An increasing demand for metals in general, and higher purity metals in particular, decreasing ore grades and more stringent environmental regulations have driven, and will continue to drive, research into finding more effective and efficient methods for processing the ores available to us, and recycling previously used metals. Hydrometallurgy has provided, and will continue to provide, many of the new processes and solvent extraction technology will certainly play an important role in many of these new processes.

Solvent extraction as a simplified technique to purify and recover a metal goes back to at least 1842 when Peligot extracted uranyl nitrate into ethyl ether.

The chemical literature has hundreds of references to the use of solvent extraction as a technique in analytical chemistry and a large amount of fundamental knowledge in solvent extraction, particularly in the areas of solution chemistry and organic based extractants, comes from analytical chemistry.

Solvent extraction technology has been applied on a commercial scale to the recovery of uranium, vanadium and molybdenum for about forty years and to copper and nickel for almost thirty years. For this same length of time Henkel Corporation, (formerly General Mills Chemicals, Inc.), has been committed to the development of solvent extraction technology in general, and solvent extraction reagents in particular. This commitment has resulted in the development and successful marketing of the LIX, Alamine and Aliquat Reagents used by numerous, commercially successful metal recovery operations around the world. These reagents are noted for the wide range of conditions in which they operate and for their pollution free flowsheets. It is within the realm of technical feasibility that one day the hydrometallurgist could have an unlimited selection of commercially available chemical reagents at his disposal in order to make efficient and economic separations of virtually all metals in solution.

It is important for the reader to understand that solvent extraction is only one unit process in a series of unit processes needed to win metal from ore. For this reason solvent extraction must be compatible with, and complimentary to, the leaching process which precedes it and the final metal recovery process which follows (Figure 1). The success of the whole operation then is dependent of the success of each individual unit process.

This booklet is designed to familiarize the reader with some of the basic concepts of solvent extraction technology as well as with the LIX, Alamine and Aliquat Reagents and systems marketed by Henkel Corporation. More detailed information about specific reagents, metals, or process systems is available on request.

LIX , Alamine and Aliquat are registered trademarks of Henkel Group.

Description of the Solvent Extraction Process

Solvent extraction (SX) is applicable in any instance where it is desirable to selectively remove or extract a species from one solution into another. This can apply either to the removal of a valuable component from contaminants or to the removal of contaminants from the valuable component. The solution originally containing the desired species and the solution into which this species or the contaminants are to be extracted must be immiscible to effect separation.

In metal recovery operations the valuable component is normally a metal ion or a metal ion complex contained in an aqueous solution. This aqueous solution is mixed with an immiscible organic phase containing the active extractant at which time the active extractant transfers the desired component from the aqueous phase into the organic phase. The aqueous/organic mixture, which is called a dispersion, then passes from the mixer to a settler where the phases disengage. The “loaded organic” phase, now containing the extracted metal, is then transferred from the extraction section of the SX circuit to the stripping section where the extracted metal is stripped from the organic phase. The “stripped organic” is then recycled back to extraction. In most instances, stripping the extracted species from the loaded organic is accomplished by mixing the loaded organic with an aqueous solution. After the stripping operation final metal recovery takes place from the aqueous solution. The recovery of metal directly from the loaded organic phase is also possible, but, currently not commercially in use. Because the solvent extraction process resembles metal extraction with solid ion exchange resins, it is sometimes called liquid ion exchange.

Figure 2 shows a typical counter-current, mixer-settler solvent extraction unit. An individual mixer-settler is known as a stage. In some systems specific problems make it desirable to add one or more organic “wash” or “scrub” stages in order to prevent the transfer of deleterious species from extraction to strip or vice versa.

Solvent extraction as part of an overall metal recovery process has three main objectives:

1. The purification of a metal(s) from unwanted impurities either by extracting the desired metal(s) from the impurities or by extracting the impurities from the desired metal(s).
2. The concentration of metal in order to reduce downstream processing costs.
3. The conversion of the metal values to a form which simplifies eventual recovery.

In any given SX process, one, two or all three objectives may be accomplished.

These objectives can usually be met by choosing the proper reagent and operating under the optimum conditions. A variable number of mixer-settler stages, adjustable flow rates and the ability to use wash stages allows the optimization of operating conditions as well as design flexibility.

Metal Species Extractable by SX Solvent Extraction

If a metal is to be extracted from impurities or vice versa it is important to know what species are present in solution. With respect to metals, the extractable species can be divided into four categories:

1. Metal cations such as Cu2+, Ni2+, and Co2+
2. Complex metal anions, for example UO2(SO4)3, and Mo8O26 4-
3. Complex metal cations such as MoO2+
4. Neutral metal specials like UO2(NO3)2

The examples above represent metal species which are recovered by solvent extraction in commercial metal recovery plants. There are many others.

SX Solvent Extraction Reagents

It is obvious that if a metal species is to be transferred from an aqueous leach solution into an organic solution, there must be some chemical interaction which causes this to happen. The component in the organic phase which chemically interacts with the metal is properly called the extractant, but, commonly called the reagent.

At the present time there are organic extractants known for virtually all metals in one form or another. As stated previously, this body of knowledge is due in large part to analytical chemists, however, the requirements for a successful extractant in analytical chemistry are much different than the requirements for a reagent to be successful in large scale metal recovery operations, especially as they relate to process continuity and economics. In order for a reagent to be commercially successful it must:

a. Extract the desired metal(s) selectively from the metal-containing solution.
b. Be strippable into a solution from which eventual metal recovery can take place.
c. Be stable to the circuit conditions so it can be recycled many times.
d. Be nonflammable, nontoxic, noncarcinogenic, etc.
e. Be soluble in an inexpensive organic diluent or be able to function as the diluent.
f. Load and strip metal at a rate fast enough to allow for the use of economical mixing times.
g. Not promote stable emulsions.
h. Not transfer deleterious species from strip to extraction.
i. Have an acceptable cost.

Normally, reagent behavior with respect to the above list is not a black and white, pass or fail situation. No one reagent is the best with respect to all of the properties in the list; rather, successful reagents possess a good balance of all of the properties in the list.

There are misconceptions about the selectivity of reagents which merit discussion. It should be realized that no reagent is selective for only one metal under all conditions, but, that many reagents are selective for only one metal under certain conditions.

Selectivity is dependent upon conditions and the challenge is to match the conditions presented by a given leach solution to the selectivity characteristics of the reagents available. A perfect match is not always possible and is, in fact, rare. Most of the time, the researcher either settles for a reagent which, even though not perfect, functions well, or he tries to alter the leach procedure to give a leach solution from which a reagent will be more selective. In order to do this successfully, the chemistry of extraction processes must be understood.

Types of Extractants

There are five classes of metal extractants as characterized by structure, extraction mechanism and the metal species extracted: chelation, organic acids, ligand substitution, neutral or solvating and ion pairing. Henkel markets chelation reagents and ion pairing reagents and has research interests in ligand substitution reagents.

Chelation Extractants

A. General Comments

Chelation refers to “claw” which is a graphic description of the way in which the organic extractant binds the metal ion, in that, the extractant chemically bonds to the metal in two places in a manner similar to holding an object between the ends of the thumb and the index finger. Upon bonding with the metal ion the extractant releases a hydrogen ion into the aqueous solution from which the metal was extracted. The simplified equation shown below is for the extraction of a +2 metal cation:

One of the important parameters controlling the equilibrium position of this reaction is the acid content of the aqueous phase. A graphic representation of this behavior is referred to as a pH isotherm, with some typical pH isotherms for the reagent LIX 84-I shown in Figure 3. These pH isotherms can be used to predict the extraction characteristics of the reagent with respect to the metals shown under a variety of conditions. For example, at pH 2.0 copper(II) is strongly extracted, ferric iron is slightly extracted, while nickel(II) and cobalt(II) are not extracted. However, at pH 5 all four of these metals would be strongly extracted, except that this fact is not always important for iron(III) since it is only slightly soluble at pH 5.

It is important to realize that in order to make a direct comparison of the pH isotherms of two different reagents with the same metal, or of two different metals with the same reagent, the pH isotherms must be determined under exactly the same conditions.

A tabular method to express these extraction characteristics is shown in Table I.

The investigator screens the reagent against the metals most likely to be found in the aqueous feed solutions of interest under conditions similar to those expected in the solvent extraction process. This is normally referred to as selectivity data and holds only at the conditions under which it is determined. These conditions include the metal and reagent concentration, the oxidation state of the metal, the pH of the aqueous feed solution, and the organic-aqueous contact time. From Table I it is obvious that Cu(II) is the only metal listed which is strongly extracted by LIX 84 at pH = 2, and that only Fe(III) would present any potential problems to the preparation of a pure copper solution from a “normal” feed at pH = 2. This data is in agreement with the pH isotherms shown in Figure 3.

The extraction of metals from ammoniacal solution is also an area of interest, thus, the extractive behavior of some metals from ammoniacal solutions has been examined. In general, these data are expressed as ammonia dependent extraction isotherms, several of which are shown for LIX 84-I in Figure 4. The tendency for extraction to decrease with increasing ammonia concentration can readily be explained by the extraction stripping equation for a typical +2 metal cation shown below:

In essence, the ammonia in solution functions as a coordinating ligand and competes with the organic extractant for the metal ion. The reaction is reversible and stripping with ammonia might be practical in some situations. The number of ammonia molecules coordinated to the metal will vary depending on the conditions; four being a favored number.

By looking at extraction isotherms and selectivity data, it is possible to develop metal separations schemes. For example, a look at the pH isotherms of LIX 84-I with Cu(II) and Ni(II) (Figure 3) suggests that a good Cu-Ni separation can be made with LIX 84-I by controlling the pH in the extraction stages. All of the copper can be extracted at low pH (1.5 to 2.0) before the system is adjusted with ammonia to higher pH (9.0 to 9.5) where the nickel is extracted. This type of scheme has been used on a commercial scale at the S.E.C. Corporation in El Paso, Texas, for the recovery of copper and nickel from a refinery bleed stream.

The LIX 84-I ammonia dependent extraction curves in Figure 4 show that copper(II) is more strongly extracted from an ammoniacal solution than is nickel. Thus copper-nickel separations from ammonia leach solutions should be possible and indeed have been studied for a similar reagent. The study concentrated on a flow scheme where copper is preferentially extracted from an ammonia solution containing copper and nickel before the nickel is extracted. A high copper/nickel ratio on the copper loaded organic is obtained by using an extra extraction stage in which copper crowds nickel from the loaded organic, i.e., the copper in an ammoniacal copper and nickel containing solution, when contacted with an organic loaded with nickel, will replace some of the loaded nickel. After the copper and nickel are loaded in separate extraction operations, each “metal loaded organic” is washed and then stripped to give a copper-rich and a nickel-rich electrolyte, respectively. Circuit complexity, overall operational ease and the number of stages required are dependent upon the aqueous feed.

A different separation scheme on the same type of feed has been reported and has also been studied in this laboratory. This scheme involves the co-extraction of nickel and copper with the same organic stream, then washing followed by pH controlled, selective stripping of each metal, to give copper-rich and nickel-rich electrolytes, respectively. This flow scheme is based on the ammonia isotherms, which show that copper and nickel are strongly extracted from ammoniacal solutions by LIX 84-I, and the pH isotherms, which show that while LIX 84-I loads copper very well at a pH of 2.5 to 3.0, nickel is readily stripped at that pH.

It is well indeed, that the study of pH isotherms, selectivity data and ammonia isotherms leads to the conceptual development of metal separation processes. However, this type of data does not tell in a detailed way how to do a given metal separation, nor the way the circuit must be set up in order to do the separation in the most effective and least costly manner. For this type of information some research effort is required.

Laboratory Evaluation Program for a Copper Leach Solution

Assuming a metal containing solution is available, what is the next step? The first thing is that the contents of the solution must be known, (metals, pH, NH3, etc.) as well as whether that solution is typical of the feed solution to be treated. The evaluation of a feed solution which is not typical of the solution to be treated can at times be a wasted effort. Once the solution contents are known, judgments about which metals can be separated should be made based on the extraction isotherms available. If data on some of the contained metals is not available, the reagent(s) to be used should be screened against these metals. At this point it should be known which metals are considered desirable to recover, and which metals are possible to recover via solvent extraction, under the given set of conditions. The circuit development work can now begin with the realization that all parts of the total metal recovery scheme must fit together in a workable manner.

For example, consider a solution obtained from a typical copper leach operation containing 2.5 g/I Cu(II) and 1.30 g/I total Fe at a pH of 1.80. The overall goal of the laboratory evaluation program is to determine the conditions required so that about 90% copper recovery is realized. It is also important to generate data on copper over iron selectivity.

From experience it is known that LIX 984, a mixed aldoxime-ketoxime reagent, is a good reagent for the leach liquor under discussion. Previous work tells us that in a 2 extraction—1 strip stage circuit, where – 90% copper recovery was obtained, LIX 984 transferred about 0.24 to 0.30 g/I Cu per volume percent reagent. Since the leach liquor of interest has a pH of 1.8 assume that LIX 984 will transfer in the middle to lower end of the range. The following simple calculation then gives the reagent concentration which should be suitable:

2.50 g/I Cu x (0.90) / 0.26 g/I Cu per v/v% reagent = 8.7 v/v% reagent

An organic solution containing 8.7 v/v% LIX 984 in a suitable diluent is prepared and then vigorously contacted once for about three minutes in a separatory funnel with the feed solution at an organic/aqueous (O/A) ratio of about 1. After the phases have separated and the aqueous phase is discarded the copper loaded organic phase is vigorously contacted twice, for about one minute each contact, with fresh pregnant strip solution (the aqueous solution expected to exit the strip stage of the circuit, 50 g/I Cu and 140 g/I H2SO4 in this case), also at an O/A=1. This results in an organic solution which closely approximates the stripped organic (S.O.) to be expected in the actual continuous circuit operation. The newly prepared stripped organic and the leach solution are now equilibrated at various organic to aqueous volume ratios by vigorous shaking for 3-10 minutes in separatory funnels or by rapid stirring in mixer boxes. After phase separation the respective organic and aqueous layers are filtered and saved for analysis. The organic layer is analyzed for copper (and perhaps iron) and the aqueous layer for copper only. The data is shown in Table II and is plotted as illustrated in Figure 5.

Stripping isotherm data is generated in a similar fashion by equilibrating the stripping aqueous phase (S.E.), in this instance a typical copper electrowinning tankhouse electrolyte containing -30 g/I Cu and 170 g/I H2SO4, with copper loaded organic (L.O.) at various O/A ratios (Table II, Figure 6).

Isotherms are unique to the conditions under which they are generated. If one of the parameters is changed, for example, the reagent concentration, the copper concentration of the respective organic or aqueous phase, the pH of the leach liquor, the acid concentration of the strip liquor, etc., then a different isotherm will be generated. In instances where the change in one or two of the parameters is very small it may not be necessary to run a second isotherm.

Properly generated extraction and stripping isotherms represent equilibrium conditions and, as such, predict the best extraction and the best stripping which can be obtained. These isotherms can be used to set the staging in a circuit. Consider, for example, the extraction isotherm in Figure 5 and suppose that the stripped organic entering into the last extraction stage contains 1.80 g/I Cu and that the advance flowrates of the leach solution and the organic phase are equal. Knowing this, an operating line can be constructed by starting at the point where the stripped organic intersects the isotherm and drawing the line up and to the right with a slope equal to the ratio of the organic/aqueous flow rates (one in this instance) until the operating line intersects the vertical line representing the copper content of the feed. Next, a horizontal line to the isotherm curve and then a vertical line to the operating line are drawn creating a “step”. This process is repeated, creating a second step and completing a two stage “McCabe-Thiele” diagram. Each triangle represents a single stage of extraction. In this system, a raffinate of 0.22 g/I Cu and a loaded organic of 4.24 g/I Cu are predicted in two stages of extraction.

Even though the McCabe-Thiele diagram shown in Figure 5 does not represent true equilibrium, but only a first approximation, it is still quite useful. For example, if a third stage of extraction were to be added to the McCabe-Thiele diagram in Figure 5, (note the dotted line) a near perfect equilibrium McCabe-Thiele diagram would result. In addition, a more accurate two stage McCabe-Thiele extraction diagram can be drawn by taking the two stage McCabe-Thiele construction as shown in Figure 5 and choosing as the point from which a new operating line is to be drawn a distance about ½ way between the isotherm line and the raffinate line. When this is done, and then a second two stage McCabe-Thiele diagram constructed as described above, a raffinate of about 0. 15 g/I Cu and a loaded organic of 4.17 g/I Cu are predicted. The construction of a true equilibrium McCabe-Thiele diagram is an iterative process. With the example under discussion the second step of iteration is all that is needed to produce a near equilibrium McCabe-Thiele diagram. This will not always be the case.

The construction of an equilibrium McCabe-Thiele diagram depicting one stage of stripping is very simple (Figure 6). A horizontal line is drawn from the loaded organic line (3.90 g/I Cu) to the isotherm line at the value of the pregnant electrolyte (P.E.) desired, 51 g/I Cu in this case. Then a vertical line is dropped from the point where the P.E. line intersects the isotherm to the horizontal axis. This gives the stripped organic (S.O.) to be expected (1.77 g/I Cu). Next, the strip electrolyte (S.E.) line, 30.7 g/I Cu, is drawn horizontally from the vertical axis of the graph to the S.O. line. The line connecting this point with the intersection of the loaded organic and pregnant electrolyte lines is the operating line. The slope of the operating line (2.13/20.3 = .105) is equal to the ratio of the advance strip aqueous flow to the advance organic flow needed to obtain the desired pregnant strip solution.

A two stage McCabe-Thiele strip diagram, constructed as described above for extraction, predicts a stripped organic of 1.12 g/I Cu when building a pregnant electrolyte of 51 g/I Cu and operating at an O/A of 7.3/1.

One of the important decisions the designer or builder of an SX plant must make is to decide on the staging requirements. The capital cost of a stage must be weighed against the benefits the stage provides. For example, consider the benefit of a second strip stage for the example cited above. Two stages of stripping will give a stripped organic of 1.12 g/I Cu and result in a net copper transfer for the organic (the difference between the loaded and stripped organic) of about 2.78 g/I Cu, while one stage of stripping gives a copper net transfer of about 2.13 g/I Cu. Because of the difference in net copper transfer, a one strip stage circuit requires a reagent concentration about 1.3 times greater than a two strip stage circuit in order to have equal copper extraction performance. Therefore, the organic losses expected in a one strip stage plant should be about 1.3 times greater than those in a two strip stage plant. By comparing the capital cost for a second strip stage with the increased operating cost (due to higher reagent usage) for the one strip stage plant, a purely economic decision on the strip stage requirements can be made.

In recent times most companies have opted to treat dump leach liquors in 2 extraction—1 strip (2E-1S) stage plants rather than 2E-2S or 3E-2S plants. For this reason, a 2E-1S circuit was set up and operated at the advancing O/A ratios shown in the McCabe- Thiele diagrams in Figures 4 and 5. A mixer retention time of 2.6 minutes was used simply because most copper SX plants have two or three minute mixers, depending on the leach solution conditions, mixer design and company philosophy.

In order to run a circuit properly, resulting in good metallurgy and maximum information, several things are important. Flow rates must be accurately set and continually monitored. Mixer dispersions should have an O/A ratio near one in order to get proper mixing, thus, recycles should be employed where needed.

The mixing turbines must rotate fast enough for good mixing, but, not so fast that heavy entrainment results. Another important feature of running a good circuit is frequent sampling with rapid and accurate analyses. This allows the operator to monitor the circuit closely, easily observing circuit behavioral differences as operating parameters are changed. Once an operator has changed operating conditions, the circuit should be allowed to run until circuit equilibrium is established. When a circuit is in equilibrium it means that there is a good metal balance and a good metal-acid balance across the whole circuit, as well as in each stage. Thus, the metal extracted from the aqueous phase in an extraction stage is equal to the acid equivalent gained by that same aqueous phase, and also equal to the metal loaded by the organic in that stage. In a stripping stage the metal stripped from the organic phase is equal to the metal gained, and to the acid equivalent lost, by the strip aqueous phase in that stage.

Obviously, in order to balance metal and acid values, solution analyses and solution flow rates must be known. Furthermore, solution samples for a given stage should be taken from the rear of the respective settler for that stage. When possible it is best to pull samples at those points easiest to access and least likely to upset the circuit.

It is important to realize that circuit behavior will reflect the interdependence of the extraction stripping operations. For example, a loss of stripping efficiency will result in a higher stripped organic which in turn leads to a higher raffinate. The circuit must then be operated with a slightly higher reagent concentration in order to maintain copper recovery or a slightly lower recovery must be accepted.

Because of the design of mixers currently used in copper solvent extraction plants, metal transfer in a stage is usually about 85 to 95% of theoretical. For this reason, the values predicted by equilibrium McCabe-Thiele diagrams for the aqueous and organic phases exiting a given stage are seldom realized in an operating circuit. For example, compare the organic and aqueous values which are predicted by the McCabe-Thiele diagram in Figure 5 with the values generated in a continuous laboratory circuit using the same aqueous, organic and strip aqueous solutions (Table III, Figure 7).

The circuit produced a raffinate of 0.28 g/I Cu and a loaded organic of 4.08 g/I Cu as compared to the raffinate of 0.22 g/I Cu and the loaded organic of 4.24 g/I Cu predicted by the first approximation McCabe-Thiele extraction diagram (Figure 5). The stripped organic of 1.80 g/I Cu is close to the 1.77 g/I Cu predicted by the McCabe-Thiele strip diagram (Figure 6).

The natural question arises: Why perform the work required for an equilibrium isotherm when the circuit will not produce the predicted results? There are several good reasons. First of all, generating an isotherm helps to develop a “feel” for the system. Secondly, if the system has any unusual metallurgical behavior characteristics, the isotherm will usually reflect them. Thirdly, a good idea of the actual circuit results can be obtained by drawing in the operating line as previously described. However, instead of filling in the horizontal and vertical lines to reflect 100% mixer efficiency, draw the horizontal lines in to represent only 95% efficiency. When this is done using Figure 5, a two stage raffinate of 0.28 g/I Cu is predicted—a value identical to that realized in the circuit operation.

The feed chosen for the above example is typical of many copper leach operations and represents one of the simplest feeds to treat. Copper recovery is very good, the selectivity for copper over iron is high (in the above circuit it is about 1100/1 on a transfer basis), and there is little tramp metal contamination. More complicated feed solutions are evaluated in a manner similar to that described above, but, the extraction and stripping isotherms and the staging in the circuit can be, and often are, more complicated. Thus, the initial circuit may have to be modified several times before the best recovery scheme is worked out.

Computer-Generated McCabe-Theile Diagrams for Copper

Henkel has developed a computer program, called Isocalc, which will generate isotherms for the extraction of copper from typical sulfuric acid leach solutions. This computer program is based on the equilibrium constants for the extraction of copper from aqueous acidic sulfate solutions with various Henkel reagents, according to the equation on page 3. The program also incorporates the effects of sulfate buffering, the initial pH of the leach liquor, the reagent concentration and the stripping conditions for the loaded organic.

After generating an isotherm, the program will use that isotherm to calculate McCabe-Thiele extraction diagrams under specified conditions. Finally, the program will calculate and print mass balances across both the extraction and stripping stages.

This program is useful for:

a) predicting reagent behavior under varying conditions,

b) comparing the performance of one reagent with another under varying conditions,

c) predicting metallurgical changes as conditions in a circuit are changed, for example, staging, the pH of the leach liquor, reagent concentration, mixer efficiency and stripping conditions.

However, the computer program only predicts those circuit behaviors associated with the copper extraction ↔ stripping equilibrium. It does not predict copper over iron selectivity nor does it address the physical properties of a circuit such as:

d) phase separation—how fast the aqueous/organic dispersion exiting a mixer separates into organic and aqueous phases,

e) entrainment—the very small droplets of one phase carried in the other phase after primary phase separation has taken place,

f) crud generation—the term crud or gunk is used to describe an interfacial material which is present in almost all solvent extraction circuits. This material usually contains finely divided solids (either carried into the circuit with the leach liquor or precipitated in the circuit), some stable emulsion, bulk leach solution and/or bulk organic, colloidal silica and at times bacteria. Crud is a very difficult thing to describe as crud in one plant may be quite different from crud in another plant.

The Isocalc program should be used in conjunction with, but, not as a total substitute for, laboratory evaluations. As such, it can save a lot of time by very quickly providing information on “what if” type questions. This is particularly true early in a project.

A version of the program is available for Henkel customers, but, it is not available for general distribution. Under certain conditions your Henkel solvent extraction specialist will be happy to run Isocalc programs for your projected sulfuric acid copper leach solutions.

Commercial Applications for Copper Recovery

Sulfuric Acid Leach Solutions

The use of chelation type extractants in liquid ion exchange has enjoyed its greatest commercial success for the recovery of copper from dilute sulfuric acid leach solutions which are generated from the leaching of oxide and/or sulfide ores using a variety of leaching techniques including dump, heap, in-situ, vat, thin layer and agitation. The leach-SX-EW technology begins with leaching and Henkel advises that companies considering this technology thoroughly investigate the leaching characteristics of their ore. Money spent up front in leaching studies usually pays handsome dividends.

Typical leach solutions may contain less than 1 g/I Cu up to about 35 g/I Cu over a pH range of 1.1 to 3.0. The leach solution may also contain up to 50 g/I chloride and a host of other impurities depending on the ore, available water and evaporation rate. Typical sulfuric acid leach solutions successfully treated by solvent extraction are summarized in Table IV.

Other common constituents: Fe, Mo, Mn, Al, Mg, Na, K, sulfate ions and at times chloride ions.

Typical solutions from which high purity copper is electrowon contain from 30 to 38 g/I Cu and 140 to 180 g/I H2SO4. Most copper tankhouses coupled with SX in the acid sulfate system will have copper and acid concentrations in these ranges.

At the present time (December 1996) there are forty seven commercial operations that recover copper from dilute sulfuric acid leach solutions and there are several plants under construction with several others planned. Actual operating costs for one of the largest of these, the tailings leach plant of the Chingola Division, Zambia Consolidated Copper Mines, Ltd., in Zambia, Africa, and one of the smaller, the Cyprus Bagdad Copper Company plant in Bagdad, Arizona, have been published. The advantages of solvent extraction-electrowinning (SX-EW) over the cementation process for the recovery of copper from these types of leach solutions is well documented in these papers. These papers also demonstrate the applicability of solvent extraction on a very large scale, since ZCCM produces over 7,000 tons per month of high quality cathode copper, and on a smaller scale, as Cyprus Bagdad produces about 1100 tons per month of superior quality cathode copper. In addition, the ability to forecast fairly accurate operating costs for the SX-EW operation, as reported in the above papers, is a great advantage to the parties involved during feasibility studies.

Commercial Copper Solvent Extraction Reagents

Even though a large number of molecules with a wide variety of extractive functionalities have been proposed as extractants to be used for the recovery of copper from sulfuric acid leach solutions, only the hydroxy oximes have been used in commercial copper SX-EW plants. The basic structure for these copper extractants is shown in Figure 8.

The extractants of the general structure can be subdivided into two distinct classes based on their structure and properties: the ketoximes, which are normally copper extractants of moderate strength and the salicylaldoximes, which are very strong copper extractants. The strength of a copper extractant is based on the degree to which the copper extraction stripping equilibrium previously shown is driven toward extraction by the reagent. In simple terms, very strong copper extractants extract substantial amounts of copper at pH values less than 1.0, while moderate strength copper extractants are most useful above a pH of 1.6-1.8.

Another class of extractant combines a salicylaldoxime with a ketoxime in an approximate 1/1 mole ratio. This “third” class is not based on structure, but rather on the distinct and advantageous properties the mixtures exhibit. These patented mixtures (15) are classified as strong copper extractants and as such are useful at pH values as low as -1.2.

Ketoximes: A ketoxime was the first hydroxy oxime extractant to be used commercially for the extraction of copper from dilute sulfuric acid leach liquor and ketoximes were used exclusively for about 11 years. The most outstanding feature of the ketoximes, as represented by LIX 84, is the good physical performance they display under a wide variety of conditions, especially with respect to aqueous solutions that are known to be sensitive to certain organic solutions. One example would be agitation leach solutions which contain colloidal silica, some solids and/or residual flocculent. A second example are leach solutions with dissolved organics, often present due to rotting vegetation. The ketoximes show excellent phase separation, low entrainment losses to the raffinate and do not promote excessive crud formation. Because ketoximes are only moderately strong copper extractants and kinetically slow at cold temperatures, the number of copper SX plants using a ketoxime extractant exclusively is limited. However, in the circuits which are designed to use ketoximes, the operator enjoys a low cost, trouble-free copper extractant.

Salicylaldoximes: The salicylaldoximes were developed to overcome the perceived shortcomings of the ketoximes. Their outstanding characteristics include rapid copper transfer kinetics and high extractive strength. However, the salicylaldoximes by themselves are such strong copper extractants that they are most often used in combination with an equilibrium modifier or with a ketoxime so that they can be efficiently stripped with an acidic copper solution from which high quality copper can be electrowon.

The use of equilibrium modifiers leads directly to certain shortcomings in that some modifiers are known to accelerate reagent degradation. It has also been reported that equilibrium modifiers contribute significantly to the amount of crud generated in some solvent extraction circuits. Another problem sometimes associated with modified reagents is the contamination of electrolyte due to the excessive entrainment of leach liquor in the loaded organic stream. Finally, nonylphenol is known to have deleterious effects on certain materials of construction.

As a group salicylaldoximes are less stable than ketoximes and within the salicylaldoxime subgroup, the nonyl derivative is less stable than the dodecyl derivative. In operating circuits at normal temperatures, reagent degradation with 5-nonylsalicylaldoxime has been calculated to be equal to about 10% of the total reagent makeup. Actual degradation results on operating circuits have not been published; however, for some plants, degradation is known to be higher than the calculated value depending on the modifier, the temperature and the acid content of the aqueous stripping solution. With heated circuits, reagent degradation could be even more significant and should be determined experimentally. Still, overall reagent loses are quite low in a well operated plant so the loses due to degradation are not excessive in any case.

In spite of their shortcomings, modified salicylaldoxime reagents such as LIX 622 have been quite successful in commercial circuits which are designed for their use, and where the leach liquor is compatible with the reagent.

Ketoxime-Salicylaldoxime Mixtures: The properties of ketoxime-salicylaldoxime mixture, such as LIX 984 and LIX 984N, which contain no added modifier, reflect the most desirable characteristics of the components: the extractive strength and fast kinetics of the salicylaldoximes combined with the proven, excellent physical performance and stability of the ketoximes.

Although the extractive strength of the mixtures is not quite as high as the modified salicylaldoximes, it is considerably greater than the ketoximes. Copper transfer kinetics are a little slower than salicylaldoximes, but are still very fast.

One interesting feature of these mixtures is the greater than expected copper transfer these mixtures give because of their lower than expected stripped organic value. It is felt this property results from the ketoxime component of the mixture functioning as a modifier for the salicylaldoxime component in the strip stages of the circuit. The mixtures were introduced to the industry in 1982 and are now the most widely used copper extractant system in the world. Since the ketoxime and salicylaldoxime components can be mixed in any ratio, a reagent mixture can be tailored to the demands of almost any leach liquor. Indeed, these mixtures could prove in time to be a universal reagent system for acid sulfate copper leach solutions since, in most instances, a properly tailored mixture will function both physically and chemically very well with virtually any leach solution of this type.

Since all three reagent classes of oxime extractants are commercially successful it follows that each class of extractant meets the reagent requirements previously given. In reading the successful reagent requirements it is obvious that they are somewhat general in nature. For example, the first requirement says that the reagent must extract the desired species selectively (in this case copper), but, it says nothing about how much copper must be extracted, nor the degree of selectivity the reagent must have. In fact, it is virtually impossible to be more specific because each leach liquor, each plant and each reagent is unique. Whereas a given reagent requirement may be critical in one situation, in another it can be relatively unimportant. What should be understood is that all three classes of hydroxy oxime reagents possess the successful reagent requirements to such an extent that all three classes are commercially successful, yet, no one reagent is the best for each and every copper leach liquor or copper solvent extraction circuit.

b. Henkel Copper Solvent Extraction Reagents

Henkel produces three oxime extractants: 5-nonylsalicylaldoxime, 5-dodecylsalicylaldoxime and 5-nonyl- 2-hydroxyacetophenone oxime. These three oximes form the basis for a wide range of reagents which are derived by blending the oximes in a variety of ways, for example, with diluent, with one another or with a modifier, to give LIX Copper Solvent Extraction Reagents with properties that best fit the particular leach solution and plant design under consideration.

Furthermore, there are operating copper SX plants which purchase the individual reagent components separately and mix these components in a way that best meets the individual needs of the particular leach solution, plant or season. The ability to purchase the reagent components gives copper SX plant operators great flexibility in the operation of their plant to meet changing conditions, for example, summer verses winter operation, while still maintaining excellent physical properties.

Henkel makes available both normally formulated reagents and concentrated reagents. The concentrated reagents contain the same oximes as the normally formulated reagents. However, because the concentrated reagents are formulated with less diluent in order to save packaging, shipping and handling costs, the concentrated reagents are more viscous than normally formulated reagents and they do not flow well at temperatures less than about 5°C.

Table V shows the contents for some of the oxime copper extractants available from Henkel. If you would like more information about Henkel oxime copper solvent extraction reagents, or if you would like to discuss the extraction of copper from dilute sulfuric acid leach solutions in more depth, contact your local Henkel representative.

The oxime copper extractants marketed by Henkel Corporation, either alone or mixed, are highly selective for copper over iron when mixed with almost all copper leach solutions. As a result minimum tankhouse bleeds are needed for iron control.

Ammonia Leach Solutions

The extraction of metal from ammoniacal leach solutions is a second commercial application for chelation type extractants. Ketoximes, such as LIX 84, and the beta-diketone reagent LIX 54 have both been used. In one example, a ketoxime was used as the extractant for a leach liquor derived from the ammonia leaching of copper sulfide concentrates in a process termed the “Arbiter Process”. In another example, LIX 54 is used as the extractant for copper from a leach liquor derived from the ammonia leaching of copper-lead dross.

LIX 54 was designed for the extraction of copper from ammoniacal leach liquors. Its ease of stripping, low ammonia loading, fast kinetics, good phase separation, high copper loading, and low viscosity when fully loaded, make it ideally suited for this task. However, LIX 54 is not applicable to sulfuric acid leach solutions since it does not extract copper well below pH -3.0.

E. Henkel Chelating Reagents and Some of Their Possible Uses

LIX 54: A beta-diketone based reagent, used for the selective extraction of copper from ammoniacal leach solutions, which strips with very low concentrations of residual sulfuric acid and operates at a high net copper transfer. It may be possible to strip copper from this reagent using concentrated ammonia solutions.

LIX 63: This hydroxy oxime has been proposed for germanium recovery (20), for copper recovery from typical tankhouse electrolytes (21), in synergistic mixtures with other reagents for cobalt and nickel recovery and for the separation of molybdenum from uranium in dilute sulfuric acid solution.

LIX 84: This ketoxime based reagent requires about 150 g/I H2SO4 for copper stripping and contains no added modifier. It is widely used for copper recovery from dilute sulfuric acid leach solutions, and has application for the recovery of copper and nickel from ammonia solutions.

LIX 87QN: A ketoxime based reagent developed specifically for the extraction of nickel from ammonia, followed by a stripping operation using a concentrated solution of ammonia and carbon dioxide.

LIX 860: This aldoxime based reagent, containing no added modifier, has been proposed as a reagent for the coextraction and selective stripping of copper and zinc. In addition, this reagent when coupled with a copper sulfate crystallization circuit can be used to extract copper efficiently at pH values less than 1.0.

Whereas the above list is impressive, it should not be considered as limiting. Henkel is constantly working on the development of new chelating functionalities, while Henkel, and others in the industry, are continually working on new applications for existing chelation type extractants.

F. Capital Costs for Copper SX-EW

It is difficult to set out a formula by which to predict capital costs for solvent extraction-electrowinning plants. Raw material costs, local labor markets, the type of construction, the nature of the feed and the final plant design are variables important in predicting capital costs. Henkel Corporation does try to keep current in these areas and will do some preliminary capital cost analysis for people interested in solvent extraction. To this end, Henkel has developed its Metcalc computer program, which can be used in conjunction with the Isocalc program previously discussed, or by itself. The Metcalc program is not available for distribution, but, Henkel technical representatives will be happy to run the program according to customer needs.

Ion-Pair Extractants

A second class of extractants is based on the principle of ion association; whereby a large, positively charged organic moiety causes the extraction of a large, anionic metal complex into the organic phase, with concomitant expulsion of a small common anion to the aqueous phase. Henkel Corporation markets two classes of compounds which fall into this general classification.

A. Alamine Series

The Alamine Series of Reagents all contain a basic nitrogen capable of forming amine salts with a wide variety of inorganic and organic acids. The amines which find the widest use in metals recovery processing by solvent extraction are the tertiary amines of the general formula R3N, where R can represent a variety of hydrocarbon chains. The equations below are representative of the extraction chemistry of the Alamine Series of Reagents:

Equation a represents simple amine salt formation while equation b represents true ion exchange. The extent to which B will exchange for A is a function of the relative affinity of the two anions for the organic cation and the relative stability of the anions in the aqueous medium. Since tertiary amines can form salts with a wide variety of acids, the choice of reagents for a liquid ion exchange process is quite large, i.e., the bisulfate salt, the chloride salt, etc., of any of the Alamine Reagents.

Amine type extractants can be stripped with a wide variety of inorganic salt solutions such as NaCl, Na2CO3, and (NH4)2SO4. The choice of stripping agent depends on the overall recovery process, but in general, basic stripping agents which deprotonate the amine, give the best stripping in the fewest stages. The equation below shows the stripping action of Na2CO3 on an amine salt:

The greatest commercial acceptance for Alamine type reagents has been in uranium recovery processes, however, any metal capable of forming anionic complexes in aqueous solutions is a candidate for extraction by an amine type extractant. Figure 9 shows some of the extraction data which has been generated using Alamine 336 and/or similar amines. The data is far from all-inclusive, but, it does show the wide variety of metal-aqueous systems which are amenable to treatment with amine type extractants.

B. Aliquat Series

The Aliquat Series of Reagents is based on the methyl chloride quaternization of a respective Alamine reagent. Because a quaternized amine is always positively charged, anion extraction with Aliquat Reagents is not pH dependent like it is with the tertiary amines. As a result, some basic metal leach solutions may be successfully treated by Aliquat type reagents without pH adjustment. The biggest dis-advantage of using an Aliquat reagent is that these reagents will not deprotonate, therefore, stripping is usually more difficult than with the parent amine type reagent.

C. Laboratory Evaluation Program for Ion- Pair Extractants

Laboratory evaluation programs with amine type extractants are carried out much like those described earlier for chelating type reagents. Even though amine systems tend to be more complicated than chelating systems for a variety of reasons:

a. the large number of aqueous anionic systems where an amine extractant may be applicable,

b. the large variety of potential stripping agents as compared, for example, to copper loaded chelates, where stripping with sulfuric acid is the norm,

c. the three variables: pH, the oxidation state of the metal and the concentration of the anion contributing to the anionic metal complex, all of which are important in metal separation schemes with amines,

The goals of the laboratory evaluation program remain the same; to develop the best conditions for the extraction, stripping and final product recovery consistent with the overall metallurgical flowsheet.

An example which illustrates the importance of the variables cited above is the separation of Co, Fe and Ni from an acidic chloride solution. Refer to Figure 10 and note that Ni(II) is only slightly extracted at 200 g/I chloride ion, Co(II) and Fe(II) are strongly extracted at 150-200 g/I chloride ion and Fe(III) is strongly extracted at 50 g/I chloride ion. As a result, an excellent Co, Ni, and Fe separation can be effected by taking an alloy of the metals into a concentrated HCl solution, oxidizing the iron to ferric, co-extracting the Co(II) and Fe(III) while leaving the Ni in the original solution, and then selectively stripping the Co(II) from the Fe(III) with water. This flowsheet has been studied in our laboratory. Several slightly different flowsheets have also been reported. The best flowsheet for this type of feed is dependent upon the relative amounts of Co, Ni and Fe present.

In general, the use of organic phase modifiers, normally polar organic molecules such as long chain alcohols, is required with amines in order to keep the amine metal complexes soluble, prevent third phase formation and/or to have acceptable phase separation.

D. Present and Possible Future Uses of Amine Extractants

One of the best ways to illustrate the value of amines as ion-pair metal extractants is to list the high purity amines commercially available from Henkel Corporation and some of their known uses:

1. Alamine 300 (tri-n-octylamine): cobalt extraction from chloride leach solutions.
2. Alamine 308 (tri-isooctylamine): cobalt-nickel separation from hydrochloric acid liquors.
3. Alamine 336 [tri-(C8C10)amine]: uranium recovery from sulfuric acid leach liquors or sulfuric acid resin eluate solutions, cobalt-nickel separations from hydrochloric acid, iron extraction from aluminum chloride solutions, vanadium extraction, chromium extraction, platinum group metals-separations and tungsten recovery.
4. Alamine 310 (tri-isodecylamine): uranium and vanadium extraction from acidic sulfate leach liquors and platinum group metals separations.
5. Alamine 304 (trilaurylamine): uranium and molybdenum extraction from acidic leach liquors, the advantage being that the molybdenum amine complex is highly soluble.
6. Aliquat 336 [tri-(C8C10)methylammonium chloride]: vanadium extraction, separation of platinum group metals, rare earth extractions, rhenium recovery, and arsenic extraction from refinery electrolytes.
7. Aliquat 336 Nitrate [tri-(C8C10)methylammonium nitrate]: a specialty product for rare earth separations.

Future uses for these types of reagents may lie in the areas of:

a) reclaiming metals from spent catalysts,
b) niobium, tantalum, titanium and zirconium separations,
c) inorganic acid purification,
d) separation of cobalt, copper and zinc from nickel.

Neutral or Solvating Type Extractants

A third class of extractants are known as neutral or solvating type extractants. Extractants of this class are basic in nature and will coordinate to certain neutral metal complexes by replacing waters of hydration, thereby causing the resulting organo-metal complex to become aqueous insoluble, but organic soluble. Solvating extractants have an atom capable of donating electron density to a metal in the formation of an adduct, and are classified according to that ability:

R3PO >(RO)3PO >R2CO >ROH >R2O

trialkylphosphine oxides>trialkylphosphates> ketones>alcohols>ethers.

It takes little imagination to see that the above list is only a brief representation of the organic compounds which could function as solvating extractants. In general, extractions with solvating extractants are limited by: 1) the metal’s ability to form neutral complexes with anions, 2) the co-extraction of acid at high acid concentrations, and 3) the solubility of the organo-metal complex in the organic carrier.

An important extractant of this type in tri-n-octylphosphine oxide (TOPO), (C8H17)3PO. The extraction characteristics of TOPO with a wide variety of metals have been investigated and are summarized (Figure 11). The most important commercial application of TOPO in solvent extraction is its synergistic combination with di-2-ethylhexylphosphoric acid (D2EHPA) for the extraction of uranium from wet process phosphoric acid.

Other important extractants in this class are tributylphosphate (TBP), di-butyl butylphosphonate (DBBP) and 2,2’-dibutoxy diethyl ether (trade name dibutyl carbitol).

It is of interest to note that all of the equilibrium and phase modifiers commonly used come from this class of extractants. This is no surprise since both require a group having the ability to donate electron density and in some cases to be a hydrogen bond acceptor and/or donor.

Henkel Corporation does not currently market reagents of this type.

Organic Acid Extractants

A fourth class of extractant is the so-called acid extractants. The chemistry of this type of extractant has some characteristics which resemble chelating extractants and some which are similar to neutral or solvating extractants. Di-2-ethylhexyl-phosphoric acid (D2EHPA), shown below, is representative of this class of extractant.

Reagents which belong to this class are the organophosphoric, -phosphonic and -phosphinic acids, their respective mono- and di-thio derivatives, organosulfonic acids and carboxylic acids.

The following equation shows the dual behavioral characteristics:

Note that two of the D2EHPA molecules lose a proton much like a chelating extractant while two other D2EHPA molecules solvate the zinc similar to solvating type extractants. The pH extraction characteristics of D2EHPA with respect to several metals is shown in Figure 12.

One of the more interesting characteristics of the extraction behavior of non thio acid extractants is their strong affinity for iron(III) over other base metals. This is unfortunate since it limits the use of these reagents for two reasons:

1. Insufficient selectivity for most base metal systems,
2. Iron is so strongly extracted that stripping the iron can be a problem.

In spite of these limitations, the acid extractants have found some actual and proposed use for metal recovery. D2EHPA has been used for the recovery of uranium(3,52), vanadium (3, 52), and zinc. All three types of acidic organo-phosphorous compounds have been used for the selective extraction of cobalt over nickel. There are remarkable selectivity differences for cobalt over nickel between certain types of these compounds resulting from very interesting and subtle chemical and steric differences.

Reagents in Development, Gold

At the time this edition of the Red Book was published, development work on several new extraction reagents and/or metal recovery processes was taking place in our laboratories. A brief discussion about one of these developments follows. Inquiries from interested parties about the progress of these developments are invited.

Gold Extraction

Henkel is developing liquid reagents for the extraction of gold from typical cyanide leach solutions which can be stripped with an aqueous solution of 2-4 w/v% NaOH and 0.5 w/v% of NaCN. These reagents are LIX 79, a guanidine-based extractant, and XI 78. The selectivity of these reagents for metal cyanide complexes has the following order:

Au-Zn-Hg > Ag > Ni>Cu-Fe.

The recovery of gold from clarified cyanide leach solutions may offer certain advantages over existing gold recovery processes, particularly:

1. where very high gold recovery is needed or,
2. in the case of leach solutions carrying either high levels of copper or organic compounds which foul carbon.

In addition, the gold inventory carried in the SX-EW recovery process is less than that in typical carbon in column or Merrill Crowe recovery processes.

The guanidine functionality is also available as a solid ion exchange resin called Aurix 100. This resin has shown excellent gold extraction kinetics and high gold loading in laboratory and pilot plant tests. It has also been shown to be resistant to breakage in typical resin in leach or resin in pulp applications. Aurix 100 may show significant advantage over activated carbon for gold extraction from preg robbing pulps or from bio-leach liquors which foul carbon.

Solvent Extraction Diluents

The discussion to this point has been centered around the reagents used for metal extraction, yet, in many commercial SX plants the diluent used as a carrier for both the reagent and the reagent metal complexes is present in far larger amounts than the reagent. Furthermore, the diluent can alter both the chemical and physical properties of an SX system. Given this, a brief discussion on diluents is appropriate.

In many cases, the role of the diluent on the chemistry of the system is minor, and the choice of diluent is made simply on economics and on the diluent’s physical properties such as phase separation, flash point, volatility and the solubility of the reagent and the reagent metal complexes. It is fortunate that with many of the large SX systems, for example, the copper-oxime system, the above holds true. In other cases, where the diluent has a major role in the chemistry of the system, the choice of diluent depends on both chemical and physical properties. In either case, efforts to choose a proper diluent, and to monitor the quality of continuing supplies of the diluent, are important. These efforts must be included as part of any good laboratory evaluation program.

It should be noted that in this discussion, the function of the diluent is assumed to be separated from that of the modifier. In real systems this is not a valid assumption, but, solvent-modifier interactions are complex and for reasons of brevity this topic will not be discussed.

How does solvent extraction work

The potential of solvent extraction is well documented by the wide variety of metal-containing solutions which are amenable to recovery, purification, and concentration by solvent extraction. However, much work in both reagent synthesis and process development remains. Henkel plans to continue as an industry leader in this field via its vigorous research and development program. This program includes close working relationships with existing and potential customers in a joint effort to solve problems of common interest in metals recovery. Efforts in the area of technical service to existing and potential customers have made a significant contribution to our overall success in solvent extraction. Henkel realizes it does not have all the answers nor it is aware of all the problems. But Henkel firmly believes that by cooperating with the customer we can become aware of, and help find solutions to, many of these problems. Henkel’s 40 years of experience in solvent extraction is an important asset to our problem-solving ability.

Information contained in this technical literature is believed to be accurate and is offered in good faith for the benefit of the customer. Henkel, however, cannot assume any liability or risk involved in the use of its chemical products since the conditions of use are beyond our control. Statements concerning the possible use of our products are not intended as recommendations to use our product in the infringement of any patent.