Industrial Use of Solvent Extraction

1.4.3 Industrial Use of Solvent Extraction

A surge in interest in solvent extraction occurred in the decades of the 1940s and 1950s initiated by its application for uranium production and for reprocessing of irradiated nuclear materials in the U.S. Manhattan Project. The first large-scale industrial solvent extraction plant for metals purification was built in 1942 by Mallinckrodt Chemical Co., St. Louis, for the production of ton amounts of uranium by selective extraction of uranyl nitrate by ether from aqueous solutions. The high degree of purity (99.9%) required for use of uranium in nuclear reactors was achieved. An explosion led to the replacement of the ether by other solvents (dibutylmethanol and methylisobutylketone). At the same time new types of more efficient metal extractants were introduced, e.g., tri-n-butylphos- phate in 1945 and trioctylamine in 1948. This activity became a great stimulus to the nonnuclear industry, and solvent extraction was introduced as a separation and purification process in a large number of chemical and metallurgical industries in the 1950s and early 1960s. For example, by leaching copper ore with sulphuric acid followed by extraction of this solution with an organic hydroxyaryloxime dissolved in kerosene, several million tons of copper (30% of world production) is now produced annually. This and many other processes are described in later chapters.

For these applications, the technique of solvent extraction had to be further developed and with this a new terminology was also developed. This can be illustrated by considering a process where a desired component in an aqueous solution is extracted with an organic reagent (extractant) dissolved in another organic liquid; note here that the term “organic solvent” is not used because of possible confusion. The term “solvent” could be used for the whole organic phase or for the organic liquid in which the organic extractant is dissolved. Thus the term generally given to the latter is (organic) diluent.

While in laboratory experiments the extraction vessel may be a test tube, or more conveniently some kind of separation funnel (Fig. 1.1), this is not suited for industrial use. Industry prefers to use continuous processes. The simplest separation unit is then the mixer-settler, or some clever development of the same basic principle, as described in Chapter 9. Figure 1.5 pictures a simple mixersettler unit, here used for the removal of iron from an acid solution also containing nickel and cobalt (the same systems as in Fig. 1.3). The mixer (or contactor) is here simply a vessel with a revolving paddle that produces small droplets of one of the liquid phases in the other. This physical mixture flows into and slowly through the separation vessel, which may be a long tank; through the influence of gravity the two phases separate, so that the upper organic kerosene-octanol-amine phase contains the Fe(III) and the lower aqueous CaCl2 phase contains the Co(II) and Ni(II). Numerous variations of the construction of mixersettlers (or MS-units, as they are abbreviated) exist (Chapter 9), often several joined together into MS-batteries. Many such will be described later on.

A diagram of a full basic process is given in Fig. 1.6 to illustrate the common terminology. The incoming aqueous solution is called the feed. It is contacted with the (recycled) solvent phase in a mixer-settler unit. Here we do not indicate the exact type of unit, but only its function (extraction), as commonly is done. After extraction and separation of the phases, the depleted phase becomes the raffinate and the enriched solvent phase becomes the extract or loaded (or pregnant) solvent. The raffinate may undergo a solvent recovery stage to remove any entrained solvent before exiting the process. The extraction process is rarely specific so that other solutes may be co-extracted with the main component. These impurities may be removed with an aqueous scrub solution in a scrub stage producing a scrub extract and a scrub raffinate containing the impurities. The latter may return to the feed solution to maintain an overall water balance. The scrubbed extract is now contacted with another aqueous solution to strip or back-extract the desired component. The stripped solvent then may undergo some regeneration process to prepare the solvent phase for recycle. The loaded (pregnant) strip solution then is treated to remove the de- sired product and the strip solution is recycled. One of the important aspects of this flowsheet is that, wherever possible, liquid phases are recovered and recycled. This is important from both an economic and an environmental standpoint.

Figure 1.6 shows a situation where each process-extraction, scrubbing, and stripping-occurs in a single operation or stage. This is generally not efficient because of the finite value of the distribution coefficient. Thus if D has a value of 100 (i.e., E = 99%), then after one extraction, the organic extract phase will contain approximately 99 parts and the aqueous phase 1 part. To achieve a greater extraction, the aqueous raffinate should be contacted with another portion of the solvent after which the new organic phase contains 0.99 parts, and the aqueous raffinate now only 0.01 part. Therefore, two extraction stages will provide 99.99% extraction (with “2 volumes” of the organic phase, but only “1 volume” of the aqueous phase).

Three different ways of connecting such stages are possible: namely, cocurrent, crosscurrent, and countercurrent (see Fig. 1.7). In cocurrent extraction, the two phases flow in the same direction between the various contactors. A simple inspection of the diagram will show that with this configuration no advantage is gained over a single contact because, providing equilibrium is reached in the first contactor, the separated flows are in equilibrium when entering the second contactor so no change in relative concentrations will occur. In the second configuration (b), crosscurrent, the raffinate is contacted with a sample of fresh solvent. This is the classical way of extracting a product in the laboratory when using a separatory funnel and will give an enhanced recovery of the solute. However, on an industrial scale, this is seldom used because it results in the production of a multitude of product phases containing a reducing concentration of the desired solute. These have to be combined before stripping resulting in a much larger volume of loaded solvent to be treated with consequences for plant size and economics. The third configuration (c), countercurrent, is the one generally chosen by industry. The phase volumes remain constant and by feeding the two phases, feed and solvent, at opposite ends of the bank of contactors, the driving force for extraction, i.e., the solute concentration difference between the two phases, is maximized. Chapters in the second part of the book will extend this discussion.

The need to use multiple extraction to achieve efficient extraction required the development of new types of continuously working extractors, especially mixersettlers and pulsed columns, which were suitable for remotely controlled operations. These new extractors could be built for continuous flow and in multiple stages, allowing very efficient isolation of substances in high yield. A good example is the production of rare earth elements in >99.999% purity in ton amounts by mixer-settler batteries containing hundreds of stages. These topics will be further developed in Chapters 6 and 7.

In the early analytical applications of solvent extraction, optimal extraction or separation conditions were obtained empirically. This was unsatisfactory and general mathematical descriptions were developed by a number of researchers in many countries. This was especially important for largescale industrial use and is an activity that continues today almost entirely with computers.

Soure: Solvent Extraction Principles and Practice, Revised and Expanded edited by Jan Rydberg