master's thesis_silvia diener.pdf

7/29/2019 master's thesis_silvia diener.pdf

1/117


2/117

Faculty of Forestry, Geosciences and HydrosciencesInstitute of Waste Management and Contaminated SitesChair in Waste Management Prof. Dr.-Ing. habil. B. Bilitewski

Mineral phases of steel industry slags used in a landfill

cover construction

MASTER THESIS

TECHNISCHE UNIVERSITT DRESDEN

Submitted by: Silvia Diener

Matriculation number: 2800 264

Tutors: Dr.-Ing. Lale Andreas, Division of Waste Science & Technology,

Lule University of Technology, Sweden.

Prof. Dr.-Ing. habil. Bernd Bilitewski, Institute of Waste Management

and Contaminated Sites, Technische Universitt Dresden, Germany.

Dresden, the 21

st

of March 2006


3/117

I

Aufgabenstellung fr die Diplomarbeit

Thema: Mineralphasen in Stahlschlacken beim Einsatz

in der Dichtungsschicht einer Deponieoberflchenabdichtung

(Mineral phases of steel industry slags used in a landfill cover construction)

Pfannenschlacke und Elektroofenschlacke sind Abfallprodukte aus der Stahlindustrie.

Die derzeitige Praxis besteht in einer Deponierung dieser Stahlschlacken. Je nach

Stahlwerk fallen jhrlich Tausende Tonnen verschiedener Schlacketypen an.

Der Einsatz dieser Schlacken beispielsweise als alternative Baustoffe im Deponiebau

wird in Zukunft strkere Bedeutung und Akzeptanz erfahren.

Durch die zementhnlichen Eigenschaften von Pfannenschlacke nach Zugabe von

Wasser und anschlieender Verdichtung rcken Einsatzgebiete bei denen hohe Dichte,

niedrige Permeabilitt und mechanische Stabilitt vonnten sind, in den Vordergrund,

beispielsweise in der Dichtungsschicht einer Deponieoberflchenabdichtung. Die zu

untersuchenden Stahlschlacken stammen von Uddeholm Tooling AB, einem Stahlwerk

in Hagfors, Schweden.

Zielsetzung:

1) die Untersuchung der Abbindeeigenschaften von Pfannenschlacke, Elektro-

ofenschlacke und Mischungen aus denselbigen durch kalorimetrische Messungen,

2) die Untersuchung der kristallinen und amorphen Mineralzusammensetzung der

Schlackeproben mit Hilfe IR-Spektroskopie und Rntgenbeugung,

3) die Auswertung der Ergebnisse im Hinblick auf Langzeitverhalten und Stabilitt

der Mineralphasen sowie

4) die Analyse einer mglichen Korrelation zwischen Abbindeeigenschaften und Mi-

neralzusammensetzung der Materialien.


4/117

TABLE OF CONTENTS II

TABLE OF CONTENTS

LIST OF ABBREVIATIONS .......................................................... IIILIST OF FIGURES........................................................................... V

LIST OF TABLES ......................................................................... VII

TERMINOLOGY .................................................................................... VIII

SUMMARY ............................................................................................. XIII

1 INTRODUCTION......................................................................................1

2 MATERIAL AND METHODS.................................................................2

2.1 Material....................................................................................................................... 2

2.2 Methods ...................................................................................................................... 5

2.2.1 Calorimetry ...................................................................................................... 6

2.2.3 X-ray diffraction .............................................................................................. 7

3 RESULTS................................................................................................... 8

3.1 Calorimetry .................................................................................................................8

3.2 IR spectroscopy........................................................................................................... 9

3.3 X-ray diffraction ....................................................................................................... 13

4 DISCUSSION ..........................................................................................15

5 CONCLUSIONS......................................................................................22

6 REFERENCES.........................................................................................25

APPENDICES

APPENDIX I Sample preparation and Measurement Data

APPENDIX II Steel slags - General data, mineralogical compositionand suitable analysis methods for determining heat ofhydration and mineral phases - Literature Review

APPENDIX III Extended summary in German -Mineralphasen in Stahlschlacken beim Einsatz in derDichtungsschicht einer Deponieoberflchenabdichtung


5/117

LIST OF ABBREVIATIONS III

LIST OF ABBREVIATIONS

AOD Argon Oxygen Decarburisation

BF Blast Furnace

BOF Basic Oxygen Furnace (LD converter)

CEN Comit Europen de Normalisation

(European Committee for Standardization)

DRIFTS Diffuse Reflectance Infrared Fourier Transform Spectroscopy

DTA Differential Thermal Analysis

EAF Electric Arc Furnace

EDX Energy Dispersive X-Ray AnalysisFEhS FEhS - Institut fr Baustoff-Forschung e.V.

(former Forschungsgemeinschaft Eisenhttenschlacken;

German research institute for building materials)

FTIR Fourier Transform Infrared (Spectroscopy)

KBr Potassium Bromide

MID-IR Mid range of infrared radiation

IR Infrared (Spectroscopy)MSWI Municipal Solid Waste Incineration

NMR Nuclear Magnetic Resonance

OECD Organisation for Economic Co-operation and Development

OPC Ordinary Portland Cement

SEM Scanning Electron Microscopy

TG/TGA Thermal Gravimetric Analysis

XRD X-Ray Diffraction

A special cement nomenclature, called cement chemist notation (CCN) is used to

simplify cement formulas. Important abbreviations are:

C = CaO lime (calcium oxide)

S = SiO2 silica (silicon oxide)

A = Al2O3 aluminate (aluminium oxide)

F = Fe2O3 ferrite (iron oxide)

S = SO3 sulphate (sulphur oxide)


6/117

LIST OF ABBREVIATIONS IV

H = H2O in cement: hydrated water

C-S-H Calcium silicate hydrate

C-A-H Calcium aluminate hydrate

With the help of that nomenclature, formulas as for the mineral ettringite

6CaO*Al2O3*3SO3*32H2O can be written simpler as C6A S 3H32.


7/117

LIST OF FIGURES V

LIST OF FIGURES

Figure 1. EAF slag ............................................................................................................ 2

Figure 2. Ladle slag .......................................................................................................... 2

Figure 3. Chemical composition of EAF slag and ladle slag............................................3

Figure 4. Particle size distribution of EAF 1, EAF 2 and ladle slag

(Herrmann et al., 2005) ..................................................................................... 4

Figure 5. Sample cup in the calorimeter ........................................................................... 6

Figure 6. Temperature changes per gram slag during hydration measured with

calorimetry ........................................................................................................ 8

Figure 7. IR spectra of 100 % ladle slag (hydrated sample)............................................. 9Figure 8. FTIR spectra of two hydrated slags and three hydrated slag mixtures ........... 10

Figure 9. Comparison of area ratio of peaks caused by absorption of OH-bonds.......... 11

Figure 10. Diffraction pattern of 100 % ladle slag and 100 % EAF slag .......................13

APPENDIX I

Figure 1. IR spectra for all hydrated slag samples (LS = ladle slag)................................ 8

Figure 2. IR spectra for hydrated ladle slag and non hydrated ladle slag.........................9Figure 3. IR spectra of 100 % EAF slag and 100 % ladle slag (hydrated) ....................... 9

Figure 4. Diffraction pattern of 100 % ladle slag and 100 % EAF slag ......................... 10

APPENDIX II

Figure 1. Flow chart of the steel making process at Uddeholm Tooling Ltd.

(Beskow and Du Sichen; 2004)......................................................................... 3

Figure 2. Types of steel slags generated in Europe 2004

(Euroslag, 2006) ................................................................................................ 5

Figure 3. Utilisation of steel slags in Europe 2004 (Euroslag, 2006)............................... 6

Figure 4. Reuse of steel slags in Germany in 2004 (FEhS, 2005)....................................7

Figure 5. Layers and their functions in a cover construction with a mineral liner;

examples for possible alternative materials (Andreas et al., 2005)...................8

Figure 6. Changes in cement microstructure during hydration (Wenk and Bulakh, 2004)

........................................................................................................................................ 15

Figure 7. Phase diagram of CaOSiO2Al2O3 system.................................................... 17


8/117

LIST OF FIGURES VI

Figure 8. Differential calorimetric analyses of CA, CA2 and corresponding mixtures

(Mhmel et al., 1998)...................................................................................... 19

Figure 9. Compilation of literature data about main elements in EAF slags

(BMU data [6])................................................................................................ 21

Figure 10. Composition of Uddeholm ladle slag and BMU data about ladle slag [6].... 23

Figure 11. Mineralogical evolution taking place during weathering of MSWI

bottom ash (Piantone et al., 2004)...................................................................26

Figure12. The rate of heat evolution of cement hydration at 25 C

(Mostafa and Brown, 2005) ............................................................................ 30

Figure 13. X-ray diffraction of EAF sample (Shen et al., 2004).................................... 34

Figure 14. Vibration modes of the CO2 molecule .......................................................... 37

Figure 15. Schematical draw of the Michelson interferometer ...................................... 39

Figure 16. Comparison of interferogram and spectrum..................................................40

Figure 17. Sample accessory of the DRIFTS method .................................................... 40

Figure 18. Model spectrum for explaining fundamental terms

(Gnzler and Gremlich, 2002) ........................................................................42


9/117

LIST OF TABLES

VII

LIST OF TABLES

Table 1. Annual slag generation at Uddeholm Tooling Ltd (Herrmann et al., 2005)....... 3

Table 2. Mixtures of EAF and ladle slag used in the experiments ................................... 4

APPENDIX I

Table 1. Sample data.........................................................................................................2

Table 2. Recorded data of 100 % ladle slag sample ......................................................... 2

Table 3. Recorded data of 35 % EAF slag sample ........................................................... 3



Table 6. Recorded data of 100 % EAF slag sample ......................................................... 4

Table 7. Mixtures for preparation of specimens ...............................................................5

Table 8. Data of the cylinder ............................................................................................ 5

Table 9. Data of modified Proctor device......................................................................... 5

Table 10. Main IR peaks of 100 % EAF slag, 50 % EAF slag and the two ladle slag

samples including literature data .................................................................................... 11

APPENDIX II

Table 1. Steel slag generation in EU, Germany and Sweden ...........................................6

Table 2. Clinker and mineral phases in OPC according to Knoblauch

and Schneider (1992).......................................................................................... 12

Table 3. Main mineral phases in cement with chemical compounds and corresponding

cement nomenclature..........................................................................................14

Table 4. Hydraulic behaviour or different materials (Rhling et al., 2000) ...................16Table 5. Chemical composition of certain materials used in cement industry

(according to Knoblauch and Schneider, 1992).................................................17

Table 6. Chemical composition of two EAF slags (Motz and Geiseler, 2001) .............. 20

Table 7. Predominant mineral phases associated with steel slag

(Murphy et al., 1997)..........................................................................................22

Table 8. Regions of Infrared radiation............................................................................ 35


10/117

TERMINOLOGY

VIII

TERMINOLOGY

Amorphous Minerals that have no crystal lattice, i.e. no structure consisting

of crystal units. The antonym is crystalline.

Blast furnace In a blast furnace, iron ore is reduced to iron in the form of pig

iron. Raw materials are coke and iron ore. Furthermore, air is

blown into the furnace. Products of this continuous process are

molten iron and blast-furnace slag.

Carbonation Chemical reaction leading to formation of carbonates. An

example is the reaction of calcium hydroxide with carbon

dioxide in the air to calcite.

Cash minerals Minerals that have the ability to incorporate heavy metals in

their crystal structure (also called reservoir minerals). They can

be formed by thermal processes (primary cash minerals) or by

hydraulic reactions (secondary cash minerals). Certain cement

phases, carbonates and iron-hydroxides are known as cash

minerals.

Curing Hardening of cement due to hydration.

Diffraction When a beam of light is directed towards a crystal, light is

diffracted under a certain conditions, e.g. that distancesbetween the crystal lattice are equivalent to the wavelength of

the light beam. Diffraction means that the light of the original

light beams is scattered, many beams are formed that all have a

regular pattern including information about the crystal

structure.


11/117

TERMINOLOGY

IX

Fly ash Artificial puzzolan originating from coal combustion or MSWI

incineration. Fly ash is collected in electric filters and bag

filters.

Hydration1 The curing of cement and cementitious materials directly after

addition of water is based on the chemical and physical

integration of water in the structure of hydraulic minerals.

Chemically bound water cannot evaporate and belongs to the

hardened cement paste, whereas physically bound water present

in e.g. capillary pores can evaporate.

Ladle A vessel for transporting molten steel or metals.

Latent hydraulic Property of calcium silicate aluminate compounds. After

addition of water, hydration takes place but just in presence of

an activator, e.g. Ca(OH)2. An example for a material

consisting of calcium silicate compounds is blast furnace slag.

The cause for this property is a defective (metastable) latticestructure (Knoblauch, 1992).

Liner A liner is a designed containment layer. It can be built from

natural or synthetic materials. Here, the term is used for a part

in the top cover construction of a landfill, also called the barrier

layer.

Mineral phase A phase is any part of a system that is physically homogeneous

within itself and is mechanically separable from the other parts.

Any pure mineral is a single phase; any rock is a system in

which the phases are the individual minerals. In cement for

example the main mineral phases are tricalcium silicate,

dicalcium silicate, tricalcium aluminate and tetracalcium

alumino ferrite.

1 German: Hydratation


12/117

TERMINOLOGY

X

Molecular vibrations The chemical bonds of a molecule will vibrate, if a molecule

absorbs infrared radiation. The bonds can perform different

vibrations. They can stretch, bend and contract. Therefore,

infrared spectroscopy belongs to the methods of vibrational

spectroscopy.

Puzzolanic materials Natural or artificial materials that - together with water and the

activator calcium hydroxide (slaked lime) - form cement stone

similar products. Beneath alumina, puzzolans generally contain

silica amounts between 50 and 80%. The silica is present in an

amorphous state, otherwise no reaction would occur. Puzzolans

just form a hydraulic binder after the chemical reaction with

calcium hydroxide. The activator calcium hydroxide is needed

for the hydraulic reaction of puzzolanic materials even if there

is enough calcium oxide present in the material to build the

same hydration products.

They are therefore distinguished from hydraulic or latent

hydraulic substances. The reaction rate for a puzzolanicreaction is slower than for a hydration reaction (Grbl et al.

2001).

Solid solution2 According to IUPAC(1997), a solid solution is a mixed crystal.

Another constituent - apart from the ones belonging to the

original crystal - must be present that fits into the crystal

structure and is distributed in the host crystal.

Steel slags Also called steelmaking slags. The term includes EAF slags,

ladle slags and AOD slags. They are formed during steel

production by adding slag formers. The slag formers react with

the undesirable elements, e.g. carbon, silicon, sulphur or

phosphorous. These elements are oxidised and rise to the

2 German: Mischkristall


13/117

TERMINOLOGY

XI

surface of the molten steel, where they are removed

(deslagging).

Tapping of steel Here: The process of pouring the steel from the EAF furnace

into the ladle.

Uphill casting For solidification, different steel casting processes exist. The

typical ingot mould casting process is known as uphill casting,

in which molten metal fills a mould from the bottom.

Vacuum degassing After the steel has been treated in the ladle, the ladle is brought

to a degassing station where it is put under a vacuum lid. The

vacuum treatment and the additional stirring with injected

argon gas or blowing of oxygen shall a. o. reduce several

elements as e.g. hydrogen in the steel.


14/117

PREFACE

XII

PREFACE

I would like to express my gratitude to Dr. Lale Andreas for giving me not only the

opportunity to write this thesis at Lule University of Technology (LTU), but also forhelpful answers to all my questions, the organization of my stay, the supervision of this

thesis and her efforts to make me settle in.

I thank Prof. Dr.-Ing. habil. Bernd Bilitewski for being my supervisor in Dresden.

I also thank Inga Herrmann for her help, for her correction tips and last but not least for

bailing me out at Christmas 2005.

The master thesis was performed in co-operation with the division of process

metallurgy at LTU. For the laboratory analyses, data interpretation and sharing theirknowledge about metallurgy and minerals, I especially thank Margareta Lidstrm

Larsson and Fredrik Engstrm.

I really enjoyed my stay at the division of Waste Science and Technology in Lule and

always felt a very friendly atmosphere there. Therefore, I want to thank Roger Lindfors,

Ulla Brit Uvemo, Igor Travar, Lisa Dahln, Dr. Holger Ecke and Prof. Dr. Anders

Lagerkvist.

Furthermore, I got to know a lot of friends in Lule who tried to make me see the things

besides work and studies.

I also thank Rita Ougolnikova and Irene Schneider for last comments on this work.

I thank my family, for their continual support in everything that I do. I will always be

grateful to you.


15/117

SUMMARY XIII

SUMMARY

In 2004, the European steel industry generated about 15.2 million tonnes of different

steel slags. Out of these, electric arc furnace (EAF) slags and secondary metallurgical

slags account for almost 6 million tonnes (Euroslag, 2006). Steel slags can potentially

be reused, e.g. as construction material in landfill liners or cover constructions. Their

physical and chemical properties have been investigated by e.g. Herrmann et al. (2005),

Shen H. et al. (2004), Shi (2002), Motz and Geiseler (2001) and Fllman (1997).

However, not much is known with regard to their long-term behaviour. The presented

project deals with alterations of the mineralogical composition of steel slags as a base

for estimations of the long-term stability.

Two types of steel slags, electric arc furnace (EAF) slag and ladle slag were mixed in

different proportions and analysed with calorimetry, infrared (IR) spectroscopy and X-

ray diffraction (XRD). In the mixture ladle slag reacts with water by hydration, while

the EAF slag works as filler.

The study focused on the cement reaction as well as on the mineralogical composition.

The cement reaction was tested with the aid of calorimetry, whereas IR spectroscopy(FTIR spectrometer) and XRD were used to analyse the amorphous and crystalline

mineral phases. A method for preparing specimens was developed.

The temperature development of the different mixtures of EAF and ladle slag recorded

by the calorimeter showed an activation of the mixture by Al-rich ladle slag: higher

portions of ladle slag resulted in an increased development of heat. However, higher

portions of ladle slag also involve a delay of the maximum heat build-up.

The IR analyses showed that changes in the sample spectra were proportional to the

content of EAF slag. Additionally, one IR analysis of non-hydrated (dry) ladle slag was

performed. During qualitative analysis of the IR spectra, absorption bands were

identified in the wave number ranges of inorganic carbonates, calcium silicates and

aluminium oxide. The biggest differences between the samples could be seen in the

region from 3500 to 3200 cm-1. Ladle slag showed clear absorption peaks, while EAF

slag did not. This wave number range represents the absorption of O-H bonds and

therewith the integration and absorption of water molecules in the mineral structure.


16/117

SUMMARY XIV

With the help of XRD, the main minerals in a 100 % ladle slag were identified as

- Calcium silicate ( -Ca2SiO4) and Mayenite (Ca12Al14O33), whereas Merwinite

(CaMgSiO4) and Monticellite (Ca3Mg(SiO4)2) were most common in 100 % EAF slag.

New mineral phases after mixing both slag types could not be detected with XRD.

Mineralogy determines steel slag properties and liner performance. Mineral alterations

include the formation of secondary minerals through weathering of the analysed

primary minerals. Estimations for these aging reactions of primary minerals are

outlined.


17/117

INTRODUCTION

1

1 INTRODUCTION

Electric arc furnace (EAF) slag and ladle slag are by-products of steel produced in an

electric arc furnace and the following secondary metallurgy processes in a ladle vessel.

Their reuse is not only of environmental benefit as natural aggregates do not have to be

exploited, but also of economical benefit as many landfills in Europe need a final cover

in the future.

In order to assess if EAF slag and ladle slag can be utilised as construction materials in

the liner barrier of a landfill, their physical and chemical properties have to be known,

as well as their environmental impact and long-term behaviour. The mechanical and

chemical stability of the construction is an important question.

The aims of the study are to estimate the reactivity of the slags after water addition, to

determine amorphous and crystalline mineral phases in the slags and slag mixtures, and

to evaluate the stability of the major mineral phases. Another question is the formation

of new mineral phases after mixing EAF slag with ladle slag and water.

The reactivity of the slags after water addition is together with the gradient of heat

generation over time an indicator for the formation of more or less stable mineral phases

during the curing process. The stability of the steel slag minerals is important for the

prediction of mineral transformations that may occur in a landfill environment and

therewith affect the stability and long-term behaviour of the cover construction itself.

The suitability of the analysis methods calorimetry, IR spectroscopy and X-ray

diffraction (XRD) for answering the above stated research questions was assessed in a

literature review (Appendix II). A method to prepare samples for IR and XRD analyses

was developed and is described in Appendix I, together with the sample preparation for

the calorimetric experiments and the collected data.


18/117

MATERIAL AND METHODS

2

2 MATERIAL AND METHODS

In the following chapter, the well-known properties of the steel slags are reported about.

In addition, the measuring procedure of the various analytical techniques is described.

2.1 Material

Five steel slags were sampled from Uddeholm Tooling Ltd., a company producing a. o.

high-alloyed tool steel. The slags were investigated at the Division of Waste Science

and Technology at Lule University of Technology (LTU) during autumn and winter

2004/05. The results of these investigations are published in Herrmann et al. (2005) and

Andreas et al. (2005) yet those results being important for the understanding of this

study are shortly described in the following.

Two EAF slags and one ladle slag were investigated in this study, whereas both EAF

slags (called type 1 and 2) are mixed in equal shares due to their similar chemical

composition. Figure 1 and 2 show the two steel slag types.

Figure 1. EAF slag Figure 2. Ladle slag

The investigated EAF slag and ladle slag had some major differences in their chemical

composition, which is presented in Figure 3. Not only does ladle slag contain 26 %

more calcium oxide and 47 % more aluminate, it also reaches only 39 % of the silicate

content of EAF slag.


19/117


3

0

5

10

15

20

25

30

35

40

45

CaO SiO2 Al2O3 MgO MnO Fe2O3 FeO

content[wt.-%

]

Electric Arc Furnace slag 1/2

Ladle slag

Figure 3. Chemical composition of EAF slag and ladle slag

EAF 1 was produced under reducing conditions while there was still a high silicon

content in the steel melt. EAF 2 was formed under a reducing atmosphere as well, but

with addition of silicon and iron chromate (FeCr). Furthermore, oxygen was blown into

the melt.

The ladle slag was tapped after the secondary steelmaking process at the ladle furnace

station. In the ladle, a. o. processes as deoxidisation with aluminium, addition of slag

formers (dolomite and lime) and alloys take place. The production steps in the EAF andin the ladle are further explained in Appendix II (Chapter 2).

Uddeholm Tooling Ltd. is a small steel mill. The annual slag output can be seen from

table 1.

Table 1. Annual slag generation at Uddeholm Tooling Ltd (Herrmann et al., 2005)EAF 1 EAF 2 LS

amount per year

[ton] 5200 2000 1300

After tapping, the slag is stored indoor and just cooled by air contact. The slag types are

piled separately. A representative sample was taken from every heap in summer 2004.

EAF 1 and 2 were crushed to a size smaller than 20 mm. They contain much coarser

particles than the ladle slag which disintegrates into fine powder because of the high

Ca/Si ratio. As can be seen in Figure 4, two thirds of the ladle slag consists of particles

smaller than 0.25 mm, while only 18 to 24 % of the EAF slag particles are below that

size.

CaO SiO2 Al2O3 MgO MnO Fe2O3 FeO


20/117


4

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.001 0.01 0.1 1 10 100

particle size [mm]

passedamount

EAF 1

EAF 2

Ladle

Figure 4. Particle size distribution of EAF 1, EAF 2 and ladle slag(Herrmann et al., 2005)

Curing, a special feature of materials with cementitious properties occurred after mixing

the slags with water. This hardening of the material originates from the hydration

reaction and is further explained in Appendix II.

Hardening tests described in Herrman et al. (2005) showed that ladle slag hardens very

quickly in contrast to EAF slag. In order to be able to use it as a construction material,

EAF slag 1 and 2 can be added to decelerate the hardening process that there is at

least one day time for construction. Curing of the material originates from hydration

which is further explained in chapter 4 of Appendix II. All three types if steel slags,

EAF slag 1, 2 and ladle slag, have been categorised as non hazardous waste according

to EU legislation (EU, 2002). The mixtures of ladle slag and EAF slag 1 and 2 (in equal

shares) used in all measurements are shown in table 2.

Table 2. Mixtures of EAF and ladle slag used in the experiments

EAF slag [%] 100 0 35 50 65

Ladle slag [%] 0 100 65 50 35


21/117


5

2.2 Methods

In order to quantify the heat release during hydration, the slags are mixed with water

and analysed in a calorimeter. The method of IR spectroscopy was taken to observe

changes in the chemistry of the hydrated slags, in case of a major part of amorphous

minerals in the slags. Furthermore IR spectra are easier to analyse than e.g. X-ray

diffractograms. The spectra can be analysed without the use of databases, but

identification of a compound is usually not done only with IR analyses. The most

certainty is achieved by combining information derived from IR spectra with knowledge

of the chemical composition, i.e. other analysis methods. About combinations of IR

spectroscopy with other analysis methods has been reported about in Appendix II.

In combination with IR spectroscopy, XRD was applied. XRD is especially of use for

phase identification of crystalline solids. It is therewith possible to distinguish, e.g. two

minerals with the same chemical composition but different crystal structure. The

detailed applied methods of IR spectroscopy and XRD are explained in chapter 6 and 7

of Appendix II.

Before analyses, the slags were sieved. From EAF slags, the fraction 8 < x > 19 mm and

from ladle slag, the fraction < 19 mm was taken and analysed. For the calorimetry

measurements, the slag was taken in the original state. Slag samples for IR and XRD

analyses were prepared. Each sample was mixed with 10 % of water and compacted in

three layers. This water content was determined as optimum water content for a

compacted slag mixture with a maximum density. The corresponding Proctor-

experiment is found in Herrmann et al. (2005).

The specimens were stored in a bucket with 100 % humidity for two days in order toallow hydration of as many minerals as possible during these two days. A complete

hydration of the material would take much longer, but not more time than one week of

sample preparation was available. After that, the specimens were dried for five days in

an exsiccator. It is important that he samples for the IR measurements are dry because

otherwise, water would absorb IR radiation and the spectra would be altered. After

drying the samples were grinded (particle size required for IR and XRD: < 200 m) and

iron particles were removed by a magnet (see also Appendix I for preparation of


22/117


6

specimens). One ladle slag sample (100%) was analysed with IR spectroscopy without

mixing it with water before. Therefore, this sample is called non hydrated ladle slag.

2.2.1 Calorimetry

The experiments were carried out with a Parr 1455 Solution Calorimeter. The sample

chamber shown in Figure 5 is a fully silvered glass dewar within a stainless steel air

can. For isolation the whole assembly is surrounded by a block of plastic foam which

lies in a rugged aluminium case. Temperatures are measured with a temperature probe,

which has to be immersed to the water in the dewar. The calorimeter is connected to a

strip chart recorder. The anticipated temperature range was fixed with the help of the

key functions of the calorimeter. The temperature change is simultaneously measured

and recorded on a paper sheet. The measurement was cancelled when the temperature

decreased up to the initial temperature. The heat outputs of the slags are compared by

relating the temperature change to the mass of the slag in the calorimeter.

The water in the dewar (ca. 40 ml) and the water later mixed with the sample (water

content ca. 30 %) should both have approximately room temperature. The high water

content was taken to have an excess of water, because hydration is limited by water

content. The water is filled in the measurement assembly before sample addition to

record the initial temperature. When the initial temperature is stable, sample is added

and mixed with water.

Figure 5. Sample cup in the calorimeter

water in dewar

thermometer

sla sam le in lastic containerwater


23/117


7

2.2.2 IR spectroscopy

The samples were taken from the same material which has been prepared for the XRD

analyses, except of an additional grinding with a mortar before IR analyses. Beforemeasuring the actual samples, a background spectrum needs to be recorded. For this,

Potassium bromide (KBr) is used as a so called window material, because it does not

absorb in the MID-IR, i.e. it is infrared transparent. To not take up humidity from the

air, KBr is stored in an exsiccator. It has to be grinded with a mortar in a bowl as well.

The spectrometer automatically subtracts the background spectra of KBr from all

measured spectra. The performance of a quantitative analysis was not undertaken. For a

quantitative analysis, the analysis has to be performed with a standard included in thesample.

2.2.3 X-ray diffraction

For the diffraction experiments, a powder diffractometer was used. A powder

diffractometer uses a detector to register the positions of the scattered X-rays. The

detector is moved around the sample on a circular plane. Each lattice plane of a crystal

in the sample can diffract X-rays at all angles that fulfil the Bragg-equation (see chapter6.1, Appendix II) and therewith produces a diffraction cone. These diffraction cones

consist of very close beams diffracted by crystallites in the powder and are recorded by

the detector. Consequently, the received powder diffractogram is determined by the

crystal structure and unique for each material.


24/117

RESULTS

8

3 RESULTS

In the following, analyses results of calorimetry, IR spectroscopy and XRD are

presented. Further IR spectra are displayed in Appendix I.

3.1 Calorimetry

The heat development of the different slag mixtures is shown in Figure 6 below. The

x-axis shows the time, the y-axis the temperature difference per gram slag. The heat

released decreases with increasing EAF slag content. For one sample with 100 % EAF

slag, the heat development over 60 hours was recorded and it showed a continuous heat

release. However, the temperature change was only 0.22 C.

0.00

0.05

0.10

0.15

0.20

0.25

0 5 10 15 20 25 30 35time [hrs]

[dT/g]100%LS

35%EAF

50%EAF

65%EAF

100%EAF

Figure 6. Temperature changes per gram slag during hydration measured with

calorimetry

The diagram shows that the heat development in the dewar occurred under partly

adiabatic3 conditions. One part of the released heat of hydration leads to heating of the

slag sample, another part is evolved to the environment (depending on isolation

properties) and the third part increases the temperature in the calorimeter. At the

beginning, conditions are almost adiabatic. Heat generation and heat outflow are in

equilibrium at the maximum of the reaction (Zement-Taschenbuch, 2000).

3 Adiabatic conditions: no heat exchange between sample and environment


25/117

RESULTS

9

3.2 IR spectroscopy

The IR results are compiled in chapter 3 of Appendix I. The IR spectra of EAF and ladle

slag plus the spectra of three mixtures of them, i.e. 35 / 50 / 65 % and one additional

sample of non hydrated ladle slag were recorded. The IR spectrum of 100 % (hydrated)

ladle slag is given as an example in Figure 7. The two below explained peaks A1 and

A2 are marked.

4000.0 3000 2000 1500 1000 500 400.0

0.000

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0.60

0.65

0.70

0.75

0.80

0.85

0.900

cm-1

K-M

36233526

3318

1656

1422

970

929919

866

818

568

523

476

442

426411

Figure 7. IR spectra of 100 % ladle slag (hydrated sample)

The spectrum in Figure 6 is shown in the form wave number vs. Kubelka-Munk units.

For qualitative interpretation, just the relation of the peaks is of interest, not the

respective Kubelka-Munk value of each peak itself (see chapter 7.4, Appendix II). The

range on the x-axis comprises the wave numbers 40,000 to 400 cm-1

. In Figure 8, acompilation of all sample spectra is presented.

peak with area A2peak with area A1


26/117

RESULTS

10

Figure 8. FTIR spectra of two hydrated slags and three hydrated slag mixtures

The biggest difference between all spectra is seen in the wave number range from 3500-

3200 cm 1 . The spectrum of hydrated ladle slag shows several peaks, at e.g. 3623 and

3526 cm 1 , whereas the 100 % EAF slag has none there. This wave number range

reflects the vibrations of O-H bonds, which can be bound to several components.

During the hydration reaction, the water molecules took up spaces in the crystal lattice.

If an OH

-

ion is bound at a different place of the molecule or mineral after reacting withthe water, the O-H bond absorbs IR radiation at a slightly different frequency.

The mixtures show similar peaks as the ladle slag, but these peaks are not as intense as

the ladle slag peak at 1600-1300 cm-1. The latter one is explained further down but the

decrease of the ratio between the two peaks is evaluated in Figure 9 below. It shows,

that the more ladle slag the sample contains, the more intense is the peak A1 from 3750-

2450 cm-1 compared to peak A2 from 1600-1300 cm-1.

4000 3000 2000 1500 1000 500 400wave number/cm-1

K-M

3299 1462

983 905

653

590

523

866

865

35263338

1660

1418

9 78 951

653

524

35281655

1415

867

817 669

525

36223527

33281655

1418

972

929

919867

818

524

3623352633181656

1422

970

929

919

866

818568

523

100%EAF, 0% LS

65%EAF, 35% LS

50%EAF, 50% LS

35%EAF, 65% LS

0%EAF, 100% LS


27/117

RESULTS

11

0

35

65100

0

1

2

3

4

5

6

0 20 40 60 80 100

ladle s lag [%]

A1/A2

Figure 9. Comparison of area ratio of peaks caused by absorption of OH-bondsA1 = area below the graph from 3750-2450 cm-1

A2 = area below the graph from 1600-1300 cm-1

There is also a broad band around 3200 cm-1 (accompanied with a weak peak at

1655 cm-1), which is caused by physically bound water respectively its asymmetric

stretches. That means that water molecules are absorbed to some compound.

A general rule is that the lower the wave number, the stronger the bond. The reason is

that atoms having a larger mass, e.g. metals, result in higher bond forces and absorb at a

lower frequency, i.e. wave number.

DRIFTS (Diffuse Reflectance Infrared Fourier Transform spectroscopy) measures the

radiation scattered at the surface of a material. Therefore small impacts of the surface,

can be seen clearly in the spectrum. But also changes of the air, respectively atmosphere

in the sample compartment during the analysis can influence the spectrum. Same

spectra, as the one of ladle slag slag show a smaller peak at 2359 cm-1 for example. CO2

is known for having an antisymmetric C=O stretching at 2349 cm -1 (from the rotational

spectrum). Probably, the small peak in the steel slag represents the changing CO2

content in the atmosphere during the recording of the KBr as background spectrum and

the sample itself.

In that so called fingerprint region (from 1500 to 650 cm-1), a spectrum may have

hundred or more absorption bands present, but according to Gnzler and Gremlich

(2002) only the most intense bands serve as indicators. Therefore, only the most

intense peaks can be assigned to a bond. For the other ones, often overlapped by eachother, assumptions can be made.


28/117

RESULTS

12

A broad band present in all spectra lies at 1500-1350 cm -1, originating from C-O

stretching vibrations in the functional group CO3 of inorganic carbonates as CaCO3

(further bands at 880 and 710 cm-1).

Also, calcium silicate and calcium aluminate phases are likely to be present. Generally,

they show bands from 1100 to 800 cm-1 (CaxSiyOz) as well as from 600 to 500 cm-1

(CaxAlyOz). One problem of analysing them with certainty is that many bounds show

absorption in that area. Resolving the broad (overlapping) bands of spectra is

complicated. However, the 970 cm-1 is considered to be originated from calcium silicate

phases as this peak can be seen rather clearly. The stretching vibration of the Si-O bond

in calcium silicates absorbs in the region 1100-800 cm-1 as well as from 600 to

500 cm-1.

Anhydrous calcium silicates anhydrous refers to no crystalline water absorb at

920 cm-1, 536-526 and 464-458 cm-1 (Gomes and Ferreira, 2005). Furthermore, they are

known for having puzzolanic properties. Hydrated ladle slag shows the 920 cm-1 band.

But with an increasing amount of EAF, this peak is getting harder to identify.

Because oxides of iron, aluminium and small amounts of magnesium are present in the

slags, those are worth to look for as well. Al-O stretching vibrations occur in the regionfrom 950-800 cm-1 and the strong peak at 866 cm-1 in the ladle slag spectrum is

probably due to Al-O stretching vibration. This peak is shifted towards 860 cm-1 in the

non hydrated ladle slag and also present at the 100 % EAF slag samples, but with much

lower intensity. The aluminium peak is much weaker in the EAF spectra, because ladle

slag has a three times higher aluminium content than EAF slag. Possible compounds

containing aluminium can be calcium alumino silicate hydrates (C-A-S-H).

Speaking about differences in the spectra, it can be said, that the relation of the peak

around 866 cm-1 and the peak around 970 cm-1 changes with EAF content. The more

EAF the sample contains, the higher is the intensity of the around 970 cm-1 compared to

the peak around 866 cm-1. As said before, the first peak is assumed to be aluminium

while the second is related to calcium silicates.

Another peak shifting between hydrated and non hydrated ladle slag has occurred from

568-576 cm-1. Possible compounds causing these bands are again aluminium or iron

oxides (610-520 cm-1 and 580-560 cm-1), whereas magnesium oxide bands should be


29/117

RESULTS

13

placed at slightly lower wave numbers (~560-550 cm-1). But calcium silicates and

aluminates absorb in this range as well. Molecules absorbing at wave numbers lower

than 700 cm-1 cannot be identified with certainty.

3.3 X-ray diffraction

Steel slags are a very complex material, which means they can contain a variety of

minerals. The XRD measurements showed that most of the minerals are present as

crystals, i.e. very few amorphous phases have been detected. The presence of

amorphous structures can be detected with XRD as an increase of the baseline from

2 (2-Theta) = 20 on. But the pattern produced of an amorphous phase cannot be

assigned to a compound.

For identification, the diffractogram is compared with minerals in a database. If a

sample consists of more than one mineral, as it is the case for steel slags, the different

peaks of each substance overlap each other and get harder to identify. A mineral only

can be detected if the sample contains more than 4 % of it. EAF slag and ladle slag

consist of different minerals.

In Figure 10, the diffraction pattern of 100 % ladle slag and 100 % EAF slag are shown.

X-ray diffraction pattern of steelslags

10 20 30 40 50 60 70 80 90

2 - Theta - Scale

Inten

sity

100 % EAF

100 % Ladle

Figure 10.Diffraction pattern of 100 % ladle slag and 100 % EAF slag

100 % ladle slag

100 % EAF slag


30/117

RESULTS

14

The mixtures of EAF and ladle slag only contained mixtures of the minerals of both

samples indicating that no new minerals have been formed.

The minerals detected have been:

Ladle slag

Periclase MgO

Dicalcium silicate - Ca2SiO4

Iron Fe

Spinel MgAl2O4

Mayenite Ca12Al14O33

Dicalcium Silicate - Ca2SiO4

EAF slag

Iron Fe

Merwinite Ca3Mg(SiO4)2

Monticellite CaMgSiO4

Clinoenstatite MgSiO3

Dicalcium Silicate Ca2SiO4

Magnesium Aluminium Oxide MgAl2O4

Out of these, the main minerals were

- Dicalcium silicate ( - Ca2SiO4) and Mayenite in ladle slag and

Merwinite and Monticellite in EAF slag.

The minerals have different properties considering their hydraulic behaviour. Periclasefor example forms hydraulic minerals under addition of water.

In XRD analysis, the peaks of high intensity can be assigned to a certain mineral. The

strong peak in the EAF slag diffractogram for example originates from periclase.

Furthermore, different phases of dicalcium silicate have been recorded. They are formed

during cooling of the molten slag. Each of the five polymorphs of Calcium silicate has a

different crystal structure and different properties.


31/117

DISCUSSION

15

4 DISCUSSION

Method development for preparing sample specimens

For the IR and XRD analyses, a method of sample preparation including hydration of

slags and compaction was developed. Five different slag specimens according to five

mixtures were prepared. They were stored in a 100 % humid atmosphere for two days

(see description in chapter 2 of Appendix I). This short period of time restricts the

curing processes since the hydration is limited by the water content. If water is

available, the hydration will continue until all silica molecules have reacted with the

OH- ions. The short curing time affected certainly the results of the mineralogical

analyses but the time for preparing samples was limited by external circumstances and,

also, the study had a survey character, i.e. the results are considered sufficient for a first

overview and more investigations are planned.

Analysis methods

The common area for the application of IR spectroscopy is analysis of the chemical

structure of organic compounds. However, no reason could be identified - by studying

the literature (Appendix II) - to not use IR spectroscopy for inorganic compounds as

well. Furthermore, the possible presence of a considerable amount of amorphous mate-

rial in the slags suggested the IR method.

The evaluation of the IR spectra of steel slags was based on personal experience and

literature data. However, available databases for IR data were difficult to use because

steel slags are very complex and contain a broad variety of minerals. Absorption data

from databases mostly refer to pure chemical compounds.

IR measurements of such complex materials as steel slags can not give evidence for thepresence of a compound with absolute certainty. Together with the knowledge of the

chemical composition and the absorption data, it only allows statement s like that the

present elements and functional groups absorb in a certain wave number range. If

absorption in that wave number range was measured, it can originate from the assumed

compound, but it also can originate from the absorption of other compounds or

influences as sample preparation or changing conditions while the analyses were

performed. A quantitative analysis has not been performed. In such an analysis astandard is measured together with the window material (KBr) and the slag.


32/117

DISCUSSION

16

The analyses of hydrated and non hydrated ladle slag proved the integration of hydra-

tion water in the slag structure. The absorption bands below 1000 cm -1 and particularly

below 700 cm-1 are difficult to assign to a chemical compound as not only carbonates,

calcium silicates and calcium aluminates absorb in that range but also iron, aluminium

and magnesium compounds.

The differentiation of mineral phases as calcium silicates as tricalcium silicate and

dicalcium silicate is not possible with IR measurements.

Furthermore, conditions in the atmosphere of the sample compartment in the apparatus,

as for example CO2 and air humidity can have an influence on the measurements. This

is because water vapour and carbon dioxide absorb in the MIR. If the atmospheric

conditions in the sample compartment are the same during the analysis of both the

sample and the background, the instrument will ratio out these bands and they are not

seen anymore in the final spectra (Smith, 1996). The IR spectrometer had to be moved

to another room between the first experiments (including measurement of background)

and the repeated analyses of 50 % EAF slag and non hydrated ladle slag. However, only

a little disturbance around 2350 cm-1 has been caused by this.

The investigation of the crystal structures was done by XRD analyses. Even if it is not

possible to identify amorphous phases with XRD, the diffractogram of a sample contai-

ning amorphous phases shows a hump of the background in the range 2 (2 Theta) ~ 20

to 40. In the diffractograms, this hump was very low. Hence, the majority of the

minerals in the slags are present in a crystalline state. The time for crystallisation after

tapping of the slag is determined by the slag treatment. The Uddeholm slags have been

stored and cooled by air contact and could therefore crystallise during storage.

As diffraction occurs at the sample surface, the diffracted X-rays only contain

information about the minerals detected at the surface. For that reason the sample

preparation has to be very careful, so that crystals are evenly distributed in the sample.

The information about the main minerals in the EAF slag and ladle slag was attained by

thermo dynamic calculations. The diffractograms were evaluated by a database. Except

of some very intensive, clear peaks, it is impossible to analyse them visually, because of

the number of peaks caused by each mineral. XRD as well as IR peaks can overlap each


33/117

DISCUSSION

17

other, the peaks can be shifted on the x-axis or they can differ in intensity from sample

to sample.

Some compounds have very similar diffraction patterns. They can consist of the same

crystalline phase but include different irregularities in their crystal structure. This

phenomenon is called solid solution. Under certain conditions, minerals can exchange

atoms. For example calcium silicate phases as Ca(Mn)Si2O6 can have iron atoms at

some places in the crystal structure. XRD is not able to detect those small differences at

the atomic level.

XRD analyses further proved that no mineral formations were induced by mixing EAF

and ladle slag. Only minerals that were present in the pure slags could be detected in the

mixtures. One possible explanation is that no new phases have been formed (at the

present conditions, e.g. 25 C), but it is also possible, that new formed phases were

below the detection limit of 4 wt-%. As only the surface is scanned by X-rays, small

varieties of the particle distribution originating from the sample preparation could cause

errors too.

Heat generation and curing behaviourThe curing properties and the released heat of hydration have been analysed with the

calorimeter. The heat generation was highest for pure ladle slag while pure EAF slag

released no or very little heat. The ladle slag specimen hardened very fast and, looking

at the different mixtures, curing was the faster the higher the portions of ladle slag.

Ladle slag has a CaO/SiO2 ratio of 3 and hence, is considered as a hydraulic material.

The larger this ratio, the higher is the hydraulic reactivity of the material.

Even if aluminate activates the hydration and the formation of cement phases, the

maximum heat output occurred slightly later, the more ladle slag, i.e. aluminate, the

sample contained. The hydration reaction of pure ladle slag occurs 5 hours later

compared to the mixture containing 65 % EAF slag. Probably, the presence of a cement

activator can retard the time of maximum heat release.

The function of aluminate in the mixtures of EAF and ladle slag was weak not just

because of a lower amount of ladle slag but also because of the low amount of reactive


34/117

DISCUSSION

18

lime and high silica in EAF slag4 (see composition in chapter 2). That means that EAF

slag of that composition can be seen as a latent hydraulic material even though almost

no heat output has been recorded.

From cementitious materials it is known that a slow curing process with heat output

over longer period leads to higher long-term stability. For answering the question,

whether the reaction of the samples containing EAF slag proceeds faster but stops

earlier, or if the low reaction rate leads to a continuous reaction, longer measurements

would be needed.

Strength and stability of phases

Although ladle slag released the most heat of hydration, it is not sure that the sample

develops the highest final strength during curing.

The strength developing characteristics depend primarily on the mineral phases: they

contribute either to early strength (e.g. tricalcium silicate) due to higher heat of

hydration or to later strength (e.g. dicalcium silicate) due to less heat of hydration

(Dobrowolski, 1998).

It has been observed during sample handling that the cured specimen consisting of

100 % ladle slag collapsed when touched after three months. It had been stored in a

room together with the other samples. As the sample with 100 % EAF slag did not cure,

it had not formed a stable specimen after hydration. All specimens consisting of slag

mixtures were still stable after three months. The alumina content of ladle slag of

23 wt-% is compared to EAF slag relatively high. The instability of the ladle slag

specimen asserts the hypothesis of very low final strength of cementitious materialscontaining high aluminate (see Appendix II, chapter 4.1).

Aluminous cement shows a (compared to OPC) fast and intensive hydration reaction

with monocalcium aluminate being the main mineral phase. Low final strength of

aluminous cement is caused by mineral transformations of monocalcium aluminate

during curing. The minerals developed at the end of the hydration reaction (e.g.

dicalcium aluminate hydrate) have a different crystal structure and are less stable than

4 CaO/SiO2 ratio EAF slag = 0.9


35/117

DISCUSSION

19

monocalcium aluminate. The same hydration products could have been formed in ladle

slag and instead of reacting with silicates, calcium could have reacted with aluminate,

which in the following could have caused low final strength.

Another possible reason for the disintegration of the ladle slag specimen could be the

expansion of mineral phases. As been detected with XRD analyses, a substantial

amount of periclase is present in ladle slag. Periclase is known for expansion after

hydration. Also, the phase - Dicalcium silicate ( -Ca2SiO4) was measured. This

phase is formed of - Ca2SiO4 during cooling of the molten slag. The different phases

of dicalcium silicate all have a different crystal structure. During phase transformation,

the change in the crystal structure leads to a disintegration of the mineral. This also can

be a reason for the fine structure of ladle slag. However, no complete explanation of the

observation can be provided.

Mineral alterations

Even though no long-term behaviour has been measured in this study, some additional

observations after three months of storage could be made.

The colour of the three months old specimens was lighter than directly after thepreparation which could be caused by carbonation and calcite formation on the surface.

The formation of carbonates and hence the development of a high buffer capacity is one

assumption for mineral alterations in a liner consisting of steel slags.

The mineral alterations that are estimated for slags consisting of calcium silicates and

calcium aluminates also include the formation of C-S-H and C-A-H phases. These

phases have a good adsorption capacity for heavy metals due to their large reactionsurface (see Appendix II, chapter 4.3). Other types of so called cash minerals that could

hinder the release of contaminants are iron and aluminium phosphates/hydrates/hy-

droxides and silicates.

It is hard to draw conclusion from literature data of ashes and slags regarding mineral

alterations. How much of a material will be transformed and which secondary phases

are developed, has to be investigated for each type of material separately. Ashes and

slags can differ very much in their composition which is reflected also in the literaturewhere considerable variations of data for mineral alterations of MSWI ashes are

reported.


36/117

DISCUSSION

20

Application of steel slags in a liner

As ladle slag adds the reactive compounds to a mixture and addition of EAF slag leads

to higher long-term stability, both slag types have their own function and are necessary

in construction applications. The 50/50 mixture is most favoured in a liner application

because its hydraulic behaviour, curing time and stability showed to be the best for the

target function.

Principally, the chemical and mineralogical composition encourages the application of

steel slags in a top cover construction of a landfill. Through compaction, an increase in

buffer capacity will be achieved, as gas and water fluxes need more time to reach the

reactive surface of the material. Furthermore, construction materials containing

carbonates have advantages considering the expected chemical and mineralogical

changes.

Future experiments

Further analyses with regard to the stability of mineral phases, the formation of

secondary mineral phases and the long time behaviour need to be done. Humidity and

carbon dioxide influence these mineral alterations. Hypotheses about possible

secondary mineral phases have been outlined in the literature review. The expansion offree CaO and MgO due to late hydration is one problem of applying steel slags as a

construction material. If a construction material does not have sufficient volume

stability, the stability of the whole construction is endangered.

If longer measurements will be performed, much more efforts are necessary. A bigger

number of specimens (including parallels) has to be stored under defined conditions,

e.g. fixed CO2 content and humidity, and more IR and XRD measurements have to beperformed after different periods of sample storage. There are good reasons to assume

that the described mineral alterations can be detected with these two analyses methods.

Quantitative IR as well as XRD measurements should be considered. If possible, the

quantitative methods should be favoured. Suggested time spans for sample storage are

30 days, half a year and one year.

Furthermore, SEM analyses could complete the picture about the mineralogical slag

structure as phases below 4 wt-% and non crystalline material cannot be measured with

XRD. The presence of crystalline structures could be unveiled by this type of


37/117

DISCUSSION

21

microscopy. One could illustrate the topography of the surface. Simultaneously, data of

the crystal units are measured.

Models estimating the time for decomposition of ash layers consisting of carbonates by

carbonation result often in time spans of thousands of years (Ecke, 2003). Results of

models performed with steel slag compositions, layer thickness and climate data would

be interesting. Also, further investigations about the stability of the estimated secondary

phases under landfill conditions are necessary. At present thermo dynamical data and

secondary mineral phases are not available for all of the measured minerals.


38/117

CONCLUSIONS

22

5 CONCLUSIONS

Calorimetry, IR spectroscopy, XRD as well as sample handling and experiences with

preparing slag specimens provided the following findings.

In Europe steel slags are often landfilled or temporarily stored, although

applications for their reuse already exist (e.g. as additives in a road construction).

The target of nowadays research is to identify further applications for reusing steel

slags and to investigate the environmental impact that steel slags would have. The

conditions present in the liner have to be estimated before using alternative

construction materials as steel slags.

Up to today, the long-term behaviour is still unknown. It is affected by processes as

weathering, including carbonation (carbonic acid weathering) and pH changes,

leaching, aging of mineral phases and therewith stability changes of the minerals.

XRD and IR spectroscopy are suitable methods to analyse the mineral phases of

steel slags. The reaction of the slags under addition of water is called hydration

which has been described in this literature review. To determine the heat

development and conditions of this reaction, calorimetry is a suitable technique.

Ladle slag releases the most heat during hydration, while the heat release of EAF

slag is very little. The maximum heat output occurred the later the more ladle slag

was in the mixture.

The ladle slag specimen cured fastest. The EAF slag sample was not curing to a

rock-like compound and developed no mechanical strength.

A conclusion regarding the correlation of curing behaviour and mineral phases

formed is possible in the way that a high amount of certain mineral phases as

dicalcium silicates results in limited cementitious properties. Such mineral phases

have been recorded by XRD in both slag types.


39/117

CONCLUSIONS

23

The heat release during hydration does not necessarily correlate to the stability of

the hydration products, i.e. there is no correlation between heat of hydration and

developed final strength.

Ladle slag showed a higher reactivity than EAF slag, but probably has a lower final

strength, which is concluded from the high alumina content and the properties of

similar cement types. However, the investigation of the final strength after curing

was not part of the measurements.

A mixture of EAF slag and ladle slag is recommended for a landfill cover.

A sample preparation method was developed for mineral analyses, which worked

well and can be applied in further tests. With this method explained in Appendix I

the mineral phases after hydration could be measured by IR and XRD analyses.

IR spectroscopy revealed the different chemical bonds and compounds formed after

hydration. The IR spectra and the intensities of the peaks changed proportional to

the ladle slag content in the mixture. For example, the detection of hydrated waterwas the lower; the more ladle slag was present.

Both steel slags consist predominantly of crystalline phases. It is likely that solid

solutions are formed during slag cooling. However, these irregularities of the

crystalline phases cannot be detected with XRD. According to the XRD analyses,

no new minerals have been formed after mixing the two steel slag types.

Further research

Slag storage and how slag properties change with different storage methods or slag

treatments should be investigated more detailed. The stability of mineral phases present

in steel slags, e.g. merwinite, monticellite, mayenite and phases consisting of calcium

silicate, ferrite, magnesia and aluminate is mainly influenced by humidity, carbon

dioxide and pH changes. Humidity and carbonation are assumed to be correlated but

this has not been analysed for steel slags up to now.


40/117

CONCLUSIONS

24

Secondary mineral phases of the analysed phases are assumed to be carbonates and

bicarbonates. The formation of carbonates will be a short-term process, while

bicarbonates are formed over very long time periods (in the range of thousands of

years). The amount of minerals that will react and be transformed into secondary phases

is unknown. Compaction will increase the buffer capacity and contribute to longer

reaction times.

More knowledge of stability and long-term behaviour of mineral phases of steel slags

can be achieved with long-term experiments and by modelling and further thermo dyna-

mical data about the analysed minerals.


41/117

REFERENCES

25

6 REFERENCES

Andreas, L., Herrmann, I.; Lidstrom-Larsson, M.; Lagerkvist, A. (2005). Physical

properties of steel slag to be reused in a landfill cover,Sardinia 2005, Tenth

International Waste Management and Landfill Symposium, S. Margherita di

Pula, Cagliari, Italy; 3-7 October 2005.

Dobrowolski, J. A. (1998). Concrete Construction Handbook. 4th ed., Mc Graw-Hill

Inc., New York, p.1.15, 1.17 et seq.

Ecke, H. (2003). Sequestration of metals carbonated in municipal solid waste incine-

ration (MSWI) fly ash. Waste management vol. 23, p. 631- 640.

EU (2002). "Council Decision establishing criteria and procedures for the acceptance of

waste at landfills pursuant to Article 16 and Annex II of Directive

1999/31/EC." Document 14473 ENV 682.Council of the European Union,

Brussels.

Fllman, A-M. (1997). Charaterisation of Residues Release of contaminants from

slags and ashes. Doctoral Thesis Linkping University, Sweden, p. 9.

Gadsden, J. A. (1975). Infrared Spectra of Minerals and Related Inorganic Compounds.

London, Butterworth, p. 6et seq., 23et seq., 62.

Gomes, C. E. M.; Ferreira, O. P. (2005). Analyses of Microstructural Properties ofVA/VeoVA Copolymer Modified Cement Pastes. Polmeros: Cincia e

Tecnologia, vol. 15, n 3, p. 193-198.

Gnzler, H.; Gremlich, H-U. (2002). IR Spectroscopy - An Introduction. WILEY-VCH

Verlag GmbH, Weinheim, Germany, p. 176.


42/117

REFERENCES

26

Herrmann, I.; Lidstrm Larsson, M.; Andreas, L. (2005). Anvndning av stalslagg i

sluttckningen av Hagfors kommunala deponin. Delrapport, MIMER, Lule

Tekniska Universitet, Sweden.

Hirschmann, G. (1999). Langzeitverhalten von Schlacken aus der thermischen

Behandlung von Siedlungsabfllen. Fortschr. Ber. VDI Reihe 15 Nr. 220.

VDI Verlag, Dsseldorf, p. 153.

IUPAC (1997). Compendium of Chemical Terminology. 2nd Edition. Blackwell

Science. International Union of Pure and Applied Chemistry IUPAC.

Online version of IUPAC compendium:

http://www.chemsoc.org/cgi-shell/empower.exe?DB=goldbook

Knoblauch, H.; Schneider, U. (1992). Bauchemie. Werner-Verlag GmbH, Dsseldorf,

p. 126, 160.

Krenkler, K. (1980). Chemie des Bauwesens. Band 1: Anorganische Chemie. Springer-

Verlag, Berlin, Heidelberg, p. 405.

Motz, H. and Geiseler, J. (2001). Products of steel slags as an opportunity to save

natural resources. Waste Management, Vol. 21, p. 285-293.

Shen, H. et al. (2004). Physicochemical and minerological properties of stainless steel

slags oriented to metal recovery. Resources, Conservation and Recycling 40,

p.245-271.

Shi, C. (2002). Characteristics and cementitious properties of ladle slag fines from steel

production. Cement and Concrete Research vol. 32, p. 459-462.

Smith, B. C. (1996). Fundamentals of Fourier transform infrared spectroscopy. CRC

Press LLC, Boca Raton, Florida, USA, p. 30.

Zement-Taschenbuch (2000). Verein Deutscher Zementwerke e.V. (ed.). 49th ed. Verlag

Bau+Technik GmbH, Dsseldorf, Germany, p. 322.


43/117

APPENDIX I

APPENDIX I

SAMPLE PREPARATION AND MEASUREMENT DATA

1 CALORIMETRY.......................................................................................2

2 PREPARATIONS OF SPECIMENS FOR IR SPECTROSCOPY AND

XRD EXPERIMENTS..................................................................................4

2.1 Sample preparation............................................................................................4

2.2 Storage of specimens and grinding.................................................................... 6

2.3 Observations...................................................................................................... 73 IR SPECTROSCOPY ................................................................................ 7

4 X-RAY DIFFRACTION.......................................................................... 10

5 REFERENCES.........................................................................................12


44/117

APPENDIX I 2

1 CALORIMETRY

The EAF slag was already crushed to a size smaller than 20 mm at Uddeholm Tooling

AB. Sieving at 20 mm for ladle slag was planned, but a 20 mm sieve was not available

at the university laboratory. That is why EAF 1, EAF 2 and ladle slag (sample no. 5)

were sieved with a 19 mm sieve to remove the fraction > 19 mm. Furthermore, the

fraction < 8 mm of EAF slag 1 and 2 was removed by sieving. Each slag type was

homogenised and reduced. EAF slag 1 and 2 were mixed in equal shares to form the

EAF fraction. These slag fractions were used for all performed measurements. The

exact data of the mixtures measured in the calorimeter are presented in table 1.

Table 1. Sample data

EAF fraction0 (ladle

slag)35 50 65 100

in [%]

mass EAF slag in [g] - 10.44 15.75 19.54 30.32

mass ladle slag in [g] 30.04 19.48 15.72 10.54 -

mass of water inplastic container in [g]

10.27 10.04 10.47 10.57 10.52

mass of water in dewar

in [g]40.66 40.11 40.52 40.10 40.22

The recorded values measured by the calorimeter for each sample are presented in table

2 to 6.

Table 2. Recorded data of 100 % ladle slag sample

time T dT dT/g slag[hrs] [C] [C] [C/g]

0 21.25 0 0.005 22.95 1.70 0.0610 23.35 2.10 0.0715 24.75 3.50 0.1220 27.35 6.10 0.2021 27.4 6.15 0.2025 26.43 5.18 0.17

30 24.92 3.67 0.1235 23.65 2.40 0.08


45/117

APPENDIX I 3

Table 3. Recorded data of 35 % EAF slag sample

time T dT dT/g slag[hrs] [C] [C] [C/g]

0 22.94 0 02.5 23.35 0.41 0.015 23.94 1.00 0.037.5 24.2 1.26 0.0410 24.3 1.36 0.0515 25.25 2.31 0.0820 26.68 3.74 0.1325 26.08 3.14 0.1030 24.95 2.01 0.0735 24.05 1.11 0.04

Table 4. Recorded data of 50 % EAF slag sampletime T dT dT/g slag[hrs] [C] [C] [C/g]

0 21.98 0 0.005 22.83 0.85 0.0310 23.00 1.02 0.0315 24.74 2.76 0.0917.5 25.24 3.26 0.10

20 25.00 3.02 0.1025 24.07 2.09 0.0730 23.17 1.19 0.0435 22.58 0.60 0.02


0 21.88 0.00 0.00

5 22.65 0.77 0.0310 23.23 1.35 0.0415 24.19 2.31 0.0820 23.80 1.92 0.0625 23.07 1.19 0.0430 22.80 0.92 0.0335 22.55 0.67 0.02


46/117

APPENDIX I 4


0 21.45 0.00 0.005 21.72 0.27 0.0110 21.68 0.23 0.0115 21.64 0.19 0.0120 21.93 0.48 0.0225 22.12 0.67 0.0230 21.90 0.45 0.0135 21.70 0.25 0.01

2 PREPARATIONS OF SPECIMENS FOR IR SPECTROSCOPY AND

XRD EXPERIMENTS

Specimens for IR and XRD analyses had to be prepared in order to detect the chemical

and mineral structure of the hydrated samples. The used spectroscopic method was the

DRIFTS method which cannot be applied for wet samples, because the water absorbs

radiation and the spectra would be altered.

2.1 Sample preparation

The sieving of the original slag fractions has been explained above (see chapter 1,

Appendix I). They were further used for preparation of specimens. The different

mixtures shown in table 7 were formed and the corresponding mass of each sample

fraction was recorded. The mass of EAF and ladle slag needed for e.g. the specimen

consisting of the 100 % mixture differed because the density of EAF slag is less than for

ladle slag. The height of all specimens was about 4.7 cm.


47/117

APPENDIX I 5

Table 7. Mixtures for preparation of specimens

sample no.fraction of EAF

slagmass of EAF

slagmass of ladle

slagmass ofwater

in [%] in [g] in [g] in [g]

1 0 - 144.32 14.43

2 35 63.85 117.51 19.43

3 50 78.86 78.86 16.18

4 65 117.14 63.44 18.18

5 100 198.54 - 16.15

Each slag sample was separated in three equal parts, mixed with about one third of the

corresponding amount of water and compacted. The necessary knocks of one layer with

the proctor device were calculated according to the following.

Calculation

The sample compaction was carried out according to SS 02 71 09 (SIS, 1994) exceptthat a different proctor device (than the standard proctor device) was used. The number

of beats from the proctor device was calculated according to equation 1 below in order

to compact the samples with a certain energy. The energy applied was assumed to be

2.5 Nm/cm3 (according to SS 0271 09).

Table 8. Data of the cylinder

height [cm] 9.0

height of sample [cm] 4.7

diameter [cm] 5.0

sample volume [cm3] 92.28

Table 9. Data of modified Proctor device

height [m] 0.295

falling weight [kg] 2.099

diameter of knock area [cm] 3.0

energy applied [Nm/cm3] 2.5


48/117

APPENDIX I 6

The following formula was used to calculate the necessary knocks with the proctor

device for the requested energy input of 2.5 Nm/cm3.

Equation 1 for calculating energy input with proctor devise:

EV

hgmnn=

**** 21

Before getting the necessary knocks for obtaining the stated energy input, unit

conversion was done. The result was again multiplied with a factor of 0.82, because the

relation of cylinder diameter to diameter of knock area of the standard proctor device

(102mm/50mm) in SS 02 71 09 is bigger than the one of the used proctor devise

(50mm/30mm). Therefore 12.6 * 0.82 10 knocks with the bearable proctor device had

to be applied for specimen compaction.

2.2 Storage of specimens and grinding

The specimens were stored in a bucket with 100 % humidity to allow a full hydration

reaction over this time. However, the reaction will not continue if there is no humid

atmosphere. For a longer hydration reaction, a longer storage time with 100 % humidity

must be taken. For these measurements, not more time than one week of sample

preparation was available.

After two days in the bucket, the specimens were put in an exsiccator. This was done

because the sample material for the IR measurements had to be dry. After, five days of

curing in the exsiccator, the specimens were taken out.

One day before X-ray measurements, the specimens were pulverized. At first, some

material was scratched off the specimen with big pliers. This material was put in a

grinding apparatus. Before starting the grinding process, a magnet was held over the

n1 number of slag layers in specimen

n2 number of knocks with proctor devise

m mass of falling weight at Proctor devise[kg]

g acceleration due to gravity [m/s2)]

h height of falling weight before compaction [m]

V sample volume [cm3]

E applied energy [Nm/cm3]


49/117

APPENDIX I 7

sample and small iron particles were taken out. Then grinding started. It was also

possible to look inside the apparatus and adjust the grinder. Between several grinding

steps, the apparatus was opened and the magnetic fraction was removed again.

Otherwise, these particles could disturb the analyses. After grinding, the sample was put

in a little plastic bag.

2.3 Observations

It was observed from earlier compaction experiments that the 100 % EAF slag specimen

had a crumble consistency after curing. The specimens containing both EAF and ladle

slag exhibited higher strength after standing in room conditions for three months. The

100 % ladle slag specimen became less stable over this time. This indicates that EAF is

basically stabilizing the whole structure, e.g. as aggregate in cement. However, no

strength measurements were performed.

3 IR SPECTROSCOPY

The five slag samples were grinded with a mortar in a bowl a second time. The nonhydrated ladle slag and the KBr, which was used as a background material, were

grinded in that way as well.

For the measurements, the samples were diluted in KBr (3.5 % of sample). The material

was weighed and mixed with the KBr by using a little piece of paper. The material is

diluted because only a small sample amount (in the range of some mg) is needed. A

particle size of 2 m (Gadsden, 1975) is to be preferred. Afterwards, the KBr and the

slag sample are carefully put in the sample cup. As for the XRD measurements as well,shaking or pressing of material has to be avoided. Instead, a small spatula is pushed

over the plane of the sample cup. If no even surface was obtained, the sample cup has to

be filled again.

The sample cup (filled with about 20 to 25 mg of material) is than put in the specimen

holder and the data (e.g. sample name, scanning range) are given into the computer.

Before starting the measurement, waiting of five minutes was necessary to have stable

atmosphere conditions in the sample compartment. The measurement itself included

130 scans of the sample and took about 20 minutes. After each measurement, the

sample cup was cleaned with cotton wool.


50/117

APPENDIX I 8

The spectra of the five hydrated slag samples is shown in Fig. 1, whereas the

comparison between hydrated and non hydrated ladle slag performed by IR

spectroscopy is shown in Figure 2.

Figure 1. IR spectra for all hydrated slag samples (LS = ladle slag)

4000 3000 2000 1500 1000 500 400wave number/cm-1

K-M

3299 1462

983905

653590

523866

865

35263338

1660

1418

978 951

653

524

35281655

1415

867817 669

525

36223527

3328

1655

1418

972929

919867

818

524

3623352633181656

1422

970

929919

866

818 568523

100%EAF, 0% LS

65%EAF, 35% LS

50%EAF, 50% LS

35%EAF, 65% LS

0%EAF, 100% LS


51/117

APPENDIX I 9

Figure 2. IR spectra for hydrated ladle slag and non hydrated ladle slag

Figure 3. IR spectra of 100 % EAF slag and 100 % ladle slag (hydrated)

4000 3000 2000 1500 1000 500 400wave number/cm-1

K-M

3299 1462

983905

653590

523

866

36233526

33181656

1422

970929

919

866

818 568523

100%EAF, 0% LS

0%EAF, 100% LS

4000 3000 2000 1500 1000 500 400

K-M

362335263318

1656

1422

970929919

866

818

568

523

476

3300

1459

973928919

860

818576

521492473

hydrated ladle slag

non hydrated ladle slag


52/117

APPENDIX I 10

4 X-RAY DIFFRACTION

For X ray diffraction analysis just a small amount of the solid sample is needed. Before

putting the sample in the diffractometer, the sample was put in a small round deepening

of a thin plastic plate. The difficulty is to have an even sample surface but not to press

or shake the material in the plate. This would change the result, because the premise of

the analysis is that the crystals are randomly orientated over the surface. The flat plate

with the sample is than put on the circumference of an X-ray focusing circle. One

analysis took approximately 1.5 hours.

X-ray diffraction pattern of steelslags

10 20 30 40 50 60 70 80 90

2 - Theta - Scale

Intensity

100 % EAF

100 % Ladle

Figure 4. Diffraction pattern of 100 % ladle slag and 100 % EAF slag

100 % ladle slag

100 % EAF slag


53/117

Table 10.Main IR peaks of 100 % EAF slag, 50 % EAF slag and the two ladle slag samples including liteslag type and wave number Literature [cm- author corresponding

date of measurement vibration

(nh = non hydrated)

100%EAF (12.12.) 3300 3500-3400 Smith (1999) Silanol SiO-H stretch silica

100%EAF (21.12.) 3468 ~3440 O-H stretches comb

50%EAF (21.12.) 3528 3700-3200 wate

100%LS (12.12.) 3526 3440-3446 Gomes and Ferreira (2005) alum

100%LS_nh (21.12.) 3300 (ancl.1640-

100%EAF (12.12.) 1458 1436-1424 Gomes and Ferreira (2005) C-O stretch carb

100%EAF (21.12.) 1459 1510-1410 Smith (1999) C-O stretch inorg

50%EAF (21.12.) 1415 ~1463 Smith (1999) antisymm. CO3 carb

100%LS (12.12.) 1422 1520-1320 Gnzler and Williams (2001) stretching vibrat. in ino

100%LS_nh (21.12.) 1459

1

master's thesis_silvia diener.pdf

Documents