Extraction pilot plant, solvent

Feather, A. Bouwer, W. Swarts, A. Nagel, V. Pilot-plant solvent extraction of cobalt and nickel for Avmin s Nkomati project. International Solvent Extraction Conference, Cape Town, South Africa, Mar. 17-21, 2002, 946-951. 

A hydrocarbon mixture containing 10% aromatics and at the rate of 55.5 m3/hr is to be treated with a solvent at the rate of 173.6 m3/hr. Ten stages are needed for the extraction. Pilot plant data are available for the HTU and the slip velocity they are shown on the graphs for solvent either continuous or dispersed. The procedure of Table 14.5 was applied by Kosters (in Lo et al., 1983, pp. 391-405) with the following results ... 

Alonso, A.I., Galan, B., Gonzalez, M. and Ortiz, I. (1999) Experimental and theoretical analysis of a nondispersive solvent extraction pilot plant for the removal of Cr(VI) from a galvanic process wastewaters. Industrial Engineering Chemistry Research, 38, 1666. 

Flowsheet testing and data collection were performed using the General Atomic Company solvent extraction pilot plant equipment shown in Figure 1. Included in this equipment are several 5.1 to 7.6 cm (2 to 3 in.) diameter cylindrical glass pulse columns, a 15-2 cm (6 in.) diameter annular pulse column, a centrifugal contactor (Robatel Co.) and associated tanks, feed systems and concentrators. 

Table 7.12 Flow rates and compositions in TBP solvent extraction pilot plant for hafnium-zirconium separation...

The only Canadian uranium refinery is operated at Port Hope, Ontario by Eldorado. Tonnage quantities of oxide (U3O3) were produced first in 1942 from ore concentrates. A solvent-extraction pilot plant was operated in 1950 and 1951 to investigate methylisobutylketone (hexone) and then tributylphosphate as extractants for uranium to obtain a high-purity product. The present refinery was designed and built by the Catalytic Construction Company in 1955 (11). Yellow cake is digested in nitric acid, the resultant slurry extracted with tributylphosphate dissolved in kerosene, and the uranium, after purification, transferred back to water. This solution is decomposed thermally to UO3. Capacity is about 5 Gg U/a. 

Aromatic and Nonaromatic Hydrocarbon Separation. Aromatics are partially removed from kerosines and jet fuels to improve smoke point and burning characteristics. This removal is commonly accompHshed by hydroprocessing, but can also be achieved by Hquid-Hquid extraction with solvents, such as furfural, or by adsorptive separation. Table 7 shows the results of a simulated moving-bed pilot-plant test using siHca gel adsorbent and feedstock components mainly in the C q—range. The extent of extraction does not vary gready for each of the various species of aromatics present. SiHca gel tends to extract all aromatics from nonaromatics (89). 

The initial bench-scale experimental investigations into solvent extraction processes are conducted with small apparatus, such as separating funnels. Following the successful completion of these tests, when the best reagent and other conditions for the system have been established, small-scale continuous operations are run, such as in a small mixer-settler unit. The data so obtained are used to determine scale-up factors for pilot plant or plant design and operation (see Chapters 7 and 8). 

Inclusion of this practice in all solvent extraction studies will ensure that the solvent is not discarded as being unsuitable, and that problems in pilot plant or continuous operations are not the result of impurities that were present in the bench-scale test work. 

Because of the diversity of contacting equipment available, it is unlikely that all these contactors will be available in any one laboratory or pilot plant. Consequently, unless test work is carried out on similar contactors, the system may not be optimized. Since mixer-settlers are the easiest to construct, are simple to operate, and require little room and low-flow rates, these contactors are, in many cases, the only ones used to investigate a continuous solvent extraction process. This is by no means ideal and may result in abandonment of a process that, using another type of contactor, could be found to be entirely satisfactory. 

Pilot plant operations, as we have noted previously, can vary between extremes of flow rates. It is necessary, therefore, that the feed and reagent volumes be large enough that the pilot plant may be operated for a sufficient length of time to obtain meaningful data. For example, if the aqueous feed to the solvent extraction (SX) circuit is being produced batchwise, variations between batches are bound to occur. Such variations should be controlled as much as possible. Batchwise production of the feed solution may be very different from feed to the actual plant, especially if the plant process involves continuous production of the feed to the SX circuit. 

The solubility of the sodium salt of DEHPA in basic (NaOH) solution has been reported, together with the effect of temperature on the water solubility of this salt [18], (Figs. 7.12 and 7.13). It is evident that the presence of salts in the aqueous phase depresses the solubility of this extractant in water (Fig. 7.11). This has been confirmed in the extraction of cobalt with DEHPA(Na) at pH 5-6, for which a solubility of the extractant was found to be <50 ppm. Furthermore, the use of DEHPA in the extraction of cobalt from an ammoniacal (pH 11) system containing sodium sulfate showed no apparent loss of extractant after 10 contacts of a DEHPA-ker-osene solvent with fresh aqueous solution [1]. Operation of pilot plants using DEHPA(NH4) and DEHPA(Na) for the extraction of cobalt, at pH 5-6 and at 60°C, showed the loss of DEHPA to be less than 50 ppm [3]. Temperature also has a significant effect on the solubility of DEHPA(Na) (Fig. 7.13). 

Much of the optimization of the solvent extraction plant can be achieved in the pilot plant testing. As noted earlier on the subjeet of proeess design, one must investigate the dependence of the dispersion and eoaleseence char-aeteristies and their effect on extraction and phase separation. Also, such variables as metal concentration, equilibrium pH (or free aeidity or free basieity), salt concentration, solvent concentration (extraetant, diluent, and modifier), and temperature have to be studied to determine their effect on mass transfer. Although many of the variables can be tested in the pilot plant, many circuits are not optimized until the full-scale plant is in operation. 

Experience in solvent extraction processes has shown that such processes can be scaled up from pilot plant—or even bench-scale—data quite reliably. This is particularly evident in processes employing mixer-settlers. However, scale-up will only be as reliable as the data on which it is based, and time spent in obtaining the correct and relevant data will always pay dividends. 

With liquid feed solutions, however, it is possible to work in a manner analogous to traditional solvent extraction. Pressurized columns can be of the packed-bed type or agitated by magnetic stirrers. Because of the efforts of pilot plant tests, much of the scale-up work has to be carried out in laboratory extractors. From solubility measurements, it is possible to determine parameters in thermodynamic models (e.g., equations of state), which can be used for the simulation of large-scale applications. 

The process flow sheet was first tested for direct leaching of steel mill flue dust and production of zinc metal by electrowinning. The tests were performed in a continuously operating pilot plant, producing 10-20 kg/day zinc metal. The same pilot plant was then used for treating copper/zinc-rich brass mill flue dust in a closed loop operation, recycling all the zinc solvent extraction raffinate to the copper circuit leach section. In the zinc circuit leach section, only the amount of zinc rich dust necessary for neutralization of the copper solvent extraction raffinate was used. The results obtained from the pilot plant tests indicated contamination problems within the solvent extraction loops. The estimation of economic data showed a weak return on the assets compared with the alkali route, and sensitivity toward the raw material price. 

All the novel separation techniques discussed in this chapter offer some advantages over conventional solvent extraction for particular types of feed, such as dilute solutions and the separation of biomolecules. Some of them, such as the emulsion liquid membrane and nondispersive solvent extraction, have been investigated at pilot plant scale and have shown good potential for industrial application. However, despite their advantages, many industries are slow to take up novel approaches to solvent extraction unless substantial economic advantages can be gained. Nevertheless, in the future it is probable that some of these techniques will be taken up at full scale in industry. 

Palladium catalyst stability, recovery and recycle are the key to viable commercial technology. Continuous palladium recovery and recycle at 99.9% efficiency is critical to the economics of the process. Traditional catalyst recovery methods fail since the adipic acid precursor, dimethyl hex- -enedioate, is high boiling and the palladium catalytic species are thermally unstable above 125 C. Because of this problem, a non-traditional solvent extraction approach to catalyst recovery has been worked out and implemented at the pilot plant scale. Since patents have not issued, process detail on catalyst separation, secondary palladium recovery, and product recovery cannot be included in this review. 

Many licensors and contractors who must obtain data from which to design and build solvent extraction plants utilize large, elaborate pilot plants which simulate commercial plant conditions. Many refiners have constructed such plants to obtain data from which to guide commercial equipment and prepare samples for engine test evaluations. Plants of this sort cost 200 to 300 daily to operate. An evaluation of a single stock may require weeks and cost thousands of dollars. 

During recent years pilot scale equipment, smaller than the prototype pilot plants described and capable of operating with exceedingly high efficiency, has been designed. Such equipment as the York-Scheibel solvent extraction tower and the Podbielniak countercurrent centrifugal mixer and extractor are typical. Data from this equipment may be correlated with commercial performance. 

From adsorption analysis data, it is possible to calculate theoretical yields of deoiled wax, asphalt, resins, and solvent extraction yields of waxy raffinate. Nearly 100 different stocks have been so evaluated by California Research Corp. during the past few years and many of these data have been correlated with operations in refinery equipment and with pilot plant operations. 

Pilot plants can be used to predict solvent dosages and other operating conditions, but such operations are expensive and should be minimized. It is, therefore, desirable to establish correlations of operating variables. Kalichevsky (16) describes correlations of solvent extraction equilibrium data which indicate that the percentage dissolved in the extract layer, L, is related to the solvent dosage, S, by the expression... 

By using Thurman s empirical formula, it was calculated that a column of 45 L of XAD-8 would be required to completely retain these compounds. This size column would require 135 L of solvent for elution. Collecting five samples at once would require a pilot plant for distillation and extraction comparable in magnitude with the water treatment pilot plant being sampled. Therefore, a decision was made to use smaller columns at the risk of losing more soluble organic compounds. 

The extraction of the PGM by amine salts was first reported some 30 years ago,235 shortly after which a process for the recovery of the PGM by solvent extraction was patented in the Soviet Union.236 The results of pilot-plant operations for the recovery of PGM from copper and nickel refinery slimes have also been described,237-23S and it is probable that such processes are currently in full-scale commercial use. Further extensive studies by Soviet workers have been described in a recent review.239 It has been shown that, in addition to the anion-exchange mechanism, e.g. 

The solvent extraction process has not yet undergone pilot plant investigation, and all the above estimates are based on small laboratory or bench scale experiments. If further testing under practical conditions substantiates the laboratory observations, it appears that the solvent extraction process definitely has an area of specialization in the over-all saline water conversion program. 

---