pH measurement and control in solvent extraction using column contractors

pH measurement and control in solvent extraction using column contractors

Minerals Engineering, Vol. 14, No. 1, pp. 13-23, 2001 Pergamon 0892--6875(00)00157--6 © 2000 Elsevier Science Lid All fights reserved 0892-6875/01/$...

567KB Sizes 54 Downloads 96 Views

Minerals Engineering, Vol. 14, No. 1, pp. 13-23, 2001

Pergamon 0892--6875(00)00157--6

© 2000 Elsevier Science Lid All fights reserved 0892-6875/01/$ - see front matter

pH MEASUREMENT AND CONTROL IN SOLVENT EXTRACTION USING COLUMN CONTACTORS*

Y. ZHUANG, A. SIEMON, N. IRELAND and G. JOHNSON Western Minerals Technology, Unit 1, 45 Edward Street, Osborne Park, WA 6017, Australia. Email [email protected] (Received 13 May 2000;accepted 3 October 2000)

ABSTRACT pH measurement and control in a solvent extraction process using column extractors is a considerable challenge. This paper summarises the progress of WMT's understanding of the emulsion characteristics, hydraulic and kinetic regimes existing in the column extractors and their effects on pH measurement and control. Three sets of pilot scale test results are provided for a case study of nicke{ solvent extraction using Versatic I0 in a column extractor, with ammonia solution as the p n modifier. Initially the pH in extraction columns was controlled by directly dosing ammonia into the column to achieve the pH set point. The pH was measured in-situ by a pH meter inserted directly into the column. It was found that the ammonia tended to flow upwards with the organic rather than downwards w~th the aqueous flow, as would be expected, in an organic continuous operation. In the second test, the barren organic was pre-equilibrated with ammonia via an in-line mixer prior to entering the column. The pre-equilibrating ratio was proportional to the nickel pregnant liquor solution (PLS) tenor and the flowrate of the various phases. The final aqueous pH was used to tune the ratio. This resulted in a significant improvement in control of the pH in the column. In the third test, a combination of pre-equilibration and direct ammonia injection was used. The pH control in the column was very stable and a desirable pH profile in the column was achieved to maximise nickel recovery and reduce impurity co-extraction. © 2000 Elsevier Science Ltd. All rights reserved.

Keywords Solvent extrac,tion; pH control; process control

INTRODUCTION Solvent extraction (SX) has been used in the minerals processing industry since the 1950s. It involves contacting two imrniscible phases, usually an organic phase containing an extractant and an aqueous phase containing the metal values. With energy input into the SX system, the two immiscible phases form an emulsion to provide the interfacial area for chemical reaction or mass transfer. As a result, the metal ions are extracted from the aqueous phase into the organic phase to achieve separation, purification and concentration. The process usually involves chemical reactions.

* Presented at Hydromet 2000, Adelaide, Australia, April 2000

13

14

Y. Zhuang etal.

In any solvent extraction reaction that involves hydrogen or hydroxide ions, either as a product or reactant, control of the aqueous pH can be effectively used to adjust the equilibrium position of the reaction. For example, some of the commonly used extractants can be represented by the formula HR, where HR can be oximes, carboxylic acids, sulfonic acids, organophosphoric acids etc. The reaction of the extractant with metal ions is shown in Eq (1) below:

(1)

nHR(org) + Mn+(aq) ~ MRn(org ) + nH+(aq)

By adjusting the pH, the position of the equilibrium can be altered and as a consequence the following process parameters can be controlled: Metal recovery The extent of metal extraction is dependent on the equilibrium pH of the aqueous solution. It can be seen from Eq (1) that with increasing hydrogen ion concentration, the equilibrium position shifts to the left, and the metal ions in aqueous solution will stop loading onto the extractant. To keep the reaction proceeding to the right, alkaline reagents, such as sodium hydroxide or ammonia, are added to the system to neutralise the acid generated during extraction. Selectivity The equilibrium of the reaction is also dependent on the specific metal cation involved. By changing the equilibrium pH, different metals can be selectively extracted using the same extractant. For example, as shown in Figure 1, Cyanex 272 can be used to selectively extract zinc at pH 3.5, and cobalt at pH 5.5. Cyanex 272 is commonly used to effect a separation between nickel and cobalt from a leach solution containing both nickel and cobalt ions. ZR 2+

CO 2.

Ni2.

c 0

X o o =E I

I

I

I

2

4

6

8

pH Fig. 1 Extraction of metals by Cyanex 272 from sulfate solutions. Precise pH control is also necessary to limit the co-extraction of impurities, such as calcium and magnesium in the case of nickel extraction using Versatic 10.

Solubility of solvent Some extractants, for example Versatic 10, will form a salt by reacting with an alkali such as NaOH or NH3. The salt form of Versatic 10 is capable of partially dissolving in the aqueous phase. The solubility of Versatic 10 dramatically increases with increasing pH. This was once considered an impediment to the use of this extractant, particularly for the extraction of nickel where, at the pH required for extracton, the solubility of Versatic 10 in aqueous phase is significant. Tight control of the pH is necessary to operate this solvent extraction system, otherwise the depletion of Versatic 10 from the organic phase will be significant. Even with good pH control, a further extractant recovery step is needed to recover the dissolved extractant.

pH measurementand controlin solventextraction

15

According to Burkin (1987), the organic recovery stage has been used in the SMM Process developed by Sumitomo Metal Mining Company Limited. A dilute acid solution with 5% sulfuric acid was used to reacidify the rafflnate to lower the pH, which converted the dissolved Versatic back to the free acid form. It can then be collected in organic form for reuse in the circuit.

pH MEASUREMENT IN SOLVENT EXTRACTION SYSTEMS To achieve effective control of a solvent extraction process based on the chemistry described above, fast and reliable pH measurement is essential. For the solvent systems under consideration here, the extent of metal extraction, and the selectivity achieved, are directly related to the pH of the aqueous phase. There are two possible ways in which to measure the pH of the reactions in question, (1) by inserting the pH probe directly into the emulsion phase (be it in a column type contactor or the mixer box of a mixersettler combination) and correlating the measured emulsion pH with that of the aqueous pH, (2) by removing the mixed phase emulsion from the system and separating it into its components in some way (for example in an on-line settler), and measuring the pH of the aqueous phase. According to Kordosky et al. (1981), directly measuring the emulsion pH offers a significant practical advantage in that i.t gives the fastest response to the pH changes caused by the extraction reactions. This is obviously important if fast and accurate pH control is critical. The second method needs extra equipment to separate the emul,;ion, and acts only on a sub-sample of the main process flow. Also, it adds a lag time to the control loop as the external separation of the phases takes some time. For the first method to be effective, the correlation between the measured pH of the emulsion and the actual pH of the aqueous phase needs to be well understood. There are many factors that need to be considered in establishing this correlation.

Mixed phase emulsion pH and aqueous pH To measure the pH of a SX emulsion, the hydrogen ions in the aqueous phase have to diffuse through the glass membrane of the pH probe. If the organic phase covers part or all of the surface of the membrane, the pH reading may not reflect the true hydrogen ion concentration in the aqueous phase. In many practical solvent extraction systems it has been observed that there is an offset between the measured pH of the mixed emulsion phase and that of the separated aqueous phase. This offset is not necessarily a constant value. Several of the factors affecting the offset between the measured emulsion pH and actual aqueous pH are related to the nature of the dispersion in which the measurement is being undertaken. These include the phase continuity and the extent of dispersion.

Phase continuity There are two tyl:,es of phase dispersion in forming an emulsion, i.e., organic continuous (or water in oil, W/O) and aqueous continuous (or oil in water, O/W) emulsion. For an in-situ pH measurement, in an organic continuous emulsion, the probe is only in contact with droplets of aqueous phase, as in Figure 2. Reference electrode

Glas~ nrnhe

Fig. 2 pH probe in W/O emulsion

16

Y. Zhuanget al.

In an aqueous continuous emulsion, the probe is in contact with a continuous aqueous phase, as in Figure 3.

Reference electrode

Glass nrobe Organic

Aqueous

Fig. 3 pH probe in O/W emulsion. In Ni solvent extraction using Versatic 10, piloting tests have been conducted at WMT's demonstration plant using a column extractor, which can be operated with either organic or aqueous as the continuous phase. The pH of the emulsion when measured in-line, and the off-line measurement of the aqueous pH, for the two different continuities, is compared in TABLE 1.

TABLE 1 The comparison of emulsion and aqueous pH at W/O and O/W operation in column

Organic Continuous Emulsion pH AqueouspH 7.40 7.90 6.67 7.38

6.0 6.4 2.34 2.89

@ pH 1.40 1.50 4.33 4.49

Aqueous Continuous Emulsion pH AqueouspH @ pH 5.60 7.10 7.40 7.17

6.04 7.40 8.02 7.57

0.44 0.30 0.62 0.40

It was found that the difference between emulsion pH and aqueous pH is more significant when the organic phase is continuous than when the aqueous phase is the continuous phase.

Extent of dispersionmmixing intensity The energy input or mixing intensity will affect the drop size and holdup of the dispersed phase. With the increase of energy input, the holdup of the dispersed phase in the emulsion increases and the drop size decreases. For an in-situ pH measurement, it is necessary to have a reasonable amount of aqueous phase in the emulsion to obtain a reliable aqueous pH measurement. The selection of a suitable type of pH probe is also very important.

pH Sensors The measurement of pH in solvent extraction systems is usually undertaken with a glass membrane pH electrode. The pH-sensing electrode consists of a glass membran~ which is sensitive to hydrogen ions, and a reference junction. Both the glass membrane and the referencejunction are prone to errors when used to measure the pH of systems containing some organ!c phase. It is beyond the scope of this paper to describe these effects in detail, but it needs to be recognised that the organic phase does have an unpredictable effect on pH measurement. It has also been observed that there are significant differences between the performance of pH probes from different manufacturers, in particular their stability through time.

Type of contactor Kordosky e t al. (1981), described the control of pH in mixer-settlers operation. There is little information available in the literature regarding the pH control in columns. Based On WMT's understanding from the

pH measurementand controlin solventextraction

17

operation of both mixer-settlers and columns in WMT's pilot plant, there are a number of differences in the hydraulic and chemical regimes that exist in mixer-settlers and column extractors. Understanding these differences is important when designing a pH control system for either application. Mixer-settler

In a mixer-settler the designer aims at achieving perfect mixing of the two phases in the mixer box, as well as providing enough retention time for the mixture to approach equilibrium. The mixer, of a mixer-settler combination, is usually operated at an O:A ratio of 1:1 by recycling either organic or aqueous phase (as appropriate). Thi,; is primarily to achieve good contact between the phases and prevent phase inversion if the advancing flowrates of organic and aqueous are very different. The result is that there is always a significant portion of aqueous in the mixed phase emulsion. The combination of phases close to equilibrium, together with a large volumetric fraction of aqueous, greatly assists in making reliable pH measurement, especially if the mixture is aqueous continuous. When themixer runs organic continuous, the pH measurement is always more problematic, as the pH probe has to contact with droplets of aqueous rather than the continuous phase. Despite this limitation it is possible to make reasonable measurements of the aqueous phase pH using a mixer-settler. The results in TABLE 2 show that after an extended period of mixing, when the organic and aqueous phases approached equilibrium, the emulsion pH and aqueous pH were very close for both aqueous and organic continuous operation. TABLE 2 Comparison of the pH of Emulsion and Aqueous Phase Using a Mixer pH Reading (O:A =1:1~ M i x i n g time(s) 0 60 120 180 240 0 60 120 180 240

Co ntinuity

Emulsion

O/W W/O O/W W/O

3.11 1.90 2,45 2.40

O/W O/W W/O O/W

5.51 5.48 4.78 4.65

Aqueous phase 3.50 2.21 2,20 2,35 2.36 7.00 4.64 4.65 4.79 4.76

Organic phase 2.12 2.06 2.22 2.28 4.63 4.64 4.65 4.66

The situation where the reaction approaches equilibrium and the operating O:A ratio is close to 1:1 in the emulsion occurs in the mixer-settler, but generally does not occur in the column. Column extractors

A column extractor is a differential multi-stage extractor. In a continuous courttercurrent column design, the organic and aqueous phases flow simultaneously in opposite directions. Mixing is achieved by a combination of external energy input, the pulsing action, and the design of the column internals. Within the column the extent of m ~ l extraction exhibits a continuous profile rather than the stage-wise extraction equilibrium of a mixer-settler train. Metal extraction along the column is driven by the displacement of the reaction system from equilibrium. This means that under steady state conditions the phases may not reach equilibrium at any point in the column within the mass transfer height required by the number of theoretical unit (NTU). If the column is operated organic continuous, the aqueous holdup under certain energy input levels can be as low as less than 5%. Under these conditions, the inline measured emulsion pH can be very different from that of the 5% aqueous phase in the emulsion. This makes it very difficult to obtain reliable aqueous pH measurement and control using inline pH measurement. The suitability of running organic continuous in a column, which needs very close pH control, has to be carefully considered in this situation.

18

Y. Zhuangetal. pH MODIFIER ADDITION Case study of ammonia addition in nickel solvent extraction using Versatic 10

There are two ways to add the pH modifier- ammonia into the nickel solvent extraction system. The first method is direct injection of ammonia into the emulsion. The second method is a pre-equilibration of the barren organic with ammonia. This is performed through an in-line mixer in the organic feed line prior to injection into the column. Different ammonia addition methods lead to different metal extraction mechanisms. The ammonia mixing and flow pattern in different contactor (mixer-settler or column), or under different phase continuity, will be different.

Extraction mechanism If ammonia is directly injected into the emulsion, the ammonia is to be utilised for neutralising acid in the aqueous phase. According to Ritcey et aL (1984), assuming that the ionisation of extractant occurs in the aqueous phase, the metal extraction reaction mechanism is proposed as follows: HR(o~g)~HR(,q)

(2)

2HR(~q) + Ni2+(~q)~ NiR2(~q)+ 2H+(~q)

(3)

NiR2(aq) ~ NiR2(org)

(4)

NH3(aq) + H+(aq)~NI"I4+(aq)

(5)

If the ammonia is used to pre-equilibrate the barren organic, the metal extraction reaction mechanism is proposed as follows: NH3(aq) + HR(org)~ NH4R(org)

(6)

NH4R(org) ~NI-hR(~q)

(7)

2 NH4P~aq)+ Ni2+(aq) -~ NiR2(aq) + 2NI-I4+(aq)

(8)

NiRE<~q)~ NiR2(o~g)

(9)

It can be seen from the above equations that in the ammonia pre-equilibrated extraction, the aqueous reaction does not generate an acidic product. There is no pH change in the extraction reaction.

Ammonia flow pattern in pulsed columns under different operating phase continuity The ammonia flow and mixing pattern in the column is different under organic continuous and aqueous continuous operation. Figures 4 and 5 illustrate the ammonia flow and mixing under different conditions.

Densities of organic and aqueous solutions The reasons that the ammonia flows upward with the organic phase rather than downward with the aqueous phase, could be explained as follows: the density of ammonia is close to that of the organic phase (refer to Table 3). In an organic continuous environment, when a small amount of ammonia solution is dispersed into the organic continuous emulsion, the ammonia solution droplets will flow upwards with the organic unless they get the chance to meet and mix with the aqueous phase droplets. In an aqueous continuous operation, as soon as the ammonia solution enters the column, it will mix with the aqueous PLS, which has a heavier density, and flow downwards along the column.

pH measurementand controlin solventextraction

19

)rganic

NH 3 Injection

Ammonia

Aqueous

Fig. 4 Ammonia mixing in the column--organic continuous.

OUS

At

3rganic

Fig. 5 Ammonia mixing in the column--aqueous continuous If the ammonia is added discretely through multiple injection points, the amount of ammonia added in one section is that required to maintain the pH in the desired range for metal extraction within this section of column. The pH a!long this section of column will therefore exhibit a profile related to the extent of the reaction that occurs. This raises the question of where to position the pH probe relative to the injection point as well as the value that should be assigned to the set point of the pH controller. TABLE 3 Densities of organic and aqueous solutions

Sample Description Ni organic solvent at 25 C Co organic solvent at 25 C A m m o n i a solution 20% at 25 C A m m o n i a solution 32% at 15 C Co PLS and Ni PLS at 25 C

Density (kg/m 3) 824 920 930 889 1100 - 1200

20

Y. Zhuanget al. PILOT PLANT TESTS USING COLUMN EXTRACTORS

In WMT's pilot plant, the nickel solvent extraction using Versatic 10 has been conducted in column extractors. The pH control in the extraction columns was developed over three trials: 1. Multiple injection with pH probe control. 2. Single point injection with pre-equilibration in the barren organic phase. 3. Combination of a pre-equilibration injection point with direct ammonia injection points utilising pH control within the column.

Multi-point ammonia injection There were three ammonia injection points along the column in Test 1. Three pH probes, situated below each of the injection points controlled the pH. Ammonia was directly injected to the column by a dosing pump regulated by the pH probe (Figures 6 and 7).

Loadedorganic I,

Scrub Soln

pH Top PLS Ammonia pH Upper

Ammonia pH Middle

Ammonia

~H Lower

Barrenorganic~I .)

Aqueous P~aff

Fig. 6 Multi-point ammonia injection to column.

pH top

H upper H middle pH lower pH Fig. 7 pH profile of the column. The pH control was erratic due to many reasons such as unstable level control of the bottom decanter. However, the main reason was that the ammonia flowed upwards and the ammonia injection pump was controlled by the pH probe situated underneath the injection point. This means that when the probe reading was below the set point, for example pH 6.5, the pump starts to add ammonia, but the ammonia flows upwards instead of downwards, the pH below the injection point did not increase, and so the pump kept pumping, causing an overdose of ammonia.

pH measurementand controlin solventextraction

21

The overdosing of ammonia caused a very high pH on the top section of the column, upsetting the scrubbing section of the column. At the high pH, calcium and ammonium were co-extracted into the organic then stripped into the nickel electrolyte in the strip column. Ammonia could be physically entrained by the organic or chemically bonded as the salt form of Versatic 10. The calcium formed gypsum precipitate under supersaturating conditions.

Pre-equilibration of barren organic with ammonia Since the ammonia tends to flow upwards in an organic continuous operation, it was decided to add ammonia to the b~a'ren organic. The barren organic was pre-equilibrated with ammonia via an in-line mixer and injected into the column at the lower decanter. The amount of anlmonia addition was regulated by a controller to achieve a pH set point at the lower section of the column (Figures 8 and 9).

Loaded organic Scrub Soln

P pH top

PLS pH upper

pH middle pH lower ~

Barrenorganic Aqueous ~aff Fig. 8 Pre-equilibration of barren organic.

H top per middle pH lower

pH Fig. 9 pH profile of the column. The performance of pH control using the pre-equilibration method was stable. The raffinate pH levels were maintained at appropriate set points, and the pH shows the desired profile along the column: lower at the top section for scrubbing and higher at the lower section to ensure maximum metal extraction.

A combination of pre-eqnilibration and direct ammonia injection In this testwork, ~the amount of pre-equilibrating ammonia added to the barren organic was controlled by the lower pH probe through a ratio controller. The second ammonia addition was at the middle of the

22

Y. Zhuanget al.

column controlled by the upper pH probe situated above the ammonia injection point through a PID controller. The scrub section was removed from the extraction column, and scrubbing was performed externally in a mixer-settler (Figures 10 and 11).

I Loadedorganic PLS

pHlop pHupper Ammonia oH middle

pHlower Ammonia .

Barren organm~I I AqueousFIilaff Method 3

Fig. 10 Pre-equilibration and direct injection.

eX

pH top X ~ p H upper ~x~H middle

-

.

p.:low

pH Fig. 11 pH profile of the column. The pH control was reasonably stable. About 80% of the ammonia required was added into the middle of the column and 20% was used in~pre-equilibrating the barren organic. The pH profile in the column was lower at the top section for scrubbing co-extracted impurities, and higher at the lower section to achieve maximum nickel extraction.

CONCLUSIONS It can be concluded from the above discussion that the pH measurement in the solvent extraction system depends on many factors, such as the type of pH sensor, the operating continuity, the energy input or extent of dispersion and methods of pH modifier addition. The design and operation of different types of extractors, for example mixer-settlers and columns create a very different hydraulic and chemical environment. To develop a robust solvent extraction Control system, a deep understanding of the physical and chemical interaction of the process is required.

pH measurementand controlin solventextraction

23

REFERENCES

Burkin A. R., Extraction Metallurgy of Nickel, 1987, John Wiley & Sons, pp. 98-147. Kordosky, G., A., Champion, W, H., Dolegowski, J., R., Olafson, S., M., Jensen, W., S., Use of pH Control in Solvent Extr~LctionCircuits, Mining Engineering, 1981, March, 291-299. Ritcey, G., M., Ashbrook, A., W., Solvent Extraction Principles and applications to process metallurgy, 1984, Elsevier, pp. 12.

Correspondence on papers published in Minerals Engineering is invited by e-mail to bwills @min-eng.com