THE SYSTEM DIOPSIDE-ANORTHITE-AKERMANITE

A DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Dep-ree Doctor of Philosophy in the Graduate School of the Ohio State University

' '' E. Christiaan de Wys, A.B., M.A.

The Ohio State Universitv 1955

Approved by:

0 Advisor department of Kineralopy A c knowled gment s

.Vith deep gratitude the writer acknowledges his indebtedness to Dr. W.R. Foster, without whose constant encouragement, and

friendly guidance completion of this investigation would not have been possible. The writer also wishes to express his appreciation to the faculty of the Department of Mineralogy of Ohio State University,

for the research facilities afforded and training received. The author especially wishes to extend his thanks to his wife, J.B.N. de Wys,

for her patience, constant assistance, and typinr of the dissertation.

Thanks are due also to H.H. Uotila for the final preparation of the diagrams. Table of Contents

page

Acknowledgements ii

List of Tables iv

List of Illustrations v

Introduction 1

Some Occurences of the Minerals 5

Characteristics of the Minerals 11

Experimental Procedure 20

Thermal Data 24

Thermodynamic Analysis of the System Anorthite-Akermanite 39

Discussion and Conclusions 45

Bibliography 49

Autobiography 57

iii List of Tables

Table No. paye

1 Optical data of CaAl2Si20g 12

2 Melting point of anorthite 14

3 Heat of fusion of anorthite 15

4 Meltinp point of diopside 16

5 Melting point of akermanite 18

6 Heat of fusion of akermanite 18

7 Thermal data for the system anorthite-akermanite 31

8 Thermal data for the system diopside-anorthite-akermanite 36 List of Illustrations

Figures page

1 The lime-magnesia-silica tetrahedron 4

2 Index of refraction chart for glasses of the system

anorthite-akermanite 22

3 Index of refraction chart for glasses of the system

diopside-anorthite-akermanite 23

4 Equilibrium diagram of the system diopside-anorthite 26

5 Equilibrium diagram for the system diopside-akermanite 27

6 Equlibrium diagram for the system anorthite-akermanite 28

7 Equilibrium diagram for the system diopside-anorthite-

akermanite 33

8 Equlibrium diagram for the system diopside-anorthite-

akermanite on the basis of the data of Osborn et al.

(1954) and Prince (1954) 35

9 Comparison of theoretical and experimental liquidus

curves for the system anorthite-akermanite 40

v Introduction

This study presents the results of an investigation of the high temperature phase equilibrium relationships in the binary system

anorthite (CaAl2 Si2 0 g) - akermanite (Ca2 MgSi2 0 y) and in the ternary system diopside (CaMgSigO^) - anorthite (CaAl2 Si2 0 g) - akermanite

(Ca2MpSi20y). The well known quenching technique was employed. The two sjrstems lie within the quaternary system CaO-MgO-A^O-j-SiC^. The interior of this four component system is made up of numerous sub­ systems, at least some of which are true binary, ternary and quaternary systems. The present work deals with that portion of the large quaternary system in which the composition can be expressed in terms

of the component-'- Ca0 .MgO.2 Si0 2 , CaO.A^O^. 2 Si0 2 , and 2Ca0.Mg0.25i02«

Such ternary planes serve to partition the interior of the tetrahedron into smaller tetrahedra which can be isolated and investi­ gated individually. If the plane diopside-anorthite-akermanite proves to be ternary it will serve as a partition separating liquids which, during crystallization, must remain on one side of this partition from those which must remain on the other side. This is true whether frac­ tional or equilibrium crystallization takes place. An illustration of the general quaternary system CaO-KgO-AlgO^-SiC^ is presented in

Figure 1. This figure indicates the position of the diopside-anorthite- akermanite plane in the tetrahedron. A hollow, transparent tetrahedron has been used to permit a clearer view of the plane under consideration.

1. Ca0.MgO.2SiC>2, CaO.A^O^.2Si02 and 2Ca0.Mg0.2Si02 are, of course, merely the oxide formula equivalents of CaMgSi^O^, CaA^SigOg, and

Ca2 NpSi0 y respectively.

1 2

The system CaO-MgO-SiOg-A^O-j has considerable petrologic and industrial improtance. The three silicates investigated in this research were chosen since they occur in basic igneous rocks, metamorphic hybrid zones, and in industrial slags. Anorthite and diopside are well known in rocks of common petrologic interest, while all three of the silicates are important in the composition of slags.

One of the purposes of experimental research of this nature has always been to clarify the relationships of mineral associations, of magmatic or metamorphic origin, observed by the petrographer. With the exception of the iron and alkali content the basic magmas may be reason­ ably well expressed in terms of the components GaO, MgO, A^O-^, and SiC^.

In metamorphism the phase assemblages of some of the calcareous-aluminous metamorphosed rocks fall within this quaternary system.

Technologically, some of the industrial glasses, cements, and virtually all usual iron blast furnace slaps are included in the lime- magnesia-alumina-silica system, according to McCaffery et a l . (1927),

Koch (1933), Osborn et al. (1954), and Prince (l95l)« Since apparently the properties of slag are not mere composites of the properties of

Si02, A^O-^, MgO, and CaO, but rather of the mineral compounds present

(McCaffery et al.), an investigation of the interrelationships of some of the minerals should be useful. According to Martin and Derpe (1943) a knowledge of possible structural units in the slaps would clarify greatly the nature of chemical reactions in such fluid, and would help to explain reactions between the slag and metal. Such knowledge would facilitate a greater control of slag chemistry. Accordingly, a thermo­ dynamic analysis of the system anorthite-akermanite was undertaken in 3 an attempt to formulate possible structural units present in the liquid phase,

A number of investigators (McCaffery et al,, 1927), (Janecke, 1933),

(Nurse and Midgley, 1951), (Osborn et al,, 1954), (Prince, 1954) have indicated that anorthite, pyroxene, and melilite form a stable equilibrium phase assemblage. But no detailed investigation of the liquidus relations for the diopside-anorthite-akermanite system has ever been carried out. There has been no clear indication as to whether anorthite can coexist with pure akermanite. Since the stability of akermanite below a certain temperature has been seriously questioned

(Carstens and Kristofferson, 1931), (Bowen et al,, 1933), (Osborn and

Schairer, 1941), (Neuvonen, 1952), it appeared desirable to consider this matter in some detail. u

A I2 ° J

0 Mu 9 Si

Co

lSIO.

Ak Fo Mor Mo

CaO MgO

Kir. 1 The lime-marnesia-alumina-silica tetrahedron.

The shaded area represents the anorthite-akermanite-diopside ternary plane. The abbreviations have the following simi- ficance: An - anorthite; D - diopside; Ak - akermanite;

Mu - mullite; Si - sillimanite; Sp - spinel: Co - cordierite; n - <-ehlenite; L - lamite; R - rarkinite; W - vrollastonite;

Mer - merwinite; Mo - monticellite; Fo - forsterite; Cl - elinoenstatite. Some Occurences of the Minerals

Anorthite

This mineral belongs to the important feldspar group which constitutes about 59*5$ of all the igneous rocks (Burgess, 1949)*

It is the lime rich member of the plagioclase, or soda lime, feldspars and is relatively rare compared with the other members of the series.

Impure anorthite is present as an essentiall constituent in such basic igneous rocks as andesite, basalt, diorite, gabbro, and norite (Barth,

1936). Good crystals of anorthite occur at Mount Somma, in Italy, where they occur as glassy crystals in the old lavas in association with spinel and leucite. In India anorthite is found as a gangue of corundum.

It is also observed in the druses of ejected volcanic blocks, as well as in the granular limestones of contact deposits, near Trenlino, Italy.

In Japan it is found in the lavas of the island of Miyake. Some anorthite is found in meteorites (Grout, 1932).

Anorthite also occurs in metamorphic assemblages. Turner (1933) states that anorthite is an antistress mineral. Anorthite is apparently typically found in the amphibolite facies (Turner, 1948), a critical mineral assemblage formed during progressive metamorphism of basic igneous rocks. It is then associated with hornblende. .Viseman (1934) found anorthite associated with the red ^arnets in the epidorites of the Scottish highlands. Grubenmann and Niggli (1924) found anorthite in the kata-metamorphic zone where it is associated with garnets, vesuvianite, calcite, periclase, diopside, hornblende and others.

Anorthite is also present in a variety of technicological products.

In basic slags it is an important constituent (McCaffery et al. (1927), (Koch, 1933). It is sometimes found as scum on stone-ware bodies made of white firing clay (Knizek and Fetter, 1945). Rigby and Green,

(1942) observed the formation of anorthite in fireclay or fireclay prop containing additions of dolomite or asbestos. Hugill and Green

(1939) drew attention to the ease with which anorthite forms whenever fireclay materials are attacked either hy pure lime or by lime-bearing slaps. St. Pierre (1955) noted the formation of anorthite during the firing of bone-china bodies. De Ville (1955) found that anorthite develops in talc whiteware bodies when wollastonite is substituted in whole or in part for talc.

Diopside

Dionside is a member of the important pyroxene group of minerals.

Of the two main reaction series encountered in the basaltic lavas, the pyroxene series constitutes one and plarioclase the other (Bowen).

Consideration of these two series goes far towards explaining the essential features of the crystallization process in the basaltic lavas. Barth (1936) concludes that most natural basalts lie on the boundary curve separating the primary fields of pyroxene and plagio­ clase. Some diopside is found in the granite family as phenocrysts in the aphanites (Grout, 1932). Diopside occurs in the alnoitic rocks at Isle Cadieux in Canada (Bowen, 1922). There it is associated with biotite, olivine, melilite, and oerovskite. It is also found in the melilite-nephelite basalts of Moilili-Oahu, Hawaiian Islands (3owen,

1923). Dionside associated with wollastonite is found in Essex County,

New York. This diopside is used for ceramic purposes (Amberr, 1945). 7

Metamorphism of silica-containing dolomites generally results in the formation of diopside (Grout, 1932), (Bowen, 1940). In his account of contact metamorphism of marly and pelitic sediments of the Oslo district, Goldschmidt (1911) states that diopside is associated with

Grossularite, wollastonite, anorthite and andalusite. Eskola (1922) observed a diopside limestone in the pre-Cambrian limestone formations of Finland. Daly (1933) mentions the diopsidation of a granite in

Idaho. Turner (1948) observed the metasomatic introduction of CaO into recently emplaced small intrusives of rranite, pegmatite and syenite had apparently caused the development of diopside in granites by the replacement of some of the primary igneous minerals. He be­ lieves, however, that in general when granitic rocks contain diopside or wollastonite, these minerals originate by primary crystallization from magmas that have absorbed limestones.

Akermanite and the genesis of the Melilites

The mineral akermanite is a member of the melilites, a group of tetragonal minerals appearing in igneous rocks, in metamorphic assembl­ ages and in crystallized blast furnace slags. Akermanite was first de­ scribed in slags by Vogt (1842) long before its discovery in nature by

Zambonini (1910) in rocks at Vesuvius.

In igneous rocks the melilites are confined to magmas of a highly basic nature, in which olivine is always present in abundance and in which the silica content scarcely rises above 35p (Shand 1947). Rocks of such a t'npe are rather rare and accordin" to Grout (1932) do not make up more than O.Ol'o of the igneous rocks. Indications are that assimilation of limestone into magma is an important cause of the 8 appearance of the melilites in igneous rocks. Since melilite is rich in lime and poor in silica, it is not surprising to find among the melilite basalts the highest recorded percentage of CaO (about 27%) and almost the lowest recorded percentage of silica {35% or less).

This in itself tends to point toward a limestone assimilation effect in the basic magmas. Shand (1947) states that the addition of CaCO^ to diabase fused in the laboratory caused the crystallization of melilites.

Field evidence appears to connect melilite formation with lime­ stone. Tilley (1929) examined some melilites occuring at Scawt Hill, Co.

Antrim, Ireland, where dolerite had intruded into a chalk. The dolorite was of the normal type containing essentially diopsidic pyroxene and basic plagioclase withsome olivine. The chalk was almost wholly calcium carbonate. The product formed as a result of the interaction of dolorite and chalk consisted mainly of augite and melilite.

Tilley plotted the composition of the doleritic magma, insofar as its constituent molecules of diopside, anorthite and accessory forsterite are concerned, as a point situated close to the diopside- anorthite .join in the plane diopside-anorthite-forsterite. He believed that the addition of lime to such a composition would then bring the new composition into the field of the melilites.

The original alnoite, a biotite-melilite basalt, occurs at Alno,

Sweden, where there is an intimate mingling of carbonate and magmatic rocks (Shand, 1947). In Germany, at Wartemberg, the occurance of melilite-basalts is closely related to the distribution of thick

Jurassic limestones (Becker, 1907). Scheumann (1922) noted that the melilite rocks of the Polzen district, in Bohemia, are also in a region of thick limestones. Cross and Larsen (193?) both noted that f

9 uncompahgrite, a rock consisting up to 66$ of melilite is found in close contact with limestone. The Ohaite of Quebec, another predominantly melilite rock, is in contact with crystalline limestone (Stansfield,

1923). A pipe of alnoite near Avan, Missouri, cuts dolomite

(Singewald and Milton,1930). Daly (1933) mentions that the 1883, 1886,

1892, and 1910 lavas from Mount Etna absorbed limestone with resulting crystallization of melilite.

But not all occurances of melilites show such a close limestone association and this has led to conflicting opinions as to their origin.

In South Africa there are numerous basalts which do not appear to be clearly related to limestones. Shand (1947) observed in the district of Namaqualand one of the preatest developments of melilite basalts in the world. It occurs in a gneiss which locally contains only insignificant lenses of limestone. Shand states that, in all probability there is a deep-seated body of limestone in this repion since, in his opinion, only the addition of lime or the subtraction o r silica can convert ordinary basalt into a melilite basalt.

Bowen (1923), in a synthetic mixture of diopside and nephelite, noticed the initial separation of diopside, which then reacted with the liquid to form forsterite and akermanite. This experimental result correlated well with what he found in the alnoitic rocks at Isle Cadieux

(Bowen 1922). The relations of Isle Cadieux led him to believe that an alkalic liquid, rich in nepheline, had reacted with diopsidic pyroxene, to produce monticellite and melilite. He felt that the alkaline liquid therefore had reacted as a desilicating fluid. Examination by Bowen of some melilite-nephelite basalts from Hawaii indicated that there, also, 10 aupite was attacked by alkaline liquid forming a melilite. Bowen (1923)

reasoned that if the addition of lime to ordinary basalts resulted in melilite production, this mineral group should be just as common in deep-seated rocks as in effusive rocks. The fact that melilite is practically absent from deep seated rocks suggested to him that the main manner of melilite production is the interaction of nephelite and pyroxene. The rapid cooling accompanying lava eruption accounts

for the failure of the reversal of this reaction at lower temperatures, whereas the slow cooling of the deep-seated rocks causes the melilite

to revert to nephelite and pyroxene. Daly (1933), however, observed

the presence of melilite in frranular plutonic rocks of Tasmania, which would tend to weaken BowenTs arrument. Bowen (1940) noted the occur­

rence of akermanite in metamorphosed siliceous limestones associated with diopside. Turner (1910) found akermanite associated with lamite,

spurrite, and merwinite in the sanidinite metamorphic facies which

includes local metamorphic rock developments at high temperatures and low pressures in the immediate vicinity of near-surface intrustions.

It was already noted that akermanite was first described by Vogt.

McCaffery et al. (192?) found akermanite to be important in the composi­ tion of slags. Koch (1933) also stated that it is not of the most important components of cyrstallized slags. Indeed, modern blast

Turnace slags are calculated to contain 60 to 90^ of an akermanite-

p'ehlenite solid solution. Characteristics of the Minerals

Since in phase equilibrium studies accurate optical identification of the crystalline phases is essential, sufficient diagnostic optical data must be available to permit distinction between the phases encountered. Crystalline phases of interest in the present study have all been repeatedly described in textbooks and papers. Their properties have been determined on natural as well as on synthetic chemically pure substances.

In the discussion which follows attention will be restricted to the pure synthetic varieties, since these are the most pertinent for experimental studies. Since a thermodynamic analysis of the binary system anorthite-akermanite will also be attempted, some thermodynamic data will be included in the following tabulations of previously publish­ ed data. Melting point data, to be used for both phase equilibrium and thermodynamic treatment will likewise be presented.

Anorthite

This mineral belongs to the group of tectosilicates (Morey, 1954)*

It is an end-member of the triclinic solid solution series comprising the plagioclase feldspars. Taylor (1933) proved that Machatski’s

(1928) hypothesis, according to which a feldspar consists of a framework of linked (SiO^) ancj (AIO^) tetrahedra with cations of K, Na, Ca, or

Ba situated in the interstices, is essentially correct. The basic feature of this framework is a ring of four tetrahedral groups in which two adjacent tetrahedra have vertices pointing up and the other two tetrahedra have their vertices pointing down. This type of arrangement

11 12 enables the rings to be connected in such a way as to form a frame­ work type of chain structure parallel to the a crystallographic axis.

These chains are connected to each other by means of common oxygen ions and the particular cations present. It appears from Gay’s work

(1953) that anorthite ordinarily exists as a superlattice, and changes with temperature increase from an ordered to a disordered structure. Apparently, synthetic anorthite that has been rapidly cooled from high temperatures preserves the disordered structure. It may therefore be assumed that in high temperature quench experiments the structure of the anorthite crystals is of the disordered type.

L.P. Wyckoff of Union Carbide accidentally synthesized two new non-triclinic forms of CaAl23i20g (Goldsmith and Ehlers, 1952).

The new phases, one hexagonal and the other orthorhombic, were des­ cribed in more detail by Davis and Tuttle (1952). Chemical analysis showed the three phases to be represented by the same formula,

CaAlpSipOg.

A comparison of optical properties as shown by Davis and Tuttle

(1952) is as follows:

Table 1 Optical data of C a A ^ S ^ O g

Triclinic Orthorhombic Hexayonal (anorthite) 2V 77° 39° 0° s i m -- Nw - - 1.585 Ne - - 1.590 Na 1.5755 1.533 1.5832 1.580 n b 1.5885 1.584 ny 13

Goldsmith and Ehlers (1952) felt that the similarity of re­ fractive indices of the triclinic and hexagonal anorthite indicated the likelihood of closely similar densities. Morey (1954) gave the density of anorthite as 2.765, while Da,Tis and Tuttle listed the density of thehexagonal phase as 2.74 and that of the orthorhombic variety as 2.70.

Dairis and Tuttle (1952) suggested that the orthorhombic and hex­

agonal types of CaAl2 Si2 0 g are metastabls, since on heating they both change to triclinic anorthite. Their laboratory studies repeatedly demonstrated that a very viscous fluid environment promotes the for­

mation of the metastable forms of CaAl2Si2 0 g.

The indices of refraction of pure synthetic anorthite show but slight variation in the literature, and are as follows: Na, 1.576;

Ng, 1.585; Ny, 1.589 (Rankin and ./right, 1915).

For the index of anorthite glass Bowen (1916) gives a value of

1.575, Goldsmith (1950 gives 1.5755, and Kracek and Neuvonen (1952)

1.5763.

A search of the literature shows the following melting point determinations of anorthite. The most widely accepted value is

1553°C., and this has been accepted for the present study. 14

Table 2

Keltiny ooint of anorthite

Authors Date m.p. Method °C

Voyt 1903 1220 optical Brun 1903 1500 optical Day and Allen 1905 1532 ouench Freis 1907 1275-1345 optical Dittler 1911 1330 optical Day and Sosman 1911 1550 pas thermometry Ginsberp 1921 1400 optical Rankin and Wrirht 1915 1550 quench Schairer and Bowen 1938 1550 quench Osborn 1942 1553 quench Goldsmith 1950 1553 quench Goldsmith 1955 1552 quench

The literature indicates that synthetic anorthite observed in quench runs usually shows polysynthetic twinning. Generally the crystals are lath-like, tabular, and elongated parallel to the a axis. The average extinction anrle is 34*5• It ranyes from 27°

(Anderson, 1915) to 40° (Rogers and Kerr, 1942). The optic axial anple (2V) measurements ranye from 77° (Davis and Tuttle, 1952) to

80° 3° (Anderson, 1915).

The thermodynamic analysis of the system anorthite-akermanite renuires the use of heat of fusion data. A search of literature showed the following such data for anorthite: 15

Table 3

Heat of fusion of anorthite

Author Year A'.1, (as given) A Hr(Cals/mole) f Vovt 1903 105 cal/gram 29200 Van Laar 1906 105 cal/gram 29200 Ooranson 1942 375-505 joules/^ram 25000-33600 Goldsmith 1950 29000 cal/mole 29000 Kracek and Neuvonen 1952 63.7 cal/gram 17700

Biopside

Dipside is a typical triclinic member of the pyroxene ino-

silicate group (Morey, 1954). The silicon atoms are surrounded by-

four oxygen atoms as in anorthite. Two oxygen atoms of each group

are held in common with neighboring groups in accordance with a three

to one ratio of oxygen to silicon atoms. The tetrahedra are thus

linked together by shared oxygen atoms to form endless chains

parallel to the c axis of the mineral. They lie side by side and

are connected to one another by calcium and magnesium atoms (barren and Bragg 1928). The refractive indices of synthetic diopside are:

Na , 1.664; Ng# 1.671; N , 1.694 (Allen et al. 1909). The optic axial angle is 60°.

The published values of the melting point of diopside, as ^iven

in the literature, are indicated in Table 4 below. The commonly accepted value is 1391.5°C. 16

Table 4

Meltine Point of Dionside

Authors Date m.p. Method °C. Vogt 1903 1225 optical Doelter 1908 1330-1350 optical Allen et al. 1909 1380 optical Schumoff et al. 1911 1300-1320 optical Day and Sosman 1911 1391 ras thermometry Dittler 1911 1290-1250 optical Schairer and Bowen 1938 1391.5 quench

Rowen (1916) states that the index of refraction of diopside r-lass is 1.607» Dane (1941) yives the density of synthetic diopside as 3.265 and that of diopside glass as 2.846 which is a density difference of 12.9%,

Generally, the literature indicates that diopside is a good crystallizer and shows in melts the prismatic habit, often short and thick, Allen et al. (1909) mention that a characteristic microscopic feature of diopside crystallized from melts is the presence of bubble-like inclusions or cavities throughout the crystallized material. They attribute this phenomenon to shrinkage accompanying cyrstal formation from the melt.

Akermanite

Akermanite is an end-member of the melilite solid solutions, a tetragonal series belonpinr to the sorosilicates (l'*orey, 1954).

Basically the structure consists of 0i20y groups connected by magnesium ions in fourfold coordination with oxygen, and calcium ions in eightfold coordination with oxygen (Brarr, 193?) (barren, 1930), 17

A survey of the literature indicates that the refractive in­ dices have the following values: 1.631; Ng 1.638. The mineral therefor is uniaxial positive. In silicate melts akermanite generally appears as short square prisms. Ferguson and Merwin (1919) noted that frequently the crystals did not appear in quenches of the

system CaO-MgO-SiC> 2 unless the melt was considerably undercooled.

Dane (1941) makes the interesting observations that although

the density of synthetic akermanite crystals is 2 .9 4 , akermanite glass has a density of 2.95. Buddington (1922) mentions the density of

synthetic akermanite as 2 . 9 4 4 and states that it is considerably less than that of natural akermanite, which has a density of 3.12. Ervin and Osborn (1949) calculated the density of synthetic akermanite to

be 2 .9 2 2 +0 .0 0 1 .

Vogt (1903) presented two distinct formulae for akermanite:

(Ca,Mg )Si^ 0-^q and (Ca,Mg)Si^0^. Schaller (1916) proposed the

orrnula 4Mg0 • 8Ca0 . 93102* Ferguson and Merwin (1919) were unable to prepare a compound agreeing with Schaller*s formula but succeeded

in obtaining a compound 2 CaO . MgO . 2 Si0 2 with optical character­ istics similar to those of akermanite. They assumed, therefore, that

the correct formula for this mineral is 2CaO . MgO . 2 Si0 2 »

V/inchell (1924) noints out that natural akermanite contains 4.5$ more

Si0o than does the synthetic' variety. * This he attributes to free

3i0o occup”ing spaces between atoms of the space lattice. All recent investigators are agreed that the Ferguson and Merwin formula 2 CaO .

MpO . 2 Si0 2 satisfactorily represents the composition of synthetic akermanite. 18

The followinp table shows the different melting point de­

terminations of akermanite. The value accepted for the present

investigation is 1 4 5 4 °C»

Table 5

Melting point of akermanite

Authors Date m.p. Method °C Voct 1903 1175 optical Fercuson and Merwin 1919 1453 quench Osborn and Schairer 1914 1 4 5 4 quench

The heat of fusion data for akermanite are shown in Table 6 .

Ta'^le 6

teat of fusion of akermanite

Author Year Zkt^(as given) AH^(cals/mole)

Vogt 1903 9 0 cal/pram 2 4 7 0 0 Ooranson 1942 315-445 joules/ 20500-2900 pram

Carstens and Kristofferson (1931) found that on coolinc very

slowly a melt of akermanite composition, the crystalline phases

obtained consisted of akermanite and diopside. Bowen et al. (1933)

observed that class of akermanite composition crystallized at 1 3 7 5 °C

contained akermanite as the only crystalline phase. The class

crystallized at 1050°C showed akermanite with CapSiO^ crystals.

Osborn and Schairer (1941) found only akermanite crystals to be present in a charge held above 1325°C, but below that temperature diopside appeared to form within the akermanite crystals. 19

Work by Taylor and Williams (1935) on solid state reaction in the system CaO-MgO-SiC^ indicates akermanite formation at 800°C.

The increase of the formation of akermanite with rising temperatures

(1 0 0 0 - 1 2 0 0 °C ) is rather puzzling in view of the alleged instability of this compound as observed by the previous authors. Clark (1946) also succeeded in synthesizing akermanite through solid state re­ action, though at a much higher temperature.

Neuvonen (1 9 5 2 ) reinvestigated this apparent instability of akermanite below 1325°C, His x-ray studies indicated that akermanite breaks up into merwinite and diopside when subjected to a hydrothermal

environment of 1 atmosphere pressure and temperatures ranging from

400-900°C. He suggests the following course of events:

akermanite 1 3 2 5 ° diopside+merwinite

______diopside+monticellite+Ca2 5 iO^

______monticellite+wollastonite

Nurse and Midgley (1953) attempted to duplicate the experiments of

Osborn and Schairer but were unsuccessful. Their failure led them to question the alleged instability of akermanite at temperatures be­ low 1325°C. Experimental Procedure

The experiments undertaken in this investigation were conducted with artificial mixtures made up of chemically pure CaCO^, MgO,

and A12 0^. All thermal work was done under atmospheric pressure.

For each composition studied, ten-gram mixtures were prepared.

The raw materials were carefully weighted, mixed and three times fused in graphite crucibles heated in an Ajax-Northrup induction furnace.

No indications of any reduction nor volatilization of the oxide- components was encountered in this treatment. The indices of the glasses were checked to assure homogeneity and correctness of compo­ sition. The index of refraction charts of the glasses for the system anorthite-akermanite, and diopside-anorthite-akermanite are presented in Figures 2 and 3. A portion of each glass was devitrified prior to thermal study.

A platinum resistance furnace was used in all the thermal work.

The temperatures in the furnace were measured with platinum-platinum

9 0 rhodium 1 0 thermoelements in connection with a potentiometer in­ stallation, The thermoelements were standardized at suitable interval

The thermal data were gathered by the usual quenching technique.

This technique has been extensively described in numerous publications

(Oreene and Morgan, 1941) (Roedeer, 1951). In brief, the procedure is to hold a charge of devitrified mixture wrapped in a platinum foil en elope, at a constant temperature long enough to assure a condition of eouilibrium throughout the charge. The charge is then suddenly cooled to room temperature by causing it to drop into a bowl of mercur”. The product is then examined under a polarizing microscope, 21 and in some instances by means of x-ray diffraction, for the purpose of identifying the phases present.

In order to check the attainment of equilibrium the length of time of the runs v;as in some instances varied. If these runs show agreement in the type of phase present it can be assumed that a state of equilibrium has been established. In general, runs were of thirty minutes duration.

Since anorthite is reported to be reluctant to crystallize

(Osborn, 1942) it was deemed advisable to assure that anorthite was present in the devitrified mixture before a run was begun. According­ ly, the devitrified glasses were carefully examined by means of a polarizing microscope to ascertain the presence of this feldspar. 22

1.941 I 64

1.95

i.ei iso

1.59

1.98

1.57, 2 0 30 40 50 70 BO 90 100 . .. CoO AI.O. 2 SIO. 2C«0 M«0 2510*

Kir. 2. index of refraction chart for

classes of the system anorthite-akermanite. 23

ANORTHITE

20 BO

30 JO

40 60

50

60 40

TO, 30

80, 2 0

90

OIOPSIDE '0 20 30 40 60 7050 BO 90 A k e r m a n i t e

Fip. 3. Index of refraction chart for passes of

the system diopside-anorthite-akemanite. Thermal Data

Thermal data have been fathered on a total of forty mixtures of varied composition. A total of one hundred and forty-nine separate runs varyinf from one half to eleven hours were made. Of these mixtures ten fell within the binary system anorthite-akermanite, while thirty are contained in the ternary system diopside-anorthite-akermanite.

In yeneral only the results of those runs needed to determine the liquidus relations of the systems are presented in the tabulations, charts and equilibrium diagrams in the following pages. In order to facilitate the understanding of the ternary system, diagrams of the two binary systems diopside-akermanite and diopside-anorthite have been included, although they did not form a part of this investi­ gation.

The binary system diopsihe-anorthite

This system was previously studied by Bowen(1916) and by

Osborn (1942). Figure 4 shows the binary system after Osborn. The eutectic between diopside and anorthite lies at a composition of forty-two percent of anorthite and fifty-eight percent of diopside.

The temperature of the eutectic is given as 1274 3°0.

Osborn believes that a small amount of aluminum, ions enters the diopside structure and that therefore the system is not completely binary and the minimum liquidus point is not strictly invariant.

However, departure from binary nature is slight. The slightly downward sloping eutectic horizontal and the slight shift in h-spacings

24 25

of the diopside as compared to those of pure diopside, formed the

basis for the conclusion that the diopside formed is slightly

aluminous.

The binary system diooside-akermanite

The diagram for this system was published by Ferguson and Merwin

(1919)» It is presented in Fig. 5 and, as can be readily seen, it

is a simple binary system. The eutectic temperature was not specified, o but from their diagram it appears to be about 1365 C. The eutectic

lies at a composition of forty-two percent akermanite. The authors

of this system obtained the binary diagram solely by interpolation

from the liquidus temperatures determined from compositions lying

in the ternary system CaO-MgO-SiO^.

The binary system anorthite-akermanite

■'fork by McCaffery and Oesterle (1924) and by McCaffery et a] •

(1 9 2 7 ) on the constitution of iron blast furnace slags indicated

that akermanite and anorthite are compatible. Similarly Janecke

( 1 9 3 1 and 1 9 3 5 ) indicated compatibility between akermanite and

anorthite. However, neither these nor' any later investigators presented any equilibrium diagram for this system nor the thermal

data from which such a diagram might be constructed.

The data of the high temperature phase equilibrium investigation

of this system are presented in Table 10 and Fibure 6 . The melting

point of anorthite and akermanite were not redetermined for this

study. Osborn’s (1942) value of 1553° was accepted for anorthite,

and Osborn and Schairer’s (1941) value of 1454°C for akermanite. 26

isoo

1300 Liquid

1400 Anorthite + Liquid Diopside + 1300 Liquid

1200 Diopside + Anorthite

2 0 30 4 0 30 60 70 so •o Diopside Anorthite Weight Per Cent

Fig. l + . Equilibrium diarram of the system diopside-

anorthite (after Osborn). 27

isoo

•450 MM

XCaO M«|0 1 >i0 , ♦ Malt (ISO

CaO- M9O- XSiO, 4 XCaO M jO XS.O .

1 1 0 0

wt. per cent.

Fig. 5. Equilibrium diagram for the system diopside- i akermanite (after Ferguson and Merwin). Tamp. C o- f0, 2SI . I0 S 2 , 0 lf A CoO- 0 0 3 1 1000 0 0 4 1 1200 1500 0 0 1 1 1553 1600 akermanite . Fi/r • 6 Equilibrium diagram• for the system anorthite- N QUID U IQ L ♦ AN 20 NORT ITE TH R O AN 4 CO 54. 0 5 0 3 0 4 28 t par t n a rc a p wt. C * 4 3 2 1 QUID U IQ L A emanite t i n a kerm 0 7 • 0 0 Ml-28Oj 8IO 2 MflO- - «0 C 2 UID 100 0 9 29

The eutectic for this system is located at fifty-four percent akermanite, at a temperature of 1234°C. This point was determined by the intersection of the two liquidus curves, in addition to the determination of the lowest temperature at which one of the two crystalline phases disappeared from the completely devitrified charges of several different compositions.

The crystals obtained throughout the work on this system were in nearly all cases very small, and no precise measurements of optical properties were possible. The anorthite crystals, particularly, were extremely small, and did not show distinct faces. They were generally lath-like and never showed polysynthetic twinning. A

special run of the composition 7 5 $ anorthite 2 5 $ akermanite was made at 1250°C. for seventeen hours. The purpose of this run was to grow larger crystals to permit unequivocal identification of the

phase as anorthite, rather than one of the other forms of CaAl2 3 i2 0 g.

Although the expected characteristic polysynthetic tweinning of anorthite was still not observed, the inclined extinction and the x-ray powder diffraction pattern left no doubt but that the phase was anorthite. Since the anorthite crystals were usually warped it was difficult to determine accurately the extinction angle on many specimens, although on the more favorable crystals an extinction anple of about thirty-five degrees could be measured.

In this system akermanite crystals were equant to slightly club- shaped. '.'/here good crystal outlines were available, parallel extinction was apparent. The interference colors were moderate to 30 low. Almost invariably, the crystals of akermanite had small bubble­ like inclusions. Interference figures obtained from some of the larger akermanite crystals indicated this synthetic compound to be uniaxial positive.

The apparent stability of akermanite down to 1234°C, contrasting with the results published by the numerous authors discussed previously, caused some concern. Besides the runs critical to the determination of the liquidus relations of the binary system a number of other runs were made in the akermanite field. These runs are tabulated in Table 10.

From all the available experimental data gathered from these compositions there is strong indication that akermanite is stable at least down to 1234°C, which is considerably below the 1325°C stability limit in­ dicated by others. As mentioned previously, work by Taylor and

Williams (1935) also appears to indicate greater stability of aker­ manite than previously believed.

To verify the composition of the crystals x-ray powder photo­ graphs were obtained. Comparison of the d-spacings to those of pure synthetic akermanite clearly indicate that pure akermanite rather than an akermanite-gehlenite solid solution, coexists with anorthite in this system.

The system dionside-anorthite-akermanite

The data, critical for the determination of this three component system, are presented in Table 11. These data are graphically presented in Figure 7. This figure indicates the limits of the primary fields of the three components. 31

Table 10

Thermal data for the system Anorthite-akermanite

Weight Percent Time Temp. Final Condition CaAl2 Si2 0 g Ca2 MgSi2 0 y (min) °C

Primary phase CapMgSi2 0 y

1 0 90 30 1,412 Ak and glass 30 1,416 Glass

2 0 80 30 1,369 Ak and glass 30 1,373 Glass

30 70 30 1 , 3 2 8 Ak and glass 30 1,332 Glass

40 60 30 1,273 Ak and glass 30 1,277 Glass

45 55 6 0 1 , 2 4 0 Ak and glass 6 0 1,244 Glass

Frimary chase CaAl2 3 i2 0 g

47 53 1 2 0 1,244 An and glass 1 2 0 1,248 Glass

50 50 30 1,265 An and class 30 1,269 Glass

60 40 60 1,340 An and class 30 1,344 Glass

70 30 30 1 , 4 0 8 An and class 30 1,412 Glass

80 2 0 30 1,462 An and class

30 1 , 4 6 6 Glass

Data for eutectic horizontal

30 70 1 2 0 1 , 2 3 2 Ak and An

1 2 0 1 , 2 3 6 Ak and Glass

45 55 60 1,230 Ak and An 6 0 1,236 Ak and glass

47 53 90 1 , 2 3 2 An and Ak 1 2 0 1 , 2 3 6 An and class 32

Table 10 (continued

Thermal data for the system Anorthite-Akermanite

Wftipht Percent CaA^S^Og Ca2 MgSi2 0 r7 Time Temp. Final Condition (min) °C

Data for Hutectic horizontal (continued) o CP o p- 1 2 0 1 , 2 3 2 An and Ak

1 2 0 1 , 2 3 6 An and plass

Additional confirmatory data, primary phase Ca^pSi^r, o CP o p- 3 0 0 1 , 2 5 0 very larpe Ak crystals and plass. O CP o P- 285 1,280 very larpe Ak crystals and plass.

4 0 6 0 30 1,236 Ak and plass 30 1,244 Ak and plass 30 1,252 Ak and plass 30 1,267 Ak and plass 30 1,265 Ak and rlass

S-i nee the system diopside-anorthite is reported to be not truly binary throughout, it would follow that this ternary system is not quite truly ternary. However it is believed that any departure from ternary nature is slirht. As mipht be expected on the basis of known relations in the three binary systems, the ternary eutectic lies closer to the binary eutectic of anorthite and akermanite than to either of the other two binaries.

The ternary eutectic of this system is located at a composition of eirht percent diopside, forty-four percent anorthite and forty- ei^ht percent akermanite. The temperature of this eutectic is 1226°C. ^epunoq put? SuUd^ o s T *S9AJno

"et'TSdOIP » » & Sq, . ' ‘a’IU,'“U® i * - 3 n liWoIJ

Oi UQ a ^ ’*!„ Os LIto OS US 01 i£!Sdoia

09frl

£C 34

From the ternary diagram with isotherms it is apparent that the anorthite field has a steeper temperature gradient than the other data

of Osborn et al. and of Prince are presented in Fig. 8 .

It could scarcely be expected that such interpolations could give as precise an equilibrium diagram as this present experimental study of the liquidus relations in the diopside-anorthite-akemanite plans.

However, it was felt that a diagram derived from such data would furnish a guide as to what relations might be expected. In view of the admitted difficulty involved in the interpolation of the data of Osborn et al.

and of Prince to provide the diagram shown in Fig. 8 -it is considered that the correlation with Fig. 7 is good. On the other hand, it is f,elt that an accurate diagram for this system can be expected only from an actual experimental study of the system in question. Fiirthermore the actual composition of the crystalline phases encountered in the system diopside-anorthite-akermanite cannot be directly deduced from the data of the investigators cited above. Thus one could not deduce that the melilite encountered would actually be pure akermanite nor that the pyroxene would be pure or essentially pure diopside.

The crystals obtained from the thirty ternary mixtures were in nearly all cases very small. Anorthite and akermanite were similar to those found in the binary system previously described, diopside crystals in general were well formed and appeared lath-shaped with an extinction anrle ranrinr from 32° to 40°• Although Allen et al. (1909) mention that a characteristics microscopic feature of diopside crystallized from melt is the presence of bubble-like inclusions, none were observed in diopsi:e formed in these experimental melts. 35

CaO *1 ,0 ,' 2 810,

Dadaaad tram data at Prlnaa-- Dadaaad tram data at Oabarn atal

ANORTHITE

7T PYROXENE (Oaborn atal) MELILITE OIOPSIOE ( Prlnea)

2 CoO MfO 2SI0,

Fifr. 8 . Fiquilibrivun diagram for the system anorthite- diopside-akermanite on the basis of the data of Osborn et al. (1954) and Prince (1954). 36

Table 11

Thermal data for the system Anorthite-Diopside-Akermanite

V/eirht Percent CaA^Si^Og CaMfrSi2 0 £ Ca2 MpSi2 0 r, Time Temp. Final Condition (min) °C

Primary rihase CaAl?Si2 0g

50 40 1 0 30 1308 An and plass 30 1312 Glass

6 0 30 1 0 30 1378 An and plass 30 1 3 8 2 Glass

57 2 0 23 30 1338 An and plass 30 1342 Glass

45 30 25 30 1 2 6 2 An and plass 30 1 2 6 6 Glass

50 15 35 30 1280 An and plass 1284 Glass

45 1 0 45 30 1236 An and plass 30 1 2 4 0 Glass

Primary Phase CallpGi^ ° 6

40 50 1 0 30 1 2 7 0 0 i and class 30 1274 Glass

42 43 15 30 1 2 6 0 Oi and plass 30 1264 Glass

1 0 70 2 0 30 1346 Oi ar/ 1 plass 30 1350 Glass

50 30 2 0 30 1288 Oi and plass 30 1 2 9 2 Glass

55 2 0 25 40 1314 Oi and "lass 45 1 3 1 8 Glass

40 30 30 70 1 2 5 6 Oi and rlass 30 1 2 6 0 Glass

2 0 45 35 30 1304 Oi and plass 40 1308 Glass

35 30 35 30 1274 Oi and class 1278 Glass 37

Table 11 (continued)

______Thermal data for the system Anorthite-Dio-pside-Akermanite_____

ieipht Percent CaAl^Si 0rt CaMpSio0, Ca^MpSi^Oo Time Temn. Final Condition 2 2 8 2 6 2 2 7 (min) oG‘

Primary Phase CaMpSi2 0 6 (continued)

2 0 42 38 40 1 3 0 0 Di and plass 30 1304 Glass

43 19 38 30 1 2 4 2 Di and plass 30 1246 Glass

2 0 40 40 40 1 3 0 2 Di and plass 30 1 3 0 8 Glass

40 2 0 40 40 1244 Di and plass 30 1248 Glass

a 50 42 30 1336 Di and class 30 1340 Glass

38 18 44 30 1 2 4 0 Di and plass 40 1244 Glass

43 1 2 45 30 1230 Di and plass 30 1234 Glass

43 1 1 4 6 30 1226 Di and plass 50 1230 Glass

Primaryr Phase Ca MpSi 2°7

2 0 35 45 30 1304 Ak and class 30 1308 Glass

40 15 45 30 1236 Ak and plass 45 1 2 4 0 Glass

43 9 48 30 1226 Ak and class 30 1 2 3 0 Glass

2 0 30 50 50 1314 Ak and plass 40 1 3 1 8 Glass

DO 2 0 50 30 1284 Ak and class 30 1288 Glass 38

Table 11 (continued)

Thermal data for the system Anorthite-Diooside-Akermanite

Weight Percent CaAl^SipOg CaMpSi^O^ CapMgSi?Oy Time Tsurp • Final Condition (min) °c

Primary Phase Ca^IlpSi^O^, (continued)

40 10 50 30 1246 Ak and glass 30 1250 Glass and exceed­ ingly rare Ak crystals 30 1254 Glass

10 25 65 30 1376 Ak and glass 30 1380 Glass

Eutectic horizontal

43 11 46 40 1228 Di and glass 30 1224 Di and Ak and An

45 10 45 30 1228 An and glass 30 1224 An and Di and Ak

44 8 48 30 1228 Glass 30 1224 Di and An and Ak Thermodynamic- Analysis of the System Anorthite-Akermanite

A thermodynamic analysis of a phase equilibrium system may be made with various objectives in mind. In general, thermodynamic infor­ mation will provide a quantitative basis for better understanding of

the characteristics of a particular phase equilibrium system. Also,

by combining thermodynamic calculations with the results of electrical

and x-ray investigations, it may be possible to develop a more detailed

picture of the structure of the liquid phase present.

Although some authors feel that a liquid may be considered as a

continuation of the baseous phase into a region of stronger inter-

molecular attractions and small volumes (Prutton and Maron, 195l), others

(Darken and Curry, 1953) state that there is a general relationship be­

tween the structure of solids and liquids. This appears to be true in

the more viscous silicate melts.

The Clausius-Clapeyron equation L is

applicable to gas-liquid equilibrium. It may also be applied to solid-

liquid equilibrium, when the ratio of vapor pressure is replaced by

the mole fraction of the solvent.

Experimentally it has been established that the freezing point

depression of a melt depends both on the nature of the solvent and

on the degree of dissociation or association of the solute. Assuming

an ideal solution, an equation relating these two factors to each

other follows directly from the Clausius-Clapeyron equation. The

equation may be expressed as follows:

(Prutton and Maron, 1951)

39 40

1900

IIOO LIQUID

1400

j iioo * LIQ U ID AK 4 LIQUID

1234* C \ ✓ 1200

IIOO ANORTHITE

20 10 40 •070 00

Fig. 9. Comparison of theoretical and experimental liquidus curves for the system anorthite-akermanite. Upper zone lying between the dashed lines is the calculated liquidus zone on the basis of a A value for anorthite ranging from 375-505 joules per pram and a A H value for akermanite between 315-445 joules per rram, and an i factor of 1. The lower liquidus zone, lying between the dotted lines re­ present the results of freezing point lowering calculations with similar values and an i factor of 2. Note that the experi­ mental curves fall within the lower liquidus zone. where:

= freezing point depression of the solution

K = gas constant in calories per mole

= meltinr point in °K of the pure compound-1

which in the liquid state suffers freezing

point depression by the addition of another

component -2.

T = actual freezing point of solution

= h eat of fusion of component 1

= mole fraction of component-1 in the solution

However, it is known that the freezing point lowering of

electrolytic solutions is greater than the corresponding effect for

solutions of nonelectrolytes of the same reneral concentrations.

In order to compare this collipative property of electrolytes to that

of nonelectrolytes, vanft Hoff introduced the S7yTmbol i, the so-called

vanft Hoff factor. This may be defined as the ratio of the colligative

effect produced b?/- a mole fraction of an electrolyte to the effect ob­

served for an equal mole fraction of nonelectrolvte. Hence ‘ , .j!*" L (Prutton and Maron, 1951), where ( is the freezing point

lowerinr produced by the nonelectrolyte and ( A?/)* is that

produced by an electrolyte. Also ( an<^ since ( a

we may now rewrite the equation for freezing point depression of an

electrol7 rte as follows: 42

It is obvious that if we have a nonelectrolyte solution, i is 1.

The last equation may be utilized for the calculation of a theoretical diagram for the system anorthite - akermanite with the view of comparing experimental and calculated liquidus curves. Since

% and are known for both anorthite and akermanite it is a simple matter to obtain the A ^ for any selected value of Ht or ^ and for various i factors.

Fig. 9 shows two calculated liquidus zones. The upper zone is the result calculated with an i-factor of 1, while the lower zone shows the result of a calculation with an i factor of 2. Since the lower calcuated zone shows an excellent correlation with the ex­ perimental curve, it is reasonable to conclude that an i-factor of 2 is applicable. An i-factor of 2 may be taken as an strong indication that the silicate melt is electrolytic in nature. Such a conclusion appears to be quite conservative in that it arrees with numerous ob­ servations of other authors, such as Doelter (1907 and 1908), Farup et al. (1924), Martin and Derge (1943), Bockris et al. (1948) and

Barth and Rosenquist (1949)• All of the above authors stated that the numerous silicate melts with which they were concerned acted as stronv electrolytes.

The foregoing treatment permits the construction of a tentative model of the anorthite-akermanite silicate melt. Darken and Curry

(1953), Bockris et al. (1948), Barth and Rosenquist (1949), and others believe that the essential difference between the solid and molten silicate lies in the decree of order or disorder. Accordinr to this 43 belief the liquid state is thus characterized by an extended network lacking symmetry and periodicity. Bockris et al. consider the molten silicates as consisting of irregularly arranged silicate rings and chains with cations fitting in the interstices. Martin and Oerge in experimenting with liquids of anorthite compositions indicated that at 1450°C calcium had migrated to the cathode while a complex alumina- silicate had migrated to the anode. They concluded, therefore, that in the molten slag such a composition is present as Ca++ and a complex silico-aluminate. Balu (1951) indicated that in supercooled silicate melts there are two distinct types of oxygen ions. He divided them into those which bridge two network forming cations, and those which are connected to only one '’network-forming” cation and which have valence charges balanced by the considerably more loosely held and distant "network-modifying” cations. Blau then appears to indicate that when the ratio 11 network-forming” cations to oxygen is 1:2 all the oxygens in the silicate melt are bridging. This agrees with tables by Evans (1952) concerning similar ratios in silicate structures.

Blau proceeds to provide further ratios and their connection to the types of oxygen present. A melt of anorthite has definitely a ratio of 1:2 and thus such a liquid would tend to ha^e all bridging oxygens.

A complex as well as Ca++ would therefore tend to result from such an ionic melt. Such a point of view would then agree with the observations by Martin and Derge. Also, such dissociation of the dissolved anorthite into ions of that type would explain an i factor of 2 for the akermanite liquidus curve. 44

In our brief discussion on the structure of akermanite we stated that it belongs to the sorosilicates, the basic structure of which

is the Si2 0 y group. Eitel (1952) prefers to classify akermanite among the ph-'fLlosilicates. He states that the Si20y groups are combined with (MgO^) tetrahedra to form fivefold annular structures. Eitel then states that complexes of the composition will result. Since 2N = 4NQa++ + N it follows akermanite (MggSj^Q^ ^ )°~ that N = 2M +++f;N g_. The i factor of aker- akermanite 0^2^^4^14^ manite electrolyte would thus be 2*. Considering Blaufs calculation and looking upon Mg as a network forming ion in this case, the ratio would be 0.428. This would allow such a composition to fall into the caterory of 3-4 bridging oxygens per tetrahedron of (SiO^) or (MgO^).

According to Evans such a ratio would be classified as sheet structures.

The fact that an i factor of 2 appears to fit the experimental curve while structural considerations tend to indicate a factor of 2y might be due to incomplete dissociation of all the units. Discussion and Conclusions

This investigation has indicated that CaAl2Si20g and Ca^gSi20y

form a simple eutectic binary phase equilibrium system without inter­

mediate compounds. Also the indications are that CagMgSi20y is stable

at least down to a temperature of 1 2 3 4 °C, in contrast to the observations

of numerous other authors. The hexagonal and orthorhombic forms were not

found to be present among the CaAl2 Si2 0 g crystals of these synthetic

melts. This is consistent with the view of Davis and Tuttle that these

two forms are metastable.

From thermodynamic considerations of the system anorthite-akermanite

it appears that the melts of such a system are ionic in nature. This

can be hypothesized on the basis of the approximation to an i factor of 2.

The liquidus relations in this system would thus seem to afford confirm­

ation to the theory, based on conductivity measurements, that silicate

melts such as molten anorthite dissociate into such ions as Ca++ and

(A^SipOg)^ .

The experimental data of this study indicate that CaMgSi20^,

CaAl2Si20g, and Ca2MgSi20y form essentially a simple ternary system. The

slight departure from a true binary system observed by Osborn in the

system CaMrSi2 0 ^ - CaAl23i20g did not have a perceptible effect on the

three component system. On the basis of the work of Osborn et al.

(1954) and of Prince (1954) it was possible to construct a similar

ternary diarram. However it is apparent that there is a significant

difference in the trend of the diopside-akermanite boundary curve and

in the position of the eutectic.

45 US

Since the diopside-anorthite-akermanite system appears to be essentially a true ternary system, it is felt that a new partition of liquids crystallizing in the system Ca0-Mg0-Al20^-Si02 has been established. It is also apparent from these data that the intrusion of a basaltic fluid into a limestone should lead to the formation of melilite. This observation would agree with the numerous field obser­ vations of others.

By a simple extrapolation of the data of this investigation, it is possible to account satisfactorily for the scarcity of the natural igneous rocks composed essentially of pyroxene, plagioclase, and melilite. There is reason to believe, from compatibility relations recorded by previous authors, that diopside, anorthite, akermanite, and pehlenite are linked together to constitute a subordinate tetrahedron, or sub-system, within the general system Ca0-Mg0-Al20^-

SiO^. This tetrahedron comprises a very thin composition wedge in the general tetrahedron. This is tantamount to saying that only a very narrow range of compositions is so located as to yield the phase assemblage diopside-anorthite-melilite. Hence, the probability of frequent occurrence of pyroxene-plagioclase-melilite rocks is slight.

The primary basaltic magma of course lies outside this sub-system as regards its Ca0-Mg0-Al20^-Si02 contents. Haematic differentiation would have to carry the magma composition into this system, and a fortuitous interruption of the process at this juncture, would be called for, before such an assemblage would be forthcoming. On the basis of the limestone-assimilation hypothesis, too, the probability of such rock being formed is low, since just the critical amount of 47

carbonate rock to carry the basaltic magma into the sub-system would have to be digested. In discussing the formation of melilite in nature

from the decomposition of the carbonate of lime or magnesia and sub­

sequent reaction with the intruding magmatic fluid, it should be remember­

ed that such a decomposition process is endothermic in nature. The ability of an intruding magma to supply the quantities of heat necessary

for such a process on an extensive scale still has to be established.

As has been discussed previously the field evidence indicates that on

a limited scale the magma is capable of supplying the necessary energy.

In general it should be kept in mind that in applying the results

of any investigation of this type to petrologic problems involving minerals as they occur in nature one should realize that it is a con­

siderable ”extrapolation” from relatively simple laboratory systems to the significantly more complex and varied magmatic system. While in a laboratory the system is maintained at atmospheric pressure, under geologic conditions the reactions often take place under con­ siderably higher pressures which moreover are not uniform throughout the system itself. Also, the complexity of reactions caused by the presence of volatile substances should be kept in mind. The completion of a phase equlibrium system on a laboratory scale and under care­ fully controlled conditions is nevertheless one step further in un­ raveling the complexity of magmatic processes. However, the complete picture will not be available until systems approaching natural magmas in complexitv and in range of temperature and pressures have been in­ vestigated. f

48

According to Osborn et al. (1954) the presence of substantial

Ca++ increases the desulferizing potential of blast furnace slag. The thermodynamic treatment of the binary system anorthite-akermanite led to the conclusion that both phases dissociate in the molten state, and in mutual solutions to yield Ca+4 . It therefore appears probable that a high concentration of anorthite and akermanite in the slag would have a salutary effect on slag chemistry.

The system diopside-anorthite-akermanite is somewhat removed from that portion of the CaO-MgO-A^O^-SiC^ system which has received the attention of ceramic technologists. Spark plug insulators porcelains are confined to the A^O^-apex; and Portland cement to near the CaO-

A^O^-SiC^ face of the tetrahedron. Recent interest in whitewares has centered on wollastonite-talc-clay compositions which lie closer to the Si02 anex than does this system. Nevertheless, rather low-firing bodies could be expected from the diopside-anorthite-akermanite system, and it might be well worthwhile to test the ceramic possibilities of such compositions. Such bodies could readily be compounded of clay, talc, whiting and wollastonite. Bibliography

Allen, E.T., White, W.P., Wright, F.E., and Larsen, E.S. "Diooside and its relation to calcium and magnesium silicates” .

Am. Jour. Sci. (4 ), 27, pp. 1-47, 1909.

Amberg, C.R. "Properties of wollastonite and diopside with admixtures of TiO , Zr02 and SiO". 2 Jour. Amer. Cer. 5oc.. 28, pp. 137-143, 1945.

Anderson, 0. "The system forsterite-anorthite-silica"• Am. Jour. Sci. (4). 39. pp. 407-454._1915.

Barth, T.F.W. and Rosenquist, T. "Thermod^Tiamic relations of immiscibility and crystallization of molten silicates". Am. Jour. Sci. (5), 247, pp. 316-323, 1949.

Becker, E. "Cie Basalte des Wartemberps bei Oeisingen in Baden". Z. deutsch. geol. Gesellsch.. 59,pp. 242-274, 1907.

Berman, H. "Composition of the Melilite group". Am. Ilin.. 14, pp. 309-389, 1920.

Blau, H. "The relations of thermal expansion, composition and structure of -lasses". Part I : "The sodium-silica glasses". Trans. Soc. Class. Techn.. 35, pp. 304-317, 1951.

Bockris, J.O.M., Kitchener, J.A., Imatow'cz, S., and Tomlinson, J.W. "The electrical conductivity' of silicate melts : svstems containing Ca, Mn, and Al". discussions Far. 5oc«. 4, pp. 265-281, 1948.

Brar1-, W.L. Atomic structure of minerals. Cornell Uni”ersitv Press, Ithaca, II.Y. 1937, pp. 184-190.

Brun, A. "Etude sur le point de fusion des minereaux".

Arch, de 3c. Phvs. et nat. de Oeneva (4 ), 13, pp. 24-38, 1903.

Bowen, N.L. "The crystallization of haplobasaltic, hanlodioritic and related magmas".

Am. Jour. Sci. (4 ), pp. 161-185, 1915.

49 50

Bowen, N.L. "The ternary system diopside-forsterite-silica".

Am. Jour. Sci. (4 ), 38, pp. 207-264, 1914.

Bowen, N.L. "Progressive metemorphism of siliceous limestone and dolomite". Jour. Oeol. 48, pp. 225-274, 1940.

Bowen, N.L. "Oenetic features of Alnoitic rocks at Isle Cadieux, Quebec". Am. Jour. Sci. (5), 3, pr. 1-34, 1922.

Bowen, N.L. "The Genesis of melilite". Jour. Nash. Acad. Sci.. 13, pp. 1-5, 1923

Bowen, N.L., Schairer, J.F., and Posn.lak, E, "The system CaO-FeO-SiOp" Am. Jour. Sci. (5), 26, pp. 193-284, 1933

Buddin^on, A.F, "On some natural and synthetic melilite” . Am. Jour. 3ci. (5), 3, pp. 35-87, 1922

Burress, B.C. "Feldspar" Industrial Minerals and Rocks. 1949, A.I.M.E., New York

Carstens, C.'J. and Kristofferson, K. "Schlackenstudien mit besonderer Berucksichtifmnr der manrenhaltiren Schlacken"• Neues Jahrb. Min. Oeol.. Beil. Bd. 62A, pr>, 150-196, 1931

Clark, C.B. "X-ray diffraction data for compounds in the system CaO-M^O-SiOp". Jour. Am. Cer. Soc«. 29, pp. 25-31, 1946

Cross, J., and Larsen, E.S, "A brie*' review 0 ° the ^eolorv of the 3an Juan Rerion of Southwestern Colorado". U.S.0.5. Bull.. 843, nr. 1-135, 1935

'ana, j.j , o. "A textbook of Mineralonr". 4th ed., revised bv Ford, N.F. J. Wiley and Sons, N.v . 1932, pn, 548-607

Dane, S.B. Jr. "Densities of Molten rocks and Minerals"

Am. Jour. Sci. (5 ), 39, pp. 804-810, 1941 51

Daly, R.A. Ipneous Rocks and the Depta. of the Earth. McGraw-Hill Book Co., N.Y., 1933, pp. 482-544.

Darken, L.S. and Curry, R.W. Physical Chemistry of Metals McGraw-Hill Book Co., N.Y., 1953

Davis, G.L. and Tuttle, O.F. "Two new crystalline phases of the anorthite compositions, CaO.A^Oo. 2SiO " Am. Jour. Sci. Bowen Volume, 1952

Da'r, A.L., and Allen, E.T. "The isomorphism and thermal properties of the feldspars".

Am. J^ur. Sci. (4 ), 191. np. 1-1A4, 1905

Day, A.L., and Sosman, R.B. "The meltiny points of minerals in the lirht of recent investigations of the p u s thermometer".

Am. Jour. Sci. (4 ) 31, pp. 341-349, 1911.

De Ville, R.V/. "The substitution of wollastonite for talc in a hirh talc whiteware body". M.S. Thesis. Ohio State University, 1955

Dittler, E, "Deitrar zur Thermochemie der Silicate". Anor^. Chem.. 69, no. 273-304, 1911.

Doelter, C. "Ueber die Dissoziation der Silikatschmelzen". Monatsch. Chem.. 28, pr>. 1313-1379, 1907

"Ueber die Dissoziation der Silikatschmelzen". Monatsch. Chem.. 29, 1908, pp. 607-643

Eitel, H. The Physical Chemistry of the silicates. Univ. Chicavo Press, 1954, p. 47

Ervin, 0., Jr. and Osborn, A.F. "X-ra,r data on synthetic melilites" Am. Min.. 34, pp. 717-722, 1949.

Eskola, P. "On contact phenomena between Gneiss and limestone in Uestern Massachusetts"• Jr>nr. Oeol.. 30, pn. 265-294, 1922 52

Evans, R.C. An introduction to crystal chemistry Univ. Press, Cambridge, 1952, p. 241

Farup, F., Fleisher, V/., and Holtan, F. "La conductibilite electrique des Laitiers fondus" Chim. et Industrie, 12, pp. 11-1$, 1924

Fermuson, J.B. and Merwln, H.E. "The ternary system CaO-MpO-SiO "

Am. Jour. Sci. (4 ), 48, pp. 81-122, 1919

Ferguson, J.B, and Merwln,H.E. "Wollastonite and related solid solutions in the ternary system lime- mapnesia-silica",

Am. Jour. Sci. (4 ), 48, pp. 16$-189, 1919

Freis, R. "Ueber die Ausscheidunpspefolpe von Silikaten" Neues Jahrh. Min. 23, pp. 70-85, 1907

Bay, P. "The structure of the Plapioclase feldspars III" Min. Ma*.. 30, 1953

Uinsberp, A.S. "Ueber einire kuenstliche Aluminosilikate vonT’-pus RO AlpOo ^SiO," Z. Anorp. Chem.. 73, pn. 277-293, 1912. ^

Uoldsmith, J.R. "Callium and Oermanium substitution in S-'mthetic Felspars". J. Oeol.. 58, pp. 529-^38, 1950

C!oldsmith, J.R. and Ehlers, E.O. "The stability relations of anorthite and its hexaponal polymorph

in the system CaAl2 Si2 0 g - HpO" Jour. Oeol. 60, 4, pp» 386-397, 1952

Golschmidt, V.M. "Die Kontakt metamorphose im Kristianiapebiet" Vidensk. Skr. I.. Mat. Naturv.. Kd., no 11, 1911

fold smith, J.R. "The system CaAl2Si20p-Ca2Al2Si0y-NaAlSi0 " Jour. Oeol.. 55,pp. 381-404, 1947

Ooranson, R.V/. "Heat capacity, Heat of fusion" U.S.A. special oaoer. 36, 1942 53

Cossner, B. "Die chemische Zusaimiensetzung in der Gehlenite-Melilith Gruppe". Chemie der Erde. 2, pp. 103-113, 1926.

Greene, K.T. and Morgan, tf.R. "The system sodium disilicate-barium disilicate". Jour. Am. Cer. Soc.. 24,4,p p .111-116, 1941

Grubernmann, U. and Nipgli, P. "Die Gesteins Metamorphase". Dorntraerer, Berlin, 1924.

Grout, F* Petrography and P e t r o l o ^ . McGraw-Hill, 1932, N.Y., pp. 143-259.

7upill, W. and Green, A.T. "The action of slaps of refractory minerals". Iron and Steel Inst.. Spec. Report No. 26, pp. 295-357, 1939*

Janecke, E. "Mineralorie and Phasenlehre" Forschr. Min. Krist. Petrop.. 15,pp. 7 5-lll, 1931.

Janecke, E, "Ueber die Zerlerunr des Systems CaO-MpO-Al^Oo-SiO^ in drei, Phasentheoretisch natuerliche Gruppen mit Melllith, Anorthi und Periklas". Neues Ja’Tb . Min. Geol. 3al«. 69A, pp. 296-308, 1935.

Koch, L, "Konstitution und Kristallisations verlauf der basischen Eisenhocvio "en- schlacken und i'~re Bezierhunpen zu silikatischen Mehrstoff systemen". Neues Jahrb. Min.. Geol.. Pal.. Beil.,Bnd. 67, Abt.A,pp.401-428,1933.

Kracek, F.C., and Neuvonen, K.J. "Thermochemistry of plagioclase and alkali feldspars". Am. Jour. Sci..nowen Vol. pp. 293-318, 1952.

Knizek, J.O. and Fetter, M. "Alunite as a source of scummin" on stoneware body". Jour. Am. Cer. Soc.. 28, pp. 256-261, 1945.

Machatski, F. "Fur Frace der Struktur und Konstitution der Feld spate". Centralbl. Min.. Geol.. Pal.. Abt.A, pp. 97 -104, 1928.

Martin, A.E. and Uerre, G. "The electrical conductivity of molten blast furnace sla^s". Trans. A.I.M.E., 154, 1943, pp. 104-115. 54

McCaffery, R.S. and Oesterle, J.R. "The constitution of iron blast furnace slaps" S.I.M.E. Yearbook, pp. 285-307, 1924.

McCaffery, R.S., Oesterle, J.R., and Schapiro, L, "Composition of iron blast furnace slaps". A.I.M.E. Tech. Pub. 19, pp. 1 42, 1927.

Morey, O.V. "Silica and Inorganic Silicates" Encyclopedia Chem. Tech..12. pp. 268-303, 1954.

Neuvonen, K.J. "Thermochemical Investigation of the akermanite-rehlenite series". Oeolopinen Tutkimulaites. Bull, de la Commiss. Oeol. de Finlande. 158, pp. 5-49, 1952.

Nurse, R.V/. and Mid "ley, H.G. "Studies on the melilite solid solutions". J. Iron and Steel Inst.. 174, pp. 121-131, 1953.

Osborn, E.F, "The system CaSiO„-diopside-anorthite". Am. Jour. Sci.. 240, pp. 7^1-788, 1942.

Osborn, E.F. and Schairer, J.F. "The ternary system Pseudo-wollastonite-akermanite-pehlenite". Am. Jour. Sci.. 239, pp. 715-763.

Osborn, E.F., de Vries, R.C., Gee, K.H., and Kraner, H.M, "Optimum composition of blast furnace slavs as deduced from liquidus

data for the Quaternary 37Tstem Ca0 -i'r0 -Al2 0 ^-Si0 2 ". Jour. Metals. 6, pp. 33-45, 1954.

Prince, A.T. "Liquidus relationships on 10^ MrO plane of the system lime-mamesia- alumina-silica" • Jour. Am. Cer.Soc. 37, pp. 402-408, 1954.

Prince, A.T. "Phase Equilibrium relationships in a portion of the system KrO-AlpOo- 2Ca0.Si02". ^ Jour. Am. Cer. Soc.. 34, 1951, pp. 44-51.

Prutton, C.F, and Maron, S. Fundamental Principles of Physical Chemistry. MacMillan Co., N.Y. 1951, pp. 193-212. 55

Ramdohr, P. Klockmanfs Lehrbuch der Mineralogie. Snke Verlap, Stuttgart, 1948.

Rankin, O.A. and Wright, F.E, "The ternary system CaO-Al 0-,-SiO

Am. Jour. Sci. (4 ), 39, pp. 1-79, 1915.

Roedder, E.VJ. "The system F^O-MpO-SiC^" • Am. Jour. Sci. 249, pp. 81-130, 1951.

Rogers, A.F. and Kerr, P.E. Optical Mineralogy McCraw-Hill, N.Y. 1942, p. 262.

Ri^bT', O.R. and Oreen, A.T, "The thermal expansion characteristics of some calcareous and ma^- nesian minerals". Trans. Brit. Cer. Soc.. 41, pp. 123-143, 1942.

Schairer, J.R., and Bowen, N.L. "The system leucite-diopside-silica". Am. Jour. Sci. (5), 35A, pp. 289-304, 1938.

Schaller, v/.T. "The Melilite group". U.S.O.S. Bull.. 610,pp. 109-128, 1 9 1 6 .

Schumoff-Deleano, V. and Dittler, E. "Finite Versche zur Bestimmunv des Kristallisations - Vermo<-en von Mineralien". Centralbl. Min.. Oeol. Pal.. Abt.A, pp. 753-758, 1911.

St. Pierre, P.D.S. "Const it Tit ion of bone china : II, Reactions in bone china bodies". Jour. Cer. Soc. 38, 6, pp.217-222, 1955.

Shand, S.J. Eruptive Rocks. John Wiley and Sons, N.Y., 1947, pp. 87-325.

Scheuman, K.H. "Fur Oenese alkalisch-lamprophyrischer C-any^esteine" • Centralbl. Min. Ceol. Pal.. 16, pp. 495-520, 1922.

Sin^ewald, J.T. and Milton, C. "An Alno’tenipe, its contact phenomena, and ore deposition near Avon, Missouri". Jour. Ceol.. 38, pp. 54-66, 1930. 56

Stansfield, J. "Chemical characters of Okaite".

Am. Jour. Sci. (5 ), 1926, pp. 396-415.

Taylor, W.H. "The structure of sanidine and other feldspars” . Z. Krist.. 85,pp. 425-442, 1933.

Tailor, N.W. and Williams, F.J. "Reactions between solids in the system CaO-MpO-SiCU in the temperature ran ye 600-1200°C". O.S.A. Bull. 46, pp.1121-1136, 1935.

Tilley, C.E. "On melilite as a product of interaction of limestone and basaltic liquid". Oeol. Mar.. 66, pp. 347-354, 1929.

Turner, F.J. "Structural evolution of the metamorphic rocks". O.S.A. Memoir. 30, 1948

Van Laar, J.J. "Ueber den verlauf der Schnelzkurven bei festen Loesunren (oder isomorphen Oemiscben) in einem speziellen Fall” . Z. Ph^s. Chem. Stoch. Verwandsch. 55, pp. 435-441.

Vorrt, J.H.L. "fie Silikatschmelzloesunre" . Brokers, 1903.

Warren, 8.E. "The Structure of Kelilite". Z. Krist.. 74, up. 131-138, 1930.

■Warren, B.E. and Bra nr, W.L, "The structure of diopside".

Z. Krist.. 6 9 , up. 168-193- 1928.

Winchell, A.M. "The composition of melilite".

Arn. J^ur. Sci. (5 ), 8, up. 375-387, 1924.

Winchell, A.M. "The Optic and microscopic characters of artificial minerals". Univ. 'Wisconsin studies in Sci.. 4, 1927.

'Wiseman, J.U.H. "The central and southwest Highland Epidiorites". Oeol. Soc. London 90, pp. 354-417, 1934. 57

Zambonini, F. Mineralopia Vesuviana Napoli dell. Acad. Delle Sci. fisiche, matem,1910. AUTOBIOGRAPHY

i' I, Epbert Christiaan de 'Wys, was born in Soerabaja, Java,

Netherlands, East Indies, April 9,1924. My rrade school education was obtained at the fs Ora’renhaapsche Genootschap Lapere School at the :larue, Netherlands, Europe. I received a hiph school diploma at Utrecht, Holland, and one at V/est Tech Hiph School at Cleveland,

Ohio, U.S.A. At the end of military service I became a freshman at

Hestern Reserve University in Cleveland in 1948. In 1950 I was rranted a Bachelor of Arts derree in Geolory at Miami University. At the same University I was awarded the Master of Arts depree in Geolopv in 1951. After two years of employment in Venezuela with the Creole

Petroleum Corporation of the Standard Oil of New JerseTr Oil Company as a surface and then subsurface ^eolorist I became a rraduate student at the Ohio State University, where I was a Bovmocker scholar and a graduate assistant.