Table of Contents
- Basic Process Reaction and Design Parameters
- High Selectivity of LIX 64n for Copper
- Stripping Section — Two-Stage Mixing-Settling
- Leach Liquor Throughput Determines SX Solvent Extraction Capital Cost
- Why Mixer-Settler Units are Preferred Contactors
- Choosing the Number of Extraction Stages
- Maximum Stripping Efficiency is Essential
- How to Minimize Losses of Organic Solvent
- Coalescence Beds versus Flotation Units
- Acid Consumption & Overall Operating Costs
Development of Suitable Reagents was Vital Key: Solvent extraction techniques were first applied to the recovery of metals for atomic energy projects in the decade 1945-55. They were subsequently applied to the large scale production of such refractory metals as hafnium and zirconium. Since then the technique has been applied to the recovery of the rarer metals including columbium, tantalum, and rare earths.
In the early 1960’s General Mills Chemicals, Inc. developed a solvent designated LIX 63, for the extraction of copper from splutions of pH greater than 4.0. It was thus very suitable for treating the product from the ammoniacal leaching of copper. It could not be applied to the much more common acidic copper leach solution, the pH values of which are in the range 1 to 3. In 1965, however, General Mills announced a new solvent, LIX 64, which had been developed to fill this latter need. Up to that time, there was no alternative but to treat acid leach liquors by cementation techniques and although these had been much improved, the product still required final treatment in a smelter. What solvent extraction offered was the opportunity to produce cathode copper directly from the leach liquor on the same site as the leaching operation. This gave advantages to both the large producer who would not have to consider extra smelting capacity when extending his leach operations and to the small producer who would be able to produce a finished grade of copper.
The development of LIX 64 and its improved successor LIX 64N has led to the application of the process in several locations in the United States.
The world’s first commercial scale copper solvent extraction plant was that of Ranchers Development and Exploration Corporation, Miami, Arizona. Bagdad Copper Corporation, Bagdad, Arizona, has recently commissioned a plant which will treat 2,750 gallons per minute of leach liquor.
In addition to these, pilot plants have been operated by Duval Corporation, Esperanza, Arizona; and Inspiration Consolidated Copper Company, Inspiration, Arizona.
Plans have now been announced for the construction of the largest solvent extraction plant in the world to be designed by The Power-Gas Corporation for Nchanga Consolidated Copper Mines Ltd, at Chingola in Zambia. This plant will have a throughput of 12,000 gallons per minute of leach liquor with a copper content in excess of 60,000 annual tons.
It can be justly said therefore that solvent extraction now stands alongside traditional methods as a major copper winning technique.
Basic Process Reaction and Design Parameters
The extraction of copper from aqueous solution by the application of organic solvents such as LIX 64N or the Ashland Chemical products Kelex 100 and 120 proceeds chemically, according to the following equation:
(2RH) org. + CuSO4 aq. = (R2Cu) org. + H2SO4 aq.
It will be seen that the extraction of copper into the organic phase is accompanied by an increase in the acid level of the aqueous phase. This fact must be borne in mind when considering the type of liquor which can be treated effectively by these solvents.
It may be said that LIX 64N is a very effective extractant for copper from solutions with a pH in the range 1.5 to 2.8. Below a pH of 1.5 the loading capacity falls off quite sharply and at a pH of 0.5 it is about one-fifth of that at 1.5. Provided the initial pH of the aqueous solution is sufficiently high it is possible to achieve high percentage extractions even for copper contents as high as 30 grams per liter.
In the normal range of leach solutions which contain 2 to 3. grams per liter Cu and have a pH in the region of 2.0 to 2.5, each 1 percent by volume of LIX 64N, dissolved in kerosene carrier to form the organic phase, will pick up at least 0.25 grams per liter of copper from the aqueous feed, when operating with an organic/ aqueous ratio of 1:1. This is the fundamental design parameter for determining the concentration of LIX to be used in the organic phase. Because the organic phase returning from the stripping section is never completely stripped of copper, it is necessary to add a small amount of extra LIX to maintain the free LIX at a satisfactory level.
High Selectivity of LIX 64n for Copper
The selectivity of LIX 64N for copper is very high in comparison with that for the metals with which it is commonly associated. The only metal ion picked up to a measurable extent is ferric iron. Even so, the ratio of copper to ferric iron extracted is of the order of several hundreds to one. The level of ferric iron in the feed liquor is not as important as the pH of the solution and the concentration of LIX 64N in the organic phase. The latter is important because the extraction system operates in a counter-current fashion. Thus the strongest, unloaded
FLOWSHEET of a combined solvent extraction and metal electrowinning plant. Copper passes from the aqueous leach liquor phase to the organic phase in the extraction section and back from the organic to the aqueous electrowinning plant feed in the stripping section.
LIX 64N solution comes into contact first of all with the aqueous solution most depleted in copper and therefore having a higher ferric iron/copper ratio and lower pH than at any other part of the system. Although the distribution coefficient for ferric iron between organic and aqueous phases is lower at lower pH, the net effect is to pick up iron in this extraction stage. Once picked up, the iron remains in the solution until the stripping stage of the process. Because of these conflicting effects it is impossible to formulate a general rule for the amount of ferric iron pick-up in any particular case. It may be said, however, that the degree of iron pick-up will be of the order of only 10 to 20 parts per million in the organic phase (compared with 1,000 to 3,000 parts per million of copper), for a typical leach liquor having a copper concentration in the range 1 to 3 grams per liter, a ferric iron content in the same range and a pH of 1.8 to 2.8.
It is preferable to limit the ferric iron content of the electrolyte feed to an electrowinning process to about 2 grams per liter. Because all the iron picked up in the extraction stage is stripped from the organic phase by the spent electrolyte it is necessary to take a bleed from the electrolyte to maintain the iron content of the electrolyte at the preferred level. The bleed flowrate required is usually less than 1 percent of the initial leach liquor flowrate. For many applications, the bleed passes back into the leaching operation so that only a very small fraction of its copper and acid content is lost.
The acid used in the leaching of the copper values is regenerated allowing the raffinate to return to the leaching operation with a high useful acid content. In addition it does not have the high iron content of returned cementation liquor which may produce an iron “capping-off” effect in the dump being leached.
Fig. No. 1 is a typical flowsheet where the leach liquor is pumped from the leaching circuit through polishing filters into the extraction section of the solvent extraction plant. The filters may or may not be necessary according to the quality of the liquor leaving the leach plant.
The extraction section of the solvent extraction plant consists of a number of mixer-settlers in series. The
aqueous leach liquor flows through these counter-current to an organic stream which is made up of LIX 64N dissolved in a kerosene carrier, and the copper content of the leach liquor is transferred into the organic phase. In order to provide the necessary head to drive the two solutions through the mixer-settlers, the mixing impellers are designed with pumping characteristics.
It is usual to operate the extraction mixer-settlers with an organic/aqueous ratio of 1:1. At higher ratios, lower concentrations of extractant could be used but the overall solvent inventory of the plant would be increased, thus increasing the size of most of the equipment. At lower ratios, the higher organic extractant concentrations which would be required would aggravate iron pick-up problems.
It is preferable to operate the final stage extraction mixer-settler with an organic continuous dispersion so that losses of organic material are reduced to a minimum. There is, however, always some dispersed organic, material in the aqueous raffinate.
The aqueous raffinate leaving the final stage mixer-settler is pumped into a hold-up tank, where any accidental massive carryover of organic material can be caught.
The raffinate then flows to the solvent recovery plant where flotation units or coalescing beds separate the greater part of the dispersed organic material and return it to the extraction circuit. The aqueous stream is then returned to the leach operation.
Stripping Section — Two-Stage Mixing-Settling
The organic solution, loaded with copper, leaves the first stage extraction settler and is pumped to the stripping section of the plant. This usually consists of two mixer-settlers through which the loaded organic solution and the aqueous electrolyte flow countercurrently. The electrolyte has a free H2SO4 content in the range 150 to 200 grams per liter and is loaded to about 40 to 50 grams per liter of copper in the stripping plant. Depending on the preferred method of operation of the electrowinning plant the copper drop can vary from 2 grams per liter to 20 or even 30 grams per liter. If a very low copper drop is chosen, the flow of electrolyte is high and the organic/aqueous overall flow ratio in the stripping section is low, with the result that the solvent extraction equipment is larger. In general this method of operation is not preferred, even though very stable operation of the mixer-settlers is possible with high aqueous throughput, because high flows can produce high organic carryover to the cells impairing cathode quality. The usual copper drop is of the order of about 10 grams per liter, operating in the range 30 to 40 grams per liter. This gives an organic/aqueous flow ratio of about 5:1. It is usual to operate the first stripping mixer-settler with an organic continuous dispersion to reduce carryover of organic material into the aqueous phase. Overall ratios greater than 5:1 are required if higher copper drops are used and these conditions can be achieved by including an aqueous recycle stream which stabilizes mixer-settler operation while maintaining overall stripping characteristics.
The organic stream stripped of the major part of the copper, leaves the second stage stripping mixer-settler and passes back to-the extraction section. The aqueous, loaded electrolyte leaving the first stage stripping settler passes first to a surge tank where initial separation of entrained organic material takes place and then to the solvent recovery section where flotation units or coalescence beds are used to separate the bulk of the remaining entrained organic solvent. It then enters the electrowinning plant via
a heat exchanger where it is heated by the spent electrolyte which itself is returning to the second stage stripping mixer-settler.
Leach Liquor Throughput Determines SX Solvent Extraction Capital Cost
The capital cost of a solvent extraction plant depends on the throughput of leach liquor and not simply on the copper production rate. Higher concentrations of copper in the liquor are dealt with in the extraction section by increasing the concentration of the organic extractant in the organic phase while maintaining the same overall organic/aqueous ratio. Because the size of each mixer-settler is proportional to the combined flow of aqueous and organic streams through it, it is therefore proportional to the aqueous leach liquor flowrate.
As has been described earlier, it is usual to operate the stripping stage with a fixed organic/aqueous ratio to achieve operational stability and high stage efficiencies in the stripping mixer-settlers. The size of the stripping mixer-settlers is therefore also proportional to the organic flowrate and hence to the leach liquor flow.
Fig. No. 2 shows the typical variation in capital cost
of solvent extraction plants as a function of leach liquor throughput. The figures are for a three-stage extraction two-stage stripping plant with full solvent recovery facilities, but including no part of the electrowinning plant. The lower line shows the cost of equipment and initial solution inventory only, while the upper line represents a typical range for total costs including supply of equipment, engineering, erection and civil works for a green field site.
Assuming a copper concentration in the leach solution of 2 grams per liter, the installation cost/assumed ton of copper produced is about $130.00 in the upper ranges of capacity and even at flows of 1,000 gallons per minute this only rises to about $260.00. Based on f.o.b. equipment and initial fill of chemicals only, the figures are $85.00 and $130.00 respectively.
Why Mixer-Settler Units are Preferred Contactors
For the larger plants, the mixer settler units themselves, with their initial fill of organic solution contribute almost 40 percent of the f.o.b. value of the plant and, of course, account for a high proportion of the erection and civil costs for the total plant.
It may therefore be asked why mixer-settlers are used for this duty. Solvent extraction techniques are used in a wide variety of industries other than hydrometallurgy and a large range of extraction equipment has been developed over the years, many types having apparent advantages over mixer-settlers from the point of view of both solvent inventory and plot area. The short answer is that no other type of contactor has been shown to be as reliable, or as well able to cope with upset process conditions as the simple mixer-settler. There are fundamental differences between mixer-settlers and most column contactors which render the former applicable very widely and the latter usually suitable for special conditions only.
In analysing systems in the laboratory the possibility of solids build up must be examined. The source of these solids may be (a) bacteria which can be either air-borne, or leach liquor borne, and whose breeding is promoted by the favorable conditions prevailing at the organic/aqueous interface, (b) the spores of fungi such as Cladosporium Resinae which are frequently present in the kerosene carrier, and which also find conditions suitable for growth at the phase interface and (c) very fine solid particles entering the plant in the leach liquor. The fungal growth acts as a breeding ground for the bacteria and as a net to trap solid material.
Mixer-settlers are more suitable than column contactors for the study of the effects of the build-up of solids because they lend themselves to the separate study of mixing, mass transfer, settling and separation of the phases and the effect of solids on each operation can be examined. Also the design of a mixer-settler makes it easy to remove solids from the interface should any build-up occur.
Choosing the Number of Extraction Stages
A number of factors in combination are used to determine the number of mixer-settler extraction stages required for a specific duty. Fig. No. 3 shows a typical McCabe-Thiele diagram for the extraction of 2.35 grams per liter of copper in five extraction stages from a solution at a pH of 2.0, using an 11 percent LIX solution. The organic solution returning from the strip system has a copper content of 0.2 grams per liter. From the isotherm, it is possible to calculate the theoretical overall extraction efficiencies for various numbers of stages. These are:
for a three-stage system 94.9 percent
for a four-stage system 97.0 percent
for a five-stage system 97.8 percent
Assuming an overall extraction efficiency of say 95 per-cent compared with what is theoretically possible the actual extraction is:
for a three-stage system 90.0 percent
for a four-stage system 92.5 percent
for a five-stage system 93.0 percent
Fig. No. 4 shows the variation in capital cost for a plant treating 6,000 gallons per minute, of leach liquor containing 2.35 grams per liter copper, depending on whether the plant has three, four or five stages of extraction. The capital cost including civil works and erection per annual ton of copper recovered, taking account of the increased efficiency with the higher number of stages is as follows:
three stages $151.00
four stages $156.00
five stages $165.00
The extra operating costs involved in running additional stages are minimal, being those of maintenance and of the power required in the mixers. For the capacity under consideration this can be taken as 75 kilowatts for each extra mixer.
The final decision on whether to include extra stages or not will hinge partly on the economics of the preceding leach operation, but if a very high proportion of the raffinate has to be purged from the leach circuit the economics will favor maximum copper recovery in the solvent extraction plant.
Maximum Stripping Efficiency is Essential
The efficiency of the stripping section has an important bearing on the extraction efficiency due to the level of copper returning in the stripped organic solution. Fig. No. 5 shows the matched stripping isotherm for the same system as that for the extraction isotherm shown in Fig. No. 3.
The operating line has been drawn for an organic/aqueous ratio of 6.25 and it can be seen from the diagram that one stage will reduce the Cu++ content of the organic phase to 0.14 grams per liter and two stages to 0.1, this latter representing very nearly 100 percent stripping. Assuming an overall efficiency of 95 percent of that theoretically possible the actual copper content of the organic phase leaving the strip circuit is 0.2 grams per liter. The effect of this copper content on the extraction efficiency is given in Fig. No. 6 which shows the percent extraction as a function of the copper content of the organic phase returned from the stripping section. It will be seen that the stripping efficiency can have a most significant effect on extraction efficiency and it is imperative to design the stripping mixer settlers for maximum stage efficiency. Using specially developed techniques, Power Gas Corporation has improved overall stripping to 95 percent, where many other systems can only achieve 90 percent or less, making it possible in the case described, to achieve a copper content in the stripped organic of 0.2 grams per liter, which enables a high extraction efficiency to be maintained.
As the addition of extra extraction stages increases the capital and running costs to only a small extent, it is often possible to aim for maximum possible extraction. It is possible to achieve higher degrees of extraction by increasing the concentration of LIX 64N in the circulating organic phase, and it has been suggested that inefficiencies in the stripping section can be overcome in this way, by providing a higher concentration of unloaded LIX in
the extraction stages. However, operation in this way can increase the ferric iron pick-up in the extraction section. This in turn leads to increased acid and copper losses in the electrolyte bleed and nullifies any advantage to be gained from extra copper pick-up.
How to Minimize Losses of Organic Solvent
The removal of droplets of organic phase from the aqueous raffinate and the loaded electrolyte is desirable not only economically but also because the presence of organic material in the aqueous streams can be harmful in subsequent processing steps such as electrowinning.
It is difficult to distinguish between entrainment and solubility because some of the droplets are extremely small, but work carried out by Power-Gas indicated that true total organic soluble losses are less than 20 parts per million, compared with a typical total organic content of the aqueous streams of about 100 parts per million.
Careful design of the mixer settler units can substantially lower organic loss by entrainment and special attention must be given to the design of the mixing unit. Mixing must be sufficient to produce an adequate mass transfer area without at the same time giving rise to a high proportion of very fine droplets. Mixers designed by Power-Gas produce very low rates of shear while maintaining adequate pumping head for the transfer of the two phases through the system. In the settler itself, sufficient area must be provided to allow for the maximum degree of coalescence while minimizing solvent inventory as far as is feasible. It is possible to design a last-stage settler conservatively relative to the other stages as a certain degree of entrainment in intermediate stages does not affect the performance of the overall unit. However well the mixer-settler is designed there will nevertheless always be some entrainment and equipment for the recovery of some of the organic material from the aqueous raffinate is generally installed.
Coalescence Beds versus Flotation Units
Coalescence beds, which consist of high surface area packed filaments through which the aqueous streams flow, are very effective in promoting coalescence of the entrained droplets and, because there is no power requirement for this type of equipment, operating costs can be very low. However, the interstices in the packing need to be so small that solids in the stream are trapped causing an increase in the pressure drop across the bed leading to eventual failure either by channelling or blockage. Replacement of the beds is then necessary. If the solids content is high, the cost of renewing the filters necessary is also high.
Flotation units have been shown to produce good recovery rates from typical raffinates and electrolytes and are not prone to blocking by solids. In this, they have an advantage over coalescence beds but on the other hand, they have the disadvantage of requiring power. Very good recovery rates are possible even with initial organic contents as high as 1,000 parts per million. Optimisation curves shown in Fig. No. 7 have been produced which indicates that the optimum flotation time is about 2.5 minutes for initial organic entrainments of 50 to 100 parts per million and slightly longer for entrainments of 1,000 parts per million. It may be possible to use shorter flotation times to recover entrained organic material from the highly acidic condition occurring in the advance electrolyte solution.
Acid Consumption & Overall Operating Costs
The only significant acid loss from the system is that in the iron control bleed. Typically, this figure will be about 0.01 gallon of 150 grams per liter acid per gallon of feed liquor. Assuming a copper content of 2 grams per liter in the feed liquor the acid consumption/ton of copper treated is 0. 75 tons i.e. a cost of approximately $29.00 per ton copper. A high proportion of the acid in the bleed can be re-used in the leaching operation, reducing the net cost of acid in a typical operation to only $9 to $10 per ton copper.
Since it is common for the copper solvent extraction plant to supply feed electrolyte to an electrowinning plant, it is useful to combine the operating costs for the two plants when comparing costs of copper recovery by solvent extraction techniques with those for cementation. These have been summarized in the table which shows typical operating costs for a plant producing 30,000 annual tons of copper. These can be compared with figures given recently by McGarr et al.
Assuming total labor costs of about $24 per ton of copper produced, the overall operating costs of solvent extraction plus electrowinning can be as low as $72 per ton of copper. For cement copper production freight and smelting charges alone usually exceed this figure. In addition, acid consumption in the cementation process exceeds that in solvent extraction/electrowinning by about 1.5 tons for every ton of copper extracted and it is possible for the consumption of iron to be as high as 4 tons per ton of copper.
The authors wish to acknowledge the assistance of colleagues in the R and D Division of The Power-Gas Corporation Ltd in the preparation of this paper and also their gratitude to the directors of The Power-Gas Corporation for their permission to publish.
- Throughout this article gallons refer to Imperial gallons. 1.0 Imperial gallon = 1.2 U.S. gallons.