DEPARTMENT OF CHEMISTRY AND 
CHEMICAL ENGINEERING 

CHALMERS UNIVERSITY OF TECHNOLOGY 
Gothenburg, Sweden 2020 

www.chalmers.se  

Investigation of Cu and Zn 
interactions with ilmenite during 
waste incineration 
Modelling of phases and reactions using a 
thermodynamic modelling approach 
Master’s thesis in Chemical Engineering 
 

 
 
ELIN BORGMAN 
 



 
 
 
 
 
Investigation of Cu and Zn interactions with ilmenite during waste incineration 
 

© ELIN BORGMAN, 2020. 

 

Department of Chemistry and Chemical Engineering  

Chalmers University of Technology  

SE-412 96 Göteborg  

Sweden  

Telephone + 46 (0)73-430 9554 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Göteborg, Sweden 2020 

 

 

 

 

 

 



Investigation of Cu and Zn interactions with ilmenite during waste incineration 

ELIN BORGMAN 

Department of Chemistry and Chemical Engineering 

Chalmers University of Technology 

 

SUMMARY 

In 2018, approximately 2.01 billion tons of waste was generated in the world. By 2050, 

this number is expected to increase drastically to 3.40 billion tons of waste annually as 

lower income countries develop and the world population increases (Kaz, Yao, Bhada-

Tata, & Van-Woerden, 2018). Some of the waste can be combusted to produce heat and 

power, leaving residual ash behind. This also decreases the volume of the waste flow by 

90 % (Lam, Ip, Barford, & McKay, 2010). As waste generation is expected to increase 

while the land available for landfilling decreases, it is of interest to recycle as much as 

possible from waste, as well as from ash, and move towards a more circular economy in 

which no new raw materials are needed. 

 

One technology that can be used for waste incineration is Oxygen Carrier Aided 

Combustion (OCAC). OCAC is a type of fluidized-bed combustion, in which the 

conventional inert bed material is replaced fully or in part with metal oxides, called 

oxygen carrier. The oxygen carrier is reduced in fuel rich parts of the reactor and oxidized 

in oxygen rich parts during the combustion, thereby improving the distribution of oxygen 

in space and time. This improves the combustion efficiency. The interaction between the 

oxygen carrier and fuel ash can lead to an uptake of certain elements such as zinc, copper, 

potassium, calcium and phosphorous among others.  

 

The aim of the thesis was to investigate how major ash components (Si, Al, Fe, Mg, Ca, 

K, Na and Cl) and the minor ash components Cu and Zn interact with the oxygen carrier 

bed material ilmenite (FeTiO3) during waste incineration. A focus on Cu and Zn was 

chosen since these usually have the highest concentrations out of the trace metals that can 

be valuable for recycling, as they are high-cost metals and energy intensive to produce. 

The investigation was carried out by simulating different boiler conditions and 

performing thermodynamical equilibrium calculations (constructing phase diagrams) 

using the software FactSage 7.2.  

 

The simulations indicated that ilmenite interacts with Al2O3, CaO/CaCO2, Na2O, K2O, 

MgO, ZnO and Cu/Cu2O but does not interact with SiO2. The compounds formed from 

interactions between ilmenite and ash which could be relevant for OCAC applications are 

FeAl2O4, CaTiO3, (Na2O)(TiO2)3, (Na2O)(TiO2)6, ZnFe2O4, (ZnO)2TiO2, (Cu2O)(Fe2O3) 

(s2) and (CuO)(Fe2O3) (s3). Experimentally, potassium has been found as KTi8O16 in 

ilmenite particles rather than as K2Ti3O7 and K2Ti6O13 as indicated by the simulations. 

The compound KTi8O16 is missing in the software database, therefore giving inconclusive 

results not corresponding to experimental findings. 

 
Keywords: FactSage, copper, zinc, major ash elements, interactions, bed material, combustion, 

ilmenite, thermodynamical equilibrium calculations, phase diagram , oxygen carrier, ash 

interaction



 

Contents 
1. Introduction .................................................................................................................................. 1 

1.1. Background ............................................................................................................................... 1 

1.2. Ash formation ........................................................................................................................... 3 

1.3. Aim............................................................................................................................................ 5 

1.4. Limitations ................................................................................................................................ 5 

1.5. Specification of issue under investigation ................................................................................ 5 

2. Methodology ................................................................................................................................ 6 

2.1. SEM-EDX-analysis of obtained sample ................................................................................... 6 

2.2. Simulation conditions ............................................................................................................... 6 

2.2.1. FactSage Phase Diagram .................................................................................................... 6 

2.2.2. Inputs to the software ......................................................................................................... 7 

3. Results & Discussion ....................................................................................................................... 9 

3.1. SEM-EDX-analysis of obtained sample ................................................................................. 10 

3.2. Oxidation/reduction behavior of ilmenite ............................................................................... 12 

3.3. Interactions between ilmenite and ash-components ................................................................ 15 

3.3.1. Interactions with major ash component SiO2 ................................................................... 15 

3.3.2. Interactions with major ash component Al2O3 ................................................................. 16 

3.3.3. Interactions with major ash component CaO ................................................................... 18 

3.3.4. Interactions with major ash component CaCO3 ............................................................... 19 

3.3.5. Interactions with major ash component Na2O (and Na2O/Cl) ......................................... 22 

3.3.6. Interactions with major ash component K2O (and K2O/Cl) ............................................. 24 

3.3.7. Interactions with major ash component MgO .................................................................. 26 

3.3.8. Interactions with minor ash-component ZnO .................................................................. 28 

3.3.9. Interactions with minor ash-components CuO/Cu2O....................................................... 29 

3.4. Interactions between ilmenite and multiple ash components .................................................. 31 

3.4.1. Interactions with Zn and Cl .............................................................................................. 31 

3.4.2. Interactions with Cu and Cl ............................................................................................. 33 

3.4.3. Interactions with Zn and S ............................................................................................... 35 

3.4.4. Interactions with Zn and SiO2 .......................................................................................... 37 

3.4.5. Interactions with Zn, H2O and CO2 ................................................................................. 39 

3.4.6. Interactions with Zn and Al2O3 ........................................................................................ 41 

4. Conclusions .................................................................................................................................... 43 

Acknowledgements ............................................................................................................................ 45 

References .......................................................................................................................................... 46 

Appendix i - Elemental analysis ........................................................................................................ 48 

Appendix ii – Solids, gases and solution phases included in system simulations ............................. 49 

Appendix iii – Numbering of phases ................................................................................................. 69 

 
  



 
 
 
 
 

1 
 

1. Introduction  
Global climate change has already had observable effects on the environment such as loss of sea ice, 

accelerated sea level rise and more intense heat waves (NASA, 2020). The global temperature is also 

expected to rise as an effect of greenhouse gases produced by human activity (NASA, 2020). At the 

same time, the world population is increasing and is producing increasingly more waste that must be 

managed (Kaz et al., 2018). With the world becoming more aware of the effects of climate change 

and the waste generation increasing; while the land available for landfilling decreases, it is of interest 

to focus on process effectivization and to recycle as much as possible from waste and move towards 

a more circular economy in which no new raw materials are needed. This also, since the planet’s 

resources are limited and spent materials must be reused. As more emphasis is put on recycling, there 

has been an increasing interest in finding new ways to recycle materials from  Municipal Solid Waste 

(MSW) incineration ashes, mainly with a focus on precious metals such as copper, zinc and vanadium 

as these are expensive and energy intensive to produce. Solid metal pieces are already separated, but 

metals bound in different forms in ash particles are not. Previous studies have investigated leaching 

of heavy metals such as Fe, Al, Mg, As, Ba, Cd, Co, Cu, Mn, Ni, Pb, Sr, V and Zn with different 

acids; not only for their monetary value but also to decrease the volume going to landfill, and since 

some metals are toxic (such as Cr, Cu, Ni and Pb) and would be beneficial to remove before 

landfilling (Tang, 2015), (Karlfeldt-Fedje, 2010), (Huang, Inoue, Harada, Kawakita, & Ohto, 2011), 

(Pöykiö, Mäkelä, Watkins, Nurmesniemi, & Dahl, 2016), (Hykš, 2008). Some studies have focused 

on Cu and Zn in particular as they often are the trace metals found in the highest concentrations in 

waste (Tang, 2015), (Karlfeldt-Fedje, 2010). Leaching has been found to be more efficient for fly 

ashes (volatile ash compounds) than the heavy bottom ashes, as the bottom ash is more heterogeneous 

and contains larger particles (Tang, 2015), (Karlfeldt-Fedje, 2010). 
 

1.1. Background  
The definition of waste according to the EU Waste Framework Directive is “any substance or object 

which the holder discards or intends or is required to discard” (Falkenberg, 2012). Waste occurs when 

any organism returns substances to the environment. Usually, these substances are recycled by other 

organisms (EnvironmentalLiteracyCouncil, 2015). Humans, however, produce a large flow of 

material residues (items consumers throw away) that would overload the capacity of natural recycling 

processes and therefore must be managed. Solid and fluid, hazardous and non-toxic wastes are 

generated in households, offices, schools, hospitals, and industries. Municipal Solid Waste (MSW) 

is waste collected from households, commercial buildings, hospitals and schools. MSW consists 

primarily of paper, containers and packaging, food wastes, yard trimmings, and other inorganic 

wastes (EnvironmentalLiteracyCouncil, 2015). In 2018, approximately 2.01 billion tons of waste was 

generated in the world. By 2050, this number is expected to increase drastically to 3.40 billion tons 

of waste annually as lower income countries develop and the world population increases (Kaz et al., 

2018).  

 

For managing waste, there is a waste hierarchy: a set of priorities for the efficient use of resources 

based on the Waste Avoidance and Resource Recovery Act 2001. The waste hierarchy is (as 

reproduced from EPA (2017): 

1. avoidance including action to reduce the amount of waste generated by households, industry 

and all levels of government 

2. resource recovery including re-use, recycling, reprocessing and energy recovery, consistent 

with the most efficient use of the recovered resources 

3. disposal including management of all disposal options in the most environmentally 

responsible manner. 

https://www.epa.nsw.gov.au/legislation/Actsummaries.htm#waarra


 
 
 
 
 

2 
 

MSW (as well as other wastes and biomass) can be combusted to produce heat and power. For 

example, in 2017 approximately 240 million tons of municipal solid waste was generated in the U.S., 

(EPA, 2020). Out of the 240 million tons, 12.7 % were combusted with energy recovery (EPA, 

2020). The incineration of waste, which falls under resource recovery in the waste hierarchy, is 

desirable over disposal as it (of course) recovers energy instead of directly sending it to landfill. In 

addition to this it also decreases the volume of the waste flows with the remaining ash being on 

average 10% of the volume of the input, and around 30% of its weight (Lam et al., 2010). Although 

the volume decreases significantly, the resulting ash (the residual material left after combustion), 

which must be disposed is still left behind. The ash can be seen as a new waste flow that can follow 

the waste hierarchy. The ash is usually landfilled or reused for secondary materials (Lam et al., 2010).  

 
 

The combustion/incineration of waste and biomass is usually done with a grate-fired or fluidized-

bed boiler. Fluidized beds offer advantages such as better contact between the fuel and air, rapid heat 

and mass transfer rates (created by the turbulence of the bed) and reduction of sulphur dioxide 

emissions (when using limestone). As a result of this, the combustion temperature can be lower; 

which in turn reduces the formation of nitrogen oxides (Dryden, 1982). In fluidized-bed combustion, 

an inert bed material of for example coal ash, silica sand and/or limestone 

is conventionally used (Dryden, 1982). For combustion of waste and biomass, silica sand is the 

usual choice of bed material (Rydén, Hanning, Corcoran, & Lind, 2016). There are two main 

categories of atmospheric fluidized bed combustors: bubbling fluidized-bed combustors (BFBC) and 

circulating fluidized-bed combustors (CFBC). The main difference between the two is that the 

fluidizing velocity is low in a bubbling bed which keeps most of the bed in the combustor (at a depth 

around 1 m); whereas the fluidizing velocity is significantly higher in a circulating bed. The higher 

velocity in the CFBC results in a higher uniform bed and some of the particles are entrained out of 

the combustor. They are then guided back to the bed through a cyclone. The heat and mass transfer 

rates are higher in a CFBC as a result of a higher slip velocity (difference between gas and solid 

velocity), longer residence and contact times and more intense mixing (Sarkar, 2015). Apart from the 

different fluidizing velocities and circulation, the different combustors operate similarly. As 

mentioned, they operate in the same temperature interval (Caneghem et al., 2012). Air is introduced 

below the bed in the combustor and goes through a perforated plate on which the fluidized bed rests. 

The fuel is introduced into the bed where it burns. Volatile compounds (fly ash) exit in the top of the 

boiler and is lead into the convective section where they are cooled and removed. Heavier compounds 

(bottom ash) in the bottom  (Miller, 2017), (Sarkar, 2015). A simplified schematic of an atmospheric 

fluidized-bed combustor can be seen in Figure 1 below.   

 
Figure 1. Simplified schematic of an atmospheric fluidized-bed combustor 



 
 
 
 
 

3 
 

In a fluidized bed, the air and thereby oxygen is dispersed by mixing as a result from turbulence of 

the bed. Insufficient mixing in the bed results in oxygen rich and oxygen poor zones, which can cause 

local temperature variations and emission of unburned species.  

 

Another technology for fluidized-bed combustion is called Oxygen Carrier Aided Combustion 

(OCAC). In this process, the conventional inert bed material in the fluidized combustion reactor 

is replaced fully or in part with oxygen carriers. In OCAC, the oxygen carrier bed material is reduced 

in fuel rich parts of the reactor and oxidized in oxygen rich parts during the combustion. This 

improves the distribution of oxygen in space and time, which is especially beneficial for managing 

heterogeneous fuels such as waste and biofuels. OCAC has been found to have positive effects on 

emissions compared to other methods, whilst also having the advantage of increasing combustion 

efficiency and capacity (Gyllén, 2019), (Ryden, Hanning, & Lind, 2018). Although the dispersion of 

oxygen is improved, the oxygen carrier itself is subjected to both oxidizing and reducing conditions; 

which corresponds to the partial pressure of oxygen.  

 

One oxygen carrier that is used in the OCAC process is ilmenite (FeTiO3), which is the oxygen carrier 

that will be studied in this thesis. It can be found in natural mineral ores. Although ilmenite cannot 

achieve as high of a conversion rate as some other oxygen carriers, such as for example Fe2O3, 

Al2O3 and MnO, it has the advantage of having a lower cost (Breault, 2018) and is the most studied 

oxygen carrier in the world. It is also the only one that has been applied in industrial scale. 

  

The intent behind the thesis is to investigate interactions between ilmenite (oxygen carrier bed 

material) and ash-forming components from the waste fuel in a fluidized-bed combustor. For 

simulating the system, a fluidized bed combustor, some conditions must be specified. There are 

different types of fluidized-bed combustors, but they all usually operate at a temperature between 

800-900 °C (Caneghem et al., 2012). There are vessels operating under pressurized and atmospheric 

conditions respectively, but for this study an atmospheric boiler has been chosen as it has the widest 

application globally (Sarkar, 2015). The operating pressure in the boiler is thereby 1 atm. Chemical 

looping combustion (CLC) is a technology in which two interconnected fluidized beds, a fuel reactor 

and air reactor are used. Metal oxides called oxygen carriers are circulated between the two, 

transporting oxygen, so that direct contact between fuel and air is avoided. This results in an outlet 

stream consisting of CO2 and H2O allowing for CO2 capture (Lyngfelt, 2013). The air reactor in CLC 

can be thought to correspond to the maximal oxidizing conditions (the highest partial pressure of 

oxygen) and the fuel reactor in CLC the most reducing conditions. Thereby the highest partial 

pressure of oxygen in the interval investigated is that of oxygen in air: 0.21 atm. A partial pressure of 

10-15 atm can be said to represent the most reducing conditions. 

 

1.2. Ash formation 
 

The interaction between the oxygen carrier and fuel ash can lead to an enrichment of certain elements 

such as zinc, copper, potassium, calcium and phosphorous among others in the oxygen carrier 

particles. As previously mentioned, there has been an increasing interest in recycling trace metals 

from ashes. As the bed material in fluidized bed boilers can contain a preconcentration of the metals 

found in waste, it is interesting to evaluate the possibilities to extract trace metals from it as well. 

Although research has shown that leaching is generally more effective for fly ash than bottom ash 

(Tang, 2015), (Karlfeldt-Fedje, 2010), spent ilmenite presents an opportunity as it can be 

magnetically separated from the rest of the bottom ashes and thereby present a more uniform fraction. 

For better understanding of how metals can be extracted from the ilmenite and how the different ash 

components interact with the bed material there are two different approaches: experiments with 

material characterization and theoretical simulations. In this thesis, theoretical simulations are 



 
 
 
 
 

4 
 

performed. Experiments were also initiated but not able to be finished. They can however be 

continued as future work.  

 

For investigating how ash forming components interact with ilmenite, the fuel chemistry must be 

understood in order to choose the components to simulate. When a fuel is burned, it can be said to 

undergo three different steps: drying, pyrolysis and char burning. After the particles have been heated 

up and dried, pyrolysis starts and organic volatile species are released and burn with a visible flame. 

In this step, some reactive ash-forming elements are released. In the next step, char burning, most of 

the ash-forming elements end up in the residual ash. The released ash-forming matter reacts with flue 

gas components and with each other (Zevenhoven, Yrjas, & Hupa, 2010). When investigating 

interactions with ilmenite in the boiler, both ash-forming matter and compounds formed from 

interactions between the ash-forming matter and flue gas components should be considered.  

 

In waste, the major elements are Si, Al, Fe, Mg, Ca, K, Na and Cl. The most common oxides found 

in ash from waste are SiO2, Al2O3, CaO, Fe2O3, Na2O, K2O and MgO (Lam et al., 2010). Silicon is 

present in the fuel and/or bed material as silica (SiO2) and does not react substantially under the 

combustion conditions but can have some interactions with the ash. Aluminum is present in waste in 

many forms, and in the furnace processes they all form alumina, Al2O3. The organic calcium 

compounds will be converted into calcium oxide (CaO) with CaCO3 as an intermediate middle step. 

All the forms of iron will yield iron oxides, and finally Fe2O3. Magnesium behaves similarly to 

calcium and forms MgO. Sodium and potassium mainly primarily react with Cl (Zevenhoven et al., 

2010), but can also form oxides as Na2O and K2O (Lam et al., 2010). For investigating interactions 

between ilmenite and major ash components, the components of interest are SiO2, Al2O3, CaO, 

CaCO3, Na2O, K2O and MgO. Fe2O3 is unlikely to have any significant interaction since the system 

is already iron rich and does therefore not need to be investigated. CaCO3 needs to be investigated as 

it is a possibility that it might significantly interact with ilmenite before forming CaO. Other 

carbonates such as Na2CO3, K2CO3 and MgCO3 are not investigated as Ca is expected to be in higher 

concentrations in MSW. K2CO3 which is considered as a main component in biomass, should under 

the conditions in the boiler mostly exist in the form of KOH and K2O. When investigating potassium 

and sodium, it is especially important to also look at the subsystems of potassium/sodium, ilmenite 

and Cl as reactions with this component might be favored over interactions with ilmenite. Sodium 

and potassium form NaCl and KCl.  

 

Waste also contains minor elements such as Cu and Zn, which are the main focus in this study as 

these trace metals usually have the highest concentration in waste and are valuable in recycling, as 

they are high-cost metals and for saving energy. Copper is used mainly for electrical purposes in 

power transmission and generation, wiring for buildings, telecommunication and 

electronics/electrical products (USGS, 2019). A lot of pigments also contain copper (Sward, 1972), 

but these compounds are unfortunately not available in the FactSage database for simulation. Zinc 

predominantly exists as ZnO as it is used as an additive in a range of products such as rubber and heat 

resistant glass among others (Jones, Bisaillon, Lindberg, & Hupa, 2013). Zinc as metal is primarily 

used for galvanizing steel (stainless steel), die casting machine parts, in batteries and other electrical 

applications. It is also used alloyed with copper to form brass (MadeHow, 2020).  

 

Ash from a fluidized bed boiler fired with municipal solid waste has been shown to contain copper 

metal, Cu2O, CuO and mixed oxides, such as CuCr2O4 (Lasseson & Steenari, 2013). Fly ash from 

municipal solid waste incineration has been shown to contain ZnS, ZnO, hydrozincite 

(Zn5(OH)6(CO3)2), gahnite (ZnAl2O4), and willemite (Zn2SiO4) (Struis, Ludwig, Lutz, & 

Scheidegger, 2004). Both zinc and copper also readily react with Cl to form ZnCl2 (Struis et al., 2004) 

(Jones et al., 2013) and CuCl (Lasseson & Steenari, 2013). From all compounds mentioned, the main 



 
 
 
 
 

5 
 

compounds of interest to investigate for interaction with ilmenite are ZnO, Cu2O and CuO since they 

are volatile ash-forming components. Secondary products ZnS, hydrozincite (Zn5(OH)6(CO3)2), 

gahnite (ZnAl2O4), willemite (Zn2SiO4), ZnCl2 and CuCl indicate that the ash-forming components 

interact with each other and flue gas components to form other stable zinc and copper compounds. 

The interactions between Zn-S, Cu/Zn-Cl, Zn-H2O-CO2, Zn-SiO2 and Zn-Al2O3 also have to be 

investigated to evaluate which interaction is favored. If one of these interactions is favored, it might 

significantly affect to what extent Zn and Cu interact with ilmenite. Cl can be considered to be in the 

forms NaCl, KCl and Cl2. 

 

1.3. Aim 
The aim of the thesis is to investigate how major ash components (Si, Al, Fe, Mg, Ca, K, Na and Cl) 

and the minor ash components Cu and Zn interact with ilmenite (FeTiO3) during waste incineration. 

A focus on Cu and Zn was chosen since these usually have the highest concentrations out of the trace 

metals that can be valuable for recycling in the waste fuel. 

 

The investigation was carried out by simulating different boiler conditions and performing 

thermodynamical equilibrium calculations using the software FactSage 7.2. The results from the 

calculations were compared to an elemental analysis of a sample obtained from an industrial OCAC 

plant. 
 

1.4. Limitations  
This thesis will not:  

• Investigate interactions of other minor ash elements than Cu and Zn  

• Investigate other bed materials or compare them to ilmenite 

• Verify the simulated results experimentally 

• Discuss the influence of process type on the accumulation of Cu and Zn.  

 

1.5. Specification of issue under investigation  
During the thesis, the following questions are to be answered:  

• How does ilmenite interact with major ash components (Si, Al, Fe, Mg, Ca, K, Na and Cl), 

what phases are formed? 

• What phases may be formed when Cu and Zn interacts with ilmenite during waste 

incineration? 

• How do the formed phases vary with different boiler conditions?  

• What reactions may the copper and zinc phases be formed by? 

• How may the ilmenite oxygen carrier properties be affected practically based on the 

theoretically calculated phases formed from interactions? 

 

 

 

 

 

 

 

 

 



 
 
 
 
 

6 
 

2. Methodology  
In this section, the method for how the project was executed is presented; which includes SEM-

EDX-analysis of a sample of spent ilmenite, how parameters were defined and how the simulation 

and calculations were carried out. A sample of ilmenite used during waste incineration was 

provided by the company E.ON from their CHP plant at Handelö. A short description of how the 

software and its modules work is also included.  

 

2.1. SEM-EDX-analysis of obtained sample 

A sample of ilmenite used during waste incineration was provided by the company E.ON from their 

CHP plant at Handelö. As mentioned in the introduction, experiments were also initiated in the 

thesis but not able to be finished. A SEM-EDX-analysis (Scanning Electron Microscope - Energy-

dispersive X-ray spectroscopy) of the ilmenite sample obtained from Handelö was however carried 

out. SEM-EDX-analysis with a Phenom ProX was used in an attempt to qualitatively map the trace 

elements and look at the distribution throughout the particles. Both samples prepared by 

immobilizing in epoxy and on carbon tape were analyzed as they give different types of 

information. The epoxy samples were grinded to a flat surface and coated with Au. The 

samples enclosed in epoxy gave information about the cross section of the particles, whereas 

the samples immobilized on carbon tape gave information about the particle surface. Both a 

mapping of the whole particles and point analysis in points of interest were performed. The settings 

were set to maximum obtainable HV and the intensity to “map”. Detector “BSD Full” was 

used. With the SEM analysis, images of the particles and the distribution of chemical 

elements were obtained.  For the mapping, the map resolution was set to 256, the pixel time to 

20 ms and number of passes to one.  

 

2.2. Simulation conditions 
This section describes how the simulation of the system and how the calculations were carried out. It 

also includes short descriptions of how the software and its modules work. The software FactSage 

7.2 (FactSage, 2020a) is used for the simulation and calculation of the system in this thesis. FactSage 

is an integrated database computing system in chemical thermodynamics. It can be used for many 

applications of which one is combustion. The program has access to data for thousands of compounds 

as well as to evaluated databases for solutions of metals, liquid and solid oxide solutions, mattes, 

molten and solid salt solutions, aqueous solutions, etc. From the databases available in the program, 

the most relevant ones for the combustion boiler system is FactPS (pure elements) and FToxid (oxide 

mixtures). FTsalt is also relevant when investigating subsystems with Cl. With FactSage it is possible 

to calculate conditions for multiphase, multicomponent equilibria (thermodynamical equilibrium 

calculations) and there are several tabular and graphical output modes (FactSage, 2018). The program 

contains different “modules” for different types of calculations, as for example FactSage phase 

diagram which is used in this thesis. It is however important to keep in mind that the results are 

heavily dependent on the chosen database and input (ratio between species) and that the calculations 

do not consider chemical limitations such as activation energies, diffusion limitations and reaction 

rates. 

 

2.2.1. FactSage Phase Diagram 

In the FactSage Phase diagram module unary, binary, ternary and multicomponent phase diagram 

sections can be calculated plotted and edited. The axes can be various combinations of T, P, V, 

composition, activity, chemical potential, etc. The results can be presented in a Y vs X or Gibbs 

triangle plot. The phase diagram calculations work in the same manner as in the Equilib module, but 

over intervals (FactSage, 2010). The FactSage Equilib module uses Gibbs free energy minimization 

to calculate the concentrations of chemical species when specified elements or compounds react to 



 
 
 
 
 

7 
 

reach a state of chemical equilibrium (FactSage, 2020b). A point in the phase diagram corresponds 

to an Equilib calculation. The diagram shows in which phases the compounds/products exist in the 

intervals that were specified before the calculations (FactSage, 2010). 

 

2.2.2. Inputs to the software 

The sample of ilmenite used during waste incineration (provided by the company E.ON from their 

CHP plant in Handelö) was used to estimate the fractions of the different ash components (by 

elemental analysis) for finding the region of interest in the phase diagrams. The elemental analysis 

can be found in Appendix i.  
 

The different component systems that were simulated in the present work (as motivated under 1.1. 

Background) are summarized in Table 1 below, in which the components, number of ideal gases, 

solids and species/solutions can be seen. For a full list of what compounds were included in the 

simulation of the different systems, see Appendix ii.  

 
Table 1. Summary of simulated component systems, including components, number of ideal gases, solids and species/solutions. 
Detailed information about all components included in the simulation can be found in Appendix ii. 

Component system # ideal gases # pure solids #solution phases: species/solutions 
FeTiO3 – O2 8 29 32/9 

FeTiO3 – SiO2 – O2 13 50 38/10 

FeTiO3 – Al2O3 – O2 15 40 43/9 

FeTiO3 – CaO – O2 11 43 39/11 

FeTiO3 – CaCO3 – O2 21 53 39/11 

FeTiO3 – Na2O – O2 21 52 37/9 

FeTiO3 – K2O – O2 11 39 34/9 

FeTiO3 – Na2O – Cl – O2 35 63 41/10 

FeTiO3 – K2O – Cl – O2 35 51 34/9 

FeTiO3 – MgO – O2 11 35 50/9 

FeTiO3 – ZnO – O2 9 32 44/10 

FeTiO3 – CuO/Cu2O – O2 11 39 34/9 

FeTiO3 – Zn – H2O – CO2 – O2 49 46 44/10 

FeTiO3 – Zn – SiO2 – O2 14 54 58/13 

FeTiO3 – Zn – Al2O3 – O2 16 43 80/12 

FeTiO3 – Zn – Cl – O2 32 39 56/13 

FeTiO3 – Zn – S – O2 24 56 54/10 

FeTiO3 – Cu – Cl – O2 34 47 45/12 

 

Firstly, a system of only ilmenite and oxygen was simulated to investigate the oxidation/reduction 

behavior of ilmenite and for using as a reference for the other phase diagrams. Then, interactions with 

the major ash components SiO2, Al2O3, CaO, CaCO3, Na2O, K2O and MgO were investigated with 

three component systems of ilmenite – major ash component – oxygen. For the major ash components 

Na2O and K2O four component systems including Cl were also investigated as sodium and potassium 

are known to readily react with Cl and to evaluate which interactions are favored.  

 

The minor ash components copper and zinc were also investigated by three component systems of 

ilmenite – ash component – oxygen as well as four component systems including Cl as copper and 

zinc have been known to form ZnCl2 and CuCl. CuO and Cu2O can be investigated by the same 

simulation as the systems both contain copper and oxygen. The software is then expected to determine 

in what areas the respective compounds are stable. In addition to investigating the favored reactions 

between Cu/Zn-Cl/ilmenite, four component systems of FeTiO3 – Zn –S – O2, FeTiO3 – Zn – H2O – 



 
 
 
 
 

8 
 

CO2 – O2, FeTiO3 – Zn –Al2O3– O2 and FeTiO3 – Zn – SiO2– O2 were investigated as zinc has also 

been known to form ZnS, hydrozincite (Zn5(OH)6(CO3)2), gahnite (ZnAl2O4) and willemite 

(Zn2SiO4). 

 

For the investigation of the oxidation/reduction behavior of ilmenite, the temperature range of 800 - 

900°C in which a fluidized bed boiler usually operates was chosen as the x-axis. This since the 

component fraction is known (=1) and to investigate how the behavior of the ilmenite varies with 

temperature. A temperature of 900°C was chosen for the other systems, as the simulation of the 

system ilmenite – oxygen showed that this is the temperature that captures the largest number of 

possible stable phases formed from the ilmenite.  

 

The three-component systems were simulated, and the phase diagrams constructed at 900°C, at 

atmospheric pressure with a logarithmic oxygen partial pressure (log10p(O2)) ranging from 0 to -20 

as the y-axis. An additional diagram with a temperature of 800°C was constructed for the system 

ilmenite – CaCO3 – oxygen as CaCO3 forms CaO and CO2 (gas) at 893°C. The ash component 

fraction between 0 and 1 was used for the x-axis, as the concentrations can vary significantly in the 

heterogeneous fuel. This, to get an overview of what happens as the concentration of the ash 

component increases. The fraction of ash component obtained from the elemental analysis of the 

sample obtained from Händelö (Appendix i) was marked on the images as an estimation of the area 

of conditions that the ilmenite particles are exposed to. If the phase diagrams were difficult to interpret 

in the marked area, an additional phase diagram was constructed within the intervals suggested by 

the sample content (estimation with included error margins) as the x-axis, thus enlarging the same 

phase diagram to the area of interest. 

 

Four component systems were constructed in a similar manner, but with the additional component 

having a constant fraction. 

 

For easier labeling and understanding of the resulting phase diagrams, the stable phases appearing in 

the diagrams were numbered according to the list in Appendix iii. All diagrams were marked with an 

estimation of the ash component fraction based on the elemental analysis of the sample obtained from 

Händelö, seen in Appendix i. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 
 
 
 
 

9 
 

   3. Results & Discussion 
The following sections present and interpret the phase diagrams resulting from the simulations. In the 

thesis “oxygen carrier performance” is discussed based on the stable redox-phases of ilmenite, 

referring to the oxygen release between different phases and conditions. It is important to note that 

the diagrams can be interpreted differently if different purposes are defined from the beginning. The 

results obtained from the simulations are applicable for OCAC, and analyzed from that perspective, 

but could be used for understanding interactions in the CLC systems, where the highest oxygen partial 

pressure corresponds to the air reactor and the lowest to the fuel reactor.  

 

The y-axis displaying the logarithmic oxygen partial pressure (in atm) can be interpreted as locations 

in the boiler with different concentrations of oxygen (depending on the fuel density and burning) but 

also as different locations within the particle (different depths of the formed layers). At the surface 

of the particle, the oxygen access is high, but moving inward in the particle, the conditions become 

increasingly reducing with less free oxygen present. An oxygen partial pressure of -20 is highly 

reducing, and very unlikely to occur for the circulating ilmenite particles as the bed is fluidized and 

the particles are in constant motion. It is however still of interest to investigate this pressure, as it can 

be said to represent the internal parts of the particles. The molar fraction of ash component on the x-

axis can be interpreted as the ash component concentration surrounding the ilmenite particles in the 

boiler, but also as a concentration increasing due to accumulation of ash compounds on the particle 

surface. Lines in the plots that are believed to be representative for the conditions that the ilmenite 

particles are subjected to in an industrial application, will be marked in the figures based on ash 

component concentrations determined for sample obtained from Händelö by elemental analysis. 

 

Once again it is important to keep in mind when analyzing the results, that the software does not take 

reaction rates or diffusion limitations into account. It is therefore possible for compounds that are not 

stable over the entire diagram to still be found in the sampled as bottom ash ilmenite fraction, due to 

the compound having a low reaction rate or getting encapsulated in a layer built up on the ilmenite 

particle surface and thereby not being able to react with other compounds in the boiler. The formation 

of a layer could be observed with SEM-analysis of the Händelö sample. Unexpected phases not 

showing up in the phase diagrams may also be found experimentally on the ilmenite due to missing 

data for those components in the software database. This was for example found for the compound 

KTi8O16, that has been the main potassium phase found experimentally on used ilmenite particles. 

This compound is missing from the software database and other phases were therefore predicted as 

stable by the software. During combustion, iron has also been found to migrate to the surface of the 

ilmenite particles (Corcoran, Knutsson, Lind, & Thunman, 2018b). Although the stable phases shown 

in the phase diagram contain titanium, interactions that form iron containing phases may be favored 

instead if only iron is available for reaction.  

 

 

 

 

 

 

 

 

 

 



 
 
 
 
 

10 
 

3.1. SEM-EDX-analysis of obtained sample 
With the SEM-EDX, an ash-layer on the ilmenite particles’ surface could be observed, see an example 

in Figure 2 below. The EDX-analysis indicated that the layer consisted mainly of calcium, as 

previously described by Corcoran, Knutsson, Lind, and Thunman (2018a). 

 

 
Figure 2. SEM and EDX image of an ilmenite particle with a visible ash-layer formed on the surface 

The layer formation on the particles means that they are not uniform and that there are concentration 

gradients. This must be kept in mind when analyzing the simulated system, as the software does not 

take layer formation or gradients into consideration; rather it considers a “perfectly mixed” system.  

 

The SEM-EDX-analysis was mainly intended to map the Cu and Zn trace elements (along with the 

other elements) and to observe their distribution within the ilmenite particles. Upon analysis of the 

samples it was however discovered that the trace element concentration was too low to be reliably 

detected/mapped. If the software of the microscope did not identify the copper or zinc by default and 

a mapping was requested, the obtained information was misleading, where according to the obtained 

results copper and zinc were present throughout the particles and all over the image. An example of 

an intensity map over a cross-section of particles can be seen in Figure 3 below.  

  

 
Figure 3. A SEM-EDX  mapping of Zn and Cu showing the distribution of Cu and Zn throughout the particles which was considered  
incorrect  



 
 
 
 
 

11 
 

As can be seen in the figure, Zn and Cu appear all over the image, including in the epoxy background. 

As this is considered as impossible, the results were not further used. As the mapping time was only 

approximately an hour and a half, this background noise could possibly have been decreased (and the 

image become more reliable) if the time was increased. This would result in a higher picture 

resolution.  

 

In addition to the mapping, point analysis was used to follow the concentration of Cu and Zn in 

various parts of the particles. According to the point analysis no Zn nor Cu were detected within the 

particles. A small unidentified peak could however be seen for one point-analysis in one of the EDX-

spectra, which likely is a copper or zinc signal. This point was located toward the outer surface of the 

ilmenite particle. Looking at Figure 3, it appears that there is a brighter green layer (higher 

concentration) toward the particle surfaces. This observation together with the unidentified peak may 

indicate that there is an enrichment of zinc in the outer layer of the particles. 

 

For another sample, an aggregation of Cu could however be detected, see Figure 4, confirming that 

Cu can accumulate on the ilmenite particles.   

  

 
Figure 4. SEM-EDX mapping displaying an aggregation of Cu, seen as a bright red spot indicated by the red arrow 

The Cu aggregation (rightmost image in Figure 4) could be seen as a bright red spot in the mapping, 

as indicated by the red arrow. It appears that the mapping works better when there is a more intense 

reference point. Although this SEM-EDX worked for detecting larger aggregations of Cu, it did not 

work for the purpose intended; mapping the trace elements throughout the particles. It appears that 

the concentration of the elements is too low to be able to separate them from background noise. The 

results may however have been improved if an increased mapping time was used. 

 

 

 

 

 

 

 

 

 

 

 

 

 



 
 
 
 
 

12 
 

3.2. Oxidation/reduction behavior of ilmenite 
The resulting phase diagram for the system ilmenite – oxygen can be seen in Figure 5 below.  

 

 
Figure 5. Phase diagram for the system ilmenite-oxygen over the boiler temperature interval, for investigating oxidation/reduction 
behavior of ilmenite. The M2O3 (corundum) structure is Fe2O3, the spinel Fe3O4 and the titania spinel Fe2TiO4, found as the main 
stable phases by FactSage. 

The partial pressure intervals in which the phases are stable vary with approximately log10p(O) = 1 

atm over the entire temperature interval. At a log10p(O2) of under -18, there is a variation in which 

phases are stable. Ilmenite is stable between a log10p(O2) of approximately -19 to -12. Note that a 

log10p(O2) of 0 corresponds to 1 atm which is the conditions in the boiler at which the ilmenite is in 

contact with the maximum amount of oxygen. Starting by looking at the interval where ilmenite 

(FeTiO3) and the spinel phase are stable, it can be seen when through equilibrium calculations that 

the stable spinel phase is principally negligible. More than 99 % of the formed stable phase is ilmenite. 

Looking at what happens when the oxygen partial pressure changes, the ilmenite becomes unstable 

when the oxygen pressure increases. Based on the diagram it should then instead form rutile (TiO2) 

and spinel (together with oxygen). The spinel structure has the general form AB2O4 where A, B= Al-

Co-Cr-Fe-Mg-Mn-Ni-Zn. As the system only contains iron out of these compounds, the spinel 

structure in this case is Fe3O4, which is also called magnetite. A general reaction for this formation is 

presented in Equation (1):  

 

6 FeTiO3 + 19 O2 → 2 Fe3O4 + 6  TiO2     (1) 

 

As can be seen in the equation, through this reaction the ilmenite is oxidized. Upon further increase 

of the oxygen partial pressure, it can be seen in the diagram that the spinel phase becomes unstable 



 
 
 
 
 

13 
 

and transforms into M2O3 (corundum), which in this case is Fe2O3, also called hematite. The rutile 

phase remains stable. A general reaction for this can be described by Equation (2): 

 

4 Fe3O4 +  17 O2 → 6 Fe2O3      (2) 

 

As can be seen in Equation (2), magnetite is further oxidized. Going back to the interval in which 

ilmenite “and spinel” are stable, it can be seen that the spinel structure (Fe3O4) becomes unstable and 

instead forms titania spinel, which in this case is Fe2TiO4. The amount of titania spinel is very small 

compared to the amount of ilmenite, with there being roughly 355 times more stable ilmenite (in 

moles). For this titania spinel to be formed from magnetite, it must react with some ilmenite according 

to the general Equation (3) below:  

 

2 Fe3O4 +  6 FeTiO3 → 6 Fe2TiO4 + 25 O2    (3) 

 

As can be seen in Equation (3), oxygen is released when magnetite reacts with ilmenite and therewith 

the compounds are reduced. When looking at a further reduction in low oxygen pressure, the resulting 

phases vary with temperature. At temperatures below 815°C, there is no further reduction of the 

compound (further release of oxygen) but at temperatures between approximately 815 – 840°C there 

is one additional reduction. At temperatures above 840°C there are two further reductions, which are 

described by Equation (4) and (5) below. Looking at the first reduction, the titania spinel becomes 

unstable and forms Fe and ilmenite:  

 

2 Fe2TiO4 → 2 FeTiO3 + 2 Fe +  O2     (4) 

 

At lower oxygen partial pressures, ilmenite decomposes to Fe and rutile (TiO2): 

 

2 FeTiO3 → 2 TiO2 + 2 Fe +  O2      (5) 

 

From these results it can be concluded that there are two higher levels of oxidation of ilmenite, and 

three lower levels / levels of reduction. Looking at the phase diagram, oxygen partial pressures above 

approximately -12 can be said to result in oxidizing conditions, and partial pressures below -14 in 

reducing conditions. Ilmenite is oxidized to magnetite, which is further oxidized to hematite. Ilmenite 

is reduced by reaction with magnetite to form titania spinel, TiFe2O4, which is further reduced by 

decomposition to ilmenite and Fe. At the lowest pressures for the present study, ilmenite is reduced 

by decomposition to rutile and Fe. This is however not desirable practically because there is a larger 

energy barrier to oxidize Fe back to Fe3O4 than when oxidizing Fe3O4 to Fe2O3. Furthermore, 

defluidization of ilmenite has been linked to highly reduced oxygen carriers (Leion, Lyngfelt, 

Johansson, Jerndala, & Mattisson, 2008). In addition to this, an oxygen partial pressure of -20 is 

highly reducing, and very unlikely to occur for the ilmenite particles as the bed is fluidized and the 

particles are in constant motion. The highly reducing conditions can however be seen as 

representative for the inner intact parts inside of the ilmenite particles, depending on how the oxygen 

migrates in the particles (which is dependent on for example cracks in and on the particle). 

 
It is also important to note, as earlier mentioned, that the software does not consider any chemical 

limitations such as diffusion limitations and reaction rates. In full scale application, the different 

compounds in the phase diagram can therefore exist simultaneously at the different pressures if 

limited by diffusion or reaction rates to form the predicted stable phases. The software assumes that 

every molecule is available for reaction (not limited by diffusion). Practically, the whole bed particle 

volume would not be available for reacting with oxygen, as the particle is only in contact with the 

surrounding oxygen at the surface. 



 
 
 
 
 

14 
 

 

Leion et al. (2008) investigated the reactivity of ilmenite towards methane and syngas experimentally 

in a laboratory setup, simulating a CLC process at temperatures in the range of 970 – 980°C, 

alternating between reducing and oxidizing conditions. They found through X-ray diffraction that the 

phases indicated for particles after oxidation were Fe2TiO5, Fe3Ti3O10, Fe2O3 and possibly some 

FeTiO3. For particles after subsequent reduction the indicated phases were FeTiO3, Fe3O4, some TiO2 

and FeO and possibly some Fe3Ti3O10. This is somewhat consistent with the results in Figure 5. The 

phases shown in the figure are stable, but kinetics play a role practically, making different reactions 

between the oxidation levels possible. This results in phases not predicted by the thermodynamical 

equilibrium calculations being able to form. The experimental findings also show that the maximal 

reducing atmosphere (as in a CLC reactor) is not enough to initiate the reaction in Equation (5). In 

addition to Equation (1) – (5), Equation (6) – (14) below are also possible, as reproduced from Leion 

et al. (2008). Level  0 corresponds to the oxidation state 2+ for iron (FeII) and level 2 has the oxidation 

state 3+ for iron (FeIII),whereas level 1 is a mixture of 2+ and 3+, Fe2
IIIFeII.  

 

• Oxygen release from level 1 to 0: 

2 Fe3Ti3O10 →  6 FeTiO3 +  O2     (6) 

2 Fe3Ti3O10 →  6 FeO +  6 TiO2  + O2    (7) 

2 Fe3O4 →  6 FeO +  O2       (8) 

 

• Oxygen release from level 2 to 1: 

6 Fe2TiO5  + 6 TiO2 →   4 Fe3O4  + 12 TiO2  + O2        (9) 

6 Fe2TiO5  + 6 TiO2 →   4 Fe3Ti3O10  + O2             (10) 

6 Fe2O3  + 12 Ti →  4 Fe3Ti3O10  + O2             (11) 

 

• Oxygen release from level 2 directly to 0: 

 2 Fe2TiO5  + 2 TiO2 →  4FeTiO3  + O2    (12) 

 2 Fe2TiO5  + 2 TiO2 →  4 FeO +  4 TiO2  + O2   (13) 

 2 Fe2O3 →   4 FeO +  O2      (14) 

 

The additional reactions should not affect the simulated results significantly with regards to ilmenite 

– ash component interactions, as the additional compounds (Fe2TiO5 and Fe3Ti3O10) are very similar 

to the stable phases considered from the simulation in Figure 5. 

 

 

 

 



 
 
 
 
 

15 
 

3.3. Interactions between ilmenite and ash-components 

3.3.1. Interactions with major ash component SiO2 

The resulting phase diagram for the system ilmenite – SiO2 (silica sand) – oxygen can be seen in 

Figure 6 below. SiO2 is not only an ash component but can be mixed with the ilmenite as bed material.

 
Figure 6. Phase diagram for system ilmenite – SiO2 – O2 for investigating interactions between SiO2 and ilmenite. The M2O3 
(corundum) structure is Fe2O3, the spinel Fe3O4 and the titania spinel Fe2TiO4. 

When comparing the phase diagram (Figure 6) to Figure 5, ilmenite seems to display exactly the same 

behavior, apart from spinel not showing up as a stable compound in the oxygen partial pressure 

interval between -12 – -10. A possible explanation for this is that the number of moles of spinel is 

low and thus is neglected in the calculations. The titania spinel formed from the spinel in the adjacent 

lower partial pressure interval is however included (although it is formed from the spinel phase), as 

3 moles of titania spinel form from 1 mol of spinel according to Equation (3) leading to a number of 

moles sufficiently high to be included in the calculations and show up on the phase diagram. It does 

not seem probable for ilmenite (FeTiO3) to decompose to titania spinel (TiFe2O4) as additional iron is 

needed to form titania spinel.  

 

Looking at the behavior of SiO2 (silica sand), it is stable over the entire interval and no new 

compounds appear. It is therefore not expected to interact with ilmenite under the specified boiler 

conditions.   

 



 
 
 
 
 

16 
 

3.3.2. Interactions with major ash component Al2O3  
The resulting phase diagram for the system ilmenite – Al2O3 – oxygen can be seen in Figure 7 below. 

The green line indicates the aluminum fraction (approximately 2.4 %) found in the sample obtained 

from Händelö. 

 

 
Figure 7. Phase diagram for system ilmenite – Al2O3 – O2 used for investigating interactions between Al2O3 and ilmenite. The number 
2 denotes rutile, and 6 - M2O3 (Corundum) where M = Al, Fe. The green line indicates the Al fraction (approximately 0.024) found in 
the sample obtained from Händelö. Note that the fraction based on the elemental analysis was divided by two as the compound 
(Al2O3) contains two Al atoms. 

As the phase diagram in Figure 7 is more complex, it can immediately be seen that Al2O3 is not inert 

in its behavior. M2O3 (corundum) in the diagram denotes both Fe2O3 and Al2O3. After performing 

equilibrium calculations in various points in the diagram, the main spinel structures were identified 

as Al3O4 and Fe3O4. The main titania spinel structures are FeAl2O4, and FeTi2O4. This means that 

there are interactions between ilmenite and aluminum oxide forming the titania spinel structure 

FeAl2O4. The FeAl2O4 forms under the same conditions as FeTi2O4, which is stable under reducing 

conditions. Looking at the phase diagram, it can be seen that Al2O3 behaves very similar to ilmenite, 

forming similar structures. At oxidizing conditions, the FeAl2O4 is again oxidized to Al3O4 and/or 

Al2O3, which suggests that aluminum should not be stable and is not expected to accumulate on the 

ilmenite particle surface over time. However, some aluminum may accumulate in the system as the 

titania spinel FeAl2O4 is stable under reducing conditions and can be blocked from oxygen contact 

through layer formation on the particle. Besides the interactions forming FeAl2O4, the 

oxidation/reduction behavior of the ilmenite seems practically unchanged. The expected oxygen 



 
 
 
 
 

17 
 

carrier performance should also be unchanged as FeAl2O4 contains the same amount of oxygen as 

FeTi2O4 and therefore behaves similarly over the same pressure interval.  

 

As it is hard to distinguish the phases in the region around the aluminum fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 8 below. 

 

 
Figure 8. Phase diagram for system ilmenite – Al2O3 – O2 within the intervals suggested by the sample content (estimation with error 

margins) as the x-axis for investigating interactions between Al2O3 and ilmenite. 

In the figure, the oxygen partial pressure intervals can be seen for the region of most interest based 

on the sample obtained from Händelö. However, the same compounds are formed as earlier 

mentioned: Fe2O3, Al2O3, Al3O4, Fe3O4, FeAl2O4 and FeTi2O4 with an addition of the compound 

(FeO)2(TiO2) in a small interval.  
 
 
 
 
 



 
 
 
 
 

18 
 

3.3.3. Interactions with major ash component CaO  

The resulting phase diagram for the system ilmenite – CaO – oxygen can be seen in Figure 9 below. 

The green lines indicate the calcium fraction interval (approximately 14 – 20 %) found in the sample 

obtained from Händelö. Some problems occurred with the FactSage software when trying to construct 

this curve. Not all tie-lines were drawn on the diagram and the phases displayed were incorrect. As 

the software indicated that the specified conditions did not result in any stable gas phases, the problem 

could be solved by simulating the system without ideal gases. This resulted in a complete phase 

diagram.  

 

 
Figure 9. Phase diagram for system ilmenite – CaO – O2 used for investigating interactions between CaO and ilmenite. The green 
lines indicate the Ca fraction interval (approximately 0.14- 0.20) found in the sample obtained from Händelö. 

 

In the phase diagram it can be seen that calcium interacts with the titanium in the ilmenite and forms 

the various structures CaTiO3, Ca3Ti2O7, Ca5Ti4O13,Ca3Ti2O6, Ca2Fe2O5, CaFe2O4 in different parts 

of the diagram with CaTiO3 being the most predominant one. At lower concentrations calcium 

interacts with the titanium in the ilmenite, and only at concentrations above roughly 50 % it reacts 

with the iron. One or more of these calcium compounds are stable through all pressures and 

concentrations. This indicates that calcium should interact readily with ilmenite when it is present in 

the boiler and accumulate on the particles over time, mainly in the form of CaTiO3. The results from 

the modelling correspond well to experimental findings, as an outer layer consisting of mainly 

calcium and phosphorus has been shown to form on the ilmenite particles during combustion (also 

confirmed by the SEM-EDX analysis performed on the sample obtained from Händelö) and the main 

phase found with X-ray diffraction is CaTiO3 (Gyllén, 2019). Calcium has also been shown  to 



 
 
 
 
 

19 
 

sometimes increase the reactivity/catalytic ability of ilmenite. This increase in catalytic ability can be 

explained by the calcium rich layer promoting the water-gas shift reaction and reducing tars in the 

produced gas (Gyllén, 2019). 

 

The diagram also shows a change in the oxidation/reduction behavior of ilmenite. When calcium is 

present in high concentrations, above 50 %, (at the particle surface) the pressure intervals in which 

spinel is stable are shifted, and the ilmenite and titania spinel structures become unstable. At very 

high concentrations, the spinel structure also becomes unstable. This indicates that the oxygen carrier 

performance may be affected if a too thick layer of calcium builds up on the ilmenite particles. 

Ca2Fe2O5, which is instead stable, does itself work as an oxygen carrier but with only two levels of 

oxidation. Highly reducing conditions are required to release oxygen. It is hard to determine how this 

would impact the oxygen carrier practically. The elemental analysis however indicated that the 

calcium content in the sample obtained from Händelö of spent ilmenite was approximately 14 – 20 

%. Within this interval calcium is only stable as CaTiO3. 

 

3.3.4. Interactions with major ash component CaCO3 

The resulting phase diagram for the system ilmenite – CaCO3 – oxygen can be seen in Figure 11. The 

green lines indicate the calcium fraction interval (approximately 14 – 20 %) found in the sample 

obtained from Händelö. Some problems occurred with the software when trying to construct the 

curve. Not all tie-lines were drawn on the diagram and the phases displayed were incorrect. As CaCO3 

forms CaO and CO2 (gas) at 893°C, as determined from Figure 10 below, including ideal gases in the 

simulation is essential and the problem could not be solved in the same way as for the CaO curve in 

Figure 9 (by excluding ideal gases).  

 
Figure 10. Phase diagram displaying the temperature at which CaCO3 decomposes to CaO and CO2, as indicated by the green 
dashed line. Gas-ideal denotes CO2. 



 
 
 
 
 

20 
 

As CaO is formed from CaCO3 at 893°C and the curve is constructed at 900°C, the curve should be 

almost identical to the CaO curve. The temperature used for the construction of the phase diagram 

(900°C) being so close to the reaction temperature of for CaCO3 to form CaO and CO2 may be related 

to the software having problems constructing the correct curve. The problem was solved, and the 

correct curve produced, by superimposing the CaO diagram on the “unfinished” CaCO3 diagram (a 

function in the software). The finished diagram for a temperature of 900°C can be seen in Figure 11. 

 

 
Figure 11. Phase diagram for system ilmenite – CaCO3 – O2 used for investigating interactions between CaCO3 and ilmenite at a 
temperature above 893 °C. The green lines indicate the Ca fraction interval (approximately 0.14- 0.20) found in the sample obtained 
from Händelö. 

As can be seen in Figure 11, the phase diagram for CaCO3 at temperatures above 893°C looks fairly 

similar to the one for CaO. The ilmenite behaves in a similar way with regards to oxidation/reduction 

behavior and oxygen carrier performance, compared with the behavior illustrated by Figure 9. In the 

fraction interval based on the sample obtained from Händelö the calcium is stable as CaTiO3 over the 

entire pressure interval. The carbon present in the system only interacts with the ilmenite and forms 

Fe3C under very reducing conditions, below an oxygen pressure of -19, corresponding to the inside 

of the particle. This means that the carbon would have to diffuse inward to be able to react with the 

ilmenite. This indicates carbon is unlikely to significantly interact with ilmenite. Apart from this small 

interval carbon is otherwise stable as CO2.  

 

As an OCAC-boiler usually operates closer to 850 °C, at which the CaCO3 is stable as a solid, it is in 

this case insufficient to only look at the phase diagram constructed at 900 °C for evaluating the 

interactions with ilmenite. Therefore, an ilmenite – CaCO3 – oxygen phase diagram was constructed 

at the lower temperature of 800 °C for investigating how the CaCO3 interacts with the ilmenite below 

the temperature of 893 °C. Again, the software had problems constructing the curve. This was solved 



 
 
 
 
 

21 
 

by excluding the gas phase in the simulation. This phase diagram can be seen in Figure 12 below. 

The solid carbon in the graph would form CO2. 

 

 
Figure 12. Phase diagram for system ilmenite – CaCO3 – O2 used for investigating interactions between CaCO3 and ilmenite at a 
temperature below 893 °C. The green lines indicate the Ca fraction interval (approximately 0.14- 0.20) found in the sample obtained 
from Händelö. C in this graph corresponds to CO2. 

As can be seen in the phase diagram in Figure 12, a temperature below 893 °C results in stability of 

CaCO3 in the oxidizing region and only interacting with ilmenite under reducing conditions, forming 

CaTiO3, Fe3C and a Ca3Ti2O7-Ca3Ti2O6 solid solution (at very high concentrations). 

 

 

 

 

 

 

 

 

 

 



 
 
 
 
 

22 
 

3.3.5. Interactions with major ash component Na2O (and Na2O/Cl)  

The resulting phase diagram for the system ilmenite – Na2O – oxygen can be seen in Figure 11 below. 

The green line indicates the sodium fraction (approximately 1.1 %) found in the sample obtained 

from Händelö. 

 
Figure 13. Phase diagram for system ilmenite – Na2O – O2 used for investigating interactions between Na2O and ilmenite. The green 
line indicates the Na fraction (approximately 0.011) found in the sample obtained from Händelö. Note that the fraction based on the 
elemental analysis was divided by two as the compound (Na2O) contains two Na atoms. 

In the phase diagram for the system ilmenite – Na2O – oxygen (Figure 13) it can be seen that there 

are many compounds formed as a result of interactions between ilmenite and Na2O. These are: 

(Na2O)(TiO2)3, (Na2O)(TiO2)6, Na2TiO3, Na4TiO4, Na8Fe2O7, Na8Ti5O14 and NaFeO2. The software 

indicated that Na3Fe5O9 also is a stable phase that can be formed, but the area in which it is stable 

was too narrow to mark on the plot. At low and intermediate concentrations (Na2O)(TiO2)3, 

(Na2O)(TiO2)6 and Na8Ti5O14 are the major phases containing sodium and at higher concentrations 

the stable phases from ilmenite interaction are mainly Na8Fe2O7 and NaFeO2. At lower concentrations 

the sodium seems to mainly interact with the titanium in the ilmenite, and at higher concentrations 

the iron. This means that the sodium will mainly interact with the titanium in ilmenite, and only react 

with the iron if it is already is present at the surface in very high concentration. There are however 

stable sodium compounds containing iron or titanium in practically all areas in the phase diagram 

(through the entire condition intervals specified in the study). This indicates that sodium will 

accumulate on the ilmenite particles over time. 

 

Already at low concentrations the sodium starts affecting the oxygen carrier behavior and 

performance of the ilmenite. Rutile, TiO2 becomes unstable under the studied oxidizing conditions 

and the intervals in which the spinel and titania spinel are stable shift. With increasing concentration 

of Na2O, slag-liquid phase stable over the whole oxygen partial pressure interval starts forming and 

ilmenite reacts with sodium and forms other stable compounds. This means that accumulation of 

sodium should affect the performance of the oxygen carrier.  



 
 
 
 
 

23 
 

 

When adding Cl to the system and constructing a phase diagram (not shown) it could be seen that Cl 

was stable as NaCl over the entire diagram. This indicates that the Na – Cl interactions are favored 

over Na – ilmenite when Cl is present in the system. As the concentration of sodium is low, Cl heavily 

impacts what phases are formed by the interactions between sodium and ilmenite.  

 

As it is hard to distinguish the phases in the region around the sodium fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 14 below. 

 

 

 
Figure 14. Phase diagram for system ilmenite – Na2O – O2 within the intervals suggested by the sample content (estimation with 
error margins) as the x-axis for investigating interactions between Na2O and ilmenite. 

As can be seen in Figure 14, sodium forms (Na2O)(TiO2)3 and (Na2O)(TiO2)6 in the region based on 

the sample obtained from Händelö. If the sodium concentrations are similar in other units during 

operation, sodium is expected to mainly accumulate as the enumerated phases.  

 

 



 
 
 
 
 

24 
 

3.3.6. Interactions with major ash component K2O (and K2O/Cl) 

The resulting phase diagram for the system ilmenite – K2O – oxygen can be seen in Figure 15 below. 

The green line indicates the potassium fraction (approximately 0.4 %) found in the sample obtained 

from Händelö. 

 

 
Figure 15. Phase diagram for system ilmenite – K2O – O2 used for investigating interactions between K2O and ilmenite. The green 
line indicates the K fraction (approximately 0.004) found in the sample obtained from Händelö. Note that the fraction based on the 
elemental analysis was divided by two as the compound contains two K atoms. 

The phase diagram shows that K2O forms compounds from interactions with ilmenite during almost 

all the studied oxygen partial pressures and concentrations. According to the phase diagram, K2O is 

not stable under the studied conditions. This means that K2O interacts readily with ilmenite when 

present. The main stable compounds formed are K2Ti3O7 and K2Ti6O13. The software also indicated 

that K8Ti5O14 was stable in part of the diagram, but the area was too narrow to mark on the plot.  

 

Already at low concentrations the potassium starts affecting the oxygen carrier behavior and 

performance of the ilmenite, in a similar way to sodium (both are alkali metals). Rutile, TiO2 becomes 

unstable under all specified oxidizing conditions and the intervals in which the spinel and titania 

spinel are stable shift. At concentrations above roughly 30%, slag-liquid phase stable over the whole 

pressure interval starts forming. This indicates that the oxygen carrier performance of ilmenite is 

significantly affected if potassium accumulates on the surface of the particles and if it is present in 

significant concentrations in the combusted waste. 

 



 
 
 
 
 

25 
 

When adding Cl to the system (not shown), potassium behaves like sodium. Cl is stable as KCl over 

the entire diagram. This indicates that the K – Cl interactions are favored over K – ilmenite 

interactions as all existing Cl in the system binds with potassium.  

 

Experimentally, potassium has been found as KTi8O16 in ilmenite particles rather than as K2Ti3O7 

and K2Ti6O13. Corcoran et al. (2018b) who received the same results proposed that this is due to the 

ilmenite particles not being fully saturated with potassium and thereby KTi8O16 is formed as an 

intermediate. During the simulations in this project it was however discovered that data for KTi8O16 

is missing in the software. This is most likely the explanation for the inconsistent results between the 

simulations and the previous experimental findings. This highlights that although the software using 

thermodynamic calculations may give a good estimation of formed phases, physical and chemical 

limitations or compounds missing in the software databases can result in predictions of stable phases 

differing from the stable phases found in practice. 

 

As it is hard to distinguish the phases in the region around the potassium fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 16 below. 

 

 
Figure 16. Phase diagram for system ilmenite – K2O – O2 within the intervals suggested by the sample content (estimation with 
error margins) as the x-axis for investigating interactions between K2O and ilmenite. 

As can be seen in the phase diagram (Figure 16), potassium forms K2Ti3O7 and K2Ti6O13 in the region 

indicated by the sample obtained from Händelö.  



 
 
 
 
 

26 
 

3.3.7. Interactions with major ash component MgO  

The resulting phase diagram for the system ilmenite – MgO – oxygen can be seen in Figure 17 below. 

The green line indicates the magnesium fraction (approximately 3 %) found in the sample obtained 

from Händelö. 

 
Figure 17. Phase diagram for system ilmenite – MgO – O2 used for investigating interactions between MgO and ilmenite. The green 
line indicates the Mg fraction (approximately 0.03) found in the sample obtained from Händelö. 

In the phase diagram, spinel and titania spinel denotes both the previously mentioned phases 

consisting of iron and titanium as well as phases formed mainly between iron and magnesium; which 

means that the magnesium interacts with the ilmenite. The major spinel phase containing magnesium 

and the major titania spinel both are FeMg2O4 (different structures). The software also indicated that 

MgTiO3 was a stable phase in part of the phase diagram, but this area was too narrow to mark on the 

image. 

 

At low concentrations of MgO, the structures formed in the different oxygen partial pressure intervals 

are similar to those formed when the system consist only of ilmenite and oxygen (see Figure 5) with 

an exception of pseudobrookite formed under highly oxidizing and reducing conditions within the 

studied interval. At the highly oxidizing conditions, the formed pseudobrookite consist mainly of 

MgTiO5 and under the highly reducing conditions it is a solid solution of MgTi2O5, FeTi2O5 and 

Ti3O5 with the compounds enumerated according to the expected concentrations in descending order. 

As the compounds under the oxidizing conditions behave similarly to ilmenite with regards to uptake 

of oxygen, the oxidation of the ilmenite compounds should not be significantly affected by the 

interaction with magnesium. However, a significant amount of pseudobrookite is formed under the 



 
 
 
 
 

27 
 

highly reducing conditions. The ratio between rutile and pseudobrookite is roughly 1:1 and the ratio 

of rutile and solid iron 1:3. The formation of pseudobrookite from titania spinel takes up oxygen 

rather than releasing it, thus possibly affecting the oxygen carrier performance under the highly 

reducing conditions. As these reducing conditions however correspond to the inner parts of the 

ilmenite particles, the magnesium would have to diffuse inward in the particles for this to occur which 

is unlikely to happen. 

 

At higher concentrations of MgO, Fe2O3 and rutile become unstable and the ilmenite only displays 

reducing behavior, with spinel and titania spinel forming in different oxygen partial pressure 

intervals. This indicates that the oxygen carrier performance should be heavily affected as the ilmenite 

does not pick up and release as much oxygen as if the MgO was not present as a phase. As MgO is 

in excess, it becomes a stable phase (monoxide). One or more phases formed from interactions 

between magnesium and ilmenite are stable in all areas of the phase diagram, which means that 

magnesium should accumulate on the particles with time.  

 

As it is hard to distinguish the phases in the region around the magnesium fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 18 below. 
 

 
Figure 18. Phase diagram for system ilmenite – MgO – O2 within the intervals suggested by the sample content (estimation with 
error margins) as the x-axis for investigating interactions between MgO and ilmenite. 

In the diagram the intervals in which magnesium forms spinel, titania spinel (both have the formula 

FeMg2O4) and Pseudobrookite as earlier described can be more easily seen. 



 
 
 
 
 

28 
 

3.3.8. Interactions with minor ash-component ZnO 

The resulting phase diagram for the system ilmenite – ZnO – oxygen can be seen in Figure 19 below. 

The green line indicates the zinc fraction (approximately 2 %) found in the sample obtained from 

Händelö. 

 
 

 
Figure 19. Phase diagram for system ilmenite – ZnO – O2 used for investigating interactions between ZnO and ilmenite. The green 
line indicates the Zn fraction (approximately 0.02) found in the sample obtained from Händelö. 

 
As zinc is a trace element/minor ash element, the leftmost part of the phase diagram is of greatest 

interest. As can be seen in the phase diagram, the formed phases and oxygen partial pressure intervals 

are similar to those of the system with only ilmenite and oxygen (Figure 5). Zinc however interacts 

with the ilmenite and forms spinel and (ZnO)2TiO2. The main spinel structure containing zinc is 

ZnFe2O4.  

 

 

 

 

 

 

 

 



 
 
 
 
 

29 
 

 

Possible reactions for interactions (also considering between levels) are given as Equation (15) – (20) 

below. 

 

6 ZnO + 3 FeTiO3 →  3 (ZnO)2𝑇𝑖𝑂2 +  Fe3O4 + 1½ O2   (15) 

ZnO + 2 FeTiO3 + ½ O2  →  ZnFe2O4 + 2 TiO2    (16) 

5 ZnO + 2 FeTiO3 + ½ O2 →  ZnFe2O4 + 2 (ZnO)2TiO2   (17) 

3 ZnO + 2 Fe3O4 + ½ O2 →  3 ZnFe2O4     (18) 

2 ZnO + TiO2  →  (ZnO)2𝑇𝑖𝑂2      (19) 

ZnO + Fe2TiO4 + ½ O2 →  ZnFe2O4 +  TiO2    (20) 

One or more phases formed from zinc and ilmenite interactions are stable over the entire oxygen 

partial pressure interval, indicating that zinc should accumulate on the ilmenite particles over time. 

 

3.3.9. Interactions with minor ash-components CuO/Cu2O 

The resulting phase diagram for the system ilmenite – CuO – oxygen can be seen in Figure 20 below. 

The green line indicates the copper fraction (approximately 0.6 %) found in the sample obtained from 

Händelö which corresponds to the fraction CuO. The diagram can also be read for Cu2O, for which 

the fraction corresponding to the sample instead would be 0.3 % (which is still at the very left end of 

the diagram and represented with the same line). 

 
Figure 20. Phase diagram for system ilmenite – CuO/Cu2O – O2 used for investigating interactions between CuO/Cu2O and ilmenite. 
The green line indicates the Cu fraction (approximately 0.006) found in the sample obtained from Händelö which corresponds to the 
fraction CuO. The diagram can also be read for Cu2O, for which the fraction corresponding to the sample instead would be 0.003. 



 
 
 
 
 

30 
 

As copper is a trace element/minor ash element, the leftmost part of the phase diagram is of greatest 

interest (lower concentrations of Cu). As can be seen in the phase diagram, the ilmenite 

oxidation/reduction behavior is unaffected by copper present in the system. Under oxidizing 

conditions, copper forms (Cu2O)(Fe2O3) (s2) and (CuO)(Fe2O3) (s3) (where s2 and s3 refers to 

different solid structures possible for the same chemical formula) by interaction with ilmenite. 

Under reducing conditions, these phases are reduced to solid copper. This indicates that copper 

should still accumulate on the ilmenite particles over time as it first forms (Cu2O)(Fe2O3), is 

integrated into the ilmenite structure and then solidifies on the surface. Possible reactions for the 

interactions (also considering between level) are given as Equation (21) – (24) below. 

 

6 CuO +  2 Fe3O4 →  3(Cu2O)(Fe2O3)     (21) 

2 CuO +  Fe2O3 →  (Cu2O)(Fe2O3) + ½ O2    (22) 

3 Cu2O + 2 Fe3O4 + ½ O2 →  3 (Cu2O)(Fe2O3)    (23) 

Cu2O +  Fe2O3 →  (Cu2O)(Fe2O3)      (24) 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 
 
 
 
 

31 
 

3.4. Interactions between ilmenite and multiple ash components 

3.4.1. Interactions with Zn and Cl  

The resulting phase diagram for the system ilmenite – Zn – Cl – oxygen can be seen in Figure 21 

below. The green line indicates the zinc fraction (approximately 2 %) found in the sample obtained 

from Händelö. A constant Cl fraction of 20 % was chosen as this was believed to be a high enough 

fraction to observe all possible phases of interaction but not be too high as to give unreasonable 

results. A decrease and increase to 5/10 and 30 % respectively give the same results (same stable 

phases) but somewhat shifted along the x-axis as the equilibrium concentrations are affected. The 20 

% was chosen over the 5/10 % for the figure as the phases were easier to distinguish. 

 
 

 
Figure 21. Phase diagram for system ilmenite – Zn – Cl – O2 used for investigating favored interactions between Zn and ilmenite/Cl. 
The green line indicates the Zn fraction (approximately 0.02) found in the sample obtained from Händelö. The ideal gas is ZnCl2 and 
the salt-liquid is FeCl2. 

As can be seen in the phase diagram, at low zinc fractions (close to the green line) the phases from 

interactions between zinc and ilmenite as found in Figure 19 are not stable when Cl is present in the 

system. Instead it forms ideal gas together with Cl, ZnCl2. With an increase in concentration (to a Zn-

fraction between 0.1 – 0.2), when half of the available Cl is bound to zinc, part of the gas becomes 

unstable and the Zn starts interacting with ilmenite and forms (ZnO)2TiO2. It is however very unlikely 

that zinc would reach these higher concentrations anywhere in the boiler, as it is a trace element and 

as indicated by the content in the sample obtained from Händelö. Above a zinc fraction of 0.2, 

(ZnO)2TiO2 forms as zinc and ilmenite are in excess to Cl.  
 



 
 
 
 
 

32 
 

As it is hard to distinguish the phases in the region around the zinc fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 22 below. 
 

 

 

Figure 22. Phase diagram for system ilmenite – Zn – Cl – O2 within the intervals suggested by the sample content (estimation with 
error margins) as the x-axis for investigating favored interactions between Zn and ilmenite/Cl. The ideal gas is a composition of 
ZnCl2, FeCl3 and FeCl2. The salt-liquid is FeCl2. 

As can be seen in Figure 22, the zinc is stable as gas, ZnCl2, rather than (ZnO)2(TiO2). This indicates 

that zinc will favor interactions with Cl over interactions with ilmenite under the conditions of low 

zinc concentrations in the boiler. 

 

 

 

 

 

 

 



 
 
 
 
 

33 
 

3.4.2. Interactions with Cu and Cl  

The resulting phase diagram for the system ilmenite – Cu – Cl – oxygen can be seen in Figure 23 

below. The green line indicates the copper fraction (approximately 0.6 %) found in the sample 

obtained from Händelö. A constant Cl fraction of 20 % was chosen as this was believed to be a high 

enough fraction to observe all possible phases of interaction but not be too high as to give 

unreasonable results (as for zinc). Increasing or decreasing the Cl content does not change the phases 

stable but shifts their intervals along the x-axis according to the new equilibrium concentrations.  

 

 

 
Figure 23. Phase diagram for system ilmenite – Cu – Cl – O2 for investigating favored interactions between Cu and ilmenite/Cl. The 
green line indicates the Cu fraction (approximately 0.006) found in the sample obtained from Händelö. The ideal gas consists of a 
majority (2/3) of Cu present as (CuCl)3 with approximately a third of FeCl2. The salt-liquid is FeCl2. 

As can be seen in the phase diagram, the copper behaves quite similar to zinc when Cl is present in 

the system. At low copper fractions (close to the green line) the phases from interactions between 

copper and ilmenite, as found in Figure 20, are not stable when Cl is present in the system. Instead it 

forms ideal gas, (CuCl)3, together with Cl. Only when all Cl has reacted with copper (at a Cu-fraction 

above 0.2), the Cu interacts with ilmenite and forms (CuO)(Fe2O3) (s2) and (CuO)(Fe2O3) (s3) (where 

s2 and s3 refers to different solid structures possible for the same chemical formula). This behavior 

is different than that of zinc, seen in Figure 21, as zinc starts to interact with ilmenite at lower zinc 

concentrations (0.1) whereas copper starts to interact with ilmenite first after reaching Cu fraction 0.2 

(which is the fraction of Cl in the entire system).  

 

As it is hard to distinguish the phases in the region around the copper fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 



 
 
 
 
 

34 
 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 24 below. 

 

 

 
Figure 24. Phase diagram for system ilmenite – Cu – Cl – O2 within the intervals suggested by the sample content (estimation with 
error margins) as the x-axis for investigating favored interactions between Cu and ilmenite/Cl. The ideal gas consists of a majority 
(2/3) of Cu present as (CuCl)3 with approximately a third of FeCl2. The salt-liquid is FeCl2. 

As can be seen in Figure 24, the copper is stable as ideal gas, (CuCl)3 over a majority of the oxygen 

partial pressure interval rather than (CuO)(Fe2O3) (s2) and (CuO)(Fe2O3) (s3). Otherwise it is present 

as solid copper, Cu (s). This indicates that copper will favor interactions with Cl over interactions 

with ilmenite under the conditions of low copper concentrations in the boiler. Decreasing the Cl 

content does not change the phases stable. 

 
 
 
 
 
 
 



 
 
 
 
 

35 
 

3.4.3. Interactions with Zn and S 

The resulting phase diagram for the system ilmenite – Zn – S – oxygen can be seen in Figure 25 

below. The green line indicates the zinc fraction (approximately 2 %) found in the sample obtained 

from Händelö. A constant sulphur fraction of 10 % was chosen as this was believed to be a high 

enough fraction to observe all possible phases of interaction but not be too high as to give 

unreasonable results. Increasing or decreasing the sulphur content does not change the phases stable 

but shifts their intervals along the x-axis according to the new equilibrium concentrations. The 

elemental analysis on the sample obtained from Händelö indicated that approximately 1.7 % sulphur 

was present in the sample. However, it is important to keep in mind that this is the content of sulphur 

that has interacted with the ilmenite particles (and phases formed on them) and not necessarily 

representative for the surrounding system. 

 
 

 
Figure 25. Phase diagram for system ilmenite – Zn – S – O2 used for investigating favored interactions between Zn and ilmenite/S. 
The green line indicates the Zn fraction (approximately 0.02) found in the sample obtained from Händelö. 

In the phase diagram for the system ilmenite – Zn – S – O2, Figure 25, it can be seen that ZnS is 

formed under reducing conditions. Under oxidizing conditions, sulphur is stable as gas, SO2. This 

indicates that interactions between zinc and sulphur are favored under reducing conditions, but 

interactions between zinc and ilmenite are favored under oxidizing conditions. This means that part 

of the available zinc should still accumulate on the ilmenite particles over time, but another part will 

form ZnS in the presence of sulphur. At very high concentrations of zinc (solid zinc pieces), ZnSO4 

can also be formed as a stable phase. 

 

As it is hard to distinguish the phases in the region around the zinc fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 



 
 
 
 
 

36 
 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 26 below. 

 

 
Figure 26. Phase diagram for system ilmenite – Zn – S – O2 within the intervals suggested by the sample content (estimation with 
error margins) as the x-axis for investigating favored interactions between Zn and ilmenite/S. 

In Figure 26, likewise as observed in Figure 25, ZnS is formed under reducing conditions. Under 

oxidizing conditions, the sulphur is stable as gas, SO2. Decreasing the sulphur content does not change 

the phases stable. 

 

 

 

 

 

 

 

 

 

 

 

 



 
 
 
 
 

37 
 

3.4.4. Interactions with Zn and SiO2 

The resulting phase diagram for the system ilmenite – Zn – SiO2 – oxygen can be seen in Figure 25 

below. The green line indicates the zinc fraction (approximately 2 %) found in the sample obtained 

from Händelö. A constant SiO2 fraction of 34 % was chosen so that the stable phases of zinc could 

be observed when ilmenite, Zn and SiO2 are in equal parts in the system (for investigating favored 

interactions between the compounds when they are in equal fractions) as well as for a low zinc 

fraction and high ilmenite and SiO2 fractions. A lot of SiO2 is present in the system; both as bed 

material and as an ash component. The elemental analysis on the sample obtained from Händelö 

indicated that approximately 18 % Si was present in the sample. However, this content does not only 

represent the Si that has interacted with the ilmenite particles (and phases formed on them) but the 

sample contains SiO2 which was not successfully magnetically separated from the sample. The Si 

fraction present in only the ilmenite particles is lower. The sample fraction can however be used as 

an indication of an upper limit of Si. Looking at the figure, fractions below 18 % all result in the same 

stable phases formed.  

 

 

 
Figure 27. Phase diagram for system ilmenite – Zn – SiO2 – O2 used for investigating favored interactions between Zn and 
ilmenite/SiO2. The green line indicates the Zn fraction (approximately 0.02) found in the sample obtained from Händelö. 

As it is hard to distinguish the phases in the region around the zinc fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 



 
 
 
 
 

38 
 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 28 below. 

 

 

 

Figure 28. Phase diagram for system ilmenite – Zn – SiO2 – O2 within the intervals suggested by the sample content (estimation with 
error margins) as the x-axis for investigating favored interactions between Zn and ilmenite/SiO2. 

 
As can be seen in the phase diagram, zinc forms spinel, ZnFe2O4, under oxidizing conditions. Under 

reducing conditions it forms willemite, Zn2SiO4, instead. Willemite is stable rather than the phase 

(ZnO)2(TiO2) observed in Figure 19. This indicates that zinc favors interactions with ilmenite under 

oxidizing conditions, and interactions with SiO2 under reducing conditions. Therefore, SiO2 should 

lead to a decrease in zinc accumulation on the ilmenite particles when it is present in the system. 

 

 
 
 
 
 
 
 



 
 
 
 
 

39 
 

3.4.5. Interactions with Zn, H2O and CO2 

The resulting phase diagram for the system ilmenite – Zn – H2O – CO2 – oxygen can be seen in Figure 

25 below. H2O and CO2 are the two compounds expected to together with Zn form hydrozincite 

(Zn5(OH)6(CO3)2), and therefore this system was chosen. The green line indicates the zinc fraction 

(approximately 2 %) found in the sample obtained from Händelö. A constant H2O and CO2 fraction 

of 25 % was chosen so that the stable phases of zinc could be observed when ilmenite, H2O and CO2 

are in equal parts in the system as well as for a low zinc fraction and high ilmenite and H2O/CO2 

fractions. This, since a lot of H2O and CO2 may be present in the system. The important compound 

hydrozincite (Zn5(OH)6(CO3)2), was found to be missing in the database, making the results 

inconclusive. 

 

 

 

 
Figure 29. Phase diagram for system ilmenite – Zn – H2O – CO2 – O2 used for investigating favored interactions between Zn and 
ilmenite/H2O/CO2. The green line indicates the Zn fraction (approximately 0.02) found in the sample obtained from Händelö. 

 

 

 

 

 

 



 
 
 
 
 

40 
 

As it is hard to distinguish the phases in the region around the zinc fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 30 below. 

 

 

 
Figure 30. Phase diagram for system ilmenite – Zn – H2O – CO2 – O2 within the intervals suggested by the sample content 
(estimation with error margins) as the x-axis for investigating favored interactions between Zn and ilmenite/ H2O – CO2. 

 
As can be seen in the phase diagram, zinc is not stable as hydrozincite (Zn5(OH)6(CO3)2), the 

compound expected from Zn/H2O/CO2 interactions. Instead, it is present as the same phases as in 

Figure 19 (spinel, ZnFe2O4 and (ZnO)2(TiO2)). However, it was found that the important compound 

hydrozincite (Zn5(OH)6(CO3)2) was missing in the database, making the results inconclusive. Any 

conclusions regarding the favored interactions between zinc, ilmenite and H2O/CO2 (to form 

hydrozincite) could therefore not be drawn. 

 

 

 

 

 

 

 



 
 
 
 
 

41 
 

3.4.6. Interactions with Zn and Al2O3 

The resulting phase diagram for the system ilmenite – Zn – Al2O3 – oxygen can be seen in Figure 31 

below. The green line indicates the zinc fraction (approximately 2 %) found in the sample obtained 

from Händelö. A constant Al2O3 fraction of 33 % was chosen so that the stable phases of zinc could 

be observed when ilmenite, Zn and Al2O3 are in equal parts in the system as well as for a low zinc 

fraction and high ilmenite and Al2O3 fractions. 

 

 

 

 

Figure 31. Phase diagram for system ilmenite – Zn – Al2O3 – O2 used for investigating favored interactions between Zn and ilmenite/ 
Al2O3/CO2. M2O3(Corundum) corresponds to a majority of Fe2O3 with some Al2O3. M2O3(Corundum)#2 corresponds to a majority of 
Al2O3 with some Fe2O3. Spinel denotes FeAl2O4 together with ZnAl2O4, and spinel#2 Fe3O4 together with FeAl2O4. The green line 
indicates the Zn fraction (approximately 0.02) found in the sample obtained from Händelö. 

 

 

 

 

 

 

 

 



 
 
 
 
 

42 
 

As it is hard to distinguish the phases in the region around the zinc fraction found in the sample 

obtained from Händelö (green line), an additional phase diagram was constructed within the intervals 

suggested by the sample content (estimation with error margins) as the x-axis, thus enlarging the areas 

of particular interest of the same phase diagram. This diagram can be seen in Figure 32 below. 

 

 

 
Figure 32. Phase diagram for system ilmenite – Zn – Al2O3 – O2 within the intervals suggested by the sample content (estimation 
with error margins) as the x-axis for investigating favored interactions between Zn and ilmenite/ Al2O3. 

 
In the phase diagram for the system ilmenite – Zn – Al2O3 – O2, Figure 31, it can be seen that the 

compound (ZnO)2(TiO2) seen in Figure 19 is no longer stable. Instead, the zinc forms M2O3 and 

spinel structures. The M2O3(Corundum) structure corresponds to a majority of Fe2O3, approximately 

90 %, with 10 % Al2O3. M2O3(Corundum)#2 corresponds to a majority of Al2O3 with some Fe2O3 

(approximately 90 and 10 % respectively). Spinel denotes FeAl2O4 together with ZnAl2O4, and 

spinel#2 Fe3O4 together with FeAl2O4. Over the entire diagram, the zinc is stable as ZnAl2O4. This 

indicates that interactions between zinc and Al2O3 should be favored over interactions between zinc 

and ilmenite.  
 

 



 
 
 
 
 

43 
 

   4. Conclusions 
This thesis has investigated how major ash components (Si, Al, Fe, Mg, Ca, K, Na and Cl) and the 

minor ash components Cu and Zn interact with ilmenite (FeTiO3) during waste incineration, by 

simulating different boiler conditions and constructing phase diagrams using the software FactSage 

7.2. The results obtained from the simulations are applicable for OCAC but could be used for 

understanding interactions in the CLC systems. Different phase formations depending on boiler 

conditions have been discussed throughout this thesis along with possible reaction paths for Cu and 

Zn. It was also discussed how the ilmenite oxygen carrier properties may be affected practically based 

on the theoretically calculated phases formed from interactions. 

 

For the major ash components, simulations indicate that: 

• Ilmenite interacts with Al2O3, CaO/CaCO2, Na2O, K2O, MgO, ZnO and Cu/Cu2O but does 

not interact with SiO2.  

• Interactions between ilmenite and aluminum oxide should form the titania spinel structure 

FeAl2O4.  

• Calcium (CaO) should interact readily with ilmenite when it is present in the boiler and 

accumulate on the particles over time, mainly in the form of CaTiO3 (a calcium rich ash-layer 

was observed on the ilmenite particles during SEM-EDX analysis). The interactions with 

calcium can be desirable, as calcium has been shown  to sometimes increase the 

reactivity/catalytic ability of ilmenite.  

• CaCO3 behaves similar to CaO at temperatures above 893°C, but a temperature below 893°C 

results in stability of CaCO3 in the oxidizing region and only interacting with ilmenite under 

reducing conditions, forming CaTiO3.  

• Sodium (Na2O) should interact with ilmenite and form mainly (Na2O)(TiO2)3 and 

(Na2O)(TiO2)6.  

• Potassium forms K2Ti3O7 and K2Ti6O13, which does not correspond to experimental findings, 

where potassium has been found as KTi8O16. The compound KTi8O16 is missing in the 

software database, therefore giving inconclusive results not corresponding to experimental 

findings.  

• Potassium displays behavior similar to sodium with increasing concentration. With increasing 

concentration, slag-liquid phase stable over the whole studied pressure interval starts forming. 

This indicates that the oxygen carrier performance of ilmenite is significantly affected if 

potassium and sodium accumulate on the surface of the particles and if they are present in 

significant concentrations in the combusted waste. These interactions are therefore especially 

undesirable.  

 
For the minor ash components Cu and Zn, simulations indicate that: 

• Zinc interacts with the ilmenite and forms spinel phase and (ZnO)2TiO2. The main spinel 

structure containing zinc in the simulations is ZnFe2O4.  
• Under oxidizing conditions, copper should form (Cu2O)(Fe2O3) (s2) and (CuO)(Fe2O3) (s3) 

by interaction with ilmenite.  
• The Cu and Zn systems simulated with Cl in them indicate that both Cu and Zn favors 

interactions with Cl over ilmenite, forming gas, (CuCl)3 and ZnCl2.  
• Interactions between zinc and sulphur are favored under reducing conditions, while 

interactions between zinc and ilmenite are favored under oxidizing conditions.  
• Similarly, zinc favors interactions with ilmenite under oxidizing conditions but interactions 

with SiO2 to form willemite (Zn2SiO4) under reducing conditions.  
• Interactions between zinc and Al2O3, forming ZnAl2O4, should be favored over interactions 

between zinc and ilmenite.  



 
 
 
 
 

44 
 

• No conclusions regarding whether interactions between zinc and H2O/CO2, to form 

hydrozincite (Zn5(OH)6(CO3)2), are favored over interactions between zinc and ilmenite can 

be drawn as data for the compound hydrozincite was found to be missing in the software 

database.   
 
Proposed reaction paths for Zn based on the stable phases indicated by the simulations are: 

 

• 6 ZnO + 3 FeTiO3 →  3 (ZnO)2𝑇𝑖𝑂2 +  Fe3O4 + 1½ O2    

• ZnO + 2 FeTiO3 + ½ O2  →  ZnFe2O4 + 2 TiO2     

• 5 ZnO + 2 F