Assessment of Ceramic Raw Materials in Uganda for Electrical Porcelain

Peter Wilberforce Olupot

Licentiate Thesis in Material Science Department of Materials Science and Engineering

Royal Institute of Technology (KTH) Stockholm, Sweden 2006

Akademisk avhandling som med tillstånd av Kungl Tekniska Högskolan framlägges till offentlig granskning för avläggande av teknisk licentiatexamen fredagen den 2 juni 2003 kl. 10.00 vid Institutionen för Materialvetenskap, Kungl Tekniska Högskolan, konferensrummet K408, 4:e våningen, Brinellvägen 23, Stockholm. Fakultetsopponent är docent Thommy Ekström, Lönnviksvägen 66, 178 90, Ekerö. ISBN 91-7178-408-X ISRN KTH/MSE--06/45--SE+MEK/AVH © Peter Wilberforce Olupot, April 2006

i

Abstract Clay, quartz and feldspar are widely available in Uganda. The location and properties of various clay deposits are reported in the literature, but little is reported on feldspar and quartz deposits. In this work an extended literature on ceramics and porcelains in particular, is documented. Samples from two deposits of feldspar and two deposits of quartz are characterised and found to possess requisite properties for making porcelain insulators. Sample porcelain bodies are made from materials collected from selected deposits using different mixing proportions of clay, feldspar and quartz. Their properties in relation to workability, firing temperature, dielectric and bending strengths are studied. It is found that a mixture consisting of 30% Mutaka kaolin, 15% Mukono ball clay, 30% Mutaka feldspar and 25% Lido beach flint yields a body with highest mechanical strength (72MPa) and dielectric strength (19kV/mm) when fired at 1250°C. The strength (both mechanical and dielectric) is found to decrease with increasing firing temperature. At high firing temperatures, the undissolved quartz in the body decreased, the glass content increases and pores are formed. Mullite content on the other hand does not change at temperatures above 1200°C but there are significant differences in the morphologies of the mullite crystals in the samples. Optimum mechanical and electrical properties are found at maximum virtification and a microstructure showing small closely packed mullite needles. Keywords: Porcelain, characterisation, bending strength, dielectric strength, Uganda.

ii

Contents 1. Introduction................................................................................................................. 1 2. Presentation of the Thesis ........................................................................................... 2 3. Porcelains.................................................................................................................... 3

3.1 Raw Materials ..................................................................................................... 3 3.2 Strength Considerations ...................................................................................... 3 3.3 Source of the Raw Materials............................................................................... 4

4. Experimental Techniques............................................................................................ 7 4.1 Characterisation of Feldspars and Quartz Raw Materials................................... 7 4.2 Formulation of Porcelains................................................................................... 7

5. Summary of Results.................................................................................................... 8 5.1 Characterisation of Raw Materials...................................................................... 8

5.1.1 Chemical composition of the deposits......................................................... 8 5.1.2 Microstructure of the minerals ................................................................... 9 5.1.3 Thermal analysis ......................................................................................... 9 5.1.4 Gravimetry ................................................................................................ 10 5.1.5 Mineralogy ................................................................................................ 11

5.2 Formulated Porcelains ...................................................................................... 11 5.2.1 Properties after Firing ..................................................................................... 11 5.2.2 Formability ...................................................................................................... 11 5.2.3 Microstructure and phase analyses of fired samples....................................... 12

6. Conclusion ................................................................................................................ 16 7. Proposals for Further Investigations ......................................................................... 17 Acknowledgments............................................................................................................. 17 References......................................................................................................................... 18 Appended Papers .............................................................................................................. 20

1

1. Introduction Porcelains are vitrified and fine grained ceramic whitewares, used either glazed or unglazed. They are widely used in household, laboratory and industrial applications. For technical purposes, porcelain products are designated as electrical, chemical, mechanical, structural and thermal wares. Porcelains are primarily composed of clay, feldspar and a filler material, usually quartz or alumina. The clay [Al2Si2O5(OH)4], gives plasticity to the ceramic mixture; flint or quartz (SiO2), maintains the shape of the formed article during firing; and feldspar [KxNa1-x(AlSi3)O8], serves as flux. These three constituents place electrical porcelain in the phase system [(K,Na)2O-Al2O3-SiO2)] in terms of oxide constituents, hence the term triaxial porcelains (Buchanan 1991). The fired product contains mullite (Al6Si2O13) and undissolved quartz (SiO2) crystals embedded in a continuous glassy phase, originating from feldspar and other low melting impurities in the raw materials. By varying the proportions of the three main ingredients, it is possible to emphasize thermal, dielectric or mechanical properties, as illustrated by Thurnauer (1954). For electrical insulation applications, porcelains are expected to meet minimum specifications of the latter two. Electrical porcelains are widely used as insulators in electrical power transmission systems due to the high stability of their electrical, mechanical and thermal properties in the presence of harsh environments (Kingery, 1967). These are the reasons for their continued use over the centuries despite the emergence of new materials like plastics and composites. They form a large base of the commonly used ceramic insulators for both low and high tension insulation. They are considered to be one of the most complex ceramic materials and represent the most widely studied ceramic system (Dana et al, 2004). Still there remain significant challenges in understanding porcelains in relation to raw materials, processing science, phase and microstructure evolution (Carty & Senapati, 1998). Clay, quartz and feldspar are widely available in Uganda. Due to demand and its wide application, most of the previous studies revealed the location and properties of various clay deposits (Nyakairu and Kaahwa 1998; Nyakairu and Koeberl, 2001; Nyakairu, Koeberl and Kurzweil 2001; Kirabira, et al, 2005) whereas little evaluation of the feldspar deposits were carried out due to low demand of this mineral in East Africa. This picture was pointed out long ago by Kabagambe-Kaliisa (1983) and Engelthaler and Engena (1972) and has not changed much up to present date. As a result, the present thesis focussed on feldspar and quartz deposits in Uganda by collecting and characterising feldspar samples from Lunya (Mukono district) and Mutaka (Bushenyi district) and quartz samples from Mutaka and Lido beach (Entebbe). Powder material samples from these deposits were characterised for their chemical and physical properties in order to establish their potential for use in electrical porcelain insulators. Previous characterisation studies on Ugandan clay deposits by Kirabira et al (2005) are quite exhaustive and reveal good properties of Mutaka kaolin and Mukono ball clay for application in porcelain development. As a result, porcelain bodies were formulated from clays from these deposits, Mutaka feldspar and Lido beach sand in order to establish the

2

optimum mixing proportions of kaolin, ball clay, feldspar and flint and optimum firing temperature for producing superior mechanical and dielectric properties. 2. Presentation of the Thesis The thesis deals with the processes of evaluation of properties of raw materials for production of electrical porcelains. It includes an extended literature survey explaining the various developments relating to porcelain properties, composition and production process. Results of the properties of sample porcelains made from the raw materials studied, and fabricated by a process suitable for industrial production are also included. The thesis is organised in the following papers. 1. “State of the art paper on development of electric porcelain insulators from Ugandan

raw materials” Peter Wiberforce Olupot, Stefan Jonsson, Joseph Kadoma Byaruhanga. Unpublished report

2. “Characterization of Feldspar and Quartz Raw Materials in Uganda for Manufacture

of Electrical Porcelains” Peter W. Olupot, Stefan Jonsson, Joseph K. Byaruhanga. J. Aust. Ceram. Soc. 41[1] (2006) 29-35.

3. “Optimization of composition and firing temperature for high-strength electric

porcelains from Ugandan materials” Peter W. Olupot, Stefan Jonsson, Joseph K. Byaruhanga. Manuscript ready for submission

Paper 1, is a “state of the art paper” and covers the classifications of ceramics and porcelains, the major properties of porcelains for insulation requirements and the manufacturing processes for porcelains. The recent studies on triaxial porcelains, especially those emphasising improvements in various aspects of manufacturing and mechanical properties, are covered. The different methods for characterisation of raw materials are discussed. Raw material deposits in Uganda are mentioned and the benefits of exploiting them for the manufacture of electric porcelains are highlighted. Paper 2 contains results on the characterisation of selected quartz and feldspar deposits Samples from two deposits of each feldspar and silica were investigated to assess their potential as raw materials in the manufacture of electric porcelains. Raw samples ground to powder form were investigated by means of X-ray diffraction, thermal analysis, and scanning electron microscopy. In addition, the chemical composition, particle size distribution and density of the powders were determined. Paper 3 contains the results of the variation of the properties of sample porcelain formulations from selected raw material deposits in Uganda. The mechanical, dielectric strength and water absorption together with workability properties of the samples were investigated in relation to composition and firing temperatures. This study was carried out so as to identify the optimum composition and the appropriate firing temperatures for optimum dielectric and mechanical strength, with good formability characteristics using a forming process that is quite suited for industrial production.

3

3. Porcelains 3.1 Raw Materials The basic raw materials for electric porcelains are quartz, feldspar, ball clay and kaolin. These materials are also used in the production of various whiteware products. The distinguishing factor in the properties of different porcelain products are brought about by variations in the proportion of these materials, the processing and the firing schedule adopted. For electric porcelains, the quest over the period of time has been to increase mechanical strength, and to reduce the production costs. A number of such studies are cited in Paper 1. In most efforts to increase strength, emphasis has been placed on minimisation of quartz in the porcelain formula because of the β to α phase inversion of quartz which occurs at 573°C during cooling. The inversion results into decrease of quartz particle volume and may lead to cracks in the body. So far, there are reports of improvements in the mechanical properties by reducing/eliminating the use of quartz. These include replacements of quartz with kyanite (Schroeder 1978), alumina (Kobayashi et al, 1987, Das and Dana 2003), rice husk ash (Prasad et al, 2001), sillimanite sand (Maity and Sarkar 1996), fly ash (Dana et al, 2004), partial replacement of feldspar and quartz by fly ash and blast furnace slag (Dana et al, 2005), silica fume (Prasad et al, 2002), with a mixture of rice husk ash and silica fume (Prasad et al, 2003). In this context, it can also be mentioned that an attempt to substitute part of quartz with fired porcelain by Stathis et al (2004) did not result in a positive effect on the bending strength. Other modifications on the triaxial porcelain system, which have proven successful include, replacement of clay with aluminous cement (Tai et al, 2002), substitution of feldspar with nepheline syenite (Esposito, et al, 2005), use of soda feldspar in preference to potash feldspar (Das and Dana 2003), partial substitution of feldspar by blast furnace slag (Dana and Das 2004), use of recycled glass powder to replace feldspar to reduce firing temperature (Bragança and Bergmann 2004). On the other hand, there is evidence that under optimized conditions of firing and for a particle size of 10-30μm (Norton, 1970; Ece and Nakagawa, 2002; Bragança and Bergmann, 2003), quartz has a beneficial effect on the strength of porcelain, in conformity with the pre-stressing theory. For small particle sizes, the dissolution is more rapid leaving less quartz crystals in the glass and hence yielding a low pre-stress and low strength of the material. For large particle sizes an interconnected matrix with favourable crack path is formed leading to low strength (Carty and Senapati, 1998). Hence, quartz grain size affects bending strength in two ways, that is, directly through the induction of compressive stresses to the vitreous phase and indirectly through the development of a favourable microstructure (Stathis et al, 2004). Thus, there is an optimum particle size of quartz for mechanical strength. 3.2 Strength Considerations The great interest in strength of porcelain for power transmission installation and the wide research on the porcelain system have resulted in three major hypotheses describing the strength properties of porcelain formulations. These were described by Carty and

4

Senapati (1998) as the mullite hypothesis, the matrix reinforcement hypothesis and the dispersion strengthening hypothesis respectively. The mullite hypothesis suggests that porcelain strength depends on the felt-like interlocking of fine mullite needles. Specifically, the higher the mullite content and the higher the interlocking of the mullite needles, the higher is the strength. Hence the strength of porcelain depends on the factors that affect the amount and size of mullite needles, like the firing temperature and composition of alumina and silica in the raw materials. The matrix reinforcement hypothesis concerns the development of compressive stresses in the vitreous phase as a result of the different thermal expansion coefficients of dispersed particles, or crystalline phases, and the surrounding vitreous phase. The larger these stresses are, the higher is the strength of the porcelain body. The phenomenon is known as the pre-stressing effect. The dispersion strengthening hypothesis, on the other hand, states that dispersed particles in the vitreous phase of a porcelain body, such as quartz and mullite crystals in the glassy phase, limit the size of Griffith flaws resulting in increased strength. There is evidence supporting each of these hypotheses (Maity and Sarkar 1996, Stathis et al 2004, Islam et al, 2004). Carty and Senapati (1998) concluded that the typical strength controlling factors in multiphase polycrystalline ceramics are thermal expansion coefficients of the phases, elastic properties of the phases, volume fraction of different phases, particle size of the crystalline phases and phase transformations. Islam, et al (2004) conclude that the best mechanical and dielectric properties can be achieved by high mullite and quartz content with low amount of the glassy phase and in absence of micro cracks. However, a high amount of SiO2 leads to a high amount of the glassy phase which is detrimental to the development of high dielectric strength. Adherents of the matrix reinforcement theory suggest that the composition of porcelain should be such that the batch should contain as little clay as conformable with the workability of the body, as little feldspar as conformable with the impermeability of the fired porcelain, and as much quartz of uniform grain size as possible (Mattyasovszky-zsolnay, 1957, Stathis et al, 2004). Indeed Mattyasovszky-zsolnay (1957) reported maximum strength with a body with quartz content of 39% while Stathis et al (2004) kept the filler content to 29%. Strength aside, the other limiting factor is the forming/shaping process adopted and the particle size of the starting powders. On the basis of the above, in this study the materials of focus were quartz, kaolin, ball clay and feldspar. The entire research was based on the conceptual framework indicated in Figure 1. 3.3 Source of the Raw Materials

All materials in the present study were sourced from deposits in Uganda. Kaolin, quartz and feldspar were from the Mutaka deposit, flint from Lido beach and ball clay from

5

Mukono. Another feldspar deposit in Lunya, Mukono district was also analysed. These are indicated on the Map of Uganda shown Figure 2.

Figure 1: Conceptual Framework

Mechanical and dielectric strength, shrinkage

Binding and particle agglomeration, porosity

Chemical analysis; Components of interest, Al2O3, SiO2, Fluxes

Mineralogical analysis (XRD, SEM, DTA TGA). Physical analyses

Raw Materials Ball clay, Kaolin,

feldspar, quartz/Flint

Mixture proportions

Material Preparations and mixes

Mixture properties, particle size, chemical composition

Forming process

Drying and Firing

Density, Porosity

Vitrification, sintering

Cause/Effect

Properties of Interest • Shrinkage • Porosity • Density • Water

absorption • Modulus of

Rupture • Dielectric

strength

6

Figure 2: Map of Uganda showing location of the minerals studied

1

3 2

1 Mutaka deposit 2 Lido beach

3 Mukono ball clay deposit 4 Lunya feldspar deposit

4

7

4. Experimental Techniques 4.1 Characterisation of Feldspars and Quartz Raw Materials A number of techniques were employed in characterising the raw materials. The details of each of these methods are given in paper 2 appended. Table 1 is an outline of the methods used. Table 1: Methods for characterisation of raw material sample powders Property Method Chemical composition Inductively coupled plasma-Atomic Emissions

Spectroscopy (ICP-AES) Phase constitution X-ray diffraction (XRD) analysis Texture and morphology Field Emission Gun-Scanning electron microscopy

(FEG-SEM) Density Pycnometer Weight change and phase transformation on heating TG-DTA Particle size BI-90 particle sizer. 4.2 Formulation of Porcelains Five sample porcelain insulator bodies S-1 to S-5 were formulated from Mutaka feldspar, Mutaka kaolin, Mukono ball clay and Lido beach flint in proportions indicated in Table 2. The raw materials were wet milled separately; ball clay and kaolin sieved through 45μm, flint through 25μm and feldspar through 53μm, respectively. The resultant materials were dried with the exception of ball clay which was kept in slip form after wet sieving. The amount of dry ball clay in the slip was estimated from Brogniart’s formula given in equation 1 (Norsker and Danisch, 1993). The mixture was wet milled for 3h to form a uniform mix, which was later made into a paste that was extruded through a vacuum pug mill into cylindrical specimens of 15mm diameter and 70mm length. The remaining materials in the pug mill after each formula was pulverised and used to make discs of 25mm diameter and about 3mm thickness by pressing at a pressure of 100MPa. The resulting samples were characterised by the methods given in Table 3. The details are appended in Paper 3.

)1..(1

1000 materialdryofxDensitymaterialdryofDensity

liquidlitreoneofgraminWeightgramsinweightDry ⎟⎟⎠

⎞⎜⎜⎝

⎛−−

=

Table 2: Composition of sample porcelains (wt %)

Sample S-1 S-2 S-3 S-4 S-5 Kaolin 35 25 30 30 30 Ball clay 15 25 20 15 15 Feldspar 25 30 20 30 25 Flint 25 20 30 25 30 % Total 100 100 100 100 100

8

Table 3: Methods of characterisation of porcelain bodies Property Method Shrinkage Measurement of dimensions Phase constitution X-ray diffraction (XRD) analysis on pulverised samples Texture and morphology

Field Emission Gun-Scanning electron microscopy (FEG-SEM)

Bulk density Measurement of dimensions and mass Water absorption Measurement of weights of samples before and after boiling and soaking in

water Bending strength Measurement of dimensions and breaking load in a 3 point bending load fixture Dielectric strength Measurement of voltage required to break insulation of a given thickness of a

specimen Particle size of materials

Sieving through standard sieves.

5. Summary of Results 5.1 Characterisation of Raw Materials

5.1.1 Chemical composition of the deposits Table 3 gives the chemical composition of the sample powder characteristics of the feldspar and quartz deposits. The results reveal Mutaka quartz and Lido beach sand as almost pure SiO2. The Mutaka quartz sample however, had a higher Fe2O3 content than the Lido beach sand. Consequently, the Lido beach sand was preferred in order to reduce gas formation during firing as a result of the Fe2O3 to Fe3O4 transformation. The feldspar samples reveal that the Lunya and Mutaka deposits have constituents in the ranges comparable to many deposits cited elsewhere (Norton, 1970). In comparison of the two deposits, although Lunya feldspar has a relatively higher flux content, its Fe2O3 is higher than that of Mutaka feldspar. As before, the deposit with lowest Fe2O3 content was preferred. Previous characterisation studies by Kirabira et al (2005) revealed the properties of Ugandan clay deposits of relevance for porcelains. The chemical composition of the deposits relevant for this study is given in the last two columns of Table 4. Table 4: Chemical Composition (Weight %)

Compound Lunya

Feldspar Mutaka

Feldspar Mutaka Quartz

Lido Beach Sand

Mutaka kaolin

Mukono ball clay

SiO2 65.7 62.9 101.0 100.0 48.8 67.2

Al2O3 18.3 22.5 0.193 0.127 36 18.2 CaO <0.1 <0.09 <0.09 <0.09 <0.09 0.306 Fe2O3 1.65 0.065 2.57 0.201 0.238 2.83 K2O 12.3 11.8 <0.06 <0.06 1.140 0.975 MgO 0.0436 <0.02 0.0251 <0.02 0.038 0.363 MnO 0.0219 <0.003 0.0199 0.0092 0.0277 0.0262 Na2O 1.84 0.409 <0.04 <0.05 0.0481 0.185 P2O5 0.0366 0.0874 0.0177 0.0171 0.0094 0.049 TiO2 0.0053 0.0036 0.0051 0.1720 0.0041 1.38 LOI1 0.1 3.1 -0.4 0.4 12.6 8.1

1 LOI is loss on ignition at 1000ºC

9

5.1.2 Microstructure of the minerals The microstructures in Figure 3 show Lunya feldspar as densely-packed layered sheets of aggregate, while Mutaka feldspar consists of rather loosely held sheets of aggregate, suggesting that the two deposits are at different levels of geological transformation with the Mutaka feldspar in the more advanced stages of transformation to kaolinite. The SEM image of the Mutaka feldspar closely resembles the morphology of Mutaka kaolin presented by Kirabira et al (2005). The microstructure of Mutaka quartz reveals a non homogeneous structure consisting of dense and porous regions while that of the Lido beach sand sample reveals a more homogeneous, porous structure. Features of the microstructure of Mutaka quartz are similar to those reported by Wright (2004) of freshly fractured quartz fragment. The difference in these two microstructures thus reflects the level of abrasion as a result of the weathering and transportation history of the beach sand.

Figure 3: SEM of (a) Lunya feldspar, (b) Mutaka feldspar, (c) Mutaka quartz, (d) Lido Beach flint 5.1.3 Thermal analysis Results of thermal analysis of the samples revealed distinctively different DTA signals for the two feldspar samples. While Lunya feldspar did not show any distinct peaks on heating, Mutaka feldspar had a strong endothermic peak at 528°C and an exothermic peak at 1000°C. This difference can be attributed to the significantly differing ratios of Al2O3: SiO2 in the two samples. The high ratio accompanied with the high loss on ignition suggests the presence of a considerable amount of crystalline water in Mutaka feldspar, which upon heating led to the formation of metakaolin at about 530°C followed

ba

c d

10

by a spinel phase formation at 1000°C due to the decomposition of metakaolin. These particular peak temperatures are typical of kaolinite thermal characteristics (Carty and Senapati, 1998; Kirabira et al, 2005). This feature further re-affirmed the kaolinitic tendency of Mutaka feldspar shown in the SEM analysis. The endothermic peaks observed in the silica samples at 576°C in Figure 4 are typical for the α-β quartz inversion. Other investigators have reported this transformation to occur at 573°C (Carty and Senapati, 1998; Iqbal and Lee, 2000; Hand et al, 1998).

4

3

2

1

-22

-12

-2

8

18

28

38

48

58

68

78

0 100 200 300 400 500 600 700 800 900 1000 1100 1200Temperature, Celsius

Hea

t flo

w in

mic

ro v

olts

1- Lido beach sand2- Mutaka quartz3- Lunya feldspar4- Mutaka feldspar

Figure 4: DTA signals of heat flow during heating of the samples. For clarity the curves have been displaced as follows: Lunya feldspar +10, Mutaka quartz +20, and Lido beach sand +30 5.1.4 Gravimetry Mutaka feldspar, in contrast with Lunya feldspar revealed a unique weight change profile upon heating (Figure 5) with a steep weight drop in the temperature range of 450°C to 600°C. The weight drop profile exhibited by Mutaka feldspar in the temperature range of 450°C to 600°C is identical to that exhibited during the dehydration of the hydroxyl groups in kaolinite reported by Carty and Senapati (1998), leading to formation of metakaolin (Al2O3·2SiO2). The resultant weight loss agrees well with the results reported for the loss on ignition in Table 3. The TGA, DTA, and SEM analyses all point to kaolinitic character in Mutaka feldspar.

Lunya feldspar

Mutaka quartz

Lido beach sand

Mutaka feldspar

-4.0

-3.5

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

0 100 200 300 400 500 600 700 800 900 1000 1100 1200Temperature, Celsius

Mas

s cha

nge,

[%]

Figure 5: Mass change on heating of samples. For clarity the curves have been displaced as follows: Lunya feldspar +0.3, Mutaka quartz +0.5, and Lido beach sand +1.5.

11

5.1.5 Mineralogy XRD studies (Figures 6 and 7) show that Lunya feldspar contains microcline, albite and quartz whereas the Mutaka feldspar is predominantly composed of ordered microcline. Mutaka quartz and the Lido beach sand are both predominantly composed of quartz mineral.

12 14 16 18 20 22 24 26 28 30 32 340

5000

10000

15000

20000

25000

30000

35000

40000

45000

50000

Mutaka Feldspar

Lunya Feldspar

M Microcline ordered, KAlSi3O8

A Albite, NaAlSi3O8

Q Quartz

A

A

A

MMM

M Q

M

M

M M

MMM

M

MMM

M

M MM

M

MM

M M

M

MMM

M

Cou

nts

2 Theta [0] Figure 6: XRD of Lunya feldspar and Mutaka feldspar. Lunya feldspar was offset by 15000 counts on the vertical axis

12 16 20 24 28 32 36 40 44 48 52 56 600

5000

10000

15000

20000

25000

30000

Q Quartz

Q

Q

Q

QQ

Q QQ

Q QQ Q

Q

Q

Q

Q

Q Q

Lido beach sand

Mutaka Quartz

Cou

nts

2 Theta [0] Figure 7: XRD of Mutaka quartz and Lido beach sand. Lido beach sand was offset by 7500 counts on the vertical axis. 5.2 Formulated Porcelains

5.2.1 Properties after Firing Upon drying and firing, the formulated samples listed in Table 2 showed the trends in Figures 8-12. The trends suggest that a dense body results at 1250°C. Firing beyond this temperature results in progressive deterioration of the properties of the samples. Quite favourable, samples exhibited highest bulk density at the firing temperatures where they achieved the highest bending strength, dielectric strength and least water absorption. Thus, optima for the different properties coincide at a firing temperature of 1250°C. 5.2.2 Formability It was also found during extrusion of the samples that samples with quartz content of 30% were difficult to work with. However, at 25% of quartz, the plasticity of the bodies was good enough.

12

5.2.3 Microstructure and phase analyses of fired samples Microscopic studies revealed that at temperatures above 1250°C, the samples became more porous (Figure 13) due to bloating. This assertion is complemented by the decrease in bulk density, strength and shrinkage. Detailed SEM indicates that samples exhibit different shapes of the mullite needles upon firing depending on the amount of glass forming constituent in the composition (Figure 14). The different shapes of mullite needles influenced the strength. Samples with acicular needles showed low strength values compared to rounded needles. It was also noted that at higher temperatures of firing, the glass content increased at the expense of quartz in the body as shown in Figure 15. An optimum temperature and composition at which considerable amounts of quartz and mullite crystals existed in the glass phase maximised the physical, mechanical and dielectric properties. XRD of all samples fired at 1250°C in Figure 16, shows that the samples are practically identical with respect to constituent phases and amounts. Thus, the major differences are found in the morphology of the microstructure as indicated in Figure 14.

0

2

4

6

8

10

12

14

1175 1200 1225 1250 1275 1300 1325 1350 1375

Firing temperature (oC)

Per

cent

shr

inka

ge

S-1S-2S-3S-4S-5

Figure 8: Shrinkage of samples

1.4

1.6

1.8

2

2.2

2.4

2.6

1175 1200 1225 1250 1275 1300 1325 1350 1375

Firing temperature (oC)

Bul

k de

nsity

(g/c

m3 )

S-1S-2S-3S-4S-5

Figure 9: Bulk density of samples

13

00.5

11.5

22.5

33.5

44.5

5

1175 1200 1225 1250 1275 1300 1325 1350 1375

Firing temperature (oC)

Wat

er a

bsor

ptio

n (w

t %)

S-1S-2S-3S-4S-5

Figure 10: Water absorption

20

30

40

50

60

70

80

1175 1200 1225 1250 1275 1300 1325 1350 1375

Firing temperature (oC)

Ben

ding

stre

ngth

(MP

a)

S-1S-2S-3S-4S-5

Figure 11: Modulus of rupture (MOR)

4

6

8

10

12

14

16

18

20

22

1175 1200 1225 1250 1275 1300 1325 1350 1375

Firing temperature (oC)

Die

lect

ric S

treng

th (k

V/m

m) S-1

S-2S-3S-4S-5

Figure 12: Dielectric strength of samples

14

Figure 13: S-4 fired at 1200 and 1350°C respectively. Polished and Etched in 40% HF acid for 25s

1200°C

1350°C

15

S-1 1200°C

S-2 1250°C

S-3 1300°C

S-5 (a) Samples fired at 1250°C

1350°C (b) S-4 fired at different temperatures

Figure 14: Microstructures of selected samples fired at different temperatures, polished and etched in 40% HF acid for 25s.

16

0 10 20 30 40 50 60 70 80 90 100

3

4

5

6

7

8

Q

M Mullite Q Quartz

QMMMM

MMMMQQ

QQQ

QQQ

Q

Q

Q

13500C

13000C

12500C

12000C

Log[

Coun

ts]

2Theta (degrees) Figure 15: Powder diffractograms of S-4 fired at different temperatures. For clarity, the three upper curves are shifted vertically by equal proportions from each other.

0 10 20 30 40 50 60 70 80 90 100

3

4

5

6

7

8 M Mullite Q Quartz

MQQMMMM

MMMMQQ

QQQQQ

QM

Q

S-2

S-5

S-4

S-3

S-1

Log[

Coun

ts]

2 Theta (degrees) Figure 16: Powder diffractograms of all samples fired at 1250°C. For clarity, the four upper curves are shifted vertically by equal proportions from each other. 6. Conclusion The literature presented shows that triaxial porcelains are still very useful materials for insulation. The use of quartz in the bodies can be carried out with careful control of the production parameters. Results from characterisation studies suggest that Lunya feldspar consists of microcline and albite minerals. This deposit also has a relatively high amount of Fe2O3. Its use in the production of porcelain insulators requires the removal of Fe2O3. Mutaka feldspar is purely microcline and is suitable for use with very minimal beneficiation. The analyses have revealed Mutaka feldspar to show strong kaolinitic characteristics. Both Mutaka and Lido beach quartz deposits have the qualities desirable for the production of various whiteware products in their raw form. The deposits studied together with the clay deposits studied by Kirabira et al (2005) revealed them to be of good prospects for their exploitation in the production of electric porcelains.

17

From the formulated samples of porcelain bodies, it was found that for a plastic forming process, use of quartz content of 30% or above cannot be carried out effectively yet it is one of the recommendations for improvement of strength of porcelains by various researchers. High porosity developed in samples at firing temperatures above 1250°C is responsible for deterioration of both bending and dielectric strengths. Firing above the sintering temperature also reduces the strength due to a reduction in quartz and increase in glass phase. The shape of mullite needles formed in the microstructure influences the strength of porcelains. Acicular needles are associated with a low-viscosity body at high temperatures resulting from high flux content and leading to low strength. An optimised strength of porcelain can be achieved from a composition which gives needle-like mullite crystals as opposed to acicular crystals and with sufficiently high amounts of quartz and as little glass as is necessary to hold the crystalline phases together. A porcelain composition of 30% Mutaka kaolin, 15% Munono ball clay, 30% Mutaka feldspar and 25% Lido beach quartz yielded a body with highest mechanical and dielectric strength at a firing temperature of 1250°C. 7. Proposals for Further Investigations From the foregoing, the following work is suggested for further investigation; • Study of the mechanism of strengthening of porcelains by glazing • Determination of the optimum conditions of glazing in terms of thickness of glaze,

and firing temperature schedule for optimum mechanical properties (strength, toughness and hardness).

• The properties of a full-size sample made and subjected to typical product specification tests.

Acknowledgments With much pleasure and gratitude, I extend my sincere thanks to all those that helped me in all the activities leading to this thesis, both in their individual and institutional capacities. First and foremost, I acknowledge the financial support from the Sida/SAREC-Makerere University Collaborative Research Programme. My gratitude goes to the coordinators of the above program both at Makerere University, Uganda and at the Royal Institute of Technology (KTH), Sweden. Special thanks also go to the Head of Department of Mechanical Engineering at Makerere University for the support. I also thank the staff of the Department of Ceramics, Uganda Industrial Research Institute for their cooperation. My heart felt gratitude goes to my supervisors, Prof. Stefan Jonsson of KTH and Eng. Dr. Joseph K. Byaruhanga of Makerere University for the invaluable guidance and support throughout this work. It is also my pleasure to thank Mr. Hans Bergqvist of KTH for help with FEG-SEM, Mr. Chimtawi of ICIPE- Nairobi, Kenya for the SEM, John Gitta, Fredrik Carnsjö and all the

18

colleagues both, in the division at KTH and those on the Sida/SAREC Programme at Makerere University for their support. Special thanks to Rosette for her patience. My appreciation also goes to my parents and siblings for the continued encouragement. Thank you all! References 1. Bragança, S.R. and Bergmann, C.P. (2003) “A View of Whitewares Mechanical

strength and microstructure” Ceramics International 29, pp 801-806. 2. Bragança, S.R. and Bergmann, C.P. (2004) “Traditional and glass powder

porcelain: Technical and Microstructure analysis” Journal of the European Ceramic Society 24, pp 2383-2388.

3. Buchanan, R.C. (1991) “Ceramic Materials for Electronics” (R.C. Buchanan, Ed.) Chap. 1. Dekker, New York.

4. Carty, W.M and Senapati, U. (1998) “Porcelain-Raw Materials, Processing, Phase Evolution and Mechanical Behaviour” J. Am. Ceram. Soc. 81 pp 3-20.

5. Dana, K., Das, S. and Das, K. S. (2004) “Effect of substitution of fly ash for quartz in triaxial kaolin-quartz-feldspar system” Journal of the European Ceramic Society 24, pp 3169-3175.

6. Dana, K., Dey, J. and Das, K. S. (2005) “Synergistic effect of fly ash and blast furnace slag on the mechanical strength of traditional porcelain tiles” Journal of the European Ceramic Society 31, pp 147-152.

7. Das, K.S. and Dana, K. (2003) “Differences in densification behaviour of K- and Na-Feldspar-containing porcelain bodies” Thermochimica Acta 406, pp 199-206.

8. Ece, O. I. and Nakagawa, Z. (2002) “Bending Strength of Porcelains” Ceramics International 28, pp 131-140.

9. Engelthaler, A. Z. and Engena, M. Y (January, 1972) “Ceramic raw materials in Uganda, preliminary report” African Ceramics Company Ltd, Kampala, Uganda.

10. Esposito, L., Salem, A., Tucci, A., Gua1tieri, A.,and Jazayeri (2005). “The use of nepheline-syenite in a body mix for porcelain stoneware tiles” Ceramics International 31, pp 233-240.

11. Hand, R.J., Stevens, S.J. & Sharp, J.H. (1998). “Characterisation of fired silicas”, Thermochimica Acta, 318, 115-123.

12. Iqbal, Y and Lee, W.E (2000) “Microstructural evolution in Triaxial porcelain” J. Am. Ceram. Soc. 83 [12] pp 3121-27.

13. Islam, R. A., Chan, Y.C, and Islam M. F. (2004) “Structure-property relationship in high-tension ceramic insulator fired at high temperature” Materials Science and Engineering B106, pp 132-140.

14. Kabagambe-Kaliisa, F.A (1983) “Industrial mineral deposits of Uganda” Report of the Geological Surveys and Mines Department, Entebbe Uganda.

15. Kingery, W.D (1967) “Introduction to Ceramics” John Wiley and Sons Inc., New York.

16. Kirabira, J.B, Jonsson S. and Byaruhanga, J.K (2005). “Powder Characterization of High Temperature Ceramic raw materials in the Lake Victoria Region” Silicate Industriels 70 [9-10], pp 127-134.

19

17. Kobayashi, Y., Ohira, O., Ohashi, Y. and Kato, E. (1987). “Strength and Weibull distribution of alumina strengthened whiteware bodies” Journal of the Ceramic Society of Japan, International Edition 95, pp 837-841.

18. Mattyasovszky-zsolnay, L. (1957) “Mechanical Strength of porcelain” J. Am. Ceram. Soc. 40 [9] pp 299-306.

19. Norsker, H. and Danisch, J. (1993) “Glazes: for the self- reliant potter” Friedr. Vieweg & Sohn Verlagsgesellschaft mbH, Braunschweig. Germany.

20. Norton, F.H. (1970) “Fine Ceramics, Technology and Applications.” McGraw-Hill Book Co. New York.

21. Nyakairu, G.W.A. and Kaahwa, Y. (1998) “Phase Transformation in Local Clays” The Am. Soc. Bull, 77(6).

22. Nyakairu, G.W.A and Koeberl, C. (2001) “Mineralogical and chemical composition and distribution of rare earth elements in clay-rich sediments from Central Uganda” Geochemical Journal, 35, pp 13-28.

23. Nyakairu, G.W.A., Koeberl, C. and Kurzweil, H. (2001) “The Buwambo kaolin deposit in Central Uganda: Mineralogical Composition” Geothermal Journal. 35. pp 245-256.

24. Prasad, C.S, Maiti, K.N. and Venugopal R. (2001) “Effect of rice husk ash in whiteware compositions” Ceramics International 27 pp 629-635.

25. Prasad, C.S, Maiti, K.N. and Venugopal R. (2002) “Effect of silica fume addition on the properties of whiteware compositions” Ceramics International 28 pp 9-15.

26. Prasad, C.S, Maiti, K.N. and Venugopal R. (2003) “Effect of substitution of quartz by Rice husk ash and silica fume on the properties of whiteware compositions” Ceramics International 29, pp 907-914.

27. Schroeder, E. J. (1978) “Inexpensive high strength electrical porcelain” Am. Ceram. Soc. Bull. 57, 526.

28. Stathis, G., Ekonomakou, A., Stournaras, C.J, and Ftikos, C. (2004) “Effect of firing conditions, filler grain size and quartz content on bending strength and physical properties of sanitary ware porcelain” Journal of the European Ceramic Society 24, pp 2357-2366.

29. Tai, Weon-Pil. Kimura, K. and Jinnai K. (2002) “A new approach to anorthite porcelain bodies using non-plastic raw materials” Journal of the European Ceramic Society 22, pp 463-470.

30. Thurnauer, H. (1954) “Dielectric Materials and Applications” (A.R.V. Hippel, Ed.) Ceramics, Chapman & Hall, London

31. Wright S.J. (2004) “The role of weathering in the production of Quartz silt”, Int. Workshop HWK Delmenhorst. 15-18 April (2004).

20

Appended Papers 1. “State of the art paper on development of electric porcelain insulators from Ugandan

raw materials” Peter Wiberforce Olupot, Stefan Jonsson, Joseph Kadoma Byaruhanga. Unpublished report

2. “Characterization of Feldspar and Quartz Raw Materials in Uganda for Manufacture

of Electrical Porcelains” Peter W. Olupot, Stefan Jonsson, Joseph K. Byaruhanga. J. Aust. Ceram. Soc. 41[1] (2006) 29-35.

3. “Optimization of composition and firing temperature for high-strength electric

porcelains from Ugandan materials” Peter W. Olupot, Stefan Jonsson, Joseph K. Byaruhanga. Manuscript ready for submission