the properties of feldspars and their use in whitewares, · semi-vitreous ware 15 15. hotel china...

II LLINOI SUNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN

PRODUCTION NOTE

University of Illinois atUrbana-Champaign Library

Large-scale Digitization Project, 2007.

The Properties of Feldspars and Their

Use in Whitewares

by

Joseph C. Kyonka

FORMERLY RESEARCH ASSOCIATE IN CERAMIC ENGINEERING

Ralph L. Cook

PROFESSOR OF CERAMIC ENGINEERING

ENGINEERING EXPERIMENT STATION BULLETIN NO. 422

UNIVERSITY

5050--1-54--53757 ,. eRss0,

CONTENTS

I. INTRODUCTION 51. Definition 5

2. Mode of Occurrence 53. Geographic Occurrence 54. Commercial Use 55. Acknowledgments 5

II. FUNDAMENTAL PROPERTIES 66. Chemical Composition 67. Mineralogical Composition 78. Structure 7

9. Thermal Properties 910. Solubility in Water 12

III. USE IN WHITEWARE COMPOSITIONS 13

11. Physico-Chemical Behavior 1312. Purpose and Scope of Experimental Investigation 14

13. Properties of Feldspars Used 14

14. Semi-Vitreous Ware 15

15. Hotel China 20

16. Electrical Porcelain 21

17. Sanitary Ware 24

18. Floor Tile 26

IV. SUMMARY OF RESULTS 28

APPENDIX- DETAILED TEST PROCEDURE 29

19. Raw Materials Used in Experimental Bodies 29

20. Details of Body Preparation 29

21. Specimen Formation 2922. Test Procedures 31

BIBLIOGRAPHY 33

FIGURES

1. Spacing of (201) Planes in a Series of Alkali FeldsparsCrystallized at 900 deg C and 300 kg/cm2 Pressure of Water 7

2. Relation Between Refractive Index and Anorthite Content ofNatural Plagioclase 8

3. The Alkali Feldspar Join, Showing the ProbableSub-Solidus Relations 9

4. Equilibrium Diagram for the Plagioclase Feldspars 95. Portions of the Equilibrium Diagrams, Soda-Alumina-Silica

and Potash-Alumina-Silica 116. Fired Properties of Semi-Vitreous Bodies 177. Properties of Semi-Vitreous Bodies at Maturity

(10 percent Porosity) 188. Linear Thermal Expansions of Series I and II Semi-Vitreous

Bodies Fired to Cone 9 209. Linear Thermal Expansions of Semi-Vitreous Bodies Containing

Tremolitic Talc and Fired to Cone 9 2010. Fired Properties of Hotel China Bodies 2211. Fired Properties of Electrical Porcelain 2312. Fired Properties of Sanitary Ware Bodies 2513. Fired Properties of Floor Tile Bodies 27

TABLES

1. Theoretical Compositions of Pure Feldspars 62. Range of Compositions of Commercial Feldspars 63. Properties of Feldspars Studied 144. Compositions of Semi-Vitreous Bodies 155. Comparison of Fired Properties of Semi-Vitreous Bodies as

Obtained from Laboratory and Commercial Firings (2240 deg F) 166. Cristobalite Content of Fired Semi-Vitreous Bodies 197. Thermal Shock Resistance of Semi-Vitreous Bodies from

480 deg F to Room Temperature (70 ± 2 deg F) 208. Compositions of Hotel China Bodies 219. Compositions of Electrical Porcelain Bodies 23

10. Dielectric Properties of Electrical Porcelain Bodies 2311. Compositions of Sanitary Ware Bodies 2412. Casting Characteristics of Sanitary Ware Bodies 2413. Composition of Floor Tile Bodies 2614. Chemical Analyses of Raw Materials Used in

Experimental Bodies

I. INTRODUCTION1. Definition

Feldspars are the most common constituents incrystalline rocks and make up about 60 percent ofthe earth's crust. They may be technically definedas aluminosilicates of sodium, potassium, calciumand barium; most commonly, the feldspars are con-sidered as solid solutions of three limiting com-pounds, NaAlSi 8Os, KAISi3O,, and CaA12Si208,which are respectively known as soda feldspar,potash feldspar and lime feldspar. Natural depositsof feldspar are generally solid solutions of eitherthe soda and potash feldspars or the soda and limefeldspars.

2. Mode of OccurrenceAs the main constituent of the igneous rocks

making up the earth's surface, feldspars are abun-dantly distributed; the chief commercial sourcesare found in pegmatite dikes associated with otherpegmatite minerals such as quartz and the variousmicas as well as minor amounts of tourmaline,beryl, garnet, spodumene, pyrite and magnesite.Although pegmatite deposits are widely distributedgeographically, feldspars sufficiently free from im-purities and occurring in large mineable quantitiesare not commonly found. For many years it hasbeen the practice to separate the feldspar mineralsfrom the associated impurities by a method of handselection, but recent technological developmentshave led to flotation methods for the separation ofquartz and mica. The flotation process has provedto be an effective method of producing large quan-tities of feldspar relatively free of undesirableimpurities.

3. Geographic OccurrenceCommercial sources of feldspar are found in all

the states of the Appalachian Region from Georgiato New York and in the New England States. SouthDakota, Colorado, Minnesota, Arizona, California,Nevada, New Mexico and Texas also contain de-posits. In recent years, North Carolina has been thechief feldspar-producing state, followed by Colo-

rado, Virginia and South Dakota. In Canada, peg-matites containing commercial feldspar are foundin the provinces from Nova Scotia and Labrador toManitoba, and through the Rocky Mountain area.

In Europe, Sweden and Norway are the mostimportant sources of feldspar; the Norwegian de-posits are noted for their high purity. Great Britaindoes not have any appreciable amounts of purefeldspar although British potteries use a feldspathicmaterial known as "Cornish Stone" which is a typeof decomposed granite containing feldspar andquartz with varying amounts of kaolin, muscovite,fluorspar, and topaz. Other feldspar deposits inEurope are found in Czechoslovakia, Germany,France, Italy, Rumania, Russia, and Finland. Else-where in the world, pegmatite deposits are locatedin China, India, Japan, Australia, New Zealand,Egypt, South Africa, and Argentina.

4. Commercial UseThe chief commercial value of the feldspars is in

their use by the ceramic industries for the manu-facture of glass, whiteware, and porcelain enamelproducts. Their fundamental characteristics andtheir behavior in the presence of other constituentsare important considerations in the formulation ofceramic compositions. The following discussions willattempt to describe the basic features of feldsparsand to show their behavior and influence when usedas fluxing agents in whiteware bodies.

5. Acknowledgments

Funds for this project were furnished under acooperative arrangement with the Engineering Ex-periment Station by the Consolidated Feldspar Cor-poration, which became the Consolidated FeldsparDepartment of International Minerals and Chem-ical Company. This work has been carried outunder the general administrative direction of DeanW. L. Everitt, Director of the Engineering Experi-ment Station and Professor A. I. Andrews, Headof the Department of Ceramic Engineering.

II. FUNDAMENTAL PROPERTIES

6. Chemical CompositionIn a consideration of the properties of feldspars,

the chemical composition is of primary importance.The chemical analysis is an important key to thebehavior of the material under various conditionsand is the criterion most often used in evaluatingthe use of a particular feldspar in a whitewarebody. The theoretical chemical compositions for thethree major feldspars in their pure state are shownin Table 1.

TaTheoretical Compositi

K 20

16.9Potash Feldspar

K 20 Ah03-6SiO0Soda Feldspar

Na2O Al20 3 6Si0 2Lime Feldspar

CaO Al03 2Si0 2* Weight percent.

ble Iions of Pure Feldspars*

NasO CaO AlsO1

18.3

11.8 19.4

20.2 36.6

Since deposits of pure feldspar are rare or non-existent, the naturally occurring minerals will havecompositions which deviate somewhat from thoseshown in the table. The range of chemical compo-sitions of typical commercial feldspars in theUnited States is shown in Table 2. The most sig-nificant variations in composition are found to be inthe contents of Si0 2, Al 203, NaO2 and KO2 ; there-fore, it may be expected that these oxides will con-tribute to the variability of behavior of differentfeldspars. Although the amounts of CaO, MgO andFeO,3 are small in relation to the other constituents,their effects on the thermal behavior are quitepronounced.

The chemical compositions of feldspars andother body materials have been used to estimatethe quantities of ceramic bond in fired bodies andto differentiate the various phase constituentsformed in the firing process. This method proposesto refer the chemical composition of the body toappropriate equilibrium diagrams and to relate thecalculated phases to the fired physical propertiesof the body.

The conventional wet methods of chemical anal-ysis are still widely employed for the determina-

tions of feldspar composition; however, recent in-strumental procedures have been introduced for theanalyses of silicates and have proven to be morerapid than the older methods and to give results ofhighly reproducible accuracy. Two of the instru-mental methods particularly applicable for theanalysis of feldspars are colorimetry and flamephotometry. Colorimetric methods may be em-ployed for the complete feldspar analysis with ex-cellent reproducibility. 1 * Flame photometry hasbeen used mostly for the determination of the alkaliand alkaline earth metals, 23 '4, 5) although com-plete silicate analysis methods are proposed.

Several proposals have been made for the clas-sification of commercial feldspars on the basis oftheir chemical composition, physical properties andfineness of grind. An early basis for their classifica-tion and standardization was the alkali content andfusion temperature. ( 6 ) Another system for commer-cial standards of quality was introduced by theBureau of Standards to cover the specifications ofground feldspar used in ceramic processes. ( 7 ) Thismethod of classification was based on the particlesize, chemical composition and end use of thematerial.

Table 2Range of Compositions of Commercial Feldspars*

Oxide Range in percent by weightNasO 2.0 - 9.0K0O 0.5 -13.5CaO 0.1 - 2.5MgO tr - 0.5FeOs3 0.02- 0.4A1203s 15 - 21SiOs 65 - 75

* For use in whiteware compositions.

Although the proposed methods of feldspar clas-sification were steps forward in the standardizationof materials, their utilization was not fully realized.This lack of realization was no doubt largely dueto the advent of feldspar blending. The introductionof blending and chemical control resulted in pro-moting standardization of composition and enabledthe producers to supply large quantities of uniformmaterial.

SParenthesized superscripts refer to correspondingly numbered entriesin the Bibliography.

Bul.422. PROPERTIES OF FELDSPARS AND THEIR USE IN WHITEWARES

While feldspar blending is chiefly based on theuniform control of chemical composition, the modeof occurrence and the inherent nature of each feld-spar are also important considerations in judgingthe feasibility of blending. (8)

7. Mineralogical CompositionThe mineralogical compositions of feldspars may

often be determined by calculation from the chem-ical analysis to an accuracy suitable for certainapplications such as those concerned with the freesilica content. Since commercial feldspars consistessentially of microcline, albite and anorthite mix-tures with some associated quartz, muscovite andkaolinite, the mineral composition may be calcu-lated on a 6-component basis. This method hasbeen simplified to such an extent that mathematicalformulae are available for direct computation ofthe mineral content. Formulae for the calculationare shown below. (9)

Percent Albite =Percent Microcline =

Percent Anorthite =

Percent Muscovite =

Percent Quartz =

where:A= percent H20; B

8.458E15.442A - 5.459C +9.923D + 8.977E +

11.815F4.960D

7.813C - 22.097A -14.201D -12.847E - 8.455F

B + 1.178C -6.666A - 4.284D -7.751E - 5.101F

=percent SiO 2 ; C=percent A120l ; D=percent CaO; E=per-cent Na20; F = percent K20, obtained fromchemical analysis.

The mineralogical compositions of feldspars maybe determined by direct methods such as opticaland X-ray analyses, but the complex nature offeldspathic crystallization makes a review of theoptical and X-ray methods beyond the scope ofthis discussion. Nevertheless, recent studies havegreatly simplified these methods and are note-worthy of mention. In the soda-potash feldsparseries a linear relationship has been found betweenthe spacing of the (201) planes and the weightcomposition of the end members in solid solution. (10 )

This relationship is illustrated by Fig. 1. Therelations between the refractive indices and theanorthite content of plagioclases have also been

simplified and a mathematical expression has beenproposed which describes the relation in the rangesof 0-30 percent and 64-100 percent anorthite. Thisis shown in Fig. 2. (11)

8. StructureAs the most important group of rock minerals,

the geology and mineralogy of feldspars have longbeen the subjects of extensive study. Although thereis probably more information regarding feldsparsthan any other group of minerals, they are not

IC~

Potash & d4 6 80 SodaFeldspor Weight Percent Feldspar

Fig. 1. Spacing of (201) Planes in a Series of Alkali Feldspars

Crystallized at 9000 C and 300 kg/cm2 Pressure of Water

(Reprinted from The Journal of Geology, Vol. 58, p. 493, 1950)

completely understood. Optical methods have beendeveloped for their characterization, but the rela-tionships between optical properties and chemicalcomposition are yet to be definitely determined.Some of the optical properties are understood onthe basis of crystal structure, and certain funda-mental features have been established, even thougha detailed knowledge of the complete structure isyet unknown.

All feldspars consist of a three-dimensional net-work of [Si0 4 ] and [A10 4] tetrahedra in which allthe tetrahedra share their oxygen atoms with theirneighbors. (12) The fundamental units of [Si04] and[AlO] are linked together in a four-ring frameworkconsisting of tetragons and collapsed octagons asviewed along the crystallographic a-axis. Positivelycharged ions of sodium, potassium or calcium aresituated in the octagonal interstices of the nega-tively charged framework. (l3) The fundamentalstructure of these tetrahedra is elastic to somedegree and can adjust itself to the sizes of thecations in the octagonal openings. The crystalsymmetry of the feldspars is dependent upon thesize of these cations; relatively large cations such

ILLINOIS ENGINEERING EXPERIMENT STATION

as K + give a symmetry which is monoclinic ornearly monoclinic while the small cations, such asNa+ and Ca +, cause a slight distortion of the struc-ture and triclinic symmetry results.( 14

Commercial feldspars are considered on thebasis of a three-component system, the endmembers of which are KAISiOs,, NaAlSi3Os andCaAl 2SiO08, or potash, soda and lime feldspars,respectively. The extent of solid solution betweenthe components and the manner in which the solidsolutions are affected by the conditions which pre-vail during and after their formation remains par-tially unsolved. The potash and soda feldspars andtheir solid solutions are grouped together under thedesignation of alkali feldspars to set them off fromthe soda-lime feldspar solid solutions which areknown as the plagioclases. No solid solution seriesexists between the potash and lime members sincetheir mutual miscibility is practically zero at alltemperatures. (3)

The feldspars may be mineralogically classifiedon the basis of their crystal symmetry as follows:

A. Monoclinic or nearly (pseudo-) mono-clinic-

Orthoclase KAlSi 3OsMicrocline KAlSiOs8Soda Orthoclase (KNa) AlSisOs

B. Triclinic-Anorthoclase (NaK) AlSi3OsAlbite NaAlSi 3OsAnorthite CaAl2Si20s

The structural relationships between orthoclaseand microcline are not completely known. Theexistence of dead bonds between some 0= and K+

ions in orthoclase has been determined so that nowit is generally believed that the difference betweenmicrocline and orthoclase is one of difference inatomic arrangement;( 15 it is assumed that in ortho-clase the aluminum and silicon are randomly dis-tributed, while in microcline these elements areconsidered to be in a particular set of latticepositions.

A potash feldspar approaching the compositionof KAlSi 3Os is rarely found in commercial deposits;the orthoclase and microcline are always found tocontain some soda feldspar. When relatively largeamounts of soda feldspar are present, thesefeldspars are referred to as soda orthoclase orsoda microcline. The molecules of KAlSisOs and

NaAlSiOs form a continuous series of solid solu-tions at high temperatures, but at temperaturesbelow 600 deg C there is a gap in the isomorphicseries. At these lower temperatures, the solid solu-tions between potash and soda feldspars are meta-stable and under conditions of slow cooling show

56

40 60Anorthite (Weight Percent)

80 /00

Fig. 2. Relation Between Refractive Index and AnorthiteContent of Natural Plagioclase


alterations into an oriented growth of sub-parallellamellae which are alternately rich in soda andpotash feldspar. Such intergrowths are called per-thites or antiperthites. In the perthites, potashfeldspar is the more abundant mineral, with sodafeldspar occurring as uniformly-oriented films,veins or patches. In the antiperthites the sodafeldspar is the more abundant mineral with potashfeldspar interspersed.

Upon heating the perthites at 1000 deg C forseveral hundred hours, a homogeneous material willresult. The mineral nature of feldspar is thereforedependent upon the temperatures of the magmafrom which the feldspar crystallized. Data havebeen obtained on the crystallization of one andtwo feldspar fields in the soda-potash feldsparsystem. (10) Figure 3 illustrates the conditions underwhich soda-potash feldspars will crystallize as asingle or a mixed feldspar. Any point on the curvedividing the two-feldspar field from the one-feldspar field represents the minimum temperatureat which a feldspar of that composition will remainin stable equilibrium; if equilibrium is maintainedbelow that temperature, unmixing will occur.

The soda and lime feldspar series has long beenregarded as the example of an ideal isomorphic

I I I IFor O<An <30, An=25473/Y-39/9.27r-, 64Z*As/011] A =190f7 209On CI

7

4

or n

n

-

'''/.,

> ̂ ·


QI7'

KA/Si,0, 20 40 60 80 NaA/Si 0,Potash Weight Percent Soda

Feldspar Feldspar

Fig. 3. The Alkali Feldspar Join, Showing theProbable Sub-Solidus Relations


series. This system, known as the plagioclase series,has been customarily designated as mixtures ofdifferent albite and anorthite ratios as follows:

Mineral Name Molecular Ratio

Percent PercentAlbite Anorthite

Albite 90-100 0-10Oligoclase 70-90 10-30Andesine 50-70 30-50Labradorite 30-50 50-70Bytownite 10-30 70-90Anorthite 0-10 90-100

Plagioclases crystallized at high temperaturesexhibit the perfect solid solution characteristicsattributed to this system; however, there is evidencethat in natural plagioclases there is a considerablemiscibility gap between 30 and 70 mole percentanorthite. This gap is apparently affected by analbite inversion at about 700 deg C as shown inFig. 4.' 16 ) Thus, although all appearances would

I clu ur realpoar

Fig. 4. Equilibrium Diagram for the

Plagioclase Feldspars


indicate a perfect isomorphism in the plagioclases,and unmixing may take place during the coolingof the magmatic crystals and result in a mixtureof highly ordered end members at room tem-peratures.

9. Thermal PropertiesThe melting characteristics of the soda, potash

and lime feldspars have been determined for thepure materials and for their mixtures. It should benoted that the melting temperatures of feldsparsare extremely difficult to obtain since the meltingphenomena are very sluggish. The melting beginsat the surfaces of crystals and proceeds so slowlythat much of the crystal can exist in the presenceof the melt for long periods of time even thoughthe temperature is somewhat above that of themelting point. The melting point is considered asthe temperature at which the crystal and the meltmay exist in equilibrium and is determined bylocating the temperature above which crystals show

_=


melting tendencies and below which the crystalstend to grow. The exact temperature of thisequilibrium is not easily determined and it has beennecessary to express feldspar melting points interms of temperature ranges.

Soda feldspar has been found to melt con-gruently to a very viscous liquid at a temperatureof 1118 ± 3 deg C. 17, 18) From the phase equi-

librium diagram of the system NaO0-A120 3-SiO 2, asshown in Fig. 5a, a binary mixture of pure sodafeldspar and silica is found to have a minimummelting temperature of 1062 ± 3 deg C (I); thismixture is equivalent to 68.5 percent soda feldsparand 31.5 percent silica. A ternary mixture of 66.0percent soda feldspar, 33.3 percent silica and 0.7percent alumina is shown to form an eutectic at1050 + 10 deg C (M).

Potash feldspar has been established as meltingincongruently at 1150 + 20 deg C to form crystals

of leucite (K,0-A1203-4SiO2 ) and a viscous liquid(12.5 percent KO, 13.5 percent A1203, 74.0 percentSiO 2) which is more siliceous than the feldspar. (19)

With potash feldspar there is a long temperatureinterval during which leucite and the liquid maycoexist at equilibrium; above temperatures of 1530deg C the leucite crystals disappear. As a com-pound in the system K 20-Al103sSiO, (Fig. 5b),potash feldspar theoretically forms a binary eutec-tic with silica at 990 + 20 deg C (I) when thecomposition is 58.0 percent potash feldspar and42.0 percent silica; a ternary mixture of 56.2 per-cent potash feldspar, 43.2 percent silica and 0.6percent of alumina forms a theoretical eutectic at985 ± 20 deg C (M).

The melting relations of binary mixtures of thefeldspars have also been determined. In the sodafeldspar-potash feldspar binary system as shown inFig. 3 a minimum melting temperature of 1063 -t30 deg C is obtained at three compositions (60, 65and 70 percent of soda feldspar), but the compo-sition of 65 percent soda feldspar and 35 percentpotash feldspar is generally considered as the lowmelting mixture. (22)

In the plagioclase series (Fig. 4), the soda andlime feldspars form a series of solid solutions witha continuous rise in the liquidus and solidus tem-peratures from pure soda feldspar to pure limefeldspar.

A recent study has been made to determine thenature of the liquidus surface of the ternary sys-tem soda feldspar-potash feldspar-lime feldspar. (23 )

The data obtained to date has been concerned withthe completion of melting; however, determinationsare being made for the beginning of liquid forma-tion.

The melting behaviors which have been dis-cussed above are those of the pure feldspars andtheir mixtures under equilibrium conditions; thethermal relations of commercial feldspars, such asthose used for the manufacture of whitewares, maybe expected to be somewhat different due to theinfluence of impurities. In ceramic processes, equi-librium states are seldom achieved although thereactions taking place tend to approach thosestates; therefore, the rates at which the reactionsproceed toward equilibrium must be considered.

It has been stated that feldspar melts approachequilibrium conditions very sluggishly due to theirhigh viscosities. These viscosities of melts have beenthe basis of evaluating the thermal behavior offeldspars for ceramic use; it is a common practiceto express this behavior in terms of fusibility orthe plastic deformation of a feldspar cone whenheated at a specified rate (i.e. 20 deg C per hr).When the feldspar cone deformation is comparedto the deformation of standard pyrometric conesheated at the same rate, the fusibility is generallyexpressed in terms of pyrometric cone equivalents(p.c.e.).

At equivalent temperatures, the viscosities ofsoda feldspar melts have been found to be lowerthan the viscosities of potash feldspar melts; themixtures of these alkali feldspars having inter-mediate values. 24 ) Accordingly, the p.c.e. ranges oftypical feldspars used in whiteware compositionsare: (25)

Cone 4-5(1165-1180 deg C) for high sodafeldspars

Cone 5-8(1180-1225 deg C) for intermediatealkali feldspars

Cone 8-10(1225-1260 deg C) for high potashfeldspars

Binary mixtures of 65 parts soda feldspar and35 parts potash feldspar have been found to de-form at temperatures slightly below the deforma-tion temperature of the soda feldspar.(2 6) Ternarymixtures consisting of 70 parts soda feldspar, 25parts potash feldspar and 5 parts lime feldspargive a lower deformation temperature than anyother feldspar mixture.(27)

Viscosity studies have shown that the presenceof uncombined SiO 2 will increase the viscosity of


E=767 1 3°F= II/8 3G

= 740 ± 5°

H =

867 1 30I = 1062 3J r 1470 /0°K = 1545L = 1470 /00M = /050 /0°N /1104f 3j0 1108 30P = /063 5°0

= 068t 5°

T 732 t 50

Weight Percent

A'= /3/5 /0°H 867 3°I 990 20°J /417010 /L /4 1470 /10M 985 ±20°N= 1/40 t2000 1//50 20P = 710 -20°0

= 725 1 50

R /81O 5°S 695

f 50

W /530 5S

Fig. 5. Portions of the Equilibrium Diagrams, Soda-Alumina-Silica and Potash-Alumina-Silica


a feldspar melt and also deter the drop in viscositywith increasing temperatures;28) accordingly, thepresence of free SiO2 increases the p.c.e. value offeldspars. This effect is most pronounced in highsoda feldspars; the presence of 10 percent of un-combined SiO, in a soda feldspar is sufficient toincrease its refractoriness, whereas a high potashfeldspar may tolerate up to 20 percent of free SiO 2before the p.c.e. is materially increased.' 29)

Amounts of Fe2Os in the order of 0.3 percent aresufficient to lower the viscosity of a feldspar meltand thereby lower its p.c.e. value. The addition ofthe oxides of Ca, Mg, Ba and Zr in amounts of 2to 5 percent have also been found to lower theviscosities of feldspars melts and reduce the de-formation temperatures.(

28, 30)

Studies of binary systems of feldspars andclay have not revealed any deformation eutectics;however, it has been established that as a feldsparmelts it takes the decomposition products of clayinto solution at a rate dependent upon the tempera-ture and the surface areas. In general, it has beenshown that clay is more soluble in soda feldsparthan in potash feldspar.( 31, 32

)

The relative solubilities of clay and quartz infeldspar melts have not been definitely established;however, it is known the presence of both materialsaffects their mutual solubility; as the quartz con-tent is increased the solubility of clay in the feld-spar melt is diminished. (33)

The thermal expansion coefficients of feldsparglasses may be predicted from the expansion fac-tors proposed by Hall (34) if the free quartz contentis not in excess of 4 percent. (35) In the crystallinestate soda feldspars have a higher coefficient ofexpansion than potash feldspars; in the fused statethe thermal expansion of the fused potash feldsparis considerably greater than that of fused sodafeldspar chiefly due to the formation of leucite inthe potash feldspar melt. The thermal expansionof fused potash feldspar is effectively reduced bythe presence of free SiOz or soda feldspar due to agreater solution of the leucite.

The density of feldspar is reduced when itchanges from the crystalline to the fused state. Ahigh soda feldspar with a density of 2.635 in itsraw state will fuse to a density of 2.37, and as aresult will occupy a 12 percent greater volume.Similarly a high potash feldspar of 2.572 densitywill fuse to a density of 2.37 with a volume in-crease of 9 percent./ 25)

The thermal properties of feldspars have beenthe subject of considerable study in the past andwill doubtless be the object of many future investi-gations. Several factors have been established assignificant for the thermal behavior of feldsparsand may be summarized as:

(a) chemical composition, which determinesthe ultimate equilibrium condition for anyspecified temperature

(b) mineralogical composition, which deter-mines the initial point from which reactions willproceed and the nature of melting (i.e. con-gruent or incongruent)

(c) particle size of the mineral constituents,which determines the extent of surface area andthe rate at which melting will take place

(d) viscosity of the melt formed, whichultimately determines the rate at which re-actions will proceed toward the equilibriumstate(36)

10. Solubility in WaterIn addition to the thermal behavior of feldspars,

the solubility of feldspars in water is of interestin ceramic processes. No definite conclusion hasbeen made as to whether a high soda or a highpotash feldspar is more soluble in water. It is wellknown that when finer ground fractions of feld-spar are placed in water, decomposition takes placeand alkalies are extracted from the feldspar. Thealkalinity of the water is immediately increasedand continues to increase with time although therate of increase gradually becomes smaller withlonger exposure.(7)

III. USE IN WHITEWARE COMPOSITIONS

11. Physico-Chemical BehaviorThe function of feldspar in whiteware bodies is

that of a flux and as such it takes part in physico-chemical reactions with other crystalline phases.The old conception that the feldspar serves as abond for the crystalline phases is being rejected andcurrent theories consider that the bonding of grainsand formation of a dense mass is due to a deepinter-diffusion of phases. (38 )

During the firing of a whiteware body composedof feldspar, clay and flint, the first glassy phase toform is due to a ternary eutectic. With pure mate-rials and equilibrium conditions the temperatureof the eutectic formation would be 990 ± 20 deg Cwith a pure potash feldspar or 1050 ± 10 deg Cwith a pure soda feldspar, and the amount of meltformed would vary with the amount of feldsparpresent. In bodies which contain somewhat impurematerials, the eutectic temperatures may be some-what lower. Commercial bodies are fired at a ratewhich is necessarily much too rapid for the achieve-ment of equilibrium conditions; and while someeutectic formation may occur at the theoreticaltemperature, the amount of glass formed at thatpoint will be very minute. The eutectic formationmay be more readily detected at somewhat highertemperatures. Partial fusion has been observed at1075 - 1085 deg C in bodies fired at a rate of 10deg C per minute. ('" As the temperature is in-creased more liquid melt is formed which begins todraw particles together by surface tension and pro-gressive solution takes place. Mullite, which wasformed from the decomposition of clay, may diffuseinto the melt or the crystals may continue to growat higher temperatures and increased time of heattreatment. Mullite is also formed by recrystalliza-tion as the melt becomes saturated or when thetemperature is lowered. Above the normal vitrifica-tion range, air, which had been entrapped withinthe pores of the body by the melt, will build upsufficient pressure to expand against the viscousglass and cause bloating or blistering.

The fluxing influence of a feldspar used in abody does not necessarily follow the order antici-

pated by the fusion behavior of the feldspar aloneor combinations of the feldspar with one of theother body ingredients. Each body composition isan individual system and its thermal reactions aredependent on all of the components collectively.It might be expected that a body containing a highsoda feldspar of low p.c.e. value would reachmaturity at a temperature considerably lower thana similar body containing a high potash feldspar ofhigh p.c.e. value; actually, the difference in thematuring temperatures of the two bodies may be ofsmall magnitude (i.e. 10 deg C) in contrast to thedifference in fusion temperatures (i.e. 70 deg C) ofthe individual feldspars. It has been shown thatbodies containing potash feldspar may even maturebefore those containing equivalent amounts of sodafeldspar if free calcia is present in the compo-sitions. (3)

It has been generally accepted that finer particlesizes of feldspar and flint lower the maturing tem-peratures of feldspar-flint-clay bodies and thatonly the finest particles of feldspar form a glassymatrix. The larger feldspar particles fuse and be-come isotropic but do not lose their original shapeto any great degree. (4 0, 41) The advantage of finerparticle size is most effectively realized with con-trolled firing procedures; a more gradual heatingrate or longer soaking time at the maturing tem-perature will effect a more representative saturationof the glassy phase. ( 42) A body which achieves aporosity of zero percent in 10 min at 1250 deg Cmay reach the same degree of vitrification at 1200deg C in a longer time period. (43)

In addition to the effect of finer particle size onthe maturing temperature of bodies, the finenessalso governs the degree of firing shrinkage, strength,thermal expansion and warpage. Finer grinds offeldspar tend to increase the firing shrinkage andstrength; however, greater warpage and lowerthermal expansion result. (44) The warpage of bodiesis a function of the viscosity and the proportionof glassy phase formed at elevated temperatures.Greatest warpage is obtained in bodies of highflint and high feldspar contents; increasing clay


content at the expense of flint effectively reduceswarpage.'4 5 46) Soda feldspars which form lessviscous glasses have been found to cause morebody warpage than potash feldspars. The thermalexpansion of fired bodies is largely dependent uponthe amount of uncombined quartz and will there-fore decrease with higher firing temperatures."( 4

In general, higher thermal expansions have beenobtained in bodies containing soda feldspars.(25)

12. Purpose and Scope of Experimental InvestigationA laboratory investigation was conducted to

determine the comparative effects of high potashand intermediate soda-potash feldspars on the firedproperties of various types of whiteware bodies.The need for this study has been accentuated bythe rapid depletion of readily available supplies ofhigh grade potash feldspar. In the past, feldsparsof highest potash content have been sought as thebest type for use in ceramic whitewares, but recentstudies have shown that feldspars of considerablesoda content may be used advantageously as re-placements for very high potash feldspars withlittle or no significant changes in the fired proper-ties of whiteware bodies.' 481

Four commercial feldspars were used in thestudies of bodies. The body compositions selectedfor the investigation were those which are typicalof:

semi-vitreous dinnerwarehotel chinasanitary wareelectrical porcelainfloor tile

Compositions were formulated to study the effectsof different types of feldspars, variable feldsparcontent and the effect of auxiliary fluxes in com-bination with the feldspars.

Bodies were prepared and formed into testspecimens according to commercial processingmethods. The details of body preparation, specimenformation, firing and testing procedures are de-scribed in the Appendix.

13. Properties of Feldspars UsedThe properties of the four commercial feldspars

were determined and are shown in Table 3. Thefeldspars Buckingham and Custer are of the highpotash type with KO to NaO ratios of more than3 to 1 and microcline to albite ratios greater than

2 to 1. These feldspars may be expected to givesomewhat similar firing properties in typicalfeldspar-flint-clay bodies due to their high potashcontent. A-3 and F-4 feldspars have KO to NazOratios of 1.53 and 1.07 respectively, so that they

Table 3Properties of Feldspars Studied

a. Chemical AnalysesWeightPercent Buckingham Custer A-3 F-4

SiOJ 66.7 67.5 71.1 65.6A120o 18.4 17.8 16.4 20.3Fe2Oa 0.05 0.08 0.07 0.04CaO 0.1 0.1 0.5 2.2MgO tr tr tr trNa2O 3.0 3.0 4.5 5.7KzO 11.4 10.8 6.9 6.1Ignition Loss 0.3 0.3 0.3 0.2

b. Mineralogical Analyses (calculated weight percent)Albite 25.4 25.4 38.1 48.2Microcline 66.8 63.0 42.0 37.3Anorthite 0.5 0.5 2.5 10.9Free Quartz 4.6 7.7 16.5 3.5MuscoviteKaolinite

and Assoc.Minerals 2.7 3.4 0.9 0.1

c. Particle Size Distribution by Sedimentation Analyses (weight percent)Less than 30 microns 71 69 65 78Less than 20 microns 58 58 54 66Less than 15 microns 48 50 45 56Less than 10 microns 33 38 34 42Less than 5 microns 16 21 17 21Less than 2 microns 11 7 5 5

d. Fusibilityp.c.e. (heating rate

50 deg F per hr) 103 103

91 83e. Solubility in Water (milliequivalent per liter*)

NasO 0.12 0.10 0.15 0.19K 2 O 0.39 0.28 0.17 0.16Total KNaO 0.51 0.38 0.32 0.35

* After 24 hrs of exposure; 1 part by weight of feldspar to 25 parts byweight of distilled water; determined by flame photometer.

are classified as intermediate soda-potash feldspars.Actually, F-4 may be considered as a plagioclase-potash feldspar due to its appreciable anorthitecontent. It also should be noted that A-3 contains aconsiderable amount of free quartz and thereforethe actual feldspathic content per unit weight issomewhat reduced. F-4 is a flotation-processedfeldspar which is relatively free of the associatedminerals such as muscovite.

The chemical analyses of the feldspars arethose furnished by the Consolidated Feldspar De-partment of the International Minerals and Chem-ical Corporation. The mineralogical compositionswere calculated from these analyses according tothe methods outlined by Koenig. 9)

The particle size distribution was determinedby the Andreasen Pipette method as proposed byLoomis' 49' and Russell and Weisz,"5o) and is pre-sented in tabular form.

Fusibilities of the individual feldspars were de-termined by firing several small slender trihedral

Bul. 422. PROPERTIES OF FELDSPARS AND THEIR USE IN WHITEWARES

pyramids of each at a heating rate of approxi-mately 50 deg F per hr in an electric furnace andobserving their deformation in relation to the de-formation of standard pyrometric cones which werefired simultaneously. The reported values of fusi-bility are expressed in terms of p.c.e. and representthe pyrometric cone equivalent at the temperaturewhen each feldspar specimen had deformed suchthat the tip of the specimen was level with thesupporting plaque.

The relative solubilities of the feldspars inwater were obtained using particle sizes between44 and 53 microns. Ten grams of each feldspar,collected between Nos. 270 and 325 mesh sieves,were placed in 300 ml pyrex bottles containing 250ml of double distilled water. The bottles weresealed and tumbled end over end for 24 hr; afterthe agitation period the feldspar water mixtureswere centrifuged and the supernatant liquid drawnoff for solubility tests. Ninety ml of each liquidwere added to 10 ml of a standard Li solution.These solutions were examined with a Perkin-Elmer Flame Photometer for Na20 and KO con-tent. The values expressed in the table of solubili-ties represent the milliequivalents of NaO and KOper liter taken into solution from each feldsparafter 24 hr of exposure in water.

14. Semi-Vitreous WareComposition

Four series of compositions were formulated forthe purpose of studying the fired characteristics oftypical semi-vitreous dinnerware bodies as in-fluenced by high potash and intermediate soda-potash type feldspars, varying feldspar content, andcombination of feldspar with small amounts ofauxiliary fluxes. A total of 21 bodies was studied;the compositions are shown in Table 4.

A typical semi-vitreous body which was knownto achieve a porosity of approximately 10 percentat cone 10 was selected for a base composition.This composition contained:

33.5 percent of flint36.0 percent of mixed Tennessee and Kentucky

ball clays21.0 percent of mixed Georgia and North Caro-

lina kaolins13.5 percent feldspar

Bodies of Series I (Table 4) were of this composi-tion; succeeding series of bodies represent variationsin the base composition such as increased feldsparcontent, addition of auxiliary flux, or both. Thedeviations from the base composition were, for themost part, compensated for by appropriate altera-tions in the kaolin content.

Four commercial feldspars were used for thecompositions of Series I; the body in which eachwas contained may be readily identified by thecode letter following the body composition number.Thus, bodies SV1B and SV1C respectively con-tained Buckingham and Custer feldspars, whichare of the high potash type; bodies SV1A andSV1F contained A-3 and F-4 feldspars respectively.These feldspars may be considered as intermediatesoda-potash types. The analyses arid properties ofthe feldspars used have been shown in the previoussection. The Custer (C) and F-4 (F) feldsparswere selected as representative of the high potashand intermediate soda-potash type feldspars andwere used for comparative purposes.

Bodies of Series II represent increases in thefeldspar content over that of the base composition.Increases to 15.0 and 16.5 percent feldspar weremade at the expense of Pioneer kaolin. Series IIIwas composed of bodies in which 2.0 percent addi-tions of various fluxes were added to the base

Body No.Feldspar AFeldspar BFeldspar CFeldspar FC & C Ball ClayImperial Ball ClayTenn. No. 5 Ball ClayOld Mine No. 4 Ball ClayPioneer KaolinKamec KaolinE.P.K.Ottawa FlintWhitingLow Lime TaleTremolitic TalcMagnesite

Series I1A 1B 1C 1F

13.513.5

Table 4Compositions of Semi-Vitreous Bodies

Series II Series III2C 2F 3C 3F 40 4F 5C 5F 6C 6F 70 7F

16.5 13.516.5 13.5

7.0 7.0 7.0 7.05.0 5.0 5.0 5.07.0 7.0 7.0 7.07.0 7.0 7.0 7.09.0 9.0 10.0 10.09.0 9.0 9.0 9.06.0 6.0 6.0 6.0

33.5 33.5 33.5 33.52.0 2.0

13.5 13.513.5 13.5

7.0 7.0 7.0 7.05.0 5.0 5.0 5.07.0 7.0 7.0 7.07.0 7.0 7.0 7.0

10.0 10.0 10.0 10.09.0 9.0 9.0 9.06.0 6.0 6.0 6.0

33.5 33.5 33.5 33.5

2.0 2.0

Series IV8F 9F 10F 11F 12F

16.5 13.5 15.0 16.56.5 7.0 6.5 6.05.0 5.0 5.0 5.06.5 7.0 6.5 6.06.5 6.5 6.0 6.09.0 9.0 9.0 9.09.0 9.0 9.0 9.05.5 5.5 5.5 5.0

33.5 33.5 33.5 33.5

2.0 2.0 2.0 2.0 4.0 4.0 4.02.0 2.0


compositions. The fluxes used were whiting, lowlime talc, tremolitic talc and magnesite.

In Series IV only F-4 feldspar was used in thebody compositions for the purpose of determiningthe effects of varying combinations of feldspar withtremolitic talc. These compositions contained 15.0and 16.5 percent F-4 feldspar in combination with2.0 percent tremolitic talc, and combinations of13.5, 15.0 and 16.5 percent F-4 feldspar with 4.0percent tremolitic talc.

Processing and FiringThe semi-vitreous dinnerware bodies were pre-

pared by blunging and filter pressing. Test speci-mens consisted of extruded 7/% in. diam rods. Setsof 15 specimens of each body composition werefired to temperatures of 2140 deg F, 2200 deg F,2230 deg F, 2280 deg F and 2340 deg F in anelectrically heated kiln. These temperatures cor-responded to cone equivalents of 56*, 76, 86, 96, and116. The temperature rise was controlled at approxi-mately 100 deg F per hr up to 2000 deg F and 50deg F per hr thereafter until the peak temperaturewas reached and held for 11/ hr. In addition to thelaboratory firings, several bodies were fired in acommercial tunnel kiln to cone 93 (See Table 5).

Tests were conducted on the fired specimens toobtain the fired porosity, modulus of rupture andvolume shrinkage. The thermal expansion andcrystalline nature of bodies fired to cone 96 werealso determined.t

ResultsPhysical Properties. The physical properties of

the fired semi-vitreous dinnerware bodies are pre-sented in Fig. 6. The data in this form are con-venient for examining the relations between thefiring temperatures and the changes in the proper-ties of each body; however, a comparison of theproperties of all bodies at some common conditionsuch as maturity is more desirable. For convenience,therefore, the authors have arbitrarily assumedmaturity in the semi-vitreous bodies to be thatcondition at which the body has a fired porosity of10 percent. In order to obtain the body propertiesat the point of 10 percent porosity, the followingprocedure was followed: A vertical line is drawnthrough the 10 percent point on the porosity curve;the intersections of the vertical line with the tem-

*The superscript refers to the degree of bending of the cone tipcorresponding to the numerals on a clock face between 1 and 6.

tFor detailed description of the equipment and procedures employedin the processing and testing of bodies, see Appendix.

perature scale and the other physical propertycurves determine the maturing temperature andphysical properties at that temperature. The resultsthus obtained with all bodies are shown as bargraphs in Fig. 7.

In compositions of Series I, it was observed thatthe bodies containing high potash feldspars (SV1Band SV1C) matured at about the same tempera-ture with similar strengths and shrinkages. Thesoda-potash feldspar bodies (SV1A and SV1F)matured at slightly higher temperatures with less

Table 5Comparison of Fired Properties of Semi-Vitreous Bodies as Obtained

from Laboratory and Commercial Firings (2240 deg F)

Body No. Percent Porosity Modulus of Rupture (psi)Lab. (cone 91) Comm. (cone 93) Lab. (cone 91) Comm. (cone 98)

1C 16.43 17.95 5814 5714IF 17.45 18.07 5583 55494C 12.21 13.40 6010 56584F 13.27 14.01 5780 59365C 11.96 12.62 6234 62315F 12.56 13.10 6523 63356C 11.88 12.48 6520 64256F 12.21 12.95 6413 65137C 8.93 10.14 6541 65747F 7.87 8.75 6605 6513

Note: Laboratory firing in electric kiln with 100 deg F temperature rise;commercial firing in gas fired tunnel with 29-hr total cycle.

strength than the potash feldspar bodies. Theshrinkage of SV1F was similar to the shrinkage ofSV1B and SV1C, while SV1A showed less shrink-age, possibly due to the high free quartz contentof A-3 feldspar.

Increasing feldspar content lowered the matur-ing temperature of the bodies somewhat as wouldbe expected; however, it was observed that whilebody SV1F matured at a slightly higher tempera-ture than SVIC, the two bodies, SV2F and SV2C,matured simultaneously, and SV3F matured earlierthan SV3C. The increased feldspar contents causedno appreciable change in the strengths of the

.matured bodies. The shrinkages of bodies madewith high potash feldspar were essentially un-changed when the feldspar content was increasedfrom 13.5 to 16.5 percent; however, in the soda-potash bodies an increase in shrinkage was notedwhen the feldspar content was raised from 13.5 to15.0 percent.

From the results of physical properties obtainedwith the first two series of compositions, it may beconcluded that direct weight substitutions of soda-potash feldspar may be made for high potashfeldspar in a semi-vitreous type body with (a) littleor no change in maturing temperature, (b) slightlyless transverse strength, and (c) little or no in-crease in shrinkage.


Deg i ii i i ii i

F 2/50 2200 2250 2300 2150 2200 2250 2300Firing Temperature

Fig. 6. Fired Properties of Semi-Vitreous Bodies

0

04

--

18 ILLINOIS ENGINEERING EXPERIMENT STATION

254 25.4

-23 020

202

22.2 pq22. L l

23.6

22.4 l142228

225 2-

24.2

Fig. 7. Properties of Semi-Vitreous Bodies at Maturity (10 percent Porosity)

I

20

/8-

16

220

2/.4

~··I··I

i::iii

:::: ::"-'- --'-"

i·-i

ii_- :ii--i

Body IA /B IC IF 2C 2F 3C 3F 4C 4F 5C 5F 6C 6F 7C 7F 8F 9F IOF IIF 12F% Feldspar 13.5 /3.5 /35 135 /5.0 /5.0 /65 /6.5 /35 13.5 135 /3.5 /3.5 /3.5 /3.5 /3.5 5.0 /6.5 /35 5.0 16.5%Auxi/iory 0 0 0 0 0 0 0 0 o 20 20 0 0 2.0 20 4.0 4.0 4.0

Flux Whiling Low Lime Tremolific Magnesite Tremo/itic TolcTalc Talc

. "

:~::;

i~·iiii···:·I"'::3

i~z:::::::--::--:·:,:i::-:-

I .

":::'

::a::-

i L

91·iiii

-:::i:-:

::-ii--

~II: ii:,-::: :s~i-ii~ni-:is:s~

::::::::"

iE-·l

:::·::::::--:--

ilLii

::-::::

:-::::.::::·-:::;::i:::

:;-:::::;_

.... •JQ

::i:::::::·~~iiiiiiil

: ::: :::-'

:::::::"

-:_--:-i·:::i::·::·i:::i:::·:~·:::::~;:: :··:·1::j::::~·::;:::·

_i-ii:--1;·::::::i:·;:::;:·

:: : :::

ri iii:-

:;-_:::

:::::-I:·i:

-i':-----

:: ::

::::::::::

::::i:: ::_::-----

::::::::::

20 1 /9.8

19.2


The addition of 2 percent whiting to the basecomposition was effective in reducing the maturingtemperature of the potash feldspar body (SV4C) toabout one cone below that of the base body. Theeffect of a like whiting addition to the soda-potashbody (SV4F) was not quite as pronounced. Thewhiting addition to the potash feldspar body re-sulted in reduced strength at maturity. Littlechange of matured shrinkage was noted.

The effect of 2 percent low lime talc with thepotash feldspar was very similar to that obtainedwith whiting; with soda-potash feldspar slightlymore fluxing was obtained but with little changein the matured properties. The addition of 2 per-cent tremolitic talc to the base composition wasonly slightly more effective in its fluxing actionthan the whiting or low lime talc; however, some-what greater shrinkage resulted.

Bodies SV7C and SV7F contained 2 percentadditions of magnesite. The maturing temperatureof SV7C was about 2 cones lower than the basebody SV1C, while SV7F matured about 3 coneslower than the base body SV1F. These markedreductions in maturing temperature were accom-panied by increases in volume shrinkages but noappreciable changes in strength.

Combined increases in feldspar content and talcadditions reduced the maturing temperature verymarkedly. A 3 percent increase in feldspar contentand an addition of 4 percent of talc reduced thematuring temperature by 5 cones. At the lowermaturing temperatures, slightly less strength andreduced shrinkages were obtained.

A comparison of the results obtained from acommercial firing and a laboratory kiln firing areshown in Table 5. It is noted that in all bodies thefired porosities obtained with the laboratory kilnwere lower than those obtained from the commer-cial firing. This difference may be attributed to thedifference in firing rates, the commercial rate beingalmost twice as rapid as the laboratory cycle. Thevalues of strength obtained from the two firingswere in close agreement.

Crystalline Content. The crystalline nature ofthe specimens fired to cone 96 were determined byX-ray analysis.* The only crystal phases identifiedin the bodies were quartz, mullite, and in somecases, cristobalite. The crystalline content of SeriesI bodies made with potash feldspar was entirelyquartz and mullite; the soda-potash feldspar bodies

* See Appendix for description of X-ray unit and techniques employed.

showed a slight amount of cristobalite present butnot in excess of 5 percent. Increasing feldspar con-tent was found to decrease the amount of freequartz and to increase the mullite formation.

The addition of whiting also decreased thequartz content and slightly increased the mulliteformation. A small degree of cristobalite formationwas noted in the soda-potash body. In those bodiescontaining 2 percent of talc, the amount of freequartz was considerably reduced and some cristo-balite development was noted in the potash feldsparbodies; while in the soda-potash bodies consider-able cristobalite development occurred. Magnesitewas found to cause extensive cristobalite formationwith the soda-potash feldspars and to a lesserdegree with high potash feldspars. The quantitativedetermination of cristobalite for several fired bodiesis shown in Table 6.

Table 6Cristobalite Content of Fired Semi-Vitreous Bodies

Body No. Percent Percent Percent CristobaliteFeldspar Tremolitic (± 2 percent) After

Talc Firing to Cone 96

IC 13.5 CIF 13.5 ... tr.6C 13.5 C 2.0 tr.6F 13.5 F 2.0 128F 15.OF 2.0 119F 16.5 F 2.0 10

10F 13.5 F 4.0 1611F 15.0 F 4.0 1512F 16.5 F 4.0 12

The compositions which showed high cristobaliteformation were fired in an X-ray furnace to -deter-mine the temperature at which the cristobaliteformed. The formation was found to occur duringthe cooling cycle at temperatures between 1100deg C and 1000 deg C. The cristobalite was formedby the process of separation with crystal depositionof SiO, or devitrification.

Thermal Expansion. The linear thermal expan-sion curves for bodies of Series I and II fired tocone 96, are shown in Fig. 8. The expansions ob-served with the soda-potash feldspar bodies wereslightly higher than those noted in bodies contain-ing potash feldspar. The high expansion of bodySV1A is undoubtedly due to the high free quartzcontent of A-3 feldspar. An increase of feldsparcauses more crystalline silica to dissolve, thus de-creasing the expansion.

The combination of auxiliary fluxes and potashfeldspar caused a reduction in the thermal expan-sion of bodies fired to cone 96 . In bodies containingsoda-potash feldspar, the addition of auxiliaryfluxes which contained considerable percentages of


IF

0.2

S 0.5

0.4

/,4 ·

o F\ /.ff \l /llll

Temperature, C

Fig. 8. Linear Thermal Expansions of Series I and IISemi-Vitreous Bodies Fired to Cone 9

MgO were found to cause a hump in the thermalexpansion curves between 150 - 250 deg C, indi-cating the presence of cristobalite. The expansiondata obtained with bodies containing the soda-potash feldspar and tremolitic talc are shown inFig. 9. The most pronounced cristobalite inversionwas noted in body SV10F which contained 13.5percent of F-4 feldspar and 4 percent tremolitictalc. This data agrees favorably with the X-rayanalysis since body SV10F was found to have thegreatest cristobalite development among thosebodies containing tremolitic talc as an auxiliaryflux.

Thermal Shock. The influence of cristobalite onthe thermal shock properties of the bodies wasstudied using glazed 6-in. coupe plates. Five platesof each of the compositions were heated to 480 degF in an electric oven and quenched in a dye solu-tion at 70 _ 2 deg F; the number of cycles re-quired to cause a body or glaze defect was recorded.Table 7 shows selected results of the test. The

Table 7Thermal Shock Resistance of Semi-Vitreous Bodies from

480 deg F to Room Temperature (70 ± 2 deg F)Body No. Percent Percent Trial

Feldspar Tremolitic I II IIITalc

c1 13.5 C ... 4c*

5 4c

IF 13.5 F ... 4C 5 5C6C 13.5 C 2.0 4

c 5

c 5c

6F 13.5 F 2.0 5C

4ED 3E D

10F 13.5 F 4.0 3ED 3D 4D

* Number indicates the cycltheuring which the defect was noted: C-crazed; ED-edge dunt; D-dunted.

bodies containing extensive cristobalite develop-ment were found to have a reduced thermal shockresistance although their crazing resistance wasimproved. Such bodies are likely to dunt in thekilns unless special precautions for cooling areobserved.

15. Hotel ChinaComposition

Two series of hotel china compositions wereformulated as shown in Table 8. A typical compo-sition containing 21 percent feldspar was selectedas a base composition and four bodies of this typewere made to show the influence of the four com-mercial feldspars, A-3, Buckingham, Custer andF-4, when each is used as the total flux content.These bodies are represented by compositions from2A to 2F in Series I. All other bodies contain eitherCuster or F-4 feldspars.

cz

A

B

0 IoU zoo iuu o U U OU/u SUUTemperature, C

Fig. 9. Linear Thermal Expansions of Semi-Vitreous Bodies

Containing Tremolitic Talc and Fired to Cone 9

In Series I, the feldspar content is varied from18 to 25 percent at the expense of kaolin. In seriesII additions of auxiliary flux are made to bodiescontaining Custer and F-4 feldspars.

Processing and FiringThe hotel china bodies were prepared by a

process of ball milling and filter pressing. After oneweek of aging, the bodies were formed into 7/-in.diam rods by extrusion. Specimens were fired in anelectrically-heated kiln at a rate of approximately75-80 deg F per hr up to 2000 deg F and 50 deg Fper hr thereafter to temperatures of 2180, 2220,


2250, 2280, 2310, 2350, 2400, and 2450 deg F. A 11/2hr soak at these temperatures was used to deformstandard pyrometric cones to 73, 83, 93, 103, 113, 122,133, and 136 respectively.

ResultsThe physical properties obtained with the fired

specimens of hotel china bodies are shown in Fig.10. In the bodies which contain 18 percent of eitherCuster or F-4 feldspar, vitrification was observedto occur at a temperature of approximately cone126, where the bodies developed zero porosity anda transverse strength in excess of 10,000 psi. Thevolume shrinkages of the two bodies were found tobe similar with a value of approximately 30 per-cent at vitrification. Maximum values of strengthand shrinkage were observed at cone 133; the bodycontaining Custer feldspar had slightly greater ulti-mate strength than the corresponding F-4 feldsparbody.

With 21 percent feldspar (2A-2F), the firingbehavior of bodies containing A-3, Buckingham,Custer and F-4 feldspars may be compared. Thebody containing A-3 feldspar was vitrified at cone126, while the bodies containing Buckingham,Custer or F-4 feldspar vitrified at a half conelower, cone 121. The potash feldspars promotedslightly more strength and somewhat less shrinkagethan the soda-potash feldspars. The A-3 feldsparcomposition did not show evidence of overfiring atcone 136, while the potash feldspars bodies indi-cated overfiring at this temperature. The body con-taining F-4 feldspar was found to begin overfiringat cone 133.

Bodies which contained 25 percent of eitherCuster or F-4 feldspar were found to vitrify be-tween cones 11" to 123 with evidences of overfiringat cone 133. The F-4 composition approached vitri-fication at a more rapid rate than the Custer feld-spar body and also showed a little more shrinkage.

The additions of talc to the hotel china bodiescontaining 21 percent feldspar increased the rate of

vitrification, thus lowering the maturing tempera-ture. An addition of 2 percent talc to the Custerfeldspar body promoted vitrification at cone 11Iwhich was one cone lower than the maturing tem-perature with no talc present. In a similar bodycontaining F-4 feldspar, a 2 percent talc additionreduced the vitrification temperature from cone 123to about cone 106. The bodies containing 2 percenttalc began to overfire at cone 126. An increase inthe talc addition to 4 percent was effective inpromoting the vitrification of the Custer body atcone 103 and the F-4 body at about cone 96. Over-firing took place at cone 123. The bodies of in-creased feldspar content with 2 percent talc addi-tions vitrified at cone 103 and overfired at cone 126.In general it was observed that when talc wasadded, the rate of fluxing was more rapid in F-4feldspar bodies than in corresponding Custer feld-spar bodies, and as a result had shorter firingranges.

The fluxing effect of whiting with Custer andF-4 feldspars was observed in bodies 6C and 6F.Two percent additions to bodies containing 21 per-cent of feldspar effected vitrification after firing tocone 123; overfiring was noted at cone 133. Thewhiting was found to be a more efficient flux withCuster feldspar than with F-4 feldspar.

A qualitative analysis of the crystalline phasespresent in the matured hotel china bodies investi-gated showed evidence of some cristobalite develop-ment in all bodies which contained F-4 feldspar,while little or no cristobalite was noted in Custerfeldspar bodies.

16. Electrical PorcelainComposition

One series of electrical porcelain-type bodieswas investigated. The composition of this series ofbodies is shown in Table 9; the only variation inthe compositions of the four bodies is the type offeldspar used. As in previous studies, the fourcommercial feldspars, A-3, Buckingham, Custer

Table 8Compositions of Hotel China Bodies

Body No. 1CA-3 FeldsparBuckingham FeldsparCuster Feldspar 18F-4 FeldsparOttawa Flint 35Old Mine No. 4 Ball Clay 8.5Kamec Kaolin 15Pioneer Kaolin 13.5Florida Kaolin 10Tremolitic Talc No. 1Whiting

Series IIF 2A 2B 2C 2F 3C 3F

21

Series II4C 4F 5C 5F 6C

21

35 35 358.5 8.5 8.5

14 14 1412.5 12.5 12.59 9 9

6F 7C 7F

21

358.51411.58

2 2


Firing Temper

Fig. 10. Fired Propertie

and F-4 are employed as the total feldspathic con-tents, and the bodies in which they are containedmay be identified by the code letters following thecomposition designations.

Processing and FiringThe electrical porcelain bodies were prepared

by blunging and filter pressing. Test specimens for

ature

es of Hotel China Bodies

shrinkage, porosity and modulus of rupture deter-minations were formed by extrusion. The testspecimens consisted of 7/-in. diam rods. Samplesfor the determination of dielectric properties wereformed by plastic pressing in a plaster mold; thesesamples were made in the shape of a 6%-in. diamdisc 56 in. thick, the surfaces of which wereground approximately parallel after drying.

'' '"'

CMn

Deg

Bul. 422. PROPERTIES OF FELDSPARS AND THEIR USE IN WHITEWARES

Test specimens were fired in an electricallyheated kiln at a heating rate of approximately 80deg F per hr to temperatures of 2230, 2260, 2300,2340, 2375, and 2410 deg F and held at these tem-peratures for 2 hr. Cone plaques set among the

Composi

Body No.A-3 FeldsparBuckingham FeldsparCuster FeldsparF-4 FeldsparOttawa FlintVictoria Ball ClayH.T.P. Ball ClayKentucky Old Mine

No. 4 Ball ClayPioneer KaolinKamec KaolinE.P.K. Florida Kaolin

Table 9tions of Electrical Porcelain Bodies

EP1A EP1B EPIC30

20 2010 1012 12

EPlF L<

30201012

specimens in each burn indicated the above firingsto be respectively equivalent to cones 8

6, 96, 112,123, 126, and 131. The dielectric test samples werefired to the vitrification temperatures determinedfor each body composition.

ResultsThe physical properties determined for the elec-

trical porcelain bodies are shown in Fig. 11. BodyEP1A, which contained A-3 feldspar was found tomature later than any of the other bodies showinghighest values of porosity and lowest volumeshrinkage throughout the firing range. As shown inother type bodies, this behavior was probably dueto the high free quartz content of A-3 feldspar.The vitrification of body EP1A was shown to occurat a temperature between cones 126 and 133.

The two potash feldspar bodies, EP1B andEP1C, had similar fired physical characteristics

Table 10Dielectric Properties of Electrical Porcelain Bodies*

Body Range of Storage Dielectric Loss DielectricNo. Capacitance Factor Constant Factor Strength

(uuf) (percent) (volts permil)

EP1A 71-85 111.1 5.08 4.57 289.0EPIB 77-90 117.0 5.40 4.62 284.3EP1C 77-94 118.0 5.37 4.55 307.3EP1F 80-86 106.0 5.60 5.28 207.7

* Measured by Locke Department, General Electric Corporation. Theseare the average of values obtained on 6 specimens. The specimens werecircular discs approximately 6 in. in diam by .20 in. in thickness.

throughout the firing range studied. The vitrifica-tion of both of these bodies was complete at cone123-126.

The body showing the earliest vitrification inthis series of electrical porcelain bodies was EP1Fwhich contained the flotation feldspar, F-4. Thisbody was shown to achieve complete vitrificationat temperatures between cone 112 to 123.

Deg F 2200 2250 2300Firing Temperature

2350 2400

Fig. II. Fired Properties of Electrical Porcelain

The modulus of rupture values obtained withthis series of bodies indicated the F-4 feldsparbodies gave higher strengths than the potash feld-spar bodies at equivalent firing temperatures. Thefiring shrinkage of the F-4 feldspar body wasgreater than any of the other bodies in this seriesthroughout the firing range.

The results of the electrical tests conducted onthe dielectric samples are shown in Table 10. Themeasurements were made by the Locke Departmentof the General Electric Company. The electricalproperties which were of chief concern in evaluatingthese bodies were the dielectric strength and lossfactors. The dielectric strengths of conventionalhigh voltage porcelains range between 250 and 300volts per mil; all of the bodies tested show favor-

C IC

7/1

8I 96 // /2 s /26' /3S

.<?

30

/


Body No. SIA S1BA-3 Feldspar 32Buckingham Feldspar 32Custer FeldsparF-4 FeldsparMartin No. 5 Ball Clay 15 15C & C Ball Clay 7.5 7.5Royal Ball Clay 7.5 7.5Pioneer Kaolin 9 9Kamec Kaolin 9 9Ottawa Flint 20 20Tremolitic Talc No. 1

Series I

Table 11Compositions of Sanitary Ware Bodies

Series IISiC S1F S2A

32S2B S2C S2F

Series III Series IVS3A S3B S3C S3F S4C S4F S5C S5F

Table 12Casting Characteristics of Sanitary Ware Bodies

Specific Flow TestGravity sec per

100cc

1.7951.8011.7971.8001.8011.800

1.8001.8011.7971.7981.8001.798

1.7981.8011.8021.8001.7981.797

1.7971.8041.7961.7981.8011.797

Wet Wt Dry Wtof Cast of Cast

282.2280.9264.8255.3245.6245.2

291.3266.3250.0243.5235.7233.7

289.6256.8245.0241.4236.3235.0

277.4249.6246.3239.5237.3236.1

239.8240.3227.2219.6211.9211.7

240.1221.7208.6203.6196.8195.5

239.6216.2204.8201.6197.1196.5

236.1213.5211.7205.6203.5202.2

the highest breakdownstrength was shown by the Custer feldspar bodyand the lowest breakdown strength in the F-4 feld-spar samples. .The F-4 feldspar body also had thehighest loss factor of the series.

17. Sanitary Ware

CompositionIn the study of sanitary ware bodies, four series

of compositions were formulated as shown in Table11. A typical sanitary ware composition containing32 percent of feldspar was selected for the firstseries of compositions. This series consisted offour bodies, each of which was made with a differ-ent commercial feldspar - A-3, Buckingham, Cus-ter and F-4. In Series II, 2 percent of tremolitictalc was added to each of the above bodies. Thetalc addition was increased to 4 percent in seriesIII. In Series IV, bodies containing Custer and F-4feldspars were made at a reduced feldspar level(28 percent) with and without a 2 percent talcaddition. All variations from the base composition

were compensated for byin the kaolin content.

appropriate alterations

Processing, Casting and FiringThe sanitary ware bodies were prepared by

blunging and filter pressing. After one week ofaging, the casting characteristics of the variousbody compositions were determined at six electro-lyte levels. These casting trials were determinedwith slip batches containing 1000 grams of drybody. The sodium carbonate content was held con-stant at 0.05 percent in all trial slips but the sodiumsilicate content was varied from 0.05 to 0.175 per-cent in increments of 0.025 percent.*

The results of the casting trials indicated thatelectrolyte additions of 0.05 percent sodium car-bonate and 0.10 percent sodium silicate ("N"Brand) were suitable for all body compositions.Approximately 10 gal of slip were prepared fromeach body composition with this addition of electro-lyte. The slips were adjusted to 1.80 specificgravity and de-aired in a vacuum mixer.

Details of procedures used in conducting the initial casting trials maybe found in the Appendix.

PercentSodium

CarbonateA. Body S1A

0.050.050.050.050.050.05

B. Body SIB0.050.050.050.050.050.05

C. Body SIC0.050.050.050.050.050.05

D. Body S1F0.050.050.050.050.050.05

PercentSodiumSilicate

0.050.0750.1000.1250.1500.175

0.050.0750.1000.1250.1500.175

0.050.0750.1000.1250.1500.175

0.050.0750.1000.1250.1500.175

PercentWater

Retention

17.7016.8516.6016.2515.9015.85

17.617.116.516.416.516.4

17.517.016.416.516.616.4

17.516.916.416.516.616.75

Drainage

GoodGoodExcellentExcellentExcellentExcellent

FairGoodGoodExcellentExcellentExcellent

FairFairGoodGoodExcellentExcellent

FairGoodGoodGoodExcellentExcellent

TypeCast

SoftGoodGoodVery GoodVery GoodVery Good

SoftSoftGoodGoodExcellentExcellent



able dielectric strengths:


Test specimens were formed by casting 7 in.lengths of % in. diam rods in plaster molds. Setsof 15 specimens were fired to six temperatures inan electrically-heated kiln at a firing rate of 75deg F per hr to 2000 deg F and 50 deg F thereafteruntil the desired temperatures were reached; a 2-hr soak was maintained at the peak temperatures.No cooling control was attempted.

ResultsThe results of casting trials made with Series I

sanitary ware bodies given in Table 12 show that

(a) Bodies Containing Potash Feldspar

8

4

2\ k

6 \ -- -- - - V-------- ---- -

^ ^\^ _____

Q

!Y

-_I.~~~ ___ i 1I

Custer and Buckingham feldspars may be usedinterchangeably with very little change in the cast-ing characteristics of the body slip. The use of F-4feldspar as a replacement for a potash feldsparwould cause slightly more rapid casting rates at anequivalent electrolyte level. The best casting slipsof the series were found to be those containing A-3feldspar.

Figure 12 presents the results obtained from thephysical tests on fired specimens of sanitary warebodies. In Series I, the data show that the body

(bl Bodies Confaining Intermediate Alkali Feldspar

___ _\N-%.\ •'-. .

Deg F 2/50 2200 2250 2300 2350 2/50 2200 2250 2300 2350Firing Temperature

Fig. 12. Fired Properties of Sanitary Ware Bodies

i I I I I I ;


containing F-4 feldspar achieved maturity in therange of cone 93 to 96. Bodies containing Custeror Buckingham feldspar matured at a slightlyhigher temperature, between cone 96 and cone 103.The body containing A-3 feldspar was found tomature in the range of cone 103 to cone 106. Themaximum moduli of rupture in this series werefound at temperatures slightly above the point atwhich zero porosity was indicated.

The tests indicated that more shrinkage maybe anticipated with F-4 feldspar than with equiva-lent amounts of high potash feldspars such asBuckingham or Custer. The ultimate transversestrengths of the potash feldspar bodies were some-what greater than those obtained with F-4 feldsparbodies, although F-4 feldspar bodies showed higherstrengths at equivalent firing temperatures up tothe point of overfiring. It should be noted, however,that at equivalent firing temperatures the porosityof the F-4 feldspar body was lower.

Results obtained with Series II and Series IIIbodies illustrated the effective fluxing action in-duced by small additions of tremolitic talc. Incompositions containing 32 percent A-3 feldspar, a2 percent talc addition reduced the maturing tem-perature range from cone 103-106 to cone 93-9',while a 4 percent addition of talc effected maturityin the range of cone 76-83. In a like manner, 2 per-cent talc, added to similar bodies containing Buck-ingham or Custer feldspars, reduced the maturingranges from cone 96-103 to cone 86-93, and 4 percenttalc caused maturity at cone 76-83. The maturingranges of F-4 feldspar bodies were reduced fromcone 9-9cone 93 tone 86-93 and cone 73-76 by additionsof 2 and 4 percent tremolitic talc. The addition ofauxiliary flux material apparently had little effecton the ultimate transverse strength of the bodiesinvestigated but caused a slight increase in shrink-age.

Bodies of reduced feldspar content (reducedfrom 32 to 28 percent) indicated that the 4 percentreduction raised the maturing temperature byslightly more than one cone. A 2 percent talc addi-tion to bodies containing 28 percent feldsparaffected maturity in about the same temperaturerange as those bodies containing 32 percent feld-

spar with no auxiliary flux additions. The reducedamount of feldspar was observed to yield bodiesof slightly lower ultimate strength but of similarshrinkage.

18. Floor Tile

CompositionFour bodies of a representative floor tile compo-

sition were prepared using the four commercialfeldspars A-3, Buckingham, Custer and F-4 as totalflux contents. The specific compositions are shownin Table 13.

Con

Body No.A-3 FeldsparBuckingham FeldsparCuster FeldsparF-4 FeldsparVictoria Ball ClayKamec KaolinE.P.K. Florida KaolinTremolitic Talc No. 1Ottawa Flint

Table 13nposition of Floor Tile Bodies

FT1A FT1B55

55

FT1C FTIF

55

Processing and FiringThe floor tile bodies were prepared by dry mix-

ing with 8 percent of water. The mixed materialswere aged for 24 hr and granulated through a 10-mesh screen, after which specimens were formedby pressing at 2000 psi. Specimens consisted of 1 in.square bars 6 in. in length. Firings were made in anelectric kiln at a temperature rate of 100 deg Fper hr to 2120, 2150, 2170, 2200, and 2230 deg F.These temperatures represented pyrometric coneequivalents of 46, 56, 65, 76, and 86.

ResultsThe results obtained from tests on the floor tile

bodies are shown in Fig. 13. The data show a rapiddrop in the porosities of all bodies in the tempera-ture range of cone 4 to cone 8. The body containingF-4 feldspar reached vitrification at a temperatureslightly above cone 76. The potash feldspar bodiesachieved vitrification at cone 86, while the A-3feldspar body was still incompletely vitrified atthat temperature.

The shrinkages obtained with all bodies wereapproximately of the same magnitude; this wasalso observed to be the case with the modulus ofrupture values.

55


2\2 K. FTIA

\". \ FrIR

5

4

3

Cone

Deg F 2

SI T 1 I 1

% . . ........ .. -

__FT/F

8•

6--

4

2

n - --- - --- ^ -'*

'00 2150 2200

Firing Temperature

Fig. 13. Fired Properties of Floor Tile Bodies

F

B

b

i

IV. SUMMARY OF RESULTS

The results of this extensive study on the useof high potash and intermediate alkali feldspars ina wide range of whiteware bodies may be summa-rized as follows:

1. Intermediate alkali feldspars may be usedinterchangeably with high potash feldspars in alltypes of whiteware bodies with minor changes inthe firing characteristics and properties.

2. In whiteware bodies of low feldspar content,such as semi-vitreous bodies, intermediate alkalifeldspars with low free quartz content may beinterchanged with high potash type feldspars withlittle or no change in the firing characteristics orfired properties if no other fluxes are present.

3. In whiteware bodies of moderate and highfeldspar contents, such as electrical porcelain, sani-tary ware and floor tile, an intermediate alkalifeldspar such as F-4 may be substituted for potashfeldspars to effect maturity at a lower temperaturewith only a slight reduction in strength and a smallincrease in shrinkage.

4. Feldspars of high potash content may be usedinterchangeably at all levels with small differencesin the firing characteristics of bodies.

5. Whiting and talc are beneficial in reducingthe maturing temperatures of whiteware bodieswhen added in amounts of 2 to 4 percent.

6. Whiting is a more active flux in bodies con-taining high potash feldspars than in similar bodiescontaining intermediate alkali feldspars.

7. Talc is a more active flux in bodies contain-ing intermediate alkali feldspars than in similarbodies made with high potash feldspars.

8. The over-all fluxing action of talc was foundto be more effective than equivalent amounts ofwhiting in bodies containing either type of feldspar.

9. Magnesite was found to be a powerful fluxwhen added in amounts of 2 percent, but causedappreciable increases in the fired shrinkage.

10. The addition of auxiliary fluxes was foundto shorten the firing range of bodies in which theywere used.

11. Considerable cristobalite formation was ob-served in semi-vitreous bodies containing inter-mediate alkali feldspars and MgO-bearing fluxessuch as talc and magnesite. Little or no cristobalitewas found in similar bodies made with high potashfeldspars.

12. Some cristobalite formation was found inall whiteware bodies containing intermediate alkali-feldspars.

13. X-ray studies revealed that the formationof cristobalite took place during the cooling cyclewith crystal deposition of SiO, or devitrification.

APPENDIX

19. Raw Materials Used in Experimental BodiesThe chemical analyses of the feldspars used in

the experimental bodies are shown in Table 14. Theanalyses of other materials used in making variousbody compositions are also given in Table 14.

20. Details of Body Preparation

BlungingBlunging was used for the blending of raw

materials in preparation of semi-vitreous, electricalporcelain and sanitary ware bodies. The blungingequipment consisted of a 20-gal tank with two setsof paddles which were rotated in opposite directionsto create turbulent counterflow blunging. Approxi-mately 12 gal of water was used for the blungingof bodies; 120 lb of dry weight of each body com-position was slowly added to the water with thepaddles rotating. The materials of each batch wereadded in the order of decreasing plasticity; the ballclays were added first, followed by the kaolins andthen the non-plastic materials. Sufficient time wasallotted between the various additions to allow theslip to become smooth. After all materials had beenadded, the slip was allowed to blunge for approxi-mately 1 hr and then passed through a 120-meshlawn and over a set of magnets to remove anyparticles of iron. The processed slip was stored insmooth blungers for filter pressing.

Ball MillingBall milling was used to prepare the hotel china

bodies. A 35-gal Patterson direct motor-driven ballmill was used. The mill charge consisted of 120 lbof dry batch and 10 gal of water. Each batch wasmilled for a period of 4 hr after which it waspassed through a 120-mesh lawn and over magnets.The ball-milled batches were placed in smoothblungers for storage until ready for filter pressing.

Dry MixingDry mixing was employed for the blending of

raw materials in the preparation of floor tile bodies.Twenty-five lb of dry batch was added to a No. OSimpson Muller-mixer and allowed to mix for 5min, after which 1 qt of water was added and

mixed an additional 15 min. The water additionmade was approximately 8 percent of the dryweight of the body. The mixed material was placedin sealed jars and aged for 24 hr, after which itwas granulated through a 10-mesh screen and re-placed in the storage jars until ready for use inpressing.

Filter Pressing and PuggingAll of the bodies prepared by blunging or ball

milling were filter pressed and subsequently pugged.The processed slips of each body were pumped fromthe smooth blungers and filter pressed to a pressureof 140 psi for 45 min; the resultant filter cakescontained approximately 20 percent of water. Thefilter cakes of each body composition were puggedinto slugs 3 in. in diam by 12 in. long and placedin sealed damp jars for one week of aging. Thepugging operation was used to assure a uniformmoisture distribution within each body and also tofacilitate the storage during the aging period.

21. Specimen Formation

ExtrusionSpecimens for the testing of semi-vitreous, elec-

trical porcelain and hotel china bodies were pre-pared by vacuum extrusion with an InternationalVac-aire De-airing Extrusion Machine. Specimenswere extruded as round rods 7 in. in diam; then80 rods cut 7 in. in length and 40 rods cut 3% in.in length for each body composition. The specimenswere allowed to air dry in V-shaped troughs andthen placed in an oven at 220 deg F to completethe drying.

CastingTest specimens of sanitary ware bodies were

made by casting in plaster of Paris molds. Previousto the preparation of the casting slips of each body,six small slip batches were prepared from eachbody composition to determine the casting char-acteristics. Each trial slip batch was prepared fromslugs which had been aged for one week. The watercontent of each body slug was determined by thedifference in weight before and after thorough dry-


ing of 100 grams of plastic slug. The percent mois-ture content based on the weight of dry solids wasobtained as follows:

% Moisture (dry basis)Plastic Weight - Dry Weight 100

- ------- T^ -- vr -* T" -------- X 1UUDry Weight

The weight of plastic material equivalent to 1000grams of dry body was calculated as follows:

1000 (1)÷ m1000 (i + '

%1 ) grams

It was desired to prepare slips with a specific grav-ity of approximately 1.80; by experiment it wasfound that when the solid to water ratio was ap-proximately 72 to 28 a specific gravity of slightlymore than 1.80 was obtained. Slips were preparedwith this solid to water ratio. The electrolyte con-tents of the trial slips were based on the dryweight of solids and were as follows:

PercentSodium Carbonate

0.050.050.050.050.050.05

PercentSodium Silicate

("N" Brand)

+ 0.05+ 0.075+ 0.100+ 0.125+ 0.150+ 0.175

The electrolytes were added to the water re-quired for each body slip and thoroughly dispersedbefore adding any of the plastic body. A smallmotor-driven propeller type mixer was used formixing the plastic body into the electrolyte solu-tion. The trial slips were mixed to a smooth con-sistency, passed through a 120-mesh screen andadjusted to a specific gravity of 1.80. The slips werethen transferred to Mason jars which were placedin a vacuum chamber for de-airing of the slip. Afterde-airing, the jars of slip were set in a tumblingapparatus and slowly tumbled end over end for48 hr.

At the end of the tumbling period, the specificgravity of each slip was determined by weighing100 cc of slip. The flow properties of each weredetermined by the time required for 100 cc of slipto flow through a % in. diam orifice of a Mariotteflow tube. Three determinations were made witheach slip sample.

The relative casting rates of the various slipswere determined by casting two 1 in. square bars

and allowing each slip to "set up" for 40 min afterwhich time the molds were inverted and allowed todrain for 5 min. The molds were opened after theend of the draining period and the cast specimensremoved; the nature of drainage and type of castwas noted and a 7 in. length of each cast bar wascut and immediately weighed to obtain the wet castweight. The cast specimens were dried completelyand reweighed to obtain the dry cast weight. Thepercent water retention of the cast specimens wascalculated as follows:

Percent Water RetentionSWet Cast Weight - Dry Cast Weight

Wet Cast Weight

The molds used for the casting trials and all othercasting were prepared from a mixture of 65 per-cent gypsum plaster and 35 percent water.

From the casting trials described above, it wasfound that an electrolyte addition of 0.05 percentsodium carbonate + 0.10 percent of "N" Brandsodium silicate, based on the dry weight of thebody, was suitable for the casting of all body com-positions. Using this electrolyte content, the plasticbodies were made into slips of 1.80 specific gravity

Chemical Analyses of

SiO2

TremoliticTale No. 1* 56.72Sierramic LowLime Talet 59.62Wash. DeadBurnedMagnesitel 6.7E.P.K. Fla.Kaolin¶ 46.75KamecKaolin§ 46.18PioneerKaolin** 45.34Tenn. No. 5Ball Claytt 46.85Old Mine No. 4Ball Claytt 51.65H.T.P.Ball Claytt 50.15Martin No. 5Ball Claytt 60.72Champion-ChallengerBall Claylt 54.08VictoriaBall Clay¶¶ 58.44ImperialBall ClayI¶ 54.24Royal BallClay¶¶ 56.06OttawaFlint§§ 99.6

Table 14

Raw Materials Used in Experimental BodiesA120l Fe2Oa CaO MgO KNaO TiOs

0.70 0.18 4.80 30.81 0.50

2.05 0.94 0.91 29.91 0.49

1.8 3.5 5.0 82.7

36.75 0.80 0.80 0.20 0.24

Ing.Loss

5.95

5.98

0.5

14.95

38.38 0.57 0.37 0.42 0.68 0.04 13.28

37.29 0.61 0.25 0.22 0.45 1.54 14.39

36.15 2.04 0.50 0.40 0.71 16.48

31.24 1.17 0.20 0.50 0.94 1. 72 12.13

34.17 1.06 0.12 0.08 1.33 1.21 11.85

25.53 0.74 0.08 tr 2.12 1.39 9.35

28.90 1.06 0.14 0.20 0.49 1.74 13.25

25.89 0.84 0.36 0.15 0.51 1.60 11.98

26.84 0.86 0.53 0.42 0.65 1.70 14.90

27.61 1.12 0.33 0.56 1.58 1.62 11.10

0.10 0.017 0.02

* W. H. Loomis Talc Corp., Gouverneur, N.Y.t Sierra Tale Co., Los Angeles, Calif.SHarbison-alker Refractories Co., Pittsburgh, Pa.Edgar Plastic Kaolin Co., Metuchen, N.J.

§ Harris Clay Co., Spruce Pine, No. Car.** Georgia Kaolin Co., Elizabeth, N.J.

i Kentucky Tennessee Clay Co., Mayfield, Ky.Spinks Clay Co., Paris, Tenn.

¶I United Clay Mines Corp., Trenton, N.J.§ Ottawa Silica Co., Ottawa, Ill.

0.02 0.1


using the same type calculations as described forthe casting trials. The slips were mixed for 4 hrand allowed to age in stoneware jars for 24 hr,after which they were remixed for 30 min andpassed through a 120-mesh lawn. De-airing of theslips was accomplished in a vacuum mixer.

Specimens were cast into rods % in. in diam and7 in. in length. Plaster of Paris molds were usedand the specimens were allowed to cast solid. Uponremoval from the molds, eighty 7 in. lengths andforty 31/2 in. lengths were cut and allowed to airdry for 3 days; drying was completed at 200 degF in an oven dryer.

Dry PressingFloor tile specimens were prepared by dry

pressing using a Denison "Hydroilic" press.Seventy-five bars 1 in. square by 6 in. long werepressed from each body composition at a pressureof 2000 psi.

22. Test ProceduresVolume Shrinkage

The firing volume shrinkages of all bodies ex-cept those of floor tile were obtained with the 3%in. length specimens. Prior to firing, the dryvolumes of these specimens were obtained from thesaturated and suspended weights in kerosene. Eachpiece was placed in kerosene until saturated; dryvolume was calculated as shown below:

Dry VolumeSaturated Weight - Suspended Weight

Specific Gravity of Kerosene

After firing, the fired volumes were obtained ina similar manner using the saturated and suspendedweights in water. The volume shrinkage was de-termined as follows:

Percent Volume ShrinkageDry Volume - Fired Volume

Dry VolumeThe reported values of volume shrinkage are theaverage of five specimens.

The volume shrinkages of floor tile bodies werecalculated from linear shrinkage determinations.Linear shrinkage was calculated from the equa-tion:

Percent Linear ShrinkageDry Length - Fired Length X 100- ----- T -- T - n ----- ^ 100

Dry Length

The conversion from linear to volume shrinkage isdetermined as follows.

Percent Volume Shrinkage= 100 [( 1 -- l)-1],

where a equals the percent linear shrinkage.

Fired Modulus of RuptureThe modulus of rupture values reported were

obtained as the average of a minimum of 12 speci-mens. Specimens were stored in dessicators imme-diately after firing to prevent any moistureabsorption. A Dillon Dynamometer with a 5 in.span between knife edges was employed for thedetermination of the cross-breaking load.

If P is the cross-breaking load in lbs, 1 is thespan between knife edges in in. and d is the diamof a round specimen at the point of fracture; thenthe modulus of rupture (M) is calculated as:

8P1M - ps'

(for specimens of round crosssection)

For square bars, such as the floor tile specimens,

3P1M = 3 p1M 2b d2 p s

(where b and d are the breadthand depth of the square bar atthe point of fracture)

Fired PorosityThe fired porosities were determined on the

fractured portions of modulus of rupture specimens.The fractured portions were thoroughly dried in anoven at 220 deg F and allowed to cool to roomtemperature in a dessicator. The weights of thedry pieces were determined after which the pieceswere placed in water which was brought to a boiland allowed to boil for 8 hr. After cooling to roomtemperature, the specimens were removed from thewater, wiped with a damp cloth to remove anyexcess moisture and weighed. The suspendedweights of the saturated pieces were also deter-mined. The porosity was calculated as follows:

Percent Porosity =Fired Saturated Weight-Fired Dry Weight

Fired Saturated Weight - Fired Suspended WeightX 100

X-Ray AnalysisThe powder diffraction method was used for

qualitative and quantitative analyses of the crys-talline content of fired specimens. This study re-quired a Norelco Geiger-Counter X-ray Spectrom-eter with automatic recorder.


The method of sample mount was that proposedby McCreery. (51 The sample holder consisted ofan alloyed aluminum cell, 1% in. x 1 in. x % in.with a %6 in. hole at the center.

Samples for X-ray analysis were obtained bycutting a % in. x 1/2 in. section from the centers ofspecimens. These segments were crushed and groundin an agate mortar to pass a 325-mesh screen. Ironimpurities were removed by passing a permanentmagnet through the finely ground material. Samplesdesignated for quantitative analysis were thor-oughly dried after grinding and 1.000 gram of eachwas mixed with chemically-precipitated calciumfluoride in an agate mortar with 5 ml of methylalcohol. The alcohol was removed by evaporationand the samples were mixed dry for an additional10 min.

Standard samples for the determination of cali-bration curves were prepared from Ottawa flint,electric furnace mullite, prepared cristobalite andcalcium fluoride with fused F-4 feldspar glass as adiluent. Considerable difficulty was encountered inanalyzing the mullite content because of its tend-ency toward preferred orientation.

Quantitative analyses were made according tomethods described by Tuttle and Cook (5 2 ) in whichthe ratio of the height of a peak to the height ofthe internal standard peak (calcium fluoride) wascompared to the calibration ratio curve obtainedwith standard mixtures. Peak intensities were alsodetermined by the counting method and also bymeasurement of peak heights on the recorder chartoperating at 10 (2 theta) per min. Using the chartmethod, three runs for each of three mixings ofeach sample were made.

The major quartz and mullite peaks at d =3.35 and 3.43 respectively could not be used due totheir juxtaposition and the tendency of each toreinforce the other. In most cases the mullite peakcould not be separated from the more intensequartz peak. The peaks used in the analyses werethose found at d-values of 4.24 for quartz, 4.05 forcristobalite and 2.70 for mullite.

Both copper and iron radiation were used forthe X-ray analyses, but iron radiation was foundto give more resolution and also shifted the peaklocations to higher angles on the goniometer arcwhere the background interferences were lower.

BIBLIOGRAPHY

1. Rune Hedin, "Colorimetric Methods for Rapid Anal-ysis of Silicate Materials," Swedish Cement and ConcreteRes. Inst., Proc. N: R8 (1947).

2. G. G. Marvin and L. B. Woolaver, "Determinationof Sodium and Potassium in Silicates," Ind. and Eng.Chem., Anal. Ed., 17, p. 554 (1945).

3. P. J. Elving and Po Chuan Chao, "Determination ofAlkali Metals in Silicates and Similar Materials," Anal.Chem., 21, p. 507 (1949).

4. A. L. Conrad and W. C. Johnson, "Flame PhotometerTechniques," Anal. Chem., 22, p. 1530 (1950).

5. J. L. Gilliland, "Applications of Flame Photometryfor the Analysis of Alkalies in Silicates, Waters andMetals," A.S.T.M. Spec. Bull. 116, p. 33 (1951).

6. A. S. Watts, "Specifications for Commercial Feldsparand Flint," Jour. Am. Ceram. Soc., 3, p. 722 (1920).

7. Commercial Standards CS23-30, National Bureau ofStandards (Oct., 1930).

8. C. W. Parmelee and C. R. Amberg, "Comparison ofBodies Containing Blended Feldspars and One-Mine Feld-spar of Similar Composition," Jour. Am. Ceram. Soc., 14,p. 309 (1931).

9. E. W. Koenig, "Calculation of Mineralogical Compo-sition of Feldspar by Chemical Analysis," Jour. Am. Ceram.Soc., 25, p. 420 (1942).

10. N. L. Bowen and O. F. Tuttle, "The SystemNaAlSiAOs- KAlSiaOs- HO0," Jour. Geol., 58, p. 489 (1950).

11. Felix Chayes, "Notes on the Feldspars: On the Re-lation Between Anorthite Content and y-Index of NaturalPlagioclase," Jour. Geol. 58, p. 593 (1950).

12. W. L. Bragg, Atomic Structure of Minerals, pp. 230-50, Cornell Univ. Press, London (1937).

13. Brain Mason, Principles of Geochemistry, pp. 89-95,John Wiley and Sons, Inc., New York (1952).

14. K. Rankama and Th. G. Sahama, Geochemistry, pp.131-40, Univ. of Chicago Press, Chicago (1952).

15. S. H. Chao, A. Hargreaves and W. H. Taylor, "TheStructure of Orthoclase," Mineralog. Mag., 25, p. 498 (1939).

16. O. F. Tuttle and N. L. Bowen, "High TemperatureAlbite and Contiguous Feldspars," Jour. Geol., 58, p. 572(1950).

17. J. W. Greig and B. T. W. Berth, "The SystemNazO A120Ls2SiO2 (Nephelite, Carnegieite) - NasO-A120h6SiO (Albite)," Am. Jour. Sci., 35A, p. 93 (1938).

18. J. F. Schairer and N. L. Bowen, "Melting Relationsin the Systems NazO-Al20SSiO 2 and K2-Al20S'SiO2," Am.Jour. Sci., 245, p. 193 (1947).

19. G. W. Morey and N. L. Bowen, "The Melting ofPotash Feldspar," Am. Jour. Sci., 5th Ser., 4, p. 1 (1922).

20. A. L. Day and E. T. Allen, "The Isomorphism andThermal Studies of Feldspars, Part I: Thermal Studies,"Carnegie Inst. of Wash., 31, p. 1 (1905).

21. G. A. Rankin and F. E. Wright, "The Ternary Sys-tem CaO-AlO-SiO2," Am. Jour. Sci., 4th Ser., 39, p. 1(1915).

22. J. F. Schairer, "The Alkali-Feldspar Join in theSystem NaAlSiO,-KAISiO,-SiO,," Jour. Geol., 58, p. 512(1950).

23. R. R. Franco and J. F. Schairer, "Liquidus Tempera-tures in Mixtures of the Feldspars of Soda, Potash andLime," Jour. Geol., 59, p. 259 (1951).

24. W. M. Cohn, "Viscosity Measurements in FeldsparMelts at High Temperatures," Ber. Deut. Keram. Ges. 15(11) 551 (1934), Am. Ceram. Soc. Abstr., 14, p. 152 (1935).

25. R. F. Geller, "Bureau of Standards Investigation ofFeldspars: Second Progress Report," Jour. Am. Ceram.Soc., 10, p. 411 (1927).

26. C. B. Tauber and A. S. Watts, "A DeformationStudy of Mixtures of the Deformation-Eutectics of Potash-Soda Feldspars-Cao-AlhO3-SiOz and MgO-AlOi-SiO 2, Bothin the Raw and Prefused State," Bull. Am. Ceram. Soc., 31,p. 458 (1952).

27. J. A. Martz, "Potash-Soda-Lime Feldspar EutecticStudy," Jour. Am. Ceram. Soc., 16, p. 299 (1933).

28. R. Riecke and L. Mauve, "Softening of FeldsparGlass and Influence of Different Additions Thereon," Ber.Deut. Keram. Ges., 12 (8), p. 379 (1931), Am. Ceram. Soc.Abstr., 11, p. 188 (1932).

29. C. R. Amberg and J. L. Gallup, "Melting Relationsof Potash Feldspar, Soda Feldspar and Flint Mixtures,"Jour. Am. Ceram. Soc., 14, p. 733 (1931).

30. J. M. Teague, Jr., and A. S. Watts, "A Study ofDeformation-Eutectic Development in Mixtures of Feld-spar and CaCO,, MgCOs, BaCO0 and ZnO," Bull. Am.Ceram. Soc., 31, p. 457 (1952).

31. C. W. Parmelee and C. R. Amberg, "Solubility ofQuartz and Clay in Feldspar," Jour. Am. Ceram. Soc., 12,p. 699 (1929).

32. A. S. Watts, "Crystalline Structure in Porcelains,"Trans. Am. Ceram. Soc., 11, p. 185 (1909).

33. A. H. Fessler and H. Navratiel, "The Effect ofDifferent Fluxing Oxides on the Constitution of Porcelain,"Jour. Am. Ceram. Soc., 13, p. 332 (1930).

34. F. P. Hall, "The Influence of Chemical Compositionon the Physical Properties of Glazes," Jour. Am. Ceram.Soc., 13, p. 182 (1930).

35. R. F. Geller and A. S. Creamer, "Investigation ofFeldspar and Its Effect in Pottery Bodies," Jour. Am.Ceram. Soc., 14, p. 30 (1931).

36. I. C. McDowall and W. Vose, "Feldspars and TheirUse in Ceramic Bodies of the Type Feldspar-Clay-Quartz,"New Zealand Pottery and Ceramic Res. Assoc. Inc., Tech.Rpt. 6 (May 1950).

37. C. W. Parmelee and A. J. Monack, "The Solubilityof Feldspars in Water," Jour. Am. Ceram. Soc., 13, p. 386(1930).


38. P. P. Budnikov and Kh. O. Gevorkyan, "Theory ofthe Formation of a Porcelain Body," Jour. Applied Chem.U.S.S.R., 24, p. 141 (1951), Chem. Abstr., 46, p. 4189 (1952).

39. G. A. Kovelman, "Vitrification of Feldspar Duringthe Firing of Ceramic Masses," Stekolnaya i Keram. Prom.,No. 7-8, p. 11 (1946), Am. Ceram. Soc. Abstr., 31, p. 105(1948).

40. H. M. Kraner, "Importance of the Glass Phase inPorcelain," Jour. Am. Ceram. Soc., 12, p. 383 (1929).

41. R. F. Geller, D. N. Evans and A. S. Creamer,"Effects of Particle Size of Potter's Flint and a Feldsparin Whiteware," Bur. Stds. Jour. Res., 11, p. 327 (1933).

42. E. A. Durbin and A. S. Watts, "Effect of GrainSize and Firing Schedule on Properties of Feldspar as aVitrifying Agent," Ceramic Age, 26, p. 135 (1935).

43. F. H. Norton and F. B. Hogdon, "The Influence ofTime on the Maturing Temperature of Whiteware Bodies,"Jour. Am. Ceram. Soc., 14, p. 177 (1931).

44. S. C. Sane and R. L. Cook, "Effect of Grinding andFiring Treatment on the Crystalline and Glass Contentand the Physical Properties of Whiteware Bodies," Jour.Am. Ceram. Soc., 34, p. 145 (1951).

45. R. A. Snyder, "Progress of Fluidity in PorcelainBodies," Jour. Am. Ceram. Soc., 14, p. 855 (1931).

46. A. V. Bleininger and P. Teetor, "The Viscosity ofPorcelain Bodies," Trans. Am. Ceram. Soc., 15, p. 328(1913).

47. W. R. Morgan, "Relation Between UncombinedQuartz and Thermal Expansion of Ceramic Bodies," Jour.Am. Ceram. Soc., 17, p. 117 (1934).

48. G. A. Loomis and A. R. Blackburn, "Use of SodaFeldspar in Whiteware Bodies," Jour. Am. Ceram. Soc.,29, p. 48 (1946).

49. G. R. Loomis, "Grain Size of Whiteware Clays asDetermined by the Andreasen Pipette Method," Jour. Am.Ceram. Soc., 21, p. 393 (1938).

50. R. Russell, Jr., and R. S. Weisz, "Some ParticulateCharacteristics of Commercial Flint and Feldspar," Jour.Am. Ceram. Soc., 27, p. 1 (1944).

51. G. L. McCreery, "Improved Mount for PowderedSpecimen Used on the Geiger-Counter X-ray Spectrometer,"Jour. Am. Ceram. Soc., 32, p. 141 (1949).

52. M. A. Tuttle and R. L. Cook, "Fundamental Studyof Crystalline and Glassy Phases in Whiteware Bodies,"Jour. Am. Ceram. Soc., 32, p. 279 (1949).

the properties of feldspars and their use in whitewares, · semi-vitreous ware 15 15. hotel china...

Documents