particle size distribution and suspension stability in aqueous

ABCDEFG

UNIVERSITY OF OULU P .O. B 00 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

S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Esa Hohtola

University Lecturer Santeri Palviainen

Postdoctoral research fellow Sanna Taskila

Professor Olli Vuolteenaho

University Lecturer Veli-Matti Ulvinen

Director Sinikka Eskelinen

Professor Jari Juga


Publications Editor Kirsti Nurkkala

ISBN 978-952-62-0549-6 (Paperback)ISBN 978-952-62-0550-2 (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


TECHNICA

OULU 2014

C 501

Katja Ohenoja

PARTICLE SIZE DISTRIBUTION AND SUSPENSION STABILITYIN AQUEOUS SUBMICRON GRINDING OF CaCO3 AND TiO2

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

C 501

ACTA

Katja O

henoja

C501etukansi.kesken.fm Tuesday, September 2, 2014 9:41 AM

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 SC Te c h n i c a 5 0 1

KATJA OHENOJA

PARTICLE SIZE DISTRIBUTIONAND SUSPENSION STABILITY IN AQUEOUS SUBMICRON GRINDING OF CaCO3 AND TiO2

Academic dissertation to be presented with the assent ofthe Doctoral Training Committee of Technology andNatural Sciences of the University of Oulu for publicdefence in Auditorium IT116, Linnanmaa, on 10 October2014, at 12 noon

UNIVERSITY OF OULU, OULU 2014

Copyright © 2014Acta Univ. Oul. C 501, 2014

Supervised byProfessor Jouko NiinimäkiDoctor Mirja Illikainen

Reviewed byDoctor Ergün AltinDoctor Carsten Schilde

ISBN 978-952-62-0549-6 (Paperback)ISBN 978-952-62-0550-2 (PDF)

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

Cover DesignRaimo Ahonen

JUVENES PRINTTAMPERE 2014

OpponentProfessor Kari Heiskanen

Ohenoja, Katja, Particle size distribution and suspension stability in aqueoussubmicron grinding of CaCO3 and TiO2. University of Oulu Graduate School; University of Oulu, Faculty of TechnologyActa Univ. Oul. C 501, 2014University of Oulu, P.O. Box 8000, FI-90014 University of Oulu, Finland

Abstract

During the past decade submicron and nanoparticles have aroused a wide interest and gained newapplications due to their high surface area and strength. Grinding with a wet stirred media mill isusually the last process step before the submicron or nanoparticles are added to an application, andthe step where the final particle size distribution is achieved. Since stirred media milling is anenergy-intensive process, energy efficiency should be optimized. This can be done by determiningthe optimum operational parameters for the mill and using the highest possible solidsconcentration. The solids concentration can be increased by controlling particle-particleinteractions with stabilization chemicals, e.g. polymers.

This thesis concerns parameters and grinding aids affecting the particle size distribution andsuspension stability of the aqueous submicron grinding of calcium carbonate (CaCO3) andtitanium dioxide (TiO2) in stirred media mills. TiO2 particles are aggregates produced via abottom-up method, while CaCO3 are primary mineral particles produced by a top-down method.

The most energy efficient grinding of TiO2 to a 300 nm particle size with the narrowestpossible particle size distribution was obtained with the lowest stress energy, implying thesmallest grinding medium size. It was observed that electrosteric stabilization with sodiumpolyacrylates was effective for TiO2, and sodium polyacrylate with a molecular weight of 12500g/mol was found to be the most effective for reducing the viscosity of the suspension.

As with TiO2, electrosteric stabilization with sodium polyacrylates was also found to beeffective for CaCO3, but in this case sodium polyacrylate with a lower polydispersity index wasmore effective, showing a better stabilization potential in micron and submicron grinding andreducing the viscosity and particle size to a greater extent. Nanogrinding experiments wereperformed for a CaCO3 suspension with low PDI sodium polyacrylate and it was found to bepossible to obtain a particle size of 26 nm, smaller than any size previously reported when grindingCaCO3.

Keywords: dispersing agent, grinding, grinding limit, industrial mineral, milling,nanoparticle, polymer, stabilization

Ohenoja, Katja, Partikkelikokojakauma ja lietteen stabiilisuus CaCO3:n ja TiO2:nvesipohjaisessa submikronijauhatuksessa. Oulun yliopiston tutkijakoulu; Oulun yliopisto, Teknillinen tiedekuntaActa Univ. Oul. C 501, 2014Oulun yliopisto, PL 8000, 90014 Oulun yliopisto

Tiivistelmä

Viimeisen kymmenen vuoden aikana alle yhden mikrometrin partikkelit ovat herättäneet kiin-nostusta ja niille on kehitetty uusia sovelluksia niiden suuren pinta-alan ja lujuuden ansiosta.Ultrahienojauhatus märkähelmimyllyllä on useimmiten viimeinen prosessivaihe ennen partikke-lien lisäämistä sovelluskohteeseen ja siinä saavutetaan partikkelien lopullinen partikkelikokoja-kauma. Helmimyllyjauhatuksen energiankulutus minimoidaan etsimällä optimioperointipara-metrit kullekin jauhatusprosessille ja käyttämällä korkeinta mahdollista suspension kuiva-ainepi-toisuutta. Suspension kuiva-ainepitoisuutta voidaan nostaa hallitsemalla partikkelien välisiä vuo-rovaikutuksia stabilointiaineilla, kuten polymeereillä.

Tässä väitöskirjassa tutkittiin operointiparametrien ja jauhatusapuaineiden vaikutusta titaani-dioksidin (TiO2) ja kalsiumkarbonaatin (CaCO3) partikkelikokojakaumaan ja lietteen stabiili-suuteen submikronijauhatuksessa. Tutkitut TiO2-partikkelit olivat aggregaatteja, jotka oli val-mistettu sulfaattiprosessilla saostamalla, ja tutkitut CaCO3-partikkelit olivat primäärisiä mine-raalipartikkeleita.

TiO2-partikkeleille saavutettiin energiatehokkain jauhatus ja samalla toivottu partikkelikoko-jakauma, eli mediaani 300 nm ja mahdollisimman kapea jakauma, pienillä helmillä, jotka aiheut-tavat partikkeleihin pienimmän puristusenergian. Elektrosteerinen stabilointi käyttämällä natri-umpolyakrylaatteja stabilointiaineena havaittiin tehokkaaksi menetelmäksi hallita TiO2-partikke-lien välisiä vuorovaikutuksia. Natriumpolyakrylaatti, jonka molekyylimassa oli 12500 g/mol, olitehokkain TiO2-partikkeleille alentaen suspension viskositeettiä eniten. Myös CaCO3-partikke-leille elektrosteerinen stabilointi natriumpolyakrylaatteja käyttäen oli tehokkain stabilointimene-telmä. Myös natriumpolyakrylaattien polydispersiteetti-indeksin vaikutusta tutkittiin CaCO3-suspensioille. Tulokset osoittivat matalan polydispersiteetti-indeksin olevan tehokkaampi alenta-en viskositteettia ja pienentäen partikkelikokoa tehokkaammin kuin natriumpolyakrylaatti, jollaoli korkeampi polydispersitetti-indeksi. Tämän vuoksi natriumpolyakrylaatti, jolla oli matalapolydispersiteetti-indeksi, valittiin nanojauhatuskokeisiin. Kokeissa CaCO3-partikkelit saatiinjauhettua 26 nm kokoon, joka on pienin koskaan aiemmin jauhamalla saavutettu koko CaCO3-partikkeleille.

Asiasanat: dispergointiaine, hienonnus, jauhatus, jauhatusraja, nanopartikkeli,polymeeri, stabilointi, teollisuusmineraali

“A scientist in his laboratory is not a mere technician:

he is also a child confronting natural phenomena

that impress him as though they were fairy tales.”

(Marie Curie)

Dedicated to my dear, sweet son Oskari

9

Acknowledgements

The present work was performed at the Fibre and Particle Engineering

Laboratory, University of Oulu, during the years 2011-2013. Three research

exchanges took place with the Institute for Particle Technology at TU

Braunschweig for a total of 5 months. The financial support from the Finnish

Funding Agency for Technology and Innovation (TEKES), industrial partners in

the EFFSTA project and the Academy of Finland (AKA) in the form of a German

Academic Exchange Service (DAAD) grant is gratefully acknowledged. I also

greatly appreciate the personal grants received for this work from the Jenny and

Antti Wihuri Foundation, the Riitta and Jorma J. Takanen Foundation and the

Tauno Tönning Foundation, and similarly travel grants from the University of

Oulu Graduate School (UniOGS) which enabled me to participate in two

international conferences.

In addition to the financial support, this thesis would not have been possible

without the support received from many people whom I wish to remember here.

First and foremost, I wish to express my deepest gratitude to my principal

supervisor, Prof. Jouko Niinimäki, for giving me the opportunity to produce this

thesis and encouraging me throughout the work. Secondly, I would express my

warmest thanks to my second supervisor, Dr. Mirja Illikainen, for her support and

helpful advice whenever needed. Then I would like to express my particular

thanks to Prof. Dr.-Ing. Arno Kwade for giving me the opportunity to visit his

institute and teaching me a lot about stirred media mills. I would also like to thank

Dr.-Ing. Sandra Breitung-Faes for advice, cooperation and leisure-time activities

in Braunschweig.

Official reviewers of this thesis, Dr.-Ing. Carsten Schilde from TU

Braunschweig and Dr. Ergün Altin from Hosokawa Alpine are greatly appreciated

for their efforts and valuable commentary on the manuscript of this thesis. I

would also like to express my gratitude to Malcolm Hicks for revising the English

language of this thesis.

I would like to express my thanks to Juha Saari for his helpful advice and

cooperation in writing some of the papers. I would also like to thank Esa Latva-

Nirva, Kaarina Heikkilä, Petteri Piltonen and the other people who worked on the

EFFSTA project. All the colleagues, friends, technical personnel and students I

have worked with in Oulu and Braunschweig are acknowledged. Particular thanks

go to Jarno Karvonen and Helmut Seeland for their help in the laboratory and to

10

Pasi Karinkanta, with whom I have shared an office all these years, for many

fruitful conversations.

Finally, I wish to express my thanks to my parents, Päivi and Ilpo, for their

encouragement throughout my whole education, to my sister Sari for her

friendship, and to my dear husband Markku, without whose support, patience and

love this would not have been possible.

Oulu, July 2014 Katja Ohenoja

11

Abbreviations

CaCO3 Calcium carbonate

GCC Ground calcium carbonate

IEP Isoelectric point, ζPOT=0

NaPa Sodium polyacrylate

NaPa4000 Sodium polyacrylate, molecular weight 4000 g/mol




NaPaPDI>2 Sodium polyacrylate, polydispersity index over two

NaPaPDI=1.2 Sodium polyacrylate, polydispersity index 1.2

PCC Precipitated calcium carbonate

PSD Particle size distribution

TiO2 Titanium dioxide

b Burger vector length [m]

cm Solids concentration by weight [-]

D Dose of sodium polyacrylate [wtNaPa/TiO2% or wtNaPa/CaCO3%]

dGM Grinding media size [m]

d10 Particle size for cumulative 10% finer [µm]

d50 Median particle size [µm]

d90 Particle size for cumulative 90% finer [µm]

Em Specific energy [kJ/kg]

EFeed Young´s modulus of the feed material [Pa]

EGM Young´s modulus of the grinding media [Pa]

G Shear modulus [Pa]

H Hardness [Pa]

L Minimum separation distance [m]

mFeed Mass of dry limestone [kg]

Mp Mass flow rate [kg/s]

Mn Number-averaged molecular weight of sodium polyacrylate [g/mol]

Mw Weight-averaged molecular weight of sodium polyacrylate [g/mol]

Npass Number of passages through the mill [-]

Pn Power input [kJ]

P0 No-load power [kJ]

PDI Polydispersity index of sodium polyacrylates, Mw/Mn [-]

12

SEGM Maximum stress energy of the grinding media [J]

SEP Stress energy of the product particle [J]

SI Stress intensity, SEGM/d503 [N/m2]

tG Grinding time [s]

v Poisson´s ratio [-]

vt Stirrer tip speed [m/s]

WPSD Span value indicating the width of the particle size distribution, (d90-

d10)/d50 [-]

x10 Primary particle size for cumulative 10% finer [nm]

x50 Median primary particle size [nm]

x90 Primary particle size for cumulative 90% finer [nm]

xcrys Crystallite size [nm]

σCS Compression strength of material [N/m2]

Viscosity of a suspension [Pas]

γ Shear rate [s-1]

GM Grinding media density [kg/m3]

ζPOT Zeta potential [mV]

GM Grinding media filling ratio of the mill [vol.%]

τ Shear stress [Pa]

τB Bingham yield stress [Pa]

PL Plastic viscosity [Pas]

13

List of original publications

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

text by their Roman numerals:

I Ohenoja K, Illikainen M & Niinimäki J (2013) Effect of operational parameters and stress energies on the particle size distribution of TiO2 pigment in stirred media milling. Powder Technol 234: 91–96.

II Ohenoja K, Saari J, Illikainen M & Niinimäki J (2014) Effect of molecular weight of sodium polyacrylates on the particle size distribution and stability of a TiO2 suspension in aqueous stirred media milling. Powder Technol 262: 188–193.

III Ohenoja K, Saari J, Illikainen M & Niinimäki J (2013) Effect of the molecular structure of polymer dispersants on limestone suspension properties in fine grinding. J Chem Chem Eng 7: 606–612.

IV Ohenoja K, Saari J, Illikainen M, Breitung-Faes S, Kwade A & Niinimäki J (2014) Effect of polydispersity index on the grinding limits of highly concentrated limestone suspensions. Chem Eng Technol 37(5): 833–839.

V Ohenoja K, Breitung-Faes S, Kinnunen P, Illikainen M, Saari J, Kwade A & Niinimäki J (2014) Ultrafine Grinding of Limestone with Sodium Polyacrylates as Additives in Ordinary Portland Cement Mortar. Chem Eng Technol 37(5): 787–794.

All the publications listed were written by the author of this thesis, whose main

responsibilities were the experimental design, experimental work, data analysis

and reporting of the results. The co-authors participated in the experimental

design and writing of the papers by making valuable comments. In Paper V the

experimental design and data analysis were carried out together with the co-

authors.

15

Table of contents

Acknowledgements 9

Abbreviations 11

List of original publications 13

Table of contents 15

1 Introduction 17

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

1.2 Outline of the thesis ................................................................................ 19

2 State of the art 21

2.1 Wet grinding with stirred media mills ..................................................... 21

2.1.1 Introduction to stirred media mills ............................................... 21

2.1.2 Evaluating the grinding process: particle size and width of

the particle size distribution ......................................................... 22

2.1.3 Operational parameters of the mill ............................................... 22

2.1.4 Stress energy model ...................................................................... 23

2.1.5 Material properties and required stress intensity .......................... 25

2.1.6 Grinding limits ............................................................................. 26

2.2 Stabilization of mineral suspensions ....................................................... 28

2.2.1 Effect of particle-particle interactions on grinding success .......... 28

2.2.2 Electrostatic stabilization .............................................................. 29

2.2.3 Steric stabilization ........................................................................ 29

2.2.4 Electrosteric stabilization ............................................................. 30

2.2.5 Measuring stability ....................................................................... 31

2.2.6 Polymers in a stirred media mill ................................................... 32

2.3 Stirred media milling and stabilization of TiO2 ....................................... 33

2.3.1 Introduction to TiO2 ...................................................................... 33

2.3.2 Production and properties of TiO2 pigment .................................. 33

2.3.3 Stabilization of a TiO2 suspension ................................................ 34

2.3.4 Factors affecting the particle size distribution of TiO2 in

stirred media milling .................................................................... 35

2.4 Stirred media milling and stabilization of CaCO3 ................................... 35

2.4.1 Introduction to CaCO3 .................................................................. 35

2.4.2 Grinding and stabilization of CaCO3 ............................................ 36

2.4.3 Sodium polyacrylate as a grinding aid for CaCO3 ........................ 38

2.4.4 The grinding limit for CaCO3 ....................................................... 39

3 Aim and questions to be resolved 41

16

4 Materials and methods 43

4.1 Materials .................................................................................................. 43

4.1.1 Milling materials .......................................................................... 43

4.1.2 Solvent .......................................................................................... 43

4.1.3 Grinding media ............................................................................. 43

4.1.4 Grinding aids ................................................................................ 43

4.2 Experimental setups for grinding ............................................................ 44

4.2.1 Multiple passage mode for TiO2 ................................................... 44

4.2.2 Micron and submicron grinding experiments for CaCO3 ............. 45

4.2.3 Nanogrinding experiments for CaCO3 .......................................... 46

4.3 Analysis ................................................................................................... 47

4.3.1 Particle size distribution ............................................................... 47

4.3.2 Viscosity ....................................................................................... 47

4.3.3 Zeta potential measurements ........................................................ 47

4.3.4 X-ray diffraction measurement ..................................................... 48

5 Results and discussion 49

5.1 Stirred media milling of TiO2 pigment .................................................... 49

5.1.1 Factors affecting the particle size distribution .............................. 49

5.1.2 Factors affecting TiO2 suspension stability .................................. 52

5.2 Stirred media milling of CaCO3 particles ................................................ 54

5.2.1 Micron grinding of CaCO3 ........................................................... 55

5.2.2 Submicron grinding of CaCO3...................................................... 58

5.2.3 Nanogrinding of CaCO3 ............................................................... 63

6 Conclusions 67

References 69

Original publications 81

17

1 Introduction

1.1 Background

Most everyday products nowadays contain inorganic mineral particles with a

certain, product-specific particle size distribution. Around 30 percent of all

standard copy paper, for example, consists of inorganic filler, and some coated

paper grades may contain over 50 percent inorganic matter. Inorganic mineral

particles also appear in concretes, composites, plastics, paints, printing inks,

medicines and cosmetics such as sun lotions, for instance, and even chewing gum

contains mineral particles. The desired particle size depends on the application,

but is typically in the micron range. During the past decade submicron and

nanoparticles have also aroused a wide interest and have gained new applications

due to their high surface area and strength. In addition to product particle size, the

width of the particle size distribution also affects the final properties of the

product, such as transparency, as shown by Breitung-Faes & Kwade (2011), and

the particle size distribution should be taken into account when considering the

decisive particle properties.

Particle size reduction from the micrometre range to a nano-scale results in

totally different properties for the same material (Krumpfer et al. 2013). A good

example is titanium dioxide (TiO2), which is a good but expensive pigment at

particle sizes around 200 nm, with a high refractive index, substantial scattering

properties and a good hiding power, whereas when its particle size is reduced to

the nano-range, particularly around 10 nm (Auvinen et al. 2011), it is transparent

at the wavelengths of visible light, absorbs UV light and has significant

photocatalytic properties, so that it can be used as a self-cleaning material (Wang

et al. 2013).

Grinding or dispersing is usually the last process step before the mineral is

added to an application, and the step in which the final particle size distribution

for the mineral is achieved. Considering only wet processes, stirred media mills

are the most frequently used grinders for micro, submicron and nanoparticle

production. In addition to this “true grinding”, in which coarser solid particles are

comminuted, particles can also be produced by “bottom-up” processes in which

the material is synthesized via chemical reactions, allowing the particles to grow

to the desired size. These particles produced by bottom-up process are usually

aggregated or agglomerated so they need to be deagglomerated or dispersed.

18

Stirred media mills have been found to be effective for this purpose as well

(Schilde et al. 2011).

Submicron grinding by means of stirred media mills is nevertheless an

energy-intensive process, and it is thus necessary to optimize energy usage by

determining the optimum operational parameters for the mill, e.g. using the stress

energy model presented by Kwade (2003). Energy efficiency can also be

improved by using the highest possible solids concentration, which depends on

the target product particle size, i.e. where solids concentrations of up to 75 weight

percent are used in micron grinding (He et al. 2006a) while concentrations of only

around 5 to 20 wt% are typically used in nanogrinding, since a decrease in

particle size will lead to an enormous increase in the number of particles. Kwade

& Schwedes (2007) expressed that the grinding of a single spherical particle of

size 1 mm down to 100 nm will generate one trillion (1·1012) spherical fragments.

It is obvious, therefore, that particle-particle interactions play an important role in

submicron range grinding and that controlling these interactions is an important

issue when reducing the particle size to the submicron scale. These particle-

particle interactions can be controlled by adding stabilization chemicals, e.g.

various salts, polymers or polyelectrolytes.

This thesis will consider the parameters affecting particle size distribution

and suspension stability in the submicron grinding of calcium carbonate (CaCO3)

and titanium dioxide (TiO2). TiO2 particles are aggregates produced via the

bottom-up method, while CaCO3 are primary mineral particles produced by the

top-down method, and therefore a comparison of the behaviour of these materials

will be of interest on account of their differences in nature and strength. Some of

these differences can be appreciated from the electron microscope images

presented in Fig. 1.

19

Fig. 1. Electron microscope images of a) TiO2, and b) CaCO3.

1.2 Outline of the thesis

This thesis is organized into six chapters. Chapter 1 briefly traces the background

to the work, while Chapter 2 sets out present understanding of stirred media

milling, the stabilization of mineral particles and stirred media milling of TiO2

and CaCO3. The aim of this thesis and the questions to be solved are detailed in

Chapter 3, and the materials, experimental setups and analytical devices used in

the research are presented in Chapter 4. The results obtained concerning the

stirred media milling of TiO2 and CaCO3 particles are presented in Chapter 5, and

the conclusions reached on the basis of these are contained in Chapter 6.

21

2 State of the art

After setting out the fundamentals of wet grinding with stirred media mills and

the stabilization of suspensions, this chapter will describe the present

understanding of the material properties, stirred media milling and stabilization of

the minerals to be considered here, i.e. calcium carbonate and titanium dioxide.

2.1 Wet grinding with stirred media mills

2.1.1 Introduction to stirred media mills

The first industrial stirred mills were introduced in the 1950´s in response to the

increasing need for fine particle grinders (Jankovic 2003, Kwade & Schwedes

2007). These mills have become the preferred devices for fine and even

nanogrinding (Kwade & Schwedes 2007). Stirred media mills are best suited

relative to other mills for the production of the smallest particle sizes (Wang &

Forssberg 2007). A stirred media mill is a vertical or horizontal cylinder which is

loaded with grinding beads and a suspension containing the particles to be

ground. The grinding chamber is filled to around 60–85 vol.% with grinding

beads, which can be made from steel, glass, ceramics or plastics. These are

typically in a size range between 0.05 mm and 3 mm. A rotating stirrer sets the

beads and the suspension in motion whereupon the particles are stressed by

compression, impact and shearing brought about by collision of the beads,

causing a reduction in their size. Either a pin or disc stirrer can be used, for

instance, the disc stirrer being the more commonly used type.

Stirred media mills are typically operated in wet mode, but dry mode mills

have also been introduced by Altun et al. (2013) and Mucsi et al. (2013). The

challenge in industrial dry grinding is the milling limit, which is currently in the

region of one micrometre due to agglomeration. Although submicron grinding has

proved to be possible in laboratory-scale dry mode milling with Pulvis from

Hosokawa Alpine (Droop & Stein 2011), only wet mode milling will be

considered in this thesis since the focus of interest in on the grinding of particles

to a submicron range and measurement of the stabilization of suspensions.

22

2.1.2 Evaluating the grinding process: particle size and width of the particle size distribution

The grinding process is usually evaluated by measuring the median size d50 of the

particles in the suspension. Other measures derived from the cumulative

distribution, such as d75, d80 and d90, can be used where these are of principal

importance, e.g. in the mining industry. Particle size measuring devices can be

based on a variety of methods, including light scattering and diffraction,

sedimentation, electrophoresis and ultrasound attenuation.

In addition to the median particle size, the width of the particle size

distribution is usually calculated in order to obtain information on the size

homogeneity of the particles. Various evaluation methods have been used for this.

The ratios between cumulative weight passing particle sizes d20 and d80 (d80/d20)

or d16 and d84 (d84/d16) are used to evaluate the spread of the particle size

distribution (Jankovic & Sinclair 2006, Karbstein et al. 1995, Karbstein et al.

1996, Nesset et al. 2006), while the ratio between the particle sizes d10 and d90

(d10/d90 or (d10–d90)/d50) can be more useful for representing the particle size

distribution width because it extends the comparison over a wider range (Kotake

et al. 2011). The geometric standard deviation (d84/d16)0.5 has also been used

(Johnson et al. 1997). In all these cases, a decrease in the ratio represents a

narrower particle size distribution. In this research the width WPSD will be

described in terms of the span value ((d10–d90)/d50)), since this is the commonly

used measure for products ground in a stirred media mill.

2.1.3 Operational parameters of the mill

Only a small part of the energy released into the grinding chamber of a stirred

media mill is actually used for the comminution of particles, as much of it is

dissipated into the product suspension and the grinding chamber wall as heat

through friction, for instance, as observed by Kwade (2003). The energy used for

particle breakage can be optimized by choosing the optimal operational

parameters for the mill. There are a large number of these parameters which may

affect the grinding process; in fact, up to 44 parameters have been identified by

Mölls & Hörnle (1972), although many of them can be classified as being of less

importance by Jankovic (2003). There are around ten parameters that have been

found to be significant: namely the mill geometry, operational mode of the mill,

grinding medium size, grinding medium material, grinding medium filling ratio,

23

stirrer tip speed, throughput, grinding time and aspects of the formulation of the

suspension such as its solids concentration, stabilization chemicals and solvent.

Investigations into the optimal operational parameters for the stirred media

milling of different materials have been carried out by Becker et al. (2001), Bel

Fadhel et al. (1999), Bel Fadhel & Frances (2001), Bunge et al. (1992), Choi et

al. (2007), Gao & Forssberg (1993), Hennart et al. (2012), Jankovic (2003), Joost

& Schwedes (1996a), Knieke et al. (2010), Knieke et al. (2011), Kwade et al.

(1996), Liimatainen et al. (2011), Mankosa et al. (1989), Mucsi (2013), Patel et

al. (2012), Pradeep & Pitchumani (2011), Samanli et al. (2010), Stenger et al.

(2005), Tuunila (1997), Wang & Forssberg (2000) and Zheng et al. (1996).

Generally it is the grinding bead size and rotational speed of the stirrer that have

been found to be the most important operational parameters affecting the resulting

particle size.

When considering the width of the particle size distribution, the most

important factors are the transport behaviour and the operational mode of the mill.

The transport behaviour of the suspensions can vary between plug flow and an

ideally stirred vessel. In plug flow all the particles are stressed for the same time

period, whereas the mixing action in the grinding chamber causes them to be

stressed for different lengths of time (Kwade & Schwedes 2007). The residence

time distribution represents the time for which the particles are stressed, and thus

a narrower residence time distribution means a narrower particle size distribution.

The residence time distribution is narrower for a mill operated in multiple passage

mode than in circuit mode (Stadler et al. 1990), so that the former ensures more

equal treatment for all the particles (Hond 1990), allowing a narrower particle

size distribution (PSD) to be obtained (Karbstein et al. 1996). In addition to the

operational mode, the bead size and stirrer tip speed, for instance, have also been

reported to have an effect on the width of the particle size distribution (Jankovic

& Sinclair 2006, Kotake et al. 2011, Müller & Polke 1999, Nesset et al. 2006,

Wang & Forssberg 2000).

2.1.4 Stress energy model

The first description of the stress energy of grinding media, SEGM, was presented

by Kwade et al. (1996), who took account of the effect of three important

operational parameters of a stirred media mill at the same time instead of

analysing each parameter individually. In this equation (Equation 1), the stress

energy SEGM as a measure of the maximum kinetic energy of two colliding

24

grinding beads is presented as being dependent on the size of the grinding beads

dGM, the grinding media density ρGM and the stirrer tip speed vt

23tGMGMGM vdSE . (1)

It has been noted for many materials that an optimum value for the stress energy

exists at a certain particle size; i.e. the stress energy is optimum and energy

utilization at its maximum when the stress energy is just sufficient to break the

particle (Jankovic 2001, Kwade et al. 1996, Mende et al. 2004, Stender et al.

2004). This optimal value can be found by illustrating a product property that

changes during grinding, usually the median particle size, as a function of stress

energy at different specific energies.

However, Equation 1 does not take into account the deformation behaviour of

the grinding media and the product particles, so that Becker et al. (2001) added

Young´s moduli to Equation 1 to remedy this omission. Young´s modulus needs

to be taken into account especially in the case of highly valuable materials (for

e.g. fused corundum). The following expression (Equation 2) for the stress energy

of the product particles SEp is the extended version of Equation 1 that includes the

ratio of Young´s modulus for the grinding media, EGM, to that of the product

particles EFeed:

1

23 1

GM

FeedtGMGMp E

EvdSE . (2)

In addition to these simple calculations, more exact descriptions of the processes

that take place inside a mill can be obtained by calculating its stress energy

distributions, as done by Stender et al. (2004) and later transferred and extended

to other mill types such as pin mills, annular gap mills and accelerator mills by

Breitung-Faes (Breitung-Faes & Kwade 2013b). From the stress energy

distribution it is also possible to determine the median stress energy, which

interprets the grinding process better than does the maximum stress energy.

Breitung-Faes used this median stress energy to create an enhanced stress energy

model based on the work of Kwade (2004), Becker et al. (2001) and Stender et al.

(2004) for the optimization of process parameters. The model allows the optimum

stress energies to be predicted if certain selected material and mill parameters are

known. With a knowledge of the optimum stress energy for the desired product

fineness, the most important process parameters, i.e. grinding medium size and

density and stirrer tip speed, can be chosen and the specific energy minimized. In

25

addition, Breitung-Faes & Kwade (2013a and 2013b) found grinding medium

wear, which is an economic problem and a source of product contamination

(Becker & Schwedes 1999, Breitung-Faes & Kwade 2008, Breitung-Faes &

Kwade 2011, Joost & Schwedes 1996b, Weller et al. 1997), to be lowest when the

predicted optimum process parameters were used. Grinding medium wear is

significant, especially for the grinding of very hard materials such as fused

alumina (Joost & Schwedes 1996b). Nowadays, the stress energy distributions are

also modelled using a four way CFD-discrete element method and the

distributions are different than the theoretical modelled ones (Beinert et al. 2014).

Alongside his description of the stress energy of grinding media, Kwade

(2003) presented two stressing models where the stress energies where considered

from different perspectives: a mill-related and a product-related stressing model.

The first characterizes the mill independently of the particles to be ground (SEGM),

and the second characterizes the product when the intensity of the stress applied

to the particles has to be considered. Stress intensity, SI, is the ratio of stress

energy to the particle mass d503, see Equation 3:

350d

SESI GM . (3)

2.1.5 Material properties and required stress intensity

The stress intensity required for particle breakage depends on the particle

properties. Particles produced via bottom-up processes, e.g. TiO2 particles, are

usually aggregated or agglomerated (Schilde et al. 2010) by means of chemical

bonds (aggregates) or weak physical interactions (agglomerates). The breakage of

these aggregates or agglomerates is called dispersing, and thus the stirred media

milling of TiO2 particles can be referred to as dispersing or deagglomeration

instead of grinding. As shown in Fig. 2, the stress intensity required for the

fragmentation of aggregates or agglomerates is lower than that for primary

particles breakage (Müller et al. 2004). Primary particles linked by solid bonding

include crushed and preground stones or rocks, for instance, and the breaking of

these primary particles is called true grinding.

The stress energy required depends on the particle size: the smaller the

particle size, the stronger the particle is (Schilde et al. 2007), due to the lower

number of weak points in the particle, and the higher the stress intensity required

for breakage. On the other hand, required stress energy of the grinding media

26

decreases in the case of smaller product particle sizes, as shown for limestone

(Kwade et al. 1996) and zinc oxide (Breitung-Faes & Kwade 2013b), for

example.

In addition to particle structure and particle size, the compression strength of

a material, σCS, affects particle breakage (Breitung-Faes & Kwade 2013b), as it

represents the material´s resistance to compression stresses. The stress mechanism

in a stirred media mill is dominated by compression stresses, and therefore, if the

compression stress caused by the grinding beads is higher than the compression

strength of the material, fracturing will occur (Breitung-Faes & Kwade 2013b).

Fig. 2. Typical presentation for particle structures and stress intensities required for

fragmentation.

2.1.6 Grinding limits

The grinding limit is determined by the smallest particle size that it is possible to

obtain with a certain grinding device (Jimbo 1992). The grinding limit can be the

result of a limitation inherent in the device or its operational parameters, or of

poor stabilization of the suspensions. A smallest possible fineness exists for each

grinding device, so that it is not possible to obtain a high fineness by ball milling,

for example, that is comparable to that achievable in stirred media milling (Wang

& Forssberg 2007). The operational parameters of a stirred media mill and

suspension properties such as solids concentration, stability level and viscosity

can limit the achievable fineness. These limits are called, respectively, the

Primary particles AgglomeratesAggregates

Stress intensityrequired for fragmentation

27

apparent grinding limit (dependent on the stability of the suspension) and the

process-related grinding limit (depended on the operational parameters) and

dampening-related grinding limit (dependent on its viscosity) (Knieke 2012,

Knieke et al. 2009, Knieke et al. 2010).

However, after excluding the above-mentioned limitations and when using a

device which measures primary particle sizes, it is still possible to study the true

grinding limit for materials by a stirred media milling. The true grinding limit is

reached when the particles become perfectly mono-crystalline and are no longer

able to retain defects after an energy input (Gryaznov et al. 1991, Knieke et al.

2009, Knieke et al. 2010). This limit has been found to be 50 nm for silica and

carbonate (Wang & Forssberg 2006), 30 nm for quartz (Cho et al. 1996), 10 nm

for tin dioxide (Knieke 2012) and 5 nm for zirconia (Knieke et al. 2009).

Grinding limits can be estimated theoretically by calculation, and even

though such calculations provide rough estimates, the true grinding limits for

various materials as calculated from the equilibrium distance between two edge

dislocations (Nieh & Wadsworth 1991) have shown good agreement with

experimental data (Breitung-Faes & Kwade 2013b, Knieke 2012). The formula

for calculating the minimum separation distance L is given by Equation 4:

Hv

bGL

)1(

3

, (4)

where G is the shear modulus, b is the Burger vector length, v is Poisson´s ratio

and H is hardness. As stated by Nieh and Wadsworth (1991), when the particle

size is smaller than L there will be no further dislocation pileups and the Hall-

Petch (HP) relation, which predicts that the strength or hardness of a particle will

increase with decreasing particle size until a certain critical size, will no longer

apply. Below this critical particle size the strength of the particle will decreases

and each individual particle will no longer be able to support more than one

dislocation. This point, the particle size at which the HP relation no longer holds,

should be the theoretical grinding limit.

28

2.2 Stabilization of mineral suspensions

2.2.1 Effect of particle-particle interactions on grinding success

In addition to the operational parameters of the mill (stirrer tip speed, grinding

media, etc.) described earlier (section 2.1.3), grinding success is influenced by

particle-particle interactions, which increase with decreasing particle size due to

Brownian motion (Peukert et al. 2005), increasing specific surface area, and

decreasing distance between particles. These interactions are especially

significant for particle sizes below 1 µm. These interactions have an effect on the

stability and rheology of suspensions, so that poor stability of a suspension will

increase its viscosity and detract from grinding success due to poor suspension

fluidity. High viscosity of the suspension causes damping of the grinding beads

and, thus, reduction to the stress energy of grinding media. It is also possible that

a state of equilibrium between agglomeration, deagglomeration and comminution

(Mende et al. 2003) can be reached in which no further comminution occurs even

with an increased energy input. This limitation on particle size can be called the

apparent grinding limit (see section 2.1.6) and was postulated in the past to exist

at particle sizes of around 0.5 µm, as described the history by Sommer & Peukert

(2007). Later on the effect of particle-particle interactions came to be reduced

through various stabilization mechanisms, eventually enabling nanogrinding

(Mende et al. 2003). In order to weaken the particle-particle interactions, three

stabilization mechanisms have been introduced: electrostatic, steric and

electrosteric stabilization (see Fig. 3). A short description of these stabilization

mechanisms will be given below.

Fig. 3. Simplified presentations for three mechanisms for stabilizing suspensions: a)

electrostatic, b) steric and c) electrosteric.

‐‐‐‐‐

‐

‐‐‐‐

‐‐

‐

‐‐

‐‐

‐

‐

‐

‐

‐‐

‐

‐ ‐‐ ‐

‐‐‐‐‐

‐

‐‐‐‐

‐‐

‐

‐‐

‐‐

‐

‐

‐

‐

‐‐

‐

‐ ‐‐ ‐

_ __ _ __+ ++

++

++++

__

___

__

_

___

_

_ + ++

++

++++

__

___

__

_

___

_

_

a) b) c)

29

2.2.2 Electrostatic stabilization

The first mechanism, and probably the easiest one to implement for mineral

oxides, is electrostatic stabilization. In aqueous systems of mineral oxides, the

surface potential of inorganic particles is heavily dependent on the pH, and

therefore particles can be stabilized by adjusting the pH of the suspensions so that

the particles have equal charges. Interactions between particles with equal charges

are repulsive and can be measured by means of the zeta potential, ζPOT, which is

the surface potential at an electrostatic double layer consisting of an inner charge

layer and an outer diffuse charge layer. The higher the absolute value of the zeta

potential is, the higher is the electrostatic repulsion between the particles. A zeta

potential of at least ±30 mV is required to achieve stable dispersion (Sentein et al.

2009, Vallar et al. 1999). At a zero zeta potential, i.e. at the isoelectric point IEP,

the particles tend to flocculate. It has been found in many studies that viscosity is

highest at the isoelectric point and is lower for values far away from this point

(Bernhardt et al. 1999, He et al. 2004, Liddell & Boger 1994, Morris et al. 1999).

Another way of predicting the stability of electrostatically stabilized

suspensions is by means of the DLVO theory of Derjaguin, Landau, Verwey and

Overbeek (Derjaguin & Landau 1941, Verwey & Oberbeek 1948), according to

which particle interactions are determined by calculating the total interaction

potential as a sum of single interactions, which can be attractive van der Waals

forces, repulsive electrostatic forces and additional Born repulsion.

2.2.3 Steric stabilization

Steric stabilization is another possible mechanism, especially when electrostatic

stabilization is not sufficient, i.e. at a high salt concentration or in non-polar

organic solvents (Eisermann et al. 2012). Dispersions with a high a solids

concentration cannot be dispersed efficiently with electrostatic stabilization

(Sommer & Peukert 2007). The steric stabilization mechanism has several

advantages and is based on the adsorption or chemisorption of polymer chains on

the particle surfaces (Napper 1977) thus preventing agglomeration. The thickness

of this polymer shell compared to particle size plays an important role: the thicker

it is, the more stable is the suspension, and thus when the polymer shell is too thin

agglomeration will occur.

Unlike the simplified situation presented in Fig. 3, polymer chains are usually

present in solution as random coils which are adsorbed to particle surfaces, where

30

in some cases they remain unchanged and in others they spread out, producing

tails, loops and trains at the particle surface (Farrokhpay 2009, Heusch & Reizlein

2005). The concentration of polymer also plays an important role, in that it must

be high enough to achieve a sufficient coverage of the particle surface, otherwise

bridging flocculation may occur (Fleer & Lyklema 1974).

The required amount of polymer depends on the actual surface area of the

mineral to be stabilized. The smaller the particle size, the higher the actual surface

area, hence the larger the area the polymer has to cover. Some examples of

stabilization chemicals with steric properties are poly(ethylene oxide) and

poly(vinyl alcohol) (Moudgil et al. 2002). One further example of a steric

stabilizer is the polymer DAPRAL, the influence of which on stabilization is

described in several studies published by Prof. Peukert´s group at the University

of Erlangen-Nürnberg in Germany (see, for example, Eisermann et al. 2012).

2.2.4 Electrosteric stabilization

Electrosteric stabilization is a combination of the electrostatic and steric

mechanisms, in which charged polymers, i.e. polyelectrolytes, are adsorbed onto

the particle surfaces. The charge of ionic groups existing along the polymer

chains depends on the nature and pH of the polyelectrolyte. Polyacrylic acid,

polymethacrylic acid and polyphosphoric acid, for example, are anionic

polyelectrolytes, while polyvinalamine is an example of a cationic polyelectrolyte

and long-chain amino acids are zwitterionic polyelectrolytes (Sommer & Peukert

2007). One widely used anionic polyelectrolyte is sodium polyacrylate, a sodium

salt of polyacrylic acid with negatively charged carboxylic groups in the main

chain (Eriksson et al. 2008). Its chemical formula is [-CH2-CH(COONa)-]n (Al-

Nasra 2013) and its chemical structure is as presented in Fig. 4. Low molecular

weight polymers are used for dispersing purposes, whereas polymers with high

molecular weights (Mw over 106) are used as flocculants (Farrokhpay 2009). In

the present work the electrosteric stabilization of mineral particles was ensured

using sodium polyacrylates of molecular weights ranging between 4000 and

12500 g/mol and with two polydispersity indices (PDI=Mw/Mn).

31

Fig. 4. Chemical structure of sodium polyacrylate.

2.2.5 Measuring stability

The stability of a suspension, and therefore the optimal amount of stabilization

chemicals, is usually measured rheologically - particularly viscosity - in order to

obtain information on the flocculated or deflocculated state of the suspension.

Shear-thinning suspensions, for instance, which may be related to a breaking of

the links between particles gathered in flocculates, are characterized by a decrease

in viscosity with increasing shear rate (Bernard et al. 2008). Shear-thinning

behaviour is classified as non-Newtonian (see Fig. 5), while for suspensions

having Newtonian behaviour the viscosity does not change with shear rate. Other

non-Newtonian rheological phenomena include shear-thickening and Bingham

yield.

Fig. 5. Common non-Newtonian behaviour according to Sommer (2007).

Rheological measurements can be used to determine a yield value for a

suspension (Klimpel 1999). Suspension yield refers to an excessively high

Bingham

Shear stress

Viscosity

Shear rate Shear rate

Shear‐thinning

Shear stress

Viscosity

32

viscosity due to poor stability of the suspension, and can be eliminated or reduced

by means of stabilization chemicals (He et al. 2006a). The Bingham plastic model

is one of the empirical rheological models which describe the dependence of non-

Newtonian mineral slurries on the shear stress τ and shear rate γ (He et al. 2004).

The extrapolated Bingham yield stress τB has been obtained by many authors by

fitting experimentally obtained data over a range of high shear rates (> 360 s–1) to

the Bingham model (He et al. 2006b, Morris et al. 1999, Muster & Prestidge

1995, Prestidge 1997, Tangsathitkulchai 2003, Tseng & Lin 2003). The Bingham

yield stress τB can be calculated with Equation 5:

PLB . (5)

When considering suspension yield in the context of stirred media milling, the

optimal throughput possible with a stirred media mill occurs when the suspension

is as thick as possible but still does not exhibit yield (Klimpel 1999).

An indirect characterization of the suspension rheology of an electrostatic or

electrosteric stabilization mechanism is also feasible by measuring its zeta

potential, as the relationship between the zeta potential and viscosity has been

found to be valid for many solid/additive systems (Bernhardt et al. 1999). In

addition, particle size measurements also give information about the stability

level of a suspension, in the form of an increase in particle size as a function of

grinding time, which points to agglomeration and thus poor stability.

2.2.6 Polymers in a stirred media mill

When mineral particles are stabilized with polymers during stirred media milling

it is possible for the polymer chains to be broken between colliding grinding

beads. Breakage of the polymer chains, and a consequent change in molecular

weight, is a risk, because polymers may lose their stabilization efficiency and

desorption from the particle surface may occur (Chartier et al. 1996). The

degradation of polymers has been studied in ball milling by (Chartier et al. 1996)

and in stirred media milling by Karis et al. (1993) and Sommer (2007). Sommer,

whose PhD thesis reported on the grinding of poly(ethylene oxide) (PEO) in

stirred media mills, concluded that the molecular weight of the polymer chains (as

measured by gel permeation chromatography) must be over 500 000 g/mol for

them to be broken by the stirred media milling (Sommer 2007). It is therefore safe

33

to use polymers of low molecular weight, as their degradation can be assumed to

be negligible.

2.3 Stirred media milling and stabilization of TiO2

2.3.1 Introduction to TiO2

Titanium dioxide, TiO2, is one of the main white pigments used in the world, with

a total annual production capacity of about 6 million tons in 2008 (Farrokhpay et

al. 2010). TiO2 particles are produced to the desired crystal size via a bottom-up

process. TiO2 particles with a crystal size of around 200 nm are widely used in

paints, printing inks, plastics, cosmetics and pharmaceuticals (Farrokhpay et al.

2006, Jalava 2006, Schulte 2002), the paint industry being the largest consumer,

accounting for nearly 60% of global TiO2 pigment consumption (Adam 1999).

Different applications for TiO2 nevertheless become possible when the crystal

size is reduced below 50 nm, as these nanoparticles can be used in applications

where visible light transparency and UV shielding is required, e.g. in sunscreens,

wood lacquers, plastics and cosmetics (Auvinen et al. 2013). The photocatalytic

properties of nano-sized TiO2 have also been widely studied, e.g. by Kandiel et al.

(2010), Yoshioka et al. (2013) and Zhang et al. (2014).

2.3.2 Production and properties of TiO2 pigment

Pigment titanium dioxide (crystal size around 200 nm) the form considered in this

thesis, has many excellent qualities due to its high refractive index and significant

scattering properties. It exists in two crystalline forms, anatase and rutile, of

which the latter has a higher scattering efficiency (Wang et al. 2010) because of

its higher refractive index, 2.74 vs. 2.54 (Johnson et al. 1997). Many of the

pigment’s optical properties depend on its particle size distribution (Losoi 1997),

and the stringent specifications for these pigment properties have made narrow

size distributions more and more important (Bel Fadhel et al. 1999). Optimum

scattering is obtained with TiO2 particle sizes of approximately half the

wavelength of light (Balfour 1992, Holmnäs 2000, Loebbert & Sharangpani 2000,

Saari et al. 2006). Hence it is important to obtain a particle size distribution

between 200 nm and 400 nm in order to achieve optimum scattering in the visible

range of light, i.e. 400–700 nm.

34

TiO2 pigment particles are produced via a bottom-up method in a sulphate

process in which gel-like meta-titanic acid, H2TiO3 (Wang et al. 2010) is calcined

in a rotary kiln to TiO2 crystals approximately 200 nm in size (Heine & et al,

1992). After leaving the kiln the crystals aggregate (Gesenhues 2001) to a few

millimetres in size. This calciner discharge (Gesenhues 1999) is pre-ground with

a hammer or steam jet mill, after which the particles are comminuted to the

desired product particle size distribution in a wet stirred media mill.

2.3.3 Stabilization of a TiO2 suspension

TiO2 suspensions can be stabilized with a wide range of chemicals and by all

three stabilization mechanisms as described in section 2.2. The electrostatic

mechanism can be carried out with various acids and bases (Hanaor et al. 2011,

Johnson et al. 2007, Mandzy et al. 2005, Mikulášek et al. 1997, Safaei-Naeini et

al. 2012, Widegren & Bergström 2000, Widegren & Bergström 2002, Yang et al.

2001) or with monoisopropanolamine (Barnard & Laverick 1980) and steric

stabilization with polymers (Deiss et al. 1996, Safaei-Naeini et al. 2012). The

electrosteric mechanism can be produced by various polyacrylic acids

(Farrokhpay et al. 2004, Farrokhpay et al. 2006, Farrokhpay et al. 2010,

Farrokhpay 2012, Greenwood & Kendall 1999, Heijman & Stein 1995, Ran et al.

2007, Strauss et al. 1993), including poly(methacrylic acid) (Singh et al. 2012),

sodium polyacrylates (Boisvert et al. 2001, Deng et al. 2011, Karakas & Çelik

2013, Sato et al. 2008) and sodium hexaphosphate (Greenwood & Kendall 1999),

and modified amines with long aliphatic chains (Liu et al. 2006), silane coupling

agents in an organic solvent (Joni et al. 2009) and fluorinated compounds

(trimethoxytrifluor(propyl) silane) in an organic solvent (Joni et al. 2012) have

also been used.

The isoelectric point for TiO2 has been observed naturally to be around 6, but

this can vary depending on the surface coating of the TiO2 particles, for instance.

The isoelectric point can be shifted to lower pH regions by adding sodium

polyacrylates to the suspension, as observed by Singh et al. (2012), Strauss et al.

(1993) and Tang et al. (2006).

35

2.3.4 Factors affecting the particle size distribution of TiO2 in stirred media milling

As mentioned above, the desired product particle size for TiO2 pigments is

between 200–400 nm, in order to obtain optimal scattering properties. The final

particle size, however, will depend on the operational parameters of the stirred

media mill, chiefly the stirrer tip speed, bead size, filling ratio of the grinding

media, solids concentration and the flow rate of the suspension (Bel Fadhel et al.

1999, Inkyo et al. 2006, Joni et al. 2009, Liu et al. 2006, Way 2004). In addition

to particle size, a narrow particle size distribution is essential as well. Few reports

are available on factors affecting the width of the particle size distribution of TiO2

in stirred media milling, although it is known that the scattering properties are

proportional to the width of the PSD (Bel Fadhel et al. 1999). Only dispersant

dosage is reported to have an effect on the width of the particle size distribution,

in that a narrower particle size distribution was obtained with the highest

dispersant dosage when the particles were dispersed in butyl acetate with a

dispersant, as reported by Liu et al. (2006). The effect of the molecular weight of

sodium polyacrylates on the particle size distribution of a TiO2 suspension has

been studied after TiO2 particles had been dispersed by ultrasonication (Deng et

al. 2011), but no reports involving stirred media milling exist. The factors

affecting the particle size and particle size distribution of a TiO2 pigment were

therefore studied in original papers I and II belonging to this thesis.

2.4 Stirred media milling and stabilization of CaCO3

2.4.1 Introduction to CaCO3

Limestone, which is composed mostly of calcite and aragonite minerals, i.e.

different crystalline forms of calcium carbonate CaCO3, has an abundant

occurrence, comprising around 4% of the earth’s crust (Liu et al. 2008). It is also

one of the cheapest commercially available inorganic materials (Kumar et al.

2014), and therefore has innumerable industrial applications: it is used in paints,

inks, coatings, paper products, plastics and films (Garcia et al. 2002, Tsuzuki et

al. 2000). Micron-sized CaCO3 is usually made by grinding limestone, a top-

down process, yielding a product known as ground calcium carbonate GCC, with

an edgy particle shape (see Fig. 6a). On the other hand, nano-sized CaCO3 is

usually produced by a wet chemical precipitation technique, a bottom-up process

36

(Tsuzuki et al. 2000), yielding precipitated calcium carbonate PCC with a needle-

like crystalline shape (see Fig. 6b), for instance. Since precipitation techniques

need numerous process steps, and as ultrafine wet grinding in stirred media mills

has been shown to be a suitable method for nanoparticle production (Breitung-

Faes & Kwade 2008), the direct production of ultrafine, nano-sized CaCO3 by

stirred media milling is a new, interesting option.

Fig. 6. a) Ground, and b) precipitated calcium carbonate.

The use of CaCO3 as a filler in papermaking permits the production of a brighter

paper with a greater resistance to yellowing and ageing. In addition, when it is

used as a part of the coating of the paper, it provides better opacity, printability,

ink receptivity and smoothness, while its use as a filler for plastics improves heat

resistance and hardness (Garcia et al. 2002). In addition to the paper and plastics

industries, nano-sized CaCO3 has shown to improve the properties of sealants

(Bates & Harrison 1985), nanocomposites (Kovacevic et al. 2002), lubricating

oils (Qiu et al. 2000) and concretes (Bentz et al. 2012, Camiletti et al. 2013,

Kadri & Duval 2002, Sato & Beaudoin 2011), for example.

2.4.2 Grinding and stabilization of CaCO3

The fine grinding of CaCO3 to a particle size range of 1–100 µm has studied

intensively by Kwade et al. (see, for example Kwade et al. 1996), yielding

optimal operation parameters for its stirred media milling based on a stress energy

37

model (see section 2.1.4). Operational parameters affecting the fine grinding of

CaCO3 have also been studied by Zheng et al. (1996), who derived operational

recommendations from their findings and used these later to study the

performance of grinding aids for limestone (Zheng et al. 1997).

Ultrafine or submicron CaCO3 grinding tests have been performed since the

turn of the century in response to the increasing need for suspensions with smaller

particle sizes. Choi et al. (2007) studied the ultrafine grinding of CaCO3,

obtaining a minimum particle size of 350 nm independent of the grinding medium

with sodium hexametaphosphate as a grinding aid. They also observed improved

grinding energy efficiency (m2/kJ) or an enhanced grinding rate (Em–1) when

grinding aids were added and finer grinding media were used. Later they studied

CaCO3 suspensions with different solids concentrations (10–60 wt%) and they

concluded that the lower the solids concentration, the better is the grinding rate

(Choi et al. 2011). Meanwhile, Cho et al. (2009) ground CaCO3 with poly(acrylic

acid, sodium salt) and found optimal operational parameters in the submicron

range. The minimum primary particle size they achieved was around 40 nm. Later

they dispersed these ground but agglomerated nanoparticles with primary particle

sizes of 40 nm by means of small grinding media (30 µm) and synthesized the

particles using a spray drying method (Cho et al. 2010).

It has been shown by many authors that CaCO3 particles – like other mineral

particles – require a stabilization chemical if submicron particles are to be

achieved. It is usually electrosteric stabilization that is used to reduce the

viscosity of the suspension, or conversely to increase its solids concentration in

grinding, thereby increasing throughput. There are a wide variety of stabilization

chemicals or grinding aids available for use with CaCO3, e.g. ammonium

polyacrylate (Greenwood et al. 2002), sodium hydroxide, sodium carbonate,

sodium oleate, oleic acid, poly(acrylic) acid (Zheng et al. 1997), comb-grafted

copolymer (Tobori & Amari 2003), sodium polystyrene sulphonate (Garcia et al.

2004), sodium hexametaphosphate (Choi et al. 2007), polyethylene oxides

(Mosquet et al. 1997) and tetrasodium pyrophosphate (Reinisch et al. 2001,

Reinsch et al. 2001). Poly (acrylic) acid was found to be the best grinding aid due

to the improved energy efficiency, while sodium hydroxide, for instance, reduced

the energy efficiency (Zheng et al. 1997). Greenwood et al. (2002) similarly

found polyacrylates to be good dispersants for limestone. When dispersants are

used as grinding aids, multipoint addition of the grinding aid is recommended by

(He et al. 2006a) in order to avoid the formation of cushion layers on the milling

beads, which would lower the stress intensities of the collisions between the

38

beads. Sodium polyacrylate as a dispersant for CaCO3 will be discussed in more

detail in the next section.

When stabilizing CaCO3 particles it should be borne in mind that this

material is sensitive to pH changes: calcite begins to dissolve below pH=7,

leading to a decrease in the concentration of the solid phase, whereas it is

positively charged in the pH range 7–10 and its surface becomes negatively

charged when the pH is above 10 (Vdovic & Bišcan 1998). The isoelectric point

of CaCO3 is at pH=8.3 (Roberts 1996). Thus the pH should be adjusted above 10

in order to obtain a stable suspension and minimize the dissolving of calcite.

Another characteristic of CaCO3 is the possibility of phase transformation

from calcite to aragonite during grinding, as observed in wet stirred media milling

when ethanol is used as a milling liquid (Knieke 2012), although not when water

was used.

2.4.3 Sodium polyacrylate as a grinding aid for CaCO3

Sodium polyacrylate as a dispersant for CaCO3 has been studied by numerous

authors: Bernhardt et al. (1999), Deng et al. (2010), Foissy et al. (1982), Garcia et

al. (2003), Garcia et al. (2004), He et al. (2006a), He et al. (2006b), He &

Forssberg (2007), Lamarche et al. (1983), Lynch et al. (2012), Reinisch et al.

(2001), Reinsch et al. (2001) and Tobori & Amari (2003). Garcia et al. (2003)

used sodium polyacrylate as a grinding aid and observed a well-dispersed

suspension with Newtonian rheological behaviour. Sodium polyacrylate with a

low molecular weight (Mw ~ 5500) was studied by (He et al. 2006b, He &

Forssberg 2007) and was found to be effective grinding aid for limestone, while

Garcia et al. (2004) also found sodium polyacrylate with a molecular weight

around 5000 to be effective. Lamarche et al. (1983), examining sodium

polyacrylates with molecular weights in the range 700–20000 g/mol, found

molecular weights of 2000 and 4000 to be the most effective for CaCO3. This can

be explained by the selective adsorption of polymer chains having a certain

molecular weight to the CaCO3 particle surfaces. This selective adsorption has

been observed in the range Mw=2000–5000 by many authors (Deng et al. 2010,

Geffroy et al. 2000, Loiseau et al. 2005).

In addition to molecular weight, the molecular weight distribution has also

been seen to have a significant influence on the rheological properties of CaCO3

suspensions (Lamarche et al. 1983). The molecular weight distribution of

polymers is quantified in terms of the polydispersity index PDI, the ratio of the

39

number-average molecular weight Mn to the weight-average molecular weight Mw

(PDI = Mw/Mn), a narrower molecular weight distribution having a lower PDI and

vice versa. Standard polymers have a PDI over 2, which can be reduced to around

2 by fractionating these polymers. A polymer dispersant with a lower PDI has

been produced by a method called controlled radical polymerization CRP (Qiu et

al. 2001), one example of which is the reverse addition fragmentation transfer

RAFT method (Barner-Kowollik 2008). RAFT polymerization is performed under

standard radical polymerization conditions but in the presence of chain transfer

agents producing polymer chains that have nearly the same number of monomer

units and therefore a low PDI (Lynch et al. 2012). By producing dispersants

having a low PDI it is possible to obtain polymer chains which have mostly the

same optimum length for adsorption onto the particle surface, giving better

stability. Lower viscosity values have been obtained for limestone suspensions

containing low PDI sodium polyacrylates than with those having a higher PDI

value (Loiseau et al. 2003, Lynch et al. 2012), but it is not known how the PDI

affects the stability of CaCO3 suspension, its particle size distribution or its

grinding limits (apparent or true) in the stirred media milling. These questions

were therefore addressed in original papers III and IV in this thesis.

2.4.4 The grinding limit for CaCO3

The experimental grinding limit for calcium carbonate in stirred media mills has

been studied by two groups (Knieke 2012, Knieke et al. 2009, Wang & Forssberg

2006). Knieke (2009) found a grinding limit of 55 nm for limestone when milled

in water, and Wang & Forssberg (2006) obtained almost the same limit, around 50

nm. When CaCO3 was milled in ethanol, however, the grinding limit proved to be

around 30 nm (Knieke 2012). Even so, it should be noted that these grinding

limits can vary depending on the operational parameters and analytical device

used.

A rough estimate for the theoretical grinding limit of CaCO3 can be

calculated using the equation for the minimum separation distance (see Equation

4) after Nieh and Wadsworth (1991) with factors found from the literature for

calcite (see Table 1). The minimum separation distance calculated for CaCO3 is

around 17 nm, which can be estimated to be theoretical grinding limit for CaCO3.

This theoretical limit has not yet been reached in grinding experiments, so efforts

to achieve it using the most effective sodium polyacrylate identified in the present

original papers III and IV are reported in paper V.

40

Table 1. Factors used to estimate the theoretical grinding limit for CaCO3.

Factor Unit Value Ref. for value

Shear modulus, G [GPa] 30.6 (Assefa et al. 2003)

Burger vector length, b [nm] 0.77 (Barber et al. 2010)

Poisson´s Ratio, v [-] 0.254 (Kuiper et al. 1959)

Hardness, H [GPa] 1.73 (Grübl et al. 2001)

41

3 Aim and questions to be resolved

As seen in the above literature survey (Chapter 2), there exists a lack of

knowledge as to how various parameters affect the particle size distribution of

TiO2 pigment particles and how the polydispersity index of sodium polyacrylates

affects the particle size distribution and grinding limits of CaCO3 in aqueous

stirred media milling. The aim of this research was therefore to obtain a new

understanding of the factors affecting the stability and resulting particle size

distribution of TiO2 and CaCO3 suspensions in submicron grinding. The following

questions were identified for resolution in the course of this work:

1. How do the operational parameters of a stirred media mill and the molecular

weights of sodium polyacrylates affect the particle size distribution of a TiO2

pigment?

2. How do the molecular weights of sodium polyacrylates affect the stability of

ground TiO2 pigment suspensions?

3. How does the polydispersity index of a sodium polyacrylate affect the

particle size distribution and stability of a CaCO3 suspension during stirred

media milling?

4. How does the polydispersity index of a sodium polyacrylate affect the

grinding limits of CaCO3 in a highly concentrated aqueous suspension?

5. Is it possible to grind CaCO3 particles to the calculated grinding limit in an

aqueous suspension with a sodium polyacrylate that has a low polydispersity

index?

Answers to these questions are sought and discussed in the original papers as

shown in Table 2.

Table 2. Answers to the questions posed as guides to the research.

Question Paper I Paper II Paper III Paper IV Paper V

Q1 X X

Q2 X

Q3 X X

Q4 X

Q5 X

43

4 Materials and methods

The experiments reported on here were performed at two research laboratories:

the titanium dioxide experiments at the Fibre and Particle Engineering Laboratory

of the University of Oulu, Finland, and the calcium carbonate experiments at the

Institute for Particle Technology, TU Braunschweig, Germany. Both laboratories

have a stirred media mill with a disc stirrer and grinding chamber volume of

around 1 L, but their analytical devices differ. The purpose was not to compare

exact values for two different materials, however, and it was simply necessary to

bear possible differences between the measuring devices in mind when evaluating

the work as a whole.

4.1 Materials

4.1.1 Milling materials

The milling materials were surface-unmodified rutile titanium dioxide, TiO2, and

limestone with a CaCO3 content of 99.1% (Eskal60 and Eskal500 from KSL

Staubtechnik GmbH, Lauingen, Germany). The median particle sizes d50 of these

materials were: TiO2 1 µm, Eskal60 57µm and Eskal500 4 µm.

4.1.2 Solvent

De-ionized water was used as a solvent in the grinding experiments.

4.1.3 Grinding media

Glass (SiO2), zirconium oxide (ZrO2) and yttria-stabilized zirconium oxide

((Y3O2)ZrO2) beads in size ranges from 90 µm to 1200 µm (from Sigmund

Lindner) were used as grinding media. The densities of these materials were: SiO2

2.5 g/cm3, ZrO2 3.9 g/cm3 and (Y3O2)ZrO2 6.1 g/cm3.

4.1.4 Grinding aids

The pH of the suspensions was adjusted using hydrochloric acid, sodium

hydroxide and potassium hydroxide to ensure electrostatic stabilization. Anionic

44

sodium polyacrylates from Kemira Oyj were used as polymer dispersants in the

electrosteric stabilization experiments. The doses of sodium polyacrylates D are

expressed in terms of the weight of the dry sodium polyacrylate as a percentage of

that of the dry mineral in the suspension (wtNaPa/CaCO3% or wtNaPa/TiO2%). Sodium

polyacrylates of different molecular weights Mw were tested for use with TiO2, see

Table 3, and two sodium polyacrylates with different polydispersity indices

(PDI=Mw/Mn) were used for CaCO3 (see Table 4). NaPaPDI>2 was made by

fractionation and thus had a higher polydispersity index, and NaPaPDI=1.2,

produced by controlled radical polymerization, had a lower PDI (Lynch 2011).

Table 3. Sodium polyacrylates used with TiO2.

Sodium polyacrylate, Abb. Mw [g/mol]

NaPa4000 4000

NaPa6000 6000

NaPa7300 7300

NaPa12500 12500

Table 4. Sodium polyacrylates used with CaCO3.

Abb. Mw [g/mol] PDI [-]

NaPaPDI>2 7300 >2

NaPaPDI=1.2 7800 1.2

4.2 Experimental setups for grinding

4.2.1 Multiple passage mode for TiO2

A laboratory-scale stirred media mill with a disc stirrer (Hosokawa Alpine 90

AHM hydro mill) was used in the multiple passage mode (see Fig. 7) for the TiO2

grinding experiments. The multiple passage mode is ideal for short grinding times

when aiming at a narrow residence time distribution. The mill has an effective

chamber volume of 1.12 l. A 0.1 mm screen was used for separating the product

from the grinding media. To avoid wear, the grinding chamber wall had been

coated with polyurethane PU and a PU stirrer was used. The flow rate of this

suspension was kept constant and was controlled by a feeding pump. The TiO2

suspension was circulated through the mill for a total of eight passes, and a

suspension sample of 200 ml was taken after every second pass. The power was

recorded during grinding, and the mean power was calculated for each grinding

45

pass. The specific energy Em for each grinding pass was calculated using Equation

6:

mp

nm cM

PPE 0

, (6)

where Pn is the mean active power on the nth pass, P0 is the no-load power

without grinding media or suspension, Mp is the mass flow rate of the suspension,

and cm is the solids concentration by weight. The cumulative Em is the sum of the

specific energies of the individual passes.

Fig. 7. Stirred media milling in multiple passage mode according to Kwade &

Schwedes (2007).

4.2.2 Micron and submicron grinding experiments for CaCO3

The CaCO3 grinding experiments were performed in circuit mode (see Fig. 8) in a

laboratory-scale stirred media mill (Bühler PML2) with a disc stirrer and a

grinding chamber of volume 0.9 l. The circuit mode is easy to implement and is

used for long grinding times. The grinding chamber wall had been coated with

ZrO2 and a PU stirrer was used. A 0.2 mm ceramic sieve cartridge was used to

separate the grinding media. The operational parameters of the mill were kept

constant: suspension flow rate 700 g/min, stirrer tip speed 10 m/s, filling ratio of

the grinding medium 80 vol.%, grinding medium Y3O2(ZrO2) and grinding

medium size 1.0–1.2 mm. The solids concentration of the suspension was also

kept constant: 20 wt% for the reference sample (electrostatic stabilization) and 40

wt% for the samples containing sodium polyacrylates (electrosteric stabilization).

2, 4, 6, 8 pass

2, 4, 6, 8 pass

1, 3, 5, 7 pass

1, 3, 5, 7 pass

46

The power was recorded during grinding and the specific energy Em was

calculated as the power input integrated over the grinding time tG and divided by

the mass of dry limestone mFeed, see Equation 7:

feed

Gn

m m

dtPPE

)( 0. (7)

Fig. 8. Stirred media milling in circuit mode (Paper IV, published by permission of

John Wiley and Sons).

4.2.3 Nanogrinding experiments for CaCO3

The nanogrinding experiments for CaCO3 were performed in circuit mode (Fig. 8)

in a laboratory mill equipped with a pin stirrer (LabStar, Netzsch, Germany). This

mill has different separating systems for the grinding media and product

suspension, enabling the use of a smaller grinding media size than in the Bühler

PML2 mill. Different grinding media sizes (100–800 µm) and densities (6.06

g/cm3 for (Y3O2)ZrO2 and 2.56 g/cm3 for SiO2), stirrer tip speeds and grinding

times were used in order to obtain different stress energies, but the filling ratio of

the grinding media was kept constant (80 vol.%), as also was the solids

concentration of the suspension, 10 wt%. The specific energy was determined

according to Equation 7.

M

Cooling water M

Stab.chemi‐cals

Sampling

M

pH

47

4.3 Analysis

4.3.1 Particle size distribution

Static light scattering. The particle size distributions of the finely ground

samples (> 1 µm) were analysed by static light scattering using a Malvern

Mastersizer 2000, Beckman Coulter LS 13 320 and LRS Helos by Sympatec. The

Mie optical model was used and the samples were kept in an ultrasonic bath

before analysis.

Dynamic light scattering. The particle size distributions of the submicron

samples (< 1 µm) were measured by photon correlation spectroscopy (Nanophox

NX0019 from Sympatec GmbH).

Ultrasonic spectrometer. An ultrasonic spectrometer (Dispersion Technology,

DT 1200) was also used to measure the particle size distributions. Ultrasonic

spectrometry is generally accepted for measuring primary particle sizes rather

than agglomerate sizes.

4.3.2 Viscosity

Brookfield Viscometer. The viscosity of the TiO2 suspensions was determined

using a Brookfield Viscometer DV-11+ Pro Extra, with a cup-and-bob

measurement configuration (ULA-DIN-86). The measurements were performed

at 20°C in a ramp up/ramp down mode over a shear rate range from 0.1 to 258 s–1.

Bohlin rheometer. The rheological behaviour of the CaCO3 suspensions was

determined using a Bohlin rheometer (Gemini 2, Rotometric drive 2) with a cup-

and-bob measurement geometry (C25 DIN 53019). The measurements were

performed at 25°C in ramp up/ramp down mode over a shear rate range from 1 to

1000 s–1.

4.3.3 Zeta potential measurements

Delsa Nano C. The zeta potentials of the TiO2 samples were measured with a

Delsa Nano C device (Beckman Coulter). The samples were diluted by adding an

NaCl electrolyte solution (10 mmol/g) having a conductivity of 1.19 mS/cm.

48

Ultrasonic spectrometer. The zeta potentials of the CaCO3 suspensions were

measured without dilution with an ultrasonic spectrometer (Dispersion

Technology, DT 1200), using about 200 mL of the original sample. For more

details, see (Dukhin & Goetz 2010, Dukhin et al. 1999). The same device was

also used to find the optimal dispersant dose and isoelectric point, IEP, for

CaCO3.

4.3.4 X-ray diffraction measurement

The X-ray diffraction XRD data were collected using a Siemens D5000 powder

diffractometer with a graphite secondary monochromator using Cu Kα radiation

(λ=0.1542 nm). The XRD patterns were recorded in a 2Θ range from 20° ≤ 2Θ ≤

80° using continuous scanning with a step size of 0.02° and a counting time of 1

s/step. A divergence silt and anti-scatter slit, both of 0.5°, and a detector slit of 0.2

mm were used. The crystallite size and lattice microstrain were evaluated

simultaneously using TOPAS 3R commercial software (BRUKER AXS GmbH,

Karlsruhe, Germany) and adopting the fundamental parameter approach with

whole pattern fitting.

49

5 Results and discussion

The results summarized here were published in the original papers I–V, in which

the questions identified in Chapter 3 were also discussed and resolved. Those

findings will similarly be summarized in this chapter.

5.1 Stirred media milling of TiO2 pigment

The aim of the TiO2 pigment grinding was to achieve an understanding of the

parameters affecting particle size and the width of the particle size distribution.

The desired properties for a TiO2 pigment are a particle size of around 300 nm, a

narrow particle size distribution and a stable suspension.

5.1.1 Factors affecting the particle size distribution

The parameters affecting the particle size distribution of a TiO2 pigment, as

studied in paper I, were the solids concentration of the suspensions, grinding

media size, grinding media density and stirrer tip speed. It was seen that grinding

media size had the strongest effect on the particle size distribution, so that

equation for stress energy in which the grinding media size has the greatest effect

(third power, see Equation 1) describes well the operational parameters of the mill

and how they influence the particle size distribution. Only the results concerning

the grinding media size and stress energies are presented here, except that the

effect of the molecular weight of the sodium polyacrylates on the particle size

distribution is also summarized (studied in greater depth in Paper II).

The effects of different grinding media sizes dGM on the particle sizes are

presented as a function of cumulative Em in Fig. 9a. The grinding medium was

ZrO2 with bead diameters varying from 255 µm to 897 µm. As seen in the figure,

the smaller the grinding media size was, the smaller the particle size at a given

Em. An exception to this, however, was the increase in the particle size at the

eighth pass when the smallest grinding beads were used. The effects of different

grinding medium sizes on the widths of the particle size distribution are presented

as a function of the median particle size in Fig. 9b. As can be seen, the narrowest

particle size distribution is obtained when using the smallest grinding beads: 255

or 345 µm. The width of the particle size distribution is around 1.5 when the

target particle size is around 300 nm. In addition to grinding medium size, the

number of grinding beads inside the mill may affect to the particle size

50

distribution: The higher the number of grinding beads, the more homogeneous is

the product is leading to narrower particle size distribution.

Fig. 9. a) Particle sizes and b) widths of the particle size distributions at different

grinding media sizes (modified from Paper I, published by permission of Elsevier).

The median particle sizes of the TiO2 pigment are presented as a function of stress

energy at different average specific energies in Fig. 10a. The different values for

stress energy were obtained by changing the operational parameters in the manner

represented in Equation 1. As can be seen from the figure, the lower the stress

energy is, the lower the particle size obtained. Taking the specific energy into

account as well, the optimal stress energy is seen to be around 5 µJ, due to the

low specific energy (360 kJ/kg) and the achievement of the target particle size

(300 nm).

The width of the particle size distribution is presented as a function of stress

energy in Fig. 10b. With a stress energy of 5 µJ and a specific energy of 360

kJ/kg, a slightly narrower particle size distribution was obtained. When

comparing Figs. 10a and 10b, it is obvious that the stress energy greatly affects

the resulting median particle size but its effect on the width of the particle size

distribution is negligible. This requirement for low stress energy is due to the

nature of TiO2 particles: as indicated in section 2.1.5, aggregates need only a low

intensity for breakage.

10 100 1000 100000.2

0.3

0.4

0.5

0.6

0.7

0.8TiO

2 in H

2O

cm = 52 wt%

GM

= 70 vol%

vt = 9.2 m/s

GM

= 3.87 g/cm3

Me

dia

n pa

rtic

le s

ize

d5

0 [µ

m]

Specific energy Em [kJ/kg]

dGM

[µm]

255 345 554 694 897

0.2 0.3 0.4 0.5 0.6 0.7 0.80.0

0.5

1.0

1.5

2.0

2.5

TiO2 in H

2O

cm = 52 wt%

GM

= 70 vol%

vt = 9.2 m/s

GM

= 3.87 g/cm3

dGM

[µm]

255 345 554 694 897W

idth

of p

art

icle

siz

e d

istr

ibu

tion

WP

SD [-

]

Median particle size d50

[µm]

b)a)

51

Fig. 10. a) Median particle sizes and b) widths of the particle size distributions as a

function of stress energies at different specific energies (modified from Paper I,

published by permission of Elsevier).

The effects of sodium polyacrylates of different molecular weights on the particle

size distribution of TiO2 pigment were studied by carrying out multiple passage

grinding experiments. The dose of sodium polyacrylates considered here is 0.5

wtNaPa/TiO2%, and the median particle sizes are presented as a function of the

specific energy in Fig. 11a. First, a great difference in specific energy was found

between the reference grinding and that performed with sodium polyacrylates, the

figure being around 65% higher in the former case. The target particle size was

still achieved in the reference grinding, however, and no plateau in particle size

reduction occurred. As can be seen from Fig. 11a, the molecular weight of the

sodium polyacrylates did not affect the particle size at a dose of 0.5 wtNaPa/TiO2%,

and the target particle size was achieved with almost the same specific energy

with all the sodium polyacrylates.

The width of the particle size distribution as a function of the median particle

size is presented in Fig. 11b. As indicated, the distribution for the reference

sample was broad, which was probably also due to poor stabilization efficiency,

while the sodium polyacrylate with the lowest molecular weight again differed

from the other samples containing sodium polyacrylates in having a broader

particle size distribution. As far as the width of the particle size distribution is

concerned, NaPa12500 had a slightly narrower distribution that included the target

size (300 nm).

1 10 100 10000.2

0.3

0.4

0.5

0.6

0.7

0.8

cm = 52 wt%

GM = 70 vol%

vt = 5.5 - 12.8 m/s

GM

= 2.53 - 6.10 g/cm3

dGM

= 255 - 897 µm

Em [kJ/kg]

180 360 540 720 1100

Me

dia

n p

art

icle

siz

e d

50 [µ

m]

Stress energy [µJ]

1 10 100 10000.0

0.5

1.0

1.5

2.0

2.5

cm = 52 wt%

GM = 70 vol%

vt = 5.5 - 12.8 m/s

GM

= 2.5 - 6.1 g/cm3

dGM

= 255 - 897 µm

Em [kJ/kg]

180 360 540 720 1100W

idth

of p

art

icle

siz

e d

istr

ibu

tion

WP

SD [-

]

Stress energy [µJ]

b)a)

52

Fig. 11. a) Median particle size and b) width of the particle size distribution of TiO2

suspensions at sodium polyacrylate doses of 0.5 wtNaPa/TiO2% (modified from Paper II,

published by permission of Elsevier).

5.1.2 Factors affecting TiO2 suspension stability

Zeta potential and pH

Since the zeta potential and pH exert a strong influence on the stability of a

suspension, zeta potential measurements were carried out on TiO2 suspensions at

different pH levels. Measurements were performed on a sample without sodium

polyacrylate (reference grinding with electrostatic stabilization) and the final

ground TiO2 samples (after 8 passes) containing 0.5 wtNaPa/TiO2% of sodium

polyacrylates. Zeta potential as a function of pH is presented in Fig. 12, where it

can be seen that the pH of the TiO2 suspension without sodium polyacrylates

should be lower than 2 or higher than 7 in order to obtain an absolute zeta

potential of at least 30 mV and thus a reasonably stable suspension (Sentein et al.

2009). As indicated in Fig. 12, the isoelectric point for TiO2 without sodium

polyacrylate was around 4.5, and this can be shifted to lower values, down to pH

3, by adding sodium polyacrylates to the suspension before grinding. Thus a value

of under –30 mV for the zeta potential can be obtained at pH 4.5 by adding

sodium polyacrylates.

100 400 700 4000 70000.2

0.3

0.4

0.5

0.6 c

mpH

Ref. 0.26 10 NaPa

40000.52 5.5

NaPa6000

0.52 5.5

NaPa7300

0.52 5.5

NaPa12500

0.52 5.5

D = 0.5 wtNaPa/TiO2

%

GM: ZrO2

dGM

= 345 µm

vt= 9.2 m/s

TiO2 in H

2O

Med

ian

part

icle

siz

e d

50 [

µm

]


0.2 0.3 0.4 0.5 0.60.0

0.5

1.0

1.5

2.0

2.5TiO

2 in H

2O

cm

pH

Ref. 0.26 10 NaPa

40000.52 5.5

NaPa6000

0.52 5.5

NaPa7300

0.52 5.5

NaPa12500

0.52 5.5

D = 0.5 wtNaPa/TiO2

%

GM: ZrO2

dGM

= 345 µm

vt= 9.2 m/sW

idth

of p

artic

le s

ize

dist

ribut

ion

WP

SD [-

]


[µm]

b)a)

53

Fig. 12. Zeta potential of the TiO2 suspensions as a function of pH (Paper II, published

by permission of Elsevier).

It is not only the zeta potential that affects the stability of the TiO2 suspension in

the case of an electrosteric stabilization mechanism, but the pH also has a

significant effect on the adsorption of polymer chains to the particle surfaces, as

the conformation state of the adsorbed polymer on the TiO2 particle surfaces

varies as a function of pH (Singh et al. 2012). Effective stabilization requires the

right conformation and a thick, dense layer of polymer on the TiO2 surface as well

as high coverage (Farrokhpay 2009). The stability of the suspension, and thus the

optimum pH, can be studied by measuring viscosity as a function of pH (Strauss

et al. 1993, Tang et al. 2006). The optimum pH for suspensions containing

sodium polyacrylates was found in paper II to be between 5 and 6, as the lowest

viscosity occurred in this range. Thus, in order to obtain the lowest viscosity and

the best adsorption of the polyelectrolyte chains to the TiO2 particle surfaces, the

pH of the suspensions was adjusted to around 5.5 before and during grinding.

Molecular weight of sodium polyacrylates

Since the stability of a suspension can be estimated by means of viscosity

measurements, viscosity curves were measured for the final TiO2 samples after 8

passes. These are presented together with measurements for the samples

containing different sodium polyacrylates at a dose of 0.5 wtNaPa/TiO2% in Fig. 13.

As can be seen, slight shear thinning behaviour can be observed in every sample.

The suspensions containing sodium polyacrylates had lower viscosities than the

reference suspension even though the solids concentration in the former was 50%

2 3 4 5 6 7 8 9 10-80

-60

-40

-20

0

20

40N

pass= 8

D = 0.5 wtNaPa/TiO2

%

d50

~ 0.28 µm

pH adj. with HCl/NaOH

Ze

ta p

ote

ntia

l P

OT [m

V]

pH [-]

Ref. NaPa

4000

NaPa6000

NaPa7300

NaPa12500

54

higher (52 wt% vs. 26 wt%). This indicates that the solids concentration can be

more than doubled by adding sodium polyacrylates. Fig. 13 also indicates that a

slightly lower viscosity could be achieved by using NaPa12500 as a grinding aid,

probably because this represents the right polymer chain length and the right

conformation of polymer chains and thus a higher stability of the suspension and

lower viscosity. Even though the differences in molecular weight between the

sodium polyacrylates were quite high, as can be seen from the viscosity curves, it

is obvious that other variables affecting the polymers may have had an effect,

even though no information on these other variables was provided by the supplier.

One such additional variable is the polydispersity index, the effect of which will

be considered later in this thesis in connection with CaCO3.

Fig. 13. Variations in viscosity as a function of shear rate for the final samples at a

sodium polyacrylate dose of 0.5 wtNaPa/TiO2% (Paper II, published by permission of

Elsevier).

5.2 Stirred media milling of CaCO3 particles

The aim of CaCO3 grinding was to achieve more information on how the

polydispersity index of sodium polyacrylates affects the stability, particle size

distribution and grinding limits of CaCO3 suspensions in stirred media milling.

First the optimal sodium polyacrylate dose was identified by means of zeta

potential measurements and then this optimal dose was verified in grinding

experiments by measuring particle size distributions and viscosities after grinding.

After that the milling experiments were continued in order to determine the

apparent grinding limits with both sodium polyacrylates having different

0 100 200 2601

10

100

1000

10000

100000 c

mpH

Ref. 0.26 10 NaPa

40000.52 5.5

NaPa6000

0.52 5.5

NaPa7300

0.52 5.5

NaPa12500

0.52 5.5

TiO2 in H

2O

D = 0.5 wtNaPa/TiO2

%

Vis

cosi

ty [m

Pa

s]

Shear rate [s-1]

55

polydispersity indices when the solids concentration of CaCO3 was high (40

wt%). The better sodium polyacrylate was then chosen for nanogrinding

experiments in which the aim was to see whether CaCO3 particles can be ground

down to the calculated grinding limit in an aqueous suspension.

5.2.1 Micron grinding of CaCO3

Optimal dose of sodium polyacrylate

The CaCO3 particles were first pre-ground without any stabilization chemicals at

a solids concentration of 20 wt% until the fluidity of the suspensions was good,

with a specific energy of 3500 kJ/kg. The primary particle size of the resulting

suspension was 337 nm and the zeta potential was + 20 mV as measured by

ultrasonic spectrometry. After that the effects of two sodium polyacrylates that

differed their polydispersity indices PDI, one having a wide range (PDI>2) and

the other a narrow one (PDI=1.2), were studied. The zeta potentials and pH as a

function of the dose of sodium polyacrylates (see Fig. 14) were measured by

continuous titration in order to obtain information on the optimal dose of

additives.

As shown in Fig. 14, the addition of sodium polyacrylates reduced the initial

zeta potential, as the anionic polymers neutralized the positive charges on the

particles. After adding more sodium polyacrylates, the zeta potentials fell to zero

at a dose of 0.25 wtNaPa/CaCO3% and any further addition of sodium polyacrylates

caused the particles to become negatively charged and the zeta potentials to

decrease until the curve started to plateau off at around –37 mV. This meant that

no more polymer could be adsorbed onto the particles. Thus, 1.5 wtNaPa/CaCO3% is

the optimum dose for limestone with a primary particle size of 337 nm. The

absolute values for the zeta potential have been observed to be generally higher

when a polymer is added by the single point method than with the titration

method (Eriksson et al. 2007). This phenomenon was explained by the rate at

which calcite dissolves. It should be noted, however, that the zeta potential was

measured here by continuous titration, so that he absolute values could be even

lower if a single point method were used. As can also be seen in Fig. 14, the zeta

potential and pH values are related, and a stable suspension can be obtained by

adjusting the pH to a certain value. The pH values in question are: NaPaPDI>2 ≥

10.2 and NaPaPDI=1.2 ≥ 10.5. The isoelectric point IEP (ζPOT=0) for CaCO3 can be

56

seen from Fig. 14 to be around 8.3 which corresponds to values quoted in the

literature (Roberts 1996, Somasundaran & Agar 1967).

Fig. 14. Zeta potential and pH of the CaCO3 suspension as a function of the polymer

dose (Paper IV, published by permission of John Wiley and Sons).

Effect of the polydispersity index on micron grinding

Given the optimal polymer dose for CaCO3 particles (x50=337 nm) when aiming

at a totally stable suspension, as described above, fine grinding experiments were

performed for CaCO3 at a solids concentration of 40 wt% and different sodium

polyacrylates doses (0.5, 1.0 and 1.5 wtNaPa/CaCO3%), together with a reference

grinding (electrostatic stabilization at a solids concentration of 20 wt%) without

any polymer. The median particle sizes during grinding as a function of the

specific energy at a sodium polyacrylate dose of 1.5 wtNaPa/CaCO3%, as presented in

Fig. 15a, indicate smaller particle sizes in the cases where polymers were used,

and it also seems that the difference in particle size increased relative to the

increased specific energy. This suggests that without sodium polyacrylates a state

of equilibrium between agglomeration and comminution could be reached even

before a median particle size of 1 μm was obtained. The median particle size of

the reference sample after 30 min of grinding was 1.64 μm. No significant

difference was seen between the dispersants in terms of particle sizes.

When sodium polyacrylates were used no significant agglomeration

appeared, as can also be seen in Fig. 15b, where the width of the particle size

distribution is plotted as a function of median particle size. The same figure points

0.0 0.5 1.0 1.5 2.0-40

-35

-30

-25

-20

-15

-10

-5

0

5

10

15

2020

POT

NaPaPDI>2

NaPaPDI=1.2

Zet

a po

tent

ial

PO

Tm

V]

Dose of sodium polyacrylate [wtNaPa/CaCO3

%]

3

4

5

6

7

8

9

10

11

pH NaPa

PDI>2

NaPaPDI=1.2

pH [-

]

CaCO3 in H

2O

cm= 0.20

x50

= 337 nm

57

to pronounced agglomeration in the reference sample. In addition, the widths of

the particle size distributions for the suspensions ground with sodium

polyacrylates decrease as the median particle size decreases. A slightly narrower

particle size distribution can be observed for sodium polyacrylate having a lower

polydispersity index which may suggest that low PDI sodium polyacrylate

prevents agglomeration better.

Fig. 15. Median particle sizes of the CaCO3 suspensions as a function of specific

energy, and b) width of the particle size distribution as a function median particle size

(modified from Paper III, published by permission of David Publishing).

The variations in the viscosity of the CaCO3 suspensions as a function of shear

rate after 30 min of grinding are presented in Figs. 16a–c, together with a

reference grinding having a solids concentration of 20 wt% without sodium

polyacrylates. The apparent viscosity of the reference sample after 30 min of

grinding at 100 s–1 was 190 mPas and shear thinning (pseudoplastic) behaviour

may be observed in that the viscosity decreases with increasing shear rate. The

suspension fluidity of the reference sample was poor. With a 0.5 wtNaPa/CaCO3%

dose of dispersants (Fig. 16a), the same shear thinning behaviour appeared in both

samples. The same figure also indicates that the use of dispersants (0.5

wtNaPa/CaCO3%) could increase the solids concentration of the suspension, which

could be doubled to 40 wt% without any notable change in viscosity or

rheological behaviour relative to the sample ground without dispersants at 20

wt%. It can likewise be seen from Fig. 16a that the NaPaPDI=1.2 reduced the

viscosity most. When the sodium polyacrylate dose was 1.0 wtNaPa/CaCO3% (see

Fig. 16b), the NaPaPDI=1.2 already showed Newtonian behaviour, so that the

viscosity did not change with shear rate. At the same time, the NaPaPDI>2 had a

1 2 3 4 50

1

2

3

4

5CaCO

3 in H

2O

vt= 10 m/s

GM

= 6.06 g/cm3

dGM

= 1.0-1.2 mm

pH = 10D = 1.5 wt

polymer/CaCO3%

Reference NaPa

PDI>2

NaPaPDI=1.2W

idth

of p

art

icle

siz

e d

istr

ibut

ion

WP

SD [-

]


[µm]

10 100 1000 100001

2

3

4

5CaCO

3 in H

2O

vt= 10 m/s

GM

= 6.06 g/cm3

dGM

= 1.0-1.2 mm

pH = 10D = 1.5 wt

polymer/CaCO3%

Me

dia

n p

artic

le s

ize

d50

[µm

]


cm [-]

Reference 0.20 NaPa

PDI>20.40

NaPaPDI=1.2

0.40

b)a)

58

slight shear-thinning effect and higher viscosity. At sodium polyacrylate doses of

1.5 wtNaPa/CaCO3% (see Fig. 16c), no difference in viscosities between the sodium

polyacrylates could be observed any longer, viscosities were very low and

Newtonian behaviour appeared with both sodium polyacrylates.

Fig. 16. Viscosity curves for CaCO3 suspensions ground for 30 min with sodium

polyacrylate doses of a) 0.25, b) 1.0 and c) 1.5 wtNaPa/CaCO3% (modified from Paper III,

published by permission of David Publishing).

5.2.2 Submicron grinding of CaCO3

Effects of the polydispersity index and dose of sodium polyacrylates on

the agglomerate particle size distributions

Since sodium polyacrylate having a lower PDI had been found to maintain a

lower viscosity for longer, the grinding experiments were continued to a

1 10 100 10000.1

1

10

100

1000

10000

100000CaCO

3 in H

2O

D = 0.5 wtNaPa/CaCO3

%

pH = 10

cm [-]

Reference 0.20 NaPa

PDI>2 0.40

NaPaPDI=1.2

0.40

Vis

cosi

ty

[mP

as]

Shear rate [s-1]

1 10 100 10000.1

1

10

100

1000

10000

100000CaCO

3 in H

2O


%

pH = 10

cm [-]

Reference 0.20 NaPa

PDI>2 0.40

NaPaPDI=1.2

0.40

Vis

cosi

ty [m

Pa

s]

Shear rate [s-1]

1 10 100 10000.1

1

10

100

1000

10000

100000 c

m [-]

Reference 0.20 NaPa

PDI>2 0.40

NaPaPDI=1.2

0.40

Vis

cosi

ty [m

Pas

]

Shear rate [s-1]

CaCO3 in H

2O


%

pH = 10

a)

c)

b)

59

submicron range in order to find the apparent and process-related grinding limits

for a CaCO3 suspension containing sodium polyacrylates of different PDI (Paper

IV). The same sodium polyacrylates (NaPaPDI>2 and NaPaPDI=1.2) were used as in

section 5.2.1. The sodium polyacrylate doses tested were 1.5, 3.0 and 4.5

wtNaPa/CaCO3%, and the grinding experiments were continued until the pressure

inside the mill rose over 0.5 bar, upon which the mill was stopped and the final

sample was taken out. This pressure value was selected after micron grinding

tests, as it was seen that the fluidity of the suspensions became poor when the

pressure rose to 0.5 bar or higher.

The agglomerate particle size distributions were measured by photon

correlation spectroscopy PCS without the use of either ultrasound or dispersing

agents. No forces that would break agglomerates exist in a PCS device, so that it

measures agglomerate size rather than primary size, especially when no

ultrasound or dispersing agents are used. This agglomerate particle size describes

the stability of the suspension, and therefore the efficiency of the dispersant at

preventing particles from agglomerating. If the agglomerate particle size is small,

the dispersant can be assumed to prevent agglomeration.

The agglomerate particle sizes are presented in Fig. 17a as a function of

specific energy for both sodium polyacrylates and at different doses. As can be

seen, when a high PDI sodium polyacrylate (NaPaPDI>2) was used at the lowest

dose (1.5 wtNaPa/CaCO3%) the smallest agglomerate particle obtained was around

470 nm, after which a significant growth in agglomerate size could be seen. By

comparison, when a low PDI sodium polyacrylate was used, at the same dose the

smallest agglomerate size obtained was around 420 nm and only a slight growth

in agglomerate size could be seen. By increasing the dose of sodium polyacrylates

(from 1.5 to 3 and 4.5 wtNaPa/CaCO3%), smaller agglomerate particle sizes could be

obtained for both sodium polyacrylates studied here, although the same NaPa

dose consistently led to a smaller agglomerate size when NaPaPDI=1.2 was used

than with NaPaPDI>2. Furthermore, when considering the experiments with 4.5

wtNaPa/CaCO3% of NaPaPDI>2 and 3 wtNaPa/CaCO3% of NaPaPDI=1.2, it can be seen that

NaPaPDI=1.2 reduced the agglomerate size better even when the dose was 33%

lower. The quite low agglomerate growth obtained with NaPaPDI=1.2 is related to

the fact that low PDI NaPa prevents agglomeration more efficiently. Under our

experimental conditions the smallest agglomerate sizes obtained at NaPa doses of

4.5 wtNaPa/CaCO3% were 352 nm with NaPaPDI>2 and 284 nm with NaPaPDI=1.2.

These limits for agglomerate particle size reduction, as seen in Fig. 17a, may be

referred to as “apparent grinding limits”.

60

The widths of the agglomerate particle size distributions are presented in Fig.

17b as a function of the median particle size. As can be seen, the widths of the

agglomerate particle size distributions remain constant at around 0.40, so that this

width must be independent of the median particle size for both sodium

polyacrylates and for all the doses tested.

Fig. 17. a) Agglomerate particle sizes in the CaCO3 suspensions as a function of

specific energy, and b) widths of the agglomerate particle size distribution as a

function median particle size (modified from Paper IV, published by permission of

John Wiley and Sons).

Effects of the polydispersity index and dose of sodium polyacrylates on

the primary particle size distribution

Unlike the agglomerate particle size, the primary particle size is influenced only

by breakage of the particles, and the stability of the suspension does not have

such a strong influence. The primary particle sizes of the ground CaCO3

suspensions as measured by ultrasonic spectrometry are presented in Fig. 18a as a

function of specific energy. As can be seen, no difference between the doses of

sodium polyacrylates or their PDI values appeared at first, so that the primary

particle sizes decreased linearly as a function of specific energy in all the

experiments. Differences appeared between the sodium polyacrylate doses and

their PDI values only towards the end of grinding, where the higher the specific

energy was, the smaller were the primary particle sizes. A higher specific energy

is obviously obtained when the grinding time is longer. In these experiments

grinding was continued as long as the viscosity of the suspension remained low,

1000 10000 100000200

400

600

800

CaCO3 in H

2O

dGM

= 1.0-1.2 mm

GM

= 6.1 g/cm3

vt= 10 m/s

cm= 0.40

pH = 10

CaCO3 in H

2O

dGM

= 1.0-1.2 mm

GM

= 6.1 g/cm3

vt= 10 m/s

cm= 0.40

pH = 10

Me

dian

par

ticle

siz

e d

50 [n

m]

Specific energy [kJ/kg]

NaPaPDI>2

NaPaPDI=1.2

1.5 wt% 1.5 wt% 3.0 wt% 3.0 wt% 4.5 wt% 4.5 wt%

200 400 600 8000.2

0.4

0.6

0.8CaCO

3 in H

2O

dGM

= 1.0-1.2 mm

GM

= 6.1 g/cm3

vt= 10 m/s

cm= 0.40

pH = 10

NaPaPDI>2

NaPaPDI=1.2

1.5 wt% 1.5 wt% 3.0 wt% 3.0 wt% 4.5 wt% 4.5 wt%

Wid

th o

f par

ticle

siz

e di

strib

utio

n [-

]Median particle size d

50 [nm]

b)a)

61

i.e. the pressure inside the mill did not rise to 0.5 bar. Thus the primary particle

sizes are related both to grinding time and to the viscosity of the suspensions.

The final primary particle sizes obtained at doses of 1.5 wtNaPa/CaCO3% were

still over 200 nm, but at doses of 3.0 wtNaPa/CaCO3% they were already under 100

nm with both sodium polyacrylates. At a dose of 4.5 wtNaPa/CaCO3% of NaPaPDI>2

the resulting primary particle size was 80 nm, and a still lower primary particle

size (75 nm) could be obtained only with a dose of 3 wtNaPa/CaCO3% of NaPaPDI=1.2.

On the other hand, at a dose of 4.5 wtNaPa/CaCO3% of NaPaPDI=1.2 the resulting

primary particle size was still no less than 74 nm even though the specific energy

was over 3600 kJ/kg higher. Thus a grinding limit applicable to the conditions

used in this investigation had been reached, and the primary particle size could

not be reduced any further relative to NaPaPDI=1.2 at doses of 3.0 and 4.5

wtNaPa/CaCO3%. This process-related grinding limit may equally well be called a

dampening-related grinding limit and is caused by high viscosity in the

suspension, leading to a deceleration in the movement of the grinding medium.

Fig. 18. a) Primary particle sizes of the CaCO3 suspensions as a function of specific

energy, and b) width of the primary particle size distribution as a function of median

particle size (modified from Paper IV, published by permission of John Wiley and

Sons).

The widths of the primary particle size distribution, presented as a function of

median particle size in Fig. 18b, are between 2 and 3 for all the suspensions,

becoming slightly narrower as the median particle size decreases. The narrowest

particle size distributions are obtained with sodium polyacrylate doses of 3

wtNaPa/CaCO3%.

1000 10000 10000050

100

200

300

400

500 NaPa

PDI>2 NaPa

PDI=1.2

1.5 wt% 1.5 wt% 3.0 wt% 3.0 wt% 4.5 wt% 4.5 wt%

CaCO3 in H

2O

dGM

= 1.0-1.2 mm

GM

= 6.1 g/cm3

vt= 10 m/s

cm= 0.40

pH = 10

Med

ian

part

icle

siz

e x

50 [n

m]

Specific energy [kJ/kg]50 100 200 300 400 500

0

1

2

3

4

5

CaCO3 in H

2O

dGM

= 1.0-1.2 mm

GM

= 6.1 g/cm3

vt= 10 m/s

cm= 0.40

pH = 10

NaPaPDI>2

NaPaPDI=1.2

1.5 wt% 1.5 wt% 3.0 wt% 3.0 wt% 4.5 wt% 4.5 wt%

Wid

th o

f part

icle

siz

e d

istr

ibutio

n [-

]

Median particle size x50

[nm]b)a)

62

Effects of the polydispersity index and dose of sodium polyacrylates on viscosity

The variation in viscosity as a function of the shear rate is shown for the final

ground ultrafine CaCO3 samples at sodium polyacrylate doses of 3.0 and 4.5

wtNaPa/CaCO3% in Fig. 19. The primary particle sizes of these samples were

different, but all of them were under 100 nm. As can be seen in the figure, all the

suspensions show a similar rheological behaviour, with shear thinning

(pseudoplastic) behaviour. Furthermore, a slight thixotropy can be observed in the

curves, as indicated by a hysteresis loop between the upper and lower curves, a

feature suggestive of agglomerated suspensions. These agglomerated suspensions

must have occurred due to partial coverage of the CaCO3 particle surfaces with

the sodium polyacrylates, resulting from the surface area of the CaCO3 particles

in final samples being too large relative to the quantity of sodium polyacrylate

chains available.

It is widely accepted that viscosity is a function of particle size, so that

normally the smaller the particle size is, the higher the viscosity. As can be seen

from Fig. 19, however, the viscosity curves the samples containing NaPaPDI=1.2 are

still lower than those for NaPaPDI>2 even though the primary particle size is

smaller. Thus NaPaPDI=1.2 could reduce viscosity more efficiently. This

corresponds well to earlier results that show NaPaPDI=1.2 to be a more effective

grinding aid for CaCO3 than NaPaPDI>2. Therefore, in industrial processes the

amount additive could be decrease in order to obtain the same viscosity and

stability level to the suspension. Naturally, the price of a certain additive

(NaPaPDI>2 vs. NaPaPDI=1.2) should be taken into an account.

63

Fig. 19. Final viscosities for the CaCO3 suspensions (Paper IV, published by

permission of John Wiley and Sons).

5.2.3 Nanogrinding of CaCO3

Grinding limit and particle size distribution for CaCO3

In the light of the promising results obtained using NaPaPDI=1.2 as a grinding aid

for CaCO3 (see section 5.2), this polyacrylate was used with CaCO3 (Eskal 500)

in the subsequent aqueous nanogrinding experiments (Paper V). Certain process

parameters were varied in these grinding experiments, namely stirrer tip speed (6–

12 m/s), grinding medium size (100–800 µm) and grinding medium material

(Y3O2(ZrO2) and SiO2). The main goal was to determine whether CaCO3 particles

can be ground down to their calculated grinding limit in an aqueous suspension

and to find the optimal operational parameters.

The nanogrinding results are shown in Fig. 20a, where the specific energy is

plotted as a function of particle size as measured by ultrasonic spectrometry. The

smallest particle size obtained was 26 nm, which is lower than the grinding limit

found earlier for CaCO3 in aqueous suspensions (55 nm) or in ethanol (30 nm)

(Knieke 2012, Knieke et al. 2009). The smallest particle size was obtained using

the smallest (Y3O2)ZrO2 grinding medium, having a size of 90 µm, and a stirrer

tip speed of 9 m/s. Concerning energy efficiency, it can be seen from Fig. 20a that

lower stirrer tip speeds (9 m/s vs. 12 m/s) generally yield a smaller product size

with a lower specific energy.

1 10 100 100010

100

1000

10000

Vis

cosi

ty [

mP

as]

Shear rate [s-1]

NaPaPDI>2

NaPaPDI=1.2

3.0 wt% [x50

= 98 nm] 3.0 wt% [x50

= 75 nm]

4.5 wt% [x50

= 80 nm] 4.5 wt% [x50

= 74 nm]

CaCO3 in H

2O

cm= 0.40

pH = 10

64

Fig. 20. Median particle sizes and b) widths of the particle size distributions in the

nanogrinding experiments (Paper V, published by permission of John Wiley and

Sons).

As can be seen from Fig. 20b the widths of particle size distributions become

significantly narrower as the median particle size decreases: i.e. the distribution

decreases in width from around 2.5 to 0.5 as the median particle size decreases

from 200 nm to 25 nm. This powerful effect was not seen in the fine and

submicron grinding experiments (Figs. 15 and 17) which suggests that x10 does

not decrease any further once the grinding limit for CaCO3 particles has been

reached. Therefore only d90 is left to decrease, which is seen as a narrower

particle size distribution.

Optimal stress energy

The optimal stress energy was searched for by adjusting the operational

parameters of the mill. The median particle size as a function of stress energy (see

Equation 2) at different specific energies is shown in Fig. 21, where it is evident

that the optimum stress energy is around 2.5 µJ only for the lowest specific

energy (10000 kJ/kg). For a finer product no minimum was detectable: i.e. it

seems that the smaller the stress energy, the finer the product that can be obtained

at a certain specific energy.

10 100 10000

1

2

3CaCO

3 in H

2O

cm = 0.1

pH = 10[NaPa

PDI=1.2] = 4 wt%

GM/Size vt

(Y2O

3)ZrO

2/475 µm 12 m/s

(Y2O

3)ZrO

2/180 µm 9 m/s

(Y2O

3)ZrO

2/90 µm 9 m/s

SiO2/514 µm 12 m/s

SiO2/355 µm 10 m/s

SiO2/133 µm 9 m/s

Wid

th o

f par

ticle

siz

e d

istr

ibu

tion

[-]

Median particle size x50

[nm]b)a)

1000 10000 100000 10000001010

40

70

100100

40020002000

2500250030003000

CaCO3 in H

2O

cm = 0.10

pH = 10[NaPa

PDI=1.2] = 4 wt%

GM/Size vt

(Y2O

3)ZrO

2/475 µm 12 m/s

(Y2O

3)ZrO

2/180 µm 9 m/s

(Y2O

3)ZrO

2/90 µm 9 m/s

SiO2/514 µm 12 m/s

SiO2/355 µm 10 m/s

SiO2/133 µm 9 m/s

Med

ian

part

icle

siz

e x 50

[nm

]


65

Fig. 21. Median particle sizes as a function of stress energy at different specific

energies (modified from Paper V, published by permission of John Wiley and Sons).

Diffraction analysis and crystallite size

X-ray diffraction analyses were performed on the feed material (Eskal 500) and

the suspension having the smallest particle size (x50=26 nm) in order to obtain

information about possible phase transformation during grinding and about

crystallite sizes xcrys. The diffraction patterns and crystallite sizes of the samples

are presented in Fig. 22, where it can be seen that the peaks are at the same

location, so that no phase transformation occurs: i.e. both samples are 100%

calcite. At the same time a broadening of the peaks can be observed, which refers

to decreasing crystallite sizes. This was confirmed by means of a simultaneous

evaluation of the crystallite sizes with TOPAS, the results being 108 nm for Eskal

500 and 17.6 nm for ground CaCO3. The latter figure is almost the same as the

estimated grinding limit for CaCO3 (see section 2.4.4) and it can therefore be

assumed that the theoretical grinding limit for CaCO3 was achieved in this

experiment.

0.1 1 10 1001010

40

70100

400

7001000

CaCO3 in H

2O

cm = 0.1

pH = 10[NaPa

PDI=1.2] = 4 wt%

Specific Energy Em

10000 kJ/kg 25000 kJ/kg 50000 kJ/kg

Me

dia

n p

art

icle

siz

e x

50

[nm

]

Stress energy SEP [µJ]

66

Fig. 22. X-ray diffraction patterns of CaCO3 (modified from Paper V, published by

permission of John Wiley and Sons).

20 30 40 50 60 70 80

Feed (Eskal 500)x

crys= 108 nm (100% calcite)

Inte

nsi

ty [a

.u.]

2-Theta [°]

Ground CaCO3 (x

50= 26 nm)

xcrys

= 17.6 nm (100% calcite)

67

6 Conclusions

In the first part of this work titanium dioxide (TiO2) agglomerates having a crystal

size of around 200 nm were ground in a stirred media mill in multiple passage

mode with the aim of achieving an understanding of the parameters affecting

particle size and the width of the particle size distribution of TiO2.The optimal

properties for a TiO2 pigment are generally obtained when the particle size is

around 300 nm, the particle size distribution is narrow and the suspension is

stable. The most energy-efficient TiO2 grinding with a 300 nm particle size and

the narrowest particle size distribution was obtained with the lowest stress energy

(around 5 µJ), referring to the smallest grinding medium size. Electrostatic

stabilization was not effective for TiO2 when a high solids concentration (over 50

wt%) was to be used in grinding, but electrosteric stabilization with sodium

polyacrylates was more effective, and sodium polyacrylate with a molecular

weight of 12500 g/mol was found to be the most effective for reducing the

viscosity of the suspension.

In the second part of the work grinding experiments were performed on

CaCO3 particles with a stirred media mill in circuit mode aiming at a variety of

target particle sizes in order to study the effects of two sodium polyacrylates with

different polydispersity indices and therefore different adsorption behaviour. As

with TiO2, electrostatic stabilization was not effective, so that not even a particle

size of 1 µm could be obtained with a solids concentration of 20 wt% due to the

pronounced agglomeration of particles. When sodium polyacrylates were used as

grinding aids, however, the solids concentration could be doubled with the same

rheological behaviour, and at the same time a smaller particle size was achieved.

Comparing the sodium polyacrylates with different polydispersity indices, the low

polydispersity index was more effective, showing a better stabilization potential

in micron and submicron grinding and reducing the viscosity and particle size

more. Using this sodium polyacrylate with a better stabilization efficiency it was

possible to continue the grinding experiments for longer and obtain smaller

primary particle sizes. This led to nanogrinding experiments on CaCO3

suspensions having a solids concentration of 10 wt% using the low PDI sodium

polyacrylate, in which it was possible to obtain CaCO3 particles with a particle

size of 26 nm and a crystallite size of 17.6 nm, which is the smallest ever obtained

for CaCO3 by grinding. The calculated theoretical grinding limit for CaCO3 is

around 17 nm. The fact that this size was achieved when the crystallite size was

68

measured by XRD for the smallest CaCO3 suspension obtained (x50=26nm) was

an interesting finding.

When comparing the behaviour of TiO2 and CaCO3 it can be noted that these

materials differed in their stabilization behaviour, i.e. the possible solids

concentrations for a suspension, the pH required for the suspension, the dose of

sodium polyacrylate required and the optimum molecular weight for the sodium

polyacrylate. They also differed in their grinding behaviour, as TiO2 agglomerates

can be reduced from 1 µm to 300 nm with a specific energy of around 400 kJ/kg

whereas CaCO3 particles need ten times more energy in order to achieve a

fineness in the same range. This is due to the different nature of these particles:

TiO2 particles are crystals aggregated with chemical bonds, whereas CaCO3

particles are much stronger primary particles.

69

References

Adam R (1999) Padua TiO2 Conference: Industry Takes Stock. Paint Ink Int 12(24). Al-Nasra M (2013) Optimizing the Use of Sodium Polyacrylate in Plain Concrete. Int J

Eng Res Ind Appl 3(3): 1058–1062. Altun O, Benzer H & Enderle U (2013) Effects of operating parameters on the efficiency

of dry stirred milling. Minerals Eng 43–44: 58–66. Assefa S, McCann C & Sothcott J (2003) Velocities of compressional and shear waves in

limestones. Geophys Prospect 51(1): 1–13. Auvinen S, Alatalo M, Haario H, Jalava J-P & Lamminmäki R-J (2011) Size and shape

dependence of the electronic and spectral properties inTiO2 nanoparticles. J Phys Chem C 115(17): 8484–8493.

Auvinen S, Alatalo M, Haario H, Vartiainen E, Jalava J-P & Lamminmäki R-J (2013) Refractive index functions of Tio2 nanoparticles. J Phys Chem C 117(7): 3503–3512.

Balfour JG (1992) Fine particle titanium dioxide—its properties and uses<br />. Surf Coat Int 75(21–23): 35.

Barber DJ, Wenk H-R, Hirth G & Kohlstedt DL (2010) Dislocations in minerals. In Hirth JP & Kubin L (eds) Dislocations in Solids. Oxford, Elsevier North-Holland Publishing Company: 171–232.

Barnard B & Laverick WT (1980) Tioxide Group Limited, Billingham, England, assignee. Titanium dioxide pigment. Patent United Sates Patent 4,239,548.

Barner-Kowollik C (ed) (2008) Handbook of RAFT Polymerization. Weinheim, Germany, Wiley-VCH Verlag GmbH & Co. KGaA.

Bates LP & Harrison DL (1985) Precipitated Calcium Carbonate fillers in low modulus thixotropicsealants. Adhesives, Sealants and Encapsulants<br />. Buckingham, England, Network Events Ltd 1: 131–143.

Becker M, Kwade A & Schwedes J (2001) Stress intensity in stirred media mills and its effect on specific energy requirement. Int J Miner Process 61(3): 189–208.

Becker M & Schwedes J (1999) Comminution of ceramics in stirred media mills and wear of grinding beads. Powder Technol 105(1–3): 374–381.

Beinert S, Schilde C, Gronau G & Kwade A (2014) CFD-discrete element method simulations combined with compression experiments to characterize stirred-media mills. Chem Eng Technol 37(5):770–778.

Bel Fadhel H & Frances C (2001) Wet batch grinding of alumina hydrate in a stirred bead mill. Powder Technol 119(2–3): 257–268.

Bel Fadhel H, Frances C & Mamourian A (1999) Investigations on ultra-fine grinding of titanium dioxide in a stirred media mill. Powder Technol 105(1–3): 362–373.

Bentz DP, Sato T, De La Varga I & Weiss WJ (2012) Fine limestone additions to regulate setting in high volume fly ash mixtures. Cem Concr Compos 34(1): 11–17.

Bernard J, Belnou F, Houivet D & Haussonne J-M (2008) Synthesis of pure MgTiO3 by optimizing mixing/grinding condition of MgO + TiO2 powders. J Mater Process Technol 199(1): 150–155.

70

Bernhardt C, Reinsch E & Husemann K (1999) The influence of suspension properties on ultra-fine grinding in stirred ball mills. Powder Technol 105(1–3): 357–361.

Boisvert J-P, Persello J, Castaing J-C & Cabane B (2001) Dispersion of alumina-coated TiO2 particles by adsorption of sodium polyacrylate. Colloids Surf, A 178(1–3): 187–198.

Breitung-Faes S & Kwade A (2013a) Reduction of energy input and grinding media wear in wet stirred media milling by using an enhanced stress model. 13th European Symposium on Comminution & Classification. Braunschweig, Germany, Sierke Verlag: 138–141.

Breitung-Faes S & Kwade A (2008) Nano particle production in high-power-density mills. Chem Eng Res Design 86(4): 390–394.

Breitung-Faes S & Kwade A (2011) Production of transparent suspensions by real grinding of fused corundum. Powder Technol 212(3): 383–389.

Breitung-Faes S & Kwade A (2013b) Prediction of energy effective grinding conditions. Minerals Eng 43–44: 36–43.

Bunge F, Pietzsch M, Müller R & Syldatk C (1992) Mechanical disruption of Arthrobacter sp. DSM 3747 in stirred ball mills for the release of hydantoin-cleaving enzymes. Chem Eng Sci 47(1): 225–232.

Camiletti J, Soliman AM & Nehdi ML (2013) Effect of nano-calcium carbonate on early-age properties of ultrahigh-performance concrete. Magazine of Concrete Research 65(5): 297–307.

Chartier T, Souchard S, Baumard JF & Vesteghem H (1996) Degradation of Dispersant during Milling. J Eur Ceram Soc 16(12): 1283–1291.

Cho H, Waters MA & Hogg R (1996) Investigation of the grind limit in stirred-media milling. Int J Miner Process 44–45: 607–615.

Cho K, Chang H, Kil DS, Kim B-G & Jang HD (2009) Synthesis of dispersed CaCO3 nanoparticles by the ultrafine grinding. J Ind Eng Chem 15(2): 243–246.

Cho K, Suh YJ, Chang H, Kil DS, Kim BG & Jang H-D (2010) Synthesis of mesoporous CaCO3 particles by a spray drying method from the stable suspensions achieved in a beads mill. Advanced Powder Technology 21(2): 145–149.

Choi H, Lee W, Kim S & Chung H (2011) Effect of the sample concentration on the submicrometer particles produced during a stirred ball milling of calcite powders. Int J Appl Ceram Technol 8(5): 1147–1152.

Choi H, Lee W, Lee J, Chung H & Choi WS (2007) Ultra-fine grinding of inorganic powders by stirred ball mill: Effect of process parameters on the particle size distribution of ground products and grinding energy efficiency. Met Mater Int 13(4): 353–358.

Deiss JL, Anizan P, El Hadigui S & Wecker C (1996) Steric stability of TiO2 nanoparticles in aqueous dispersions. Colloids Surf, A 106(1): 59–62.

Deng D, Boyko V, Pancera SM, Nestle N & Tadros T (2010) Rheology investigations on the influence of addition sodium polyacrylate to calcium carbonate suspensions. Colloids Surf, A 372(1–3): 9–14.

71

Deng D, Boyko V, Pancera SM & Tadros T (2011) Rheological investigations on the influence of addition of sodium polyacrylate to titanium dioxide suspensions. Colloids Surf Physicochem Eng Aspects 389(1–3): 149–157.

Derjaguin BV & Landau LD (1941) Theory of stability of highly charged lyophobic sols and adhesion og highly charged particles in solutions of electrolytes. Acta Physicochim USSR 14: 633–652.

Droop D & Stein J (2011) Pulvis - A new and energy-efficient mill for ultrafine dry powders. CFI, Ceram Forum Int 88(10): D8–D13+E32–E37.

Dukhin AS & Goetz PJ (2010) Characterization of Liquids, Nano- and Microparticulates, and Porous Bodies Using Ultrasound. Amsterdam, Elsevier.

Dukhin AS, Ohshima H, Shilov VN & Goetz PJ (1999) Electroacoustics for concentrated dispersions. Langmuir 15(10): 3445–3451.

Eisermann C, Mallembakam MR, Damm C, Peukert W, Breitung-Faes S & Kwade A (2012) Polymeric stabilization of fused corundum during nanogrinding in stirred media mills. Powder Technol 217: 315–324.

Eriksson R, Merta J & Rosenholm JB (2007) The calcite/water interface. I. Surface charge in indifferent electrolyte media and the influence of low-molecular-weight polyelectrolyte. J Colloid Interface Sci 313(1): 184–193.

Eriksson R, Merta J & Rosenholm JB (2008) The calcite/water interface II. Effect of added lattice ions on the charge properties and adsorption of sodium polyacrylate. J Colloid Interface Sci 326(2): 396–402.

Farrokhpay S (2009) A review of polymeric dispersant stabilisation of titania pigment. Adv Colloid Interface Sci 151(1–2): 24–32.

Farrokhpay S (2012) Rheology of titania pigment slurry. Appl Rheol 22(5): 552851–552856.

Farrokhpay S, Morris GE, Fornasiero D & Self P (2004) Effects of chemical functional groups on the polymer adsorption behavior onto titania pigment particles. J Colloid Interface Sci 274(1): 33–40.

Farrokhpay S, Morris GE, Fornasiero D & Self P (2006) Titania pigment particles dispersion in water-based paint films. J Coat Technol Res 3(4): 275–283.

Farrokhpay S, Morris GE, Fornasiero D & Self P (2010) Stabilisation of titania pigment particles with anionic polymeric dispersants. Powder Technol 202(1–3): 143–150.

Fleer GJ & Lyklema J (1974) Polymer adsorption and its effect on the stability of hydrophobic colloids. II. The flocculation process as studied with the silver iodide-polyvinyl alcohol system. J Colloid Interface Sci 46(1): 1–12.

Foissy A, Persello J, Lamarche JM & Robert G (1982) Stability measurement and dispersion process of CaCO3 suspensions in sodium polyacrylate aqueous solutions. J Dispersion Sci Technol 3(2): 105–127.

Gao M-W & Forssberg E (1993) A study on the effect of parameters in stirred ball milling. Int J Miner Process 37(1–2): 45–59.

Garcia F, Bolay NL, Trompette J-L & Frances C (2004) On fragmentation and agglomeration phenomena in an ultrafine wet grinding process: The role of polyelectrolyte additives. Int J Miner Process 74: S43–S54.

72

Garcia F, Le Bolay N & Frances C (2002) Changes of surface and volume properties of calcite during a batch wet grinding process. Chem Eng J 85(2–3): 177–187.

Garcia F, Le Bolay N & Frances C (2003) Rheological behaviour and related granulometric properties of dense aggregated suspensions during an ultrafine comminution process. Powder Technol 130(1–3): 407–414.

Geffroy C, Persello J, Foissy A, Lixon P, Tournilhac F & Cabane B (2000) Molar mass selectivity in the adsorption of polyacrylates on calcite. Colloids Surf, A 162(1–3): 107–121.

Gesenhues U (1999) Substructure of titanium dioxide agglomerates from dry ball-milling experiments. J Nanopart Res 1(2): 223–234.

Gesenhues U (2001) Calcination of metatitanic acid to titanium dioxide white pigments. Chem Eng Technol 24(7): 685–694.

Greenwood R & Kendall K (1999) Selection of Suitable Dispersants for Aqueous Suspensions of Zirconia and Titania Powders using Acoustophoresis. J Eur Ceram Soc 19(4): 479–488.

Greenwood R, Rowson N, Kingman S & Brown G (2002) A new method for determining the optimum dispersant concentration in aqueous grinding. Powder Technol 123(2–3): 199–207.

Grübl P, Weigler P & Karl S (2001) Beton: Arten, Herstellung und Eigenschaften. Berlin, Germany, Ernst & Sohn Verlag.

Gryaznov VG, Polonsky IA, Romanov AE & Trusov LI (1991) Size effects of dislocation stability in nanocrystals. Phys Rev B 44(1): 42–46.

Hanaor D, Michelazzi M, Veronesi P, Leonelli C, Romagnoli M & Sorrell C (2011) Anodic aqueous electrophoretic deposition of titanium dioxide using carboxylic acids as dispersing agents. J Eur Ceram Soc 31(6): 1041–1047.

He M & Forssberg E (2007) Rheological behaviors in wet ultrafine grinding of limestone. Miner Metall Process 24(1): 19–29.

He M, Wang Y & Forssberg E (2004) Slurry rheology in wet ultrafine grinding of industrial minerals: A review. Powder Technol 147(1–3): 94–112.

He M, Wang Y & Forssberg E (2006a) Parameter effects on wet ultrafine grinding of limestone through slurry rheology in a stirred media mill. Powder Technol 161(1): 10–21.

He M, Wang Y & Forssberg E (2006b) Parameter studies on the rheology of limestone slurries. Int J Miner Process 78(2): 63–77.

Heijman SGJ & Stein HN (1995) Electrostatic and sterical stabilization of TiO2 dispersions. Langmuir 11(2): 422–427.

Heine H & et al, (1992) Pigments, Inorganic. In Elvers B, Hawkins S & Schultz G (eds) Ullmann’s Encyclopedia of Industrial Chemistry, Volume A20, Photography to Plastics. Weinheim, VCH Publishers: 243–369.

Hennart SLA, van Hee P, Drouet V, Domingues MC, Wildeboer WJ & Meesters GMH (2012) Characterization and modeling of a sub-micron milling process limited by agglomeration phenomena. Chem Eng Sci 71: 484–495.

73

Heusch R & Reizlein K (2005) Disperse Systems and Dispersants. In Anonymous Ullmann's Encyclopedia of Industrial Chemistry. Weinheim, Wiley-VCH Verlag GmbH & Co. KGaA: 1–26.

Holmnäs P-O (2000) Ultrafine Titanium Dioxide as UV-Absorber for Plastics. Fillers Addit Plast 4.

Hond R (1990) Circular grinding. Eur Coat J 12: 757–765. Inkyo M, Tahara T, Iwaki T, Iskandar F, Hogan Jr. CJ & Okuyama K (2006) Experimental

investigation of nanoparticle dispersion by beads milling with centrifugal bead separation. J Colloid Interface Sci 304(2): 535–540.

Jalava J-P (2006) The use of an exact light-scattering theory for spheroidal TiO2 pigment particles. Part Part Syst Charact 23(2): 159–164.

Jankovic A (2001) Media stress intensity analysis for vertical stirred mills. Minerals Eng 14(10): 1177–1186.

Jankovic A (2003) Variables affecting the fine grinding of minerals using stirred mills. Minerals Eng 16(4): 337–345.

Jankovic A & Sinclair S (2006) The shape of product size distributions in stirred mills. Minerals Eng 19(15): 1528–1536.

Jimbo G (1992) Chemical engineering analysis of fine grinding phenomena and process. J Chem Eng Jpn 25(2): 117–127.

Johnson RW, Thiele ES & French RH (1997) Light-scattering efficiency of white pigments: an analysis of model core - shell pigments vs. optimized rutile TiO2. Tappi J 80(11): 233–239.

Johnson AM, Trakhtenberg S, Cannon AS & Warner JC (2007) Effect of pH on the viscosity of titanium dioxide aqueous dispersions with carboxylic acids. J Phys Chem A 111(33): 8139–8146.

Joni IM, Ogi T, Purwanto A, Okuyama K, Saitoh T & Takeuchi K (2012) Decolorization of beads-milled TiO2 nanoparticles suspension in an organic solvent. Adv Powder Technol 23(1): 55–63.

Joni IM, Purwanto A, Iskandar F & Okuyama K (2009) Dispersion stability enhancement of titania nanoparticles in organic solvent using a bead mill process. Ind Eng Chem Res 48(15): 6916–6922.

Joost B & Schwedes J (1996a) Comminution of white fused alumina and wear of grinding beads in stirred media mills. CFI, Ceram Forum Int 73(6): 368–371.

Joost B & Schwedes J (1996b) Comminution of white fused alumina and wear of grinding beads in stirred media mills. CFI, Ceram Forum Int 73(7–8): 432–434.

Kadri EH & Duval R (2002) Effect of ultrafine particles on heat of hydration of cement mortars. ACI Mater J 99(2): 138–142.

Kandiel TA, Feldhoff A, Robben L, Dillert R & Bahnemann DW (2010) Tailored titanium dioxide nanomaterials: anatase nanoparticles and brookite nanorods as highly active photocatalysts. Chem Mater 22(6): 2050–2060.

Karakas F & Çelik MS (2013) Mechanism of TiO2 stabilization by low molecular weight NaPAA in reference to water-borne paint suspensions. Colloids Surf Physicochem Eng Aspects 434: 185–193.

74

Karbstein H, Mueller F & Polke R (1995) Producing suspensions with steep particle size distributions in fines ranges. Aufbereitungs-Technik 36(10): 464–473.

Karbstein H, Müller F & Polke R (1996) Scale-up for grinding in stirred ball mills. Aufbereitungs-Technik/Mineral Processing 37(10): 469–479.

Karis TE, Novotny VJ & Johnson RD (1993) Mechanical scission of perfluoropolyethers. J Appl Polym Sci 50(8): 1357–1368.

Klimpel RR (1999) The selection of wet grinding chemical additives based on slurry rheology control. Powder Technol 105(1–3): 430–435.

Knieke C (2012) Fracture at the Nanoscale and the Limit of Grinding. PhD thesis. Cuvillier Verlag, Göttingen, Germany, Universität Erlangen-Nürnberg.

Knieke C, Romeis S & Peukert W (2011) Influence of process parameters on breakage kinetics and grinding limit at the nanoscale. AIChE J 57(7): 1751–1758.

Knieke C, Sommer M & Peukert W (2009) Identifying the apparent and true grinding limit. Powder Technol 195(1): 25–30.

Knieke C, Steinborn C, Romeis S, Peukert W, Breitung-Faes S & Kwade A (2010) Nanoparticle production with stirred-media mills: Opportunities and limits. Chem Eng Technol 33(9): 1401–1411.

Kotake N, Kuboki M, Kiya S & Kanda Y (2011) Influence of dry and wet grinding conditions on fineness and shape of particle size distribution of product in a ball mill. Adv Powder Technol 22(1): 86–92.

Kovacevic V, Lucic S & Leskovac M (2002) Morphology and failure in nanocomposites. Part I: Structural and mechanical properties. J Adhes Sci Technol 16(10): 1343–1365.

Krumpfer JW, Schuster T, Klapper M & Müllen K (2013) Make it nano-Keep it nano. Nano Today 8(4): 417–438.

Kuiper J, Van Ryen WML & Koefoed O (1959) Laboratory detarminations of elastic properties of some limestones. Geophys Prospect 7(1): 38–44.

Kumar V, Dev A & Gupta AP (2014) Studies of poly(lactic acid) based calcium carbonate nanocomposites. Composites, Part B 56: 184–188.

Kwade A (2003) A stressing model for the description and optimization of grinding processes. Chem Eng Technol 26(2): 199–205.

Kwade A (2004) Mill selection and process optimization using a physical grinding model. Int J Miner Process 74: S93–S101.

Kwade A, Blecher L & Schwedes J (1996) Motion and stress intensity of grinding beads in a stirred media mill. Part 2: Stress intensity and its effect on comminution. Powder Technol 86(1): 69–76.

Kwade A & Schwedes J (2007) Chapter 6 Wet Grinding in Stirred Media Mills. Handbook of Powder Technology 12: 251–382.

Lamarche J-M, Perselio J & Foissy A (1983) Influence of molecular weight of sodium polyacrylate in calcium carbonate aqueous dispersions. Ind Eng Chem Prod Res Dev 22(1): 123–126.

Liddell PV & Boger DV (1994) Influence of processing on the rheology of titanium dioxide pigment suspensions. Ind Eng Chem Res 33(10): 2437–2442.

75

Liimatainen H, Sirviö J, Haapala A, Hormi O & Niinimäki J (2011) Characterization of highly accessible cellulose microfibers generated by wet stirred media milling. Carbohydr Polym 83(4): 2005–2010.

Liu Q, Wang Q & Xiang L (2008) Influence of poly acrylic acid on the dispersion of calcite nano-particles. Appl Surf Sci 254(21): 7104–7108.

Liu Y, Yu Z, Zhou S & Wu L (2006) De-agglomeration and dispersion of Nano-TiO2 in an agitator bead mill. J Dispersion Sci Technol 27(7): 983–990.

Loebbert G & Sharangpani ASG (2000) Pigment Dispersions. Kirk-Othmer Encyclopedia of Chemical Technology. John Wiley & Sons, Inc.

Loiseau J, Doërr N, Suau JM, Egraz JB, Llauro MF, Ladavière C & Claverie J (2003) Synthesis and characterization of poly(acrylic acid) produced by RAFT polymerization. Application as a very efficient dispersant of CaCO3, Kaolin, and TiO2. Macromolecules 36(9): 3066–3077.

Loiseau J, Ladavière C, Suau JM & Claverie J (2005) Dispersion of calcite by poly(sodium acrylate) prepared by reversible addition-fragmentation chain transfer (RAFT) polymerization. Polymer 46: 8565–8572.

Losoi T (1997) Control of optical properties of TiO2 pigments. 5th International Exhibition of Paint Industry Suppliers, September 1997, Sao Paulo Brazil.

Lynch TJ (2011) Kemira Oyj, assignee. Mineral Dispersants and Methods for Preparing Mineral Slurries Using the Same. Patent US Patent 7875213 B2.

Lynch TJ, Macy PK & Mondesir B (2012) New polymeric dispersants for industrial minerals. 2012 SME Annual Meeting and Exhibit 2012, SME 2012, Meeting Preprint, 276–278.

Mandzy N, Grulke E & Druffel T (2005) Breakage of TiO2 agglomerates in electrostatically stabilized aqueous dispersions. Powder Technol 160(2): 121–126.

Mankosa MJ, Adel GT & Yoon RH (1989) Effect of operating parameters in stirred ball mill grinding of coal. Powder Technol 59(4): 255–260.

Mende S, Stenger F, Peukert W & Schwedes J (2003) Mechanical production and stabilization of submicron particles in stirred media mills. Powder Technol 132(1): 64–73.

Mende S, Stenger F, Peukert W & Schwedes J (2004) Production of sub-micron particles by wet comminution in stirred media mills. J Mater Sci 39(16–17): 5223–5226.

Mikulášek P, Wakeman RJ & Marchant JQ (1997) The influence of pH and temperature on the rheology and stability of aqueous titanium dioxide dispersions. Chem Eng J 67(2): 97–102.

Mölls HH & Hörnle R (1972) Wirkungsmechanismus der Nasszerkleinerung in der Rührwerkskugelmühle. DECHEMA, Monography 69(TI 2): 631–661.

Morris GE, Skinner WA, Self PG & Smart RSC (1999) Surface chemistry and rheological behaviour of titania pigment suspensions. Colloids Surf, A 155(1): 27–41.

Mosquet M, Chevalier Y, Le Perchec P & Guicquero J-P (1997) Functional polyethylene oxides as dispersing agents. New J Chem 21(2): 143–145.

76

Moudgil BM, Singh PK & Adler JJ (2002) Surface Chemistry in Dispersion, Flocculation and Flotation. In Holmberg K (ed) Handbook of Applied Surface and Colloid Chemistry, Volumes 1–2. New York, John Wiley & Sons.

Mucsi G (2013) Grindability of quartz in stirred media mill. Particul Sci Technol 31(4): 399–406.

Mucsi G, Rácz Á & Mádai V (2013) Mechanical activation of cement in stirred media mill. Powder Technol 235: 163–172.

Müller F, Peukert W, Polke R & Stenger F (2004) Dispersing nanoparticles in liquids. Int J Miner Process 74: S31–S41.

Müller F & Polke RF (1999) From the product and process requirements to the milling facility. Powder Technol 105(1–3): 2–13.

Muster TH & Prestidge CA (1995) Rheological investigations of sulphide mineral slurries. Minerals Eng 8(12): 1541–1555.

Napper DH (1977) Steric stabilization. J Colloid Interface Sci 58(2): 390–407. Nesset JA, Radziszewski P, Hardie C & Leroux DP (2006) Assessing the performance and

efficiency of fine grinding technologies. Proceedings of 38th Annual Canadiasn mineral Processor Operators Conference. Ottawa, Canada: 283–310.

Nieh TG & Wadsworth J (1991) Hall-petch relation in nanocrystalline solids. Scr Metall Mater 25(4): 955–958.

Patel CM, Murthy ZVP & Chakraborty M (2012) Effects of operating parameters on the production of barium sulfate nanoparticles in stirred media mill. J Ind Eng Chem 18(4): 1450–1457.

Peukert W, Schwarzer H-C & Stenger F (2005) Control of aggregation in production and handling of nanoparticles. Chem Eng Process 44(2): 245–252.

Pradeep PR & Pitchumani B (2011) Effect of operating variables on the production of nanoparticles by stirred media milling. Asia-Pacific Journal of Chemical Engineering 6(1): 154–162.

Prestidge CA (1997) Rheological investigations of galena particle interactions. Colloids Surf, A 126(2–3): 75–83.

Qiu J, Charleux B & Matyjaszewski K (2001) Controlled/living radical polymerization in aqueous media: Homogeneous and heterogeneous systems. Progress in Polymer Science (Oxford) 26(10): 2083–2134.

Qiu S, Dong J & Chen G (2000) Wear and friction behaviour of CaCO3 nanoparticles used as additives in lubricating oils. Lubr Sci 12(2): 205–212.

Ran Q, Wu S, Shen J, Ran Q & Shen J (2007) Effects of poly(acrylic acid) on rheological and dispersion properties of aqueous TiO2 suspensions. Polym-Plast Technol Eng 46(11): 1117–1120.

Reinisch E, Bernhardt C & Husemann K (2001) The influence of additives during wet ultra-fine grinding in agitator bead mills: Part I: General principles and experimental. CFI, Ceram Forum Int 78(3): E38–E42.

Reinsch E, Bernhardt C & Husemann K (2001) The influence of additives during wet ultra-fine grinding in agitator bead mills: Part 2: Results and conclusions. CFI, Ceram Forum Int 78(4): E36–E40.

77

Roberts JC (1996) The chemistry of paper. Cambridge, United Kingdom, Royal Society of Chemistry.

Saari J, Kataja K, Qvintus-Leino P, Mikkonen H & Joyce M (2006) Organic starch based pigments in paper coatings. TAPPI Advanced Coating Fundamentals Symposium, 21–31.

Safaei-Naeini Y, Aminzare M, Golestani-Fard F, Khorasanizadeh F & Salahi E (2012) Suspension stability of Titania nanoparticles studied by UV-VIS spectroscopy method. Iran J Mater Sci Eng 9(1): 62–68.

Samanli S, Cuhadaroglu D, Ucbas Y & Ipek H (2010) Investigation of breakage behavior of coal in a laboratory-scale stirred media mill. Int J Coal Prep Util 30(1): 20–31.

Sato K, Li J-G, Kamiya H & Ishigaki T (2008) Ultrasonic dispersion of TiO2 nanoparticles in aqueous suspension. J Am Ceram Soc 91(8): 2481–2487.

Sato T & Beaudoin JJ (2011) Effect of nano-CaCO3 on hydration of cement containing supplementary cementitious materials. Adv Cem Res 23(1): 33–43.

Schilde C, Breitung-Faes S & Kwade A (2007) Dispersing and grinding of alumina nano particles by different stress mechanisms. Ceram Forum Int 84(13): 12–17.

Schilde C, Kampen I & Kwade A (2010) Dispersion kinetics of nano-sized particles for different dispersing machines. Chem Eng Sci 65(11): 3518–3527.

Schilde C, Mages-Sauter C, Kwade A & Schuchmann HP (2011) Efficiency of different dispersing devices for dispersing nanosized silica and alumina. Powder Technol 207(1–3): 353–361.

Schulte K (2002) Thinking small with TiO2. Asia Pac Coat J 15: 23–26. Sentein C, Guizard B, Giraud S, Yé C & Ténégal F (2009) Dispersion and stability of

TiO2 nanoparticles synthesized by laser pyrolysis in aqueous suspensions. J Phys: Conf Ser 170.

Singh BP, Nayak S, Samal S, Bhattacharjee S & Besra L (2012) The role of poly(methacrylic acid) conformation on dispersion behavior of nano TiO2 powder. Appl Surf Sci 258(8): 3524–3531.

Somasundaran P & Agar GE (1967) The zero point of charge of calcite. J Colloid Interface Sci 24(4): 433–440.

Sommer MM (2007) Mechanical Production of Nanoparticles in Stirred Media Mills. PhD thesis. Cuvillier Verlag, Göttingen, Germany, Universität Erlangen-Nürnberg.

Sommer M & Peukert W (2007) Chapter 13 Enabling Nanomilling through Control of Particulate Interfaces. Handbook of Powder Technology 12: 551–603.

Stadler R, Polke R, Schwedes J & Vock F (1990) Wet grinding in agitated ball mills. Chem Ing Tech 62(11): 907–915.

Stender H-H, Kwade A & Schwedes J (2004) Stress energy distribution in different stirred media mill geometries. Int J Miner Process 74: S103–S117.

Stenger F, Mende S, Schwedes J & Peukert W (2005) Nanomilling in stirred media mills. Chemical Engineering Science 60(16): 4557–4565.

Strauss H, Heegn H & Strienitz I (1993) Effect of PAA adsorption on stability and rheology of TiO2 dispersions. Chem Eng Sci 48(2): 323–332.

78

Tang F, Uchikoshi T, Ozawa K & Sakka Y (2006) Effect of polyethylenimine on the dispersion and electrophoretic deposition of nano-sized titania aqueous suspensions. J Eur Ceram Soc 26(9): 1555–1560.

Tangsathitkulchai C (2003) The effect of slurry rheology on fine grinding in a laboratory ball mill. Int J Miner Process 69(1–4): 29–47.

Tobori N & Amari T (2003) Rheological behavior of highly concentrated aqueous calcium carbonate suspensions in the presence of polyelectrolytes. Colloids Surf, A 215(1–3): 163–171.

Tseng WJ & Lin K-C (2003) Rheology and colloidal structure of aqueous TiO2 nanoparticle suspensions. Mater Sci Eng A 355(1–2): 186–192.

Tsuzuki T, Pethick K & McCormick PG (2000) Synthesis of CaCO3 nanoparticles by mcchanochemical processing. J Nanopart Res 2(4): 375–380.

Tuunila R (1997) Ultrafine grinding of FGD and phosphogypsum with an attrition bead mill and a jet mill: optimisation and modelling of grinding and mill comparison. PhD thesis. Lappeenranta University of Technology, Department of Chemical Technology.

Vallar S, Houivet D, El Fallah J, Kervadec D & Haussonne J-M (1999) Oxide slurries stability and powders dispersion: Optimization with zeta potential and rheological measurements. J Eur Ceram Soc 19(6–7): 1017–1021.

Vdovic N & Bišcan J (1998) Electrokinetics of natural and synthetic calcite suspensions. Colloids Surf, A 137(1–3): 7–14.

Verwey EJW & Oberbeek JTG (1948) Theory of the stability of lyophotic colloids<br />. Amsterdam, Elsevier.

Wang Y & Forssberg E (2000) Technical note product size distribution in stirred media mills. Minerals Eng 13(4): 459–465.

Wang Y & Forssberg E (2006) Production of carbonate and silica nano-particles in stirred bead milling. Int J Miner Process 81(1): 1–14.

Wang Y & Forssberg E (2007) Enhancement of energy efficiency for mechanical production of fine and ultra-fine particles in comminution. China Particuol 5(3): 193–201.

Wang Y, Li J, Wang L, Xue T & Qi T (2010) Preparation of rutile titanium dioxide white pigment via doping and calcination of metatitanic acid obtained by the NaOH molten salt method. Ind Eng Chem Res 49(16): 7693–7696.

Wang Y, Lu L, Yang H & Che Q (2013) Development of high dispersed TiO2 paste for transparent screen-printable self-cleaning coatings on glass. J Nanopart Res 15(1).

Way HW (2004) Grinding and dispersing nanoparticles. JCT CoatingsTech 1(1): 54–60. Weller KR, Gao M & Wilhelmsson B (1997) Media wear in stirred milling. Particul Sci

Technol 15(2). Widegren J & Bergström L (2000) The effect of acids and bases on the dispersion and

stabilization of ceramic particles in ethanol. J Eur Ceram Soc 20(6): 659–665. Widegren J & Bergström L (2002) Electrostatic stabilization of ultrafine titania in ethanol.

J Am Ceram Soc 85(3): 523–528. Yang H-G, Li C-Z, Gu H-C & Fang T-N (2001) Rheological behavior of titanium dioxide

suspensions. J Colloid Interface Sci 236(1): 96–103.

79

Yoshioka K, Okuda T, Fujii H, Kamimoto Y & Kamiya H (2013) Effect of TiO2 nanoparticles aggregation in silicate thin coating films on photocatalytic behavior for antifouling materials. Adv Powder Technol 24(5): 886–890.

Zhang X, Wang DK, Lopez DRS & Diniz da Costa JC (2014) Fabrication of nanostructured TiO2 hollow fiber photocatalytic membrane and application for wastewater treatment. Chem Eng J 236: 314–322.

Zheng J, Harris CC & Somasundaran P (1996) A study on grinding and energy input in stirred media mills. Powder Technol 86(2): 171–178.

Zheng J, Harris CC & Somasundaran P (1997) The effect of additives on stirred media milling of limestone. Powder Technol 91(3): 173–179.

81

Original publications

I Ohenoja K, Illikainen M & Niinimäki J (2013) Effect of operational parameters and stress energies on the particle size distribution of TiO2 pigment in stirred media milling. Powder Technol 234: 91–96.

II Ohenoja K, Saari J, Illikainen M & Niinimäki J (2014) Effect of molecular weight of sodium polyacrylates on the particle size distribution and stability of a TiO2 suspension in aqueous stirred media milling. Powder Technol 262: 188–193.

III Ohenoja K, Saari J, Illikainen M & Niinimäki J (2013) Effect of the molecular structure of polymer dispersants on limestone suspension properties in fine grinding. J Chem Chem Eng 7: 606–612.

IV Ohenoja K, Saari J, Illikainen M, Breitung-Faes S, Kwade A & Niinimäki J (2014) Effect of polydispersity index on the grinding limits of highly concentrated limestone suspensions. Chem Eng Technol 37(5): 833–839.

V Ohenoja K, Breitung-Faes S, Kinnunen P, Illikainen M, Saari J, Kwade A & Niinimäki J (2014) Ultrafine Grinding of Limestone with Sodium Polyacrylates as Additives in Ordinary Portland Cement Mortar. Chem Eng Technol 37(5): 787–794.

Reprinted with permission from Elsevier (I–II), David Publishing (III) and John

Wiley and Sons (IV–V).

The original publications are not included in the electronic version of this thesis.


Book orders:Granum: Virtual book storehttp://granum.uta.fi/granum/

S E R I E S C T E C H N I C A

484. Mäklin, Jani (2014) Electrical and thermal applications of carbon nanotube films

485. Niemistö, Johanna (2014) Towards sustainable and efficient biofuels production :use of pervaporation in product recovery and purification

486. Liu, Meirong (2014) Efficient super-peer-based coordinated service provision

487. Väyrynen, Eero (2014) Emotion recognition from speech using prosodic features

488. Celentano, Ulrico (2014) Dependable cognitive wireless networking : modellingand design

489. Peräntie, Jani (2014) Electric-field-induced dielectric and caloric effects in relaxorferroelectrics

490. Aapaoja, Aki (2014) Enhancing value creation of construction projects throughearly stakeholder involvement and integration

491. Rossi, Pekka M. (2014) Integrated management of groundwater and dependentecosystems in a Finnish esker

492. Sliz, Rafal (2014) Analysis of wetting and optical properties of materials developedfor novel printed solar cells

493. Juntunen, Jouni (2014) Enhancing organizational ambidexterity of the FinnishDefence Forces’ supply chain management

494. Hänninen, Kai (2014) Rapid productisation process : managing an unexpectedproduct increment

495. Mehtonen, Saara (2014) The behavior of stabilized high-chromium ferriticstainless steels in hot deformation

496. Majava, Jukka (2014) Product development : drivers, stakeholders, and customerrepresentation during early development

497. Myllylä, Teemu (2014) Multimodal biomedical measurement methods to studybrain functions simultaneously with functional magnetic resonance imaging

498. Tamminen, Satu (2014) Modelling the rejection probability of a quality testconsisting of multiple measurements

499. Tuovinen, Lauri (2014) From machine learning to learning with machines :remodeling the knowledge discovery process

500. Hosio, Simo (2014) Leveraging Social Networking Services on MultipurposePublic Displays

C501etukansi.kesken.fm Page 2 Tuesday, September 2, 2014 9:41 AM

ABCDEFG

UNIVERSITY OF OULU P .O. B 00 F I -90014 UNIVERSITY OF OULU FINLAND


S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Esa Hohtola

University Lecturer Santeri Palviainen

Postdoctoral research fellow Sanna Taskila


University Lecturer Veli-Matti Ulvinen

Director Sinikka Eskelinen

Professor Jari Juga


Publications Editor Kirsti Nurkkala

ISBN 978-952-62-0549-6 (Paperback)ISBN 978-952-62-0550-2 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)


TECHNICA


TECHNICA

OULU 2014

C 501

Katja Ohenoja

PARTICLE SIZE DISTRIBUTION AND SUSPENSION STABILITYIN AQUEOUS SUBMICRON GRINDING OF CaCO3 AND TiO2

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

C 501

ACTA

Katja O

henojaC501etukansi.kesken.fm Page 1 Tuesday, September 2, 2014 9:41 AM

particle size distribution and suspension stability in aqueous

Documents