2/2/2021 Separation of Cobalt

and Nickel using

CYANEX 272 for
Solvent Extraction
In the presence of other contaminates
in sulphate based leaching solution

Caroline Kihlblom
BOLIDEN MINERAL AB
Supervisor: Mohammad Khoshkhoo
Examiner: Kerstin Forsberg
 2021-04-07

Abstract
This project aimed to examine the separation of cobalt and nickel using solvent
extraction (SX) with the extractant CYANEX 272 (C272). It was intended to
investigate the Co-Ni separation in a sulphate-based leach solution in presence of
other contaminants. This is an area of interest because of the difficulty of separating
metals of similar properties within the field of hydrometallurgy.

Batch tests, with varying modifiers and diluents, were carried out to examine the
effect of organic phase composition on phase separation. The effect of pH on
equilibrium was investigated by constructing equilibrium curves. Through various
shaking tests, different separation parameters were studied. McCabe-Thiele
diagrams were constructed to predict design parameters. In order to simulate a
continuous 3-stage countercurrent solvent extraction, batch tests were performed.
Scrubbing, as means of impurity removal was also investigated. Finally, the
product’s purity was examined by the help of crystallization.

The organic feed mixture that resulted in a sufficient phase separation consisted of
C272, tributyl phosphate and naphtha. At pH 4, equilibrium curves showed that
equilibrium was either not reached or affected by competing metal ions. A standard
equilibrium curve appearance was seen at pH 4.5, resulting in that the theoretical
required stages for extraction was calculated to 3 stages (A/O=1). However, a
McCabe-Thiele diagram did not give an accurate representation of the more
complex case (presence of contaminants). Batch simulation results gave a cobalt
recovery of 69% and 100% at pH 4.5 and 4.8, and a nickel recovery of 0% and 3%,
respectively. A recommended pH-value for solvent extraction could not be stated,
because the choice must be based on operation specifications. Therefore, several
different aspects (Co recovery, purity, and economical etc.), must be accounted for.
A similar pH-trend was shown in scrubbing, where an increase of pH resulted in an
increase of metal ions’ organic concentration. For stripping, acid test results proved
24 g/L sulphuric acid to give the highest cobalt concentration, with a marginal
difference in concentration of impurities.

An overview of the entire SX process, indicated that extraction, scrubbing, and

stripping were all successful operations. The extraction stage showed a Co and Ni
recovery of 99% and 0.02%, respectively, and a separation factor of 14250.
Distribution results indicated that Al was difficult to remove and was transferred
with Co into the product. Therefore, this element must be removed before SX. From
noticing an increase of Co:Ni ratio throughout the process, solvent extraction was
considered an effective separation method for cobalt and nickel separation. A
considerably high purity of cobalt sulphate was produced. However, impurities Al
and Ca were also detected in the product. Increasing the acetone volume in
crystallization resulted in an increase of Co purity. An increase of the cobalt
sulphate crystals formed was observed when increasing the acetone volume, where
no impurities were detected.

Key words: Solvent extraction, Cobalt-Nickel separation, CYANEX 272, pregnant leach solution

1 (48)
 2021-04-07

Table of Contents
1 INTRODUCTION ............................................................................ 4
1.1 Aim and objectives .......................................................................... 5
1.2 Scope .............................................................................................. 5
1.3 Boliden mineral ab .......................................................................... 6
2 BACKGROUND .............................................................................. 7
2.1 Solvent extraction ........................................................................... 7
2.1.1 Extraction ........................................................................................ 7
2.1.2 Scrubbing ........................................................................................ 7
2.1.3 Stripping .......................................................................................... 7
2.1.4 Solvent Extraction Chemistry .......................................................... 8
2.2 Seperation of cobalt and nickel ....................................................... 8
2.2.1 Solvent extraction in sulphate solutions .......................................... 8
2.2.2 Selectivity ........................................................................................ 9
2.2.3 Extractant and diluent ................................................................... 10
2.2.4 Impurities ...................................................................................... 12
2.3 Definitions ..................................................................................... 13
2.3.1 Distribution coefficient ................................................................... 13
2.3.2 Separation factor........................................................................... 13
2.3.3 McCabe-Thiele diagram ................................................................ 14
3 METHODS AND MATERIALS ..................................................... 16
3.1 Materials ....................................................................................... 16
3.2 Experimental set-up ...................................................................... 17
3.2.1 Process overview .......................................................................... 17
3.2.2 Effect of saponification and organic phase composition ............... 18
3.2.3 Effect of pH on extraction equilibrium ........................................... 19
3.2.4 Batch Simulation of Continuous Multistage Countercurrent Solvent
Extraction ...................................................................................... 19
3.2.5 Scrubbing tests ............................................................................. 21
3.2.6 Stripping tests ............................................................................... 21
3.2.7 Crystallization of Cobalt ................................................................ 21
3.2.8 Analytical methods ........................................................................ 22
4 RESULTS AND DISCUSSIONS ................................................... 24
4.1.1 Effect of saponification and organic phase composition ............... 24
4.1.2 Effect of pH on extraction equilibrium ........................................... 26
4.1.3 Batch Simulation of Continuous Multistage Countercurrent Solvent
Extraction ...................................................................................... 29
4.1.4 Scrubbing ...................................................................................... 33
4.1.5 Stripping ........................................................................................ 34
4.1.6 Solvent extraction overview .......................................................... 35
4.1.7 Overall distribution and process summary .................................... 38
4.1.8 Crystallization of Cobalt ................................................................ 39
4.1.9 Source of error .............................................................................. 40

2 (48)
 2021-04-07

5 CONCLUSION .............................................................................. 41
5.1 Future research............................................................................. 43
6 ACKNOWLEDGMENTS ............................................................... 44
7 BIBLIOGRAPHY .......................................................................... 45
8 APPENDIX.................................................................................... 47
8.1.1 Effect of pH on extraction equilibrium ........................................... 47
8.1.2 Batch simulation of continuous multistage countercurrent solvent
extraction ...................................................................................... 47
8.1.3 Stripping ........................................................................................ 47
8.1.4 Crystallization of cobalt ................................................................. 48

Table of Figures
Figure 1: pH-dependency of Co, Ni and Ca extracted in a sulfate solution using
CYANEX 272 ....................................................................................................... 10
Figure 2: Molecular structure of Di (2,4,4-trimethylpentyl) phosphinic acid
(CYANEX 272)11 .................................................................................................. 11
Figure 3: Different metal ion extractions (in %) with Cyanex 272 vs. pH 9 ......... 12
Figure 4: Schematic of a McCabe-Thiele diagram5 .............................................. 15
Figure 5: A flowsheet of the overall experimental solvent extraction design....... 17
Figure 6: Batch simulation of 3-stage countercurrent solvent extraction. The red
and black lines represent the aqueous and organic phase, respectively. Note that
SX A3 was not conducted, but is presented for the sake of visualization. ........... 20
Figure 7: Continuous countercurrent solvent extraction, where the aqueous
raffinate and organic extract are denoted with an R and E, respectively. The
symbol F denotes the fresh feed in and S denotes fresh extractant. ...................... 20
Figure 8: A photograph representing test 2, where a blue color is seen in both
phases .................................................................................................................... 24
Figure 9: The organic feed tests, test 8-10, from left to right ............................... 25
Figure 10: Plotted equilibrium curve for cobalt at pH 4.0. ................................... 26
Figure 11: Equilibrium curve for cobalt at pH 4.0 with a lower aqueous to organic
ratio. Notice that the red points represent Co ions being pushed out .................... 27
Figure 12: McCabe Thiele diagram for cobalt at pH 4.5 with a slope of 1 .......... 27
Figure 13: The metal ions’ concentration in the aqueous phase vs. batch
simulation cycles, for cobalt and nickel ................................................................ 29
Figure 14: Metal ion recoveries into organic phase vs. cycles at pH 4.5 .............. 30
Figure 15: Metal ion recovery into organic phase vs. cycle, at pH 4.8 ................. 32
Figure 16: Schematic of experimental design with Co and Ni in-and-out streams
presented with unit g/L, and with A:O=1 ............................................................. 35
Figure 17: A photography of all the formed crystals from crystallization ............ 48

3 (48)
 2021-04-07

1 INTRODUCTION
At Boliden’s mine in Kylylahti, Finland, ore is being processed and mined to
produce different kinds of concentrate, namely copper, zinc, and Knelson gravity
concentrate. The flotation tailings, also called pyrite- or sulphur concentrate, which
contains significant amount of pyrite, is deposited in lined ponds. In the pyrite
concentrate, cobalt (Co) and nickel (Ni) can be found, which is of high interest for
the metal company Boliden. The reason for interest of recovering of these two base
metals from the sulphur concentrate is their corrosive, resistance, and magnetic
properties, alongside their high economic value due to their applications in
manufacturing of electrical vehicles batteries. To obtain as much of the desired
product as possible, Boliden must oxidize an enormous amount of pyrite first,
which is done by bioleaching, followed by a series of purification and separation
processes. The leach solution consists of Co and Ni, alongside the main impurities
Ca, Al, Mg and Mn.

However, it is difficult to separate cobalt from nickel in a sulphate-based solution,

because of their similarity in properties. Therefore, it is of great importance to use
an effective separation method for this purpose. Applying solvent extraction by
using CYANEX 272 has shown good separation of Co and Ni in similar systems
and will therefore be used in this project.

In this master thesis project, different conditions will be investigated in batch tests,
in order to observe the effect on separation. To find out how many separation stages
are required for the specific separation process, design parameters will be
investigated by using theoretical calculations. An evaluation of where metal ions
end up after solvent extraction will also be studied.

The main research questions for this project are:

• What effect does organic phase composition have on phase separation?
• What effect does pH have on equilibrium?
• How many minimum stages are required for the separation process?
• How are other elements distributed throughout solvent extraction?
• What is the purity of a possible cobalt sulphate product?

4 (48)
 2021-04-07

1.1 AIM AND OBJECTIVES

The aim of the project is to investigate the separation of cobalt and nickel by
employment of solvent extraction using the extractant CYANEX 272. The
possibility of producing a pure cobalt sulphate solution from a sulphate-based leach
solution using this separation method will be studied in the paper.

In order to achieve the aim, some objectives are formulated, which are as follows:
• Planning and design of the experimental work by the means of a literature
study.
• To test different diluents and modifiers to find a suitable organic feed
composition for phase separation
• To examine different separation parameters (pH and acidity) to find
recommended values for obtaining maximum cobalt recovery without co-
extraction of impurities.
• To investigate how many extraction stages are required for the operation,
by using McCabe-Thiele diagram.
• Conducting an analysis of metal ions and contaminants using ICP-OES and
XRF, respectively.

1.2 SCOPE
The project scope was divided into three different parts. The first part consisted of
performing batch tests on the leach solution under different conditions, in order to
verify the effect on phase separation. This was done in lab scale in Boliden’s pilot
facility.

After the initial tests were conducted design parameters were investigated by the
help of a McCabe-Thiele diagram. This method was used to calculate the required
stages for the extraction process. Batch tests were conducted in this part as well, but
to simulate a continuous multistage countercurrent solvent extraction.

The third part consisted of analyses of metal ions in different streams. Through
calculating various mass balances, distribution of elements throughout the process
were studied. Impurities were studied closely, to justify if pure cobalt sulphate
could be obtained.

5 (48)
 2021-04-07

1.3 BOLIDEN MINERAL AB

Boliden Mineral AB has requested that this master thesis project is conducted at
one of their many facilities, in the Boliden area. Boliden Mineral AB, also known
as Boliden, is a well-known Swedish metal company which mines, produces and
sells valuable metals to the world’s metal market. For more than 90 years, the
company has been exploring, extracting, processing, and selling base- and precious
metals from their many mines and smelters. They have over five mines distributed
across Sweden, Finland and one mine in Tara, Ireland. Whereas, their smelters are
located in Sweden, Finland and Norway, which accounts for a total of five smelter
facilities. Their head office is situated in Stockholm. Boliden is today the European
leader of producing the base metals copper and nickel and one of the world’s
leading zinc producers. 1

6 (48)
 2021-04-07

2 BACKGROUND
2.1 SOLVENT EXTRACTION
Solvent extraction (SX) is a separation process often used in hydrometallurgy to
obtain a highly purified and concentrated metal. This specific separation technique
was introduced in connection with the Manhattan project in the early 1950s to
produce pure uranium. Later, the first commercial copper solvent extraction plant
was introduced in 1968. After that, the interest for solvent extraction grew as a unit
operation used for the recovery of a wide range of metals. The separation process
is constantly being improved by development of new extractants and equipment.2

Solvent extraction consists of contacting an aqueous phase, containing the metal of

interest, with an organic phase, containing an extractant. The metal of interest is
transferred from the aqueous phase to the organic phase after the phases are put in
contact via agitation. The phases can separate due to their immiscibility and their
difference in density. The metal chemically reacts with the extractant to form a
metal complex in the organic phase, while the aqueous phase often contains
remaining impurities. By reversing the chemical reaction, it is possible to recover
pure metal in a highly concentrated aqueous solution. Solvent extraction consists of
two operations, extraction and stripping, which are described more thoroughly
below.3 4

2.1.1 Extraction
In the extraction circuit, metal ions from the aqueous phase (SX feed) bind to the
extractant in the organic phase when they are put in contact. By the means of
stirring, the two phases are mixed, allowing them to separate. From the extraction
step a loaded organic phase is produced together with an aqueous phase called a
barren or raffinate. The raffinate is then either recycled back to the SX feed or
discharged for further treatment, while the loaded organic continues to the stripping
circuit.3–5

2.1.2 Scrubbing
If impurities are transferred to the loaded organic, it may be required to add a
scrubbing stage before the stripping step. This is in order to wash and remove all
metal impurities by treatment with a fresh scrub solution or a bleed of the recycled
strip liquor.3–5

2.1.3 Stripping
In the stripping stage, the valuable metal is recovered by reversing the chemical
reaction in the extraction circuit. In other words, the metal complex in the organic
phase is back extracted (stripped) into the aqueous phase by using a suitable
solution. The product after stripping is a strip liquor containing a high concentration
of pure metal. Here, the raffinate is the organic strip solvent, which can also be
recycled back to the extraction step.3–5

7 (48)
 2021-04-07

2.1.4 Solvent Extraction Chemistry

On a chemical level, the solvent extraction process is an equilibrium process,
distinguished by three different type of extraction mechanisms i.e. cation-, anion-
and solvation extraction. In this report, only cation extraction mechanism will be
focused on and described in more details. The extractant, the extracted metal ions
and the chemical environment determine which type of extraction process is
obtained.

In cation extraction mechanism, the metal ions (e.g. Cu2+, Fe3+, Co2+) are mainly
exchanged with an acidic hydrogen from the extractant. The extractant used for this
purpose is for example alkyl phosphoric acids, sulphonic acids and carboxylic
acids.6 The cation extraction, for both extraction and stripping, is a pH-dependent
equilibrium process and is described by the following formula:

𝑀𝑒 2+ (𝑎𝑞) + 2𝑅𝐻 (𝑜) ↔ 𝑀𝑒𝑅2 (𝑜) + 2𝐻 + (𝑎𝑞)

Where, Me and R represents the metal and extractant functional group, respectively.
The formula indicates that a high pH value is desired for extraction, whilst the
opposite is true for stripping i.e. a low pH value favors the reaction.7 3

2.2 SEPERATION OF COBALT AND NICKEL

In hydrometallurgy, the separation of cobalt and nickel in aqueous solutions has
always been problematic. The reason is their aqueous chemical similarities, due to
that they are positioned adjacent to each other in the periodic table (in the transition
metal series). Therefore, an effective separation technique is required (a high
separation factor) to achieve this separation. To separate the two base metals, one
takes advantage of their difference in water exchange rate of the metal ions, where
cobalt ions have a much higher rate of water exchange compared to nickel ions.
This implies that complex ions form more easily for cobalt.2

Today, a commonly used separation process for cobalt-nickel separation from

sulphate solutions is solvent extraction (SX). This separation alternative is preferred
over the traditional separation processes, such as selective oxidation and
precipitation. The explanation for the choice is the high degree of separation (high
separation factor) that SX offers, together with its ability to fulfill the industry’s
high separation demand2. Another advantage of using solvent extraction for the
separation of cobalt from nickel is its ability to separate metals that have very
similar properties. The separation process has a much higher degree of separation
for cobalt and nickel compared to for instance crystallization or precipitation8.

2.2.1 Solvent extraction in sulphate solutions

Solvent extraction for sulphate-based leach solutions involves mixing of the leach
solution (aqueous phase) with a liquid organic (organic phase), where the leach
solution contains Co and Ni and the organic phase contains an extractant. The cobalt

8 (48)
 2021-04-07

ions in the liquid phase are transferred to the organic liquid, forming an organic
complex. In other words, cobalt is extracted from the leach solution leaving nickel
in the aqueous phase. Later, the loaded organic, which is loaded with cobalt, can be
separated from the aqueous solution by means of gravity. The specific gravity of
the aqueous- and organic phase is 1.1 and 0.85, respectively, resulting in the organic
phase floating on top of the aqueous phase. The raffinate, containing mostly nickel,
is further processed, while the loaded organic is sent for stripping in the stripping
circuit. In the stripping stage, the loaded organic is mixed with a suitable solution,
for instance sulfuric acid, and the revers reaction to the extraction step takes place.
The product from stripping is the aqueous phase, enriched in cobalt, while the
organic phase is recycled back to the extraction circuit.6

Solvent extraction of cobalt from sulphate-based solutions is commercially

available today. At, for instance, Norilsk Nickel in Harjavalta, Finland, Minara
Resources in Murrin Murrin, Australia; and Sherritt in Toamasina, Madagascar.
Through extraction, a large amount of pure cobalt and nickel is produced each year.
The extraction reaction is described by:

2𝐻2 𝐴2 (𝑙) + 𝐶𝑜𝑆𝑂4 (𝑎𝑞) → 𝐶𝑜𝐴2 ∙ 𝐻2 𝐴2 (𝑙) + 𝐻2 𝑆𝑂4 (𝑎𝑞)

, where A represents 2,4,4-trimethyl phosphonic conjugate base in the organic

phase. The stripping reaction is described by reversing the extraction reaction. A
low concentration of sulfuric acid is preferred in extraction, whilst the opposite is
true for stripping i.e. a high concentration.6

2.2.2 Selectivity
The selectivity of a specific separation determines the outcome from the separation
process i.e. extraction and stripping. Consequently, one must be aware of the
system’s selectivity, in order to control it more easily. For separation of cobalt from
nickel in a sulphate solution the selectivity, when using an alkyl phosphorous acid,
is dependent on:

➢ Temperature
➢ pH
➢ Acid type
➢ Co concentration
➢ Diluent
➢ Modifier

Also, the separation factor for alkyl phosphoric acids changes with an increase of
10-1000 in the given order:

𝑝ℎ𝑜𝑠𝑝ℎ𝑜𝑟𝑖𝑐 < 𝑝ℎ𝑜𝑠𝑝ℎ𝑜𝑛𝑖𝑐 < 𝑝ℎ𝑜𝑠𝑝ℎ𝑖𝑛𝑖𝑐 𝑎𝑐𝑖𝑑𝑠

9 (48)
 2021-04-07

The increase of separation factor can be explained from the change of the cobalt
complex’s nature. In chemical terms, the hydrated octahedral complex changes into
the anhydrous tetrahedral complex.2

When a cation exchange chemistry exists, extraction and stripping is pH-

dependent7. Therefore, it is important to control the pH in order to obtain specified
process requirement. Hence, selective extraction is possible if the equilibrium
concentration of the two metals ions are substantially different at a specific pH-
value8.

Earlier studies have tested the effect of pH on the Co-Ni selectivity when using
CYANEX 272 as extractant, which is demonstrated in Figure 1.

Figure 1: pH-dependency of Co, Ni and Ca extracted in a sulfate solution using CYANEX 272

Figure 1 shows that depending on the solution pH, different amount of metals can
be extracted. 9

2.2.3 Extractant and diluent

2.2.3.1 Extractant
Due to the high degree of separation required for cobalt and nickel, choice of a
suitable extractant is vital. The properties of a suitable extractant are listed below10:

➢ Selective
➢ Easy to strip
➢ Easily handled
➢ Non-toxic and non-hazardous
➢ Easy to separate from water i.e. different density from water
➢ Low aqueous solubility
➢ High surface tension
➢ Has a high flash-point

10 (48)
 2021-04-07

➢ Has a low vapor pressure

➢ Does not bind too strongly to metal ion
➢ Cheap

It may be difficult for the extractant to achieve all the given properties, but it should
be carefully chosen on these basis.3, 8Therefore, one must take into consideration
what type of chemical environment is present.

If a cation exchange chemistry exists, alkyl phosphoric, phosphonic or phosphinic

acids are the preferred extractant, because of the selectivity for Co over Ni. If a
weakly acidic sulphate solution is present, an alkyl phosphorous acid has shown to
be the preferential extractant option for the separation of cobalt and nickel. The
commercially available dialkyl phosphinic acid CYANEX 272 (C272) has been
proven to give a high separation factor (>1000), which, as mentioned earlier, is
required for cobalt and nickel separation.2

Figure 2: Molecular structure of Di (2,4,4-trimethylpentyl) phosphinic acid (CYANEX 272)11

Di (2,4,4-trimethylpentyl) phosphinic acid (Figure 2), more commonly known as

CYANEX 272, is the extractant of choice when extraction of cobalt from nickel in
a sulfate or chloride solution is wanted. The extractant is non-toxic and easy to
handle, which is an advantage. However, one disadvantage is its high cost.9 The
selectivity series of C272 is:

𝐹𝑒 3+ > 𝑍𝑛 > 𝐶𝑢 > 𝐶𝑜 > 𝑀𝑔 > 𝐶𝑎 > 𝑁𝑖

, which indicates the selectivity of Co over Ni2. According to Sole, the selectivity
of cobalt over nickel depends on the organic complex of cobalt being very
hydrophobic and having a simple structure12. The extractant’s concentration is
usually in the range of 10-25% by volume, when diluted with an aliphatic diluent 6.

Soldenhoff et.al showed that cobalt recovery of 97.5% could be achieved when
using CYANEX 272 for solvent extraction in continuous mini plant trials. The
results from the trials also showed a high separation factor of >1000 (Co:Ni ratio
1000:1) and a >99% removal of impurities.13 This indicates that SX using
CYANEX 272 can be a suitable separation option for Co and Ni separation.2 Other
literature indicates a maximum separation factor of C272 at 6700 for Co and Ni,
which means that 6700 times more cobalt will be extracted from an aqueous sulfate

11 (48)
 2021-04-07

solution into the extractant C272. Note that the prior sulfate solution contains the
same amount of both metals.6

2.2.3.2 Diluent
An organic extractant is seldom used in its pure form, but instead dissolved in an
organic diluent. The diluent used possess a low density and viscosity, in order to be
immiscible with aqueous solutions and easily separated by gravity from the aqueous
phase. Typically, aliphatic diluents are mixed with CYANEX 272 due to their low
vapor pressure, a high oxidation resistance, a high flash point temperature, and,
foremost, being totally miscible with C272.6

2.2.4 Impurities
Since solvent extraction does not remove all of the impurities, this must be done
before or after the process, which is mainly done by precipitation. After the
extraction, some unwanted nickel from the aqueous phase can be co-extracted. To
purify the loaded organic, both washing and scrubbing can be conducted. The
difference between the two is that washing is done by physical removal, whereas
scrubbing is the chemical removal of the co-extracted nickel. An example of the
scrubbing of nickel from the organic phase is to add an aqueous solution containing
high concentration of cobalt, where the intention is to “push out” the nickel.
Meanwhile, Ni can be physically removed (washed) by water.6

M.B.Kime et.al concluded in their paper, that cobalt was sensitive to impurities
such as Ni, Zn, Al, Mn, which were often co-extracted in their experiments (when
using IONQUEST 290 as extractant)14. To avoid the problem a pH selective to
extracting cobalt over other metal ions (impurities) should be chosen. Figure 3
demonstrates the single metal extraction versus pH using C272.

Figure 3: Different metal ion extractions (in %) with CYANEX 272 vs. pH 9

12 (48)
 2021-04-07

Note that this figure is valid for single metal extraction under the specific
experimental conditions performed in the mentioned reference and may vary when
other conditions apply9.

2.3 DEFINITIONS
In this section some of the essential definitions will be described more thoroughly.

2.3.1 Distribution coefficient

The distribution coefficient is defined by the following equation:

[𝑀𝑒](𝑜)
𝐷=
[𝑀𝑒](𝑎𝑞)

The distribution coefficient describes how much of the metal component is

distributed between the two phases after the aqueous solution (containing the metal
of interest) is brought into equilibrium with the organic phase. In other words, D
indicates the extractant’s effectivity. Obtaining a high distribution value means that
a high amount of metal is extracted from the aqueous phase to the organic, which
is favored in the extraction stage. The opposite applies for the distribution
coefficient for stripping, D1, which is described by reversing the equation above:

[𝑀𝑒](𝑎𝑞)
𝐷1 =
[𝑀𝑒](𝑜)

A high D1 value is the result of a successful stripping, because then a high amount
of the metal has been stripped accordingly. However, if a low value is obtained
after the stripping stage, this is an indication that the metal has bonded too strongly
to the extractant, disabling the metal ion to be stripped to the aqueous phase. To be
clear, [Me](o) and [Me](aq) is the concentration of metal in the organic phase and
aqueous phase, respectively. D is dependent on different factors such as:
temperature, composition of the respective phases, and nature of solvent etc.8

2.3.2 Separation factor

The separation factor α is the ratio of distribution coefficients for the two metal
components that are to be extracted, described by:
𝐷𝑖
𝛼=
𝐷𝑗
Where, i represents one metal component and j the other metal component. The
separation factor is an indication of the ease or difficulty of a separation. The greater
the difference of the components’ distribution coefficients, the better the separation.
Thus, a separation is possible if α is not equal to one. The reason that solvent
extraction is a suitable separation process for separating and purifying certain

13 (48)
 2021-04-07

metals, is because of their great difference in distribution coefficients, which in turn

means a great separation factor.8

2.3.3 McCabe-Thiele diagram

A design of a solvent extraction operation can be made by the help of a McCabe-
Thiele diagram. This diagram consists of an equilibrium curve and an operating line
(working line), as shown in Figure 4. By equilibrating a leach solution (aqueous
phase) with an organic phase in different aqueous to organic phase ratios, an
equilibrium curve for the specific operation is determined. Thereafter, the
equilibrium curve can be drawn into the diagram from analyzing the metal content
in the two phases after each new aqueous to organic phase ratio. The equilibrium
curve is constructed by plotting metal concentrations in the organic phase against
metal concentrations in the aqueous phase. The metal concentration in the aqueous
feed is drawn as a vertical line, known as the feed line. An operating line is inserted
into the diagram by drawing a straight line of slope Aq/O, which passes through the
coordinates of the fresh aqueous stream to the coordinates of the raffinate stream
leaving. Note that a phase ratio (Aq/O) is assumed to fit the operation
requirements.3, 5, 7

In order to investigate the amount of extraction stages required for the given
process, the constructed McCabe Thiele diagram is used to calculate number of
required stages. The first extraction stage is obtained by drawing a horizontal line
from the intersection point between the working- and feed line to the equilibrium
curve and then vertically back to the working line. After one stage the metal
concentration in each phase can be obtained, where metal concentration in the
loaded organic phase is found on the y-axis, while metal concentration in the
aqueous phase (barren) is found on the x-axis. This can be given after each stage.
Similarly, the second extraction stage is acquired by continuing horizontally to the
equilibrium curve and vertically back again to the working line. This is repeated
until the desired minimum metal concentration level in the raffinate is achieved.
Worth mentioning is that the slope of the working line can be varied by varying the
Aq/O ratio, which affects the amount of extraction stages required for the separation
specifications. Note that the same procedure is true for calculating the theoretical
stages for stripping, but by reversing the axes.3,5,7

14 (48)
 2021-04-07

Figure 4: Schematic of a McCabe-Thiele diagram5

15 (48)
 2021-04-07

3 METHODS AND MATERIALS

3.1 MATERIALS
The aqueous feed used in all extraction experiments was a pregnant leach solution
(PLS) containing cobalt and nickel, where Fe, Cu, Zn were removed beforehand by
precipitation. Aluminum was partially removed from the leach solution. The leach
solution was obtained after bioleaching and precipitation of the metals mentioned
had been performed at Boliden’s pilot facility. An analysis of the different feed is
presented in Table 1 below. The strip feed for the acid tests are given in Appendix
(section 8.1.3).

Table 1: The feed composition in each experiment, determined by ICP-OES or XRF (*XRF-analyses of Mg,
Mn, Al, Ca inaccurate)

Test Co Ni Mg Mn Al Ca
(mg/L) (mg/L) (mg/L) (mg/L) (mg/L) (mg/L)
Effect of pH 1792 1002 5094 101 910 143
(pH 4)
Effect of pH 1821 1023 6462 44.8 986 103
(pH 4.5)
Effect of pH 1385 800 14933 38.0 794 96.8
(pH 4 low
A:O)
Batch 1803 990 4176 178 772 156
simulation (pH
4.5)
Batch 1716 989 2197 193 369 130
simulation (pH
4.8)
Scrub (pH- 35251 23.0 14.1 5.55 7.05 68.0
tests)
Scrub batch 30065 91.0 340 103 1738 154
(pH 3.7)*
Stripping batch 22161 13.68 4.40 3.62 0.96 11.81

Crystallization 23309 14.83 0 0 537 111

For the organic feed, CYANEX 272, provided from Solvay, was used as extractant.
Tributyl phosphate (Merck) and 1-Decanol (Aldrich) were the modifiers used.
Naphtha and Kerosene (low odor), provided from Starta and Alfa Aeser,
respectively, were used as diluents for the organic feed.

16 (48)
 2021-04-07

3.2 EXPERIMENTAL SET-UP

3.2.1 Process overview
An illustration of the solvent extraction experimental design for this project is
presented in the flowsheet below, Figure 5. The design consists of an extraction
part, where cobalt is extracted by the help of an extractant, followed by scrubbing,
and lastly stripping is carried out to strip cobalt from the organic phase back to the
aqueous phase. Scrubbing is added to the experimental set-up to “push out”
impurities, especially nickel, to obtain pure cobalt sulphate as final product. The
aqueous phases were analyzed using ICP-OES or XRF and mass balance equation
were used to calculate the concentrations in the respective organic phases. Process
efficiency was calculated through percentage of Co and Ni extracted and the
separation factor between these metal ions (α=DCo/DNi). In order to indicate if
stripping was successful, respective distribution coefficients D1 were calculated,
see 2.3.1 for more detail. Cobalt to nickel ratios were also measured to determined
process effectivity.

Figure 5: A flowsheet of the overall experimental solvent extraction design

A more thorough description of the different streams will be mentioned in

conjunction to the results, section 4.1.6. Every choice was clarified after each test
was made i.e. organic composition test, pH-tests, equilibrium tests, and acid-testing.
Temperature and contact time were held constant throughout the experiments, at 22
°C (room temperature) and 5 min, respectively. A contact time of 5 min was chosen
based on experimental grounds, because it was regarded as a sufficient time to reach
equilibrium. The pH-value was adjusted continually using a dilute ammonia
solution after each shake.

17 (48)
 2021-04-07

3.2.2 Effect of saponification and organic phase composition

These tests consisted of varying different modifiers and diluents, in order to observe
the effect on phase separation. This was done through visual observations of phase
separation and separation time. The effect of saponification was also investigated
and compared with the classical solvent extraction method.

In preliminary tests, saponification was tested, which is defined by Anne Marie

Helmenstine as “a process by which triglycerides are reacted with sodium or
potassium hydroxide (lye) to produce glycerol and a fatty acid salt called "soap"15.
In other words, a proton in the organic extractant is exchanged (saponified) against
a sodium ion, by mixing with sodium hydroxide. The reason for the saponification
test was to investigate if this process could be implemented before solvent
extraction with the intention of avoiding a control of pH during the extraction stage.
These four tests (Test 1-4) are summarized in Table 2.

Table 2: Conducted batch tests, with varying compositions, in order to examine the effect on phase
separation

Tests Extractant Modifier Diluent Dilution Saponified

(15%) (5%) factor
1 C272 TBP Kerosene 2 x
2 C272 Isodecanol Kerosene 2 x
3 C272 TBP Naphtha 2 x
4 C272 Isodecanol Naphtha 2 x
5 C272 TBP Naphtha 2 -
6 C272 Isodecanol Naphtha 2 -
7 C272 - Naphtha 2 -
8 C272 TBP Naphtha 4 -
9 C272 Isodecanol Naphtha 4 -
10 C272 - Naphtha 4 -

50 ml of organic phase was prepared based on the composition given in Table 2.

The organic feed was then mixed with 50 ml of 32 %w/v NaOH solution (VWR
Chemicals) in a 250 ml separation funnel at room temperature for 2 min. After
phase separation, the aqueous phase was tapped off. A volume of 50 ml pregnant
leach solution was added to the funnel and mixed for an additional 2 min at room
temperature. After equilibrium was reached, pH of the aqueous phase was measured
with a pH-meter (pH 3210 WTW) and adjusted to pH 4.5±0.1 by adding diluted
NH3 (14%), which was provided from VWR Chemicals.

The experimental tests, Test 5-7, were conducted without saponification, using
naphtha as diluent (Table 2). Only the modifier was varied in these tests to
investigate its effect on separation of phases. The same method was repeated for
these tests, but by removing the addition of NaOH, to avoid saponification. Three
additional tests (Test 8-10) were carried out, where the aqueous feed’s dilution
factor was doubled, to examine its effect on phase separation.

18 (48)
 2021-04-07

3.2.3 Effect of pH on extraction equilibrium

The equilibrium curves were constructed for Co extraction at pH 4.0 and 4.5. A
McCabe-Thiele diagram was constructed for cobalt extraction at pH 4.5.

Table 3: The different aqueous to organic ratios tested at pH 4.0 and pH 4.5

Test A:O Aqueous feed (ml) Organic feed (ml)

11 10 50 5
12 5 50 10
13 2 50 25
14 1 50 50
15 0.5 50 100
16 0.2 50 250

Table 4: Tests at pH 4.0 with lower aqueous to organic ratios

Test A:O Aqueous feed(ml) Organic feed (ml)

17 1 25 25
18 0.5 25 50
19 0.2 10 50
20 0.1 10 100
21 0.05 10 200

The organic feed used for these tests was 15% CYANEX 272 + 5% TBP + naphtha,
which was mixed and shook with the aqueous feed in a separation funnel for 5 min
at room temperature. Tests 11-16 (Table 3) and Tests 17-21 (Table 4) were
conducted at pH 4.0±0.1, with varying A:O ratios. The pH-value was adjusted by
adding ammonia. The same procedure and aqueous to organic ratios as Tests 11-16
were conducted at pH 4.5±0.1, under the same conditions. Cobalt concentrations in
the aqueous solutions were determined by XRF analysis, and cobalt concentrations
in the loaded organic phase were calculated by mass balance, see Appendix (section
8.1.1.)

3.2.4 Batch Simulation of Continuous Multistage Countercurrent

Solvent Extraction
Batch tests were carried out in laboratory scale to simulate the continuous
multistage countercurrent process. Figure 6 illustrates how the batch simulation
tests were performed to simulate the 3-stage countercurrent solvent extraction
process. The method consisted of continuously introducing the aqueous feed
solution with the organic feed solution into a series of batch extractions, with an
aqueous to organic feed ratio of 1:1. After the two feeds were transferred into a 500
ml separation funnel and the pH was set to 4.5±0.1, a contact time of 5 min was
used by shaking the funnel. The aqueous phase (raffinate) and organic phase
(loaded organic) were removed from the funnel by tapping them off separately. This
procedure was done for 8 cycles (A-H). The circles in Figure 6 represent a batch

19 (48)
 2021-04-07

extraction step in a separation funnel. The aqueous and organic phases are denoted
by red and black lines, respectively, in both figures below. The aqueous raffinate
and organic extractant out of the system resemble the actual streams out of a
continuous countercurrent extraction process (Figure 7). Two batch simulation of
countercurrent solvent extraction experiments were conducted, one at pH 4.5±0.1
and the other at pH 4.8±0.05. The pH-value was increased, in order to examine if a
higher cobalt recovery and a less co-extraction of impurities was possible. Note that
in the second batch simulation experiment an aqueous feed with dilution factor of
5 was used instead of 4.

Figure 6: Batch simulation of 3-stage countercurrent solvent extraction. The red and black lines represent
the aqueous and organic phase, respectively. Note that SX A3 was not conducted but is presented for the sake
of visualization.

Figure 7: Continuous countercurrent solvent extraction, where the aqueous raffinate and organic extract are
denoted with an R and E, respectively. The symbol F denotes the fresh feed in and S denotes fresh extractant.

20 (48)
 2021-04-07

3.2.5 Scrubbing tests

The loaded organic phases from the batch simulation at pH 4.8 were blended and
homogenized in a glass flask. A 2 L scrub feed was prepared by adding cobalt
sulphate heptahydrate (CoSO4·7H2O), provided from VWR Chemicals, with a
concentration of 30 g/L, into water, which were mixed by being placed on a
magnetic stir plate. Three different scrubbing tests were carried out for 5 min at
room temperature in a separation funnel containing 50 ml of respective phase. The
pH-value was varied from 3.5, 3.7 and 4.0 between each test. Thereafter, the
remaining loaded organic solution was scrubbed by the help of the scrub feed
containing CoSO4·7H2O, where a pH-value of 3.7 and a contact time of 5 min were
held constant. Even here, the scrubbing process was conducted by shaking using a
separation funnel.

3.2.6 Stripping tests

The stripping experiment was divided into two parts, an acid testing part, and a
batch simulation counter-current testing. The method for the first part was to
prepare a 2 L strip solution as base, by blending CoSO4·7H2O (concentration of 20
g/L) diluted with water in a beaker placed on a magnetic stir plate. The pH was
brought to 2.5 at room temperature. Three different 30 ml acid tests were prepared
by adding different H2SO4 concentrations to the base solution, which had
concentrations of: 12, 24, and 48 g/l. A volume of 5 ml was saved for analysis from
each sample. A volume of 25 ml of the blended and homogenized loaded organic
from scrubbing was mixed with 25 ml of the respective acid solution for 5 minutes
in a separation funnel. After equilibration, the aqueous phase was tapped off and
taken for further analysis.

For the second part of the experiment, a batch simulation multistage countercurrent
test was conducted, in order to simulate the continuous stripping process. A three-
stage system was assumed to be required. The same method for the batch simulation
test was used in stripping as earlier for the extraction process, see Figure 6. The
only difference was that the experiment was run for 7 cycles (A-G) without
controlling the pH-value. The fresh strip feed contained the prepared base solution
mixed with a 24 g/L H2SO4 concentration. The organic phase in the experiment was
the scrubbed loaded organic taken directly out from the previous scrubbing process.
For a more thorough description of the method see section 3.2.4.

3.2.7 Crystallization of Cobalt

To investigate the purity of a possible cobalt sulphate product from the stripped
solution, batch simulation tests of crystallization were made. Crystallization was
carried out by adding different amount of acetone to reduce the solubility of cobalt,
and thus precipitating it as sulphate. The purity is defined by the given equation:
𝑚𝐶𝑜
𝐶𝑜 𝑃𝑢𝑟𝑖𝑡𝑦(%) = × 100%
𝑚𝑐𝑟𝑦𝑠𝑡𝑎𝑙𝑠(𝑡𝑜𝑡)
Where the mass is determined from ICP-analyses of the aqueous phase.

21 (48)
 2021-04-07

The feed used in the crystallization test was prepared by mixing the remaining of
the aqueous phase from batch simulation of the stripping stage. Five different tests
were performed (Table 5). The solutions were stirred using a magnet stirrer for 80
min, after addition of the antisolvent, to facilitate the formation of crystals. Solid
crystals formed were separated from the liquid decantation, before drying in an
oven at 50 ºC until completely dried. The decanted solution samples were analyzed
by ICP-OES.

Table 5: Experimental solute and solvent volumes used for crystallization tests, where the solvent’s volume is
the dependent variable

Aq. solution (ml) Anti-solvent (ml)

20 5
20 10
20 15
20 20
20 25

After drying, the crystals formed for each case (10-25 ml) were weighed, see Table
19 (section 8.1.4) for more detail. Lastly, crystal samples were solubilized in 1%
nitric acid solution for analyses. As a reference, a cobalt sulphate heptahydrate
(CoSO4·7H2O) sample purchased form WVR was also prepared using similar
method for analyses. Mass balance calculations were utilized for obtaining the
crystallized mass, which was determined from ICP-OES analyses of the aqueous
phase.

3.2.8 Analytical methods

Throughout the experimental work, both XRF and ICP instrument were used as
analytical means for the sake of comparison. However, the analytical result used
depended on the given case. For instance, XRF was the analytical method of choice
when a fast analysis was desired (mainly for Co and Ni) and for an experiment
where focus was to examine a specific trend. This applied for the experiment were
the effect of pH was studied. Whereas, when an analysis of the impurities (with
smaller concentration) was required, ICP-OES was the optimal decision. These
analysis results were used for all experiments, except for section 4.1.1.

XRF
Before analysis of each sample, 5 ml of the feed or raffinate was filtered using a
syringe filter and acidified with 100 µl of nitric acid, in order to decrease the pH
and stabilize the dissolved species. The samples from stripping were already acidic
and therefore they were acidified. Thereafter, the sample was inserted into the XRF
instrument SPECTRO XEPOS 5 (60 kV, 2mA) for analysis.

ICP-OES

22 (48)
 2021-04-07

Before running the samples in the ICP instrument Agilent Technologies 5110 ICP-
OES, they were prepared by diluting procedure. The samples were prepared
beforehand by filtering each aqueous raffinate and then diluting each aliquot (a 10
potency less than aimed for) with a diluent of 2% HNO3 (nitric acid) together with
tap water, after which they are diluted an additional 10 times. The last dilution of
the samples was done by using a 2% HNO3 + 10 mg/L yttrium + H2O as diluent.
The four standards used had four different dilution factors, namely 5, 10, 20 and
30.

23 (48)
 2021-04-07

4 RESULTS AND DISCUSSIONS

4.1.1 Effect of saponification and organic phase composition
The four different organic feed mixtures that were tested with saponification all
showed poor and slow phase separations, except for Test 3. In Test 1 no phase
separation was achieved and a pH-value of 4.5 was not possible to reach, without
addition of significant amount of sulphuric acid. In Test 2 a phase separation was
observed, but not a clear one. The aqueous phase was hazy with significant amount
of organic phase entrapment (Figure 8). Likewise, Test 4 resulted in a slow and
insufficient separation of the phases.

Figure 8: A photograph representing test 2, where a blue color is seen in both phases

However, results from Test 3 gave a distinct phase separation. These preliminary
tests, generally, verified that saponification was not a suitable process for this
project, because the intention of avoiding pH-control during extraction could not
fulfill. From these tests, it was also observed that when naphtha was used as diluent,
the separation was better, compared to using kerosene. This could be due to
aliphatic nature of naphtha which, according to the producer, is a more suitable
diluent for C272. As a result, it was used in all tests during this project.

In experiment tested without saponification all tests (Test 5-7) showed poor and
slow phase separations. Aqueous phases in the three tests were cloudy, which could
possibly due to gypsum formation. To avoid this problem, the aqueous feed was
diluted four times, instead of twice. As a result of increasing the dilution factor the

24 (48)
 2021-04-07

gypsum formation was avoided, because of a lower calcium concentration. As

expected, the more diluted aqueous feed, gave better phase separation, therefore
this was kept constant in the following experiments.

Tests 8-10 all resulted in acceptable clear phase separations, indicated by clear
aqueous phase, together with a reasonable separation time. The color of the organic
phase determined the suitable organic feed mixture, where theory states that cobalt
has a dark blue color in the organic phase. From visual observation, the dark blue
color in Test 8 illustrated that there was a high concentration of cobalt present in
the organic phase (top phase in Figure 9). Test 9 had a light blue organic phase,
which confirms the absence of cobalt. A “cobalt blue” color was observed in Test
10 as well, but not as dark blue as Test 8. Therefore, Test 8 was chosen over Test
10.

Figure 9: The organic feed tests, test 8-10, from left to right

From these results, C272 + TBP + Naphtha (Test 8) proved to be the organic phase
composition that resulted in a sufficient and immediate phase separation. Therefore,
this composition was used for further experimental work.

25 (48)
 2021-04-07

4.1.2 Effect of pH on extraction equilibrium

The equilibrium curve for cobalt at pH 4.0±0.1 is presented in Figure 10. The
curve’s appearance differs from that of a standard polynomial curve (Figure 4). A
possible explanation for the deviation is the effect of competing metal ions in the
leach solution. Therefore, tests at lower A:O ratios were conducted to examine the
equilibrium phenomena further.

1200,0

1000,0

800,0
[Co org.] (mg/L)

600,0

400,0

200,0

0,0
600 800 1000 1200 1400 1600 1800
[Co aq.] (mg/L)
Figure 10: Plotted equilibrium curve for cobalt at pH 4.0.

The equilibrium curve presented in Figure 11 gives a better representation of a

standard curve and shows that equilibrium was reached. However, the two red
points in the figure, with A/O of 1 and 0.5, demonstrate that Co has been pushed
out by other metals. The competing metal ions in this case are Al and Mg because
there is a high concentration of them in the aqueous feed, which is verified in Table
1. This result was suspected due to that Al2+ is favored by C272 at pH 4.0, see
Figure 3. Cobalt recoveries into the organic phase at lower A/O ratios were greater
compared to earlier samples with a higher A/O ratio. This is an expected result. The
explanation is that with a lower A/O ratio there is more organic present, leading to
that there is more extractant for the metal ions to extract to.

26 (48)
 2021-04-07

160
140
120
100
[Co org.] (ppm)

80
60
40
20
0
0 200 400 600 800 1000 1200 1400
-20
-40
[Co aq.] (ppm)
Figure 11: Equilibrium curve for cobalt at pH 4.0 with a lower aqueous to organic ratio. Notice that the red
points represent Co ions being pushed out

Both equilibrium curves presented above make it difficult or even impossible to

construct a McCabe-Thiele diagram. Whereas, experimental data from pH 4.5 gave
a polynomial curve, similar to a standard equilibrium curve. With this equilibrium
curve it was possible to insert into a McCabe-Thiele diagram and calculate the
required extraction stages, which was one of the purposes with these tests. Figure
12 presented below, shows the constructed McCabe-Thiele diagram with an
operating line slope of 1 at pH 4.5. Note that point A/O=2 was an outlier to the
curve and was therefore removed from the plot.

3000
y = -0,001081x2 + 3,196298x
R² = 0,998558
2500

2000
[Co org.] (ppm)

1500

1000

500

0
0 500 1000 1500 2000
[Co aq.] (ppm)
Figure 12: McCabe Thiele diagram for cobalt at pH 4.5 with a slope of 1

27 (48)
 2021-04-07

By reading off Figure 12, it was evident that 3 stages were required for solvent
extraction operation to reach to 0.085 g/L cobalt in the final raffinate. Based on the
grounds discussed above i.e. similarity to standard appearance and possibility for
constructing a McCabe-Thiele diagram, a pH-value of 4.5 was chosen for the SX
operation, i.e. batch simulation experiment.

28 (48)
 2021-04-07

4.1.3 Batch Simulation of Continuous Multistage Countercurrent

Solvent Extraction
Figure 13, demonstrates the concentration of nickel and cobalt in the aqueous phase,
after each cycle. Shown in the figure, is that the nickel concentration has reached
equilibrium at around 1.20 g/L, which is represented by the blue points.
Furthermore, nickel’s concentration in the aqueous phase was higher compared to
cobalt’s, resulting in a successful separation between the two metals. The
concentration of cobalt in the aqueous phase reached equilibrium after three cycles,
see the black dots in the figure. This means that simulation tests can be run for four
cycles.

1400

1200

1000
[Me](aq.) (mg/L)

800

600

400
Co
200
Ni
0
0 1 2 3 4 5 6 7 8
Cycle
Figure 13: The metal ions’ concentration in the aqueous phase vs. batch simulation cycles, for cobalt and
nickel

From examining Figure 14 and Table 6, it is possible to notice that Co was

recovered at an average of 69 % into the organic phase. Whereas, Ni remained in
the aqueous phase and was not co-extracted into the organic phase. The blue points
in Figure 14 , verify that aluminum has a 100 % recovery and is competing against
cobalt for a spot in the organic phase. Both Al and Mn are the main impurities,
which were co-extracted into the organic phase (Table 7).

29 (48)
 2021-04-07

120

100

80
Me in org. (%)

60

40
Co recovery
20
Al recovery

0
0 1 2 3 4 5 6 7 8
Cycle

Figure 14: Metal ion recoveries into organic phase vs. cycles at pH 4.5

Table 6: Different metal ion recoveries into the organic phase after each cycle, at pH 4.5

Cycles Co Al Mg Ca Mn
(%) (%) (%) (%) (%)
B3 88 100 39 32 99
C3 76 100 36 15 99
D3 54 100 34 6 97
E3 64 100 35 10 98
F3 63 100 34 9 98
G3 73 100 37 18 99
H3 66 100 37 15 98
Average
(%) 69 100 36 15 98

When comparing organic recoveries for each metal ion from the table above with
literature (Figure 3) it is worth noticing similarities. For instance, recovery of Co
into the organic phase at pH 4.5 is around 70 % in Figure 3, which complies with
batch simulation tests at an average of 69%. The same is true for the other metal
ion recoveries when compared to literature. Hence, the experimental data at pH 4.5
agrees to a great extent with theory. It is worth noting that Al, which was partially
removed from the leach solution beforehand, shows a 100% recovery into the
organic phase.

One source of error in the batch simulation tests was the pH interval of ±0.1, where
a majority of the samples had a pH of 4.4. In other words, the results may differ
significantly between having a pH-value of 4.4 compared to 4.5. The difference in
pH between samples makes the results less reliable, due to inconsistency.

30 (48)
 2021-04-07

Therefore, a smaller interval (0.05) was used in future tests to minimize this source
of error.

Batch simulation results, at pH 4.5, showed a higher cobalt concentration in the

raffinate, at 0.56 g/L, compared to 0.085 g/L, which was read from the McCabe-
Thiele diagram (Figure 12). As a result of the difference in Co purity, it is worth
discussing if a McCabe-Thiele diagram is necessary to use for this case. In other
words, a McCabe-Thiele diagram may not be an optimal representation of the
required stages for a more complex system, where many metal ions are present. In
an ideal case it is assumed that only cobalt and nickel are present in the solution,
whereas in this specific case there were several different metal ions present in the
leach solution. This may lead to other metal ions competing against the desired
metal ions, which affects separation factor, equilibrium curve etc.

A possible factor to the difference between the real and theoretical case, can be
referred to the source of error discussed earlier, concerning the interval gap of ±0.1.
Namely, that if the experimental points to construct the McCabe-Thiele diagram
where mostly conducted at pH 4.5, whereas the batch simulation tests had a pH of
4.4, it may not give an accurate comparison between the two cases. This results in
a less reliable comparison. However, this difference should not be so significantly
large.

Due to that both dilution factor and pH-value was changed in the next set of batch
simulation tests a full comparison between the two batch tests was not possible.
With a higher dilution factor, it is evident that lower concentrations are expected.
Therefore, concentration difference between the two different tests is ignored.
However, a comparison between the recoveries is possible because a percentage
calculation is made
.
Table 7: The metal ions’ recoveries into the organic phase after each cycle, at pH 4.8.

Cycles Co Ni Al Mg Ca Mn
(%) (%) (%) (%) (%) (%)
B3 100 6 100 35 62 100
C3 99 8 100 23 43 100
D3 99 1 100 14 30 100
E3 100 2 100 14 33 100
F3 100 1 100 16 35 100
G3 100 4 100 17 35 100
H3 100 1 100 9 31 100
Average
(%) 100 3 100 18 38 100

31 (48)
 2021-04-07

100
90
80
70
Me in org. (%)

60 Co

50 Ni

40 Mg

30 Ca

20
10
0
0 1 2 3 4 5 6 7 8
Cycle
Figure 15: Metal ion recovery into organic phase vs. cycle, at pH 4.8

As predicted from literature, a higher cobalt recovery in the organic phase was seen
at pH 4.8, with an average of 100 %. However, a small percentage of Ni (3 %) was
co-extracted into the organic phase, which was not seen at pH 4.5. Operating SX at
pH 4.8 resulted in a higher recovery of impurities, which is not desired when pure
cobalt is wanted. One the other hand, less Mg was co-extracted at pH 4.8. Both Al
and Mn had a 100 % recovery into the organic phase. All these statements can be
verified in Table 7 and Figure 15 above. Experimental data at pH 4.8 compiles with
literature, see Figure 3. Figure 15 demonstrated that equilibrium was reached after
2 cycles, leading to that a maximum of 3 cycles is required.

When deciding on a pH-value for the operation, many parameters must be weighed
against each other e.g. Co recovery, co-extraction of impurities (Co purity),
economical aspects etc. For instance, when operating at pH 4.8 a high cobalt
recovery is obtained, but some unwanted nickel is also extracted into the organic
phase, followed by other impurities. In order to avoid this problem, the solvent
extraction process should be operated at pH 4.5 instead. Nevertheless, more stages
are required in this case compared to operating at pH 4.8, which is economically
unfavorable. One aspect that is worth mentioning, is that a scrubbing process is
necessary for the system operation at pH 4.8, in order to remove co-extracted Ni,
which also requires stages. Therefore, the scrubbing process was also examined in
the project. All these aspects must be taken into consideration when choosing a pH-
value for solvent extraction.

32 (48)
 2021-04-07

4.1.4 Scrubbing
The intention of scrubbing was to purify the loaded organic (after batch simulation
countercurrent tests) from impurities, especially nickel, by addition of cobalt as
sulfate. This was done in order to push out Ni for instance. Table 8 shows ICP-
OES-results from the pH-tests, which indicates that at pH 4.0 the highest cobalt
concentration in the loaded organic out was acquired, with a concentration of 7.06
g/L. However, a higher concentration of impurities was also detected here. As
expected, the trend of an increasing pH-value is that the loaded organic
concentration increases, which is verified in the table. This was suspected, due to
earlier observation from extraction, were the metal ions examined were extracted
to a higher extent at higher pH-values, and the same phenomena is true for
scrubbing.

Table 8: Concentrations of loaded organic out after scrubbing at different pH-values

pH-value Co Ni Al Mg Ca Mn
(mg/L) (mg/L) (mg/L) (mg/L) (mg/L) (m/L)

3.5 3668 21.7 373 184 79.7 80.8

3.7 4340 22.8 373 187 80.1 92.8
4.0 7058 23.2 374 190 81.5 112

Results from scrubbing will affect stripping results because the scrubbed loaded
organic continues to the stripping stage. In other words, the more Co is present after
scrubbing the more there is room to strip. However, to obtain a high purity of cobalt,
other impurities must be minimized in the scrubbing process. Both these parameters
must be taken into consideration when deciding on a pH-value.

A pH-value of 3.7 was chosen for further scrubbing, because of marginal difference
in cobalt concentration between the other pH-values. Furthermore, less impurities
were seen at pH 3.7 compared to at pH 4.0, as mentioned earlier. If the intention is
to obtain as high Co concentration as possible pH 4.0 would have been the optimal
decision. Additionally, concentrations of impurities at pH 4.0 were marginally
higher than for pH 3.5 and 3.7. However, in this report the aim is to develop a deeper
understanding of separation parameters for Co-Ni separation, opposed to obtaining
an optimal value.

33 (48)
 2021-04-07

4.1.5 Stripping
The strip results from acid testing, demonstrated that with a sulphuric acid
concentration of 24 g/L the highest cobalt concentration was obtained, which is
verified in Table 9, with a Co concentration of 24.3 g/L. Regarding the other metal
ions, there was a marginal difference in acid concentrations between the tests.
Therefore, 24 g/L H2SO4 was the acid concentration chosen to continue with for the
second part of the stripping experiment.

Table 9: Concentration of aqueous product out for each metal ion with varying acid concentrations

H2SO4 Co Ni Al Mg Ca Mn
conc.(g/l) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm)

12 23740 13.93 349 4.72 12.54 24.53

24 24333 13.95 394 4.79 12.92 24.69
48 23236 13.53 394 4.72 12.84 24.86

A 3-stage stripping process was assumed, when conducting a McCabe-Thiele

diagram was neglected (due to inaccuracy representation of the real case).
Conducting tests for the batch simulation countercurrent process for stripping
resulted in that equilibrium was reached after one cycle. This leads to the knowledge
that a minimum of 2 cycles was required to reach equilibrium, and that the average
concentration of cobalt stripped was 23.2 g/l. Table 10 indicates that a high amount
of Co was stripped compared to the other metal ions. This means that the stripped
aqueous product has a high cobalt purity, with small amounts of impurities.
However, Al concentration was 10 times higher than the other metal ions.

Table 10: Average aqueous concentrations out from batch simulation process for each respective metal ion

Co (ppm) Ni (mg/L) Al (mg/L) Mg (mg/L) Ca (mg/L) Mn (mg/L)

23180 13.45 313 4.59 12.51 20.09

34 (48)
 2021-04-07

4.1.6 Solvent extraction overview

Figure 16: Schematic of experimental design with Co and Ni in-and-out streams presented with unit g/L, and
with A:O=1

Figure 16 shows a flowsheet of the entire solvent extraction process, to demonstrate

in-and-out streams. Only Co and Ni concentrations are presented, because these
were the metals of interest to separate.

By reading off the left-hand-side of the figure, it is noticed that most of the cobalt
has been extracted into the organic phase, leaving nickel in the aqueous phase. The
calculated distribution coefficients for SX for cobalt and nickel are 294 and 0.02,
respectively, resulting in a separation factor of 14250, proving CYANEX 272 to be
an appropriate extractant for Co-Ni separation. When comparing the obtained
separation factor with earlier studies (see section 2.2.3.1) it confirms a significantly
higher value i.e. 14250 vs. 6700. Whereas, the continuous mini plant trials
conducted by Soldenhoff et.al achieved α>1000. It is worth discussing if these
results are comparable to the continuous systems. From one perspective the batch
tests throughout the experiment were made to simulate a continuous countercurrent
process and should therefore be comparable to other existing continuous systems.
However, when scaling up a chemical process many aspects are changed, and this
must also be taken into consideration. The extraction process offered a 99%
recovery of cobalt and 2.0% recovery of nickel. This result displays a similar
recovery to the one mentioned in section 2.2.3.1 , obtained from Soldenhoff et.al’s
mini plant trials, where a 97.5% Co recovery was achieved. The Co:Ni ratio after
extraction was calculated to 85.5.

Co-extracted nickel is successfully removed through scrubbing, where the cobalt

ions manages to “push out” impurities (namely Ni), as intended. Nickel from the
loaded organic in is scrubbed to half its original value, which is seen in the figure,
by observing that 0.01 g/L remains in the scrubbed organic out. This implies, that
the scrubbed organic stream that continues to stripping is nearly free from nickel
and contains practically only cobalt, leading to more room for stripping of pure Co.
Another indication of a successful scrubbing process, is a higher cobalt to nickel
ratio (Co:Ni=321) obtained after scrubbing compared to earlier observed ratio in
extraction.

35 (48)
 2021-04-07

Out of the stripping process, Co was stripped to the final product with a value of 1
g/L, while the Ni stream was unchanged. The distribution coefficients are 10.5 and
1.00 for Co and Ni, respectively. A low distribution coefficient is aimed for in this
stripping process because it is not desired to strip all cobalt due to the risk of
stripping other elements (impurities). Consequently, obtaining a low D value
minimizes this risk. A Co:Ni ratio of 2320 indicates that stripping was successful.

4.1.6.1 Distribution of elements

In order to get an overview of other metal ions (contaminants) in the streams,
distribution results were analyzed. Presented in Table 11, are metal ion distributions
between the four streams in extraction. It is of interest to give focus on the
impurities (other elements), while earlier observation can be justified for. Cobalt,
Aluminum and Manganese were evenly distributed between the streams. They were
also extracted to the organic phase to the highest extent of the elements. This
implies, that employment of scrubbing was necessary. While, Mg and Ca were not
co-extracted to the same extent and remained mainly in the aqueous phase. The
distribution of Ni confirmed earlier observations, i.e. that Ni practically remained
in the aqueous phase enabling Co-Ni separation.

Table 11: Distribution of elements in the different extraction streams

Extraction Co Ni Mg Al Mn Ca
Stream (%) (%) (%) (%) (%) (%)
Aq. feed 33.4 49.5 61.0 33.3 33.3 42.8
Org. feed 0.00 0.00 0.00 0.00 0.00 0.00
Aq. out 33.4 49.5 30.5 33.3 33.3 42.8
Org. out 33.3 0.9 8.6 33.3 33.3 14.4
Total 100 100 100 100 100 100

Distribution of different metal ion scrub streams are summarized in Table 12. As
mentioned earlier, scrubbing of Ni was successful, and this statement is strengthen
through noticing that the distribution of Ni was lower in the outgoing organic stream
compared to the ingoing. The same applies for the other impurities (Mn, Ca),
besides from Al and Mg. Distribution of Mg was even. Aluminum distribution
results show that approximately 40% is distributed to the organic stream out, which
implies that the metal ion was difficult to scrub.

36 (48)
 2021-04-07

Table 12: Distribution of elements in the different scrub streams

Scrub Co Ni Mg Al Mn Ca
Stream (%) (%) (%) (%) (%) (%)
Aq. feed 47.3 41.4 26.2 41.2 17.4 38.9
Org. feed 2.69 8.59 23.8 8.75 32.6 11.1
Aq. out 44.9 45.5 25.4 10.4 19.6 43.6
Org. out 5.07 4.49 24.6 39.6 30.4 6.41
Total 100 100 100 100 100 100

Through reading off Table 13, a remark was the high distribution of Co in the
aqueous product stream out, opposed to other elements. This proves that cobalt has
been stripped accordingly. Whereas, the other metal ions are mostly distributed
between the organic streams. Note that nickel is distributed evenly, while a
considerably high distribution of aluminum was seen in the aqueous stream out.
Table 13: Distribution of elements in different strip streams

Strip Co Ni Mg Al Mn Ca
Stream (%) (%) (%) (%) (%) (%)
Aq. feed 43.6 29.0 0.68 0.03 0.99 15.9
Org. feed 6.35 21.0 49.3 50.0 49.0 34.1
Aq. out 45.7 28.5 0.71 9.37 5.48 16.8
Org. out 4.34 21.5 49.3 40.6 44.5 33.2
Total 100 100 100 100 100 100

37 (48)
 2021-04-07

4.1.7 Overall distribution and process summary

The metal ions’ distribution, seen from an overall process perspective, is
demonstrated in Table 14. For Ni, Mg, Mn, and Ca a high percentage of the total
distribution was seen in the raffinate stream leaving extraction, resulting in less
room for co-extraction. Scrubbing proved to be effective, because a high
distribution of elements was detected in the scrub raffinate. Confirmed was that a
small percentage of impurities were distributed to the product compared to cobalt.
However, distribution results show that the product stream was not free from
contaminants. Al proved to be difficult to remove, because it was recognized to
some extent in every out stream.

Table 14: An overall distribution of elements in streams throughout solvent extraction

Stream Co Ni Mg Al Mn Ca
(%) (%) (%) (%) (%) (%)
Aq. feed 1.57 44.8 51.2 8.05 24.4 20.5
Org. feed 0.00 0.00 0.00 0.00 0.00 0.00
Aq. 1.57 44.8 25.6 8.05 24.4 20.5
raffinate
Scrub feed 27.4 4.13 7.92 37.9 13.0 24.2
Scrub 26.1 4.53 7.69 9.55 14.7 27.1
raffinate
Strip feed 20.2 0.62 0.10 0.02 0.46 1.86
Stripped 2.01 0.46 7.43 29.60 20.62 3.87
org.
Stripped 21.2 0.61 0.11 6.83 2.54 1.97
aq. product

The increase of Co:Ni ratio between each stage in solvent extraction (an increase
of 10 potency each time), verifies that solvent extraction was successful in
separating cobalt from nickel.

38 (48)
 2021-04-07

4.1.8 Crystallization of Cobalt

The purity of produced cobalt sulphate crystals is summarized in Table 15, where
no Ni, Mn and Mg was detected. The trend in Table 15 shows that with an increase
of solvent the Co purity increases, reaching a Co purity of 34.1% with a 25 ml
solvent addition. Aluminum and Calcium were detected to some extent.

From comparison with the reference, the percentage of metals in cobalt sulphate
show similar results, except for aluminum. The result gave considerably higher Al
percentage than the reference, where no Al was present. This indicates, that in order
to obtain a high purity cobalt sulphate product it is important to remove Al
beforehand. Worth discussing, is the acceptable concentration limit of impurities.
For instance, amount of Al in the product can be regarded as too high for utilizing
in laboratory work, but acceptable for production of vehicle batteries. In other
words, impurity limitations depend on the buyer’s specifications.

Worth noting is that a full comparison cannot be made between the produced cobalt
sulphate and the reference. This is because in our case cobalt sulphate monohydrate
was produced, while the reference utilized was heptahydrat. In other words, there
are less water molecules in the product compared to in the reference, resulting in
higher purities.

Table 15: Purity of the cobalt sulphate produced after stripping

Solvent added Co (%) Al (%) Ca (%)

(ml)
10 32.5 0.10 0.20
15 33.2 0.10 0.30
20 33.5 0.10 0.10
25 34.1 0.10 0.10
Ref 22.1 0.00 0.10

A similar pattern to the one above is observed in Table 16, where an increase in
solvent leads to an increase in mass crystallized. With a 25 ml acetone volume
added together with 20 ml stripped product (solute) a value of 0.42 g Co is
precipitated through crystallization. No impurities were present; therefore, they are
not presented in the table. Worth discussing, is the quantity of cobalt sulphate
crystallized as product of the entire SX process. It may not be considered that high,
but by taking the volume quantities used into account it is more reasonable to expect
these values. Additionally, it could be favorable with a lower Co amount
crystallized if it results in a higher purity product, which is the case here when no
impurities are produced.

39 (48)
 2021-04-07

Table 16: Amount of cobalt sulphate salt produced that was crystallized with different volumes of solvent
added

Solvent added Mass of Co

(ml) crystallized
(g)
10 0.17
15 0.30
20 0.30
25 0.42

4.1.9 Source of error

Many potential sources of errors are associated with human error. Some examples
of these are material loss (spill), washing, bad dilution or bad tapping off technique.
However, the errors due to the human factor can be considered constant because
the experiments were conducted by the same laborant throughout this project.
Contamination and instrumental error are other potential source of errors.

40 (48)
 2021-04-07

5 CONCLUSION
The questions which were intended to be answered after all experimental tests are
presented below:
• Using naphtha opposed to kerosene as diluent resulted in distinct and
immediate separation of phases. Both modifiers (and without) gave distinct
phase separation, but a higher amount of Co was observed in the TBP
composition.
• Equilibrium curves for cobalt extraction at pH 4, indicated that equilibrium
was either not reached or affected by the presence of competing metal ions
(Mg and Al). A standard equilibrium curve was seen from Co extraction
equilibrium curve at pH 4.5.
• A minimum of 3 stages was required for extraction at pH 4.5, with an
aqueous to organic ratio of 1.
• Distribution of Al and Mn showed that they were co-extracted into the
organic phase to a high extent compared to other elements. Out of the scrub
streams most of impurities (expect Al) were distributed in the aqueous
raffinate, resulting in a successful scrubbing process. Compared to Co, a
small distribution percentage of other metal ions were detected in the
product stream, with the exception of Al.
• The Cobalt sulphate product’s purity is regarded as high, in respect to a
reference of pure CoSO4·7H2O. However, substantial amount of Al was
detected. The acceptable amount of impurities in the solution, i.e. product
purity, should be determined through the buyer’s specification.

Worth noting is that no conclusion can be drawn regarding the recommended pH-
value for solvent extraction. The reason is that pH-value is related to process
specifications. Consequently, many different aspects must be considered and
weighed against each other before deciding on a preferable pH-value. For instance,
cobalt recovery, co-extraction of impurities, and economical factors are regarded.
The same applies for separation parameters e.g. sulphuric acid concentration in
stripping and pH-value in scrubbing.

One general conclusion was that a McCabe-Thiele diagram did not give an accurate
representation of the batch simulation case, where a more complex system was in
place. Therefore, it should not be a method as means for predicting required stages
for SX process in this project.

Solvent extraction proved to be an effective separation process for the separation of

cobalt and nickel. Furthermore, scrubbing showed to be an adequate method for
removal of impurities and is a choice to consider when a high amount of impurities
is co-extracted in the extraction circuit. Due to that Al proved to be difficult to

41 (48)
 2021-04-07

remove, and is selective over Co, it must be totally precipitated before solvent
extraction.

In conclusion, the aqueous feed’s (PLS) composition determines the operation

parameters and phase separation. In other words, every case is unique and must be
analyzed beforehand. The process specifications should also be accounted for in
that respect. Worth noting, is that these given recommendations are based on these
specific grounds and may vary on slightest change.

42 (48)
 2021-04-07

5.1 FUTURE RESEARCH

In this report a focus was given on examining if a high purity cobalt sulphate
product could be obtained. For future research it would be interesting to investigate
nickel production methods more, for instance electrowinning. Another possible
future study could be to examine this case but applied in bigger scale and using
actual continuous reactors. The scaling-up project could be conducted at Boliden’s
pilot facility by constructing a pilot scale multistage counter-current solvent
extraction. Work in this area has already started, by development engineer
Mohammad Khoshkhoo. An on-going project is to produce Ni/Co sulfide from the
same leach solution used in this project, and then conduct bioleaching again.
Therefore, it would be of interest to examine if the newly produced leach solution
could be an alternative (to the one used in this project) for SX. Investigating the
effect that kinetics and thermodynamics has on the Co-Ni separation was outside of
the project scope. Even though the chemical process is very pH-dependent it would
be of interest to examine these parameters in the future. Finding a better theoretical
method for predicting design parameters for solvent extraction could also be studied
in the future.

43 (48)
 2021-04-07

6 ACKNOWLEDGMENTS
A great gratitude is due to the company Boliden for providing me with this
exciting master’s thesis project and for the opportunity to work alongside some
great colleges within the team. I want to give an extra thanks to my supervisor
Mohammad Khoshkhoo for all the help and support and for the interesting
discussions. Lastly a thanks is given to Kerstin Forsberg, my examiner at KTH,
for recommending this company to me and for her encouragement throughout the
project. Lastly, thank you to all the ones concerned.

44 (48)
 2021-04-07

7 BIBLIOGRAPHY

1. Boliden – Metals for modern life. Accessed January 15, 2021.

https://www.boliden.com/

2. Lett DS. Cobalt-Nickel Separation in Hydrometallurgy: a Review. :11.

3. Sandström Å. HYDROMETALLURGY. Published online January 2016.

4. SX Kinetics, Inc. - Specialists in Solvent Extraction and Electrowinning Pilot

Plants. Accessed February 1, 2021. http://www.sxkinetics.com/sxprocess.htm

5. Habashi F. A textbook of hydrometallurgy. Métallurgie Extractive; 1999.

6. Crundwell FK, ed. Extractive Metallurgy of Nickel, Cobalt and Platinum-

Group Materials. Elsevier; 2011.

7. Coulson JM, Richardson JF. Chemical Engineering. Vol. 2: Particle

Technology and Separation Processes. 4. ed., reprinted (eith revisions).
Pergamon Press; 1993.

8. Venkatachalam S. Hydrometallurgy. Narosa; 1998.

9. CYANEX® 272. Solvay. Accessed September 29, 2020.

https://www.solvay.com/en/product/cyanex-272

10. Kordosky G, Virnig M, Mackenzie M. SOLVENT EXTRACTION -

REAGENTS AND SELECTIVITY CONTROL. Henkel Corporation MID; :20.

11. Souza M, Mansur M. COMPETING SOLVENT EXTRACTION OF

CALCIUM AND/OR NICKEL WITH CYANEX 272 AND/OR D2EHPA.
Braz J Chem Eng. 2019;36:541-547. doi:10.1590/0104-
6632.20190361s20170527

12. Sole KC, Cole P. Ion Exchange and Solvent Extraction. 15th ed.; 2001.

13. Soldenhoff K, Hayward N, Wilkins D. The Minerals, Metals & Materials

Society. Mishra. EPD Congress; 1998.

14. Kime MB, Kanowa EK. Valorization of low-grade copper-cobalt ore from the
Mukondo mine by heap leaching and solvent extraction. CIM J. 2017;8(4).
doi:10.15834/cimj.2017.25

15. Ph. D. BS, B. A. P and M, Facebook F, et al. Know the Definition of

Saponification. ThoughtCo. Accessed November 12, 2020.
https://www.thoughtco.com/definition-of-saponification-605959

45 (48)
2021-04-07

46 (48)
 2021-04-07

8 APPENDIX
8.1.1 Effect of pH on extraction equilibrium
When the equilibrium curves were constructed, loaded organic concentrations
were calculated by the help of the given mass balance:

𝑉𝑎𝑞. 𝑉𝑜𝑢𝑡
𝐶𝑙𝑜𝑎𝑑𝑒𝑑 𝑜𝑟𝑔. = ∗ 𝐶𝑓𝑒𝑒𝑑 − ∗ 𝐶𝑎𝑞.𝑟𝑎𝑓𝑓𝑖𝑛𝑎𝑡𝑒
𝑉𝑜𝑟𝑔. 𝑉𝑜𝑟𝑔.
, where Vout=Vaq+Vadded.
See Table 17, which demonstrated that there is more Al and Mg present in the
organic phase compared to Co. In that respect, Al and Mg are seen as competing
metal ions to Co.
Table 17: Metal recoveries into the organic phase at pH 4.0 and low A:O ratios

A:O Co (%) Al (%) Mg (%)

1 3 82 73
0.5 20 88 74
0.2 59 84 83
0.1 73 81 69
0.05 85 83 76

8.1.2 Batch simulation of continuous multistage countercurrent

solvent extraction
The calculated recoveries into the organic phase are calculated through the
equation presented below.
𝐶𝑓𝑒𝑒𝑑 − 𝐶𝑎𝑞.𝑟𝑎𝑓𝑓𝑖𝑛𝑎𝑡𝑒
%𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑦 = ∗ 100
𝐶𝑓𝑒𝑒𝑑
This equation is mostly used in the batch simulation experiment but may also be
used for other experiments throughout the report.

8.1.3 Stripping
An analysis of the strip feed for the acid tests are presented in Table 18.

Table 18: The feed analysis of stripping tests (acid testing)

Acid Co Ni Mg Mn Al Ca
conc.(g/L) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm)
12 22189 14.1 4.58 3.64 0.02 12.38
24 22160 14.4 4.75 3.78 0.06 12.67
48 22890 14.5 4.66 3.72 0.00 14.31

47 (48)
 2021-04-07

8.1.4 Crystallization of cobalt

The experimental data on the mass of crystals obtained after drying can be read
from Table 19.

Table 19: The weighed crystals after drying for different solvent volumes added

Solvent volume (ml) Mass of crystals (g)

10 0.60
15 0.96
20 0.94
25 1.11
Ref 1.01

The formed crystals are illustrated in Figure 17 below.

Figure 17: A photography of all the formed crystals from crystallization

48 (48)