CIM Lecture

The Evolution of Copper Smelting Practices in the Last Four Decades

Thank you Dave for your kind words. I want to thank the CIM for
nominating me as one of the Institute five lecturers for the period 2008-
2009. Thanks are also due to Atlas Copco for supporting this CIM program.

I was most pleased to accept Prof. Dreisinger’s invitation to present my

lecture at the University of British Columbia. Meeting good friends once
more was an important motivation, and knowing that students would be
present in the audience was further incentive. I initiated my professional
career as an academic at the University of Chile. Later in life, I had the
opportunity of teaching again at the University of Toronto, where I am still
an adjunct professor. In addition, during my entire professional life, I have
dedicated time and effort to support strong academic activity in mining and
chemical process metallurgy in universities, north and south of the equator.
The truth is that in an academic environment, I feel totally at home.

SLIDE 1

The subject of my CIM lecture is “The Evolution of Copper Smelting

Practices in the Last Four Decades”. It is based on the paper that Phillip
Mackey and I presented at the Copper-Cobre 2007 conference. The
presentation summarizes the dramatic changes that have taken place in
copper smelting during my professional life. The story exemplifies the
efforts that mining companies have made in the recent past and continue
making in the present to transform an industry that was a symbol of
environmental degradation into a sustainable economic activity.

I’ll start by reviewing a few key aspects of copper pyrometallurgy.

SLIDE 2

The main feed to copper smelters is sulfide concentrate. Main components

of this material are Cu-Fe sulfide minerals such as chalcopyrite and bornite,
Cu sulfide minerals such as chalcocite and significant amounts of pyrite. Cu
concentrates also contain minor amounts of rock minerals; precious metals,
mainly Au and Ag; and elements such as Pb, Zn, As, Sb, Bi that are
normally referred to as impurities.

SLIDE 3

The high temperature processing of copper concentrate involves the

progressive oxidation of S to SO2 and of Fe to ferrous and ferric oxides,
usually conducted in two stages, namely smelting and converting, at a
temperature of about 1250ºC.

• Smelting and converting off-gases are normally treated in acid plants to

capture SO2 as sulfuric acid.
• Iron oxides and rock minerals are collected as molten slag. As the process
of oxidation proceeds, the proportion of ferric iron in the slag increases.
Eventually, undesirable solid magnetite precipitation may occur. This is one
of the reasons why the oxidation of the concentrate S and Fe is usually
conducted in two stages.
• In the smelting furnace, copper, precious metals and variable proportions
of impurities report to matte, that is essentially a liquid solution of Cu2S and
FeS, with minor amounts of O. Matte and slag are immiscible and separate
in processing vessels according to their respective densities.
• Matte is fed to the converting vessel, where the remainder of the Fe and the
S are eliminated. The product of converting is blister copper

SLIDE 4

In the early 1970s, the dominant technology in the Americas, Africa and the
USSR consisted of the reverb furnace and the Peirce-Smith converter.

SLIDE 5

The reverb furnace is a rectangular refractory box. Charge banks are formed
along the furnace sidewalls. Fossil fuel burners, located in one of the furnace
end walls, provide the heat for smelting. The combustion gas carries most of
the heat out of the furnace. The low SO2 content of the off-gas makes it
unsuitable for processing in an acid plant. In summary, the reverb is an
energy-inefficient, labor-intensive and environmentally unfriendly smelting
furnace that usually produces a relatively low copper containing matte,
better known as low-grade matte in industry.
SLIDE 6

The Peirce-Smith converter is a cylindrical, horizontal, tilting furnace,

equipped with tuyeres for the injection of air or oxygen-enriched air. The
feed is molten smelting furnace matte. Peirce-Smith converting is a batch
operation that is conducted in two stages. In the first stage, the Fe content of
the matte is eliminated as ferrous and ferric oxides and the corresponding S
as SO2. SiO2 flux is charged to the converter to form an iron silicate slag that
it is poured out of the converter. In the second stage, the matte remaining in
the converter, that is essentially Cu2S, is blown to eliminate the remainder of
the S. The product blister copper is poured out of the converter that it is now
ready for another converting cycle.

In the early 1970s, converter hoods were poorly engineered. As a result, the
intermittent converter process gas stream was diluted by infiltrated air, thus
lowering its SO2 content. Very few smelters were processing converter off-
gas in acid plants.

SLIDE 7

However, winds of change were already blowing in some copper producing

regions of the world.

 Outokumpu flash smelting, commercialized in 1949, had been adopted in

several Japanese and Western European copper smelters in the late 1960s
and early 1970s, leading to a substantial reduction of SO2 emissions.
These smelters are located in densely populated areas.
 A number of new bath smelting routes were commercialized in the
1970s. The Noranda Reactor in Canada, the Mitsubishi continuous
copper smelting process in Japan, the El Teniente Converter in Chile, and
the Vanyukov furnace in the Soviet Union became alternatives to flash
smelting. Continuous conversion of copper matte made its commercial
debut with the commissioning of the Mitsubishi continuous copper
making process at the Naoshima smelter.

SLIDE 8

In the Outokumpu flash furnace, a blend of dry finely comminuted copper

concentrate and SiO2 sand is injected in a stream of oxygen-enriched air
through a burner at the top of a reaction shaft. The sulfide minerals ignite
upon entering the reaction shaft. The exothermic oxidation reactions are
extremely fast and are completed before the partially combusted mineral
particles and the flux hit the molten bath. Oxygen efficiency is 100%. Matte
and slag are formed on the surface of the molten bath and separate into two
distinct layers.

SLIDE 9

The Noranda reactor is an example of bath smelting. This reactor is a spin-

off of the Peirce-Smith converter. The solid feed that it is continuously
charged by means of a slinger, can be wet and coarse. Oxygen-enriched air
is injected into the bath through tuyeres. The sulfide minerals-oxygen
reactions occur in the vigorously agitated molten bath.

SLIDE 10

Common features of the flash and the bath smelting processes are:

 Use of oxygen-enriched air or just tonnage oxygen as reacting gas;

 Utilization of the heat of reaction of the sulfide minerals of the feed to
satisfy most, if not all, process heat requirements;
 High specific smelting rates;
 Production of high grade matte; and
 Production of steady, low-volume, SO2 strong process off-gas streams
amenable to sulfur fixation in acid or sulfur dioxide liquefaction plants.

SLIDE 11

 In the early 1970s, the Americas were the most important copper-
producing region in the world; 31 primary copper smelters accounted for
more than 50% of world copper smelter production.
 16 of these smelters were located in the USA, producing half of this
copper. At these smelters, input sulfur captured as sulfuric acid amounted
to less than 20%.
 This state of affairs would soon change drastically. In 1971, the
Environmental Protection Agency set 90% sulfur capture as the new
standard for the USA copper smelters.
SLIDE 12

The American industry undertook a multimillion-dollar program to meet this

challenge. Its key elements were:

 Retaining the reverb as the workhorse of primary smelting.

 Capturing particulates from process off-gas streams.
 Processing converter gas in acid plants.
 Erecting tall stacks to improve the dispersion of reverb off-gas.
 Curtailing production under unfavorable meteorological conditions.

SLIDE 13

By the late 1970s, smelter sulfur capture had reached 60% of input.
However, the potential of this sulfur fixation strategy had been exhausted.

SLIDE 14

A more powerful factor than the EPA regulation would trigger true
modernization of primary copper smelting in the USA. Two successive big
increases of the price of oil in the 1970s threw the world economy into a
deep recession. The price of copper fell sharply, and the cost of production
was rapidly increasing. The circumstances required radical changes in the
structure of the copper industry and its management practices, and in
technology. In the smelters, the energy-intensive reverb had definitely
outlived its usefulness.

SLIDE 15

 The modernization of the American primary copper smelters began in

earnest in the late 1970s, and proceeded at full pace in the 1980s.
Noranda reactors were installed in Garfield; Outokumpu flash smelting
was adopted at a new smelter, Hidalgo, and an existing smelter, San
Manuel; and Inco flash furnaces were built in Chino and Hayden
 The modernized plants, most of them operating with a single primary
smelting unit, had substantially larger capacities than the extinct reverb
smelters.
 The smelters that could not justify the investment required to modernize
were closed. The number of smelters was reduced from 15 in 1981 to 8 in
1987. However, due to the larger capacities of the modernized plants, the
American copper smelter output did not decrease by much.

SLIDE 16

By the end of the 1980s, copper smelter sulfur capture in the USA was close
to 90%.

SLIDE 17

In Chile, need to increase smelting capacity, urgency to reduce consumption

of expensive imported oil, relatively low capital cost, and expediency in the
transfer of technology contributed to the rapid adoption of the locally
developed Teniente Converter bath smelting route in the state owned
Codelco and Enami smelters. However, strict environmental smelter
standards would not be imposed until the 1990s.

SLIDE 18

Canada had pioneered the commercialization of new technology: Inco

oxygen flash smelting in 1952, and the Noranda Reactor in 1973. Kidd
Creek, a new copper smelter that started up in 1981, adopted Mitsubishi
continuous smelting. Sulfur was being captured from various process gas
streams at these smelters. However, the government “countdown acid rain
program” mandated substantial additional reductions in smelter emissions by
1994. For some operations, where copper and nickel are jointly found in the
ore, this was a tall order that would require major technological changes.

SLIDE 19

In the 1980s, a new smelter was built in Brazil and another one in Mexico.
In both cases, Outokumpu flash smelting was the technology of choice.
SLIDE 20

In summary, in the early 1990s:

 There were a number of proven energy-efficient and environmentally

sound flash and bath smelting alternatives and also a proven continuous
converting process available to the industry;
 Over 90% of smelter sulfur input was being captured in important copper
producing regions of the world;
 With a dramatic decrease in CO2 emissions, primary copper smelting was
leading the fight against global warming.

Not surprisingly, in the last two decades, industry’s R&D has mainly
focused on productivity increase, a trend that had already started in the
1980s.

SLIDE 21

These efforts have led to a new industry standard: The primary copper
smelter processing over ONE MILLION tonnes of concentrate per year
through one single smelting furnace. This used to be the combined
throughput of at least four reverb furnaces.

SLIDE 22

The Outokumpu technology led the way to this new standard. The key
factors that contributed to increasing flash furnace capacity were:

 High O2 enrichment of the reaction gas.

 Improved solids feed system and concentrate burner design.
 Water-cooling protection of furnace integrity.
 Advances in process modeling and control.
 Higher furnace operating factor.

SLIDE 23

The increase of capacity of the Toyo flash furnace over the years provides a
good example. By 2006, operating at over 70% oxygen enrichment,
throughput was almost four times bigger than the original 1,000 tonnes of
concentrate per day, when the reaction gas consisted of just preheated air.
Simultaneously, the continuous upgrading of the solids feed system and of
burner design have led to improved furnace metallurgy and substantially
reduced dusting.

SLIDE 24

These achievements have been possible thanks to:

 The implementation by Sumitomo of a focused, long term - still on-going

- program that has been carried out by researchers and operators working
in close cooperation; and
 The judicious use of physical-mathematical modeling to improve burner
design, with continuous validation and revision of model against pilot
plant and commercial plant data, practice preached by the late Julian
Szekely and Frank Jorgensen, two of the researchers who have greatly
contributed to the application of basic principles to process chemical
metallurgy.

SLIDE 25

Another prominent copper smelting technology today is “top submerged

lance” injection, TSL, in its two versions, Ausmelt and Isasmelt, both spin-
offs of the original Sirosmelt technology.

This is an Ausmelt reactor. The Isasmelt furnace has similar features. It

consists of a stationary, cylindrical, vertical vessel. Concentrate, flux and
reductant coal are gravity fed through a port located in the roof of the
furnace. Oxygen, air and fuel are injected through a single vertical
suspended lance, submerged in molten slag. The degree of oxidation or
reduction required by the process is controlled by adjusting the fuel to
oxygen ratio supplied to the lance, and the proportion of reductant, such as
coal, added to the solid feed.

SLIDE 26

TSL is highly intensive and extremely versatile. Operation is possible within

wide ranges of temperature and pO2, with wet/dry, fine/coarse solid feed.
Both Ausmelt and Isasmelt have found increased application in copper
smelting and converting in the last 10 years.
SLIDE 27

 Isasmelt was first adopted by Mount Isa for lead smelting in 1983.
Following extensive pilot plant work, two Isasmelt copper smelting
installations were almost simultaneously commissioned at Mount Isa in
Australia and Miami, Arizona, in 1992.
 In 2000, the upgraded Mount Isa furnace reached an annual concentrate
throughput of over ONE MILLION tonnes.
 Last year, the Ilo Smelter in Peru became the sixth copper smelter in the
world to commission an Isasmelt furnace.

SLIDE 28

This quotation from a recent paper on Isasmelt, explains the secret of the
success of this technology:

“Improvements in process control, achieved over more than 13 years of

operation at Mount Isa, have resulted in a highly advanced control system
that ensures that refractory wear is minimized. This sort of development can
only be achieved over many years in an operations environment”.

Just another example of the merits of developing technology in continuous,

direct cooperation with operators!

SLIDE 29

In the last few decades, substantial advances have also been made in
converting:

 Two new continuous converting processes were commercialized in the

1990s: Kennecott-OKO flash converting, and Noranda continuous
converting.
 However, improvements in engineering and operating practice have
helped the PSC to continue being the dominant industry converting
technology.
SLIDE 30

The key features of the three commercially available copper continuous

converters are presented in this slide:

 The Mitsubishi converter is a stationary, round furnace, equipped with a

set of vertical lances for the injection of oxygen-enriched air and
limestone flux. The converter is fed with a continuous stream of molten
matte.
 The Noranda Continuous Converter is a spin off of the Noranda Reactor.
Molten reactor matte is added in batches to the converter. The process
operates with three phases present in the converter: sulfur-saturated
copper, matte (essentially Cu2S), and silica slag. Blowing is done through
the matte layer.
 The geometry of the Kennecott-Outokumpu flash converter is similar to
that of the flash furnace. It is fed with dry, solid, finely comminuted high-
grade matte. The oxygen-matte particle reactions occur in the converter
reaction shaft.

The feed to each of these converters is high-grade matte containing about

70% Cu. The flash and the Mitsubishi converters that operate at higher
oxygen partial pressure produce blister copper. Lime-ferrite slag is used
under these operating conditions. The Noranda converter, with matte present
in the vessel as a third phase, operates at a lower oxygen partial pressure,
thus permitting using iron-silicate slag without excessive precipitation of
solid magnetite. But the product is sulfur-saturated copper, known as semi-
blister.

Both Noranda and Mitsubishi accept solid reverts. But only the flash
smelting/flash converting route has the potential for total decoupling of these
two unit operations.

SLIDE 31

In the last few years, both Ausmelt and Isasmelt have been developing
copper continuous converting processes, called C3 and ISACONVERT
respectively. The feed to either vessel will consist of granulated high-grade
matte.
Isasmelt will commission an ISACONVERT vessel in the Mufulira Smelter
in Zambia in 2009, where an Isasmelt furnace is already in operation.

Feeding of solid smelting matte makes it possible that, in smelters practicing

TSL continuous converting, this operation is decoupled from smelting.

SLIDE 32

Although continuous conversion will be most likely adopted in future new

smelters, and in expansions of existing smelters, the PSC is standing its
ground. The reasons are:

 It is a highly intensive, versatile pyrometallurgical reactor;

 Mechanized tuyere punching, oxygen enrichment of the blast, better
refractories, computerized converting control techniques, and improved
converting practices have led to substantial increases of converter
productivity;
 Vented tapholes and tunnels for transferring melts, secondary and tertiary
hooding plus improved practices have dramatically reduced secondary
emissions;
 Blast oxygen enrichment, longer converter time in stack, and staggering
of converters permit today feeding relatively steady, strong converter
gases to acid plants.
 In fact, Toyo in Japan and Nordeutsche Affinerie in Germany that
operate Peirce-Smith converters are among the most environmentally
sound copper smelters in the world today.

SLIDE 33

Now, I’ll make a few comments on the impact that technology has had on
size of smelters and I’ll also briefly discuss the radical changes in world
distribution of copper smelter output in the last few decades.

SLIDE 34

The size of the largest copper smelters has been steadily increasing over the
years. In the early 1970s, the copper production capacity of the largest
smelter was about 300,000 tpy. However, a typical plant would only
produce 100,000 tpy copper.
With increasing process intensity, smelters have become much bigger.
Today, the annual copper production capacity of a large plant is well over
400,000 tonnes.

SLIDE 35

This trend is better seen in this slide that presents the evolution of world
copper smelter output as a function of smelter production capacity. The data
show that the output of smelters with an annual production capacity over
200,000 tonnes, the green bars in the graph, increased from about 5% in
1975 to about 70% in 2005. Moreover, in 2005, about 20% of the world
copper smelter output came from plants with production capacities in excess
of 300,000 tpy.

SLIDE 36

In the same period, an even more dramatic change has occurred in world
distribution of copper smelter output as illustrated in this slide. In the graph,
the annual copper smelter output of each of the main producing countries or
regions is represented by the width of its respective color band.
 As discussed earlier, in the early 1970s, there were 16 smelters in the
USA producing about 1.5 million tonnes per year of copper. Today there
are only 3 active smelters south of the border, producing about 0.5
million tonnes of copper per year.
 Canadian smelter output has been close to constant throughout the whole
period. A new smelter, Kidd Creek, was commissioned in 1981, while
the Gaspé smelter, after almost 45 years of operation, was shut down in
2002.
 In Chile, annual copper smelter production has increased from about 0.6
million tonnes per year in the early 1970s to nearly 2 million tonnes
today by modernizing and expanding existing smelters.
 Plant modernization and expansion has also led to a steady increase in
copper smelter output in Japan and Western Europe. Japanese and
Western European smelters are all custom operations.
 The biggest change in copper smelter output has occurred in China, with
India - not shown in the graph - not far behind. New plants have
contributed to the rather explosive increase of Chinese copper smelter
production in the last six years to about 3,000,000 tpy at present.
The increase in copper and other metals production and consumption in both
China and India is expected to continue increasing at a fast pace in the
foreseeable future.

SLIDE 37

Technology has also helped to radically change the smoke stack image of
the copper smelter industry, so vividly illustrated by these photos of the
Caletones Smelter in Chile and the Noranda Horne Smelter here in Canada
taken in the 1960s.

SLIDE 38

The graph on the screen shows the big progress made in SO2 fixation in the
main smelter copper producing countries or regions of the world in the
period of interest.

 Smelters in Japan and Western Europe reacted earlier to social pressure

for the protection of the environment. In fact, these smelters had already
achieved substantial capturing of SO2 at the start of the period. This
process intensified further with the adoption of flash smelting.
 In the USA, SO2 capture from copper smelters started in earnest in the
1970s. Just recently, a much smaller American copper smelting industry
has reached SO2 capture levels close to those of Japan and Western
Europe.
 Chilean smelters started to move at a fast pace in the early 1990s, forced
by environmental legislation. Rapid expansion of copper leaching
operations in Chile in the last two decades has created an important
market for the sulfuric acid produced from treating smelter gases.
 In China, all the new smelters built in the last few decades have adopted
new environmentally friendly technologies. In addition, their local
industry provides a continuous expanding market for sulfuric acid. Not
surprisingly, SO2 capture from Chinese copper smelter gases has steadily
increased starting in the mid 1970s.

The brown line in the graph shows the estimated increase of the average
world SO2 capture from copper smelter gases along the period of interest.
This average is approaching 90%. With the commissioning of new TSL
smelters in Latin America and in Africa, this sulfur fixation will continue
increasing.
SLIDE 39

In my professional life, I have had the privilege of witnessing dramatic

changes in my field of work. Today, copper smelting is close to achieving
the status of sustainable technology. Work remains to be done. Those of you,
among the students present in the audience, who decide to pursue a career in
copper smelting will have the opportunity to live times as exciting as I have
lived. The important lesson to remember in future R&D endeavours is that,
in the past, the most significant advances have been achieved with
researchers and operators working in close co-operation.

I’ll end up this presentation by sharing with you Pillip Mackey’s and my
thoughts about the future

SLIDE 40

We believe that:
 The size of smelters will increase further. The average plant annual
copper production capacity will reach about 300,000 tonnes.
 The proportion of large custom smelters will also increase. High smelter
productivity is essential in this very competitive business.
 Flash smelting and TSL will compete for additional territory.

SLIDE 41
 New green-field smelters and probably expanded/modernized smelters
will incorporate continuous converting.
 However, the Peirce-Smith converter will continue having and important
place in copper smelters.

SLIDE 42

 The 400 tonne anode furnace and 100 tph twin-wheel anode caster will
become standard; inroads will be made towards continuous anode
refining, a subject that I haven’t touched this evening due to time
constrains.
 Advanced process control and automation will be introduced in all areas
of the smelter.
SLIDE 43

 In the next few years, average world SO2 capture will exceed 95%.
 Improved physico-chemical models will be available to predict impurity
behavior, and to allow sound environmental elimination and disposition
of these elements.

SLIDE 44

In summary, pyrometallurgical processes will continue having an important

place in the production of copper from sulfide feeds in the foreseeable
future.

SLIDE 45

Thank you very much!