurgy was defined as the interdisciplinary area engaged in
investigating the interaction between the process parameters and flow behaviour of metallurgical
slurries and pulps, which is determinative of the true commercial performance of key metallurgical
unit operations. Subsequently, it is the opinion of the authors that pulp flow behaviour is the main
factor determinative of the overall process risk profile. Generally, this reality sets in at advanced
stages of process development rendering, more often than not, the project to be at risk due to
technical factors that are entirely preventable. It is the objective of this paper to further contribute
to the rheometallurgical database by providing a selection of examples based on commercial
testwork focusing primarily on hydrometallurgical process development. A general overview of the
known rheometallurgical risk-factors will be attempted, with particular emphasis on
hydrometallurgical process environments. A succinct summary of terms and definitions used in
this paper is provided below. Accordingly, the majority of the mineral and metallurgical pulp
streams can be treated as particulate fluids (Healy et. al., 1993, Mezei et. al, 2011 and 2002) and as
such, they can be reliably characterized in terms of their flow behaviour (also known as rheological
behaviour) based on the Bingham model, which is characterized by yield stress. The Critical Solids
Density (CSD) is an empiric rheometallurgical parameter defined as the solids content range, above
which a small increase of the solids content causes a significant decrease of its yield stress, or
flowability. The CSD value is predictive of the maximum underflow solids content achievable in a
commercial thickener, as well as of the underflow solids content and pumpability ranges achievable
in practice and with reasonable friction pressure losses for an economically feasible operation.
Actual Specific Gravity (ASG) is the non-dimensional equivalent of the real density of the particles
(i.e. particulate flow units) in a certain pulp under real process conditions. ASG often deviates from
the dry solids specific gravity (SG). The ASG/SG ratio defines the slurry inter particle interaction
coefficient . Real Data Modelling (RDM) is a commercial testwork and flowsheet development
approach that relies on defining overall (i.e. chemical and physical) metallurgical systems
behaviour based exclusively on test data, in conjunction with subsequent verification stages, using
confirmatory test results produced through bench and pilot scale exercises. RDM encompasses
several proprietary experimental and scale-up techniques including rheological modelling
software, metallurgical balance simulations, equipment sizing calculations, etc.
Key process rheometallurgy domains can be defined as shown in Figure 1, which illustrates the
stability and flow-profiles characteristic for the various unit operations in a commercially operated
processing plant. Common to these domains is the continuous flow linking them. The degree of
stability is determined by several factors, which includes particle size, density, solution properties,
etc., whereas flow behaviour at a certain temperature is determined primarily by solids density and
the rate of shearing.

418

Pumping

Yield stress, Pa

Mixing

Ultra-stable with near zero

flow and separation:
paste, non-segregated
tailings, filter cakes
Stable: plug-flow with
marginal or zero
separation: ground pulps

Separation

CSD ~ Near stable: settlingthickening predominant at low

shear. flow decreases sharply:
underflow.

Unstable:
segregated ,
"sandingout": nonCIL, coarse,
fibrous.

Meta-stable B:
Medium shear
flow - lesscommon mass
transfer:
adsorption and
absorption: CIP

Meta-stable A: High
shear flow - common
mass transfer: mixing
(leach/float), pumping

Solids content, %

Figure 1 Key process rheometallurgy domains

For example, a gold plant feed is ground and then the stable (non-settling) ground-discharge pulp
is directed to cyanide-leaching, which often incorporates a repulp thickening stage. Only thickened
pulps at solids contents nearing their CSD can be raked and pumped effectively. Data suggest that
generally, leaching kinetics tend to be less sensitive to yield stress (thus solids content) as long as
the pulp is stable at shear rates that impact an intermediate flow regime, generally not exceeding
~150 s-1. This CSD-defined flow-behaviour can be assimilated to the meta-stable domain A. On
the other hand, carbon adsorption kinetics is highly dependent on yield stress (Mezei, 2003,
Fleming et. al, 2010) responding best to a well-defined minimum value that still ensures a stable
pulp at a reasonably-low shearing regime laminar flow generally not exceeding ~50 s-1. Practicesuggested yield stress values for this meta-stable domain B range from about 2 Pa to about 10 Pa.
The CIP circuit discharge is thickened again, the water is recycled upstream and the underflow
directed to the tailings pond or paste fabrication, completing the rheometallurgical cycle. Of note is
that the low yield stress domain shown at the bottom left corner of Figure 1 renders the process
pulp unstable (i.e. fast settling), case in which, in this example carbon-in-pulp (CIP) processing is
precluded as a process option, rendering the ore-feed more suitable for a Merrill Crowe inspired
process configuration. Unstable pulps and slurries can be kept in suspension by applying a high
rate of shear, generally exceeding ~150 s-1 in order to maintain a homogeneous particulate system,
commonly impacting a turbulent flow profile. The methodology employed in establishing the
rheometallurgical profiles along with examples typical to other processes are provided further
herein.

419

METHODOLOGY
Test procedures
The rheometallurgical process risk analysis relies on rheologically-validated engineering data
which is generated on commonly applied liquid-solid separation unit operations. The experimental
techniques are quite specific and they are based on standard safety and operating procedures
specifically designed to meet the rigors of commercial metallurgical testwork. The first part of the
testwork matrix investigates the settling-thickening response and rheological (i.e. flow) behaviour
of the resulting underflow test samples produced under optimized conditions. The test samples
(i.e. potential thickener feeds) are retrieved from discharge streams such as grinding, flotation,
leaching, precipitation, etc. which are generated by bench and/or pilot plant unit operations. The
optimum settling-thickening conditions are determined by a sequence of static and dynamic tests.
The rheology study, consisting of Concentric Cylinder Rotational Viscometry (CCRV) is carried
out at variable shear rate ranging from 0 to 600 s-1 which brackets a 300 s-1 constant shear sequence.
Vane viscometry, at a very low rate of shearing (<1 s-1), is conducted when encountering ultra-stable
pulps; this is in addition to or in lieu of CCRV, as applicable. Following rheological validation of
the optimized conditions (as to ensure that the solids contents are consistent to those predicted to
occur in the commercial equipment), the second part of the testwork investigates vacuum and
pressure filtration responses of the same underflow samples. Hydrometallurgical discharge
samples are further subjected to a detailed washing study which accurately emulates the commonly
applied (i.e. industrially applied) unit operations such as counter-current decantation and countercurrent filtration-washing, i.e. CCD and CCF, respectively. The resulting data are used to
determine the optimum and practical washing conditions through Real Data Modelling (RDM),
only in the absence of solubility related and post-precipitation related anomalies. Otherwise,
additional testing is carried out in order to quantify the effect of these interferences on both
washing and actual separation efficiencies. The updated results are fed into the RDM and the
resulting output is validated as design criteria that can be commercially applied without running
the risk of disrupting the continuous process.

RESULTS AND DISCUSSION

Process risk root causes
Data from several metallurgical projects which span over 15 years from the mid-nineties to date,
unequivocally suggest that the root cause of the process risk primarily resides in disconnects
between the optimized metallurgical parameters and the physical conditions created as a result,
which leads to severe ramp-up and operational issues. The parameters mentioned commonly
evolve around liberation and as such tend to consistently include particle size and process
chemistry. With increased fineness leading to attractive grade-recovery flotation performance as
well as rapid leaching kinetics, the temptation for over-grinding is largely predominant. The main
unintended consequence, more often than not overlooked in the majority of these situations, is the
markedly adverse effect on flow and hence separation. If undetected, the consequences could range
from a severely underperforming commercial plant through to a complete ramp-up failure due to

420

underperforming equipment and multiple blockages (i.e. tanks, rakes, pumps, piping, etc.). In this
context, data published in a previous work (Mezei et. al, 2011) indicated that the specific thickener
unit area requirements were about double vs. the end-user assumed target in the case of a 36
micron reground sample. Moreover, further re-grinding to 16 microns required a tenfold increase of
the thickener unit area. The effects on the system ancillaries (i.e. rakes, pumps, piping) were equally
drastic; on this basis, a maximum potential design error of about 1000% appeared to be clearly
quantifiable at this point, if experienced based sizing had been applied for this process. In the
case of leaching, the potential complications tend generally to be more severe because of the
combined effect of solution chemistry causing a sharp increase in the separation and handling
equipment requirements.
Extreme examples of process risk involve situations whereby the underflow is virtually un-flowable
within seconds of residence time in the thickener. The underlying cause of these occurrences may
not solely involve superfine particle size. In fact, extreme coarseness can be, at times, more
detrimental compared to over-grinding. In addition, the mere presence of traces of ultrafine
material in these coarse fractions greatly reduces the available handling options. Regardless of its
cause, an un-flowable underflow is readily detectable rheometallurgically during the incipient
stages of process flowsheet development. This, in turn, promptly clears the way forward by simply
reducing the realistic process selection options by precluding thickening. This, in turn, precludes
CCD in the case of leaching, leaving as the only option direct filtration of the discharge stream(s)
involved. Of note is that these rather complex situations occur with comparable frequency across
main unit operations within both the mineral processing and hydrometallurgical circuits of the
processing plants. Generally, the greatest associated challenge is finding an adequate solution to
pump the process discharge streams without relying on ditch-draining or trucking. A process
tested with good results for a specific case is shown in Figure 2, which involves finding the
optimum size fractions and blend ratios that ensure a stable and flowable overall combined-tailings
blend.
To the authors knowledge, universal solutions are not available. This is partly because additional
complications arise in situations when separations, whilst physically possible, are not readily
feasible due to chemical instability, footprint and several other practical limitations. Affected
processes are quite representative of many precious, light and rare metal extraction flowsheets,
saline water based operations (flotation and leaching), laterites processing, as well as certain
purification and detoxification circuits (Mezei et. al, 2011, 2002, 2003, 2006 and 2002). Here is where
the net advantage of the rheometallurgical process risk analysis comes in, as it is unique in that it
flags virtually any process risk whilst establishing the design criteria. The data package produced,
as outlined in the previous section, allows for a safe preliminary process selection in terms of key
separation, handling and related unit operations. Adverse flow properties are being flagged,
without exception, regardless of their root cause(s) or combination thereof. The direct implication of
this fact is that the rheometallurgical risk assessment can be most effective if conducted during the
earliest possible stages of the process development and adjacent metallurgical testwork. The net
overall effect is advanced screening of high quality metallurgical projects from their incipient stage.
This prevents unnecessary expenses, leading to dramatically reduced testing costs and fast turnaround times during the pre-feasibility stage. Most importantly, the rheometallurgical risk
prevention creates, with justified confidence, the premises of a fully functional metallurgical plant,
which is the ultimate objective of the process development exercise.

421

Flotation tails, Leach residue, Backwash and Combined

Primary blending

Hydrocyclone separation at 30

Undersize (Cyclone OF) <30

Oversize (Cyclone UF) >30

Thickening

Thickener OF

Thickener UF

Recycle to process

Secondary blending

Discharge to pond

Figure 2 Circuit flowsheet for the treatment of mixed residues

Effect of ionic strength on thickener performance

Ionic strength can be singled out as the most influential factor determinative of metallurgical pulps
and slurries flow behaviour, which can be comparable at times to the solids content. This topic was
discussed in great detail in previous publications, which includes the discharge treatment options
of the continuous pressure oxidation (POX) discharge stream (Mezei et. al, 2002). For the purpose of
this discussion, the relevant dependencies illustrated in Figure 2 indicated that the POX conditions
producing the best extractions (thus displaying the highest ionic strength) created the worst liquidsolid separation conditions. The specific thickener unit area requirement (TUFUA) for the POX
conditions that produced 95% average metal (Cu, Ni, Co) extraction was four times greater
compared to the requirements for the POX conditions that produced 77% metal extractions, i.e. 0.36
m2/t/day vs 0.09 m2/t/day. The corresponding specific flocculant consumption was about 13 times
greater, i.e. 164 g/t vs. 13 g/t. In addition, the rheological profiling (Figure 3) displayed a sharp
decrease of the CSD values with the increase of the ionic strength, which predicts underflow solids
contents ranging from 60% wt. to about 28% wt. across the same interval. This combined
rheometallurgical data package pointed to a reasonable, risk-minimized, optimum target extraction
of about 92%. Any increase above that would imply exponentially rising capital and operating
expenses in order to ensure continuous plant functionality.

422

0,4

52 g/L H2SO4 PLS ~

164 g/t floc

50

0,35
0,3

40

0,25

30 g/L H2SO4 PLS ~

30 g/t floc

30

0,2
0,15

20

0,1
10

0,05

21 g/L H2SO4 PLS ~ 13 g/t floc

0
70

75

80

85

90

Specific TUFUA, m2/t/d

UF Solids density, % wt.

60

0
95

100

POX extraction efficiency, %

UF Solids Density, % wt.
Specific UF Unit Area

Figure 3 Continuous POX discharge settling-thickening response

Yield stress, Pa

60

6.9% wt. AC feed-21 g/L FA disc

50

8.3% wt. AC feed-30 g/L FA disc

40

11.8% wt. AC feed-52 g/L FA disc

30
20
10
0
20

25

30

35 40 45 50 55 60
Underflow solids, % wt.

65

70

Figure 4 Continuous POX discharge underflow rheological profile

Effect of washing on filter and process performance

In most cases, washing enhances the performance of vacuum or pressure filters, which leads to
increases in dry equivalent output, as shown in the first example (Li leach discharge) in Table 1.
Exceptions to this include situations when precipitation of sparingly soluble species existing in
solution occurs on the filter due to dilution by the wash water, as shown in the second example
(iron arsenic removal discharge) in the same table. This could result in severe over- or under-sizing
of the equipment, ranging from +67% to -44%, respectively, for these two particular cases discussed.
This can be avoided by validating the design criteria as to accurately reflect the true output of the
both the fresh discharge as well as of the final cake, as exemplified. A third situation discussed
within this category involves re-dissolution of pay-metal contained in a high-value final product.

423

Lithium carbonate, which displays a solubility inverse proportional to temperature, is a typical

example of a process requiring hot washing of the precipitate to limit the losses during the washing.
Even so, a certain amount of lithium is being re-dissolved. Quantifying that this amount is critical
in order to allow for defining the wash-solution recycle rate as to prevent Li bleeding out from
the cycle filter-wash circuit. The performance of a tree-stage counter-current washing of lithium
carbonate is shown in Figure 3. Accordingly, whilst the soluble impurities were removed at about
96% efficiency, the lithium washing efficiency was practically non-existent by definition. Instead, a
lithium loss occurred during the process due to its re-leaching in fresh water. This resulted in a
7.2% recirculating load, which was determined using the real data model established based on the
data. The model returned optimized water-balance conditions that defined the fresh water entry
points and upstream advance.
Table 1 Effect of washing on filter performance

Process discharge

Li leaching

Filter feed equivalent dry

Purification at pH ~ 4.6, Fe, As, Al...

PLS Filter only specific unit area

7823

1234

dry kg/m2/hr for one stage

PLS Filter only area requirement

0.13

0.81

m2

Filter-wash feed equivalent dry

1.00

1.03

t/h if weight delta by wash

Filter-wash specific unit area

16769

478

dry kg/m2/hr

Filter-wash area requirement

0.18

6.46

m2 for three stages

PLS & Filter-wash total area

0.31

7.27

m2 without safety factor

67

-44

% + over / - under sized

Sizing error

t/h

100
80
60
Li-Predicted

Wash efficiency, %

40

Na - Real

20

Li-Real

0
-20

0,5

1,5

2,5

3,5

-40
-60
-80
Wash stage, n, with 0 = PLS

Figure 5 Three-stage counter-current washing of lithium carbonate

424

CONCLUSIONS
According to the rheometallurgical concept, the main root cause of the process risk primarily
resides in disconnects between the optimized metallurgical parameters and the physical conditions
created as a result. The parameters mentioned tend to consistently include particle size and process
chemistry, both treated preferentially during the metallurgical work-programs. However,
considering that the effect on process pulps flow behaviour is commonly ignored, the metallurgical
efficiency is maximized in isolation. Whilst promising in the laboratory, the resulting processes
tend to be non-functional in reality, which places the project at risk due to technical factors that are
entirely preventable. This is because commercial plants can operate only in continuous flow mode,
completing a well-defined, process-specific rheometallurgical cycle that can be easily determined.
Therefore, the rheometallurgical process risk analysis flags virtually any process risk whilst
establishing the design criteria. Adverse flow properties can be addressed early, without exception
and regardless of their determining factors. The direct implication of this fact is that the
rheometallurgical risk assessment can be most effective if conducted during the earliest possible
stages of the process development and adjacent metallurgical testwork. As such, the
rheometallurgical risk prevention creates, with justified confidence, the premises of a fully
functional metallurgical plant, which is the ultimate objective of the process development exercise.

REFERENCES
A. Mezei, M. Ashbury, M. Girard, M. Canizares and E. Bourricaudy, A Rheometallurgical approach to liquidsolid separation, 2011 Annual Conference of Metallurgists, October 2-5, 2011, Montreal, Canada.
T.W. Healy, D.V. Boger, L.R. White and P.J. Scales, Particulate Fluids A Key Concept in Advanced Mineral
Processing, XVIII International Mineral Processing Congress, 23-28 May 1993, Sydney, Australia.
A. Mezei, M. Ashbury and I. Todd, Rheological Aspects of Nickel Hydrometallurgy, ALTA 2002: Nickel/
Cobalt 8 Conference, May 2002, Perth, Australia.
A. Mezei, Hydrometallurgical Applications of Rheology Testing, TMS Hydrometallurgy Conference, September
2003, Vancouver, BC.
C.A. Fleming, A. Mezei, E. Bourricaudy, M. Canizares and M. Ashbury, Factors influencing the rate of gold
cyanide leaching and adsorption on activated carbon, and their impact on the design of CIL and CIP
circuits, Precious Metals Conference, June 2010, Falmouth, UK.
A. Mezei, I. Todd, R. Molnar, Can Complex Hydrometallurgical Pilot Plants Effectively Reduce Project Risks?
ALTA 2006: Nickel/ Cobalt 8 Conference, May 2006, Perth, Australia.
A. Mezei, N. Verbaan, Q. Wang and I. Todd, Effect of Salinity on Nickel Laterite Processing, Chloride
Metallurgy 2002 Conference, October 19-23, 2002, 32nd Annual CIM Meeting, Montreal, Quebec,
Canada.
A. Mezei, C. J. Ferron, M. Ashbury, Practical Aspects of Rheological Studies for the Mineral and Process
Slurries, Engineering Foundation Conferences, Rheology in the Mineral and Energy Industries II,
March 14-19, 1999, Hawaii.

425

chap. 8
Hydrometallurgical processes for producing
salt and non-metals compounds

Some improvements in
caliche heap leaching

Caliche is a conglomerate that is exploited in Northern Chile since it is an

important source of non metallics minerals such as iodine and nitrate. In
this work, caliche was leached in column experiments and it was found that
the concentration of some ions such as sulfate and chloride was initially increased reaching a peak after a few days while other more soluble ions such
as nitrate decreases its concentration rapidly. In order to understand the
processes occurring in the column, a model was developed for interpreting
these results, which showed very promising characteristics. The model includes

Javier Ordez and Silvia Valdez.

Universidad de Antofagasta, Chile
Luis Moreno. Universidad de
Antofagasta, Chile and Royal
Institute of Technology, Sweden
Luis Cisternas. Universidad de
Antofagasta and Center for Scientific
and Technological Research for the
Mining Industry (CICITEM), Chile

dissolution and precipitation of minerals and chemical equilibrium.

Regarding the comprehension of phenomena that occurs in caliche heap
leaching, the physical characteristics of this mineral are a critical issue. In
this work, some properties were studied, such as the transport of the leaching
solution, the suction capability of individual rocks by capillarity and the solid
density of caliche.
Results show that on the one hand a good representation of the interactions
between ions during the leaching was achieved through a very simple model
and the other, by physical characterization was determined that the caliche
leaching is governed by two solution transport regimes that are combined
(intra and extraparticle).

429

INTRODUCTION
Caliche, also named natural nitrate deposits, is a mineral composed by a high proportion of water
soluble materials, among which are highlighted commercially: nitrates, sulphates and iodates
(Ericksen, 1983 & Pokorny et al., 1997). The major deposits are located in northern Chile,
particularly in Antofagasta and Tarapac regions.
During early years, caliche was processed by vat leaching, but according to diminution of grades,
other methodologies have become more frequent, such as heap leaching (Lauterbach, 2004; Wisniak
and Garcs, 2001). A current area of study is the use of leaching agents different than fresh water,
resource scarce in the zones of exploitation (Taboada et al., 2012).
The processing of caliche is by leaching, where depending of granulometry, grade or objective can
be vat or heap leaching. Particularly, heap leaching is a process that involves the interactions of
physical and chemical phenomena and for this reason becomes complex to understand the subprocesses that it involves. The combination of experiments with modelling is an alternative
commonly used to address the processes involved in complex systems as the caliche leaching.
Modelling helps not only in the comprehension of these phenomena but also in the optimization
and design of the leaching operations.
In the last time, different types of models have been proposed, for approaching different tasks
(Valencia et al., 2008; Glvez et al., 2012; Ordez et al., 2013ab). The caliche leaching is a very
complex process with many species dissolving at the same time, with particles that are dissolving
and decreasing in size by collapse of the part that has been exhausted of soluble minerals.
Therefore, several parameters are needed in order to describe these processes, but they are not
known for caliche and values of the literature are taken.
Critical issues in heap leaching are the transport of the leaching solution within the heap and the
wetting of the caliche particles; i.e. if the solution flows on the particle surface or through the
particle. If the particles are small this effect is secondary, but for larges particles as those used in
heap leaching it is very important to know the wetting process within the particles. The wetting
process of caliche particles was studied, with emphasis in the capillarity of the caliche ore, which is
one of the most important mechanisms of water transport in porous materials. Despite both are not
explicitly considered into the models, they are relevant in the understanding and modelling of
caliche dissolution. Other physical parameters, which also play a certain role in the modelled
processes in heap leaching, are among others: the particle density, inter-particle porosity, and intraparticle porosity. The last one is important to define the maximum water retention within the
particle.

METHODOLOGY
Caliche and solutions used in experiments
For column leaching experiments was used caliche obtained from Antofagasta region in Chile and
in order to avoid problem of size segregation and channelling, the material was passed through
meshes to reach an average particle size of 6.3 5.5 mm. Solutions employed in leaching were
seawater, synthetic leachate prepared in laboratory, tap water and a mixture between seawater and

430

synthetic leachate. The chemical analysis of caliche (%w/w) and solutions (kg/m3) can be observed
in Table 1.
Table 1 Chemical analysis of caliche and solutions used in leaching experiments
Material

NO3-

SO42-

IO3-

Cl-

Na+

K+

Mg2+

Caliche, %w/w
Tap water (TW), kg/m3
Seawater (SW), kg/m3
Synthetic leachate diluted (LD), kg/m3
Synthetic leachate (LC), kg/m3

3.67
0.00
0.21
15.02
29.30

9.41
0.13
2.66
28.06
53.72

0.05
0.01
0.00
0.10
0.20

4.19
0.35
19.71
39.13
59.88

5.58
0.19
11.12
33.76
56.27

0.58
0.02
0.36
2.71
5.02

0.97
0.03
1.58
4.51
7.70

Density,
kg/m3
999
1004
1096
1163

For the physical characterization studies, macroparticles of caliche of different sizes were used. Tap
water, seawater, and kerosene (=811 kg/m3) are used for these tests. Kerosene was used for
determining parameters where mineral dissolution would be avoided.

Column leaching experiments

Material was loaded in 0.2 m column at three distinct heights. They were watered with three
different leaching agents and three different irrigation rates. The resume of experiments is showed
in Table 2.
Table 2 Experimental design of caliche leaching in columns
Experiment
TW16
SW63
SW66
SW13
LD66
LD84
LC16

Bed height (m)

1.0
0.6
0.6
1.0
0.6
0.8
1.0

Irrigation rate (m3/m2/h)

0.006
0.003
0.006
0.003
0.006
0.004
0.006

Leaching agent
Tap water
Seawater
Seawater
Seawater
Seawater + Leachate
Seawater + Leachate
Leachate

A first sample was taken 12 hours after starting the drip at the bottom of column, after that the time
interval was 24 hours during 20 days.

Chemical reaction modelling

The caliche mineral is formed for several soluble minerals that, in some cases, are a compound
formed by several cations. Nitrate, chloride, and sulphate are the most common anions; while
sodium, calcium, magnesium, and potassium are the most common cations. The concentration of
the sodium, at the beginning of the leaching process, is very high due to the presence of the highly
soluble sodium nitrate is leached increasing strongly the concentration of sodium. Some minerals
containing sodium, such as sodium sulphate and sodium chloride may therefore reach the
saturation and precipitate due to the effect of the common ion (Cisternas, 2009).

431

A model was developed to take into account the effect of the common ion on the less sodium
compounds as such sulphate and chloride. The model is highly simplified, but the aim is to shown
that the anomalous shale of the breakthrough curves may be due to the effect of the common ion.
The model is one-dimensional and is applied to one column located at the heap centre. The system
is described as formed by well-stirred reactors (mini-column) in series. The model considers the
chemical reactions in each reactor (dissolution and precipitation) and the transport of solute from a
reactor to the next one. The model includes three sodium minerals ( ) and four ions ( ): sulphate,
nitrate, chloride, and sodium; with the amount in moles in each reactor and the concentration
in moles per litre. The general expression referred to mass balance for the anions in each reactor is:
,

(1)

As sodium is a cation associated with all the anions studied, the equation for it in each reactor is:
,

"#$ %& ,

#' & ,

()& ,

(2)

For the minerals formed by sodium, these are: sodium sulphate, sodium nitrate and sodium
chloride, the general equation is:
+,

= ,-+

.,

/1

23+

+,

(3)

The exponent 5 corresponds to the valence of the anion. For monovalent ions the value of 5 is 1 and
for sulphate (bivalent) mineral the magnitude is 2.

Determination of physical characteristics of caliche

Three physical properties were measured for the caliche: a) water transport, b) capillary suction and
c) density of the caliche.

Water transport
When the water transport is measured, it is important to determine if the water flows around or
through the caliche stone. If the flow regimen includes both patterns, it is then important to
determine in which extent one is dominant. In the experiment, the caliche stone is irrigated with a
leaching solution, using peristaltic pump. The solution drips on a small sponge, located in the top
of particles of caliche, which is kept in place by a support. The irrigation rate is similar to that used
in heap leaching (0.006 m3/m2/h). Once the experiment is stopped, the stone is cut or destroyed in
order to determine the extent of the wet zone and the level of wetting. Two rocks with different
cross sectional areas were studied and irrigated with seawater. All the experiment is document by
taking photograph very often.

432

Capillary suction
The capillary suction is measured setting a caliche stone in contact with a liquid and determining
the height of the wet zone on the surface of the caliche as a function of time. Once the experiment is
stopped, the stone is cut or destroyed in order to determine the extent of the wet zone in its interior.
A stone per each size was used; seawater and tap water were employed as solutions. All the
experiment is document by photographs.

Mineral density
Regarding the determination of caliche bulk density, the samples were weighted and submerged in
a glass beaker with a previously determined volume of kerosene. After 3 minutes, the volume of the
liquid and particle was registered and volume of the particle was determined. The bulk density of
the caliche was determined by dividing the mass by volume. In order to obtain representative
densities, 2 particles of small and medium size were used.

RESULTS AND DISCUSSION

Evaluation of the column experiments using the chemical reaction model
In the initial period the concentration of sulphate and chloride is increased, while the concentration
of the sodium and nitrate decreases. A previously developed phenomenological model cannot
describe this behaviour; however a chemical reaction model could explain this behaviour (Ordoez
et al., 2013a). The model considers the chemical reactions occurring in the column, including
dissolution of minerals and its precipitation if the solubility product is exceeded.
The objective of this model is to reproduce the trends observed in the column experiments. The ion
sodium is associated to nitrate, sulphate, and chloride, since the solubility of the sodium nitrate is
very high could be expected that its dissolution were independent of the other ions. On the other
hand side, the solubility of the sodium sulphate and sodium chloride is lower, therefore, it is
expected that the solubility product for this minerals were reached at the initial period, when the
concentration of sodium is very high.
The amount of soluble minerals was determined by analysis of the solid samples. The solubility
constant were determined using the concentrations existent in the leachate al beginning of the
leaching process, when it is expected that the leaching solution is saturated with respect to sodium
sulphate and sodium chloride.
The dissolution constants for the sulphate, nitrate, and chloride sodium minerals (,-+ ) in the model
were fitted with the experimental data using the least square method. In the seven cases, it was able
to reproduce the trends observed in the outlet concentration with time. Some fitting results are
shown in Figure 1 for the experiments: SW13, TW16, LC16 and LD66. The model reproduces the
most important characteristics of the experimental data; the increase of the sulphate and chloride
concentration at the beginning, once that part important of the sodium nitrate has been dissolved
and removed from the column.

433

0
0

a)

Concentration, mol/L

eNO3
mNO3
eCl
mCl
eSO4
mSO4

200
Time, h

400

0
0

b)

eNO3
mNO3
eCl
mCl
eSO4
mSO4

Concentration, mol/L

eNO3
mNO3
eCl
mCl
eSO4
mSO4

Concentration, mol/L

Concentration, mol/L

200
Time, h

400

eNO3
mNO3
eCl
mCl
eSO4
mSO4

0
0

c)

200
Time, h

400

d)

200
Time, h

400

Figure 1 Experimental and simulated outlet concentration for columns: a) SW13, b) TW16, c) LC16 and d)
LD66 using the chemical reaction model. Lines are modelled data and markers are experimental data

Nitrate and chloride that forming caliche are predominantly as nitratine and halite, respectively.
The most common sources of sulphate are bloedite, gypsum, anhydrite and glaubertite. Darapskite
and humberstonite are also present as nitrate-sulphate minerals. Sodium is widely associated to
other anions, especially as common ion of nitrate, chloride and sulphate (Ericksen, 1983).

Determination of physical characteristics of caliche

Water transport
No appreciable differences in the water content of the rocks were observed, being in all cases less
than 5%; however the water fraction of Rock 2 (Figure 2) was greater. At the moment of particles
selection, it could be observed that rock 1 had a rocky appearance, which resulted in the higher
level of impermeability showed by this sample through a significant less time required to elute the
first drop from the particles, almost 4 times lower (Table 3).
Table 3 Dripping test for two different particle sizes
Description
Cross sectional area (cm2)
Dry weight (g)
Water retained (%)
Time of first drop eluted (min)

434

Rock 1
35
433.09
3.7%
54

Rock 2
56
693.70
4.4%
220

Figure 2 Progressing of wetting of rock 2. Time in minutes, A: 90, B: 150, C: 180, D: 210, E: 240 and F: 300

With respect to saturation front, in the rock 1 entered few water, finding a wet zone inferior than
1.5 cm. On the other hand, the wet front of rock 2 was more massive, reaching dimensions close to
3.5 cm in the same time. It can be said that the liquid flows through both the externally exposed
surface of the particles as well as internally, observing a mixed transport behaviour. This can be
useful in the formulation of more comprehensive models, where the solution transport is a relevant
phenomenon to consider, especially when large particles are managed.

Capillary suction
For capillary suction experiments, the advance of wet front in the surface of rocks was the indicator,
for that, the mean of three measures was annotated in the results. Particles of the three sizes were
employed (Table 4).
Table 4 Important results about capillary suction experiments using seawater
Name

Particle
mass (g)

Contact
area (cm2)

Moisture
(%)

SAMP1
SAMP2
SAMP3

290
1090
1870

20
47.5
44

4.0
2.1
0.9

Advance of wet
front at 1 min
(cm)
0.3
1.4
0.2

Advance of wet
front at 10 min
(cm)
1.2
3.3
3.2

Advance of wet
front at final (cm)
4.9
5.7
9.5

Visually, it could be observed that the water ascension is not regular due to the composition and
geometry is not homogeneous and symmetric (Figure 3a).

435

10

Height, cm

8
6
4
SAMP1

SAMP2
SAMP3

0
0

a)

b)

100
200
Time, min

300

Figure 3 Capillary suction test for SAMP3 experiment. a) Wetting progression Time in minutes, A: 0, B: 3,
C: 15, D: 37, E: 52 and F: 575 and b) Advance of wet front in capillarity suction tests

The velocity of suction was faster at initial times and in advanced stages the front of wetting was
almost unaltered for all particle sizes (Figure 3b). The results of suction are strongly dependent of
mineral composition, contact area and solution employed in tests. No important differences
between the use of seawater or tap water were observed, however when kerosene was employed,
the suction is quickly but mainly superficial. The use of any of the waters implies that a fraction of
caliche is disaggregated during suction in contrast of kerosene.

Particle density
The density of particles is an important parameter that is included in the presented model. From
the results obtained, no big differences were observed between the different particle sizes, due to in
homogenous material, the density is independent of size. For smaller stones, the value was
2430 kg/m3 and for larger 2405 kg/m3. Higher magnitudes than the values near to 1900 kg/m3
reported for caliche ores were obtained, according to expected results due to in situ determination
consider intra-particle spaces and fractures that diminish the mass per volume unit.

CONCLUSIONS
A very simple model was described that considers with a good quality the chemical interactions
that occur into column leaching and under different operational conditions and leachants,
contributing to the understanding of this phenomenon.
The ascension of liquid by capillarity was observed, supporting the affirmation that the water
transported by suction is important, but not the exclusive way. Subsequent, dripping experiences
showed a mixed transport behaviour, which is interesting to consider in subsequent modelling
works, where the solution transport is a relevant phenomenon to consider, especially when large
particles are managed. The density of particles of caliche was measured and resulted according to
expected. All the physical parameters are transversally influenced by mineralogy; therefore integral
analysis should be done, managing the mineral species that forming this caliche.

436

ACKNOWLEDGEMENTS
The authors wish to thank CONICYT for support through Project MEL 81105010 and J.I.Ordez
wish to thank CONICYT for the PhD scholarship.

REFERENCES
Ericksen, G.E., 1983. The Chilean Nitrate Deposits. American Scientist 71, 366-374.
Glvez, E.D., Moreno, L., Mellado, M.E., Ordez, J.I., Cisternas, L.A., 2012. Heap leaching of Caliche minerals:
Phenomenological and analytical models Some comparisons. Minerals Engineering 33, 46-53.
Lauterbach, A. 2004. Chapter 4 Reduction of perchlorate levels of sodium and potassium nitrates derived from
natural caliche ore. American Chemical Society.
Ordez, J.I., Moreno, L., Glvez, E.D., Cisternas, L.A., 2013a. Modelling of caliche mineral leaching with
seawater. Submitted to Computers and Chemical Engineering Journal.
Ordez, J.I., Moreno, L., Cisternas, L.A., 2013b. Seawater leaching of caliche mineral in column experiments.
Submitted to Hydrometallurgy Journal.
Pokorny, L., Maturana, I., 1997. Sodium nitrate, fourth edition. Kirk-Othmer ECT encyclopedia of chemical
Technology 22, 427-438.
Taboada, M.E., Hernndez, P.C., Galleguillos, H.R., Flores, E.K., Graber, T.A., 2012. Behavior of sodium nitrate
and caliche mineral in seawater: Solubility and Physicochemical properties at different temperatures
and concentrations. Hydrometallurgy 113-114, 160-166.
Valencia, J.A., Mndez, D.A., Cueto, J.Y., Cisternas, L.A., 2008. Saltpeter extraction and modelling of caliche
mineral heap leaching. Hydrometallurgy 90, 103-114.
Wisniak, J., Garcs, I., 2001. The rise and fall of the salitre (sodium nitrate) industry. Indian Journal of Chemical
Technology 8, 427-438.

437

Authors index

D
Albis Leo, Versiane. 79, 339

Daud, Omar. 119

Arriagada, Francisco. 191

de la Torre, Ernesto. 93, 103, 171

Ashbury, Mike. 417

de Oliveira, Debora Monteiro. 347

Ausburn, Kevin. 49

de Sousa, Adelson. 275

Aylwin, Pedro. 285

Delvalle, Florent. 143

Demergasso, Cecilia. 365

B
Basualto, Carlos. 261

Baum, Wolfgang. 49

Espinoza, Cynthia. 93

Beard, Simn. 329

Estay, Humberto. 153, 161, 191

Bravo, Luis. 409

Broschek, Ulrike. 409

Brugger, Jol. 127

Flores, Eduardo. 389

Bustos, Waldo. 191

G

C
Cceres, Luis. 311
Carvajal, Pablo. 153, 191
Casas, Jess. 389
Castillo, Carlos. 87
Castro, Osvaldo. 217, 225, 235
Castro, Sergio. 191
Cautivo, Dina. 365
Cifuentes, Gerardo. 303
Ciminelli, Virginia. 181, 275
Cipriano, Aldo. 401
Cisternas, Luis. 429
Clayton, Casey. 295
Contador, Yasna. 365
Cornejo, Jorge. 409
Correa, Mauricio. 119

Galdos, Helbert. 35
Galleguillos, Hctor. 69
Galleguillos, Pedro. 365
Garcs, Juan Pablo. 389
Gebrehiwot, Ephrem. 295
Gimnez, Patricio. 87
Goad, Robin. 59
Gonzlez, Carola. 87
Gonzlez, Karina. 153, 191
Gow, Nick. 25, 265
Graber, Tefilo. 69
Greco, Antoine. 143
Guevara, Alicia. 93
Guevara, Javier. 35
Guzmn, Amado. 375

Cotoras, Davor. 355

439

M
Herrera, Miguel. 401

Majuste, Daniel. 181, 275

Hope, Greg. 265

Mannheim, Rodolfo. 303

Hu, Christian. 389

Martins, Eder. 275

Huang, Hsin-Hsuing. 265

Melani, Patricio. 285

Mezei, Alex. 59, 417

Miller, Graeme. 375

Molnar, Ron. 59

Jara, Sebastin. 211

Morais, Carlos. 135

Johnson, Michael. 59

Moreno, Luis. 429

Joly, Philippe. 253

K
Kabugo, James. 245

Morgan, Jason. 295

N
Navarro, Patricio. 211

Karelovic, Pablo. 401

Ngothai, Yung. 127

Kratochvil, David. 201

O

Ordez, Javier. 429

Lagos, Josu. 409

Ortiz, Miguel. 161

Lagos, Nicols. 285

Osses, Mara de la Luz. 329

Lee, Keith. 59
Leinenweber, Genny. 311
Lencio, Hamilton. 339

P
Padro, Diogo de Oliveira. 347

Letelier, Hugo. 87

Pascuale, Sara. 235

Li, Kan. 127

Peixoto, Johne. 79

Ljubetic, Kresimir. 201

Pring, Allan. 127

Lobos, Jorge. 409

Lopes, Klinger. 339
Lpez, scar. 201
Lupu, Cornelia. 59

Q
Quilodrn, Alejandro. 217, 225,
235

440

Torres, Cynthia. 69

Torres, Marco. 119

Ramrez, Luis Felipe. 321

Reyes, Francisco. 253, 401
Ricaurte, Jos. 103

Rivera, Gonzalo. 225

Valdez, Silvia. 429

Robertson, Stefan. 375

Valenzuela, Fernando. 261

Rocha, Laura. 135

Vsquez, Vernica. 153

Rodrigues, Michael. 79, 339

Velarde, Guillermo. 111

Rojas, Solange. 401

Veloso, Tcia. 79, 339

Romero, Fernando. 401

Verdugo, Luis. 261

Romero, Julio. 161

Viedma, Pabla. 355

Ruiz-del-Solar, Javier. 119

Volker, Karien. 409

W

S
Saavedra, Arnaldo. 35

Watling, Helen. 69

Saavedra, Matas. 329

Weigel, Valrie. 143

Salamanca, Mauricio. 217

Salgado, Cristian. 401

Salonen, Piritta. 245

Xia, Fang. 127

Sandoval, Scot. 295

Sanguinetti, David. 201
Sapag, Jaime. 261

Y
Yez, Hctor. 191

Serrano, Juan Pablo. 171

Young, Courtney. 25, 265

Sobral, Luis. 347

Soto, Jos. 365
Souza, Clauson. 181

Z
Zahn, Randy. 49

Zambra, Rodrigo. 225, 235

Taboada, Mara Elisa. 69
Tejeda, Gabriel. 401

Zenteno, Luis. 87
Zepeda, Vctor. 365
Zhao, Jing. 127

441

Editors

Fernando Valenzuela
Dr. Fernando Valenzuela is a professor of Unit Operations and
Mining and Metallurgical Chemistry at The Universidad de Chile.
He is also a chemist, graduated from the same university, and has
a Master's degree in Chemical Engineering from Kyushu University,
Japan, specializing in hydrometallurgy and wastewaters treatment.
He is a member of the Chemistry PhD faculty at his university,
author of 77 ISI scientific publications, two chapters of technical
books and five international patents on mining wastewaters
processing.

Courtney Young
Dr. Courtney Young is Department Head and Lewis S. Prater Distinguished Professor of Metallurgical and Materials Engineering
at Montana Tech where he has been teaching courses in mineral
processing and extractive metallurgy for 20 years. He has authored
and presented more than 100 papers, most dealing with flotation
as well as metal leaching and recovery but also includes applications
in recycling and remediation, particularly solid wastes and acid-rock
drainage.

443

colop h on

This book was designed by Estudio Vicencio. The text

was output in the font Arnehm Pro designed by
Fred Smeijers. The cover was printed on
UPM 120 Fine paper and the inside
pages on Bond paper.
The first edition of 400 copies was
printed and bound in July
2013 by QuadGraphics,
Santiago, Chile.