Pyromet or hydromet? How do pyrometallurgical and hydrometallurgical processes for sulphide concentrates compare?
In the 1970s, the rise in energy prices and increasing environmental concern about sulphur-dioxide emissions spawned intense interest in hydrometallurgical routes as an alternative to smelting for the treatment of sulphide concentrates, particularly copper. Hydrometallurgy offered the prospects of eliminating air pollution, direct electrowon cathode-grade copper, and possibly lower capital and operating costs.
A number of processes, mostly chloride based, were tested at laboratory and pilot scale and although technically successful, the practical problems of operating plants with complex flowsheets in highly corrosive conditions were severe. Few produced cathode-grade copper consistently and while air pollution in the form of sulphur dioxide was eliminated, the processes presented their own environmental problems in the disposal of liquid and solid wastes. There were also difficulties in the recovery of gold and silver values, and with the purity of the elemental sulphur by-product.
Meanwhile, the same concerns (of energy conservation, environment, and reduction of capital and operating costs), accelerated development of new, high-intensity pyrometallurgical processes. These included the Outokumpu and Inco flash furnaces, the Noranda reactor, Mitsubishi process, El Teniente Modified Converter, and for lead the Kivcet and QSL furnaces.
In the developed countries, the need to reduce sulphur-dioxide emissions led to the progressive replacement of reverberatory furnaces as primary smelting units in copper smelting, as these produced low-strength off-gases unsuitable for acid production. Flash smelting, invented by Outokumpu in 1949, became the dominant technology (See E&MJ October 1989). The number of installations increased from seven in 1970 to 24 in 1980 and now stands at 35. The first Inco flash furnaces built outside Canada were commissioned at the Hayden and Chino smelters in the United States in 1983 and 1984 as replacements for existing reverberatories.
The first Noranda reactor was commissioned at the Horne smelter in 1974 and three units replaced reverberatory furnaces at Kennecott's Garfield smelter in 1977. The El Teniente Modified Converter, patented in Chile in 1977, has been installed at the El Teniente, Potrerillos, and Chuquicamata smelters.
More recent pyrometallurgical developments include Isasmelt now being marketed worldwide by MIM Holdings Ltd., KHD Humboldt-Wedag's Contop process, and Lurgi's Flame Cyclone Reactor. The latter is still at the pilot stage.
These improvements in pyrometallurgy and technical uncertainties surrounding aspects of hydrometallurgical routes have combined to reduce their relative economic attractiveness. Cost differentials have narrowed, so any hydrometallurgical process must, as an untried technology, offer significant economic advantages within environmental constraints.
Capital costs for hydrometallurgical plants are roughly proportional to throughput because many of the unit processes are modular. Electrowinning is the most expensive section of the plant, accounting for about 25% of the capital cost. Proportionately lower capital outlay for small throughputs should make hydrometallurgical plants attractive for operations producing 30,000 mt/yr copper or less. A conventional smelter, because of the number of peripherals, is unlikely to be profitable at less than 100,000 mt/yr.
On the other hand, hydrometallurgical routes are usually more energy intensive than smelters because of the electrowinning. Modern pyrometallurgical processes are much more energy efficient and produce high strength [SO.sub.2] gas suitable for acid making. In many places marketing of acid is a problem but to select a hydrometallurgical route solely because of acid marketing difficulties is unlikely to be sufficient economic justification. Apart from high purity, recovery, and marketability of the elemental sulphur, a hydromet route must produce copper which needs no further refining at high recovery, and with precious-metal credits.
Potential applications for copper hydrometallurgical processes appear limited to small operations where a conventional smelter would be uneconomic, the treatment of difficult-to-smelt "dirty" concentrates, and situations where smelting is not an option due to the inability to market sulphuric acid from the off-gases. Ever more stringent environmental regulations make it almost inconceivable that new smelters will be permitted without almost 100% sulphur recovery, at least in the developed world.
Hydromet cannot compete with modern, high-capacity smelters treating relatively clean concentrates, although a hydrometallurgical plant for incremental expansion of an existing smelter would improve the versatility of the overall complex in terms of the range of feeds, byproducts, and residues it could treat.
There remain two hydrometallurgical processes for sulphide concentrates with potential: CMEP - Cuprex Metal Extraction Process, and the GCM - Great Central Mines process. Both are chloride based.
CMEP has been successfully piloted and was described in detail in E&MJ September 1989. Copper concentrate is leached in a single stage with ferric chloride at 95 [degrees] C and atmospheric pressure to produce cupric-chloride solution. Leach residue consists of gangue, pyrite, and up to 65% elemental sulphur which can be recovered by flotation or extraction. Copper is removed from pregnant liquor by solvent extraction using Acorga CLX50, an extractant specially developed for chloride media. Loaded organic is stripped in three stages with hot water, enriched with chloride ion, and then sent to the special Metclor electrolysis cell where granular copper is recovered in the cathode compartment. Silver reports in the raffinate from the solvent-extraction stage and can be recovered by zinc-dust precipitation.
CMEP is the only hydrometallurgical process shown to have consistently produced cathode-grade copper, though in granular form. The core technology is considered proven and according to the joint-venturers ICI, Nerco Minerals, and Tecnicas Reunidas. A number of potential clients have shown serious interest for projects in the 20-30,000 mt/yr copper range. Meanwhile, investigations continue into some of the peripheral factors. Grades and recoveries of elemental sulphur, silver, gold, and molybdenum need to be quantified more accurately.
Methods for transforming the granular copper to marketable shapes are also being explored.
The GCM process was developed by Great Central Mines Ltd. and Bacon, Donaldson, and Associates. It is a chloride-based hydrometallurgical process that has been tested at the laboratory scale on both clean and dirty copper concentrates containing zinc, lead, arsenic, and antimony. In 1986-1987, Hudson Bay Mining & Smelting (HBMS) examined the process as one option for modernizing its Flin Flon smelter because it treats a number of dirty concentrates and there is no market for acid in the Manitoba location. The original process had to be modified for the dirty concentrates.
In the modified process, concentrate is ground to 80% minus-28-micron and pre-leached with hydrochloric acid to remove pyrrhotite, zinc, and lead. After solid/liquid separation, the copper-free leach solution is sent to an evaporator/oxyhydrolysis circuit where water is removed to maintain the water balance, iron is converted to hematite, non-ferrous metals to sulphates, and hydrochloric acid regenerated for recycle.
The HCl pre-leach is not necessary for clean concentrates. Solids from the pre-leach are leached with acidic ferric/cupric chloride at atmospheric pressure to extract at least 99% of the copper and 93% of the silver, and convert reactive sulphides to elemental sulphur. A two-stage reduction/oxidation leach with inter-stage thickening is used to ensure high metal-extractions. Pregnant solution going to the electrowinning circuit is in the cuprous/ferrous state. There is no solvent-extraction step to purify the solution prior to electrowinning.
After silver removal, the solution is electrolysed in diaphragm cells to produce a granular copper product. The diaphragm cells regenerate a portion of the ferric-chloride lixiviant. The rest is regenerated in two turbo-aerators.
Approximately 85% of the copper is recovered in the main electrowinning cells and the remaining 15% from a stripper-cell on the catholyte-bleed stream. Hydrogen-sulphide gas generated in the pre-leach is used to precipitate the residual copper. Excess [H.sub.2]S from the pre-leach is oxidized to elemental sulphur in the second turbo-aerator.
Final residue contains pyrite, elemental sulphur, and gangue. Sulphur can be recovered by flotation followed by hot pressure-filtration leaving a residue which must be separately treated for gold extraction.
The HBMS study concluded that the GCM process was a possible long-term option for Flin Flon because of more stringent [SO.sub.2] emission regulations, acid marketing problems, and higher zinc and copper recoveries. However, the process needed to be piloted to prove that cathode-grade quality could be consistently produced. A satisfactory method for gold recovery has been developed.
HBMS has done no further work on the process and is still finalizing its smelter modernization options. Indications are that pressure leaching will be chosen for zinc concentrates and that one of the converters will be adapted to a Noranda reactor in the copper smelter. Sulphur dioxide from the copper smelter will continue to be discharged through the stack.
In 1987, Great Central Mines Ltd. agreed to license Fenco Engineering Inc. of Toronto to design and construct a GCM process facility at Cananea, Mexico. A feasibility study was begun but nothing materialized due to ownership changes at Cananea.
The latest development is that the Canada Centre for Mineral and Energy Technology (Canmet) is conducting a feasibility into an adaptation of the process using cementation instead of electrowinning for copper recovery. The adaptation is seen as an inexpensive option for projects producing 5-10,000 mt/yr copper and has the advantage that most of the silver would be directly recovered in the cement. Such a plant would prove the leaching and regeneration aspects of the circuit on a large scale and make it easier to evaluate the full process at a later date.
Probably the most versatile and promising of the new pyrometallurgical routes is the Isasmelt process. It is applicable to an extraordinarily wide range of feedstocks and throughputs and is suitable for both small greenfield sites and for expansion of existing smelters. Other advantages include simple, low-cost plant with low energy requirements, minimal feed preparation, ability to produce a discardable slag without a slag cleaning step, volatile impurities that are readily eliminated, and emissions that are easily contained. Mt. Isa Mines will commission a 60,000 mt/yr lead bullion Isasmelt plant at the end of this year.
The Isasmelt process is based upon the Sirosmelt lance invented by researchers at Australia's Commonwealth Scientific and Industrial Research Organization (CSIRO). MIM Holdings Ltd., the licensee, refers to Isasmelt Technology as the technical innovations and knowhow developed over the last 10 years of its use of the Sirosmelt lance, including significant developments to the lance itself. Ausmelt Pty Ltd. is also a Sirosmelt licensee.
The Sirosmelt lance injects air or combustion gases into the molten bath creating very turbulent conditions as the gases rise through the slag layer and cascade over the surface. Internal swirlers enhance cooling of the lance by process air so that a protective slag layer freezes on the outside of the lance. Lances usually fail because of corrosion by matte but typically last several weeks before minor repairs are required. Metal or matte forms a layer at the bottom of the furnace and can be tapped as necessary.
These unique furnace conditions result in very high smelting rates, up to 1 mt/hr/[m.sup.3] of vessel volume, and very high fuel efficiency. Furthermore, bath conditions can be accurately and rapidly controlled by changing injection conditions or adding reactants.
CSIRO first applied the Sirosmelt lance to tin smelting and fuming but in the late 1970s also applied it to lead slag reduction, copper concentrate smelting, and reduction of copper anode furnace slag and converter slag (the latter two in conjunction with MIM).
In 1978, MIM and CSIRO initiated a project to investigate application of the lance to lead smelting. This investigation included crucible-scale testing at CSIRO and 120 kg/hr pilot-plant tests at Mt. Isa, and by 1982 had resulted in the development of the Isasmelt process, a two-stage, leadsmelting process.
In 1983, a 5-mt/hr plant was commissioned at Mt. Isa to demonstrate oxidation of lead concentrates to produce a high-lead slag. This slag was fed to the sinter plant as a partial replacement for return sinter, effectively increasing overall lead-smelter capacity. The plant is still primarily used for this purpose.
In 1985, a second Isasmelt reactor was added to demonstrate the full lead-smelting process. In the first stage, lead concentrate is added directly to a molten slag bath and oxidized by air injected down the lance. Simultaneously, high-lead slag from this furnace is transferred to a second furnace and reduced with coal. The second furnace is also used to recover lead bullion from copper drosses which had previously been shipped abroad.
The successful operation of the 5-mt/hr concentrate two-stage pilot plant was the basis for the new 60,000-mt/yr lead (20-mt/hr concentrate) Isasmelt plant designed to treat the increased concentrate production from MIM's new Hilton mine. Throughput at the existing sinter plant/blast furnace will be reduced to 150,000 mt/yr giving a total output of 210,000 mt/yr lead.
In the new plant, lead and discard slag are separated using a conventional forehearth arrangement rather than being tapped together into molds; waste-heat boilers are used rather than evaporative gas-coolers, decreasing water content of the off-gas and thus baghouse sizes, and producing high-pressure steam for power generation; and reduction coal is injected down the lance rather than dropped into the furnace.
In 1987, a 15-mt/hr Isasmelt reactor was commissioned at the Mt. Isa copper smelter. An Isasmelt furnace was chosen because of its low capital cost and ability to treat high rates of converter-slag concentrate of which there was a 10-year backlog. Copper-concentrate filter cake, dry concentrates, minus-25-mm lump coal, and a flux of crushed quartz and limestone are mixed, pelletized, and fed into the furnace through a port on the top.
The 10-m high by 2.1-m-dia cylindrical furnace is lined with chrome-magnesite bricks, the lower 3 m of which are backed by water-cooled copper jackets. The lance is made of 200-mm or 250-mm steel pipe and provides oxygen-enriched air to smelt the concentrate to the required matte grade. Slag and matte are tapped together into an existing reverberatory furnace for settling. Off-gases are cooled to 300 [degrees] C and passed through the existing electrostatic precipitator for dust removal before existing via the stack.
In April 1989, a 70-mt/d oxygen plant was commissioned which increased furnace capacity and reduced specific fuel consumption. Without oxygen enrichment, coal consumption averaged 0.68 mt/mt copper for a throughput of 15-17 mt/hr concentrate but with oxygen-enriched air furnace capacity increased to 24 mt/hr. A trial with supplemental liquid oxygen to allow oxygen input of up to 150 mt/d demonstrated the plant could treat 48 mt/hr concentrates for a coal burn of 0.09 mt/mt copper.
The Copper Isasmelt plant has increased the Mt. Isa smelter capacity by 20% or 30,000 mt/yr copper for a capital outlay of A$11 million, including the oxygen plant. Overall smelter operating costs have been reduced and there have been major improvements in plant hygiene and the operating environment. Capacity now exceeds mine capacity and the company has been buying concentrates from outside sources. Plans are well advanced for a 180,000 mt/yr copper smelting plant based on a single 110-mt/hr Isasmelt reactor to replace the existing fluosolids roaster and two reverberatory furnaces.
One of the advantages of hydrometallurgical processes is that they produce cathode-grade copper directly. Substitution of the conventional two-stage furnace/converter route by a single-stage process would reduce capital and operating costs by eliminating converters, and allow use of the entire calorific value of the concentrate for autogenous smelting. A single source of sulphur dioxide and the elimination of the cyclic nature of converter operation would also reduce investment and operating costs for acid plants (which must be designed for peak loading) and environmental-protection measures.
Direct reduction to blister in a single furnace is possible with very high-grade concentrates. The Outokumpu flash furnace at Olympic Dam in Australia is producing blister directly, but only because the chalcocite/bornite concentrate feed assays around 60% copper.
KHD Humboldt-Wedag's Contop process was originally designed for direct production of blister or high-grade matte. The process is now also seen as an efficient way of raising capacity and reducing operating costs at existing smelters. Contop cyclones are operating successfully at Palabora and Chuquicamata, and Asarco plans to rebuild its El Paso smelter using the technology.
By smelting at very high temperatures, the Contop Process reduces the problem of handling high-magnetite slags and volatalizes undesirable elements in the concentrate. Smelting takes place at 1,600-1,700[degrees]C with oxygen-enriched air in a cyclone reactor mounted over a furnace. The copper-bearing slag and matte produced fall into a slag bath which operates at 1,300-1,350[degrees]C.
The partial pressure of oxygen in the settling furnace is kept below 10[sup.-7] bar by top blowing with reducing agents to keep the magnetite level in the bath low, and to reduce the solubility and viscosity of copper in the slag. This aids settling and produces a low-copper slag.
The 150 mt/d Contop cyclone at Palabora is mounted over the existing reverbaratory furnace and was designed to run with hot air. Commissioned in 1988, it has smelted up to 180 mt/d on a regular basis. This was the first retrofit of a reverberatory furnace.
Contop technology was chosen in part because it offered a low-capital-cost solution for the small increase in smelting capacity required. Tankhouse capacity is 140,250 mt/yr copper but the long-term average smelting rate was 133,000 mt/yr. In recent times there have been prolonged periods when concentrates have been of lower than normal copper grade. Smelting capacity was thus a small but significant constraint on production levels of cathode-copper.
Other reasons for Contop selection were that it was easy to retrofit with minimal production interruption, offered the possibility of more efficient oxygen use, did not require lancing for slag cleaning, and matte grade could be controlled by varying the amount of coal fed to the cyclone. It has the further advantage of being able to be turned on and off as required and has an almost unlimited turndown ratio.
The 1.4-m-dia by 2.1 m cyclone reactor is mounted on an adaptor on top of the reverberatory furnace at the end of the green-feed charge banks. This is about 17 m from the reverb firing wall. Both adaptor and cyclone are water cooled which causes a frozen slag-layer to form on the walls, protecting them from erosion and chemical attack.
Silica and concentrate are drawn from storage bins and fed to a flash drier. Hot gas at 500-700[degrees]C, produced by coal burning, joins the concentrate in a pulverizing mixer before it enters the flash drier where the wet material is dried in the hot air stream. The drier operates at 10 mt/hr dry concentrate-silica mix. Coal consumption is 50 kg/hr. Exit gases are de-dusted in a cyclone and subsequently in a bag filter. Solids from the cyclone and filter are pneumatically conveyed to the smelting cyclone storage bin.
Dry mix from the storage bins is divided into three and routed to three small mixing chambers where additional coal can be added if required. The three streams are pneumatically transported to a common charging-pipe in the middle of the cyclone lid. Maximum charging rate is 10.5 mt/hr. After being metered, coal is blown tangentially into the three cyclone-burners with hot combustion air at 225[degrees]C.
In general, 7 mt/hr concentrate with 0.5 mt/hr silica flux is smelted using around 13% coal. No noticeable changes in slag losses and matte grades have been registered. Overall matte grade varies between 45-55% copper and cyclone matte grade has been calculated at 58-60% copper based on samples of unsettled primary slag at the cyclone outlet.
Introduction of the Contop cyclone has allowed Palabora to operate its tankhouse to capacity.
The 500-mt/d Contop at Chuquicamata is also mounted over the reverberatory and was designed for use with more than 70% oxygen enrichment. However, high temperatures produced by the autogenous reaction of the high-grade concentrates caused cooling problems with the cyclone shell. These are being solved by changing the upper part of the cyclone shell from double-wall construction to high-pressure water tubes for more efficient cooling.
Asarco plans to modernize its El Paso smelter using Contop technology. The reverberatory operation is to be substituted by two 516-mt/d concentrate Contop cyclones. The complete $30 million modernization program also includes replacement of existing roasters with a dryer and is expected to lead to operating cost savings of 17%.
FLAME CYCLONE REACTOR
Another high-intensity smelting process aimed at eliminating volatile impurities is the Flame Cyclone Reactor (FCR). Lurgi GmbH's Cyclomelt technology using this reactor has been demonstrated at a 10-mt/hr demonstration plant at Norddeutsche Affinerie's Hamburg smelter since 1982. Over 12,000 mt of concentrates have been treated. In the last two years further testwork has been performed with concentrates containing high contents of impurities such as arsenic, lead and zinc, and with relatively pure copper concentrates. These tests have shown Cyclomelt to be an economic alternative to existing pyrometallurgical processes for both impure and clean copper concentrates.
The main advantages are high matte-grades of 75% copper, increased downstream-converter capacity, high oxygen enrichment of 70% that results in high SO[sup.2] concentration in the off-gases, a totally closed system that gives no emissions, and no refractories in the smelting unit increases reactor availability.
In the Flame Cyclone Reactor, dry, fine feed predominantly less than 50 micron in size is burned in a special burner with a high degree of oxygen enrichment of the combustion air. Additional fuel may be added if the calorific value of the feed is insufficient and the gas atmosphere can be adjusted by injection of reducing agents. Oxidation, reduction, and volatilization are carried out in a small-diameter, vertical reaction shaft. Separation of the molten metal, and slag from the off-gases takes place in horizontal cyclone chamber. The reaction shaft and cyclone chamber are lined with tube panels fed by high-pressure water that provide intensive cooling and cause a protective layer of solidified melt to form on the inside wall.
Capital and operating costs of a smelter using Cyclomelt technology are estimated to be 20% less than those of a flash smelter.
The new high-intensity smelting processes developed have done much to improve the efficiency of pyrometallurgical processes, diminishing the relative cost and environmental advantages of rival hydrometallurgical routes. However, hydrometallurgical treatment of sulphide concentrates can will offer advantages for smaller operations, complex dirty concentrates, and mines in locations where marketing of sulphuric-acid byproduct would be extremely difficult. Hydrometallurgical plants could also increase the versatility and flexibility of existing smelters. It remains to be seen who will put up a commercial plant using either of the two hydrometallurgical processes described.
PHOTO : Fig. 2 - Flowsheet or the Great Central Mines (GCM) process modified for Hudson Bay Mining & Smelting (HBMS) smelter, Flin Flon.
PHOTO : Fig. 3 - Flowsheet for the 60,000 mt/yr Isasmelt lant roducing lead.
PHOTO : Fig. 4 - An Isasmelt furnace was chosen at Mt. Isa because of its low capital cost and ability to treat high rates of converter-slag concentrate of which there was a 10-year backlog.
PHOTO : Fig. 5 - The Contop Cyclone, placed on top of existing furnaces.
PHOTO : Fig. 6 - Lurgi's Cyclomelt technology uses a flame cyclone reactor. References
Coulter, M.D., Fountain, C.R., "The Isasmelt Process for Copper Smelting," Non-Ferrous Smelting Symposium - 100 Years of Lead Smelting at Port Pirie, AUSIMM, Sept, 1989. Craigen, W., Barlin, B., Krysa, B., "A Technical and Economic Evaluation of Modernization Alternatives for a Canadian Copper Smelter," Copper 87 Conference, Vol. 4. Hilbrans, H., Melcher, G., "Direct Copper Recovery from Sulphide Concentrates by the Contop Process," Copper 87 Conference, Vol. 4. Nepper, J., Ruehl, B., Emicke, E., "The Flame Cyclone Reactor," Copper 87 Conference, Vol. 4. Smith, A., Immelman, P.D., Chaudhuri, K.B., "Incremental Increase of Reverberatory Furnace Smelting Capacity at Palabora Using Contop Technology," TMS Meeting, Las Vegas, Feb. 1989. Suttill Keith R., "Solvent Extraction - A Key in Maintaining Copper Production." E&MJ, Sept. 1989. Wyllie, Robert J.M., "Flash Smelting at 40." E&MJ, Oct. 1989.
|Printer friendly Cite/link Email Feedback|
|Author:||Suttill, Keith R.|
|Publication:||E&MJ - Engineering & Mining Journal|
|Date:||May 1, 1990|
|Previous Article:||New realities, the globalization of exploration and mining.|
|Next Article:||Selecting the right shield support.|