Impact of Water on Recycling Lithium Ion Battery Cathode Material in a Deep Eutectic Solvent

(1)

Impact of Water on Recycling Lithium Ion Battery Cathode Material in a Deep Eutectic Solvent

Erik ¨ Ostlund

Degree project E in Chemistry, 1KB052 Supervisor: Reza Younesi

July 2020

(2)

Abstract

In this project, the effect of water on a deep eutectic solvent (DES) when reclaiming cobalt has been investigated. Cobalt was digested from lithium cobalt oxide (LCO) in a DES composed of ethylene glycol and choline chloride. Different amounts of water were added to the DES so that the impact on the DES and the result from the digestion could be followed. Water make the DES significantly easier to work with, while the complexes obtained after digestion remained the same.

Keywords: Recycling, Cobalt, Deep Eutectic Solvent, Ethylene Glycol, Choline Chloride, Lithium-ion Batteries, LIBs

(3)

Popular Scientific Summary

The usage of batteries is presently increasing fast world wide. This is mainly due to the large number of electric vehicles produced, which is predicted to grow exponentially in the coming years. This also poses a problem since there can be a shortage of the materials the batteries are made of. For this to not happen, the materials can be recycled. This is being done today, but not in an environmentally friendly way. That is because the two most common recycling methods utilizes a high temperature (∼ 1400 °C) and dangerous chemicals respectively.

Recently, a study showed promising results recycling in an environmentally friendly solvent at moderate temperatures (100 - 200 °C). In this project, that solvent has been used. To be able to improve upon that technique, the effects of water in the solvent has been studied. The cobalt rich material was digested in 160 °C for 72 hours. Afterwards, x-rays, visible light, infrared light, incineration in plasma, electrons and more were utilized to analyze the samples.

It was concluded that some water could be useful to improve the properties and stabilize the solvent. For example, the solvent is thick to begin with, but with addition of water it becomes thinner and flows better. This is good since it is then easier to transfer between different vessels and to work with in general. Too much water on the other hand causes the solvent to boil during digestion, which could lead to undesirable events. The most optimal water content was determined to be between 5 % and 10 %. The cobalt obtained after the digestion was found in two compounds. This did not change with water content. What did change was the ratio between them. If the solvent would be water free, only one compound would be obtained. This could be helpful in the next step when cobalt is supposed to be retrieved from the digested solution. It is, however, hard to get the solvent free from water since it pulls water from the air. To add some water is thus a promising technique, but it still needs more refinement.

(4)

Table of Content

Abstract i

Popular Scientific Summary ii

Abbreviations vi

1 Introduction 1

1.1 Cobalt . . . 1

1.1.1 What is Cobalt? . . . 1

1.1.2 Production Chain . . . 1

1.1.3 Future Outlook of Cobalt . . . 4

1.2 Recycling Today . . . 4

1.2.1 Pyrometallurgy . . . 6

1.2.2 Hydrometallurgy . . . 6

1.2.3 Other Methods . . . 7

1.3 Deep Eutectic Solvents . . . 7

1.3.1 Eutectic point . . . 8

1.3.2 Choline Chloride . . . 9

1.3.3 Ethylene Glycol . . . 10

1.3.4 Water . . . 10

1.3.5 Properties and Applications . . . 11

2 Aim of the Project 13 3 Experimental 14 3.1 Chemicals and Materials . . . 14

3.2 Sample Preparation . . . 14

3.2.1 Deep Eutectic Solvent . . . 15

3.2.2 Digestion . . . 15

3.2.3 Precipitation . . . 15

3.2.4 Analysis Preparation . . . 16

3.3 Instrumentation . . . 17

3.3.1 Centrifuge . . . 17

3.3.2 Filtration . . . 17

3.3.3 Fourier Transform Infrared Spectroscopy . . . 17

3.3.4 Heating . . . 18

(5)

3.3.5 Inductively Coupled Plasma - Optical Emission

Spectrometry . . . 18

3.3.6 Karl Fischer . . . 18

3.3.7 Thermal Gravimetric Analysis . . . 18

3.3.8 UV/Visible Spectroscopy . . . 18

3.3.9 Scanning Electron Microscopy . . . 19

3.3.10 X-Ray Diffraction . . . 19

4 Results 20 4.1 Procedure . . . 20

4.1.1 Preparation . . . 20

4.1.2 Digestion . . . 20

4.1.3 Precipitation . . . 22

4.2 Properties . . . 22

4.2.1 Density and Viscosity . . . 22

4.3 Concentration . . . 25

4.3.2 UV Visible Spectroscopy . . . 26

4.3.3 Inductively Coupled Plasma - Optical Emission Spectrometry . . . 30

4.4 Characterisation . . . 30

4.4.3 Energy Dispersive X-ray Spectroscopy . . . 34

5 Discussion 38 5.1 Procedure . . . 38

5.1.1 Sample preparation . . . 38

5.1.2 Digestion . . . 38

5.1.3 Procedure Conclusion . . . 39

5.2 Properties . . . 39

5.2.1 Density and viscosity . . . 39

5.2.4 Properties Conclusion . . . 41

(6)

5.3 Concentration . . . 41

5.3.2 UV Visible Spectroscopy . . . 42

5.3.3 Inductively Coupled Plasma - Optical Emission Spectrometry . . . 44

5.3.4 Concentration Conclusion . . . 45

5.4 Characterization . . . 46

5.4.2 Energy Dispersive X-Ray Spectroscopy . . . 47

5.4.5 Characterisation Conclusion . . . 48

6 Conclusion 50

7 Future Outlook 51

8 Acknowledgement 52

References 53

APPENDIX I - Pictures 58

APPENDIX II - Supplementary Information 60

(7)

Abbreviations

ChCl Choline Chloride DES Deep Eutectic Solvent

DRC Democratic Republic of Congo

EDX Energy Dispersive X-Ray Spectroscopy EG Ethylene Glycol

EV Electrical Vehicle

FTIR Fourier Transform Infrared Spectroscopy HBA Hydrogen Bond Acceptor

HBD Hydrogen Bond Donor HCl Hydrochloric Acid HNO₃ Nitric Acid

ICP-OES Inductively Couples Plasma - Optical Emission Spectrometry LCO Lithium Cobalt(III) Oxide

LIB Lithium Ion Battery NaCl Sodium Chloride NaClO Sodium Hypochlorite

SEM Scanning Electron Microscopy TGA Thermal Gravimetric Analysis UV/Vis Ultraviolet–visible Spectroscopy XRD X-Ray Diffraction

XPS X-Ray Photoelectron Spectroscopy

(8)

1 Introduction

Lithium-ion batteries (LIBs) are found in many electric devices. From cellphones and laptops to screwdrivers and electric vehicles (EVs), which we use every day.

The usage has increased rapidly, but also changed since LIBs first becoming com- mercialized in 1991.¹From being a convenience then until today, when they also play a vital role in decreasing the usage of fossil fuels. As EVs have become increasingly more popular two questions have become more pressing: when will a bottleneck situation arise, and what will this be? For a long time, the electric grid infrastructure has been discussed,² and it is still ongoing.³ If it should fal- ter, it would not be possible to charge the batteries, and unusable batteries would not help the environment. Another bottleneck in the battery industry could be the access to raw materials. Tran et al. recently showed that it is possible to recycle cobalt in an environmentally friendly way utilizing deep eutectic solvents (DES).⁴ This was done by heating cathode material in a DES at different temperatures and times.

1.1 Cobalt

LIBs contain multiple components which will be discussed in more depth in section 1.2. The cathode material in batteries usually contains a couple of elements.

In the case of LIBs, it is commonly made from lithium cobalt oxide (LCO), LiCoO₂. This cobalt present in the cathode material, and its recovery, has been the focus of this project.

1.1.1 What is Cobalt?

Cobalt is a non-abundant silver-gray metal that exists in natural minerals. Cobaltite (CoAsS) and skutterudite (CoAs₃) are common minerals which are mined to obtain metallic cobalt.⁵More than half (53 %) of all cobalt refined in 2019 was used for battery production. Other uses of cobalt are catalysts, magnets, hard materials and superalloys. This last one is the second largest use of cobalt, claiming about 16 % of the total production.⁶

1.1.2 Production Chain

Globally, the deposits of cobalt rich ores are geographically very confined. The largest deposits can be found in the Democratic Republic of Congo (DRC).⁶They

(9)

are by far the lead cobalt producing country of primary (mined) cobalt with 64 % of the world production (average from 2010 - 2014, see figure 1 below). This has increased to 70 % in 2018 according to U.S. Geology Survey’s ”Mineral commod- ity summaries 2020”.⁷ No other country produces over 5 % of the worlds share of primary cobalt. Of all known deposits, DRC has just shy of half of the world’s cobalt reserves. If the rate of mining in the DRC continues at the present rate, they would run out by 2055 presuming no new cobalt deposits are discovered.⁷

Figure 1: The countries with the largest worlds share of mined cobalt. Average numbers from 2010 to 2014.⁵

Another problem with the cobalt mining in the DRC, other than that the reserves are running out, is the ethical aspect. Recently, it was reported that the leakages of metals, harmful to humans, was substantial from mines in the DRC. Urine samples from people living near a mine were analyzed and compared to the American Conference of Governmental Industrial Hygienists recommended levels. Highly elevated levels were found. The highest were in children, having almost 20 times the recommended maximum amount for adults working in a mine. In these children, oxidative damage to their DNA could be proven, which causes a plethora of problems. On top of this, high concentrations of uranium and manganese were

(10)

also confirmed in the general population.⁸

In almost every mine, except one mine in Morocco,⁵cobalt is mined as a byproduct of other metals. This is the case since, as mentioned above, cobalt is most often found with other metals. Common metals which are mined when producing cobalt as a byproduct are nickel and copper.^5,7 Because cobalt is a byproduct, the production is governed by the demand and the production of other metals. When the demand for cobalt is increased, it is therefore difficult to increase production when the other metals are not in high demand. To not have a shortage of cobalt, other metals would have to be stockpiled. This could lead to unpredictable prices as demand fluctuates.

When cobalt ore has been mined, it needs to be refined. This is usually not done in the same location as it is mined. The DRC does not refine much, but rather ships the major part to China. With 42 % of the worlds share of refined cobalt, China refine the most by far (see figure 2 below). More than 90 % of the cobalt refined originating from china, comes from the DRC.⁵

Figure 2: The countries with the largest worlds share of producing refined cobalt.

Average numbers from 2010 to 2014.⁵

(11)

1.1.3 Future Outlook of Cobalt

The future of cobalt is uncertain at best, but may be mitigated with good long term planning. Even in the short term, the cobalt usage is facing some significant problems. Cobalt consumption has been predicted to increase substantially in the coming years as EV production is expected to increase exponentially over the next decade.⁶This would result in cobalt reserves being depleted much faster than what was discussed above. In the DRC, natural cobalt depletion would thus occur sooner than 2055. This fact, combined with the necessity of cobalt, has placed it on the European Union’s list of critical raw materials, where it has been since the list was first published in 2011.⁵ Another parameter that is taken into consideration while compiling the list is how self-sufficient the European Union is. As illustrated in figure 1 and figure 2, the majority of production and refinement of all cobalt produced in the world is done outside Europe. Any halt in the supply chain from countries exporting cobalt would therefore cause major problems. The history of political turmoil in the DRC is also a point of concern. In a worst case scenario, a lot of the mines could be shut down if there were to be a severe long term conflict. Since the DRC is producing 70 % of the mined cobalt in the world, this poses very serious concerns.

To combat the shortage of cobalt, there are a few things that are currently being done. In the short term, more efficient mining processes are vital. Since cobalt is mined to 94 % as a by-product, it might turn out to be hard to increase the mined production of cobalt. But if needed, there are some things that could be done. It has been estimated that until 2028 the production of primary cobalt could be increased by 20 %, from more efficient mining processes alone. Another important part is new discoveries of cobalt, which would allow new mines to be opened to increase production. There is, however, no way to know how many, if any, new cobalt deposits there may be or if they will be available for mining. Because of all the challenges discussed above, recycling will be an important and even necessary part of a stable future for cobalt. In the short term to get production to par with demand, and in the long term to have any cobalt at all.⁵

1.2 Recycling Today

It was concluded above that recycling will be necessary in the future for a stable cobalt market. Batteries are, however, chemically complex devices that contains multiple and sometimes very different components as can be seen in figure 3.

(12)

both batteries and scrap material containing cobalt. A problem when recycling scrap material is, as with mining cobalt, that the cobalt production is dependent on the production of other metals. The recycling of cobalt is dependent on e.g.

the price of tungsten. When that price is increased, the amount of recycled cobalt is increased. Moreover, most of the cobalt containing scrap recycled in Europe is used for stainless steel. The cobalt is then not recovered for reuse in batteries, rather it is repurposed.⁵

Figure 3: Components making up a normal LIB, in wt. %.⁹

Battery recycling has not been very widespread historically. Most of what have been done has been due to legislation. In 1991 the European Union passed the first directive that was directly aimed towards batteries.¹⁰ It was later revised in 2003,¹¹ to be implemented three years before it was passed as a second directive.¹² This is still as of 2020 the directive that is in force. In this directive, goals for recycling and also the amount of used batteries collected are explicitly stated.

Today there is a lot of money invested into recycling, especially from the automo- tive industry.

There are multiple different fundamental ways to recycle batteries, which can be categorized as a chemical or as a physical process. The two most common ways are using heat and strong acids respectively,¹ and these are discussed in the following sections. Before the main phase of the recycling can take place, a

(13)

pretreatment step is required. This includes discharging, dismantling and usually additional steps depending on the following recycling method.^9,13,14

1.2.1 Pyrometallurgy

Pyrometallurgy uses heat as the primary mechanism to leach metals from the used material and it falls under the category of physical processes. The main advantage with pyrometallurgy is that it is a simple method. The work that needs to be done in the pretreatment step is minimal since the heat will deal with components that would normally be a problem. This includes, for example, the organic components of the batteries that the heat will evaporate. In other methods, a first mechanical step is necessary, but not in pyrometallurgy since everything will be melted, and then separated. This also means that different types of batteries can be recycled at the same time. There is no problem recycling LIBs at the same time as other batteries, e.g. batteries containing nickel and manganese.¹

The largest downside of pyrometallurgy is the energy consumption. Depending on the method used, the temperature can vary considerably. They are, however, usually very high, ranging from 1000 °C to above 1400 °C.^9,14Because of these high temperatures, batches are also very large, so it is not really feasible to recycle small quantities using pyrometallurgy. If present, cobalt, copper, and nickel are obtained as an alloy as a result of this method, while other metals like lithium, manganese, and aluminum end up in the slag. To be able to recover these, another method like hydrometallurgy would have to be used.¹

1.2.2 Hydrometallurgy

Hydrometallurgy is a chemical process that utilizes strong acids to leach the metals from scrap material. Because of some distinct advantages over pyrometallurgy, it is believed that hydrometallurgy will be the prominent technique for recycling LIBs. These advantages are first of all that the energy consumption is much lower since temperature is kept around 80 °C. Therefore, it is possible to recycle smaller volumes without problems, while also managing the waste in better ways. How- ever, the pretreatment needed is a lot more extensive than what is needed for pyrometallurgy.¹

Different acids are used, but the most common ones are phosphoric acid together with hydrochloric acid, sulfuric acid and/or nitric acid.¹³Recovery rates of 100 % have been reported from experiments when recycling cobalt with this method.^9,15

(14)

The overarching problem is the same as with pyrometallurgy, the processes are not environmentally friendly. The acids used could cause considerable pollution, e.g. from the gas evolved during leching. New procedures are tested with more environmentally friendly acids, e.g. citric acid.¹⁴

1.2.3 Other Methods

There are also additional methods for battery recycling, one of the most prominent as of today being biometallurgy. This is a method where inorganic and organic acids are produced by microbial activity which then can be used in the leaching step.¹⁶ Other methods are mainly based on mechanical strategies combined with one or both of the methods described in the sections above.

1.3 Deep Eutectic Solvents

DESs are mixtures of two or more components that have a lower melting point than either of the individual components. The eutectic solvents reported here are liquid at room temperature. As the name suggests, they have an eutectic point (see section 1.3.1). As a concept, DESs have not been studied for very long.

The first work published was done by Abbott et al.¹⁷ and was presented in 2001.

DESs are a subcategory of ionic liquids, which are also a somewhat new area of research. They were popularized in the late 90’s.¹⁸ In the mid 90’s, the term

’Ionic Liquid’ had replaced ’Room Temperature Molten Salts’ and it was defined as: ”a liquid consisting solely of cations and anions with a melting point of 100

°C and below”.¹⁹ DESs on the other hand have a different composition. There are four types of DESs of which this project will only be concerned about one:

type III. These are the first kind of DESs described by Abbott et al. and they are also the most environmentally friendly. The other DES types (I, II and IV) all contain metal ions, which in general increase the negative environmental effects.

That being said, not all of the type III DESs are perfectly safe either. While ionic liquids only contain ions, type III DESs follow the general formula Cat⁺X⁻zY.

Where Cat⁺ is a cation, usually with a nitrogen, sulfur or phosphor ion. The counterion, X⁻, would most often be a halide ion. Lastly, zY is the number (z) of molecules (Y) that acts as a hydrogen bond donor (HBD) which can also be viewed as a Lewis or Brønsted acid.²⁰ This also leads to the cation and the halide being viewed as a hydrogen bond acceptor (HBA).²¹ In this project the DES used was made from choline chloride (ChCl) and ethylene glycol (EG), where choline is the cation, chloride the anion and EG the HBD with z=2. The structure of these

(15)

compounds can be seen in figure 4. DESs are relatively easy to prepare, since the components can be mixed and stirred at a slightly elevated temperature. The DES made from ChCl and EG is, however, commercially available under the name ethyline.²²

Figure 4: Chemical structure of a) choline chloride and b) ethylene glycol. Drawn in ChemSpider version 2020.0.12.0.

1.3.1 Eutectic point

DESs are mixtures of two or more different compounds. This results in interesting properties of the resulting liquid. Many of these properties will be discussed in section 1.3.5 along with applications. The most characteristic property is what prompted the name; there is an eutectic point. When these compounds are mixed in a certain ratio, the melting temperature of the resulting mixture is lower than that of the individual components, resulting in an eutectic point. This is illustrated in figure 5 where the eutectic point is indicated by a green star. This point occurs at a specific mole ratio that is inherent to the components that form the DES. The eutectic point can easily be identified in a solid liquid phase diagram as the point of lowest measured melting temperature.

There are many HBAs and HBDs,^22,23 making the number of possible combinations very high, with each eutectic mixture having some specific mole ratio for its eutectic point. Why this is the case is not yet fully understood. It is thought, how-

(16)

Figure 5: Illustrative diagram of a solid–liquid phase diagram that shows the eutectic point (green star) of a two component DES at a 1:2 mole fraction. The individual components melting point can be seen as a red rhomb and a blue circle.

The most common mole ratios to reach the eutectic point is 1:1 as more than 40 % of known DESs have this as the eutectic mole ratio. Other commonly seen ratios are 1:2, 1:2.5 and 1:3 of HBA:HBD. There are also many other different ratios, but they are more rare. Except for those already mentioned, no eutectic ratio is the correct one for more than 6 % of all DESs.²⁴ One of the more uncommon ratio is when fructose or glucose are used as HBD in combination with ChCl as an HBA.

The eutectic ratio is then 2:1. Other rare ratios for other HBA and HBD are 1:1.5, 1:19 and 4:5.²²

1.3.2 Choline Chloride

Choline Chloride (ChCl) is a white solid at room temperature. Since the very first paper on DESs,¹⁷ ChCl has been the focus and is the most investigated HBA. It

(17)

was singled out by Abbott et al. because it gave the lowest melting point of the HBAs tested. Most articles written are using ChCl. It is also used as a reference when other HBAs are used.^20–23,25ChCl is a quaternary ammonium salt, in which the nitrogen atom is always positively charged, as can be seen in figure 4a.

ChCl has many properties which are advantageous, except being a component in DESs. It has been used in chicken feed for a very long time.²⁶ ChCl is therefore produced on a massive scale and widely available, making it relatively cheap.

Chickens are not the only ones that can metabolize ChCl, so can humans. We can use choline as a precursor to multiple compounds that we need. An adequate intake per day has therefor been established to 550 mg for men and 425 mg for women. We should not, however, consume more than 3.5 g/day since that can lead to hypotension which could manifest as sweating, diarrhea and/or fishy body odor.²⁷ChCl would be a good chemical to work with in a laboratory or industrial environment due to its low toxicity. It is both biodegradable and relatively cheap.

This means ChCl is advantageous from an environmental and economical perspective. The only drawback working with ChCl would be that it has a somewhat foul smell.

1.3.3 Ethylene Glycol

Ethylene glycol (EG) is a colorless, odorless, liquid at room temperature with a low viscosity. Like ChCl it is a commonly used chemical in the industry. As the simplest diol, see figure 4b, it is used as a precursor in a multitude of applications. Some of which are production of hydrogen, glyoxal, acetals, coolants, heat transfer fluids and polyesters.²⁸

EG is not that toxic to plants. In our bodies, however, it rapidly metabolizes which produces oxalic acid in the end. In the presence of calcium, oxalic acid will precipitate. In tissues, like the kidney, where calcium is present, this will cause problems.²⁸ The same happens in animals, so even if EG is not toxic to the environment, it should not be dumped in nature. EG is readily available and cheap, like ChCl.

1.3.4 Water

Water can be found almost everywhere and the air is no exception. Both EG and ChCl are hygroscopic compounds, pulling water out of the air. To dry them seems to be easily achieved, reading some of the literature.²⁹ After some time trying

(18)

heating on a hot plate, azeotropic distillation with toluene, molecular sieves and vacuum oven, it is hard not to concur with Gilmore³⁰ and Vilkov´a³¹ in their as- sessments that there will always be some water molecules present. Modeling has shown that a weight ratio of <5 % water in a DES does not affect thermophysical properties.²¹

1.3.5 Properties and Applications

One of the greatest advantages with DESs is that they can be specifically designed for specific applications. Since the number of combinations of HBD and HBA is so large, there are many parameters that can be adjusted. Important properties like density and viscosity can therefore be tailored.²³ With that said, there is still a lot that is unknown about how exactly DESs work.

In general, DESs have a high viscosity, which usually is attributed to the large number of hydrogen bonds.²³ These are believed to play a critical role in how DESs work, even though it is not yet discovered how.²² The halide used in the DES directly affects the melting temperature. When urea was used with choline and the anion changed, the melting temperature varied as following: F⁻ > NO⁻₃

> Cl⁻ > BF⁻₄. This would suggest that the hydrogen bonds are vital in under- standing DESs.²³ A recent study²¹ shed some light on hydrogen bonds, showing the change of energies and number of hydrogen bonds along with their placements as DESs were formed. It was concluded that the hydrogen bonds contribute to the decreased melting point. Viscosity is a transport property, as is the electric conductivity. In the same way that viscosity is heavily affected by hydrogen bonds, likewise is the electric conductivity since it is affected by the viscosity. This means that the conductivity is generally low (<2 mS cm⁻¹)²³ in DESs, since the ions are retarded by the high viscosity.²²

The vapor pressure of DESs is a very interesting property. This thermodynamic property is usually low or very low, and in some cases it can even be viewed as negligible. The vapor pressure of a DES is often measured in nanopascal or picopascal. This results in low to no evaporation over time for DESs, even when they are not in a sealed container.²² Another thermodynamic property that is of interest is the surface tension. Depending on the DES it can vary considerably, from 77 mN m⁻¹ down to 22 mN m⁻¹.²⁰ This could be seen in perspective of water that has a surface tension of 73 mN m⁻¹.³²

One of the primary reasons DESs have become more popular is because of their

(19)

environmental friendliness. Since they can also be tailored to fit certain properties, it has led to DESs being explored in many novel applications. They have success- fully been used as a reaction medium, with both reactants and catalysts being dissolved in the DES and then the reaction occurs there. Carbon dioxide can be dissolved, which has open up new reaction pathways for artificial photosynthesis.

Biofuel production also uses DESs to be able to synthesize their products in an environmentally friendly way.^22,23 It is also possible to dissolve metal oxides in DESs, which will be of interest in this project. The metal could then, after being dissolved, be refined via a method such as electrodeposition.⁴

(20)

2 Aim of the Project

The aim of this project was to expand on the novel work of Tran et al.⁴and their recycling process of lithium ion batteries. The same process of reclaiming cobalt was used. Instead of investigating time and temperature, however, the effect of water was the investigated parameter. Due to the hygroscopic nature of the deep eutectic solvent’s components, there is always some water present. Therefore the investigation is warranted, to see if water needs to be minimized or if it could possibly introduce some beneficial properties.

(21)

3 Experimental

All calculations done were performed with constants from SI chemical data,³³ unless otherwise stated.

3.1 Chemicals and Materials

Deionized water was used in all experiments in this work that required added water. The chemicals used can be seen in table 1. They were all used as received without further purification.

Table 1: Chemicals used in this project, their purity and source.

Chemical Purity Company

Choline Chloride High Purity Grade VWR

Ethylene Glycol 99.8 % Sigma-Aldrich

Hydrochloric Acid 37 % Merck

Hydrogen Peroxide 30 % VWR

Lithium Cobalt Oxide 99.8 % Aldrich

Nitric Acid 65 % Merck

Sodium Chloride 99.5 % Fisher Chemical

Sodium Hydroxide Pellets Analysis Merck

Sodium Hypochlorite Household Colgate-Palmolive Water Blank - ASTM Type I 18 Megohm ASSURANCE

3.2 Sample Preparation

For this project a 1:2 choline chloride to ethylene glycol DES mixture was used with three different water concentrations. The water percentages (w/w) used in the DESs were 2 %, 5 % and 10 %. Three identical samples were prepared from each DES, i.e. S1, S2 and S3 for DES containing 2 %, 5 % and 10 % water respectively.

The preparation of samples was done by weighing whenever possible, using the same scale for every measurement. When weighing was not an option, the volume was measured with an air displacement pipette followed by flushing the pipette tip if possible.

(22)

3.2.1 Deep Eutectic Solvent

The masses needed of the compounds to make the DES was calculated assuming that the compounds were pure. The mixture was then heated for at least two hours at 60 °C with stirring in an air tight bottle. The heating was continued until the DES was homogeneous.

The water added was calculated according to equation 1, where m_add is the mass of water that will be added and m_in is the mass of the initial DES before adding the water. Lastly, p is the target percentage (in decimal form) of added water in the final DES.

m_add= m_in

1 − p− m_in (1)

In this way, solutions of 2 %, 5 % and 10 % (w/w) added water were prepared.

3.2.2 Digestion

All digestions were done in glass test tubes with a piece of aluminum foil on top to prevent contaminants falling in while not letting pressure build up. The digestion solution was made by first weighing 100 mg of LCO and put in a test tube. Then, from the desired DES prepared as described in 3.2.1, 5.0 g was taken and placed in the test tube the same way as the LCO. Finally, a magnetic stirring bar was added to every sample except for 10 % S3.

An oil bath was heated to 160 °C, monitored by a thermometer connected to the hotplate for automatic feedback. All test tubes were then bundled together and immersed in the oil to the same level as the solution being digested. The stirring was turned on and the samples were then digested for 72 hours. Afterwards the fluid containing the digested cobalt was decanted and put in a vial. The leftover LCO that did not get digested was also saved, for quantification.

3.2.3 Precipitation

From each sample, 500 µl of digested sample was taken and used for precipitation.

Due to high viscosity, the solution was quantitatively transferred by a mechanical pipette. After shaking and standing some time, 400 µl of sodium hypochlorite (NaClO) was added to each sample. After shaking, the samples were centrifuged

(23)

at 2500 rpm for three minutes. More NaClO was added with a subsequent cen- trifuging step, using the same parameters as the first time. This cycle was continued until no reaction could be observed with additional NaClO. The samples were then allowed to rest overnight. More NaClO was then added, followed by cen- trifugation. The speed was increased to 3500 rpm while the time of three minutes was maintained. The most NaClO used in total was 2.4 ml for the richest samples.

The supernatant was then removed with a metal syringe and kept for analysis with ultraviolet–visible spectroscopy (UV/Vis).

3.2.4 Analysis Preparation

The density measurements were performed on the three DESs with different water concentrations after the temperature was measured. This was done by taking 1000 µ l with an air displacement pipette and then weighing it. For accuracy, the average of three measurements were used for every sample to determine the density. To eliminate instrumentation bias, the same measurements were performed on water that had the same temperature as the DESs. Instead of using the average of three measurements, the average of ten was used for extra precision. The obtained value was then compared to literature to get the quotient to find any possible deviation from reported values. The quotient was then multiplied with the values obtained for the DESs to get their actual density.

After the digestion, the samples were decanted with pasteur pipettes and the digested solution stored in 20 ml glass vials. The residual powder in each test tube was then filtered out. In this way the undigested LCO was recovered in the filter paper and could be weighed.

For fourier transform infrared spectroscopy (FTIR), Karl Fischer and thermal gravimetric analysis (TGA), the digested solution can be used without any further treatment. Unused DESs were also analyzed in the same way. For UV/Vis measurements, the solution had to be diluted. The dilution was done using the same DES that was used to prepare the samples. For the 1:10 dilution, 0.100 ml of the samples was diluted with 1.000 ml DES. For the 1:40 dilution, another 3.000 ml DES was added. Since the DESs were viscous, all dilutions were weighed so they could be back calculated to their actual dilution. The 1:10 dilutions were measured with a 1 cm cuvette pathway. The 1:40 dilutions, on the other hand, used a 0.5 cm cuvette pathway. In this way the peaks of interest did not exceed an absorbance of 1 A.U.

(24)

To run X-ray diffraction (XRD), the precipitate was dried overnight at 70 °C and in a vacuum of 300 mbar and analyzed without further preparation. Afterwards the same powder was used for analysis with scanning electron microscopy (SEM).

Inductively coupled plasma optical emission spectrometry (ICP-OES) analysis preparation on the other hand, required acid digestion. This was done in a typ- ical 1:3 mix of concentrated hydrochloric acid (HCl) and concentrated nitric acid (HNO₃). The digestion was allowed to continue for at least 24 hours, where the HCl digests any residual organic components and HNO₃the inorganic. After that, 1 000 µl of the digested sample was diluted with ultra pure water from Assurance to 10 000 µl, so that a solution containing roughly 5 % acid was obtained. The solution was then ready to be analyzed with ICP-OES. For extra security, the solutions were also analyzed as diluted to make sure the correct concentrations were obtained.

3.3 Instrumentation

During this project, many different instruments were used. To the greatest extent possible, the same scale, pipette, fume hood and every day equipment were used every time to minimize bias from those sources.

3.3.1 Centrifuge

A four slotted centrifuge of the brand sigma model 2-6 with a maximum speed of 4000 rpm was used. To avoid accidents, the centrifuge was always balanced, using counterweights.

3.3.2 Filtration

To obtain the undigested LCO, filtration (gravimetric analysis) was utilized. The setup contained a B¨uchner funnel, an elastomer adapter and a B¨uchner flask which was connected to vacuum. The filter papers used were Whatmans 1450-070 quan- titative cellulose filter paper with a 2.7 µm pore size.

3.3.3 Fourier Transform Infrared Spectroscopy

FTIR was done using a Bruker Tensor 27 using an air blank. Liquid samples were dropped on top of the ATR aperture without the arm pressed down. All experiments were the results of the average of 32 scans each.

(25)

3.3.4 Heating

For heating and stirring a Heidolph MR Hei-tec hot plate with 145 mm diameter was used alongside the accompanying thermometer. When chemicals were to be dried, a Binder VD 23 Vacuum Drying Oven with article number 9030-0029 was used. It was connected to vacuum pump from Vacuubrand with model name Chemistry Diaphragm Pump MZ 2C NT.

3.3.5 Inductively Coupled Plasma - Optical Emission Spectrometry

The ICP-OES was performed on a Perkin Elmer Avio 200 with Syngistix ^® 2.0 software package. The calibration curves used had concentrations 1 mg kg⁻¹, 10 mg kg⁻¹, 100 mg kg⁻¹ and 500 mg kg⁻¹.

3.3.6 Karl Fischer

Karl Fischer was done using a Metrohm 756 KF Coulometer in combination with a Metrohm 728 Stirrer. The solution used was a Merck’s CombiCoulomat frit- less Karl Fischer reagent for coulometric water determination for cells with and without diaphragm Aquastar^®.

3.3.7 Thermal Gravimetric Analysis

TGA was done using a TA Instruments TGA Q500. Aluminum pans were used with an autosampler that had a 16 pans capacity. Nitrogen gas was used at a default flow rate of 60 mL/min. For EG two ramping rates, 5 °C min ⁻¹ and 10

°C min⁻¹, were used. For all other measurements, 5 °C min⁻¹ was used.

3.3.8 UV/Visible Spectroscopy

The UV/Vis screening was performed on a Varian Cary 50 Bio UV-Visible Spec- trometer using a scan rate of 600 nm/min. The UV/Vis data presented in this thesis were collected using a Varian Cary 5000 UV-Visible-NIR Spectrometer run at a scan rate of 400 nm/min. The cuvettes used were VWRs disposable PMMA cuvettes 634-06778P, with plastic cuvette caps. These were only used within the recommended interval of 300 nm and 900 nm.

The concentrations were calculated from Lambert Beer’s law, see equation 2,

(26)

where A is the absorbance, ε is the molar absorptivity, c is the concentration and l is the path length of sample used during analysis.

A= ε ∗ c ∗ l (2)

3.3.9 Scanning Electron Microscopy

SEM was used to obtain pictures of the surface of the precipitate. The instrumentation used was a Zeiss 1550 Gemini, which is a field emission SEM resulting in higher resolution. Energy dispersive x-ray spectroscopy (EDX) was also used to obtain the elements in the analyzed material.

3.3.10 X-Ray Diffraction

The crystal structure of one sample from each water content was analyzed with XRD. The instrument used was a Bruker D8 twin twin, Bragg-Brentano geometry.

It had fixed samples position with a movable Cu κα x-ray source. The time per step used was 0.5 seconds with a 10 rpm rotation of the sample. The 2θ range was 10-90.0014° with a step size of 0.020566°, resulting in 3891 steps.

(27)

4 Results

When a figure or table with a roman number is referenced, that can be found in the appendix with the corresponding roman number. For instance, figure I.2 would be found in Appendix I.

4.1 Procedure

4.1.1 Preparation

The stock solution of DES, without any added water, was mixed and then ho- mogenized with the lid screwed tight. The first time the bottle was opened after preparation, small white needle shaped crystals started to form. They became visible after about a minute and they continued to grow at a fast rate. Every minute the crystals could be observed to have increased in size. When additional water was added to the DESs, the small crystals dissolved within minutes. The crystals in the stock solution continued to grow for about a day, after which they stopped.

They were brittle and fell apart upon swirling the bottle. The resulting broken crystals can be seen in figure 6. To see crystals earlier in the growing process, see figure I.1. Upon taking the crystals out of the DES, they started to melt in room temperature and atmosphere. They were analyzed with TGA, figure II.1.

4.1.2 Digestion

The test tubes containing the samples were immersed in the oil bath and digestion started. A green color could be seen within the first hour and it deepened after the second hour. During the first hour, all three samples containing DES with 10 % water started to boil. After the full 72 hours, a very dark green-blue color could be observed, which can be seen in figure 7. The samples look the same visually without very close visual inspection. Upon handling them, two samples stand out. Sample S2 and S3 that contained DES with 10 % water had distinctively less color. This could be observed when smaller amounts were handled, e.g. with pipette. The test tubes where the digestion took place had a brown stain after being used, see figure I.2. This could be due to adsorbed ChCl or alternatively degraded DES from the elevated temperature during digestion.

An extra sample was prepared alongside the samples that were digested. This sample was left in room temperature to see if it would digest given enough time.

After 45 days there was no visible change, and no color could be seen.

(28)

Figure 6: Crystals in the stock solution of DES. These broken crystals could be seen after the bottle had been swirled. Before that, the crystals were longer and grew throughout out the DES.

Figure 7: Digested samples that have been decanted and put in vials.

(29)

4.1.3 Precipitation

Precipitation was tested with hydrogen peroxide and sodium hydroxide but without good results. After 24 hours there was no precipitate. When NaClO was used, a precipitate could be seen immediately (see figure I.3).

4.2 Properties

4.2.1 Density and Viscosity

To be able to convert concentrations between molar concentration (mol dm⁻¹) and ppm (mg kg⁻³), the molar mass and density are needed. The first can be looked up, but the other one is harder to find. The solution density was therefore measured and the results can be found in table 2 below.

Table 2: Density for the different DESs were measured at 22 °C, while the referenced value was measured at 25 °C.

DES Density

% H₂O (g cm⁻³) 0 1.1177³⁴

2 1.1140

5 1.1096

10 1.1086

Viscosity was not measured in this study, but it was observed to change between the DESs with different water concentrations. As the water content increased in the DES, the viscosity decreased to such a degree that it could easily be observed with the naked eye when handled.

4.2.2 Karl Fischer

The water content in the chemicals, DESs and digested samples was monitored with Karl Fischer titration since this was a vital part of this project. The results can be seen in table 3. These values are averages of two measurements, which all can be seen in table II.3. As a control sample, 5 % H₂O was added to EG in the same way it was added to the DES. Unfortunately, the Karl Fischer broke down before the digested samples and the DES with 10 % water could be measured.

(30)

Table 3: H₂O concentrations measured with Karl Fischer.

Sample Concentration Molar Ratio (mg kg⁻¹) (H₂O:DES)

EG 13

EG 5 % 5000

DES 2 % 2200 1:31

DES 5 % 5100 1:13

The DES and its components were analyzed with TGA to increase the understand- ing of how the DES works. EG was first to be examined and the result can be seen in figure 8. Two different ramp rates were used and in the third measurement an isothermal step at 105 °C. ChCl and DES with 2 % added water were both measured at 5 °C min⁻¹ and can be seen in figure 9 and 10. The crystals growing in the DES stock solution was also analyzed, which can be seen in figure II.1.

Figure 8: TGA of EG with three different parameters setup. The onset points are marked for the experiments with linear ramping temperature.

(31)

Figure 9: ChCl analyzed with TGA, with an isothermal step at 260 °C. Weight loss from initial temperature up to 150 °C marked along with the onset point.

Figure 10: DES with 2 % water added, analyzed with TGA. Two isothermal steps were performed at 115 °C and 260 °C. Weight loss at 105°C and 260°C marked.

(32)

4.3 Concentration

4.3.1 Filtration

The amount of undigested LCO was filtered so that the amount of digested LCO could be estimated. This was done by subtracting the filtered weight from the starting amount. From this amount, the percentage digested and cobalt concentration in both ppm (mg kg⁻¹) and molar concentration could be calculated. All results can be seen in table 4. These results were calculated with the initial amount of chemicals, which can be seen in table II.1. The percentage of LCO digested was calculated by dividing the digested amount with the initial amount. The concentration, in ppm, was then calculated by dividing the amount of digested LCO (in mg) with the starting amount of DES (in kg). To only obtain the cobalt concentration, the other constituents of the LCO need to be taken into account. The concentration was then calculated in molecular concentration by dividing with the molecular weight and multiplying with the density. For sample 2% S1 this was done in the following way.

Cobalt percentage in LCO calculated from molecular masses (g mol⁻¹):

58.933 g mol⁻¹

(58.933 + 2 ∗ 15.999 + 6.941) g mol⁻¹ = 0.6021 LCO concentration (ppm) in the digested sample:

49.9 mg

0.0050014 kg = 9977 mg kg⁻¹

Cobalt concentration (ppm) in the digested sample, assuming even digestion of LCO:

9977 ∗ 0.6021 = 6008 mg kg⁻¹

Cobalt concentration recalculated into molecular concentration, using density from table 2:

6008 ∗ 1.1140 ∗ mg kg⁻¹∗ kg dm⁻³

58.933 g mol⁻¹ ≈ 114 mmol dm⁻³

In this way all concentrations were calculated. For tabulated values of digested LCO in grams and concentrations in ppm, see table II.2.

(33)

Table 4: Lithium Cobalt Oxide (LCO) digested, determined by filtration of undigested LCO.

Sample LCO Undigested LCO Digested Co Conc.

(g) (%) (mmol dm⁻³)

2% S1 0.0492 50.4 114

2% S2 0.0603 40.2 92.1

2% S3 0.0382 62.6 143

5% S1 0.0534 47.0 107

5% S2 0.0489 51.1 116

10% S1 0.0435 56.8 129

10% S2 0.0756 24.7 56.2

10% S3 0.0759 25.3 58.2

4.3.2 UV Visible Spectroscopy

There is much information available from UV/Vis. If there is any structural difference between samples, that can be seen. The concentration can also be obtained if the molecular absorptivity is known. Therefore UV/Vis spectra were recorded and can be seen in figure 11-16. Large peaks can be observed in figure 11, 13 and 15. These were established to belong to [CoCl₄]²⁻, see section 5.3.2. The molar absorptivites used to calculate concentrations were 460 mol⁻¹ dm³ cm⁻¹, 640 mol⁻¹ dm³cm⁻¹ and 740 mol⁻¹ dm³cm⁻¹ for the peaks at 631 nm, 667 nm and 695 nm respectively.³⁵ The average concentration of the three peaks can be seen in table 5.

The absorbance was calculated as the values obtained from UV/Vis multiplied with the dilution, which can be found in table II.4. This was done as follows, with samples 2% S1 at the 631 nm peak as an example.

c= A

ε ∗ l = 0.45268 ∗ 43.4377

460 ∗ 0.5 ∗ mol⁻¹dm³cm⁻¹∗ cm ≈ 85.5 mmol dm⁻³

The concentrations for each individual peak can be seen in table II.4. The two smaller peaks seen in figure 12, 14 and 16 around 450 nm and 530 nm were determined to be [Co(H2O)₆]²⁺, while the third peak around 550 nm remain uniden-

(34)

reported to be 4.55 mol⁻¹dm³cm⁻¹.³⁵ When this was used to calculate the concentration, values corresponding to more LCO than what was initially used were obtained.

Figure 11: UV/Vis spectrum of all samples that used DES containing 2 % added water. It was diluted 1:40 times for analysis.

(35)

(36)

(37)

Table 5: Concentrations of cobalt in the form of [CoCl₄]²⁻, obtained from UV/Vis. The values are averages of values found in table II.4.

Sample Concentration (mmol dm⁻³)

2% S1 83.4

2% S2 58.9

2% S3 94.9

5% S1 65.6

5% S2 81.0

5% S3 104

10% S1 87.0

10% S2 35.6

10% S3 37.7

The supernatant that was removed from the precipitate and kept for analysis did not show any absorption using UV/Vis.

4.3.3 Inductively Coupled Plasma - Optical Emission Spectrometry

ICP-OES measures concentration very accurately and is a good complement to UV/Vis. The cobalt in the samples was precipitated (see figure I.3), digested with acid and measured with ICP-OES. The results can be seen in table 6. The vali- dation done with diluted samples can be seen in table II.5. Example calculations of how the concentrations were converted from ppm to molecular concentrations can be seen in section 4.3.1.

4.4 Characterisation

In this section the structure of the recovered cobalt was analyzed in an effort to characterize it. The focus has been to determine whether there is a difference between the samples or not, rather than obtaining the full structure. The result from UV/Vis is also applicable here, but can be seen in section 4.3.2.

(38)

Table 6: Cobalt concentrations obtained from ICP-OES Sample Concentration Concentration

(mg kg⁻¹) (mmol dm⁻³)

2% S1 302 5.71

2% S2 222 4.19

2% S3 369 6.98

5% S1 282 5.33

5% S2 306 5.76

5% S3 402 7.56

10% S1 354 6.62

10% S2 134 2.52

10% S3 142 2.67

4.4.1 Fourier Transform Infrared Spectroscopy

FTIR is a fast method that can detect differences between samples by utilizing infrared light, making this method a good complement to UV/Vis. Much information about the structure can also be obtained with further analysis of FTIR spectra.

Each digested sample was analyzed with FTIR, which can be seen in figure II.2- II.4. An average was then compiled for each water content from these results. In figure II.5, the results of the averages can be seen. To see clear differences, the same spectra have been stacked on top of each other, which can be seen in figure 17.

4.4.2 Scanning Electron Microscopy

SEM and EDX are done using the same instrument. The pictures obtained from SEM makes it possible to visually inspect the surface of the precipitate. The same precipitate samples that were used for XRD analysis were used for SEM and EDX.

Pictures of the samples can be seen in figures 18-20.

(39)

Figure 17: FTIR spectra comparing results of digestion done in a DES with 2 %, 5 % and 10 % (w/w) added water.

(40)

Figure 18: SEM pictures of the precipitate from the DES containing 2 wt % water.

(41)

4.4.3 Energy Dispersive X-ray Spectroscopy

The information that can be obtained from EDX is the elemental composition, which is similar to what can be obtained from x-ray photoelectron spectroscopy (XPS). EDX does not penetrate as deep as XPS, but is considerably faster. Since nothing has suggested that a deeper penetration is needed, EDX was used. This also gives access to SEM. The elemental composition can be seen in figure 21.

It is a comparison between the precipitates of the 5 % and 10 % DESs. Multiple spectra were taken on each precipitate. Spectra 1-4 were taken on the precipitate of DES with 5 % water while spectra 5-7 and 8-10 were from 2 % and 10 % water content respectively. All spectra (1-10) were obtained at different locations, which can be seen by SEM in figures II.15-II.17. All graphs (like figure 21) can be seen in figures II.6-II.14. Elements were mapped in the area seen in figure 22. The result for cobalt can be seen in figure 23, while the other elements can be seen in II.18-II.22.

(42)

Figure 21: Comparison of EDX graphs of precipitate from DES containing 5 % (Spectrum 1) and 10 % (Spectrum 10) water.

4.4.4 X-Ray Diffraction

XRD gives information about the crystal structure of the analyzed powder. The samples named S1 from each water content were precipitated and analyzed with XRD. The results can be seen in figure 24. The same diffractograms can be seen stacked on top of each other in figure II.23.

(43)

Figure 22: SEM picture of the area mapped for cobalt in figure 23.

Figure 23: Cobalt (orange) mapped with EDX of the area seen in figure 22.

(44)

Figure 24: XRD diffractograms comparing precipitates originating from a digestion of LCO in a DES with 2 %, 5 % and 10 % (w/w) added water. NaClO was used to precipitate the digested solution.

(45)

5 Discussion

The different water contents will be discussed from four different points of view.

How it affected the procedure and properties, but also the concentration and char- acter of the obtained cobalt.

5.1 Procedure

5.1.1 Sample preparation

The stock solution of DES was made with the lowest possible concentration of water. When water was added to the DES, it seemed to stabilize it since the crystals disappeared and did not regrow. When the crystals were put in room temperature and atmosphere they started to melt. This is probably due to the crystals absorbing water from the air, showing a hygroscopic nature.

When viewing the TGA result of the crystals, figure II.1, it suggests that a substantial amount of water is part of the structure. Comparing to the result of the 2

% DES, in figure 10, less water seems to be present there. This could, however, be skewed due to DESs potential unwillingness to part with its water. This will be further discussed in section 5.2.3. It could also be EG starting to leave at low temperatures that increases the early weight loss. Even if this is true, combined with the crystals releasing water easier than the DES, a substantial amount of water was present in the crystals. Finally, within the sample preparation, it can be concluded that time on its own is not enough to digest LCO. The sample left in room temperature suggests that heat is needed, since an excessive amount of time was not enough.

5.1.2 Digestion

When the digestion was started, a pale green color could be seen within the first hour. This suggests that the digestion starts to work immediately. If access to an in situ UV/Vis was possible, more information could be obtained. It could be, e.g., investigated if there was a difference between DESs with different water content.

Then a possible saturation could also be noticed.

During the digestion, the samples with 10 % (w/w) added water started to boil.

Since no facts have been discovered that boiling would be preferred, this is not desired. For lower temperatures, this composition might not boil, but for 160 °C

(46)

10 % water is too much.

The color achieved after the full 72 hours of digestion does have striking similari- ties to the color in one of the steps of Sotiles et al.³⁶They state that the blue/green color is due to a Co(II) complex. Specifically [Co(H₂O)₆]²⁺. Since there is no further information that can corroborate this at the moment, it will be put on hold for now. The continued discussion can be found in section 5.3.2.

5.1.3 Procedure Conclusion

From the procedure there are two things that can be concluded for the water content in DESs. First, it seems to stabilize the DES so that it does not crystallize.

Even though the DES would be fully liquid during digestion due to the heat added, it is not optimal to have crystals. Second, there can also be too much water. Not because it is questioned whether it is still a DES in use, but because it started to boil during digestion.

5.2 Properties

5.2.1 Density and viscosity

The density and viscosity changes reported in section 4.2.1 are in line with other reports.³⁷As stated in section 1.3.5, one of the big advantages with DESs is their tunability. With water as another way to tune properties, this advantage becomes even bigger.

5.2.2 Karl Fischer

In section 1.3.4 it was stated that water is very hard to completely remove from EG and ChCl. In table 3 it can be seen that anhydrous EG contain 13 mg kg⁻¹ water. The molar ration of EG to water is calculated by dividing number of EG moles with water moles:

n_EG n_H₂_O =

1000g

62.068g mol⁻¹

0.013g

18.015g mol⁻¹

= 16.111

7.2162 ∗ 10⁻⁴ ≈ 22 300

When the DES stock solution was prepared, the water content in EG was disre- garded. Seeing that for every mole of water, there are over 22 000 moles of EG,

(47)

disregarding water did not upset the molar ratio in the DES.

EG has a boiling point of 197.3°C.³⁸ From the graph of EG, figure 8, it can be seen that it starts to dissipate at a relatively low temperature. The isothermal step at 105 °C shows that a lot has disappeared. This is not unexpected, since water does the same when it evaporates in room temperature. At 105 °C EG has about the same vapor pressure as water has at 20 °C, 2.3 kPa.^39,40

In figure 9, the TGA graph of ChCl can be seen. The melting point, at which it also decomposes, is 305 °C for ChCl.⁴¹ In contrast to EG, ChCl is rather stable at high temperatures. This can also be seen in the isothermal step at 260 °C. This is due to ChCl being a solid when EG on the other hand is a liquid, which has higher vapor pressure. The weight loss shown at 150 °C is due to the water contained in the ChCl, corresponding to 610 mg kg⁻¹. This assumes that all water has left while none of the ChCl has. Seeing the rest of the graph, this is likely a valid assumption. The number of moles ChCl per mole water can then be calculated.

This was done in the same way as it was calculated for EG in section 5.2.2.

n_ChCl n_H₂_O =

1000g

139.62g mol⁻¹

0.61g

18.015g mol⁻¹

= 7.1623

3.3861 ∗ 10⁻² ≈ 212

For every mole of water there are thus 212 moles of ChCl. This water content is higher than in EG, but still low. When making the DES, this water content did not have an impact large enough to warrant adjusting the weight. Taken the water into account for both EG and ChCl, a ratio of 2.009 would have been obtained instead of 2.000. This is within the margin of error.

The TGA graph of the DES can be seen in figure 10. Despite containing 2 % water, there is only 1.3 % of weight lost at 105 °C. As seen above, some of the EG should also have left at this stage. This is, at least to some extent, due to the low vapor pressure of DESs as stated in section 1.3.5. At higher temperatures, the ChCl in the DES disappears faster than the pure compound does. The total weight of the EG in the 1:2 DES is calculated as follows:

2 ∗ 62 g mol⁻¹

= 0.4697 ≈ 47 %

(48)

The total weight of EG in the DES containing 2 % water is then:

0.98 ∗ 0.4697 = 0.4603 ≈ 46%

The isothermal step at 260 °C shows a much larger drop than in the graph than for pure ChCl. The weight loss shown in the graph, 60 %, suggests that more than 10 % of the lost weight comes from ChCl. This is the case since water makes up 2 % and EG 46 % of the DES. It would seem like the EG is bringing the ChCl with it when it is dissipating. The last of the DES disappears at the same time as for pure ChCl. The DES acts as if it was both one component and multiple.

One component because there is a somewhat constant loss of weight. Multiple components because the loss is not smooth as it was for EG.

5.2.4 Properties Conclusion

By all accounts and measurements, the DES used in this project behaves as other DESs that have been reported in literature. Here, the viscosity and density have been addressed. Many other properties would, however, also be affected with the addition of water. The thermal stability is something that needs to be taken into consideration when planning experiments.⁴² The DES used is shown, by TGA, to be affected at 160 °C. The condition in a TGA pan is, however, very different from bulk. In the latter the surface area to volume is much smaller. With the low vapor pressure, 160 °C did not show any signs of being a problem. Even after the relative long time of 72 hours.

5.3 Concentration

5.3.1 Filtration

There is no doubt the LCO was digested, which could be concluded from the color change observed during digestion and the noticeable sample loss as shown by gravimetric analysis. It is possible that the amount of LCO digested could be skewed with a loss during filtration due small particles as the main reason. When the LCO was digested, the particle size decreased. This could lead to some LCO not being trapped by the filter paper. Due to the visual acuity, particles this small would not be visible to the naked human eye.⁴³ These particles would therefore not have been noticed when the filtrate was inspected. In this study the filtrate was not analyzed for cobalt beyond visual inspection.

(49)

Gravimetric analysis is overall a good method to determine the amount of digested solids. It is simple and easy to perform. With that said, there can be losses as discussed above. This would lead to less undigested LCO being collected. This would then overestimate the cobalt concentration in the solutions.

5.3.2 UV Visible Spectroscopy

For the spectra obtained after dilution, the first thing needed was to identify the peaks of interest. There are limited number of chemical species in the samples that could complex with cobalt ions. There is ethylene glycol, choline, water and chloride ions present. The oxidation number of the cobalt needs to be considered too, since this could also change the peaks.⁴⁴ In LCO, cobalt (III) is found. According to some theories, as a part of the mechanism during digestion, cobalt is reduced.⁴ Even if this is not the case, the standard potential of Co²⁺/Co³⁺= +1.92 V is relatively high.³³ So if there would be something present that could act like a re- ducing agent, like water, the cobalt would then be reduced anyway. Since this is the case, it should be most likely to find Co²⁺in the digested solution.

In the original article it was stated that the blue/green color obtained was due to [CoCl₄]²⁻.⁴Since a very similar color (see figure 7) was obtained in this project, it is possible that the content also is the same. Except for one article,³⁶ all other literature found confirms the three large peaks (figure 11, 13 and 15) being cobalt tetrachloride.^45–49 Cobalt tetrachloride is, when possible, in an equilibrium with [Co(H₂O)₆]²⁺, hexaaqua cobalt, as seen below.

[Co(H₂O)₆]²⁺(aq) + 4Cl⁻(aq) *) [CoCl₄]²⁻(aq) + 6H₂O(l)

This equilibrium is easily affected by adding water and chloride ions but also by temperature. Cobalt tetrachloride has a blue color, while hexaaqua cobalt is pink.⁵⁰It is therefore common to use these compounds⁵¹to demonstrate Le Chate- lier’s principle.⁵² Since water was present in all samples in this project, it is reasonable to assume that hexaaqua cobalt is present. The equilibrium above should be strongly shifted towards the right hand side due to the concentration difference between water and chloride ions. The difference can be seen in table 3, keeping in mind that there is one mole of chloride ions in every mole of DES. Hexaaqua cobalt has two peaks at lower wavelengths than cobalt tetrachloride,35,45,47,49,53

which seems reasonable since water is a stronger ligand than chloride.⁵⁴ If a peak