UNIVERSITY OF OULU P .O. Box 8000 F I -90014 UNIVERSITY OF OULU FINLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

University Lecturer Tuomo Glumoff

University Lecturer Santeri Palviainen

Senior research fellow Jari Juuti

Professor Olli Vuolteenaho

University Lecturer Veli-Matti Ulvinen

Planning Director Pertti Tikkanen

Professor Jari Juga

University Lecturer Anu Soikkeli

Professor Olli Vuolteenaho

Publications Editor Kirsti Nurkkala

ISBN 978-952-62-2451-0 (Paperback)ISBN 978-952-62-2452-7 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

OULU 2019

C 727

Elijah D. Adesanya

A CEMENTITIOUS BINDER FROM HIGH-ALUMINASLAG GENERATED IN THE STEELMAKING PROCESS

UNIVERSITY OF OULU GRADUATE SCHOOL;UNIVERSITY OF OULU,FACULTY OF TECHNOLOGY

C 727

AC

TAE

lijah D. A

desanya

C727etukansi.fm Page 1 Friday, November 8, 2019 2:27 PM

ACTA UNIVERS ITAT I S OULUENS I SC Te c h n i c a 7 2 7

ELIJAH D. ADESANYA

A CEMENTITIOUS BINDER FROM HIGH-ALUMINA SLAG GENERATED IN THE STEELMAKING PROCESS

Academic dissertation to be presented with the assent ofthe Doctoral Training Committee of Technology andNatural Sciences of the University of Oulu for publicdefence in the OP auditorium (L10), Linnanmaa, on 13December 2019, at 12 noon

UNIVERSITY OF OULU, OULU 2019

Copyright © 2019Acta Univ. Oul. C 727, 2019

Supervised byProfessor Mirja IllikainenDoctor Päivö KinnunenDoctor Katja Ohenoja

Reviewed byProfessor Juan Manuel Manso VillalaínAssociate Professor Guang Ye

ISBN 978-952-62-2451-0 (Paperback)ISBN 978-952-62-2452-7 (PDF)

ISSN 0355-3213 (Printed)ISSN 1796-2226 (Online)

Cover DesignRaimo Ahonen

JUVENES PRINTTAMPERE 2019

OpponentAssociate Professor Maria Chiara Bignozzi

Adesanya, Elijah D., A cementitious binder from high-alumina slag generated inthe steelmaking process. University of Oulu Graduate School; University of Oulu, Faculty of TechnologyActa Univ. Oul. C 727, 2019University of Oulu, P.O. Box 8000, FI-90014 University of Oulu, Finland

Abstract

About 4 Mt of ladle slag is generated in steelmaking processes in Europe per year, a largeproportion of which (80%) is placed in landfills or stored. This pattern is expected to continuewithout further research for their valorisation due to increasing demand for quality steel productsworldwide. Ladle slag (LS) produced in Finland possesses large amounts of calcium andaluminium and mineralogical phases which can exhibit cementitious capabilities and can beutilized in applications where expensive commercial cements are currently being used. The aim ofthis thesis is to investigate the properties of ladle slag in different activation pathways, includingalkali activation and use as a hydraulic binder with gypsum.

The results showed that ladle slag can be used alone as a precursor in alkali activation or as thesole binder or a co-binder with gypsum in hydraulic binding. Depending on the activationpathway, compressive strength between 35-92 MPa can be achieved after 28 days. The reactionproperties of alkali activated ladle slag are characterized, and it is confirmed through X-raydiffraction (XRD) that the reaction product after alkali activation is mainly an x-ray amorphous(calcium aluminate silicate hydrate-like) phase. Characterization techniques (SEM, XRD, TGAand NMR) used to analyze the LS paste binder with just water showed the hydration products ofladle slag to be dicalcium aluminate octahydrate (C2AH8), tricalcium aluminate hexahydrate(C3AH6), gibbsite (AH3) and stratlingite (C2ASH8) was also identified after a prolonged period ofhydration. Furthermore, it was found that to minimize the conversion, the ideal water-to-binderratio is 0.35. The conversion mechanism is reduced at this ratio and the strength is slightlyaffected. Another pathway that can be used to annul the conversion of calcium aluminate hydratesformed in LS paste is through the addition of gypsum to the LS paste system to produce anettringite-rich binder (C6AS3H32). When ettringite is formed in place of calcium aluminatehydrates the strength increases, frost resistance is improved, and drying shrinkage is enhanced.

Lastly, a potential application of ladle slag as a refractory material was also investigated.

Keywords: alkali activation, calcium aluminate, high-alumina slag, ladle slag, refractory

Adesanya, Elijah D., Sementtimäinen sideaine terästeollisuuden alumiinirikkaastakuonasta. Oulun yliopiston tutkijakoulu; Oulun yliopisto, Teknillinen tiedekuntaActa Univ. Oul. C 727, 2019Oulun yliopisto, PL 8000, 90014 Oulun yliopisto

Tiivistelmä

Euroopassa syntyy vuosittain noin 4 Mt terästeollisuden sivutuotetta, JV-kuonaa, josta 80% läji-tetään tai kaatopaikoitetaan. Maailmanlaajuisesti syntyvän kuonan määrä tulee todennäköisestikasvamaan laadukkaiden terästuotteiden ennustetun kysynnän kanssa. Tämän vuoksi kuonalletulisi löytää hyötökäyttökohde, jota vältyttäisiin läjitykseltä. JV-kuona sisältääkin suuria määriäkalsiumia ja alumiinia sekä mineralogisia faaseja, joilla on sementtimäisiä ominaisuuksia. Näinkuonaa voitaisiin käyttää sovelluksissa, joissa tällä hetkellä käytetään kalliita kaupallisia sement-tejä. Tämän väitöskirjan tarkoituksena oli tutkia JV-kuonan ominaisuuksia sementtimäisenäsideaineena alkali-aktivoinnissa sekä hydraulisena sideaineena yksinään että kipsin kanssasekoitettuna.

Väitöskirjan tulokset osoittivat, että JV-kuonaa voidaan käyttää prekursorina alkali-aktivoin-nissa tai hydraulisena sideaineena pelkästään veden kanssa tai yhdessä kipsin ja veden kanssa.Saavutetut puristuslujuuset vaihtelivat 35 ja 92 MPa:n välillä, jotka vastaavat normaalin ja eri-tyislujan betonin lujuuksia. JV-kuonan reaktiotuotteet alkali-aktivonnin jälkeen analysoitiinXRD- ja FTIR-analyyseillä. Tuloksista nähtiin, että alkali-aktivoinnin jälkeen reaktiotuote onsementin kaltainen kalsium-aluminatti-silikaati-hydraati (C-A-S-H) –tyyppinen faasi. XRD,SEM, TGA ja NMR –analyysit osoittivat JV-kuonan hydrataatiotuotteiden olevan erilaisia kalsi-um-aluminaattihydraatteja (C2AH8, C3AH6, AH3 ja C2ASH8). Tämän vuoksi työssä tutkittiin erivesi-kuona –suhteita, ja havaittiin, että kun käytetään alhaista kuona-vesi –suhdetta (0,35), reak-tiotuoteiden muutos vähenee ja lujuus paranee. Toinen tapa, jolla voidaan estää reaktiotuottei-den muuttuminen, on kipsin lisäys: lisäämällä kipsiä tuotetaan runsaasti ettringiittiä (C6AS3H32).Kun ettringiittiä muodostuu kalsium-aluminaattihydraattien sijaan, lujuus kasvaa, pakkaskestä-vyys paranee ja kuivumiskutistuma paranee.

Väitöskirjan viimeisessä osiossa tutkittiin JV-kuonan mahdollista käyttöä tulenkestävänämateriaalina ja huomattiin, että sen tulenkestävyysominaisuudet vaihtelevat käytetyn aktivointi-tyypin mukaan.

Asiasanat: alkali-aktivointi, alumiinirikas kuona, JV-kuona, kalsiumaluminaatti,tulenkestävä

Because he lives, I can face tomorrow… Dedicated to the loving memory of my Dad, and to my

Wife and Bezalel

8

9

Acknowledgements

The research work reported in this thesis was carried out in the Fibre and Particle

Engineering Research Unit at the University of Oulu in the course of 2015-2019.

The research formed part of the MINSI project, with funding and support from

various companies (SSAB, Stora Enso, Oulun Energia, Pohjolan Voima and

Ekokem) and of the FLOW project, funded by Business Finland in the ERA-MIN

2 Innovation Programme, which is part of the EU Horizon 2020 programme. The

author would like to thank the Auramo Foundation and the Oulu University

Scholarship Foundation for their financial support.

My sincere gratitude goes to my principal supervisor, Professor Mirja

Illikainen, for giving me the opportunity to undertake my master’s and PhD

research under her guidance, with great support from Päiviö Kinnunen, PhD,

and Katja Ohenoja, DSc (Tech), both of whom guided, co-supervised and

contributed to all the research undertaken in this thesis. These brilliant people

introduced me to this field of study and nurtured me in it from an intern in 2014 to

a scholar in 2019. Also greatly appreciate the support and collaboration received

from Juho Yliniemi, DSc (Tech), during these years, and I would similarly

like to express my appreciation to Prof. Yiannis Pontikes for the opportunity

to visit his research unit at Katholieke Universiteit Leuven and to both him and

Dr. Lubica Kriskova for their supervision and support during that visit in 2017.

I would also like to express gratitude to all my past and present colleagues in

the Fibre and Particle Engineering Research Unit who have helped to create a

conducive working environment during my research. Special thanks go to Jarno

Karvonen, Jani Österlund, Elisa Wirkkala, Maria Tomperi and Johannes Kaarre for

their help in the laboratory, and to Simo Isokääntä of SSAB for providing the slag

used in this work. Malcolm Hicks is thanked for revising and proofreading the

language of this thesis. Professor Juan Manuel Manso and Associate Professor

Guang Ye are sincerely acknowledged for their thorough reviewing and

constructive feedback of this manuscript. Would also like to appreciate the

guidance from my follow-up group, chaired by Prof. Timo Fabritius and supported

by Dr. Petteri Piltonen and Pekka Tanskanen.

Finally, I wish to express my deepest gratitude to my parents, Baba Ijebu and

Iya Ijebu who went extra miles to ensure that my siblings and I would get the best

education and living conditions. I would also like to appreciate the efforts of my

siblings, most especially my mentor Gbenga Adesanya, Deborah Oguntosin,

Kayode Adesanya, Comfort Adesanya, and my Pastor Adeolu Osho, for always

10

believing in me and for their support through and through. Infinite appreciation and

thanks are due to my beloved wife Adefunke Adesanya and our son Bezalel

Adesanya, for their encouragement, support, love and patience during my studies

and during the writing of this thesis.

Oulu, May 2019 Elijah Damilola Adesanya

11

Abbreviations

AALS Alkali-activated ladle slag

AH3 Gibbsite

ASTM American society for testing and materials

BFS Blast furnace slag

BOF Basic oxygen furnace

C(N)ASH Calcium aluminate silicate hydrate with sodium composition

C2AH8 Dicalcium aluminate hydrate

C3AH6 Tricalcium aluminate hydrate

CAC Calcium aluminate cement

CAH10 Calcium aluminate decahydrate

CASH Calcium aluminate silicate hydrate

CSH Calcium silicate hydrate

d10 Particle size for cumulative 10% finer (µm)

d50 Median particle size (µm)

d90 Particle size for cumulative 90% finer (µm)

EAF Electric arc furnace

LOI Loss-on-ignition

LS Ladle slag

NASH Sodium aluminate silicate hydrate

OPC Ordinary Portland cement

W/B Water-to-binder ratio

XRD X-Ray diffraction

XRF X-Ray fluorescence

Greek symbols

β-C2S Larnite

γ-C2S Calcio-olivine

Standard cement chemistry notations were used; hence C denotes CaO, A is Al2O3,

S is SO4 and H is H2O.

12

13

Original publications

This thesis is based on the following publications, which are referred to throughout

the text by their Roman numerals:

I Adesanya, E., Ohenoja, K., Kinnunen, P., & Illikainen, M. (2017). Alkali activation of ladle slag from steel-making process. Journal of Sustainable Metallurgy 3(2), 300-310

II Adesanya, E., Ohenoja, K., Kinnunen, P., & Illikainen, M. (2017). Properties and durability of alkali-activated ladle slag. Journal of Materials and Structures, 50:255.

III Adesanya, E., Sreenivasan, H., Kantola, A.M, Telkki, V.-V., Ohenoja, K., Kinnunen, P., & Illikainen, M. (2018). Ladle slag cement – Characterization of hydration and conversion. Construction and Building Materials, 193, 128-134.

IV Nguyen, H., Adesanya, E., Ohenoja, K., Kriskova, L., Pontikes, Y., Kinnunen, P., & Illikainen, M. (2019). Byproduct-based ettringite binder – A synergy between ladle slag and gypsum. Construction and Building Materials, 197, 143-151.

V Adesanya, E., Karhu, M., Ismailov, A., Ohenoja, K., Kinnunen, P., & Illikainen, M. Thermal behaviour of ladle slag mortars containing ferrochrome slag aggregates. Advances in Cement Research, In press.

The author’s contributions to the above Papers I-V were as follows:

The author designed the study with guidance from the supervisors and co-authors,

performed the experiment, analysed the data and wrote the paper, on which the co-

authors provided valuable comments, corrections and proofreading (Papers I & II).

The author designed the study with guidance from the co-authors, performed the

experiment, analysed the data and wrote the paper, on which the co-authors

provided valuable comments, corrections and proofreading. HS, AK and VT helped

with the nuclear magnetic resonance experiment (Paper III).

The author made a partial contribution to the experimental design together with the

co-authors and did most of the experimental work, co-analysis of the results and

writing of the paper as a co-author (Paper IV).

The author designed the study with guidance from the supervisors and the co-

authors, performed the experiment, analysed the data and wrote the paper, on which

the co-authors provided valuable comments, corrections and proofreading. AI

(Tampere University) conducted the dilatometry experiments and MK (Technical

Research Centre of Finland, VTT) helped with the preliminary high temperature

experiments (Paper V).

14

15

Contents

Abstract

Tiivistelmä

Acknowledgements 9 

Abbreviations 11 

Original publications 13 

Contents 15 

1  Introduction 17 

1.1  Background ............................................................................................. 17 

1.2  Approach and aims of the thesis ............................................................. 18 

1.3  Outline of the thesis ................................................................................ 19 

2  State of the art 21 

2.1  Steelmaking slags .................................................................................... 21 

2.1.1  Production of different types of steel slag .................................... 21 

2.1.2  Physico-chemical properties of ladle slag .................................... 22 

2.2  Slag-based and high-alumina cementitious binders ................................ 24 

2.2.1  Alkali-activated slags ................................................................... 25 

2.2.2  Hydration of crystalline slag ........................................................ 27 

2.2.3  Hydration of calcium aluminate cement ....................................... 28 

2.3  Current and potential applications for ladle slag ..................................... 30 

2.4  Remarks on literature review .................................................................. 32 

3  Materials and Methods 33 

3.1  Materials ................................................................................................. 33 

3.1.1  Binders ......................................................................................... 33 

3.1.2  Alkali activators ........................................................................... 34 

3.1.3  Aggregates .................................................................................... 35 

3.2  Methods ................................................................................................... 35 

3.2.1  Characterisation of materials ........................................................ 35 

3.2.2  Preparation of binder samples ...................................................... 36 

3.2.3  Microstructural analytical methods .............................................. 38 

3.2.4  Physical analyses performed on hardened materials .................... 39 

4  Results and Discussion 43 

4.1  Properties of ladle slag ............................................................................ 43 

4.1.1  Reactivity/hydraulic potential of ladle slag .................................. 43 

4.1.2  Mineralogy ................................................................................... 43 

4.2  Effect of alkali activation on the reaction kinetics of LS ........................ 44 

16

4.3  Identification of reaction products in AALS ........................................... 46 

4.4  Effect of aggregates on strength and durability of AALS ....................... 48 

4.4.1  Compressive strength ................................................................... 48 

4.4.2  Durability properties of AALS mortars ........................................ 49 

4.5  Hydraulic activity of LS .......................................................................... 51 

4.5.1  The effect of water-to-binder ratio on compressive

strength of LS (HLS) .................................................................... 51 

4.5.2  Characterization of the conversion of HLS hydration

products ........................................................................................ 53 

4.6  Modification of ladle slag hydration products ........................................ 56 

4.6.1  Hydration kinetics of modified ladle slag (in the presence

of gypsum - LSG) ......................................................................... 56 

4.6.2  Characterization of LSG hydration products ................................ 58 

4.6.3  Microstructure of LSG binder hydration products ....................... 62 

4.7  Strength and durability of HLS and LSG mortars. .................................. 63 

4.7.1  Effect of modification of ladle slag hydration products on

strength ......................................................................................... 63 

4.7.2  Effect of ettringite formation on mortars length change ............... 63 

4.7.3  Freeze-thaw of mortars ................................................................. 64 

4.8  Fire resistance of ladle slag mortars ........................................................ 65 

4.8.1  Residual mechanical strength ....................................................... 66 

4.8.2  Dilatometry of the thermal shrinkage and expansion of

mortars .......................................................................................... 67 

4.8.3  Quality of mortars after exposure to elevated temperatures ......... 68 

5  Summary and concluding remarks 71 

List of references 73 

Appendices 85 

Original publications 89 

17

1 Introduction

1.1 Background

The amount of slag produced by the steelmaking industry has increased constantly

over the last decade due to the worldwide demand for quality steels for construction

and allied purposes. During the production of 1 ton of steel, 4 tons of waste and by-

products are generated, including slags, dusts and sludges [1]. One of these slags is

high-alumina ladle slag (LS), also known as ladle furnace slag and falling or

reducing slag. LS is generated from the secondary steelmaking operations in which

molten steel from the electric arc furnace (EAF) is further refined to obtain high

grade steel. To achieve this desired quality in the steel, fluxes such as lime and

dolomites are added as reducing agents to prevent oxides from forming on the

surface of the molten steel and to absorb impurities which are later seen as residuals

in the slag [2].

At the same time, cement production is currently responsible for 7-9% of

global anthropogenic carbon emissions, coupled with the fact that there is an

increasing demand for cement worldwide [3], [4]. Various methods have been

extensively studied and proposed with the aim of finding environmental-friendly

and sustainable alternative binders to replace ordinary Portland cement (OPC).

Partial substitution and blending of OPC with supplementary cementitious

materials such as fly ash from coal burning and biomass incineration, slags from

iron and steelmaking processes, rice husk ash and other pozzolanic materials, have

been experimented with, for example [5]–[9], while another pathway for reducing

greenhouse gas emissions accruing from OPC production and utilizing industrial

waste and by-products that has been studied is the alkali activation method [10]–

[14]. In this case industrial side-streams including fly ash and slags are mixed with

alkaline activators such as sodium hydroxide (NaOH) and/or sodium silicate to

form a hardened matrix with properties comparable or superior to OPC [12], [15].

Four million tonnes of LS annually is estimated to be generated in Europe alone,

and about 80% of this total amount of regionally generated slag is either stored or

used a landfill material [16], [17]. The environmental footprint and economic cost

of landfilling makes this method unsustainable, and it has been designated under

the European Union (EU) Framework Directive on Waste and By-products as the

least acceptable method [18].

18

Unlike other iron and steel slags such as blast furnace slag (BFS), basic oxygen

furnace (BOF) slag and EAF slag, LS is generated as a fine powder due to self-

pulverization during the cooling process and cannot be used as a coarse aggregate

(>100µm) in concrete [19]. The utilization of both BOF and EAF steel slags as

cementitious materials and aggregates has been studied widely, but the possibility

of using LS as a sole cementitious binder has received less attention and few

publications exist on the topic. Various methods have been proposed for alternative

uses of LS, such as fillers in self-compacting concrete, use in masonry mortars, also

landfill covering on filtering beds and the stabilizing soft clayey soils [17], [20]–

[22], but in view of the chemical, mineralogical and morphological properties of

this slag, the potential as a high-value binder appears promising and it is this that

forms the topic of the present thesis.

1.2 Approach and aims of the thesis

This thesis sets out to explore the possibilities for utilizing LS as a cementitious

material. The specific objectives of the work were summed up in advance as

follows:

1. To investigate the reactivity of LS as the sole precursor in an alkali-activated

mortar by optimizing the recipe/mix design of the slag and alkaline activators

to achieve high compressive strength (Paper I).

2. To examine the durability properties of the alkali-activated ladle slag (AALS)

designed here by means of freeze-thaw and drying shrinkage tests (Paper II).

3. To determine the hydraulicity of LS and evaluate the effect of water content on

the conversion mechanism of the hydration products and the similarities of

hydration with this slag to that with calcium aluminate cement (CAC) (Paper

III).

4. To find an alternative method for the modification of the hydration products of

LS through the addition of gypsum and to assess the durability of the plain and

modified binders (Paper IV).

5. To evaluate the refractory properties of LS mortars prepared through alkali

activation and just water hydration (Paper V).

19

1.3 Outline of the thesis

This thesis is organized into five chapters. Chapter 1 gives a brief introduction to

the work and states the approach adopted and the aims of the research. Chapter 2

describes the production of slags during steelmaking processes and the properties

of LS. Chapter 3 presents the materials used in the research and the experimental

and characterization methods. Chapter 4 presents the results and a discussion of the

findings, while a summary of the research and the conclusions to be drawn from it

is presented in Chapter 5.

20

21

2 State of the art

2.1 Steelmaking slags

Steel slags, i.e. industrial by-products or waste from steelmaking processes, include

EAF and BOF slags (also known as converter slag and Linz-Donawitz slag,

respectively), named after the furnaces in which they are formed. In some steel

plants other secondary metallurgical slags are also generated while refining the

molten steel to produce high-quality steel. These secondary slags include LS,

desulfurization slag, vacuum oxygen decarburization and argon oxygen

decarburization slag. A flowchart of the processes involved in the generation of

slags is shown in Figure 1.

Fig. 1. Schematic and flow chart of steel and slags production (Reprinted with

permission © SSAB Oy).

2.1.1 Production of different types of steel slag

In the basic-oxygen furnace, scrap steel (10-20%) is recycled in the furnace and

molten iron (80-90%) from a blast furnace is then injected into the furnace on top

of the scrap. Oxygen is fed into the furnace while maintaining the temperature at

1600 ºC-1650 ºC to oxidize the impurities in the mix and to promote the combustion

of the carbon in the steel [2]. Furthermore, to remove undesirable chemicals from

the molten mix, a fluxing agent such as dolomite (MgCa(CO3)2) or lime (CaO) is

22

added. The impurities compound with the burnt fluxing agent to form BOF slags,

which remain floating on top of the molten steel.

The molten steel is separated from the slag and further refined in a secondary

steelmaking process (i.e. a ladle furnace) or sent for subsequent casting, while the

slag is sent to a pit for cooling. In an electric arc furnace the steel scrap is recycled

by melting at a high temperature to produce crude steel, after which fluxing agents

are introduced into the furnace by a similar mechanism to that used in a basic-oxide

furnace. The molten steel is then removed to a ladle furnace for further refining or

sent for subsequent casting, while the EAF slag is also separated and transported to

a pit.

Depending on the type or quality of steel required, both BOF and EAF molten

steels are further refined in a ladle furnace or else sent for casting. In a ladle furnace

the molten steel is deoxidized, de-sulphured and appropriately alloyed with fluxing

agents such as calcium aluminates or calcium fluoride, calcium oxides or

magnesium oxides and calcium silicates to lower the sulphur content and other

impurities in the molten steel [23], [24]. Deoxidation forms silica and alumina

oxides which are absorbed with the ladle furnace slag generated during the refining

process [2]. Due to these modification processes and the addition of various fluxing

agents to the ladle furnace, the compositions and properties of the generated LS are

quite different from those of both BOF and EAF slags [23], having a high alumina

content.

2.1.2 Physico-chemical properties of ladle slag

The properties of LS vary from one production line to another and are dependent

on conditions during and after steel production such as; raw material/scrap used,

types of fluxing agent or alloys added, furnace conditions and the cooling procedure

in the slag pit yard. The chemistry and mineralogy of LS, as collected from various

sources literature, are shown in Table 1.

The chemical composition of LS is typically CaO (30-70%), Al2O3 (15-40%),

SiO2 (5-25%) and MgO (5-8%) [25]–[28]. The high CaO in the slag forms different

compounds with Al2O3 and SiO2, and in some cases fails to react or remains as free

CaO. Other minor components include SO3, TiO2, Fe2O3 and MnO. Typically, LS

is predominantly crystalline, with major mineral phases identified in almost all LS

reviewed being mayenite (C12A7) and polymorphs of dicalcium silicate (C2S).

Other minor phases present can include tricalcium aluminate (C3A), dicalcium

23

ferrite (C2F), Portlandite (Ca(OH)2) and periclase (MgO). C2S is abundantly

present in LS, where it exists in four distinct polymorphs: α, α՛, γ and β-C2S.

During natural air-cooling of molten LS the slag crystal undergoes self-

pulverization due to phase transformation from α-C2S to β-C2S at 630 ºC, and at a

lower temperature (400-500 ºC), this β-C2S transforms to γ-C2S. This

transformation increases the volume of the slags between 10-12% due to disruption

of their crystal structure. Due to this self-pulverization the slags exist as a fine

powder and are therefore unsuitable for use as coarse aggregates [2], [19]. By

comparison, blast furnace molten slags, after being poured into the slag pit, are

initially cooled in air before water is sprayed on the partially air-cooled solidified

slag, thereby preventing the conversion of β-C2S to γ-C2S. Thus, the mineral

content is prevented from crystallizing but form glassy minerals which are

hydraulic compared with the crystalline minerals in steel slags [29], [30]. Water

cooling is rarely use, if at all, in the cooling of molten steel slags, because of its

higher viscosity. Hence water can be confined in the slag easily and cause an

explosion [31], [32]. This mechanism differentiates steel slags from BFS in general

and affects their hydration properties.

Additionally, some LS contains free lime (f-CaO) and/or free periclase (f-

MgO). The f-CaO and f-MgO can vary with different batches of steel produced and

type of fluxing agent used. The f-CaO and f-MgO content of steel slags increases

according to whether CaO or MgCa(CO3)2 is used [33]. f-CaO in the presence of

water quickly hydrates to form calcium hydroxide (or portlandite). This product

has a lower density than CaO and thus, changes the microstructure of the concrete

through volume expansion which is deleterious to the concrete [2]. On the other

hand, f-MgO hydrates slowly with water to form magnesium hydroxide, initiating

significant volume changes over a period of months or years. This volumetric factor

is problematic and limits the use of steel slags in construction structures. In the

standard recommendation for the use of fly ash in concrete (EN 450) a maximum

limit of 1.5% of free lime is stipulated. The free lime content of the LS used here

(Papers I-V), was below 1.5%, so that no soundness problem could be envisaged.

24

Table 1. Chemical compositions (wt.%) and mineralogy of ladle slag reported in the

literature.

References Source CaO

(%)

SiO2

(%)

Al2O3

(%)

MgO

(%)

Fetotal

(%)

f-CaO

(%)

Mineralogy

[34]

Canada 57-66 6-13 16-24 3-5 0.7-4 7-11 CH, MgO, C2S/C3S, CaO,

C12A7, C3MS2, C2AS and C2F

[24] Spain 50-58 12-20 4-19 7-12 1.6-3.3 3.5-19 C3MS2, C3A, β-C2S, γ-C2S,

CA4, MgO, C2F, Spinel, CH,

C12A7etc.

Papers I-V Finland 46-51 8-10 27-29 6-8 1-5 0-0.4 C12A7, γ-C2S, C3A and MgO

[21]

Spain 58.0 17.0 12.0 10.0 - 18-22 Spinel, C3MS2, MgO, β-C2S,

γ-C2S, CaSO4, CH, Alite and

Bregidite.

[35]

Sweden 40.0 5.0 32.5 5.9 4.8 - C12A7,C3A, FeO,

C20A13M3S3, β-C2S, CM�̅�2,

C2F and C2MS2

[36]–[38]

Italy 55-42 9-18 11- 23 4-9 2-7 - C12A7, MgO, γ-C2S,

gehlenite, CaCO3, β-C2S,

quartz

[39]

Greece 56 23-26 1.3-2.0 4.3-7 1.5-1.7 0.6-0.9 γ-C2S, CH, C2F and CaCO3

[40]

Italy 54.5 16.4 11.1 4.0 8.7 - γ-C2S, MgO, CaCO3, C12A7,

CaS, β-C2S and Mg CO3

[41]

Romania 49.6 14.7 25.5 7.8 1.1 - β-C2S, Anorthite, CaS and

α-Al2O3

[42] Taiwan 48.6 23.7 4.2 8.1 1.21 - -

2.2 Slag-based and high-alumina cementitious binders

Methods for utilizing slags as binders in cementitious systems have been studied

and have gained wide acceptance in civil engineering applications and scientific

research over the past decades. This section will provide a general overview of the

alkali activation of slags, the hydration of crystalline slags and CAC hydration. It

should be mentioned that quite a number of high-quality reviews, books and papers

on the state of the art with regard to alkali-activated slags and slag hydration have

been published, providing a further understanding of this research (e.g. [13], [43]–

[45]), and the hydration and mechanism action of CAC will be introduced in this

current review, too, as high-alumina LS has proved to have similar properties.

25

2.2.1 Alkali-activated slags

The principle of the alkaline activation of slags was identified in the 1940s, but

systematic research into the synthesis of the binder was not undertaken until the

early 1960s [46]. Until the early 2000s, most of the studies on the alkali activation

of slags had been focused on blast furnace slags, and it is only in the past two

decades that attention has turned significantly towards the use of slags from other

metallurgical processes for alkali activation. The number of publications with the

keywords “alkali-activated” and “slags” published in Scopus (Figure 2) shows that

research in this field has grown immensely in recent years.

Fig. 2. Numbers of publications per year with the keywords “alkali-activated” and “slags”

in the last three decades (total: 1024). The data were collected and analysed from the

Scopus database

The basic principle behind this type of cement involves mixing finely ground slag

with a suitable alkali activator, which subsequently hardens to form a concrete-like

structure. Alkali hydroxides or alkali silicates, or both together, are popular

activators that have been used successfully in recent studies. The strength of the

resulting matrix is dependent on the type and dosage of the activator, the fineness

of the slag, the water-to-slag ratio and in some cases the curing temperature [47].

26

The reaction mechanism and products for alkali-activated slags are not

comprehensively understood, but they are dependent on the mineralogy and

fineness of the slag, the pH and type of the alkaline activator, the additives added

and in some cases the curing conditions. Nevertheless, authors of studies dealing

with the alkali activation of slag have described the process as beginning with the

dissolution of Me-O (Me=Ca,Mg) and Si-O-Si bonds to form C-S-H-type reaction

products having a low Ca/Si ratio [48], [49]. Depending on the Al content and

reactivity of the initial slag material, various levels of participation of Al in the

reaction to form a C-A-S-H-type reaction product have been reported previously

[50], [51]. In Figure 3, a schematic representation of the process of alkaline

activation of a high-calcium precursor is given and the reaction products described,

modified from two literature sources [52], [53].

Fig. 3. Mechanism for the dissolving of calcio-aluminosilicate glass (Modified from [52]

and [53]).

The activation pattern in an alkali-activated system has generally been categorised

into one of three types:

Alkaline activation of high-calcium systems: (Na, K)2O-CaO-Al2O3-SiO2-H2O,

Model 1. This model explains the synthesis in high-calcium and silicon-rich

materials, as in the alkali-activation of slags. The main reaction product in this

system is a calcium aluminate silicate hydrate (C-A-S-H) gel. This gel differs from

the C-S-H gel in hydrated OPC, as it has a low Ca/Si ratio and a distinct structure

[54]. The C-A-S-H gel chains in alkaline cements are longer, with up to 13

tetrahedral units as compared with 3 to 5 for the C-S-H gel chains in OPC systems,

in addition to which the gel structure in the former includes aluminium, which

replaces silicon in the cross-linking positions [55]. The main reaction product can

also be written as C-(N)-A-S-H with significant quantities of sodium acting as

charge balancing units or absorbed onto the gel [36]–[38].

27

Activation of low-calcium systems under alkaline conditions: (Na, K)2O-Al2O3-

SiO2-H2O, Model 2. This model describes the synthesis of materials with major

occurrences of aluminium and silicon in their composition, such as metakaolin, fly

ash and other aluminosilicate materials with a low CaO content. The main reaction

product in this model is described as a three-dimensional inorganic alkaline

polymer of sodium aluminosilicate hydrate (N-A-S-H) gel [58]. This model of

synthesis is known as geopolymerization, and the reaction gel is termed a

geopolymer or inorganic polymer [57].

Hybrid or blended alkaline cement systems: Model 3. This third model is a

combination of the previous two. The materials in the synthesis are blends of

aluminosilicates and calcium aluminosilicates (i.e. OPC + BFS + fly ash or LS +

metakaolin), and the reaction products are a complex mix of cementitious gels

which may include a C-A-S-H gel which has sodium in its gel composition and a

high calcium N-A-S-H gel that can be denoted as (N,C)-A-S-H [59].

These reaction products are accompanied by other secondary reaction products,

depending on the type of activating solution used, i.e. monosulphoaluminate

(AFm)-type phases in NaOH-activated binders [60], [61] and AFm phase

strätlingite in alkaline silicate-activated binders [56], [62]. Other factors include the

MgO content of the starting material. Hydrotalcite has been reported in high-MgO

slags [48], [56], [63] and crystalline zeolites in low-MgO precursors (less than 5%)

[53], [64] and some carbonated phases. It is proposed here that the alkali activation

of LS follows Model 1, with some new reaction products that will be discussed

later.

2.2.2 Hydration of crystalline slag

The reactivity of steel slags varies significantly from one slag to another. This is

due to the differences in mineralogy and chemical composition between the raw

materials, furnace processes and cooling methods used. When cooled rapidly with

water, dicalcium silicate (β-C2S) in the slag is stabilized and not transformed to γ-C2S, which is predominant in steel slags. With the exception of LS, some EAF and

BOF when rapidly cooled tend to crystallize further [65], [66]. Hence, most steel

slags are termed crystalline slags because of predominant crystalline mineral

phases and lower glassy content.

Steel slags are generally termed cementitious, though with low activity due to

their mineralogical phases (i.e. γ-C2S), which have lower hydraulic activity than

those in OPC [67], and they therefore require activation in a high-pH environment

28

in order for some of the mineral phases to hydrate. Another factor limiting the use

of steel slags alone in concrete is the potential volumetric instability associated with

the f-CaO and/or f-MgO reaction in the slag [2], [23].

Nevertheless, when ground to fine particle sizes, these slags have shown to be

reactive and have also been used as a co-binder with OPC, fly ash or some other

hydraulic binder [68]–[70]. Other methods that have been used to increase the

reactivity and hydration of steel slags include thermal treatment followed by rapid

cooling [66], [71], [72]. In one previous study pulverized LS was reheated at 900 °C

and rapidly cooled with high-pressure air. It was reported that after this pre-

treatment the slag had a glass content of 92% compared with 6.5% in slowly-cooled

slag. Also, the f-CaO content was negligible after this treatment [71].

In another similar study [66] the authors used water jet cooling after re-melting

EAF and LS and reported an increase in the glass content of the slags. The LS used

in the present work was deficient in f-CaO, however, and pre-treatment methods

are beyond the scope of this study.

2.2.3 Hydration of calcium aluminate cement

LS containing high alumina and calcium bares certain similarities to CAC, mainly

since the elemental composition of CAC consists of calcium, alumina and iron,

with minor proportions of silicon and titanium [73]. Also, these cements are four

or five times more expensive than OPC and thus not an economic replacement for

it. They are nevertheless used in some special applications, either as the main binder

or as a co-binder with other materials (e.g. OPC, gypsum) in a hybrid mix [73].

Other CAC applications include:

– fast-setting concrete

– refractory applications

– resistance to chemical (acid) attacks resistance to impacts and abrasion.

The basic difference between OPC and CAC lies in the main mineralogical phase

that contributes to setting and hardening. The main mineral phases in OPC are

dicalcium and tricalcium silicates (C2S and C3S), which in contact with water forms

calcium silicate hydroxide (C-S-H) and calcium hydroxide (CH). Although, unlike

OPC, there is a wide fluctuation in the chemical composition and mineralogy of

CAC, its major mineralogical phases are calcium aluminate phases such as

monocalcium aluminate (CA), dodecacalcium hepta-aluminate (C12A7), also

known as mayenite, and tricalcium aluminate (C3A). When these phases are

29

hydrated, calcium aluminate hydrates are produced [74]. These calcium aluminate

hydrate formations are chemically quite different from OPC, as their hydrates

include CAH10, C2AH8, C3AH6 and poorly-crystallized AH3. Both CAH10 and

C2AH8 are metastable, in that they form temporarily and convert to a stable product

(C3AH6 and AH3, respectively) over a period of time, as shown in the following

equations [75], [76]:

For monocalcium aluminate: 𝐶𝐴 10𝐻 → 𝐶𝐴𝐻 , (1) 2𝐶𝐴 16𝐻 → 𝐶 𝐴𝐻 𝐴𝐻 . (2)

These products subsequently convert to the following stable phases:

2𝐶𝐴𝐻 → 𝐶 𝐴𝐻 𝐴𝐻 9𝐻 , (3) 3𝐶 𝐴𝐻 → 2𝐶 𝐴𝐻 𝐴𝐻 9𝐻 . (4)

Similar reaction equations can be written for C12A7 and C3A. In C12A7-rich CAC,

C2AH8 is favoured as the metastable hydrated phase and not CAH10 [19], [73], and

the equation can thus be written as:

For mayenite:

𝐶 𝐴 51𝐻 → 6𝐶 𝐴𝐻 𝐴𝐻 , (5)

which then converts to:

3𝐶 𝐴𝐻 → 2𝐶 𝐴𝐻 𝐴𝐻 9𝐻. (6)

For tricalcium aluminate:

2𝐶 𝐴 27𝐻 → 𝐶 𝐴𝐻 𝐶 𝐴𝐻 , (7) 𝐶 𝐴𝐻 𝐶 𝐴𝐻 → 𝐶 𝐴𝐻 . (8)

In the case of equations 7 and 8, C4AH19 loses part of its interlayer water at a relative

humidity less than 88%, to form C4AH13, which is detectable in dried samples as

used for X-ray diffraction (XRD) [76]. The initial product can therefore vary

depending on the curing conditions. According to Scrivener [73], there is a change

in the microstructural morphology as a consequence of a reduction in the total solid

volume and an increment in the porosity when conversion occurs in a calcium

aluminate-rich system. Water is also released, which is then available to hydrate the

unreacted phases. This phase evolution consequently reduces the strength of the

binder.

An understanding of these different hydrates and the mechanisms behind their

conversion is necessary in order to have form a basic background for the reactions

30

of high-alumina cementitious binders. LS is depicted in the ternary system that it

forms with the other cementitious binders in Figure 4.

Fig. 4. Compositions of cements/binders within the CaO-SiO2-Al2O3 system

2.3 Current and potential applications for ladle slag

Alternative cements have been designed from slags through hydration and alkali

activation, thus offering a viable and sustainable method to partially replacing OPC

and reducing the environmental footprint. These binders are also specifically

designed for special applications where the properties of OPC are not deemed

sufficient, e.g. in refractory concrete, rapid hardening, wastewater purification,

immobilization of heavy metals and contact with radioactive waste.

The mining and extraction of non-renewable natural materials for OPC

production is not a sustainable practise, and a partial alternative would be to use

industrial side-streams such as LS to substitute for OPC usage in structural niche

applications. The cements used in the construction industry are regulated and

required to fulfil certain continental or national specifications, which means that

31

the potential for using a sustainable alternative cement is then dependent on

satisfying the specifications and requirements (i.e. strength) for the structural

applications in which the OPC and other commercial cements would have been

used. The ASTM standard requirements and specifications for cements used in

structural applications are shown in Table 2. Although not limited to the

specifications in Table 2, there are other applications of these alternative binders

developed from industrial by-products. Due to its short reaction and setting times,

LS can be used in structural repairs, where immediate strength is required, in

underwater construction, in cold weather, on foundry floors, in underground

tunnelling and in mine backfilling. Additionally, its reactivity means that it can also

be used as a reactive component in blends with other materials. In some earlier

studies, LS was used as a co-binder with other materials e.g. metakaolin, fly ash

and bottom ash in alkali-activated materials [37], [38], [40], but the present work

(reported in Papers I, II and V), was concerned entirely with the alkali activation

and durability properties of LS as a sole binder.

Table 2. ASTM standard requirements for cement materials used in structural

applications [77].

Title of specification ASTM

designation

Type/grade Minimum compressive

strength (MPa)

Maximum water

absorption (%)

Structural clay load

bearing wall tiles

C34-03 LBX 9.6 16

LB 6.8 25

Building brick C62-10 SW 20.7 17

MW 17.2 22

NW 10.3 No limit

Solid masonry unit C126-99 VC 20.7 NA

HC 13.8 NA

Facing brick C216-07A SW 20.7 17

NW 17.2 22

In view of its high calcium content and low proportion of heavy metals, a certain

amount of LS produced during steelmaking is used in agriculture as a liming and

fertilizing agent to reduce soil acidity [78]–[80]. The calcium and magnesium in

slag are readily soluble when used instead of the naturally occurring materials used

for these purposes, i.e. dolomite and limestone [81]. Apart from increasing the pH

in acidic soils, LS can be used in combination with phosphorous and silicon to

increase the nutrients in soil [80] and similarly for soil stabilization, with results

32

reported after stabilization showing improvements in the bearing capacity of the

soil and properties comparable to those achieved with lime stabilization [82]. The

use of LS in construction work has also been suggested when designing materials

for paving roads, as it is suitable for use in a soil-cement mix or as a filler in asphalt

mixes [83], [84]. Another application of an LS binder is in masonry mortars,

together with OPC, sand and admixtures [21], [85]. When 30% LS was used as a

cementitious substitute for OPC, or as 25% aggregate substitute for sand, a 90-day

strength of 19.5 MPa and a longer setting time were achieved [85].

The present results have led to the development of LS-based high-performance

fibre-reinforced composites [86]–[89], while other potential uses of LS could be in

refractory applications. The calcium aluminate cements which are largely used in

these applications are four to five times more expensive than OPC, whereas LS

possesses similar chemical and mineralogical properties to some grades of CAC

and can potentially be an economical and sustainable alternative to it. The potential

for this is studied and discussed in Papers III and V.

2.4 Remarks on literature review

From the reviewed literatures, the variances in the chemical and mineralogical

properties of ladle slags produced and utilised, differs according to the country and

steelmaking plants processes. Thus, the LS reactivity and utilization pathways are

different, and care has to be taken in reproducibility of results. For example, LS

produced from steelmaking in Canada and Spain have high content (7-22%) of f-

CaO while the LS produced in Finland and Greece have lower (< 1%) content. Also,

the chemical composition and main reactive minerals varies similarly. Nevertheless,

the reviewed publications have successfully showed methods of utilizing this slag.

Here in this thesis, a high-value application of the LS produced in Finland is studied,

its similarities with CAC, conversion mechanism and various application pathways

suggested.

33

3 Materials and Methods

3.1 Materials

The materials used as binders in this work were LS, gypsum and calcium aluminate

cement. Standard sand and ferrochrome slag were used as aggregates, while sodium

hydroxide, potassium hydroxide, diatomaceous earth and sodium silicate were used

as alkaline activators. Deionised water was added during alkali activation when

necessary and was mainly used for the hydration of slag. The properties of the

various materials are described in the following subsections.

3.1.1 Binders

Four batches of slowly-cooled LS were provided by SSAB Europe Oy (Raahe,

Finland) for use in these experiments (Paper I-V). They were collected at different

times from the cooling pit, where they have been exposed to different natural

weather conditions. To enhance their reactivity, the as-received slags were milled

for two hrs using a 10 L roller ball mill (Germatec, Germany) with a grinding

medium filling ratio of 60% to achieve average particle sizes (d50) between 8-12

µm. The LS used in the work reported in Paper V comprised equal amounts of

milled and raw slag, in order to ensure optimal particle packing in the mortars.

The CAC (CALIGHT 40) used in this research (Paper V) was supplied by

CALTRA (Netherlands). The initial setting time as indicated by the producer is 70

min and a compressive strength of approximately 66 MPa is achieved after 24 h.

The chemical composition, free lime content and particle sizes of each milled LS

used and of the CAC (provides stated by the producer) are described in Table 3.

34

Table 3. Chemical composition (in wt.%), free lime content and particle sizes of LS and

CAC.

Oxides LS Paper 1 LS Paper II LS Paper III & IV LS Paper V CAC (Paper V)

CaO 48.40 46.30 50.96 45.77 37-39

SiO2 11.20 8.60 8.27 9.50 5-6.50

Al2O3 28.50 28.30 27.87 29.83 38-40.50

Fe2O3 3.10 5.00 1.13 0.83 15-18

Na2O 0.10 0.00 0.06 0.03 -

K2O 0.10 0.10 0.06 0.05 -

MgO 6.20 7.40 6.31 6.19 -

P2O5 0.00 0.00 0.02 0.00 -

TiO2 1.30 1.00 0.40 4.53 -

SO3 1.10 0.50 0.80 0.39 -

Cl < 0.01 0.00 0.05 0.04 -

f-CaO 0.0 0.0 0.4 0.0 -

<10% (µm) 1.3 1.3 0.9 1.2 0.84

<50% (µm) 9.1 8.0 7.9 12.3 6.2

<90% (µm) 33.4 34.6 46.5 64.9 29.8

The gypsum (CaO-41%, SO3-53.8%) used as a calcium sulphate source was

procured from VWR Finland (product code 22451.360) (Paper IV). Its median

particle size (d50) was 11.5 µm.

3.1.2 Alkali activators

Potassium hydroxide (KOH) pellets, sodium silicate (Na-Sil), diatomaceous earth

(DE) and deionised water were used as alkaline activators in Papers I and II, and

sodium hydroxide (NaOH) pellets, Na-Sil solution and deionised water were used

in Paper V. To retard the reaction of the matrix, 0.8-1.0% (wt.% of water) citric acid

(Tokyo Chemical Industry Co., Ltd., Japan) was added to the solution/water as a

retardant (Papers III, IV and V).

The KOH, NaOH pellets and Na-Sil were supplied by Merck KGaA chemicals, as

was the Na-Sil solution with SiO2/Na2O=3.5 and 65.5% water.

Since the dissolving of alkaline hydroxide in water is an exothermic process,

all the alkaline solutions were prepared by stirring in a plastic container a day

before experiment to ensure equilibration and cooling of the solution.

35

3.1.3 Aggregates

CEN standard sand supplied by Normensand GmbH and conforming to EN 196-1

[90] requirements was used for the aggregates (Papers II, IV and V). Ferrochrome

slag (Fs) supplied by Outokumpu Chrome Oy (Tornio, Finland) was used as

aggregate in Paper V. The as-received Fs was dried in an oven at 100 °C for 24 h

and then sieved through a 2 mm mesh to remove slag with a > 2 mm particle size,

so that the particle size distribution after sieving would range between 0 and 2 mm.

Fs was used as an alternative aggregate to sand due to its particle structure and it is

produced as an industrial residue during stainless steel production. Thereby

enhancing the sustainability of the mortars. The chemical composition of the as-

received slag is shown in Table 4.

Table 4. Chemical composition of as-received ferrochrome slag

Oxides CaO SiO2 Al2O3 Fe2O3 Na2O K2O MgO Cr2O3 SO3 TiO2

1.8 28.9 25.1 7.5 0.14 0.22 22.8 14.9 0.2 0.5

3.2 Methods

3.2.1 Characterisation of materials

Particle size distribution

The particle size distributions of all the materials were determined using a Beckman

Coulter LS 13320 device. The average volumetric sizes of three repetitions are

reported (d10, d50 and d90).

Chemical composition

The chemical compositions of all the materials used in this research (except CAC)

were determined from X-ray fluorescence (XRF) analyses of 1.5-g samples of the

material melted at 1150 ºC with 7.5 g of X-ray Flux Type 66:34 (66% Li2B4O7 and

34% LiBO2). The elemental concentrations in the material were then determined

by means of an inductively coupled, plasma-optical emission spectrometer (ICP-

OES Thermo Electron IRIS Intrepid II XDL Duo, Thermo Scientific). The

36

materials were pre-treated by microwave-assisted wet digestion with a 3:1 mixture

of HNO3 and HCL for 0.5 g of sample.

X-ray diffraction measurement

The mineralogical compositions of the samples were determined by X-ray

diffraction (XRD) using a Rigaku SmartLab (Siemens D500 AG, Germany) 9 kW

X-ray diffraction device equipped with Cu-kα radiation (Kα1 = 1.78892 Å;

Kα2 = 1.79278 Å; Kα1/Kα2 = 0.5) at a scan rate of 3 º/min. Phase identification

was carried out using PANalytical X’pert HighScore Plus software and the samples

were quantified by Rietveld quantitative phase analysis was performed using 10%

titanium oxide as an internal standard.

Scanning electron microscopy

Scanning electron microscopy (SEM) was performed to analyse the

microstructures of the samples in both powdered and fractured form. Two

microscopes were used: a Zeiss Sigma microscope with a voltage of 5-15 kV

(Papers III and V) and a Zeiss Ultra Plus instrument fitted with energy dispersive

x-ray spectroscopy and having an accelerating voltage of 15 kV.

3.2.2 Preparation of binder samples

Alkali-activated samples

The alkali activation of LS in Papers I and II was carried out according to the

methods described in EN 196-1 [90, p. 196]. Milled LS and the already prepared

alkali activator were mixed to a homogenous paste in a Kenwood mixer. A sand-

to-binder ratio of 2:3 (represented as SS3) was used in the mortar sample and the

equivalent sand content was then added to the paste and mixed further. The sample

was then jolted to remove entrapped bubbles and to ensure matrix uniformity, after

which the sample was cast in rectangular moulds (20 x 20 x 80 mm3, for the strength

test), sealed in a plastic bag and cured at 60 ºC for 24h before being removed from

the moulds and cured further under moist conditions at room temperature until the

testing time. The samples for the shrinkage and freeze-thaw measurements were

cast in 40 x 40 x 160 mm3 moulds, cured at 60 ºC for 24 h and removed from the

37

moulds after a further 24 h and placed in a humidity chamber (50% RH) at 23 ºC.

The molar ratios of the mix of binder and activators were: SiO2/(Na2O+K2O) = 4.5

mol, CaO/SiO2 = 1.4 mol and H2O/(Na2O+K2O) = 21.3 mol. The water-to-cement

ratio was kept constant at 0.40.

Hydration of ladle slag/gypsum

The LS paste samples were prepared by mixing milled LS with varying amounts of

deionised water (water-to-cement ratios: 0.35 and 0.50). The deionised water

contained 1 wt.% of citric acid added to retard the fast setting of the slag during

hydration. The mixture was then stirred using high-shear mixer for four minutes

to achieve a uniform composition, and the samples were then cast in rectangular

stainless-steel prisms (20 x 20 x 80 mm3), sealed and stored at room temperature

before removal from the moulds after 24 h. Thereafter, the samples were kept under

moist conditions at 95% ± 3 RH and 22 ± 3 °C until the testing time (Paper III).

To modify the hydration product, 30% of the LS was replaced with gypsum using

a water-to-cement ratio of 0.45. Samples were then prepared for both neat LS mortars

and LS-gypsum mortars. Both LS and gypsum were initially mixed for a minute and

the subsequent mixing was carried out according to the methods described in EN-196-

1 [90, p. 196]. The sand-to-cement ratio was kept constant at 3:1. The samples for the

strength test were then cast in rectangular stainless-steel prisms (40 x 40 x 160 mm3),

sealed and stored at room temperature before removal from the moulds after 24 h.

For the shrinkage measurements, the matrix was cast in 40 x 40 x 160 mm3 moulds,

cured at room temperature and removed from the moulds after 24 h and placed in

a humidity chamber (50% RH) at 23 ºC. For comparison, neat LS hydration with w/b

of 0.45 was also performed (Paper IV).

Refractory mortar from ladle and ferrochrome slags

The mortars were prepared by mixing the CAC or slag with demineralized water in

a Kenwood mixer according to the EN 196-1 standard. For the alkali-activated slag,

sodium silicate and analytical-grade sodium hydroxide pellets were used as alkali

activators. The compositions of the mixes are shown in Table 5. The solution was

prepared 24 h before mixing, and 1% citric acid was added to the water to retard

the fast setting of the slag. The molar ratios of the mix of binder and activators were:

SiO2/(Na2O+K2O) = 2.0 mol, CaO/SiO2 = 2.2 mol and H2O/(Na2O+K2O) = 10.1

mol. The SiO2/(Na2O+K2O) here was reduced to avoid elevated temperature (110-

38

300 °C) degradation of silica containing reaction products. To achieve a good

packing in the slag mortar, the LS used in the mix contained equal amounts of

milled slag and as-received slag, optimized using a modification of Andreassen’s

packing model with a distribution coefficient (q) of 0.20, a maximum particle size

of 2 mm, and a slag density of 3.01 g/cm3. The water-to-cement (w/c) ratio for all

the mortars was kept at 0.35. All the samples tested here were cured at room

temperature in sealed bags (Paper V).

Table 5. Mix composition of mortars used in refractory investigation (Reprinted by

permission from Paper V © ICE publishing).

Sample name LS (g) Na-Sil (g) NaOH (g) w/c CAC (g) sand (g) FS (g)

AALS-Ss 500 246 9 0.35 - 500±5 -

HLS-Ss 500 - - 0.35 - 500±5 -

CAC-Ss - - - 0.35 500 500±5 -

AALS-Fs 500 246 9 0.35 - - 500±5

HLS-Fs 500 - - 0.35 - - 500±5

CAC-Fs - - - 0.35 500 - 500±5

3.2.3 Microstructural analytical methods

Isothermal calorimetry

Calorimetry was performed at room temperature (20 °C) to analyse the heat flow

in the samples (Paper IV) early in the hydration times. A TAM Air isothermal

calorimeter was used, and the samples were mixed in situ using an admix ampoule.

Thermogravimetry analysis

To verify the reaction and hydration products of the matrix, thermogravimetric

analyses were performed on the paste samples using the Precisa Gravimetrics

“prepASH automatic drying and ashing system”. The samples were heated from

room temperature to 1000 ºC in an inert nitrogen atmosphere.

Nuclear magnetic resonance - 27Al NMR

Nuclear magnetic resonance (NMR - 27Al NMR spectra) was used to verify the

conversion mechanism (Paper III) by means of a Bruker Avance III 300

39

spectrometer operating at 78.24 MHz for 27Al. For the MAS experiments the

samples were packed into 7 mm zirconia rotors, a rotation frequency of 7 kHz was

applied, and 2048 scans were collected at a repetition rate of two seconds. The

chemical shifts were referenced to Al(NO3)3 set to zero ppm.

Fourier transform analysis

The samples were characterized using the diffuse reflectance infrared Fourier

transform (DRIFT) method (Paper I). Spectra in the 400–4000 cm−1 range were

collected by means of a Bruker Vertex 80v spectrometer, taking 40 scans at a

resolution of 1 cm−1 for each sample.

Scanning electron microscopy

SEM observation was done for the paste samples on Zeiss Sigma using a secondary

electron detector with a voltage of 5–15 kV. The broken paste samples were coated

with platinum before analysis.

Free calcium quantification

The free calcium oxide content of the LS used here was determined according to

the recommendations laid down in the EN 451-1 standard.

3.2.4 Physical analyses performed on hardened materials

Strength tests

The compressive strength tests were performed using a Zwick testing machine with

a maximum load of 100 kN and a loading force of 2.4 kN/s for Papers II, III, IV

and V and an Instron 8500 testing machine with a maximum load of 200 kN and a

loading force of 2.4 kN/s for Paper I.

Freeze-thaw durability test

The freeze-thaw (F-T) resistance of the mortars was assessed by immersing half of

the samples in water and the other half in air. The chamber temperature was then

40

varied from -20 ºC to +15 ºC, with each cycle consisting of lowering the

temperature from 15 ºC to -20 ºC for 2 h and then maintaining it for another 2 h

and vice versa. The freeze-thaw experiment was started when the samples were

aged 35 days in Paper II and 28 days in Paper IV.

Ultrasonic pulse velocity (UPV) equipment (Matest, Italy) was used to analyse

the damage in terms of structural uniformity after the freeze-thaw routine (Papers

II and IV) and after exposure of the specimens to elevated temperatures (Paper V).

The velocity (Vp) and relative dynamic modulus of elasticity (Rd) were then

calculated using equations (9) and (10), respectively.

𝑉 , (9)

where d is the distance between the two transducers and t is the time taken for

the pulse to reach the receiver.

𝑅 𝑉 𝑉 , (10)

where Vpn is the velocity after n cycles and Vp0 is the velocity at cycle 0 [91].

Drying shrinkage

For the shrinkage measurements, the mortars with dimension 40 x 40 x 160 mm3,

stored in a humidity-controlled room (50% RH) at 23 ºC were used. The drying

shrinkage properties of the mortars were measured using Matest length comparator

equipment. The length change (L) over time was calculated using equation (11):

L 100 (11)

where Li is the initial reading after 24 h of preparing the sample, Lx is the length

of the sample at each curing age and G is the nominal effective length.

Refractory testing

The mortars were subjected to elevated temperature treatment (Paper V) by first

drying the samples at 105 ºC for 18 h to remove the capillary water and reduce the

spalling tendency. The mortars were then exposed to thermal loads of 200, 400,

600, 800, and 1000 °C at a heating rate of 6 °C/min. A dwelling time of 1h at each

temperature was used and after the last of these the samples were removed from

the furnace and allowed to cool naturally. The heating regime is shown in Figure 5.

The residual compressive strength and mortar quality using UPV were tested after

thermal exposure.

41

Fig. 5. The heating regime applied to the mortars (Reprinted by permission from Paper

V © ICE publishing).

Dilatometry

Dilatometry (Paper V) was carried out using a NETZSCH DIL 402 Expedis dilatometer, with samples approximately 8×8×10 mm3 in size. The samples were heated to 1000 ºC at 6 ºC/min and there was a one-hour dwelling time at the peak temperature, after which the furnace was set to a cooling rate of 10 °C/min for the final stage. The heating chamber was open-ended, with a constant flow of nitrogen (40 ml/min) as the purge gas to prevent unwanted gaseous/evaporated matter from entering the measurement chamber, which was separate from the heating chamber and sample holder. The in-situ volume expansion and/or shrinkage during heating was then analysed.

42

43

4 Results and Discussion

4.1 Properties of ladle slag

4.1.1 Reactivity/hydraulic potential of ladle slag

Slags are categorized into three groups according to their basicity coefficient (Mb):

acidic (Mb < 1), basic (Mb >1) and neutral (Mb = 1). Using the equation [Mb =

(CaO + MgO)/(SiO2+Al2O3)] it is clear from their chemical compositions that the

basicity coefficient of all the LS samples used in this study (see Table 3) ranges

between 1.3-1.6, and thus the slag can be classified as basic and potentially

hydraulic [57], [92].

4.1.2 Mineralogy

Quantitative X-ray diffraction (QXRD) was used to determine the amorphous and

crystalline content of the various LS samples, as shown in Table 6. Due to the slow

cooling of LS, the amorphous content of the slag is low and below the

recommended glass phase content of 30-90% reported to be essential for slag

reactivity during alkali activation [92]–[94]. Although the vitreous content of some

slags may play a role in their reactivity, it has been shown in new studies of highly

crystalline slags that there is no general correlation between the glassy phase slags

in their reactivity or hydraulicity [48], [65], [95], [96]. This claim will be

additionally verified in the later part of this section. It suffices here to state, however,

that LS consists mineralogically of mayenite (C12A7), calcio-olivine (γ-C2S),

periclase (MgO), and tricalcium aluminate (C3A), as shown in Figure 6.

44

Fig. 6. XRD pattern for LS showing the mineralogy phases: a=mayenite/C12A7, b=calcio-

olivine/γ-C2S, c=periclase/MgO, and d=tricalcium aluminate/C3A.

Table 6. Quantitative mineralogy of ladle slag.

Phase C12A7 γ – C2S C3A MgO Amorphous

Wt.% 26-35 18-22 11-19 7-13 15-22

Since the free lime content of the slag determined using the method described in

EN 450-1 [97, pp. 450–1] varied between 0.0 and 0.4% (see Table 3) of the overall

CaO content of the LS, it may be concluded that the potential volume instability

associated with steel slags is not applicable to the LS used here.

4.2 Effect of alkali activation on the reaction kinetics of LS

Alkali activation of LS results in rapid, intense heat generation, as shown in Figure

7. An exothermic peak is recorded immediately after mixing of the alkaline

activator with the slag and can be attributed to the wetting and initial dissolving of

the slag particles. This peak is similar to the well-known pre-induction period that

occurs in cement hydration, although the physicochemical reactions involved are

markedly different. The breadth of this peak may suggest initial formation of

45

reaction products immediately after adding the activator to the slag. The addition

of citric acid had no significant effect on the reactions, as the induction period was

very short, and further reactions of the binder continued after 3.5 hours. The rapid,

intense heat released in this period is mainly due to the fast-setting properties of the

mayenite phase in the slag.

Furthermore, the acceleration peak was followed by a more intense peak with

a maximum heat release of approximately 16 J/h·g after 5 hours of the reaction,

this peak being attributed to a second formation and precipitation of reaction

products such as C-A-S-H [98] and calcium aluminate hydrates in the activated

slag [99]. The cumulative heat released after 50 hours of reaction was about 125

J/g, which is lower than that observed after 50 hours in hydrated OPC, around 225

to 280 kJ/kg, and similar to the cumulative heat released in alkali-activated blast

furnace slag [100]–[102]. This calorimetry is representative of the alkali-activated

LS with no added silica source (i.e. diatomaceous silica) used in this study (Paper

V).

Fig. 7. Heat evolution in the first 50 hours of the alkali activation of LS. The first 10

hours of the reaction are shown in the inset.

46

4.3 Identification of reaction products in AALS

The dissolving of the vitreous and crystalline phases of LS resulted in formation of

x-ray amorphous reaction product in the AALS (Figure 8) determined after 28 days

of reaction. The broad hump featured between 24° 2θ and 30° 2θ after alkali

activation is consistent with the formation of amorphous phase of aluminosilicates,

typically a C-A-S-H and N-A-S-H or hybrid C-(N)-A-S-H gel, as previously

reported in the literature [103], [104]. A similar amorphous hump can be observed

with an increase in the silica content of the matrix, but it is broader and more intense

and may indicate an increase in the formation of the amorphous phase. No other

new crystalline phases were detected in the AALS diffractograms. The FTIR

spectra in Figure 9 supplement the observations of the formation of a new

amorphous gel gained from the XRD findings.

47

Fig. 8. X-ray diffractograms of raw LS, AALS (Ca/Si=1.6) and AALS (Ca/Si=1.2). a=

mayenite (C12A7) b= dicalcium silicate (ɣ-C2S) and c= periclase (MgO) (Modified from

Paper I).

The IR spectra in Figure 9 represent the bands before and after alkali activation of

the LS. The spectra of the alkali activated LS exhibit a stronger band at 1650 cm-1,

which is assigned to the stretching vibration (-OH) consistent with the presence of

a reaction product [104], [105]. This band is not intense in raw LS. Meanwhile, the

broad band at 1030 cm-1 may be attributed to asymmetric stretching vibration in the

T-O-Si (T=Al or Si) bonds [106]–[108] and shifts to a higher wavenumber after

alkali activation and with increasing silica content, which is consistent with the

formation of polymerized units in an amorphous structure [107], [109]–[111] and

may indicate reduced substitution of Si for Al in the C-A-S-H gel with increasing

silica content [105], [107].

48

Fig. 9. FTIR spectra for raw LS and AALS paste (data from Paper I).

4.4 Effect of aggregates on strength and durability of AALS

4.4.1 Compressive strength

The aggregate content of the AALS mortars had a significant effect on their strength

and durability, as studied in Paper II. As shown in Figure 10, the mortar sample

(SS3) had a greater strength at all ages than the paste samples (SS0). Due to the

low water content, high viscosity and low workability of the mix, the standard sand-

to-binder ratio of 3:1 was unattainable. Hence for SS3, the sand-to-binder ratio used

was 2.3. As observed at 28 days, both samples suffered strength loss due to changes

in the reaction product. This mechanism, known as “metastable hydrate

conversion”, will be discussed further in the next section (see subsection 4.5). In

addition, both samples continued to gain strength after 28 days of conversion, the

effect increasing until the final testing time at 90 days.

49

Fig. 10. Compressive strength of AALS paste and mortar at different curing days

(Reprinted by permission from Paper II © 2017 RILEM).

4.4.2 Durability properties of AALS mortars

The AALS mortar also showed good durability properties when subjected to

normal and extreme environments. The frost resistance test results after 60 cycles

of extreme freezing and thawing are shown in Table 6. The mortar sample exhibited

good resistance to these conditions with a minimal effect on its physical appearance,

whereas the paste sample showed serious failure after 35 cycles due to high water

retention in the microstructure of the paste compared with the mortar sample. Thus,

the superior properties during frost resistance test of SS3 over SS0 is attributed to

the sand aggregate in the mortar having low water absorption and retention

compared with the paste sample. The residual data on the samples are shown in

Figure 11.

50

Table 7. Residual data of the samples after 60 F-T cycles (Reprinted by permission from

Paper II © 2017 RILEM).

Sample name

Surface

appearance

Compressive strength (MPa)

Weight loss (%) Reference After 60 cycles Reduction (%)

SS0-paste Damaged 56 ± 1.1 Failed 100 -

SS3-mortar Good 50 ± 0.6 49 ± 1.3 0.96 1.71

Fig. 11. Appearance of the mortars after F-T (Reprinted by permission from Paper II ©

2017 RILEM).

The drying shrinkage of the AALS samples was also determined, and the length

changes are reported in Figure 12. Both samples experienced a marked length

change in the early days after activation and stabilized after 28 days of analysis, but

the AALS mortar (SS3) shrank 7 times more than did the OPC sample (with s/b:3),

while the paste sample shrank significantly more than did the mortar sample. These

findings are consistent with a previous report on the effect of aggregates on the

shrinkage of alkali-activated materials [112]. Previous studies have similarly

shown that alkali-activated slags shrinks between 3 and 6 times more than OPC

[113]–[115]. The reduction in the sand content of the AALS mortars in the present

study most likely caused the drying shrinkage recorded here to exceed this range

reported for alkali activated mortars (s/b:3) in the literatures.

51

Fig. 12. Drying shrinkage of AALS samples by comparison with OPC (data from Paper

II).

4.5 Hydraulic activity of LS

The results presented in Papers III and IV offer a low-cost alternative to the use of

LS without alkaline activators. Compared with other steel slags and blast furnace

slag, the LS used in these experiments contained large amounts of the calcium

aluminate phases C12A7 and C3A (37-54%), suggesting that a high mayenite content

in the mineral system may be of specific relevance for the hydraulicity potential of

LS. Mayenite is a fast-setting mineral phase, especially when in contact with water

[116].

4.5.1 The effect of water-to-binder ratio on compressive strength of

LS (HLS)

As expected, the HLS paste sample with a w/b of 0.35 (HLS/0.35) had a higher

compressive strength than that with a w/b of 0.50 (HLS/0.5) throughout strength

test, the maximum strength achieved for HLS/0.35 being about 43 MPa, after 7

days. Overall, however, a reduction in strength was observed during curing, as seen

52

in Figure 13, this being similar to the reduction in strength observed earlier in the

AALS samples (Figure 10). This effect may be attributed to conversion of the

hydration products in the slag, comparable with that in high-alumina cements

(equations 5 and 7). The increase in strength noted in both samples after 28 days

may be attributed partly to the formation of strätlingite (C2ASH8) later during the

curing period, as detected in the mineralogy of HLS by means of XRD (Appendix

A). The formation of this phase is dependent on the reactive silica in the slag, the

alkalinity of the binding system and the water-to binder ratio. Hydration and

hardening of γ–C2S during longer periods of curing has been reported earlier in the

literature [117], [118].

Fig. 13. Compressive strengths of the paste samples at 3, 7, 28 and 90 days of hydration

(Reprinted by permission from Paper III © 2018 Elsevier Ltd.).

The conversion that occurs in this system is accompanied by a change in the crystal

structure of the hexagonal metastable hydrate C2AH8 to that of a cubic stable one,

C3AH6. These aluminate hydrates possess different specific densities, C2AH8

53

having a density of 1750 kg/m3 and C3AH6 2520 kg/m3 [116]. Consequently, this

transformation results in a significant change in the volume of solids in the hydrated

slag, yielding a more porous structure and a decrease in strength. In addition, the

water released during the conversion (equation 6) is readily available to react with

any unreacted anhydrous slag when the w/b ratio is initially low, to form more

C2AH8, which then leads to an increase in strength, as seen for HLS/0.35 in Figure

13.

4.5.2 Characterization of the conversion of HLS hydration products

The conversion that occurs in hydrated LS is accompanied by a change in the

crystal structure of the initially formed hexagonal metastable calcium-aluminate

hydration product C2AH8 to a cubic stable hydration product C3AH6. These

aluminate hydrates differ in density, C2AH8 having a density of 1750 kg/m3 and

C3AH6 2520 kg/m3 [116]. Consequently, this transformation results in significant

change in the volume of solids in the hydrated slag, resulting in a more porous

structure, increased free water and a decrease in strength.

The preferred analytical method for determining and verifying the conversion

of these hydration products is thermogravimetric (DTG) analysis, the results of

which when applied to the HLS samples are shown in Figure 14. The comparative

range of hydration products arising in high-alumina cement hydration has been

detailed in the literature, and this information was used here for reference purposes

[119], [120]. The metastable hydration product (C2AH8) has a decomposition peak

in the temperature range 100-200 °C, while the stable hydration products AH3 and

C3AH6 have decomposition peaks at 200-250 °C and 300-380 °C, respectively. The

DTG curves in both thermograms were de-convoluted using Origin software’s

Gaussian peak fitting algorithm (Originlab Corporation, US).

The rate of conversion can be analysed in terms of the increase in the intensities

of the decomposing peaks for C3AH6 and the decrease in the decomposing peaks

for C2AH8. The DTG curves for HLS/0.50 in Figure 14a show that the broad peaks

at 3 days consisted of overlapping hydrates associated with all three hydration

products. The decomposing peaks for the C3AH6 continue to increase up to 90 days,

while a decrease occurs in the decomposing peaks for C2AH8. This is consistent

with the known conversion of the metastable hydration product to a stable hydrate.

54

Fig. 14. DTG curves for the hydrated ladle slag (HLS) using (a) w/b 0.50 and (b) w/b 0.35.

(a=C2AH8, b=AH3 and c=C3AH6) (Reprinted by permission from Paper III © 2018 Elsevier

Ltd.).

The effect of a lower water content in reducing conversion is seen in Figure 14b,

where decomposing peaks for C2AH8 increased with the duration of curing up to a

maximum at 90 days with an intensity higher than the corresponding peaks for

C3AH6. The reduction in converted phases in this system and the increment in the

C2AH8 phase may partly be attributed to the continued hydration of anhydrous LS

with water released during conversion [73]. This also contributed to the gain in

strength in the HLS/0.35 case, as discussed earlier. By contrast, in the HLS/0.50

system, where a greater amount of water is readily available for large-scale

hydration of the slag, the conversion is rapid and comprehensive.

In addition, the conversion of HLS hydrates was further analysed by means of

NMR spectra. The 27Al NMR spectra for the samples (Figure 15a and b) show a

coordination of three aluminium species (AlIV, AlV and AlVI) detectable between

the chemical shifts 52 and 80 ppm, 30 and 40 ppm, and 10 and 20 ppm, respectively

[121], [122]. The broad peak for the raw sample, with its highest intensity at 70

ppm may be assigned to the mayenite phase in the slag, while the less intense peak

55

at 10 ppm indicates slight hydration of the slag during storage at the steelmaking

plant.

Furthermore, the peak in the AlIV region of the HLS/0.35 spectrum shown in

Figure 15a reduced after hydration as a result of consumption of the mayenite phase.

At the same time, the peak in the AlVI region increased due to the formation of

hydration products. The 27Al NMR spectrum for this sample had peaks near 10 ppm,

which may be assigned to the formation of the hexagonal hydrate C2AH8 [123]. The

shift in the spectrum between day 3 (10.8 ppm) and day 90 (10.5 ppm) could

indicate continuous formation of C2AH8.

In the spectrum for the HLS/0.5 sample (Figure 15b), massive conversion of

the metastable product to a stable product may be verified from the change in signal

peaks. The spectrum contains peaks near 12 ppm associated with formation of the

cubic hydrate C3AH6 [123], [124], in addition to which the spectrum signal peaks

shifted from 11.5 ppm at 3 days towards a maximum of 12.2 ppm at 90 days, which

may suggest continuous conversion of C2AH8 to C3AH6. The spectra for both

hydrated slag samples still show minor signals in the 70 ppm range due to unreacted

slag.

Fig. 15. 27Al MAS NMR spectra of the anhydrous and hydrated slag; HLS/0.35 (a) and

HLS/0.50 (b) slags at 3, 28 and 90 days (Reprinted by permission from Paper III © 2018

Elsevier Ltd.).

The effect of water content (w/b ratio) on the impact of conversion can be described

further. When the water-to-binder ratio is set low (i.e. w/b: < 0.40), there is

insufficient water and space for all the slag to dissolve or react to form these

metastable hydrates [73]. The water released from the conversion of the hydrated

56

slag reacts with the unreacted slag and form new hydrates. Thus, the rate of

conversion is lower compared to a paste with high volume of water.

4.6 Modification of ladle slag hydration products

Another alternative for lowering the conversion of hydration products apart from

reducing the water-to-binder ratio of the LS is by modifying its hydration products.

Both C12A7 and C3A will react with gypsum to form ettringite (Aft – C6A�̅�3H32)

and later monosulphoaluminate (AFm - C4A�̅�H12) in two-stage processes as shown

in the equations below.

C12A7 with water alone: 𝐶 𝐴 51𝐻 → 6𝐶 𝐴𝐻 𝐴𝐻 (12)

In the presence of gypsum (1st stage):

𝐶 𝐴 12𝐶�̅�𝐻 113𝐻 → 4𝐶 𝐴�̅� 𝐻 3𝐴𝐻 (13)

(2nd stage): 2𝐶 𝐴 4𝐶 𝐴�̅� 𝐻 34𝐻 → 12𝐶 𝐴�̅�𝐻 6𝐴𝐻 . (14)

C3A with water alone:

2𝐶 𝐴 27𝐻 → 𝐶 𝐴𝐻 𝐶 𝐴𝐻 (15)

In the presence of gypsum (1st stage):

𝐶 𝐴 3𝐶�̅�𝐻 26𝐻 → 𝐶 𝐴�̅� 𝐻 (16)

(2nd stage): 2𝐶 𝐴 𝐶 𝐴�̅� 𝐻 4𝐻 → 3𝐶 𝐴�̅�𝐻 (17)

The formation of AFm and the rate at which it is produced are dependent on the

initial amount of gypsum used in the hydration. A previous study on the hydration

of C3A with varying amounts of gypsum [125] has reported different durations for

AFm formation following depletion of the gypsum in the initial mix.

4.6.1 Hydration kinetics of modified ladle slag (in the presence of

gypsum - LSG)

The calorimetric results for the heat that evolved during the first 50 hours of

hydration of a mix containing LS, gypsum and water (LSG) are shown in Figure

57

16, along with HLS heat evolution for comparison purposes. Three significant

peaks and a shoulder can be seen in the calorimetric curve for LSG, and these are

similar to hydration periods of OPC as described in a previous publication [126].

The first peak, denoting the initial dissolution period starting from the very first

minute of hydration, may be attributed to the combined effect of heat release from

wetting of the precursors and initial hydration of the calcium aluminate phases in

the LS and gypsum [126].

This is followed by a dormant period, or induction period, lasting up to 8 h that

is caused by the cumulative effect of retardation by citrate ions and gypsum. In

OPC, gypsum is added to the clinker to serve as a set retarder, slowing down the

hydration of calcium aluminate. It has also been proposed that the adsorption of

calcium and sulphate ions to active dissolving sites on the calcium aluminate phases

serves to retard the reaction rate [125], [127], [128].

Fig. 16. Heat evolution of LSG and HLS as a function of hydration time (Reprinted by

permission from Paper IV © 2018 Elsevier Ltd.).

The second sharp peak in LSG curve, denoting an acceleration period, can be

observed after 8 h, with its heat flow peaking after 10 h of hydration. This peak

suggests the end of mix retardation and may be attributed to heat evolving from the

precipitation of hydrates (AFt and AH3) that leads to the setting and hardening of

the binder. The sharpness of this peak may also suggest the consumption of calcium

58

aluminate phases in the slag. The second peak may be attributed to the second

formation of ettringite [129], [130], hypothetically on account of a longer delay

caused by the retarding effect of the citrate ions on C3A hydration. In addition, the

decrease of the reaction of the curve after the second peak may be because of the

formation of hydration products were covered on the surface of unreacted products.

The reaction changes in the third broad peak to the diffusion control of this covered

slag and may suggest moderate consumption of the calcium aluminate phases in

the covered unhydrated slag particles.

The broad shoulder (hump) on the third peak is due to the formation of

monosulphoaluminate (AFm). This arises when the sulphate ions in the binder

matrix decrease on account of gypsum depletion and the remaining C12A7 and C3A

starts to react with ettringite and water to form monosulphoaluminate (equation (14)

and (15) [125]. Nonetheless, this does not mean total exhaustion of the ettringite in

the matrix, a fact that is consistent with the XRD analyses and is confirmed by them.

In addition, the cumulative heat flow shows a higher amount of total heat generated

(325 J/g) during 50 h of hydration, more than in the case of LS hydration (275 J/g),

representing the cumulative contribution of the heat released by Aft, AFm and AH3.

As seen in equation (12) and (13), the quantity of AH3 produced in the hydration

of LS and gypsum is three times greater than in the hydration of LS alone.

Two distinct peaks are seen in the curve for the heat released by HLS hydration,

an initial peak attributable to the dissolving and wetting of the slag particles, which

is followed by a short dormant period of less than 2 hours, and then a second peak,

the acceleration peak, caused by the formation of hydration products in the matrix.

4.6.2 Characterization of LSG hydration products

The DTG results in Figure 17 show modification of the LS calcium aluminate

phases to ettringite after the addition of gypsum. The decomposition peaks for LSG

and the reference HLS paste after 7 and 28 days are presented in Figure 17a and b,

respectively. Ettringite, gibbsite (AH3) and possibly an AFm phase are the major

phases identifiable in the LSG thermograms. The well-defined decomposing peak

in the range 50-350 °C corresponds to ettringite decomposition [131], [132], while

the shoulder visible between 250 and 320 °C may correspond to the decomposition

of AH3 [119] and can also overlap with AFm, which has its decomposition peak

between 200 and 250 °C [133]. The quantification of ettringite by

thermogravimetric analysis is a complex process because of the possibility that the

59

ettringite peaks may overlap with those of other hydrated product phases, e.g.

C2ASH8 [119].

The DTG curves for HLS are similar to those described in the previous

subsection 4.5.2. The conversion of C2AH8 (100-200 °C) to C3AH6 (300-380 °C)

was likewise observed, along with a decrease in the metastable peak and an

increment in the stable decomposition peak.

Fig. 17. Differential thermogravimetric (DTG) analyses of HLS and LSG paste samples

at (a) 7 days and (b) 28 days (Reprinted by permission from Paper IV © 2018 Elsevier

Ltd.).

The mineralogical phases of the LSG sample were identified and quantified by

XRD/Rietveld analysis, yielding the results shown in Figures 18 and 19. As seen

in Figure 18, AFt was the dominant phase in the LSG diffractograms at both 7 and

28 days, amounting to 34.4 wt.% and 27.3 wt.% at the two time points, respectively.

The reduction in the quantity of AFt at 28 days is attributable to the formation of

an AFm phase after depletion of the gypsum. The AFm phase increased in quantity

from 0.5 wt.% at 7 days to 5.3 wt.% at 28 days, and the amorphous content of the

binding system also increased slightly (by 1%) during this period. Ettringite and an

amorphous hydration product are the main strength-imparting phases in this system.

It is clear from equation 13 and 16 that AH3 is one of the predictable hydration

products of the reaction between mayenite and gypsum, so that although it is

scarcely crystalline and is difficult to detect and identify in XRD, it may contribute

to the amorphous content of the LSG sample.

60

Fig. 18. X-ray diffractograms of LSG paste at 7 and 28 days (Reprinted by permission

from Paper IV © 2018 Elsevier Ltd.).

Fig. 19. Quantified phases of LSG at 7 and 28 days (Reprinted by permission from Paper

IV © 2018 Elsevier Ltd.).

61

The diffractograms for HLS in Figure 20 can additionally be used to verify

conversion of the hydrate phases during curing. It was not possible to quantify the

HLS paste sample by Rietveld analysis due to the number of coordination

anomalies, but a C2AH8 peak was identified at 7 days which had been diminished

significantly by 28 days, while C3AH6 was identified at both points in time but with

an increased peak intensity at 28 days. This may suggest or confirm that conversion

of the hydrate phases had taken place in the paste. This would be consistent with

previous observations regarding CAC hydration [121]. Strätlingite was observed in

the diffractograms at 28 days, and it can be postulated that this phase may be the

result of a reaction between C2AH8/C3AH6 and dissolved silica from the weakly

hydraulic calcio-olivine (γ-C2S). In any case, traces of strätlingite have been

reported earlier in CAC after prolonged hydration [134], [135].

Fig. 20. XRD of HLS phases at 7 and 28 days after hydration (Reprinted by permission

from Paper IV © 2018 Elsevier Ltd.).

62

4.6.3 Microstructure of LSG binder hydration products

The SEM analysis in Figure 21 confirms the phases detected in the LSG and HLS

samples by means of DTG and XRD. Needle-like crystals consistent with those of

ettringite in a cementitious binder could be seen in the LSG microstructure from 7

days onwards, as shown in Figure 21a. This agrees with equation 13 and 16

regarding the formation of ettringite after hydration, although no hexagonal plates

consistent with AFm phases were identified in the area analysed [127]. In contrast,

the microstructure of the HLS sample showed a different evolution at both points

in the analysis (Figure 21b). At 7 days the microstructure of HLS showed plate-like

crystals, as is consistent with the hexagonal structure of C2AH8 [136], while a

significant change in the morphology of the hydrates in the HLS sample could be

seen at 28 days, when the microstructure showed cubic crystals identifiable to the

C3AH6 crystal structure [137]. These microstructural alterations further verify the

conversion of the hydration product in the LS samples.

Fig. 21. SEM images showing the microstructural evolution of (a) LSG and (b) HLS at 7

(left) and 28 (right) days (Reprinted by permission from Paper IV © 2018 Elsevier Ltd.).

63

4.7 Strength and durability of HLS and LSG mortars.

4.7.1 Effect of modification of ladle slag hydration products on

strength

The alteration in the LS hydration products brought about by the addition of

gypsum (LSG) enhances their strength and eliminate the conversion anomaly in the

case of HLS mortars. The results of testing the strength of the LSG and comparing

it with HLS, shown in Figure 22, imply that LSG achieved a relatively good

compressive strength of 30 MPa at 7 days and 35 MPa at 28 days. In contrast, HLS,

as previously discussed, suffered a loss of 25% of its day-7 strength after 28 days.

Similar observations were made with regard to flexural strength in both samples.

This confirms modification of the hydration products in both systems. The strength-

enhancing hydration products in LSG are thought to be Aft, the AFm phases and

AH3, as confirmed earlier in the XRD and TGA analyses, as confirmed earlier from

XRD and TGA analyses.

Fig. 22. Strength test of mortars at 7 and 28 days (a) compressive strength and (b)

flexural strength (Reprinted by permission from Paper IV © 2018 Elsevier Ltd.).

4.7.2 Effect of ettringite formation on mortars length change

The LSG mortars exhibited lower shrinkage than the HLS one throughout the 90

days of the analysis, as shown in Figure 23, and the shrinkage values for LSG were

64

considerably lower than those for HLS on all the days analysed. The HLS material

shrank approximately 6 times more than the LSG, experiencing pronounced

shrinkage during the early days of curing (up to 28 days) and having almost

stabilized by 90 days. The overall shrinkage, however, was comparable to the

values recorded elsewhere for BFS mortars [138]. The lower shrinkage of LSG is

attributed to the expansive contribution of ettringite formed in the modified system

at early ages of hydration [139]. These results are in agreement with other studies

on the low shrinkage values recorded in ettringite-based binder systems [140],

[141].

However, it is noteworthy to mention that to avoid delayed ettringite formation

leading to expansion of the mortars, the starting material elemental composition

was optimised (here we used bogue’s formula) so as to have sufficient calcium

aluminate to react or deplete all the gypsum in the early days of hydration.

Fig. 23. Drying shrinkage values for LSG and HLS during 90 days of curing (Reprinted

by permission from Paper IV © 2018 Elsevier Ltd.).

4.7.3 Freeze-thaw of mortars

The residual strengths of the samples after exposure to severe weathering are shown

in Figure 24. LSG experienced a strength gain of up to 30% of that in the reference

65

mortar after 120 cycles of exposure to freezing-thaw cycling. This strength gain

under destructive conditions may be due to the further formation of ettringite in a

moist environment, which favours the continuous formation of ettringite [142].

After 120 cycles of freeze-thaw the strength decreased slightly, retaining 30 MPa

after 300 cycles, which implies a strength reduction of 15% relative to the reference

mortar. On the other hand, HLS suffered a 20% strength loss after 60 cycle,

although it showed a slight strength gain after 120 cycles. Overall, the mortar failed

(disintegrated) after 180 cycles and was not analysed further. As seen in the mass

loss curves, LSG experienced a gain in mass during the freeze-thaw cycles and

exhibited a loss of mass of less than 2% by the end of the experiment.

Fig. 24. (a) Residual strength after freeze-thaw and (b) mass loss/gain for both mortar

samples (Reprinted by permission from Paper IV © 2018 Elsevier Ltd.).

The strength exhibited by the LSG and its frost resistance values conform to the

standards for structural design [143, p. 132] and many ASTM designations.

4.8 Fire resistance of ladle slag mortars

LSG mortar sample was not analysed during the investigation into the thermal

resistance of ladle slag mortars under elevated temperature (110 -1000 °C), due to

the early decomposition of ettringite under elevated temperatures below 300°C.

From results presented earlier, it is clear that water-to-binder ratio for HLS sample

(w/b: 0.35) is ideal for all the mortar samples and was maintained in the mortars

for fire resistance. To reduce costs and the use of metallurgical by-products in the

mortar, ferrochrome slag (Fs) was used as an aggregate. The results were then

66

compared with a commercial CAC mortar and sand (Ss) as aggregates. To reduce

the spalling tendency under high temperatures, all the samples were dried at 110 °C

before exposure to elevated temperatures starting from 200 °C. The visual

appearance of the mortars after exposure to elevated temperatures is shown in

Appendix B.

4.8.1 Residual mechanical strength

LS and Fs showed significant potential as refractory materials in a hydraulic binder

and similarities to CAC and sand, respectively. The residual strengths of mortars

made with Ss and Fs after thermal exposure are shown in Figure 25. The superior

initial strength of AALS_Fs mortars at 23 °C is assisted by the reaction of the binder

with the Fs aggregates. The ferrochrome aggregates in fact acted as bilateral

materials, i.e. both as a binder and as aggregates in the AALS mortars. Similar

observations have been reported in alkali-activated materials containing

ferrochrome aggregates [144]. Usually in mortars or concretes, aggregates used are

chemically inactive. However, some aggregates can react under high alkaline

environment, this behaviour is termed alkali-aggregates reaction, and these may

either be deleterious or beneficial to the mortar properties. The latter was the case

concerning AALS_Fs in this study (See Appendix C, Fig.C1e).

Furthermore, after drying at 110 °C, both AALS mortars suffered between 57

and 71% losses of strength due to dehydration of the strength-giving phases in the

mortar, i.e. C-S-H like reaction product. The changes in strength from 200 °C

upwards were gradual, and clearly below those seen in CAC and HLS mortars,

while an increase in strength at 1000 °C was observed for AALS_Ss and could be

attributed to densification of the sample through sintering of the binder, and perhaps

also to a phase transformation in the sand, as this was not observed in the AALS_Fs

mortar. Similar increments in strength have been reported in alkali-activated

materials exposed to high-temperatures [145]–[147]. Overall, the loss of strength

in these AALS mortars may be attributed to the thermal shrinkage of the binder and

expansion of the aggregates. The drying of free water and hydroxyls at 110 °C may

also have contributed to the deterioration in strength of the mortars [148].

The HLS mortars experienced an increment in strength after drying at 110 °C,

and this continued up to 200 °C for the HLS_Ss mortars, while the CAC mortars

had a slight reduction in strength at this point. After 200 °C both the HLS and CAC

mortars experienced similar a downward trend in strength loss until 1000 °C, and

by the end of the exposure to 1000 °C the HLS_Fs, CAC_Ss and AALS_Ss

67

materials had all recorded residual strength of approximately 18 MPa, representing

52%, 70% and 76% losses of strength respectively compared with the unexposed

reference strength. It is important to note that the residual strengths of these mortars

(e.g. HLS) are all above the minimum required in a structural clay load-bearing

wall in applications designated by ASTM C34-03- [149].

Fig. 25. Residual strength of mortars with (a) sand and (b) ferrochrome aggregates after

exposure to elevated temperature (Reprinted by permission from Paper V © ICE

Publishing).

4.8.2 Dilatometry of the thermal shrinkage and expansion of mortars

The results of dilatometric studies of the mortars given in Figure 26 show that their

shrinkage/expansion is dependent on the type of aggregate used. In mortars with

sand aggregates (Figure 26a), no substantial expansion/shrinkage occurred until

100 °C, except for the AALS_Ss sample, which shrank after 80 °C due to loss of

free water in the mortar. Furthermore, AALS_Ss continued to shrink further in the

temperature range 100–230 °C, which corresponds to dehydration of the reaction

product in the binder. Similar dimensional behaviour has been reported for alkali-

activated materials [150]–[152]. Between 230 °C and 570 °C a gradual expansion

of the AALS_Ss mortar could be observed. This was due to expansion of the quartz

towards its phase transformation at 573 °C, since the net expansion of quartz

dominated over the shrinkage of the binder. Furthermore, the sample then shrank

until 750 °C due to densification of the binder at this temperature [153]. This agrees

with the strength gain recorded here, and no further shrinkage or expansion was

recorded beyond 750 °C.

68

HLS_Ss and CAC_Ss exhibited similar dimensional behaviour throughout the

experiment, expanding slightly immediately after reaching 100 °C and continuing

up to 250 °C, after which both samples shrank (at 250-300 °C) on account of the

loss of pore water and collapse of the pores in the matrix. This expansion continued

until 570 °C before stabilizing, with minimal shrinkage or expansion recorded

afterwards. The expansion seen in the HLS_Ss mortars was due to expansion of the

aggregates at high temperatures (on account of quartz transformation), which may

have caused cracking of the matrix.

Fig. 26. Dilatometric curves showing expansion and shrinkage of the mortars with (a)

sand and (b) ferrochrome aggregates at elevated temperatures (Reprinted by

permission from Paper V © ICE Publishing).

The mortars containing Fs slag as aggregates exhibited a different dilatometry

pattern (Figure 26b). The AALS_Fs experienced major shrinkage and little

expansion throughout the experiment. This difference between this pattern and that

observed with AALS_Ss is related to the reaction of the Fs aggregate to a highly

alkaline environment, as discussed earlier, while HLS_Fs and CAC_Fs, showed

similar patterns, with slight shrinkage and expansion until 1000 °C. Overall,

HLS_Fs shrank to just slightly below the initial value, whereas CAC_Fs, expanded

to slightly above the initial value.

4.8.3 Quality of mortars after exposure to elevated temperatures

A non-destructive technique was used to check the quality of the mortars after

thermal exposure. By measuring the time taken for the pulse from the transmitter

69

probe to reach the receiver probe, the ultrasonic pulse velocity (UPV) can be

calculated used to predict the microstructure and mechanical behaviour of the

sample. The higher the velocity, the fewer the internal pores and defects in the

sample. As can be seen in Figure 27, mortars containing Fs aggregate achieved

better UPV values than sand-based mortars, partly on account of the irregular

shapes of the Fs particles, which may have contributed to packing and to superior

mechanical bonding between the aggregate and the binder [144]. Bonding between

the aggregate and the binder may also have partly contributed to the result gained

with the AALS mortar. Fs bonded better with the paste than sand aggregate

attributed to aggregate reaction with the paste (See appendix C, Fig.C1e).

Fig. 27. Ultrasonic pulse velocity for mortars with (a) sand and (b) ferrochrome slag

aggregates (Reprinted by permission from Paper V © ICE Publishing).

70

71

5 Summary and concluding remarks

The aim of this work was to study the utilization of LS, a by-product of the

steelmaking process, in high-value applications by determining the potential

hydraulic activity of this slag on the basis of its chemical composition and

mineralogy. Papers I, II and V of this thesis investigated the alkali activation of this

slag and studied its durability after hardening. Since the activated sample set rapidly

by virtue of its major mineral phase, mayenite, citric acid was added as a retarder

between 0.8 and 1 wt.% (of the water content) to prolong its setting.

After alkali activation LS showed considerable mechanical strength, so that it

could be used as a sole precursor in alkali-activated materials. The reaction product

formed in AALS was analysed by XRD and FTIR and shown to be possibly a C-

A-S-H-like gel and, in view of the stoichiometry of the mix with sodium in its

structure, to be acting as a charge balance or to be absorbed onto the gel. Unlike

the situation in fly ash or other aluminosilicates alkali activation, elevated

temperatures are not required for the hardening or curing of AALS as it had no

significant influence on the strength of the matrix.

When the durability of AALS mortars under conditions of severe weathering

was studied, the mortar sample exhibited good frost resistance after 60 freeze-thaw

cycles, losing less than 1% of its strength and less than 2% of its weight. On the

other hand, the shrinkage of the AALS mortar on drying was almost twice that

reported for alkali-activated BFS and 7 times more than for OPC.

In spite of the lower strength, the hydraulic properties of LS (yielding HLS)

offers a cheaper alternative method for utilizing LS. The main hydration products

emerging from this hydraulic system were found to convert from metastable C2AH8

to stable C3AH6 during curing, thereby undergoing a loss of strength. By lowering

the water content of the system to a maximum of w/b: 0.35, the rate of conversion

could be lowered, irrespective of the w/b ratio, while all the samples tested recorded

a continuous gain in strength from 28 days onwards.

To modify and stop the conversion of hydration products in this HLS system,

the HLS was blended with gypsum to form an ettringite system. Upon adding the

gypsum, the calcium aluminate phases (mayenite and tricalcium aluminate) reacted

to form ettringite and gibbsite, and later monosulphoaluminate was also detected

as a modified hydration product. No strength loss was recorded in this modified

ettringite binder, and when the efficiency of the modified system was analysed

under severe conditions, the sample exhibited superior durability properties and

strength to those of the HLS mortars. The gypsum used in this case can also be

72

potentially replaced by gypsum produced as a by-product in the phosphoric acid

industry. Gypsum-modified LS binder is a potential low-cost cementitious material.

The last paper in this thesis studied the refractory applications of the mortars

designed here (except for LSG) using a w/b of 0.35. The results obtained by

exposing the mortars to high temperatures (110-1000 °C) showed that HLS

containing ferrochrome aggregate had a better resistance and structural appearance

than commercial refractory cement or AALS mortars. The HLS mortar sample

retained a strength of approximately 18 MPa after exposure to 1000 °C. In a high

alkaline environment (AALS), the introduction of ferrochrome slag as an aggregate

increased the strength of the mortar before its exposure to elevated temperatures.

The increment in strength compared with AALS with sand may be partly attributed

to the alkali-aggregate reaction occurring in the system, and for this reason

ferrochrome slag may not be suitable as an aggregate in high temperature

applications of alkali-activated materials.

The results of this study on the use of LS as a cementitious binder show that

LS could be used to achieve high-strength mortars. Additionally, as a cementitious

material with almost zero environmental footprints, it could be used as a cheaper

alternative to commercial calcium aluminate cement and Portland cement in

refractory and structural applications in civil engineering. The fast setting of this

slag and its rapid gain in strength can be advantageous when using it for structural

repairs. In addition, to avoid strength loss, the optimal water-to-binder ratio of 0.35

should be used, save if the composition of the mix is modified with an addition of

gypsum. And it is imperative that proper strength value is considered before use in

structural design. However, it is also necessary to point out that the free lime

content of the slags used in this thesis was below 1% which pose no deleterious

effect on the mortars. Also, free MgO (<5% recommended) content should be put

into consideration before the use of the slag in structural applications due to long

term volumetric expansion caused by carbonation and hydration of the free MgO.

Here in this thesis, the content of the ladle slag varies between 0 to 7%. The use

should be limited if it’s above the recommended value of below 5%. The presence

can however be specially utilized under suitable conditions for their gainful use in

structural applications.

To conclude, the present results contribute to the scientific understanding of

high-alumina slag hydration, including its relation to CAC cements, modification

of the reaction and potential fields for its utilization in connection with inorganic

binders.

73

List of references

[1] B. Das, S. Prakash, P. S. R. Reddy, and V. N. Misra, “An overview of utilization of slag and sludge from steel industries,” Resour. Conserv. Recycl., vol. 50, no. 1, pp. 40–57, Mar. 2007.

[2] I. Z. Yildirim and M. Prezzi, “Chemical, Mineralogical, and Morphological Properties of Steel Slag,” Adv. Civ. Eng., vol. 2011, p. e463638, Oct. 2011.

[3] L. K. Turner and F. G. Collins, “Carbon dioxide equivalent (CO2-e) emissions: A comparison between geopolymer and OPC cement concrete,” Constr. Build. Mater., vol. 43, pp. 125–130, Jun. 2013.

[4] J. S. J. van Deventer, J. L. Provis, P. Duxson, and D. G. Brice, “Chemical research and climate change as drivers in the commercial adoption of alkali activated materials,” Waste Biomass Valorization, vol. 1, no. 1, pp. 145–155, 2010.

[5] J. Rissanen, K. Ohenoja, P. Kinnunen, and M. Illikainen, “Partial Replacement of Portland-Composite Cement by Fluidized Bed Combustion Fly Ash,” J. Mater. Civ. Eng., p. 04017061, Mar. 2017.

[6] S. Z. Carvalho, F. Vernilli, B. Almeida, M. Demarco, and S. N. Silva, “The recycling effect of BOF slag in the portland cement properties,” Resour. Conserv. Recycl., vol. 127, pp. 216–220, Dec. 2017.

[7] Q. Feng, H. Yamamichi, M. Shoya, and S. Sugita, “Study on the pozzolanic properties of rice husk ash by hydrochloric acid pretreatment,” Cem. Concr. Res., vol. 34, no. 3, pp. 521–526, Mar. 2004.

[8] H. Mallisa and G. Turuallo, “The maximum percentage of fly ash to replace part of original Portland cement (OPC) in producing high strength concrete,” AIP Conf. Proc., vol. 1903, no. 1, p. 030012, Nov. 2017.

[9] G. J. Osborne, “Durability of Portland blast-furnace slag cement concrete,” Cem. Concr. Compos., vol. 21, no. 1, pp. 11–21, Jan. 1999.

[10] E. Adesanya, K. Ohenoja, T. Luukkonen, P. Kinnunen, and M. Illikainen, “One-part geopolymer cement from slag and pretreated paper sludge,” J. Clean. Prod., vol. 185, pp. 168–175, 2018.

[11] J. Davidovits, Geopolymer, Green Chemistry and Sustainable Development Solutions: Proceedings of the World Congress Geopolymer 2005. Geopolymer Institute, 2005.

[12] P. Duxson, A. Fernández-Jiménez, J. L. Provis, G. C. Lukey, A. Palomo, and J. S. J. van Deventer, “Geopolymer technology: the current state of the art,” J. Mater. Sci., vol. 42, no. 9, pp. 2917–2933, May 2007.

[13] S. A. Bernal et al., “Binder chemistry - high-calcium alkali-activated materials,” in Alkali Activated Materials - State-of-the-art report RILEM TC 224-AAM, vol. 2014, Dordrecht: Springer, 2014, pp. 59–91.

[14] P. Hlaváček, V. Šmilauer, and F. Škvára, “Alkali-Activated Fly Ash Foams – Synthesis, Chemo-Physical Properties and Microstructure Modeling,” Adv. Sci. Technol., vol. 92, pp. 14–19, Oct. 2014.

74

[15] J. L. Provis, “Geopolymers and other alkali activated materials: why, how, and what?,” Mater. Struct., vol. 47, no. 1–2, pp. 11–25, Jan. 2014.

[16] S. Kourounis, S. Tsivilis, P. E. Tsakiridis, G. D. Papadimitriou, and Z. Tsibouki, “Properties and hydration of blended cements with steelmaking slag,” Cem. Concr. Res., vol. 37, no. 6, pp. 815–822, Jun. 2007.

[17] V. Z. Serjun, B. Mirtič, and A. Mladenovič, “Evaluation of ladle slag as a potential material for building and civil engineering,” Mater. Tehnol., vol. 47, no. 5, pp. 543–550, 2013.

[18] European Commission, “Directive 2008/98/EC on waste (Waste Framework Directive) - Environment - European Commission,” 2008. [Online]. Available: http://ec.europa.eu/environment/waste/framework/. [Accessed: 29-Nov-2018].

[19] C. Shi, “Characteristics and cementitious properties of ladle slag fines from steel production,” Cem. Concr. Res., vol. 32, no. 3, pp. 459–462, 2002.

[20] T. Tracz, I. Hager, K. K. Sideris, C. Tassos, and A. Chatzopoulos, “7th Scientific-Technical Conference on Material Problems in Civil Engineering MATBUD’2015Production of Durable Self-compacting Concrete Using Ladle Furnace Slag (LFS) as Filler Material,” Procedia Eng., vol. 108, pp. 592–597, Jan. 2015.

[21] J. M. Manso, Á. Rodriguez, Á. Aragón, and J. J. Gonzalez, “The durability of masonry mortars made with ladle furnace slag,” Constr. Build. Mater., vol. 25, no. 8, pp. 3508–3519, Aug. 2011.

[22] L. Andreas, S. Diener, and A. Lagerkvist, “Steel slags in a landfill top cover – Experiences from a full-scale experiment,” Waste Manag., vol. 34, no. 3, pp. 692–701, Mar. 2014.

[23] Y. Jiang, T.-C. Ling, C. Shi, and S.-Y. Pan, “Characteristics of steel slags and their use in cement and concrete—A review,” Resour. Conserv. Recycl., vol. 136, pp. 187–197, Sep. 2018.

[24] J. Setién, D. Hernández, and J. J. González, “Characterization of ladle furnace basic slag for use as a construction material,” Constr. Build. Mater., vol. 23, no. 5, pp. 1788–1794, May 2009.

[25] E. Adesanya, K. Ohenoja, P. Kinnunen, and M. Illikainen, “Alkali Activation of Ladle Slag from Steel-Making Process,” J. Sustain. Metall., pp. 1–11, 2016.

[26] E. Adesanya, K. Ohenoja, P. Kinnunen, and M. Illikainen, “Properties and durability of alkali-activated ladle slag,” Mater. Struct., vol. 50, no. 6, p. 255, Dec. 2017.

[27] H. Nguyen et al., “Byproduct-based ettringite binder – A synergy between ladle slag and gypsum,” Constr. Build. Mater., vol. 197, pp. 143–151, Feb. 2019.

[28] E. Adesanya et al., “Ladle slag cement – Characterization of hydration and conversion,” Constr. Build. Mater., vol. 193, pp. 128–134, Dec. 2018.

75

[29] J. Zhang, H. Matsuura, and F. Tsukihashi, “Chapter 4.4 - Processes for Recycling,” in Treatise on Process Metallurgy, S. Seetharaman, Ed. Boston: Elsevier, 2014, pp. 1507–1561.

[30] C. Shi, “Steel Slag—Its Production, Processing, Characteristics, and Cementitious Properties,” J. Mater. Civ. Eng., vol. 16, no. 3, pp. 230–236, Jun. 2004.

[31] J. Ylipekkala, Quality management of chromium containing steel slags from melt phase to cooling. 2005.

[32] I. Yildirim and M. Prezzi, “Use of Steel Slag in Subgrade Applications,” JTRP Tech. Rep., Oct. 2009.

[33] H. Motz and J. Geiseler, “Products of steel slags an opportunity to save natural resources,” Waste Manag., vol. 21, no. 3, pp. 285–293, Jun. 2001.

[34] M. Mahoutian, Z. Ghouleh, and Y. Shao, “Carbon dioxide activated ladle slag binder,” Constr. Build. Mater., vol. 66, pp. 214–221, Sep. 2014.

[35] D. Adolfsson, R. Robinson, F. Engström, and B. Björkman, “Influence of mineralogy on the hydraulic properties of ladle slag,” Cem. Concr. Res., vol. 41, no. 8, pp. 865–871, Aug. 2011.

[36] F. Pietrini, V. Iori, T. Beone, D. Mirabile, and M. Zacchini, “Effects of a ladle furnace slag added to soil on morpho-physiological and biochemical parameters of Amaranthus paniculatus L. plants,” J. Hazard. Mater., vol. 329, pp. 339–347, May 2017.

[37] A. Natali Murri, W. D. A. Rickard, M. C. Bignozzi, and A. Van Riessen, “High temperature behaviour of ambient cured alkali-activated materials based on ladle slag,” Cem. Concr. Res., vol. 43, no. 1, pp. 51–61, 2013.

[38] I. Lancellotti, C. Ponzoni, M. C. Bignozzi, L. Barbieri, and C. Leonelli, “Incinerator Bottom Ash and Ladle Slag for Geopolymers Preparation,” Waste Biomass Valorization, vol. 5, no. 3, pp. 393–401, Feb. 2014.

[39] I. Papayianni and E. Anastasiou, “Effect of granulometry on cementitious properties of ladle furnace slag,” Cem. Concr. Compos., vol. 34, no. 3, pp. 400–407, Mar. 2012.

[40] M. C. Bignozzi, S. Manzi, I. Lancellotti, E. Kamseu, L. Barbieri, and C. Leonelli, “Mix-design and characterization of alkali activated materials based on metakaolin and ladle slag,” Appl. Clay Sci., vol. 73, no. 1, pp. 78–85, 2013.

[41] M. Nicolae, I. Vîlciu, and F. Zǎman, “X-ray diffraction analysis of steel slag and blast furnace slag viewing their use for road construction,” UPB Sci. Bull. Ser. B Chem. Mater. Sci., vol. 69, no. 2, pp. 99–108, 2007.

[42] W.-C. Wang, H.-Y. Wang, and H.-C. Tsai, “Study on engineering properties of alkali-activated ladle furnace slag geopolymer,” Constr. Build. Mater., vol. 123, pp. 800–805, Oct. 2016.

[43] F. Pacheco-Torgal, J. A. Labrincha, C. Leonelli, A. Palomo, and P. Chindaprasirt, Eds., “Woodhead Publishing Series in Civil and Structural

76

Engineering,” in Handbook of Alkali-Activated Cements, Mortars and Concretes, Oxford: Woodhead Publishing, 2015, pp. xvii–xix.

[44] S. Kourounis, S. Tsivilis, P. E. Tsakiridis, G. D. Papadimitriou, and Z. Tsibouki, “Properties and hydration of blended cements with steelmaking slag,” Cem. Concr. Res., vol. 37, no. 6, pp. 815–822, Jun. 2007.

[45] J. Provis and J. van Deventer, Eds., Alkali Activated Materials: State-of-the-Art Report, RILEM TC 224-AAM. Springer Netherlands, 2014.

[46] C. Shi and J. Qian, “High performance cementing materials from industrial slags — a review,” Resour. Conserv. Recycl., vol. 29, no. 3, pp. 195–207, Jun. 2000.

[47] S.-D. Wang, K. L. Scrivener, and P. L. Pratt, “Factors affecting the strength of alkali-activated slag,” Cem. Concr. Res., vol. 24, no. 6, pp. 1033–1043, 1994.

[48] M. Salman et al., “Alkali Activation of AOD Stainless Steel Slag Under Steam Curing Conditions,” J. Am. Ceram. Soc., vol. 98, no. 10, pp. 3062–3074, Oct. 2015.

[49] S.-D. Wang and K. L. Scrivener, “Hydration products of alkali activated slag cement,” Cem. Concr. Res., vol. 25, no. 3, pp. 561–571, Apr. 1995.

[50] F. Puertas and A. Fernández-Jiménez, “Mineralogical and microstructural characterisation of alkali-activated fly ash/slag pastes,” Cem. Concr. Compos., vol. 25, no. 3, pp. 287–292, Apr. 2003.

[51] I. G. Richardson, A. R. Brough, G. W. Groves, and C. M. Dobson, “The characterization of hardened alkali-activated blast-furnace slag pastes and the nature of the calcium silicate hydrate (C-S-H) phase,” Cem. Concr. Res., vol. 24, no. 5, pp. 813–829, Jan. 1994.

[52] S. A. Bernal et al., “Binder Chemistry – High-Calcium Alkali-Activated Materials,” in Alkali Activated Materials, J. L. Provis and J. S. J. van Deventer, Eds. Springer Netherlands, 2014, pp. 59–91.

[53] J. L. Provis and S. A. Bernal, “Geopolymers and Related Alkali-Activated Materials,” Annu. Rev. Mater. Res., vol. 44, no. 1, pp. 299–327, 2014.

[54] M. Chi, “Effects of dosage of alkali-activated solution and curing conditions on the properties and durability of alkali-activated slag concrete,” Constr. Build. Mater., vol. 35, pp. 240–245, Oct. 2012.

[55] F. Puertas, M. Palacios, H. Manzano, J. S. Dolado, A. Rico, and J. Rodríguez, “A model for the C-A-S-H gel formed in alkali-activated slag cements,” J. Eur. Ceram. Soc., vol. 31, no. 12, pp. 2043–2056, Oct. 2011.

[56] M. Ben Haha, G. Le Saout, F. Winnefeld, and B. Lothenbach, “Influence of activator type on hydration kinetics, hydrate assemblage and microstructural development of alkali activated blast-furnace slags,” Cem. Concr. Res., vol. 41, no. 3, pp. 301–310, Mar. 2011.

[57] I. Garcia-Lodeiro, A. Palomo, and A. Fernández-Jiménez, “An overview of the chemistry of alkali-activated cement-based binders,” in Handbook of Alkali-Activated Cements, Mortars and Concretes, 2014, pp. 19–47.

77

[58] P. Duxson, A. Fernández-Jiménez, J. L. Provis, G. C. Lukey, A. Palomo, and J. S. J. Van Deventer, “Geopolymer technology: the current state of the art,” J. Mater. Sci., vol. 42, no. 9, pp. 2917–2933, 2007.

[59] I. Garcia-Lodeiro, A. Fernandez-Jimenez, A. Palomo, and D. E. Macphee, “Effect of Calcium Additions on N-A-S-H Cementitious Gels,” J. Am. Ceram. Soc., vol. 93, no. 7, pp. 1934–1940, Jul. 2010.

[60] F. Bonk, J. Schneider, M. A. Cincotto, and H. Panepucci, “Characterization by Multinuclear High-Resolution NMR of Hydration Products in Activated Blast-Furnace Slag Pastes,” J. Am. Ceram. Soc., vol. 86, no. 10, pp. 1712–1719, 2003.

[61] J. I. Escalante-García, A. F. Fuentes, A. Gorokhovsky, P. E. Fraire-Luna, and G. Mendoza-Suarez, “Hydration Products and Reactivity of Blast-Furnace Slag Activated by Various Alkalis,” J. Am. Ceram. Soc., vol. 86, no. 12, pp. 2148–2153, 2003.

[62] W. Chen and H. J. H. Brouwers, “The hydration of slag, part 1: reaction models for alkali-activated slag,” J. Mater. Sci., vol. 42, no. 2, pp. 428–443, Jan. 2007.

[63] A. Fernández-Jiménez and F. Puertas, “Effect of activator mix on the hydration and strength behaviour of alkali-activated slag cements,” Adv. Cem. Res., vol. 15, no. 3, pp. 129–136, Jul. 2003.

[64] S. A. Bernal, R. M. de Gutierrez, J. L. Provis, and V. Rose, “Effect of silicate modulus and metakaolin incorporation on the carbonation of alkali silicate-activated slags,” Cem. Concr. Res., vol. 40, no. 6, pp. 898–907, 2010.

[65] A. S. Reddy, R. K. Pradhan, and S. Chandra, “Utilization of Basic Oxygen Furnace (BOF) slag in the production of a hydraulic cement binder,” Int. J. Miner. Process., vol. 79, no. 2, pp. 98–105, May 2006.

[66] M. Tossavainen, F. Engstrom, Q. Yang, N. Menad, M. Lidstrom Larsson, and B. Bjorkman, “Characteristics of steel slag under different cooling conditions,” Waste Manag., vol. 27, no. 10, pp. 1335–1344, 2007.

[67] Q. Wang and P. Yan, “Hydration properties of basic oxygen furnace steel slag,” Constr. Build. Mater., vol. 24, no. 7, pp. 1134–1140, Jul. 2010.

[68] E. E. Hekal, S. A. Abo-El-Enein, S. A. El-Korashy, G. M. Megahed, and T. M. El-Sayed, “Hydration characteristics of Portland cement – Electric arc furnace slag blends,” HBRC J., vol. 9, no. 2, pp. 118–124, Aug. 2013.

[69] L. Kriskova et al., “Effect of mechanical activation on the hydraulic properties of stainless steel slags,” Cem. Concr. Res., vol. 42, no. 6, pp. 778–788, Jun. 2012.

[70] Q. Wang, P. Yan, and S. Han, “The influence of steel slag on the hydration of cement during the hydration process of complex binder,” Sci. China Technol. Sci., vol. 54, no. 2, pp. 388–394, Feb. 2011.

[71] S. Choi, J.-M. Kim, D. Han, and J.-H. Kim, “Hydration properties of ladle furnace slag powder rapidly cooled by air,” Constr. Build. Mater., vol. 113, pp. 682–690, Jun. 2016.

78

[72] Y.-N. Sheen, D.-H. Le, and T.-H. Sun, “Innovative usages of stainless steel slags in developing self-compacting concrete,” Constr. Build. Mater., vol. 101, pp. 268–276, Dec. 2015.

[73] S. Karen, Calcium Aluminate cements. Butterworth-Heinemann, 2003. [74] K. L. Scrivener, J.-L. Cabiron, and R. Letourneux, “High-performance

concretes from calcium aluminate cements,” Cem. Concr. Res., vol. 29, no. 8, pp. 1215–1223, Aug. 1999.

[75] J. Bensted, “calcium aluminate cements,” in Structure and Performance of Cements, 2nd ed., 2002, pp. 114–138.

[76] E. Gartner, J. F. Young, D. Damidot, and I. Jawed, “Hydration of Portland Cement.,” in Structure and Performance of Cements, 2002, pp. 57–113.

[77] “Cement Standards and Concrete Standards.” [Online]. Available: https://www.astm.org/Standards/cement-and-concrete-standards.html. [Accessed: 14-Nov-2019].

[78] I. Anger, E. Volceanov, G. Plopeanu, L. Ilie, L. G. Popescu, and G. A. Moise, “Impact of metallurgical ladle furnace slag on acid soil remediation in agriculture,” Rev. Romana Mater. Romanian J. Mater., vol. 48, no. 1, pp. 76–82, 2018.

[79] T. A. Branca and V. Colla, “Possible Uses of Steelmaking Slag in Agriculture: An Overview,” Mater. Recycl. - Trends Perspect., Mar. 2012.

[80] A. C. F. Deus, R. M. de A. Bertani, G. C. Meirelles, A. de A. V. L. Soares, L. L. Q. Moreira, and L. T. B. and D. M. Fernandes, “The Comprehensive Utilization of Steel Slag in Agricultural Soils,” Recovery Util. Metall. Solid Waste, Dec. 2018.

[81] “Fertilizer,” Euroslag. [Online]. Available: https://www.euroslag.com/applications/fertilizer/. [Accessed: 16-May-2019].

[82] J. M. Manso, V. Ortega-López, J. A. Polanco, and J. Setién, “The use of ladle furnace slag in soil stabilization,” Constr. Build. Mater., vol. 40, pp. 126–134, Mar. 2013.

[83] J. M. Manso, M. Losañez, J. A. Polanco, and J. J. Gonzalez, “Ladle Furnace Slag in Construction,” J. Mater. Civ. Eng., vol. 17, no. 5, pp. 513–518, Sep. 2005.

[84] M. Skaf, V. Ortega-López, J. A. Fuente-Alonso, A. Santamaría, and J. M. Manso, “Ladle furnace slag in asphalt mixes,” Constr. Build. Mater., vol. 122, pp. 488–495, Sep. 2016.

[85] Á. Rodriguez, J. M. Manso, Á. Aragón, and J. J. Gonzalez, “Strength and workability of masonry mortars manufactured with ladle furnace slag,” Resour. Conserv. Recycl., vol. 53, no. 11, pp. 645–651, Sep. 2009.

[86] H. Nguyen, V. Carvelli, E. Adesanya, P. Kinnunen, and M. Illikainen, “High performance cementitious composite from alkali-activated ladle slag reinforced with polypropylene fibers,” Cem. Concr. Compos., vol. 90, pp. 150–160, Jul. 2018.

79

[87] H. Nguyen, M. Staudacher, P. Kinnunen, V. Carvelli, and M. Illikainen, “Multi-fiber reinforced ettringite-based composites from industrial side streams,” J. Clean. Prod., vol. 211, pp. 1065–1077, Feb. 2019.

[88] H. Nguyen, P. Kinnunen, V. Carvelli, M. Mastali, and M. Illikainen, “Strain hardening polypropylene fiber reinforced composite from hydrated ladle slag and gypsum,” Compos. Part B Eng., vol. 158, pp. 328–338, Feb. 2019.

[89] H. Nguyen, P. Kinnunen, V. Carvelli, and M. Illikainen, “Durability of ettringite-based composite reinforced with polypropylene fibers under combined chemical and physical attack,” Cem. Concr. Compos., vol. 102, pp. 157–168, Sep. 2019.

[90] SFS (Finnish Standards Association), “Methods of testing cement. Part 1: Determination of strength” SFS-EN 196-1:2016:en. 2016.

[91] I. Vegas, J. Urreta, M. Frías, and R. García, “Freeze–thaw resistance of blended cements containing calcined paper sludge,” Constr. Build. Mater., vol. 23, no. 8, pp. 2862–2868, Aug. 2009.

[92] S. C. Pal, A. Mukherjee, and S. R. Pathak, “Investigation of hydraulic activity of ground granulated blast furnace slag in concrete,” Cem. Concr. Res., vol. 33, no. 9, pp. 1481–1486, Sep. 2003.

[93] D. Li, Z. Xu, Z. Luo, Z. Pan, and L. Cheng, “The activation and hydration of glassy cementitious materials,” Cem. Concr. Res., vol. 32, no. 7, pp. 1145–1152, Jul. 2002.

[94] N. Y. Mostafa, S. A. S. El-Hemaly, E. I. Al-Wakeel, S. A. El-Korashy, and P. W. Brown, “Characterization and evaluation of the hydraulic activity of water-cooled slag and air-cooled slag,” Cem. Concr. Res., vol. 31, no. 6, pp. 899–904, May 2001.

[95] L. Muhmood, S. Vitta, and D. Venkateswaran, “Cementitious and pozzolanic behavior of electric arc furnace steel slags,” Cem. Concr. Res., vol. 39, no. 2, pp. 102–109, Feb. 2009.

[96] M. Salman et al., “Cementitious binders from activated stainless steel refining slag and the effect of alkali solutions,” J. Hazard. Mater., vol. 286, pp. 211–219, Apr. 2015.

[97] SFS-EN 451-1, Method of testing fly ash- Part 1: determination of free calcium oxide. 2004.

[98] S. Chithiraputhiran and N. Neithalath, “Isothermal reaction kinetics and temperature dependence of alkali activation of slag, fly ash and their blends,” Constr. Build. Mater., vol. 45, pp. 233–242, Aug. 2013.

[99] S. Klaus and A. Buhr, “Hydration of calcium aluminate cement phase CA and CA2 in refractory applications.” [Online]. Available: http://www.almatis.com/media/1355/hydration-of-calcium-aluminate-cement-phase-ca-and-ca2-in-refractory-applications.pdf. [Accessed: 01-Mar-2019].

80

[100] D. Krizan and B. Zivanovic, “Effects of dosage and modulus of water glass on early hydration of alkali–slag cements,” Cem. Concr. Res., vol. 32, no. 8, pp. 1181–1188, Aug. 2002.

[101] M. B. Haha, B. Lothenbach, G. Le Saout, and F. Winnefeld, “Influence of slag chemistry on the hydration of alkali-activated blast-furnace slag — Part I: Effect of MgO,” Cem. Concr. Res., vol. 41, no. 9, pp. 955–963, Sep. 2011.

[102] D. G. Snelson, S. Wild, and M. O’Farrell, “Heat of hydration of Portland Cement–Metakaolin–Fly ash (PC–MK–PFA) blends,” Cem. Concr. Res., vol. 38, no. 6, pp. 832–840, Jun. 2008.

[103] B. Walkley, R. San Nicolas, M.-A. Sani, J. D. Gehman, J. S. J. van Deventer, and J. L. Provis, “Phase evolution of Na 2 O–Al 2 O 3 –SiO 2 –H 2 O gels in synthetic aluminosilicate binders,” Dalton Trans, vol. 45, no. 13, pp. 5521–5535, 2016.

[104] E. Kapeluszna, Ł. Kotwica, A. Różycka, and Ł. Gołek, “Incorporation of Al in C-A-S-H gels with various Ca/Si and Al/Si ratio: Microstructural and structural characteristics with DTA/TG, XRD, FTIR and TEM analysis,” Constr. Build. Mater., vol. 155, pp. 643–653, Nov. 2017.

[105] I. Ismail, S. A. Bernal, J. L. Provis, R. San Nicolas, S. Hamdan, and J. S. J. van Deventer, “Modification of phase evolution in alkali-activated blast furnace slag by the incorporation of fly ash,” Cem. Concr. Compos., vol. 45, pp. 125–135, Jan. 2014.

[106] M. Criado, A. Palomo, and A. Fernández-Jiménez, “Alkali activation of fly ashes. Part 1: effect of curing conditions on the carbonation of the reaction products,” Fuel, vol. 84, no. 16, pp. 2048–2054, Nov. 2005.

[107] I. Lecomte, C. Henrist, M. Liégeois, F. Maseri, A. Rulmont, and R. Cloots, “(Micro)-structural comparison between geopolymers, alkali-activated slag cement and Portland cement,” J. Eur. Ceram. Soc., vol. 26, no. 16, pp. 3789–3797, 2006.

[108] D. Panias, I. P. Giannopoulou, and T. Perraki, “Effect of synthesis parameters on the mechanical properties of fly ash-based geopolymers,” Colloids Surf. Physicochem. Eng. Asp., vol. 301, no. 1, pp. 246–254, Jul. 2007.

[109] A. Fernández-Jiménez, F. Puertas, I. Sobrados, and J. Sanz, “Structure of Calcium Silicate Hydrates Formed in Alkaline-Activated Slag: Influence of the Type of Alkaline Activator,” J. Am. Ceram. Soc., vol. 86, no. 8, pp. 1389–1394, Aug. 2003.

[110] P. Yu, R. J. Kirkpatrick, B. Poe, P. F. McMillan, and X. Cong, “Structure of Calcium Silicate Hydrate (C-S-H): Near-, Mid-, and Far-Infrared Spectroscopy,” J. Am. Ceram. Soc., vol. 82, no. 3, pp. 742–748, Mar. 1999.

[111] A. D. Hounsi, G. L. Lecomte-Nana, G. Djétéli, and P. Blanchart, “Kaolin-based geopolymers: Effect of mechanical activation and curing process,” Constr. Build. Mater., vol. 42, pp. 105–113, May 2013.

81

[112] M. Chi, J. Chang, R. Huang, M. Chi, J. Chang, and R. Huang, “Strength and Drying Shrinkage of Alkali-Activated Slag Paste and Mortar, Strength and Drying Shrinkage of Alkali-Activated Slag Paste and Mortar,” Adv. Civ. Eng. Adv. Civ. Eng., vol. 2012, 2012, p. e579732, May 2012.

[113] Cartwright Christopher, Rajabipour Farshad, and Radlińska Aleksandra, “Shrinkage Characteristics of Alkali-Activated Slag Cements,” J. Mater. Civ. Eng., vol. 27, no. 7, p. B4014007, Jul. 2015.

[114] C. Duran Atiş, C. Bilim, Ö. Çelik, and O. Karahan, “Influence of activator on the strength and drying shrinkage of alkali-activated slag mortar,” Constr. Build. Mater., vol. 23, no. 1, pp. 548–555, Jan. 2009.

[115] M. Palacios and F. Puertas, “Effect of shrinkage-reducing admixtures on the properties of alkali-activated slag mortars and pastes,” Cem. Concr. Res., vol. 37, no. 5, pp. 691–702, May 2007.

[116] A. J. Majumbar, B. Singh, and R. N. Edmonds, “Hydration of mixtures of C12A7 and granulated blastfurnace slag,” Cem. Concr. Res., vol. 19, no. 6, pp. 848–856, Nov. 1989.

[117] J. Bensted, “δ-dicalcium silicate and its hydraulicity,” Cem. Concr. Res., vol. 8, no. 1, pp. 73–76, Jan. 1978.

[118] L. Kriskova et al., “Influence of mechanical and chemical activation on the hydraulic properties of gamma dicalcium silicate,” Cem. Concr. Res., vol. 55, pp. 59–68, Jan. 2014.

[119] N. C. Collier, “Transition and decomposition temperatures of cement phases – A collection of thermal analysis data,” Ceram. - Silik., pp. 1–10, Sep. 2016.

[120] N. Ukrainczyk, T. Matusinovic, S. Kurajica, B. Zimmermann, and J. Sipusic, “Dehydration of a layered double hydroxide—C2AH8,” Thermochim. Acta, vol. 464, no. 1, pp. 7–15, Nov. 2007.

[121] M. A. Chavda, S. A. Bernal, D. C. Apperley, H. Kinoshita, and J. L. Provis, “Identification of the hydrate gel phases present in phosphate-modified calcium aluminate binders,” Cem. Concr. Res., vol. 70, pp. 21–28, Apr. 2015.

[122] J. Skibsted, E. Henderson, and H. J. Jakobsen, “Characterization of calcium aluminate phases in cements by aluminum-27 MAS NMR spectroscopy,” Inorg. Chem., vol. 32, no. 6, pp. 1013–1027, Mar. 1993.

[123] X. Cong and R. J. Kirkpatrick, “Hydration of Calcium Aluminate Cements: A Solid-State 27 Al NMR Study,” J. Am. Ceram. Soc., vol. 76, no. 2, pp. 409–416, 1993.

[124] D. Müller, A. Rettel, W. Gessner, and G. Scheler, “An application of solid-state magic-angle spinning 27Al NMR to the study of cement hydration,” J. Magn. Reson. 1969, vol. 57, no. 1, pp. 152–156, Mar. 1984.

[125] A. Quennoz and K. L. Scrivener, “Hydration of C3A–gypsum systems,” Cem. Concr. Res., vol. 42, no. 7, pp. 1032–1041, Jul. 2012.

[126] C. Shi and R. L. Day, “A calorimetric study of early hydration of alkali-slag cements,” Cem. Concr. Res., vol. 25, no. 6, pp. 1333–1346, Aug. 1995.

82

[127] H. Minard, S. Garrault, L. Regnaud, and A. Nonat, “Mechanisms and parameters controlling the tricalcium aluminate reactivity in the presence of gypsum,” Cem. Concr. Res., vol. 37, no. 10, pp. 1418–1426, Oct. 2007.

[128] S. Pourchet, L. Regnaud, J.-P. Perez, and A. Nonat, “Early C3A hydration in the presence of different kinds of calcium sulfate.,” Cem. Concr. Res., vol. 39, pp. 989–996, Nov. 2009.

[129] A. Quennoz, “Hydration of C3A with Calcium Sulfate Alone and in the Presence of Calcium Silicate,” 2011.

[130] K. L. Scrivener, The development of microstructure during the hydration of Portland cement. University of London, 1984.

[131] H. El-Didamony, “Application of differential thermogravimetry to the hydration of expansive cement pastes,” Thermochim. Acta, vol. 35, no. 2, pp. 201–209, Jan. 1980.

[132] W. Sha and G. B. Pereira, “Differential scanning calorimetry study of hydrated ground granulated blast-furnace slag,” Cem. Concr. Res., vol. 31, no. 2, pp. 327–329, Feb. 2001.

[133] K. De Weerdt, K. O. Kjellsen, E. Sellevold, and H. Justnes, “Synergy between fly ash and limestone powder in ternary cements,” Cem. Concr. Compos., vol. 33, no. 1, pp. 30–38, Jan. 2011.

[134] H. G. Midgley and P. Bhaskara Rao, “Formation of stratlingite, 2CaO.SiO2.Al2O3.8H2O, in relation to the hydration of high alumina cement,” Cem. Concr. Res., vol. 8, no. 2, pp. 169–172, Mar. 1978.

[135] S. Karen and H. F. W. Taylor, microstructural development in pastes of a calcium aluminate cement, Calcium Aluminate Cements: Proceedings of a Symposium. CRC Press, 2014.

[136] C. Bradbury, P. M. Callaway, and D. D. Double, “The conversion of high alumina cement/concrete,” Mater. Sci. Eng., vol. 23, no. 1, pp. 43–53, Apr. 1976.

[137] B. Pacewska and M. Nowacka, “Studies of conversion progress of calcium aluminate cement hydrates by thermal analysis method,” J. Therm. Anal. Calorim., vol. 117, no. 2, pp. 653–660, Aug. 2014.

[138] B. S. Cho, H. H. Lee, and Y. C. Choi, “Effects of aluminate rich slag on compressive strength, drying shrinkage and microstructure of blast furnace slag cement,” Constr. Build. Mater., vol. 140, pp. 293–300, Jun. 2017.

[139] T. Bakharev, J. G. Sanjayan, and Y.-B. Cheng, “Effect of admixtures on properties of alkali-activated slag concrete,” Cem. Concr. Res., vol. 30, no. 9, pp. 1367–1374, Sep. 2000.

[140] M. García-Maté, A. G. De la Torre, L. León-Reina, E. R. Losilla, M. A. G. Aranda, and I. Santacruz, “Effect of calcium sulfate source on the hydration of calcium sulfoaluminate eco-cement,” Cem. Concr. Compos., vol. 55, pp. 53–61, Jan. 2015.

83

[141] A. Telesca, M. Marroccoli, M. L. Pace, M. Tomasulo, G. L. Valenti, and P. J. M. Monteiro, “A hydration study of various calcium sulfoaluminate cements,” Cem. Concr. Compos., vol. 53, pp. 224–232, Oct. 2014.

[142] I. Odler and J. Colán-Subauste, “Investigations on cement expansion associated with ettringite formation,” Cem. Concr. Res., vol. 29, no. 5, pp. 731–735, May 1999.

[143] “Eurocodes: Building the future - The European Commission website on the Eurocodes 132.” [Online]. Available: https://eurocodes.jrc.ec.europa.eu/showpage.php?id=132. [Accessed: 26-Feb-2019].

[144] C. Elibol and O. Sengul, “Effects of Activator Properties and Ferrochrome Slag Aggregates on the Properties of alkali-activated Blast Furnace Slag Mortars,” Arab. J. Sci. Eng., vol. 41, no. 4, pp. 1561–1571, Apr. 2016.

[145] P. N. Lemougna, K. J. D. MacKenzie, and U. F. C. Melo, “Synthesis and thermal properties of inorganic polymers (geopolymers) for structural and refractory applications from volcanic ash,” Ceram. Int., vol. 37, no. 8, pp. 3011–3018, Dec. 2011.

[146] W. D. A. Rickard, A. van Riessen, and P. Walls, “Thermal Character of Geopolymers Synthesized from Class F Fly Ash Containing High Concentrations of Iron and α-Quartz,” Int. J. Appl. Ceram. Technol., vol. 7, no. 1, pp. 81–88, Jan. 2010.

[147] J. Ye, W. Zhang, and D. Shi, “Effect of elevated temperature on the properties of geopolymer synthesized from calcined ore-dressing tailing of bauxite and ground-granulated blast furnace slag,” Constr. Build. Mater., vol. 69, pp. 41–48, 2014.

[148] T. Bakharev, “Thermal behaviour of geopolymers prepared using class F fly ash and elevated temperature curing,” Cem. Concr. Res., vol. 36, no. 6, pp. 1134–1147, Jun. 2006.

[149] ASTM C34-03, “Specification for Structural Clay Load-Bearing Wall Tile,” ASTM International, 2003.

[150] P. Duxson, G. C. Lukey, and J. S. J. van Deventer, “Thermal evolution of metakaolin geopolymers: Part 1 – Physical evolution,” J. Non-Cryst. Solids, vol. 352, no. 52, pp. 5541–5555, Dec. 2006.

[151] J. L. Provis, C. Z. Yong, P. Duxson, and J. S. J. van Deventer, “Correlating mechanical and thermal properties of sodium silicate-fly ash geopolymers,” Colloids Surf. Physicochem. Eng. Asp., vol. 336, no. 1–3, pp. 57–63, Mar. 2009.

[152] H. Y. Zhang, V. Kodur, B. Wu, L. Cao, and F. Wang, “Thermal behavior and mechanical properties of geopolymer mortar after exposure to elevated temperatures,” Constr. Build. Mater., vol. 109, pp. 17–24, Apr. 2016.

[153] S. Mohd, M. Norkhairunnisa, and F. Mustapha, “A review on thermophysical evaluation of alkali-activated geopolymers,” Ceram. Int., vol. 41, no. 3, pp. 4273–4281, 2015.

84

85

Appendices

Appendix A

Fig. A1. XRD of HLS samples with different w/b ratio (a) 0.35 and (b) 0.50. The letters

indicate A-mayenite (C12A7), B-calcio-olivine (γ-C2S), C-periclase (MgO), D-tricalcium

aluminate (C3A), S -strätlingite (C2ASH8), x-dicalcium aluminate hydrate (C2AH8) and y-

katoite (C3AH6) (Reprinted by permission from Paper III © 2019 Elsevier).

86

Ap

pe

nd

ix B

Fig

. B

1.

Vis

ua

l a

pp

ea

ran

ce

of

HL

S,

AA

LS

, an

d C

AC

mo

rtars

be

fore

an

d a

fte

r e

xp

os

ure

to

ele

va

ted

tem

pe

ratu

res

(R

ep

rin

ted

by

pe

rmis

sio

n f

rom

Pa

pe

r v

© I

CE

Pu

bli

sh

ing

).

87

Appendix C

Fig. C1. BSE-SEM image of the mortars before exposure; (a) HLS_Fs, (c) HLS_Ss, (e)

AALS_Fs and (g) AALS_Ss and after exposure to 1000 °C; (b) HLS_Fs, (d) HLS_Ss, (f)

AALS_Fs and (h) AALS_Ss. Fs: Ferrochrome aggregate; B: Binder and LS: Unreacted

ladle slag/aggregate. Red rings show the area of sintering/densification in the mortar

and rectangular box shows the interfacial zone (ITZ) between the alkali binder and the

ferrochrome aggregate (Reprinted by permission from Paper V © ICE Publishing).

88

89

Original publications

I Adesanya, E., Ohenoja, K., Kinnunen, P., & Illikainen, M. (2017). Alkali activation of ladle slag from steel-making process. Journal of Sustainable Metallurgy 3(2), 300-310

II Adesanya, E., Ohenoja, K., Kinnunen, P., & Illikainen, M. (2017). Properties and durability of alkali-activated ladle slag. Journal of Materials and Structures, 50:255.

III Adesanya, E., Sreenivasan, H., Kantola, A.M, Telkki, V.-V., Ohenoja, K., Kinnunen, P., & Illikainen, M. (2018). Ladle slag cement – Characterization of hydration and conversion. Construction and Building Materials, 193, 128-134.

IV Nguyen, H., Adesanya, E., Ohenoja, K., Kriskova, L., Pontikes, Y., Kinnunen, P., & Illikainen, M. (2019). Byproduct-based ettringite binder – A synergy between ladle slag and gypsum. Construction and Building Materials, 197, 143-151.

V Adesanya, E., Karhu, M., Ismailov, A., Ohenoja, K., Kinnunen, P., & Illikainen, M. Thermal behaviour of ladle slag mortars containing ferrochrome slag aggregates. Advances in Cement Research, In press.

Reprinted with permission from Springer Nature (I, II), Elsevier Ltd. (III, IV)

and ICE Publishing (V).

The original publications are not included in the electronic version of this thesis.

90

A C T A U N I V E R S I T A T I S O U L U E N S I S

Book orders:Virtual book store

http://verkkokauppa.juvenesprint.fi

S E R I E S C T E C H N I C A

709. Lappi, Teemu (2019) Digitalizing Finland : governance of government ICT projects

710. Pitkänen, Olli (2019) On-device synthesis of customized carbon nanotubestructures

711. Vielma, Tuomas (2019) Thermodynamic properties of concentrated zinc bearingsolutions

712. Ramasetti, Eshwar Kumar (2019) Modelling of open-eye formation and mixingphenomena in a gas-stirred ladle for different operating parameters

713. Javaheri, Vahid (2019) Design, thermomechanical processing and inductionhardening of a new medium-carbon steel microalloyed with niobium

714. Hautala, Ilkka (2019) From dataflow models to energy efficient application specificprocessors

715. Ruokamo, Simo (2019) Single shared model approach for building informationmodelling

716. Isohookana, Matti (2019) Taistelunkestävä hajaspektritietovuo kansalliseensotilasilmailuun

717. Joseph, Nina (2019) CuMoO4 : a microwave dielectric and thermochromicceramic with ultra-low fabrication temperature

718. Kühnlenz, Florian (2019) Analyzing flexible demand in smart grids

719. Sun, Jia (2019) Speeding up the settling of switched-capacitor amplifier blocks inanalog-to-digital converters

720. Lähetkangas, Kalle (2019) Special applications and spectrum sharing with LSA

721. Kiventerä, Jenni (2019) Stabilization of sulphidic mine tailings by differenttreatment methods : heavy metals and sulphate immobilization

722. Fylakis, Angelos (2019) Data hiding algorithms for healthcare applications

723. Juuso, Ilkka (2019) A cellular automaton environment for the complex system ofspeech

724. Aho, Pekka (2019) Automated state model extraction, testing and changedetection through graphical user interface

725. Bergna, Davide (2019) Activated carbon from renewable resources :carbonization, activation and use

726. Nuottila, Jouko (2019) Flexibility in agile projects : Contracting practices andorganisational arrangements

C727etukansi.fm Page 2 Friday, November 8, 2019 2:27 PM

UNIVERSITY OF OULU P .O. Box 8000 F I -90014 UNIVERSITY OF OULU FINLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

University Lecturer Tuomo Glumoff

University Lecturer Santeri Palviainen

Senior research fellow Jari Juuti

Professor Olli Vuolteenaho

University Lecturer Veli-Matti Ulvinen

Planning Director Pertti Tikkanen

Professor Jari Juga

University Lecturer Anu Soikkeli

Professor Olli Vuolteenaho

Publications Editor Kirsti Nurkkala

ISBN 978-952-62-2451-0 (Paperback)ISBN 978-952-62-2452-7 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

U N I V E R S I TAT I S O U L U E N S I SACTAC

TECHNICA

OULU 2019

C 727

Elijah D. Adesanya

A CEMENTITIOUS BINDER FROM HIGH-ALUMINASLAG GENERATED IN THE STEELMAKING PROCESS

UNIVERSITY OF OULU GRADUATE SCHOOL;UNIVERSITY OF OULU,FACULTY OF TECHNOLOGY

C 727

AC

TAE

lijah D. A

desanyaC727etukansi.fm Page 1 Friday, November 8, 2019 2:27 PM