A Study of a Group of Typical Spinels

Home , Chromite (compound), Spinel group

H ILL IN I S UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN

PRODUCTION NOTE

University of Illinois at Urbana-Champaign Library Large-scale Digitization Project, 2007.

UNIVERSITY OF ILLINOIS BULLETIN IssED TWICB A WEEK VoL XXIX June 17, 1932 No. 84 [Entered as second-class matter December 11, 1912, at the post office at Urbana, Illinois, under the Act of August 24, 1912. Acceptance for mailing at the special rate of postage provided for in section 1103, Act of October 3, 1917, authorised July 31, 1918.]

A STUDY OF A GROUP OF TYPICAL SPINELS

CULLEN W. PARMELEE ALFRED E. BADGER GEORGE A, BALLAM

SBULLETIN No. 248 ENGINEERING EXPERIMENT STATION PuaBLsgaD BY TH UtqVErrITY or ILU4Nox, URBAN

rPaic: TBrIxT CrNTr P'P HE Engineering Experiment Station was established by act of the Board of Trustees of the University of Illinois on De- qember 8, 1903. It is the purpose of the Station to conduct investigations and make studies of importance to the engineering, manufacturing, railway, mining, and other industrial interests of the State. The management of the Engineering Experiment Station is vested in an Executive Staff composed of the Director and his Assistant, the Heads of the several Departments in the College of Engineering, and the Professor of Industrial Chemistry, This Staff is responsible for the establishment of general policies governing the work of the Station, including the approval of material for publication. All members of the teaching-staff of the College are encouraged to engage in scientific research, either directly or in cooperation with the Research Corps composed of full-time research assistants, research graduate assistants, and special investigators. To render the results of its scientific investigations available to the public, the-Engineering Experiment'Station publishes and dis- tributes a series of bulletins. Occasionally it publishes circulars of timely interest, presenting information of importance, compiled from various sources which may not'readily be accessible to the clientele of the Station, and reprints of articles appearing in the technical press written by members of the staff. The volume and number at the top of the front cover page are merely arbitrary numbers and refer to the general publications of the University. Either above the title or below the seal is given the num- S er of the Engineering Experiment Station bulletin, circular, or reprint Swhich should be used in referring to these publications. For copies of publications or for other information address THE ENGINEERIN EXPERIMENT STATION, U NvnRsrry or ILLINOIS, URBANA, ILLtNoIS

.'*.*''r.*'': '.': \* ' ':\ 'l *VN& Iw"s ts- ' ,'.

^fef' f:l^ UNIVERSITY OF ILLINOIS ENGINEERING EXPERIMENT STATION

BULLETIN No. 248 JUNE, 1932

A STUDY OF A GROUP OF TYPICAL SPINELS

CULLEN W. PARMELEE PROFESSOR OF CERAMIC ENGINEERING ALFRED E. BADGER RESEARCH ASSISTANT IN CERAMIC ENGINEERING GEORGE A. BALLAM FORMERLY RESEARCH ASSISTANT IN CERAMIC ENGINEERING

ENGINEERING EXPERIMENT STATION PUBLISHED BY THE UNIVERSITY OF ILLINOIS, URBANA 5 -UNIVE32SI24 5000-6-32--2457 PRE:SSn CONTENTS

PAGE I. INTRODUCTION . . . . . 7 1. Introductory . . . . 7 2. Purpose of Investigation 7 3. Plan of Investigation 7 4. Occurrence of Spinels 8 5. Use of Spinels . . . . 9 6. Acknowledgments. 10

II. SYNTHESIS OF SPINELS 7. Previous Methods of Synthesis . : : : : 8. General Method of Synthesis Used in rThis Investi- gation ...... 11 9. Synthesis of Zinc Aluminate . S . . 12 10. Properties of Zinc Aluminate. 13 . . . . 13 11. Synthesis of Magnesium Aluminate. . . . . 14 12. Properties of Magnesium Aluminate 15 . . . . 15 13. Synthesis of Ferrous Aluminate . S. . . 15 14. Properties of Ferrous Aluminate. . .. . 16 15. Synthesis of Manganous Aluminate. 17 . . . . 17 16. Properties of Manganous Aluminate S. . . 17 17. Synthesis of Zinc Chromite S. . . 18 18. Properties of Zinc Chromite . . .. . 18 19. Synthesis of Magnesium Chromite S . . 19 20. Properties of Magnesium Chromite. . .. . 19 21. Synthesis of Ferrous Chromite . .. . 20 22. Properties of Ferrous Chromite 20 23. Synthesis of Manganous Chromite . . . . 21 24. Properties of Manganous Chromite 21 25. Synthesis of Zinc Ferrite . . . . 26. Properties of Zinc Ferrite. . S. . 22 27. Synthesis of Magnesium Ferrite. 22 28. Properties of Magnesium Ferrite . . . . 23 29. Synthesis of Ferrous Ferrite . . . . . 23 30. Properties of Ferrous Ferrite. . .. . 24 31. Synthesis of Manganous Ferrite. . .. . 25 32. Properties of Manganous Ferrite . .. . 25 CONTENTS (CONCLUDED)

III. DETERMINATIONS OF SCRATCH HARDNESSES OF SYNTHETIC SPINELS ...... 25 33. Scratch Hardness Values of Synthetic Spinels. . 25

IV. DENSITIES OF SYNTHETIC SPINELS ...... 26 34. Method of Determining Densities . . . . . 26 35. Density Values of Synthetic Spinels. . . . . 26

V. X-RAY ANALYSES OF SYNTHETIC SPINELS . . . . . 26 36. Importance of X-ray Diffraction Data . . . . 26 37. Method of Obtaining X-ray Diffraction Data . . 27 38. Computations of Lattice Dimensions of Spinels from X-ray Diffraction Data ...... 27 39. Summary of Determinations of Lattice Dimensions of Synthetic Spinels ...... 28 40. Comparison of Values of Lattice Dimensions Ob- tained by Various Observers ...... 28 41. Computations of Theoretical Densities of Synthet- ic Spinels from X-ray Diffraction Measurements 29 42. Theoretical Densities of Synthetic Spinels . . 31

VI. HEAT CAPACITIES OF SYNTHETIC SPINELS . . . . . 31 43. Method of Determining Heat Capacities . . . 31 44. Experimental Data Concerning Heat Capacities of Synthetic Spinels ...... 37 45. Summary of Heat Capacity Determinations . 37 46. Atomic Heat Capacities of Synthetic Spinels 38

VII. THERMAL EXPANSIONS OF SYNTHETIC SPINELS . . . 38 47. Method of Determining Thermal Expansions . . 38 48. Preparation of Specimens for Thermal Expansion Tests...... 41 49. Thermal Expansion Measurements . . . . . 41 50. Linear Thermal Expansion Coefficients of Synthetic Spinels ...... 41

APPENDIX ...... 45 Tables of Data for Computing Densities, Lattice Di- mensions, Specific Heats, and Thermal Expansions of Spinels LIST OF FIGURES

NO. PAGE 1. Thermal Expansion of Aluminates of Zinc, Magnesium, Iron, and Manganese ...... 17 2. Thermal Expansion of Chromites of Zinc, Magnesium, Iron, and Manganese ...... 19 3. Thermal Expansion of Ferrites of Zinc, Magnesium, Iron, and Manganese ...... 24 4. Lattice Dimensions of Synthetic Spinels ...... 29 5. Densities of Synthetic Spinels as Computed from X-ray Data . . . 31 6. Apparatus for Determination of Heat Capacities of Spinels . . . . 32 7. Effect of Mass of Vitreous Silica on Rise of e.m.f. of Thermoelements in Calorimeter ...... 35 8. Fizeau-Pulfrich Apparatus for Determining Linear Thermal Expansions of Spinels ...... 39

LIST OF TABLES 1. Densities of Powdered Synthetic Spinels ...... 26 2. Lattice Dimensions of Synthetic Spinels ...... 28 3. Lattice Dimensions of Synthetic Spinels ...... 30 4. Computed Densities of Synthetic Spinels ...... 30 5. Molar Volumes of Synthetic Spinels ...... 31 6. Determination of Heat Capacity of Calorimeter ...... 34 7. Mean Heat Capacities of Synthetic Spinels over Temperature Range of 50 to 1025 deg. C...... 38 8. Atomic Heat Capacities of Synthetic Spinels over Temperature Range of 50 to 1025 deg. C...... 39 9. Interferometric Data on Thermal Expansion of Platinum . . . .. 40 10. Thermal Expansion Values of Synthetic Spinels ...... 42 11. Coefficients of Mean Linear Thermal Expansion of Synthetic Spinels . . 43 12. True Specific Gravities and Densities of Synthetic Spinels . . . .. 45 13. Computation of Lattice Dimensions of Aluminates of Magnesium, Zinc, Iron, and Manganese ...... 46 14. Computation of Lattice Dimensions of Chromites of Magnesium, Zinc, Iron, and Manganese ...... 47 15. Computation of Lattice Dimensions of Ferrites of Magnesium, Zinc, Iron, and Manganese ...... 48 16. Determination of Mean Heat Capacities of Synthetic Spinels over Tem- perature Range of 50 to 1025 deg. C ...... 49 17. Interferometric Data on Thermal Expansions of Synthetic Spinels. . . 52

A STUDY OF A GROUP OF TYPICAL SPINELS

I. INTRODUCTION 1. Introductory.-The term spinel denotes a compound which has the type formula R" R2'" 04, and whose structure* may be represented possess the perfect symmetry of the as R" < 0 - R"R"' = 0OO Spinels isometric crystal system, and certain crystals, calcium chromite for example, which have many characteristics in common with the true spinels, are eliminated from the spinel family since they are found to be anisotropic. It is of common knowledge that many members of the spinel family intercrystallize to form one homogeneous crystal. In this investigation the common divalent metals, zinc, magnesium, iron, and manganese, were combined with the trivalent elements, aluminum, chromium, and iron, to form groups of aluminates, chromites, and ferrites, which possess the spinel type of crystal structure.

2. Purpose of Investigation.-Although various members of the spinel group have been synthesized during the past century, little information is available relative to their physical and chemical properties. Complete data are lacking even for common spinel, or magnesium aluminate, which finds extensive use as a refractory. In the cases of ferrous chromite, or chrome-iron ore, and iron ferrite, or magnetite, data are available for the natural minerals, which of necessity differ widely according to the compositions of these compounds. Gahnite, or zinc aluminate, has been described also, but here again discrepancies exist, probably due to the varying compositions of the minerals studied.

3. Plan of Investigation.-Accordingly,a study of a representative group of spinels was undertaken. This group included the aluminates, chromites, and ferrites, of zinc, magnesium, iron, and manganese, which embrace those compounds having refractory properties, and which are encountered frequently in the ceramic industry. Since the investigation was primarily concerned with the properties of spinels, the most convenient methods of syntheses which produced a fairly pure compound were adopted, and in each instance

*J. W. Mellor, Modern Inorganic Chemistry, p. 636, Longmans, Green and Co., London, 1916. ILLINOIS ENGINEERING EXPERIMENT STATION a method was evolved which could be repeated with the assurance of reproducing the desired spinel. A product containing a maximum of five per cent of uncombined oxides or other impurities was considered satisfactory where the difficulties of synthesis precluded a higher degree of purity. In each case the purity of the synthesized compound was verified by X-ray analysis, which generally indicated the presence of considerably less than five per cent of impurity. It has been the plan to report such properties of the synthetic spinels as the densities, lattice dimensions, thermal expansions, and specific heats, with such other miscellaneous information as was obtained during the course of the investigation. The resulting data were correlated in the light of those relations which exist among a group of isomorphous compounds.

4. Occurrence of Spinels.-The spinels form a group of isomorphous compounds which are found in nature as the following minerals :* Spinel ...... MgA120 4 Ceylonite (pleonaste) ...... (Mg.Fe) A120 4 Chlorspinel...... M g (Al-Fe)20 4 Picotite ...... (Mg-Fe)(Al.Fe-Cr) 20 4 Hercynite ...... FeA1AO 4 Gahnite...... ZnA120 4 Disluite...... (Zn-Fe-Mn)(A1-Fe)20 4 Kreittonite ...... (Zn-Fe-Mg) (A1-Fe)04 M agnetite...... FeFe2O4 Magnesioferrite ...... MgFe20 4 Franklinite...... (Fe-Zn-Mn) (Fe-Mn)20 4 Jacobsite...... (Mn-Mg) (Fe.Mn)2 0 4 Chromite...... (Fe.Mg)(Cr.Fe) 20 4

The compositions of the naturally occurring minerals vary greatly due to substitution by isomorphic oxides. Thus, in the case of gahnite, or ZnAl2z 4, if ferrous iron and magnesia replace part of the zinc, and if ferric iron replaces part of the alumina, the isomorphous mixture known as kreittonite, having the formula (Zn.Fe-Mg) (Al-Fe) 20 4, is formed. The mineral franklinite, or (Fe.Zn.Mn) (Fe.Mn) 20 4, represents a similar substitution. The presence of gahnite (ZnAlO 4) in the walls of zinc retorts has been observed by A. Scott,f who stated that "a microsection of the (wall) material shows it to be partly crystalline and partly glassy. The original material of the matrix has been completely changed, the alumina having reacted with oxidized zinc to form spinel, while

*W. S. Bayley, Descriptive Mineralogy, p. 196, Appleton, New York, 1917. tA. Scott, Trans. Ceram. Soc., England, 18, 512, 1918. A STUDY OF A GROUP OF TYPICAL SPINELS the silica appears partly as tridymite and partly as silicates of zinc and iron." Other spinels may be formed under the conditions existing in industrial processes; thus, manganous aluminate is found in slags* although it is not found pure in nature; magnetite, or ferrous ferrite, is a furnace product and forms the iron scale of the blacksmith.t In fact, such controlled conditions are more conducive to the formation of pure spinels than those usually found in nature. The list of naturally occurring spinels previously given attests to this truth, since most of them which occur in nature are isomorphous mixtures of various spinels. Even the common spinel, chromite, is not found pure except in meteorites.T Apparently the controlled conditions available in the laboratory are usually necessary for the formation of pure spinels.

5. Uses of Spinels.-In many metallurgical processes which re- quire refractories with exceptional qualities, advantage is taken of the resistance which spinels offer to the attack of acids, alkalies, and ordinary fusion agents, as well as of the high melting points which they generally possess. Bricks of chromite and magnesia spinel are in common use, and their application would be more widespread were it not for the com- paratively high cost of these refractories. The ordinary magnesia spinel finds a restricted use in certain refractory porcelains. It has been suggested also as a paint pigment. The transparent variety has some value as a gem stone. Magnetite has been cast for use as electrodes in the electrodeposition of copper. Heilmann¶ and Mayer and Havas§ proposed using zinc aluminate as an opacifier in enamels, and took out patents covering this usage. Quoting from a Bureau of Standards report** with regard to this application of the compound: "When artificial spinels of these compositions (ZnAl20 4 and Mg A120 4) are calcined at extremely high temperatures so as to be prac- tically insoluble in molten enamels they make satisfactory opacifying agents. Most of the spinels put on the market have not been rendered sufficiently insoluble and therefore have partially dissolved in the enamel. This has reduced their opacifying power and at the same time raised the fusion point of the enamels quite materially. They produce pure white enamels of high gloss and are both cheap and

*J. Loczka, Z. Kryst. Mineral., 43, 571, 1907. tF. W. Clarke, The Data of Geochemistry, U. S. Geol. Survey, Bul. 770, Washington, Ed. 5, p. 348, 1924. tL. W. Fisher, Am. Mineral., 14, 356, 1929. IE. Heilmann, Brit. Pat., 26, 498, 1912, taken from C. A. 8, 1652, 1914. §M. Mayer and B. Havas, U. S. Pat. 1104266, 1914, taken from C. A. 8, 3106, 1914. *H. F. Staley, Bur. Standards Technol. Paper No. 142, p. 58, 1919. ILLINOIS ENGINEERING EXPERIMENT STATION non-poisonous. The use of spinels as opacifiers has many desirable features, but at present must be considered to be in the experimental stage." The great resistance which zinc aluminate and magnesium aluminate offer to the attack of the fusion agents which are used in their chemical analyses suggests that these spinels might be of value as glass furnace refractories. Furthermore, the slow solution of these compounds would not introduce any deleterious substances in the glass since zinc, aluminum, and magnesium are often incorporated in glass batches. In the series of spinels which were a part of this investigation, the orange color of the ferrites of zinc and magnesium are pleasing to the eye, and might be useful in enamels or glazes. The olive green color of zinc chromite is also agreeable, although it presents a grayish hue resembling suede when viewed under artificial light. Magnesium chromite might be useful also. It is a very dark olive green in color.

6. Acknowledgment.-This investigation is a contribution of the Engineering Experiment Station of the University of Illinois, which is under the administrative direction of DEAN M. S. KETCHUM. The work was done in the Department of Ceramic Engineering, of which PROF. C. W. PARMELEE is the head. The X-ray diffraction patterns of the synthetic spinels were obtained with the equipment of the Chemistry Department, and the computations of their lattice dimensions were made with the advice of PROF. G. L. CLARK. A considerable part of the experimental work involved in the syntheses and the determinations of properties was carried out by ABDE ALLY, whose efficient aid is gratefully recognized.

II. SYNTHESES OF SPINELS 7. Previous Methods of Synthesis.*-There are many methods available for the synthesis of individual spinels, and some of these may be used for the syntheses of all the members of a group of spinels. The "wet" methods of Berzelius,t Sander,t and others were tried and discarded on account of the difficulty experienced in filtering and washing the gelatinous precipitates which were formed. The methods of Daubree,¶ Deville and Caron,§ and other investigators involved the vaporization of metallic fluorides in the presence of lime or boric

*In this section extensive use was made of J. W. Mellor's "A Comprehensive Treatise on Inorganic and Theoretical Chemistry," Longmans, Green and Co., London, 1924. tJ. J. Berzelius, "Lehrbuch der Chemie," 2, 310, Dresden, 1826. tT. Sander, Liebigs Ann. Chem. 9, 181, 1834. ¶A. Daubree, Compt. rend. 39, 135, 1854. §H. Ste.-C. Deville and H. Caron, Compt. rend. 46, 764, 1858. A STUDY OF A GROUP OF TYPICAL SPINELS

acid. These methods presented experimental difficulties for the preparation of any appreciable amounts of spinel, and were discarded. Some spinels may be prepared by direct union of the oxides when heated to high temperatures. This method is applicable when the component oxides are relatively stable, but fails in the cases of zinc, iron, and manganese compounds. A general account of these syntheses may be found in Meunier's treatise* on the subject. More recently Rinnef has described the syntheses of a large series of spinels.

8. General Method of Synthesis Used in This Investigation.-In 1848 M. Ebelmen published a paperS on the syntheses of certain minerals. In this and a subsequent publication¶ he described the syntheses of many spinels by an original method, which consisted in the prolonged heating of the combining oxides in the presence of boric oxide. The boric oxide acted as a flux and a solvent for the oxides which were enabled to combine at a much lower temperature than would be necessary without the admixture of this flux. During the heat treatment the boric oxide was slowly volatilized and the synthetic spinels which were formed by the union of oxides remained in the form of beautiful octahedra, attaining a length of 5-6 mm. in some cases. Ebelmen's method is clearly analogous to the formation of crystals from an aqueous solution when the liquid is allowed to evaporate slowly. His experiments were carried out in the porcelain kilns at Sevres, which provided the heat treatment of 24 to 30 hours which was necessary to volatilize most of the boric oxide. Even under these conditions some boric oxide remained at the end of the test in the form of borates, and these compounds were removed by leaching with hydrochloric acid. Ebelmen prepared the aluminates of magnesium, manganese, iron, beryllium, cadmium, zinc, cobalt, barium, and calcium, besides many chromites and ferrites. Artificial minerals of more diverse nature have since been prepared by Mourelo§ according to Ebelmen's general procedure. In the following investigations a mixture of a divalent and a trivalent oxide with boric acid was heated at temperatures ranging from 900 to 1300 deg. C. according to the compound sought, and the heating was continued for 24 to 48 hours in order to permit the boric oxide to escape. At these temperatures combination of the oxides ensued with the formation of a spinel. The reaction between the

*S. Meunier, "Les M6thodes de Synthese en Mingralogie," Baudry, Paris, 1891. tF. Rinne, Fortschritte Mineral. Kryst. Petrog. 12, 68, 1927, taken from C. A. 23, 1842, 1929. IM. Ebelmen, Ann. chim. phys. 3, 22, 211, 1848. ¶M. Ebelmen, ibid., 3, 33, 34, 1851. §J. R. Mourelo, Rev. g6n. sci. 26, 394, 1915. ILLINOIS ENGINEERING EXPERIMENT STATION oxides was accompanied by a change in color and a considerable shrinkage. The mixture of oxides was usually pressed into briquets to which a little dextrin solution was added for a binder, and the briquets were placed on a layer of oxide corresponding to the basic oxide of the spinel in order to avoid contamination by the furnace refractory. In the syntheses in which the reaction took place with great facility, the oxides were introduced in molecular proportions. If experience showed that the synthesis proceeded so slowly that appreciable amounts of uncombined oxides were likely to remain in the product, quite a large excess of the basic oxide was introduced in order that no acidic oxide remain after the synthesizing period, since any uncombined acidic oxides, notably A12 0 3 and Cr20 3, could be removed by acid leaching only with difficulty. On the conclusion of the heat treatment necessary for the synthesis the excess basic oxide was dissolved in the acid used for the leaching treatment. The leaching operation was made on the pulverized synthetic material which had been passed through a 100-mesh sieve. The concentration of the acid was varied from ten per cent upwards according to the material under treatment, and the leaching was continued on a steam bath for about 24 hours. Details concerning the syntheses of the individual spinels are considered under the separate headings.

Zinc Aluminate (ZnAl20 4) 9. Synthesis of Zinc Aluminate.-J. J. Berzelius* found that aluminum hydroxide abstracts zinc oxide from an aqueous solution of ammoniacal zinc oxide, and that zinc aluminate is precipitated from the same solution by adding a saturated solution of potassium aluminate. T. Sandert reported a precipitate of zinc aluminate on mixing solutions of potassium aluminate and ammonium zincate. J. J. Ebel- ment and J. R. Mourelo¶ synthesized gahnite by heating a mixture of zinc oxide, alumina, and boric oxide at a bright red heat for 18 hours in a porcelain muffle. The same product was obtained by S. Meunier§ on heating to bright redness a mixture of zinc oxide, alumina, cryolite, and aluminum chloride in a graphite crucible. A. Daubree** made gahnite by passing vapors of zinc and aluminum chlorides over red hot lime. H. Ste.-C. Deville and H. Carontf vaporized zinc and aluminum fluorides in the presence of boric oxide. *J. J. Berzelius, loc. cit. tT. Sander, loc. cit. J. J. Ebelmen, loc. cit. J. R. Mourelo, loc. cit. §S. Meunier, loc. cit. **A. Daubree, loc. cit. ttH. Ste.-C. Deville and H. Caron, loc. cit. A STUDY OF A GROUP OF TYPICAL SPINELS

The method of synthesis which was adopted was carried through as follows: Zinc and aluminum oxides were weighed out in the molecular proportions of 1.5 : 1, and half of the total weight of boric acid was added. The three substances were then ground intimately in a mortar. A small amount of dextrin solution was mixed in to act as a binder, and the mix was briquetted in a small press. The briquets were dried over night at 110 deg. C. and then heated at 1115-1130 deg. C. in a gas-fired muffle for about 24 hours. An oxidizing atmosphere was maintained to minimize the reduction of ZnO and its consequent loss by volatilization. A chemical analysis of the briquet after concluding the heat treatment showed that little uncombined ZnO remained. The crude spinel was crushed and ground to pass through a 100- mesh screen, and the resulting powder was then digested with twenty- per-cent HC1 for 48 hours on a steam bath, in order to remove any uncombined ZnO. After the purified spinel had been washed and dried, microscopic examination showed small inclusions of a birefringent material (probably aluminum borate) in the spinel crystals. The material was therefore ground still further and digested for several days in hydrochloric acid as before, when a very pure product was obtained as proven by X-ray analysis.

10. Properties of Zinc Aluminate.-The pure zinc aluminate was thus obtained in the form of a white powder, of hardness 7.5 on Mohs' scale. Its density at 20 deg. C. was found to be 4.54 grams per cu. cm., which agrees well with the value of 4.55 given by Larsen* and Ebelmen's value of 4.58 as listed by Groth.t X-ray analysis showed that it has a unit cell dimension of 8.062 A.U., corresponding to a theoretical density of 4.615 grams per cu. cm. Its mean heat capacity from 50 to 1025 deg. C. is 0.92 joule per gram per deg. C. The coefficient of mean linear thermal expansion from 25 to 900 deg. C. is 91 X 10- 7 cm. per cm. per deg. C. Figure 1 depicts the course of the thermal expansion curve as obtained from interferometric data,t and expansion coefficients over intervals below 900 deg. C. may be found in Table 11. Zinc aluminate resists the attack of fusion agents, such as sodium carbonate, potassium bisulphate, and sodium peroxide, but dissolves in a mixture of sodium carbonate and borax. When subjected to the test described in the following, it was found to dissolve somewhat in concentrated H2SO 4, but to be insoluble in other acids and alkalies.

*E. S. Larsen, "The Microscopic Determination of the Nonopaque Minerals," U. S. Geol. Survey Bul. No. 679, Washington, p. 180, 1921. tP. Groth, Chemische Krystallographie, Vol. 2, p. 753, Engelmann, Leipzig, 1908. JSee p. 42. ILLINOIS ENGINEERING EXPERIMENT STATION

Zinc aluminate, ground to pass a 300-mesh screen, was washed in hydrochloric acid to remove impurities introduced in grinding and screening. After the material had been calcined, samples of one gram each were placed in large Pyrex test tubes, and to each was added 100 ml. of the following reagents: (1) Hydrochloric acid, concentrated. (2) Nitric acid, concentrated. (3) Aqua regia, mixture of above acids. (4) Sulphuric acid, 60 deg. Be. (5) Sodium hydroxide, 30 per cent solution. (6) Hydrofluoric acid, concentrated (test tube ceresine coated). The tubes were allowed to stand for 30 days, being shaken occasionally. The sulphuric acid and sodium hydroxide solutions were heated periodically. The other solutions were kept at constant volume by periodic additions of the concentrated reagents. At the expiration of this time the solutions were examined for zinc, controls being examined simultaneously. In the case of the sulphuric acid a zinc content of 0.0015 g. was found; all other titrations corresponded to the blanks. Other properties of zinc aluminate which have been reported in- clude the index of refraction, which Larsen* lists as 1.80 .

Magnesium Aluminate (MgAl20 4) 11. Synthesis of Magnesium Aluminate.-Since both MgO and A1203 are stable at high temperatures, the general method used to form synthetic magnesia spinel is to heat a mixture of these oxides to a temperature near 1700 deg. C., when combination occurs. The addition of boric oxide permits the use of lower temperatures. The batch used in the synthesis of MgA1204 consisted of MgO, 181 parts by weight Al20s, 306 parts by weight HaBOs, 244 parts by weight which signifies a molecular ratio of MgO to A120 3 of 1.5 to 1. The batch was ground intimately in a porcelain mortar and then mois- tened with a ten-per-cent dextrin solution. Briquets were pressed from this mass and dried at 110 deg. C. They were then placed in a crucible, the bottom of which was covered with MgO to prevent contamination of the batch by the crucible material. The mixture was heated in a gas furnace for 36 hours at temperatures of 1275 to 1300 deg. C. On the conclusion of the heat treatment the synthetic material was ground in an iron mortar to pass through a 100-mesh sieve. Any iron

*E. S. Larsen, loc. cit., p. 178. A STUDY OF A GROUP OF TYPICAL SPINELS introduced by this grinding was removed with a magnet, and the powdered material was then leached with twenty-per-cent HC1 for 20 hours on a steam bath. After the leaching treatment the purified spinel was washed and calcined to 350 deg. C. X-ray analysis showed a high degree of purity of this product. 12. Properties of Magnesium Aluminate.-Pure magnesia spinel was obtained as a white powder resembling zinc aluminate in its appearance. Its scratch hardness is about 7.5. Its density at 20 deg. C. was found by pycnometer measurements to be 3.53 grams per cu. cm. which agrees with Larsen's* value of 3.6 ±. The side of its unit cell is 8.086 A.U., and from this value a theoretical density of 3.548 grams per cu. cm. may be computed. Rinnet found a corresponding value of 3.5775. Its mean heat capacity from 50 to 1025 deg. C. is 1.19 joules per gram per deg. C. Table 11 lists the thermal expansion data, from which its coefficient of mean linear thermal expansion is a9 = 81 X 10 - 7 cm. per cm. per deg. C. Figure 1 illustrates the course of the thermal expansion curve. Other properties of magnesium aluminate are its index of refraction, which is given by Larsent as 1.723±. Its fusion point is given by Winchell¶ as 2135 deg. C. This authority states that it is slowly soluble in H 2S0 4, but insoluble in most other reagents. A recent determination§ of the mean thermal expansion of a commercial grade magnesia spinel brick was found to be 86 X 10- 7 cm. per cm. per deg. C. between 20 and 1400 deg. C. Rieke and Blicke** give a mean coefficient of 67 X 10- 7 between 25 and 800 deg. C. on a 7-cm. sample.

Ferrous Aluminate (FeA120 4) 13. Synthesis of Ferrous Aluminate.-Considerabledifficulty was encountered in the synthesis of ferrous aluminate. Some preliminary experiments, which were made with a mixture of ferrous scale, aluminum oxide, and boric oxide, were unsuccessful, due to the tendency of magnetite to form in preference to ferrous aluminate. The synthesis of ferrous aluminate was finally effected by placing a briquet of these materials on a layer of metallic iron and enclosing the whole in a sagger to minimize oxidation. However, the conditions during this synthesis were not definitely, known, and additional experiments were made in an attempt to formulate a more satisfactory procedure.

*E. S. Larsen, loc. cit., p. 177. tF. Rinne, Neues Jahrb. Mineral. Geol. 58 A, 70, 1928. tE. S. Larsen, loc. cit., p. 177. ¶N. H. and A. N. Winchell, Elements of Optical Mineralogy, Part 2, Ed. 2, p. 63, Wiley, New York, 1927. Bur. Standards Technical News Bul. No. 152, p. 121, 1929. *R. Rieke and K. Blicke, Ber. deut. keram. Ges. 12, 181, 1931. ILLINOIS ENGINEERING EXPERIMENT STATION

If a mixture of A1203 with powdered metallic iron were heated at 1200 deg. C. in a current of steam, this treatment should produce FeO, which should in turn react with the A120 3 present to form FeAl20 4. This experiment did not result in the formation of the desired spinel, but the lack of success might have been due to the fact that air was not excluded perfectly from the charge. Another set of experiments was made, using a mixture of powdered iron, magnetite, and aluminum oxide, in the proper proportions to form FeAl20 4. These materials were heated in an iron bomb in order to exclude air, but the results of the test were not encouraging. Successful results were finally obtained by the use of the following batch:

Fe203, 479.1 parts by weight Al, 54.2 parts by weight A120 3, 511.0 parts by weight The finely powdered materials were placed in a covered crucible and heated for about 32 hours in a gas furnace in a strongly reducing atmosphere. The temperature as measured by an optical pyrometer varied from 1320 to 1420 deg. C. Under these conditions the following reaction would be expected:

2 Al + 3 Fe20 3 = 6 FeO + A120s,

and since ferrous oxide is a violent flux, it should unite with the Al 20 3 present to form FeAl20 4. The proportions in the batch were so chosen that no excess oxides were present. The material synthesized according to these directions was obtained as a crumbly black mass. It was ground in an iron mortar to pass through a 120-mesh sieve, and was then digested with twenty- per-cent HCI on a steam bath. The washed and dried product was then analyzed by X-rays to prove its purity. 14. Propertiesof Ferrous Aluminate.-Pure ferrous aluminate was prepared as a black powder having a hardness of about 7.5. At 20 deg. C. its density is 4.22 grams per cu. cm., which compares with a value of 3.9 given by Larsen.* Its theoretical density was computed as 4.392 on the basis of a lattice dimension of 8.119 A.U. Its mean heat capacity from 50 to 1025 deg. C. is 1.04 joules per gram per deg. C. Table 11 shows that the coefficient of mean linear thermal expansion from 25 to 850 deg. C. is 90 X 10- 7 cm. per cm. per deg. C. Figure 1 shows the course of the thermal expansion curve.

*E. S. Larsen, loc. cit., p. 178. A STUDY OF A GROUP OF TYPICAL SPINELS

/00

S80 ------.

Pb kF A/c .mleli

0 /00 00 300 400 500 600 700 800 900 Temp7erafcre /in de. C..

FIG. 1. THERMAL EXPANSION OF ALUMINATES OF ZINC, MAGNESIUM, IRON, AND MANGANESE

The index of refraction of ferrous aluminate (hercynite) is given as 1.80± by Larsen.*

Manganous Aluminate (MnAi20 4) 15. Synthesis of Manganous Aluminate.-The batch used for the synthesis of manganous aluminate consisted of

MnCO 3, 227 parts by weight Al203, 153 parts by weight which represents an excess of MnO over the theoretical combining proportions. It was not necessary to use a flux of boric oxide in the synthesis of this spinel. The charge was heated in an electric furnace for 24 hours at 1000 ± 15 deg. C. Microscopic examination then showed inhomo- geneities in the batch, so the substance was reground and subjected to the same heat treatment. After this second heat treatment the crude spinel was ground to a powder and leached with ten-per-cent HC1 to remove the excess manganese compound. X-ray analysis did not show diffraction patterns of any impurity in the remaining powder.

16. Properties of Manganous Aluminate.-Pure manganous aluminate was prepared as a chocolate-colored powder with a scratch hardness of about 7.5. Its density at 20 deg. C. was determined to

*E. S. Larsen, loc. cit., p. 178. ILLINOIS ENGINEERING EXPERIMENT STATION be 4.01 grams per cu. cm., which compares with Winchell's value* of 4.05 grams per cu. cm. and that of 4.12 as listed in the International Critical Tables.t The theoretical density is 4.031 grams per cu. cm., as computed from a unit cell dimension of 8.271 A.U. Its mean heat capacity is 0.89 joule per gram per deg. C. over the temperature range of 50 to 1025 deg. C. Its expansion curve is shown in Fig. 1, and Table 11 states that the coefficient of mean linear thermal expansion from 25 to 900 deg. C. is 73 X 10- 7 cm. per cm. per deg. C. Few properties of this spinel are known. Winchellt lists the index of refraction as about 1.8.

Zinc Chromite (ZnCr20 4) 17. Synthesis of Zinc Chromite.-Zinc chromite was synthesized by Ebelmen¶ in 1851 by heating zinc and chromic oxides together to a white heat. The batch which was used in this investigation consisted of an intimately ground mixture of ZnO and Cr20 3 in the molecular proportions of 1.5 : 1, with the addition of half their combined weights of H3 B0 3 . This mixture was pressed into briquets after the addition of a small amount of ten-per-cent dextrin solution. The briquets were fired in a gas furnace for about 40 hours at 1000-1100 deg. C., and they were then powdered and leached on a hot plate with twenty-per-cent HC1 for about 24 hours. X-ray analysis showed a pure product. 18. Properties of Zinc Chromite.-An olive-green-colored powder with an unctuous feel was obtained in the synthesis of zinc chromite. Its hardness is about 6.5 on Mohs' scale. At 20 deg. C. it has a density of 5.30 grams per cu. cm., a value which agrees closely with Ebelmen's old value of 5.309. A lattice dimension of 8.296 A.U. corresponds to a theoretical density of 5.393 grams per cu. cm. The mean heat capacity from 50 to 1025 deg. C. is 0.77 joule per gram per deg. C. The coefficient of mean linear thermal expansion is 920= 101 X 10 - 7 cm. per cm. per deg. C., as shown by Fig. 2 and Table 11. This spinel was found to be infusible in sodium carbonate or in a mixture of sodium carbonate and borax. When fused with potassium bisulphate, a residue of Cr 203 remained which was dissolved by prolonged fusion with sodium peroxide.

*N. H. and A. N. Winchell, loc. cit., p. 63. tI. C. T. I, 137. RN.H. and A. N. Winchell, loc. cit., p. 63. [M. Ebelmen, loc. cit. A STUDY OF A GROUP OF TYPICAL SPINELS

/00 ---

8ognes/'um Chrom/ir / , . Manglanous (C'/rom'Iwe

7 ^ ^ Ferrous C/,rom/f7/e 40 --

0 /00 200 300 400 -00 600 700 800 900 Te-'pera'/re //7 adeg'. C.

FIG. 2. THERMAL EXPANSION OF CHROMITES OF ZINC, MAGNESIUM, IRON, AND MANGANESE

Magnesium Chromite (MgCr2O 4) 19. Synthesis of Magnesium Chromite.-The batch used in the synthesis of MgCr20 4 consisted of MgO, 121.0 parts by weight CrOa, 456.0 parts by weight HsB0 3, 288.5 parts by weight These amounts represent molecular proportions of MgO and Cr20O, which were mixed with half their combined weights of H3BO 3 by grinding together in a mortar. A small amount of ten-per-cent dextrin solution was added as a binder for the powdered material, and the mass was formed into briquets in a small press. The briquets were then dried at 110 deg. C. and heated to 1275-1325 deg. C. for 18 hours to effect the combination of the oxides. The synthetic product was crushed and ground to pass through a 100-mesh sieve, and then leached with twenty-per-cent HC1 on a steam bath for 18 hours. The purified spinel was then dried at 350 deg. C., and X-rayed to prove its purity.

20. Properties of Magnesium Chromite.-Pure magnesium chromite powder has a deep olive green color with plainly visible crystalline spangles. It has a scratch hardness of about 6.5 on Mohs' scale. Its actual density at 20 deg. C. is 4.41 grams per cu. cm. which agrees with Groth's value* of 4.415. The theoretical density as obtained from X-ray measurements is 4.429 grams per cu. cm., which corresponds

*P. Groth, loc. cit., p. 750. ILLINOIS ENGINEERING EXPERIMENT STATION to a unit cell dimension of 8.305 A.U. The average heat capacity over the temperature interval of 50 to 1025 deg. C. is 0.92 joule per gram per deg. C. Its coefficient of mean linear thermal expansion is a90 = 93 X 10- 7 cm. per cm. per deg. C., as shown by Table 11 and Fig. 2. Magnesium chromite was found to resist the attack of fluxing agents, but fusion was completed in two hours using a mixture of sodium carbonate and borax.

Ferrous Chromite (FeCr20 4) 21. Synthesis of Ferrous Chromite.-Great difficulty was encountered in synthesizing this commonly occurring mineral. After the usual boric oxide method failed, it was hoped that a highly reducing atmosphere might aid in the formation of ferrous chromite. These hopes were in vain, and magnetite was usually formed instead of chromite. Finally FeCr 20 4 was synthesized by heating a mixture of FesO 3 and Cr203 in molecular proportions to a temperature ranging from 1600 to 1675 deg. C. for 3 hours in a surface combustion furnace. X-ray analysis showed about 95 per cent of FeCr20 4. These troubles in synthesizing chromite should be expected from the experiences of other investigators. For instance, Meunier used an ingenious method of synthesis by oxidizing an alloy of iron and chromium in which these metals were in molecular proportions. 22. Properties of Ferrous Chromite.-Pure ferrous chromite is a brownish-black substance of hardness about 6. Its density at 20 deg. C. was found to be 4.93 grams per cu. cm. by the pycnometer method. Larsen* gives a value of 4.5 for the density. Its lattice dimension was found to be 8.344 A.U. from which the theoretical density was computed as 5.085 grams per cu. cm. Its mean heat capacity over the temperature interval of 50 to 1025 deg. C. is 0.82 joule per gram per deg. C. Table 11 shows that the coefficient of mean linear thermal expansion is as0 = 85 X 10- 7 cm. per cm. per deg. C. This value agrees quite well with Austin's determinationt of a80 = 82.5 X 10- 7, which applies to a sample of natural chromite. The course of the thermal expansion curve of synthetic ferrous chromite is depicted in Fig. 2. The value of the index of refractionS for natural minerals varies from 2.07 to 2.16.

*E. S. Larsen, loc. cit., p. 181. tJ. B. Austin, J. Am. Ceram. Soc. 14, 807, 1931. TE. S. Larsen, loc. cit., p. 181. A STUDY OF A GROUP OF TYPICAL SPINELS

Manganous Chromite (MnCr20 4) 23. Synthesis of Manganous Chromite.-The batch used for the production of manganous chromite contained no addition of boric oxide flux. It was as follows:

MnCO 3, 172.5 parts by weight Cr 20 3 , 152.0 parts by weight

A 1.5 :1 molar ratio of MnO : CrsO 3 is represented in this batch which was briquetted in the customary fashion. The charge was fired in an electric furnace for 24 hours at 1000 + 15 deg. C. Microscopic examination showed that the batch was not altogether homogeneous, so the briquets were reground, formed into briquets again, and refired to the same extent. The briquets were then ground to pass through a 100-mesh sieve and leached with ten-per-cent HC1 to remove the excess manganese compound. X-ray examination then showed a pure product. 24. Properties of Manganous Chromite.-Pure manganous chromite was obtained as a chocolate brown powder of hardness about 6.5. Its density at 20 deg. C. is 4.80 grams per cu. cm. The International Critical Tables* list a value of 4.87. The side of its unit cell is 8.436 A.U., from which a theoretical density of 4.900 grams per cu. cm. may be computed. Its mean heat capacity over the temperature range from 50 to 1025 deg. C. is 0.74 joule per gram per deg. C. Figure 2 shows the course of its thermal expansion curve, and Table 11 indicates a coefficient of mean linear thermal expansion equal to 89 X 10- 7 cm. per cm. per deg. C. from 25 to 900 deg. C.

Zinc Ferrite (ZnFe20 4) 25. Synthesis of Zinc Ferrite.-Zinc ferrite was prepared by Swartz and Krauskopft by heating an intimate mixture of ZnO and Fe203 at about 650 deg. C. For the synthesis of zinc ferrite the following batch was weighed out: ZnO, 244.1 parts by weight Fe20, 479.0 parts by weight H3BOa, 362.0 parts by weight which represents molecular amounts of zinc oxide and ferric oxide with the addition of half of the total weight of boric acid. The substances were ground together in a porcelain mortar. A small amount of ten-per-cent dextrin solution was mixed into this mass to act as a

*I. C. T. I, 133. tC. E. Swartz and F. K. Krauskopf, Mining Met. 9, 28, 1928. ILLINOIS ENGINEERING EXPERIMENT STATION binder, and briquets were then formed in a small press. These were dried at 110 deg. C. and heated for 18 hours at 1100 deg. C. The fired product was ground so as to pass through a 100-mesh sieve and leached with ten-per-cent HC1. X-ray analysis was used in order to prove the purity of the final product. 26. Properties of Zinc Ferrite.-Purezinc ferrite obtained as a powder is deep orange in color, and has a scratch hardness of about 5.5. It is non-magnetic and thus differs from the other ferrites which were studied, since all the latter show strong ferromagnetism. Its density at 20 deg. C. is 5.27 grams per cu. cm., and its theoretical density was computed as 5.322 grams per cu. cm. from a lattice dimension of 8.423 A.U. Groth* lists variations in densities of various samples as ranging from 5.132 to 5.33. The average heat capacity was found to be 0.73 joule per gram per deg. C. from 50 to 1025 deg. C. The coefficient of mean linear thermal expansion is 99 X 10- 7 over the temperature interval of 25 to 850 deg. C. as shown by Table 11 and Fig. 3. This spinel is appreciably soluble in concentrated acids, and is readily fusible in potassium bisulphate.

Magnesium Ferrite (MgFe20 4) 27. Synthesis of Magnesium Ferrite.-This spinel was made by H. Ste.-C. Devillet who heated magnesia and ferric oxide in a stream of HC1. It was also formed by Forestier and Chaudront by precipita- tion with NaOH from a neutral mixture of ferric and magnesium chlorides. In the present series of experiments some high temperature tests were made in attempts to form MgFe20 4 from a mixture of the oxides, but it seemed that magnetite formed in preference to magnesium ferrite. The following low temperature experiment was successful: The batch consisted of powdered MgO and Fe20 3 in the molecular ratio of 1.5 : 1. The blend of oxides was mixed with a small amount of dextrin solution and pressed into a briquet. The briquet was dried and then heated to 960 deg. C. for 22 hours in an electrically heated furnace. After this heat treatment the briquet, consisting of MgFe204 with excess MgO, was ground to a powder which was passed through a 100-mesh sieve and leached with twenty-per-cent hydrochloric acid for 18 hours on a steam bath, and then the purified spinel was washed and dried at 110 deg. C. X-ray analysis proved its purity.

*P. Groth, loc. cit., p. 754. tH. Ste.-C. Deville, loc. cit. 1H. Forestier and G. Chaudron, Compt. rend. 181, 509, 1925. A STUDY OF A GROUP OF TYPICAL SPINELS

28. Properties of Magnesium Ferrite.-Powderedmagnesium ferrite resembles zinc ferrite in color, being a deep orange. It differs markedly from zinc ferrite in its magnetic properties, being strongly ferromagnetic. Its hardness is about 5.5 on Mohs' scale. An actual density of 4.61 grams per cu. cm. may be compared with a theoretical value of 4.506 computed from a lattice dimension of 8.366 A.U. Winchell* lists a specific gravity of 4.6. The average heat capacity is 0.86 joule per gram per deg. C. from 50 to 1025 deg. C. Figure 3 shows the course of the thermal expansion curve, and Table 11 affords a value of the coefficient of mean linear thermal expansion of 129 X 10-7 cm. per cm. per deg. C. from 25 to 900 deg. C. The index of refraction of magnesium ferrite is very high, as Winchellt lists a value of nLi = 2.35.

Ferrous Ferrite (FeFe20 4) 29. Synthesis of Ferrous Ferrite.-Magnetite, or ferrous ferrite, was synthesized by Ebelmen,t who prepared it by fusing an iron silicate and lime. H. Ste.-C. Deville¶ prepared magnetite by heating ferrous oxide in a stream of HC1. M. Sidot§ obtained magnetite by the calcination of ferric oxide alone, and his method was applied in the following preparation of magnetite from ferric oxide. When Fe20 3 is heated in air above about 1348 deg. C. it loses oxygen, giving products intermediate in composition and approaching magnetite as the heating continues and the temperature rises.** The dissociation pressure becomes one atmosphere at 1430 deg. C. Small briquets were pressed from pure Fe203 without the addition of any binding material. These were placed in a gas-fired furnace and heated for five hours at 1350 to 1475 deg. C. At the end of the heat treatment an iron pipe was thrust into the crucible which contained the briquets, and nitrogen gas from a cylinder was blown on the briquets in an attempt to cool them rapidly in a neutral atmosphere. Also the crucible was removed from the furnace, in order to aid in the rapid cooling of the charge. When the briquets were at a dull red heat they were dumped out of the crucible, since the latter retained a high temperature for some time. The whole cooling operation did not take over five minutes, and the cooling through the sensitive high temperature range where oxidation occurs readily must have been

*N. H. and A. N. Winchell, loc. cit., p. 63. tN. H. and A. N. Winchell, ibid. M. Ebelmen, Compt. rend. 33, 525, 1851. TH. Ste.-C. Deville, Compt. rend. 53, 199, 1861. §M. Sidot, Compt. rend. 69, 201, 1869. **0. C. Ralston, U. S. Bur. Mines Bul. No. 296, p. 8, Washington, 1929. ILLINOIS ENGINEERING EXPERIMENT STATION

/40

K/'-/

3rFerri Ferr u ferrof - 806 / 'ni ap / -iFerrc /-

40 0L

00 /00 200 300 4602 5620 600 702 802 9 2T0dperrwoon1 /t deg. C.

FIG. 3. THERMAL EXPANSION OF FERRITES OF ZINC, MAGNESIUM, IRON, AND MANGANESE very rapid. After the briquets cooled they were ground to a powder in an iron crusher and later leached with dilute hydrochloric acid. The purity of the magnetite was verified by X-ray analysis. 30. Properties of Ferrous Ferrite.-Pureferrous ferrite was obtained in the form of a sparkling black powder. This compound is the commonly known mineral magnetite, and it shows strong ferromagnetism. It has a scratch hardness of about 5.5. Its density at 20 deg. C. was found to be 5.18 grams per cu. cm. Winchell* lists a value of 5.17 and Grotht gives it as 5.2-5.25. The side of its unit cell is 8.374 A.U., from which its theoretical density was found to be 5.201 grams per cu. cm. The mean heat capacity was found to be 0.96 joule per gram per deg. C. over the temperature range of 50 to 1025 deg. C. It is of interest to note that FeO30 4 undergoes an inversionS at 593 deg. C. and the resulting heat effect is included in the average value given. This temperature¶ is closely associated with the Curie

N. H. and A. N. Winchell, loec. cit., p. 64. tP. Groth, loc. cit., p. 753. 0. C. Ralston, loc. cit., p. 68. ¶O. C. Ralston, ibid. A STUDY OF A GROUP OF TYPICAL SPINELS point, beyond which magnetite becomes non-magnetic. Figure 3 shows the course of its thermal expansion curve, and Table 11 lists the values of the coefficients of mean linear thermal expansion, from which a ',0 = 151 X 10- 7 cm. per cm. per deg. C.

Manganous Ferrite (MnFe20 4) 31. Synthesis of Manganous Ferrite.-The spinel manganous ferrite was readily synthesized by heating the following batch in an electric furnace for 40 hours at about 1000 deg. C. The relatively low temperature is advisable since magnetite tends to form at more elevated temperatures. The batch consisted of MnCO3, 459.6 parts by weight Fe203, 479.1 parts by weight H3BOs, 346.0 parts by weight The excess manganese in the resulting spinel, as well as any borates present, were removed by digesting with HC1. X-rays were used to verify the purity of the final product. 32. Properties of Manganous Ferrite.-Pure manganous ferrite was obtained as a black powder, strongly ferromagnetic, of hardness about 5.5. Its density at 20 deg. C. is 4.74 grams per cu. cm., which agrees well with Winchell's value* of 4.75. The side of the elementary unit cell is 8.457 A.U. from which a theoretical density of 5.029 grams per cu. cm. was computed. Its mean heat capacity is 0.87 joule per gram per deg. C. from 50 to 1025 deg. C. Figure 3 and Table 11 show the thermal expansions at various temperatures, from which 7 2a = 83 X 10- cm. per cm. per deg. C.

III. DETERMINATIONS OF SCRATCH HARDNESSES OF SYNTHETIC SPINELS

33. Scratch Hardness Values of Synthetic Spinels.-By using minerals of known hardness values, the scratch hardnesses of the series of synthetic spinels were determined approximately. The results indicated that the ferrites of Zn, Mg, Fe, and Mn all have a Mohs' scale hardness number of about 5.5; the chromites of these metals have values near 6.5, while the aluminates attain a hardness of about 7.5. These approximate values are in inverse order to the size of the lattice dimensions, e.g., the aluminates have the smallest-and hence the most compressed-lattices, and this results in greater hardness as a rule.

*N. H. and A. N. Winchell, loc. cit., p. 64 ILLINOIS ENGINEERING EXPERIMENT STATION

TABLE 1 DENSITIES OF POWDERED SYNTHETIC SPINELS Determined by pycnometer measurements at 20 deg. C.

Density grams per cu. cm. Divalent Element Aluminates Chromites Ferrites

Zn 4.54 5.30 5.27 Mg 3.53 4.41 4.61 Fe 4.22 4.93 5.18 Mn 4.01 4.80 4.74

IV. DENSITIES OF SYNTHETIC SPINELS 34. Method of Determining Densities.-The synthetic spinels were powdered to 100-120 mesh and their true specific gravities at 20 deg. C. obtained by means of pycnometer measurements according to a standard test method.* The densities were then computed from these values by multiplying them by the density of water at this temperature. Table 12 in the Appendix lists the true specific gravities and the densities of the spinels. The latter values are reported only to the second decimal place, since the purity of the spinels probably does not warrant a greater precision.

35. Density Values of Synthetic Spinels.-Table 1 provides a summary of the densities at 20 deg. C. of synthetic spinels as obtained by pycnometer measurements.

V. X-RAY ANALYSES OF SYNTHETIC SPINELSf 36. Importance of X-ray Diffraction Data.-In the syntheses of spinels, certain oxides were caused to combine to form the chemical compounds known as spinels. The extent to which the oxides combined, i.e., the purity of the spinels, could not well be judged by chemical analyses, since the leaching out of any uncombined oxides might not have been complete. The absence of impurities in the synthetic spinel could be checked somewhat by microscopic examination, since any birefringent material would indicate an impure product. How- ever, microscopic examination of spinels is beset with difficulties due to the. opacity of these compounds. The most reliable tool in veri-

*Standard Test Method No. 52, J. Am. Ceram. Soc. 11, 453, 1928. tA more complete account of these measurements may be found in G. L. Clark, Abde Ally, and A. E. Badger, Am. J. Sci. 22, 539, 1931. A STUDY OF A GROUP OF TYPICAL SPINELS

fying the nature of any compounds which were present is found in X-ray analysis, and the purity of all the spinels which were used in this investigation was checked by this means. In addition to the aid rendered by X-rays in verifying the purity of the synthetic spinels, the value of this method of analysis is en- hanced by the fact that the X-ray diffraction data may be used to obtain the length of a side of the elementary unit in the spinel crystal, i.e., its lattice dimension. Such data may be used in computing the theoretical densities of the materials. A consideration of these subjects is undertaken in the following account. 37. Method of Obtaining X-ray Diffraction Data.-The Hull powder method was used to obtain the X-ray diffraction data in conjunction with a multiple unit apparatus manufactured by the General Electric Company. A molybdenum target tube was utilized to produce X-radiation having the wave length 0.712 A.U. The term d - in the fundamental diffraction equation,* n d X - = , (1) n 2 sin ( was obtained with the aid of a scale which is provided with the diffraction apparatus.

38. Computations of Lattice Dimensions of Spinels from X-ray Diffraction Data.-Tables 13, 14, and 15 in the Appendix show the results of the measurements of the X-ray diffraction patterns, the d values of - for each of the twelve spinels being placed in the first n column under each compound. The estimated relative intensities of the interference lines are placed in the second columns. If the Miller indices of the reflecting crystal planes are denoted h, k, 1, it is well known that in the cubic system the ratio of the values of sin2o for any two planes, as computed from Equation (1), is equal to the ratio of the values of (h2 + k + 12)n 2 for these two planes, where n is the order of reflection. Column 3 under each spinel lists the ratios of sin 26 which have been chosen to match the expected values of (h2 + k2 + 12)n2. A fundamental law of crystallography states that the quantity (h2 + k2 + 2)n2 must be an integer. These integral values are found in column 2 of each table, and the indices of the planes to which they apply are placed in the first columns.

*G. L. Clark, Applied X-rays, p. 133, McGraw-Hill, New York, 1927. ILLINOIS ENGINEERING EXPERIMENT STATION

TABLE 2 LATTICE DIMENSIONS OF SYNTHETIC SPINELS

Lattice Dimensions Angstrom units Divalent Element Aluminates Chromites Ferrites

Zn 8.062 ± 0.001 8.296 + 0.002 8.423 ± 0.001 Mg 8.086 ± 0.003 8.305 ± 0.001 8.366 ± 0.001 Fe 8.119 + 0.002 8.344 ± 0.003 8.374 ± 0.003 Mn 8.271 ± 0.002 8.436 ± 0.001 8.457 ± 0.002

The deviations from whole numbers which are observed in column 3 under each spinel are due to the inexactness of the measurements. In the computation of the size of the unit cell, ao, as found in the fourth column under each compound, the integral values in column 2 of each table were substituted in the formula*

a, = - (h2 + k2 + 12)n 2. n

39. Summary of Determinations of Lattice Dimensions of Synthetic Spinels.-The mean values of the lattice dimensions, ao, with the probable errors of the means which result from the use of Student'st factors, are summarized in Table 2. The table shows that the lattice dimensions of chromites are generally intermediate between the lower values of the corresponding aluminates and the higher ferrite values. The unit cell of zinc ferrite is an exception to this rule. This spinel exhibits another anomaly in that it is non-magnetic while the other ferrites show ferromagnetism. The table shows that with the single exception of zinc ferrite, the lattice dimensions increase in the order Zn, Mg, Fe, and Mn. Figure 4 illustrates these relations.

40. Comparison of Values of Lattice Dimensions Obtained by Vari- ous Observers.-For purposes of comparison the results of other investigators are collected in Table 3. No previous references to the lattice dimensions of synthetic ferrous aluminate or of ferrous chromite were found in the literature. The divergences in the various values may be explained in part by the solution of contaminating substances which may distort the lattice structures of the spinels. The temperature at which the spinel was synthesized may affect the unit cell dimensions also. For example, Posnjak's value of 8.03 A.U.

*R. W. G. Wyckoff, The Structure of Crystals, p. 97, Chem. Cat. Co., New York, 1924. t"Student," Biometrika 6, 1, 1908; 11, 416, 1917. A STUDY OF A GROUP OF TYPICAL SPINELS

Chrom/'e5' t~^^-^.;J.^^ / Chrom,'f7s ': ·, I ___~ 1F-- i --- I L ' 86300 I --~- 8./00 -1 '1 %.9000 \J4~ a u Q~zV 7.900 k I FIG. 4. LATTICE DIMENSIONS OF SYNTHETIC SPINELS

in the case of MgA12O 4 refers to spinel which was synthesized at about 1700 deg. C., and it is conceivable that such material may have a lattice dimension somewhat different from the specimen studied in this paper, which was formed at about 1325 deg. C. 41. Computations of Theoretical Densities of Synthetic Spinels from X-ray Diffraction Measurements.-With the knowledge that eight spinel molecules are associated in the unit cell, the theoretical densities may be computed from the formula nM P= 3 a0 where p is the density, n the number of molecules in the lattice, ao the length of the unit cube, and M the absolute mass of the molecule (equal to 1.649 X 10- 24 of the molecular weight). The computed densities are summarized in Table 4. They represent the maximum theoretical densities of the spinels and should be greater than values which are obtained experimentally, since the latter determinations involve any voids and inclusions which may be present in the crystalline material. This difference varies from nearly zero to about 0.29 unit. The actual density of magnesium ferrite was found to be greater than the theoretical maximum value by 0.10 unit. This anomalous relation is probably due to the formation of a small amount of magnetite with the desired spinel. A comparison of these density values may be obtained by an examination of Tables 1 and 4. ILLINOIS ENGINEERING EXPERIMENT STATION

TABLE 3 LATTICE DIMENSIONS OF SYNTHETIC SPINELS Comparison of results of various investigators

Size of Unit Cell, Spinel Angstrom Units Investigator

Zn AO104 8.099 ± 0.003 Holgersson* 8.050 Passerinit 8.062 ± 0.001 Authors 8.093 ± 0.004 Hauptmann and Novakt

Mg A1 04 8.03 ± 0.01 Posnjak¶ 2 8.090 ± 0.003 Holgersson* 8.050 Passerinit 8.086 ± 0.003 Authors 8.059 ± 0.004 Hauptmann and Novakt Authors Fe AlsO 4 8.119 ± 0.002 Mn AO120 8.263 ± 0.007 Holgersson* 8.271 ± 0.002 Authors Holgersson* Zn CrsO4 8.323 ± 0.004 8.280 Passerinit 8.296 ± 0.002 Authors Mg CrsO0 8.290 ± 0.005 Passerini§ 8.32 Holgersson** 8.305 ± 0.001 Authors Authors Fe CrsO4 8.344 ± 0.003

Mn CrnO4 8.487 ± 0.002 Holgersson* 8.436 ± 0.001 Authors

Zn Fe2O4 8.403 ± 0.004 Holgersson* 8.350 Passerinit 8.423 ± 0.001 Authors Mg Fe204 8.342 ± 0.005 Holgersson* 8.360 Passerinit 8.366 ± 0.001 Authors Fe Fe2O4 8.400 Claasentt 8.417 Holgersson* 8.374 ± 0.003 Authors Mn Fe2sO 8.572 ± 0.006 Holgersson* 8.515 Passerinit 8.457 + 0.002 Authors j *S. Holgersson, Lunds Univ. Arsskrift, N. F., Avd. 2, 23, 9; Kgl. Fysiogr. Sellk. Handl. N. F. 38, 1-112; taken from C. A. 24, 804, 1930. tL. Passerini, Gazz. chim. ital. 60, 389, 1930; taken from C. A. 24, 4439, 1930. tH. Hauptmann and J. Novak, Z. physik. Chem. B, 15, 372, 1932. 1E. Posnjak, Am. J. Sci. 16, 528, 1928. §L. Passerini, Atti accad. Lincei VI, 9, 338, 1929; taken from C. A. 23, 4167, 1929. **S. Holgersson, Z. anorg. allgem. chem. 192, 123, 1930. ttA.A. Claasen, Proc. Phys. Soc. London 38, 482, 1926.

TABLE 4 COMPUTED DENSITIES OF SYNTHETIC SPINELS

Density grams per cu cm. Divalent Element Aluminates Chromites Ferrites

Zn 4.615 5.393 5.322 Mg 3.548 4.429 4.506 Fe 4.392 5.085 5.201 Mn 4.031 4.900 5.029 A STUDY OF A GROUP OF TYPICAL SPINELS

Fia. 5. DENSITIES OF SYNTHETIC SPINELS AS COMPUTED FROM X-RAY DATA

TABLE 5 MOLAR VOLUMES OF SYNTHETIC SPINELS Computed from X-ray measurements

Molar Volumes* cu. em. Divalent Element Aluminates Chromites Ferrites

Zn 39.8 43.2 45.3 Mg 40.1 43.3 44.4 Fe 39.6 44.0 44.5 Mn 42.9 45.5 45.8

*At room temperature.

42. Theoretical Densities of Synthetic Spinels.-Table 4 provides a summary of the densities of synthetic spinels which were computed from measurements of the X-ray diffraction patterns, and Fig. 5 shows the results graphically. By means of the data in Table 4, the volumes which would be occupied by one gram-molecular weight of each spinel may be computed. These molar volumes are based on the theoretical values obtained from X-ray diffraction measurements. An inspection of the values given in Table 5 shows that the molar volumes differ only by fairly small amounts, and hence it would be expected that these spinels would readily intercrystallize and form isomorphous mixtures.

VI. HEAT CAPACITIES OF SYNTHETIC SPINELS 43. Method of Determining Heat Capacities.-The mean heat capacities were determined by a comparison method which was based ILLINOIS ENGINEERING EXPERIMENT STATION

FIG. 6. APPARATUS FOR DETERMINATION OF *HEAT CAPACITIES OF SPINELS on the use of vitreous silica as a primary standard. A diagram of the apparatus is shown in Fig. 6. A weighed charge of spinel was contained in a platinum crucible at S which was suspended from a mechanism M. Rotation of the mechanism M through a small arc caused the crucible and contents to fall at the will of the experimenter. The temperature of the charge was measured by a thermocouple T which was placed about a centimeter above the top of the sample contained in the crucible. The temperature of the thermocouple was read by means of a Leeds & Northrup Type K potentiometer, and a correction was applied to this indicated temperature to obtain the average temperature of the sample. The furnace winding was of a platinum-rhodium alloy. This was A STUDY OF A GROUP OF TYPICAL SPINELS surrounded by powdered magnesia insulation B and powdered Sil-O- Cel Si. The furnace was supported on a slide which permitted the removal of the furnace from its regular position above the calorimeter. The bottom of the furnace was closed by means of an easily removable plug and shield Sh, which was held in position by a strip spring Sp. A transite shield Tr aided in decreasing the radiation of heat to the calorimeter. When the sample S had reached the desired temperature the bottom shield was removed from the furnace and the mechanism M rotated with the fingers. This rotation caused the sample to drop into the calorimeter which was placed directly below. The calorimeter consisted of a quart-size Dewar flask D in which a known mass (2000 grams) of mercury was placed. The temperature of the mercury was measured by eight copper-constantan thermoelements C which were held in small glass tubes placed sym- metrically around the interior surface of the Dewar flask. The cold junctions of these thermoelements were surrounded by an ice-water mixture, and the electromotive force developed by the series of eight couples was measured by the potentiometer already mentioned. A nichrome screen Sc extended from the top of the Dewar flask to the bottom, where it dipped below the mercury surface and prevented the hot crucible from coming in contact with the glass sides of the flask. The Dewar flask was held in a brass tube which was surrounded by a water bath W contained in a stoneware crock. The temperature of the water was maintained constant to within about 1 deg. C. by a heating coil and thermoregulator H. A thermometer Th and stirrer St were provided for the water bath. Felt packing L aided in the exclusion of outside effects on the temperature of the bath. The dropping mechanism is essentially that which has been described by W. P. White.* The efficacy of this method of determining specific heats depends on the maintenance of similar conditions when the calibrating substance (vitreous silica) and the spinel samples were used. In all tests the furnace temperature was maintained nearly constant and the temperature of the calorimeter was adjusted to approximately the same initial reading of the copper-constantan thermocouples. The calibration consisted in dropping various masses of vitreous silica in their platinum container from the same furnace temperature into the calorimeter. Each mass of vitreous silica produced a rise in the temperature of the mercury in the Dewar flask which was pro- portional to the mass of silica used, and this rise was measured by the

*See W. P. White, The Modern Calorimeter, Chem. Cat. Co., New York, 1928. ILLINOIS ENGINEERING EXPERIMENT STATION

.C N~ ONNO 00N m00O 000 00 - C 00 0O a 000000 - 00 N- 00 N )0

Co N 000000N 000 00 =O 00w -0 = 1 m0 00 00GO)1 00 o0 COo0CCOO - *-I C- 000000 COt- 00m I )N N

00 =Nt-poCO3 - o m- -^m 00 rN 00 o0 o o - a S 0;000 ,N - 00 0 '3 0 IaC 00 CO - CO N N N N

a · 00 004Cftt0000COll000 00ON0000 C00 00 tO0 COW0 CO l>N 00O d * COO G CO 00 0>~*O0 m(TDC 0 O. OO.0 t-N . 'lt O .0 'C

N 0-OiO 0 CO CO NO C *c;o cN . l . . o ooo o q- o r t» I

- 00C~iOS0 TON( CO N- 00 CO O 00I'

cO 411-^ OON 0;n >0 ® Oi t- X SCOI 0 CO Uo

N Oi 000 . . o 1 (- 00'OC'0oCO 0 0 OCO N0000 4 N N ~ 00 CO r-. .- 00 '00 00i-C

- NN G00 0O O 0 - ' 00 *'0'000 oNl NM 0'0 Nt 00 '0 00 (N 00CO -l^ 2 00 '0 0 "- 00

-~ COONlN0O 00'0 00 -- 00 '0 00 >- t0 0 CO 00 N '0 00 00 Cqa CO CdQOm C0 T'(CN0 . * O 03 CO - - 0 0 0 - 00 aot"jE NO Co 00 - N o 0S 00 ^ 000

a 0a N ~ 0

*®0 1^ CO =- a 0

Elg'so

0 . "-.5

.50ac

. C

C A STUDY OF A GROUP OF TYPICAL SPINELS

'I *s; IeK

RiXse 'inerl, of Thermoe/eelen's /i COa/or/nefer. /%y

FIG. 7. EFFECT OF MASS OF VITREOUS SILICA ON RISE OF E.M.F. OF THERMOELEMENTS IN CALORIMETER increase in the electromotive force developed by the series of eight copper-constantan thermocouples. Table 6 shows these measurements. The initial temperature of the sample was nearly the same in all cases. However, the final temperature of quenching varied with the mass of vitreous silica which was used. This difference in temperature change was adjusted to a common basis of a temperature change of 975 deg. C. by assuming a linear relation between the temperature rise and the mass of the sample. The actual temperature differences of the tests are shown in the next to the last line of Table 6. They are all near 975 deg. C. The last line indicates the rise in the electromotive force of the series of thermocouples which would result if the temperature change were 975 deg. C. The data thus obtained are shown in Fig. 7, which depicts the rise in the electromotive force of the series of copper-constantan thermoelements in the calorimeter which would be occasioned by a known mass of vitreous silica in its platinum container as it cools through a temperature range of 975 deg. C. The points of this curve do not represent a straight line theoretically, but one which is prob- ILLINOIS ENGINEERING EXPERIMENT STATION ably concave to the horizontal axis due to the increased radiation loss which occurs as the mass of silica increases. However, this con- cavity is slight and the straight line y = 1.37 x - 1.90, which was obtained by the method of least squares, fits the data quite well. The equation, y = 1.37 x - 1.90, in which y represents the mass of silica contained in the platinum crucible and x is the rise in the electromotive force of the series of thermocouples, then serves as a measure of the heat capacity of the calorimeter. If a substance of unknown specific heat is substituted for the vitreous silica, and the increase in electromotive force of the thermocouples is noted when the platinum crucible and contents cool through 975 deg. C., then the equation y = 1.37 x - 1.90 gives the amount of vitreous silica which would be necessary to bring about this temperature change under the same conditions. The heat capacity of vitreous silica at various temperatures was obtained from the International Critical Tables,* from which it may be deduced that the mean heat capacity of vitreous silica from 50 to 1025 deg. C. is 1.096 joules per gram per deg. C. This value was used as a primary standard in the heat capacity determinations which are listed in the Appendix, Table 16. This table shows how the heat capacity of one gram of each spinel is obtained in terms of the mass of vitreous silica. When this value is multiplied by 1.096, the mean heat capacity of the spinel over the range 50 to 1025 deg. C. is obtained. In general the mean of three determinations was recorded. The accuracy of the method depends on the degree to which the experimental procedure can be duplicated. Thus the upper temperature of the sample before dropping should be very nearly the same in all cases, and the furnace was held constant at this point for about 15 minutes to insure that the difference in temperature between the furnace thermocouple and the sample would be the same. It was found that a thermocouple embedded in various positions in the sample read about 18 ± 5 deg. C. higher than the thermocouple indication. This correction was applied to the thermocouple reading to obtain the average temperature of the sample. The calibrating substance should be of the same thermal conduc- tivity as the spinels in order that the radiation losses entailed while the calorimeter was rising in temperature should be equal in all cases. Although the thermal conductivities of the spinels and of vitreous silica undoubtedly differ, the error is judged to be small. Further, the physical condition of all substances should be the same, and for this reason powdered vitreous silica was used instead of larger parti-

*International Critical Tables V, 105. A STUDY OF A GROUP OF TYPICAL SPINELS cles. No devitrification of the vitreous silica took place during the calibration runs. Concerning the heat capacity of the platinum crucible and that of the mercury nothing need be known, since the masses of the crucible and of the mercury were unchanged throughout the tests. The heat losses of the calorimeter due to radiation, convection, conduction, and volatilization of the mercury do not matter providing that they are the same in each test. Attempts were made to effect this duplication in the experimental procedure. The mode of procedure indicates that the calibration of the copper-constantan thermocouples is unnecessary, the only criterion being their constancy of indication. A consideration of these factors results in a value of 2 per cent as a conservative estimate of the precision of the results. In order to compare the results obtained with the mercury calorimeter with the values obtained by other methods, the heat capacity of artificial corundum ("R. R. Alundum" from the Norton Company) was measured. It was found to be about 3 per cent less than a value which was computed from the data of Cohn,* who used a differential heating method.

44. Experimental Data Concerning Heat Capacities of Synthetic Spinels.-Table 16 in the Appendix shows the readings and final values of the heat capacity determinations. The lowest horizontal line gives the mean values of these determinations.

45. Summary of Heat Capacity Determinations.-The mean heat capacities of synthetic spinels over the temperature range 50 to 1025 deg. C. are listed in Table 7, which shows in column 2 the values in joules per gram per deg. C. The corresponding heat capacities on a gram-molecular basis are shown in column 3. Conversion to calories may be effected by dividing the given values by 4.186. Table 7 shows that the molar heat capacities of eight of the twelve spinels are very nearly equal, varying from 169 to 184 joules per mol per deg. C. The molar heat capacities of MnAI20 4 and MnCr20 4 are somewhat lower, being 154 and 165 joules per mol per deg. C., respectively. The ferrites of iron and manganese have considerably higher values for the molar heat capacities-222 and 201 joules per mol per deg. C., respectively. These ferrites show the phenomenon of ferromagnetism, but the increase in heat capacity cannot be ascribed wholly to this effect, since the ferrite of magnesium possesses the

*W. Cohn, J. Am. Ceram. Soc., 7, 482, 1924. ILLINOIS ENGINEERING EXPERIMENT STATION

TABLE 7 MEAN HEAT CAPACITIES OF SYNTHETIC SPINELS OVER TEMPERATURE RANGE OF 50 TO 1025 DEG. C.

Mean Heat Capacity, 50 to 1025 deg. C.

Kind of Spinel

joules per gram joules per mol per deg. C. per deg. C.

Zn A1204 0.92 169 Mg AhO4 1.19 170 Fe A1204 1.04 178 Mn A1iO 0.89 154 Zn Cr204 0.77 180 Mg Cr2O4 0.92 177 Fe Cr2Ol 0.82 184 Mn Cr2tO 0.74 165 Zn Fe204 0.73 176 Mg FesO4 0.86 172 Fe FeO04 0.96 222 Mn FeiO4 0.87 201 normal value and yet it is strongly ferromagnetic. Since very small changes in the atom profoundly modify its magnetizability the lack of obvious changes in the specific heat and other properties is to be expected. The similarity in the heat capacities of most of the spinels verifies to a limited extent the truth of Neumann's law, that compounds with similar molecular structures have the same molar heat capacities. Although an intimate relation must obtain between heat capacity and thermal expansion, no clear correlation is evident from these measurements.

46. Atomic Heat Capacitiesof Synthetic Spinels.-At high temperatures the atomic heat capacities of compounds-i.e., the molecular heat capacity divided by the number of atoms in the compound- approach a value of about 6.0 cal. per deg. C. The molecular heats of Table 7 may be used in this manner to compute the atomic heat capacities given in Table 8, which are average values for the temperature interval from 50 to 1025 deg. C. They should, therefore, be somewhat less than 6.0 in magnitude. This relation is verified in most cases, but the value for manganese aluminate is quite low, being 5.26 while those of the ferrites of iron and manganese are inordinately high.

VII. THERMAL EXPANSIONS OF SYNTHETIC SPINELS 47. Method of Determining Thermal Expansions.-The Fizeau- Pulfrich principle was utilized in the thermal expansion apparatus A STUDY OF A GROUP OF TYPICAL SPINELS

TABLE 8 ATOMIC HEAT CAPACITIES OF SYNTHETIC SPINELS OVER TEMPERATURE RANGE OF 50 TO 1025 DEG. C.

Atomic Heat Capacity Kind of Spinel (50-1025 deg. C.), cal. per deg. C.

Zn AlO04 5.77 Mg AlO04 5.80 Fe A120 4 6.07 Mn AlO04 5.26 Zn Cr2sO 6.14 Mg Cr2zO 6.04 Fe Cr204 6.28 Mn Cr 204 5.63 Zn Fe20 4 6.00 Mg Fe 2 04 5.87 Fe Fes0 4 7.58 Mn FeOl4 6.86

FIG. 8. FIZEAU-PULFRICH APPARATUS FOR DETERMINING LINEAR THERMAL EXPANSIONS OF SPINELS

which is illustrated by the diagram in Fig. 8. The measurements of thermal expansion were made on small pieces of spinel which were in the form of tetrahedrons with altitudes of a few millimeters each. Three such pyramids were placed between optically flat vitreous silica plates in an electrically heated furnace, and this system formed the interference element in the Fizeau-Pulfrich method of determining linear thermal expansions. ILLINOIS ENGINEERING EXPERIMENT STATION

TABLE 9 INTERFEROMETRIC DATA ON THERMAL EXPANSION OF PLATINUM

L2 = 0.2996 cm.

Interference Temperature of Expansion of Sampi Fringe Number Sample, deg. C. microns per cm.

0 27 0 5 92 5.44 6 104 6.49 14 190 14.73 15 201 15.74 24 293 24.84 25 304 25.85 34 393 34.86 35 402 35.85 45 495 45.81 46 504 46.80 55 592 55.72 56 604 56.71 66 693 66.61 67 707 67.60 77 792 77.48 78 802 78.46 89 895 89.30 90 905 90.29

Figure 8 indicates how light from a helium discharge tube at A passed through the collimating system at B. The totally reflecting prism C then caused the light to pass down into the furnace through the vitreous silica convection shields D. Interference took place between the surfaces of the vitreous silica plates separated by the samples at E, and the expansion of the sample caused a passage of the interference fringes which were viewed through the optical system F, G. A thermocouple H was used to measure the furnace temperature, from which the temperature of the sample was deduced by means of a correction term. The references* may be consulted for a detailed analysis of the method. In order to check the accuracy of the expansion tests which were made by the interference method, the expansion of platinum was determined. The data relating to this test are found in Table 9, from which the mean coefficient of linear expansion of platinum between 27 and 604 deg. C. was found to be 98 X 10- 7 cm. per cm. per deg. C. The mean coefficient from 0 to 600 deg. C. as computed from the equation 2 It = I0 (1 + 8.786 X 10- 6t + 3.118 X 10-9t - 5.27 X 10-12t3 + 4.095 X 10-15t4),

*C. G. Peters and C. H. Cragoe, Bur. Standards Sci. Paper No. 393, Washington, 1920. L. D. Fetterolf and C. W. Parmelee, J. Am. Ceram. Soc. 12, 193, 1929. A STUDY OF A GROUP OF TYPICAL SPINELS which was obtained from the International Critical Tables,* is 97 X 10- 7 cm. per cm. per deg. C. The agreement between the experimental and the computed values indicates that the thermal expansion coefficients which are given in Table 11 may be relied on to about 1 per cent.

48. Preparationof Specimens for Thermal Expansion Tests.-As a rule, small pieces of spinel were broken from the larger block of synthesized spinel in order to produce fragments suitable for the thermal expansion tests, and these pieces were then ground to the desired pyramidal form on a carborundum stone. In some cases the synthetic spinel was too friable for such treatment and another means of form- ing the test pieces was resorted to. This method utilized the spinel in a finely powdered form, which was mixed with five per cent of its weight of equivalent parts of the constituent oxides plus five per cent of boric oxide. Sufficient saccharine solution was added to facilitate molding and the mixture was then pressed, dried, and fired to such an extent that the added oxides combined to form a spinel which acted as a cement for the original spinel grains. The boric oxide was volatilized by this heat treatment and did not vitiate the test specimens, which were ready for the grinding operation after this heat treatment.

49. Thermal Expansion Measurements.-The thermal expansions of the spinels as determined by the interference method are given in the Appendix, Table 17, from which the expansion values at desired temperature intervals may be deduced by interpolation. These inter- polated values, as set forth in Table 10, were obtained by the interpolation formula of Lagrange in case the required temperature was sufficiently distant from the measured value to necessitate the use of this formula.

50. Linear Thermal Expansion Coefficients of Synthetic Spinels. Table 11 lists the values of the coefficients of mean linear thermal expansion of synthetic spinels over various temperature intervals which extend to 900 deg. C. in most cases. Figures 1, 2, and 3 illustrate the courses of the expansion curves of these spinels and provide a graphic comparison of the expansions of various aluminates, of chromites, and of ferrites. Considering the aluminates, it is apparent that up to about 450 deg. C. the expansions decrease in the order Zn, Mg, Fe, and Mn. The reverse order holds for the lattice dimensions of these spinels. Since a smaller lattice dimension denotes a greater compression of the

*International Critical Tables II, 461. ILLINOIS ENGINEERING EXPERIMEN T STATION

a C'i-Ot-O b-X' *e S0 h C0r- a 1Nm .o1 to to 00

d- O~~lmhdD-^^XO *.

W $ G

F o~mc'1c~~

( ^ COor-< OS" - -^ 0c --r-l 0

" '**^( *^ OS0 *^M *

^1as

o .E * £qC ~ O10CC-^O- ~ c v *c m Qtcoot-lc"MG~m:0 o OOOCO 00'IGC- C0 i"g 0 dadooimnd *o i 1ý1(CCOt ": t091ý if .S

0 0

1; I - riN CO (i0tt Fr

a * I 0

% -l-0~0000 ti5-< *'- M ih0 i- -l(0~0U C ^1 E

0 * 6 ddd3-lCmiXOrh'^ *OO N

~g ~~WW088~HPH A STUDY OF A GROUP OF TYPICAL SPINELS 43

9C1

t01 100 ) 0 001* m a 0 z a a 0) C) 0- )0) t O.iOS )i* 0 N x Cj2 T buO zS C) & t--haOOOOS'O a 0 0) 9c a CC Ft3 0 C) '-I *03u·~0 OO ^ 0M* z 0) t» t^ hmmo000C S ia - 0d >< C3 Z 0 c C) CO a d §0 03 0 9 0 'i;&0B z N 0

C 9 0.'013t-mC0'2~0) *0 -4 C)CO )0)0010t. *.~- 0 sz C4 8 9 0 w © h ~ i$~~$ 0 9r -4 CC D~~h-P ·

0 -

%.oI94 9XS$ 010010010010010 44 ILLINOIS ENGINEERING EXPERIMENT STATION atoms in the molecule, it would be expected that such a substance would have a greater expansion than one which possessed a larger and less compressed unit cell. At higher temperatures the expansion of ferrous aluminate increases more rapidly than that of its com- panions. In the cases of the chromites and ferrites of the metals studied, these relations do not hold. The ferrites particularly show irregular expansions which attain quite high values. Magnetite especially shows a sudden rise in expansion between 500 and 600 deg. C., and this increase is undoubtedly intimately connected with the inversion which is known to occur near 593 deg. C. Of the ferrites, only that of zinc is non-magnetic, but no effect of this property which would distinguish it from its fellows is apparent from the course of the expansion curve. APPENDIX

TABLES OF DATA FOR COMPUTING DENSITIES, LATTICE DIMENSIONS, SPECIFIC HEATS, AND THERMAL EXPANSIONS OF SPINELS

TABLE 12 TRUE SPECIFIC GRAVITIES AND DENSITIES OF SYNTHETIC SPINELS Determined by Standard Test Method No. 52 of American Ceramic Society

Aluminates Chromites Ferrites Divalent Element Specific Density Specific Density Specific Density Gravity g. per cu. cm. Gravity g. per cu. cm. Gravity g. per cu. cm.

Zn 4.545 4.54 5.307 5.30 5.273 5.27 4.547 5.302 5.277

Mg 3.527 3.53 4.404 4.41 4.612 4.61 3.532 4.441 4.629 Fe 4.220 4.22 4.937 4.93 5.175 5.18 4.234 4.944 5.147 Mn 4.000 4.01 4.798 4.80 4.757 4.74 4.033 4.810 4.728 ILLINOIS ENGINEERING EXPERIMENT STATION

1 CiNNCV i N0l NNCi 01 C N i N NNý N 000F 00 0 .0 00 30 00 00 0 0o F- F0 0

~·~3C Nm - 0? C

0" 0000C 0,o0q 000 00 0i 0c 4o r0 ,V ,:M10 0M110 t- - C0F 0

t l *F 0O\» 0-

.4. NC,01 OS

a0 af-< 000 000 0 i-> l C03 0 GO00 00 00 0000000 0X00 00 x 000 x 0000 00 0C

=o.2 i 1:] F L'; 0 I>- F- t,-7 0. ' co 0s 3 OD t- m o i0le* iD g F-- pqr r-IM cqM COCx m a ocoOfl5. n 0I B m=. da gE _~ =Nnoc-e ?00 = 00 F-F 0

s! - Q Oic0' c -. '- ^ Mc 0 0 -I 119 -ý - T- C

00 O 00S-F-0 0 I0 0 00"00 00 i- 0l CO0 (0 0 - C 00 0n - 00 0 0 0S ^ I Mi? o® tatfi~ c~ii r~ o oo i dr oooo t-- t~ md C - .0 O Ci F-< 00 00 0000 W 00 07 00 00 0 0

0 0 0000 0 0 G0 00- 0 ' 0 0 0 Ca 00 F- 0, 0o F- 0 0 C C ; .-q0 CO 0q 0 0 p'I S i 0 C0N0 m 0 0- t "0 0 0 -0 00".. 00 F-7F- 000 0 .ý .ýC 1s i1 R CýI .1 09 00- - -'400 - 0 0- - 0 - 0 0 0 0 i>cco 00Rg ^-i-0 3 tocmco-c'a0 0O r'F- 0 0 o '.a0OS co0 i osh^c 00 0

<; -000D0100 00 Go0 t-t- 0 0 - 00 00 - -00 0 0 ,-30000o0 0 0 ~ - 00 00 0000 00 0000c0 00 00 0 0 0 0 0 000 0 0

.2~ ~ O ~ rM1 C>Ci cq(NOR 1 t r-i a'*^ iO CO CO ' . coCxdc-l I t-0 '. F000cx D00' -v i 40 o 'D GoI N- 0 M c F-t t- F-00=

'*^*^* CO-^OP«OT-m 'I 0 tD O'-(INNM0-l---0 '^T-'-N C .'i 3- 3-. ^-i'N . '-» 3- a

2m00 0 """~O00D0"$ C4 , m ~~ j0 :8(cm0"00 gx00 N0g M 0CýOFCOO-F N -0 0C0 0 0 0 00O O 0 0 N i -- 1------i 0 0 000 0 0

S0 , Id q 0 0'80 F- u(rf+a+go+ ) Ima - 00sqa. týF- 00000lqOmo 0 10 10 Ft- F- 000 0 0

83858888£58£838885SSSS3'S5^SS3-0-0------' -- 0 - 0 - 6pýo Mo^0 l~ 3^^ ^ElC0e o0 0.oB0o0 a^i-iiN0ii0F N^^033C - -- 0 A STUDY OF A GROUP OF TYPICAL SPINELS

OOOOOOOOOO00 0OOO 00 0O00 00 0 0 000000 00 0o

CoýoIo ol oCo CoýCoý~~ -! ý C Co oCo o ~C oý o C Co 0 11 0 . O OC 0N 0iO0'CO 0 Co 0 mC COO - CII

0r00o 0--ff't000aCo.O- Co 0 0 0 N iO o~ 0'l - f .C,C,13 - . w "N S I-< I- 00 N CCO 0 0 0 O O 0 0 S z a; w ·E-rt ~ OS'C'sf-it-' tO i^*CM- O- OS OS oSX 00 0 aCo coCoo

Co 00 1ýCo0000Co00 0 1ý '0C00000Cý o' 0 0Co 00 00Co 0 Cý0 Co 00 Co~ Co0 Co0 2; CC Co o C Co Co Coý Co Co COo Coý Co 0 m 00 mCFRI com t CoCOCOq CO CO CCOCO N O COO O Co .2 00 00'.0C0 0 Co 0C 0 '0 COM 'Co Co m Co COS-0CO'-M. 0CO6 060COON O,-MCOOO OOO 00--i 0 00CM 000 '0 x OS00 d 000c0000Co 00 00 00Co c0 go0C Co '-l0l N Co' Co0 Co OO'N0 O 0 . c O iq O OrOo O m C 0 o Cq ,v co a" N *^4.;+ NdlO'-'^C003 00l-'-Co0 O0 O·XbXNO*''- NO'-^ *^c ^P'Co r--i' a'Cof MO ^Co CoO\IC- XmdONCSO 0 0 s u Cco ,o mmmmm mm mmnm a .o C00C O-MOI co --0 fxsooos0. O ,'.0000 co 0 0000t'co0 Co 0o 0 Co 00 II CO CO-^I'Co0C *^COi~M'l >t000 0 OS 0000iOO W 0

C? Cocootco Co Con Coo o CoCo Co Co Co CoeoCo Co Co ©Co .2mm g> o m-o mmmi-i>. o or ooos ododb-ot' od 0 Vo o .O Co '-00000 - .NO O Oi0 Co ONN- N Co m3C0*Co,- 0 0P UCo CoNE NOOl CrCo II

p^.2.o 0 oH 0-l 0lC0 Co 0 -o 0 , C S 0 0 a '43 £dO .t0 NO '-CM 0 S rl ^ ^ C O M TIC -M^ il C c 0OCo S!

Co ONO Co .co0 00 OlO0000 "Co '0Co i 00Co mi 0;o 0Go606OO Co00

m mo00od0mm0 0606ý06006 06 o0m0a6 06 ýo 06006 06 06 CoZ Go 090 o10O00 1000000- 00 0 0 0 0 0o 0 RE mo00000 0 0-00ao 0 0=0 0 1c N O" o Co0 do d S o0CO 1OCONC- -0coo o-4O ' 0 SC tt o =.ot 0 II G § n SI-'O00 COCo 0 0 O 00 0omo I (CoD0N0 Cio0'c0 Co CoOC 0 o0N 0 c~0."o i 0 0 C al t 000~S~Co'C 0Co-c~o3o00o~~o=~ Co N=Co 0 0q --4 _-; ; 60 0* 0000 o 0 I -

Nu0 0CO u ( 3 C., A 0 C -Z-7+C.l tN moco I Co LoCo oD C= N C- 000C 0 0 0N

Co 0o0 00~ CO Co0Coo0=o C Co Co o ~Co00Co CoOCO~5~o0 03630Co0Co 0 ~~CoooOCN00o000N 0 as Coom~~~~~~n~ofl~o0sCo 0Coo0 C 0 oo~~OO C o f Co Co o 1o~oo 1o~o~ go3oCo 3'0oC DooNCO,4Cio0 ILLINOIS ENGINEERING EXPERIMENT STATION

o 0 0 'I N Coo0 Co Co1 CCC$C' ggC' 00Co0Cododmodod OWO 00 Coi 00 CododoOoOQ cc 00 00 00 CO.o 0o0oCo 00

S0 o000- N Co 000m gO 00 m o t 0 iot t-

""^~ -\ H" g^~^ - ~ 10 NI C~o.>no > Co C o „-- C - - 'D to 00 q0 CoC4

Coo63~ =Lmkoý;! 0cqoi - N 10CO..C MMCOC.C C oC o-iM 0000

-~--

O 00oN00 000000 3 COCCOMCCO CO COMo^ Co00000 Co 010CoCo. 00 000o 0

o - a Co Co - CoCo Nt- : i 00 - C o CO Co Oi-'C O;3R^td-NNCo N NOOC

.OCCOC36 N- 0~~CC9'CC -l CCC C ýý-"N r1"

CoO 0 00=0~Co CoCNNCO.- -o sc'coN0· Oi>0^01> 0 CCC'C3o3 0 0C00 0 Co C ON

hConCooCoC C' 66 Co1o Coio CoCoCo CoC 00 00 xCC0o

.2 > 00 -

pr P- - 1'C'CC o o .COCC tNCo^ t00CoO tN^ 00^00 0 m3~~-~ -o

*32 .oWQo-im COo'.4 '-t Co COo '- Co'-'MCOo '-t C C E3 i- i

NO 80 O!C4

e MOCOCOCOWC WCOMCOM COCO CO COCOCO CO CO cOOocOOOo O0 00 0 OO C 0 00 CXoC0

0 o ,2 ^P 0000 0 OTCOO0 0 0i 0 N 0 <00CC0 0 II r~. aoýo. o o cooo - C:Ro

m mCN'qoN - q c4Co tNm - ro 'CCNfto - 00 00i Cooo Co

-t-T- -t-t-»- v-- 0 0000 0 0

rU (I+ zY++ s) I 0-l"m CqN t- cmmqomOMo ýWi" s?s 2 ý ta-l 1- 0 1 0I

s S 6 iBass3i -- 83 mass3-- --3 -- ?8 33858 388833---- A STUDY OF A GROUP OF TYPICAL SPINELS

0c GO GO 0 Cc <* 00 f - 00 ^ GO GO GO G 0 GO 0

GO Cc GOi GO Gm OO G Cc GO GO O GO Cc N O G c C GO a GO 0 a' GO 0 GO GO GO = Cc~od

GO GO GO CcI Cc C 0 Cc GO 0,~ tOO,-Oi GO GO GO OGO 0 b.m - oo 1- Cc O 0O Cc G OS . d. . .

0 X5ýý4M m m x C,

* *0O)0 0 .. - C . . . (O ; GO00 t-h-r50NNG0 0000O 0 0 1-N)G N 0 C c C On o GO1>)GOCO 'lM00 CO'l ll cýCc OO0d t- 0 *l Cc OM cO0> C!^"^0 c; 0c - - GO tO- CNl 0O 0 l r,1

w =N0 g 4m !8 t- g 00 l t o O'^000 G-

* *00 . 00 c w C - - GO GO N C" N> GO O8 -

* *0)0) 0 . . . . Cc DO Cc00 MCI=-' GO GO NlmO GO O - G - OO 0 O GO Gc 0 0 G C G CS GO M MNO -- 0 O GO GOO =O a C! a m"M " 0 1. d * t Cc0 0- t a'- » 0 ' c-i ©0 OC 0o O, - iN- O N IG -

cGO C) cn - 0 O GO GO GO GO Cc - GO CcNONGOO GO'1- iG uCc -i O GOO GO CcCc O . oi GOSG '-'G GO Cc O t GO-' GO GOC O 0- GO

* t Cc0^ 0oc Crc Cc^x^ w fr o g -

S P3 OS'- i-C $t (D i- OS '^ 00 T OO W T Cc NOCc ' t 0 'N GO 0C GO tS00'I l - a o GOM S rGOG O N 0 G N 0 0< CcO 0 N omI' GOa GOh GO GN 3C GO G Cc -O : God GO N N ~00 GO

- -i 13 <-*W* : -** 0 :.0: .. GOa : : ::a :3 : :C: E :.S::

0 :u s *§ llj ,E1ll jB~db |§§-1-1%· ILLINOIS ENGINEERING EXPERIMENT STATION

N N 0 CCCCCNC h C 0- CC *- C C CO = CC CO 0 N CC hCCCN * 00 C,^o riCO >r 8o o0 tOi * *00 0 od00C CCa dCO

d £ CO - Oi O-C CC CCCCC~ CC 00 C CCt- O~c C.-~0 CC CCC CC0 cCi0 0 COo , 0 0O 0 0x d CC CC '- 0 a 06 CC '- '- CCCw': CC -I's

CO CC* CC CC< 0 TC CC 0S CC 0 t *^ 00 0 0 0 1 CO CO0 flm 00 'I 0r- CC C C a CC CC3- 0 CCl 0

0 0 CO O a CGOO C C C C 0 0CC P0 0; 0' = C C CO - CO N O - O

CC a; =_ CO CC C 00 O N0 CO N c o 0

raC

O$ CC S = C W CO CO NO - 0 00 O 0aS CC CC C - CO O - O O d 3C a 0~ i 0 0 No 3 (

CC zc 0 o a ort = 0) C - CC (0 CC N m00 OS '< 0 CCCCCC 0 C CC o~CCCC0 CC CC CO - cO N0. 0 z0.EC

00. CC ,~0-CCN 0CCn CC 0 C- O CC 0 CC CO CC CC C 0~mCCCC C CC 0o N- CC N t0N OC *00 0 30

CO 0. TCC 0 01 CO CC CC .4C 0< 01 0 CO CCt C C0 CO 0 0 CO 0 N 0 0, OO io t^ T- t- O O 'i O ,,~~o oa CO CCCCCC -C O CO

CC CO a 0 CO 0 a 00 N 3 0 N CC oN C O 0.4 0 0 ,C CC O 0 ~C~CCO 0 CN O

o CO C CCCC CO CO 0 a= 0C0 CC- C CC 0mm CC CC N> CC 0 CGo 0 0 0 0. : .0o :s : 0ho z0; CC : 04 F u'* * '**s-'U '' * ' * * two ECl .3-C CO.cS g. ;u"| : : | CCCClld .g CO~~~CC CC:

.CgCOg 4 I 0 gCC .CCCC1C3a£ 1 0 *-d8s$si3 ^ ^' CO

100 tiP

S [KEchh E-i K A STUDY OF A GROUP OF TYPICAL SPINELS

0 00 -^ -^00 .4 4 0 o 0 00 u5 0 0040000 04N- "~fi"$04 04 N 04 0 0 4 00 00 **000'41, - -~3- 4 04: N 0- 04 N N N 0 0 10 to 0~ 04.4.( r-l 0o 04 0 04 04 04 04 00 04 0 - c 00 IN 04 04 0; 0 04 N- 00 00 N 0 04 00 SO t- m 000 0-4 0 00 00 04-000404 040 CO*^? - 11C

00 04.00,41 -04 C0 0 - 04O O0 0 * 000 0 .,v oo * 00 T-(o 00^ N00C 04 00 0 -^tC 00 t- 00- 00 t^oS 0 04 04 0 00 00 0 O~~-0 04 NCt O 0 O ~00t C!C0 0* C.*000 <0- '*

0 040N0 04l04 ,,> OSrl - 00 04< t-'404 -04 10~04 * 000 0 a 0 00 r- *- 00,m N C, 3 00 S04 Ný N0 0 0 04 04 00

040 04 04 04 00 . 0 "4 w. *0 .00^ 0 00C E-w N 00 0 0 00 o aoi

04 .4 0)'404^ 040 04 04 0 04 .4 04 00 04 t.- 00 .4 0 N 00 - 04 .4 04 - 0 00 = 0 , 04 04 04

6 t.4 104=04.4 0404( 04 04 00 . 4 0 0 0 04.4*f040 .404( 0 C Na '4 N .4 0 N 0 0 *.000 0 04t 0404- 00 ) O O N» 00 .- ( i. 06 cý 1 Eo

m- O M r- M L" S* 10 (0~T^ 0; mo W00 i0 mM 00Q X00 co 00c i 1 Cýi C ^ ! O O»lO-0M Cl 3 b1 -^1 u0 OSOi '-~ CO C O-0 Li-K O r- O

- u (0 .01 s No 0 0100 04 0* = 0N. . - t 0 N04 0 0 0. N0 1 § C,O X C ( OS IO m-t C r-i t C; C 0 ' a S 0444C r-«C.tO04-0 mN 04 N '4 0004 04 0S 04 O N>

w 04 04040404 CO04.4 1 04 00 04 0 .4 04 -l N 00 0 * 000 0 00i k N, 00- =-i O04 O40 '- .4 .4 O i 0 o 4) .~ .~**~4 o . o5 I .. . :. •' S. ,• ; :u?"; § °, °' :B 03 -i .. .·0~ 3 r-0 r. N. "s Ez S :l :.l s 4.0< 0 4.

• 41P$ H E " .s1 : .m : •-ssi a sE

§ " e e g . , •

- c3 o3ca c3 o , ' 3 ;u , . ILLINOIS ENGINEERING EXPERIMENT STATION

ILLINOIS ENGINEERING EXPERIMENT STATION

NN"Co00Cha oN,.)Co$$oN 4~4C CoOt-Co 0400Co"Cm ooom~~CoNN3N 04~NCoo~COC040 C 0o~o~Co~~o04-Co~oN o0Conoo Q ' Co C Co'4 ------4CoN~oNoom o '0 o CoCo C CoW

a0 Cod~~N~o0~00o'~00Coooo04N'0oooo 0N4oo,4C~~~~~ 04 'S 4 C~mO04CON0' CoOo'00'N 04 ~mO~od~04C~o'Co. ai Co13Co~noComb~~il1'0g0'0'0 Coh~~ ooo~CN~~~O O~oIComO~o~omoooO~Om o '0 CoNOo'0CoomoC4o 00 ------oN ooNoN~ om Co C lo0'''0040o~ C Co Co Co 0

CoN'0"o~oCNOQ~O~ON004O040C~oo000o04 ar 2 Co~~o0' C Cooo0C0NNN

Co 8 rI

0 0 Co II 0 ja Co o 'B CO 'i 0a NNM mm" ý 00000~ihhooo1oxc Coi~ Co

e Co an- go 04Cot-o N CoNM O Co -5 t-on**tR llý 39SESR 1 m 0· oR tWESERE~itSMS$ESCRES o***** RM844O08dd8

S i Coa 0 0 Co Cs C3NsC- Coo= Co 9.Comowo.ommCtýON " =w P~N~l~m Co Co

Co ON oOo C 0 CNot- 0OCoe04Coot- Co Os" 1 NNNrmm " " 41 = cll-- aCo 0019m*^mRN^»^m0000l^ ^~^*OiSOOOOQ~f W-0^-3'lO

Co O 3t 3 3 Nr-TWOiOmt^- h~COiÔirQD ''~iaQO dC N aos 0§ II O'^^t-OO^Mi~lmdN ntOOO'-ôtm-^ô^-~'QCSm~ tt 0.^ bs $% r8 -l -l^1m N O NCOCOM * < m^*t c o -rt m Lo < Ic N " o oc - m to ~Mot C. tlnmg~~h3 k- Cm c 0 (mCo Nt--1- 'B H *g

0 s s Ss s agii|iiiaCog l~ss§~ 0440~%$W N04 R~ tgiSislsiColon~ - a~$8~NOC- sglssN § OO

S lCo'lNCm5 4214 A STUDY OF A GROUP OF TYPICAL SPINELS

o o

0D£h~3Nr-t NNttIth1C O s ;i a-Q RE 8-*33-088 85"52 528*i**moo=

< CO C) 33NOO®'-^X'lCO'~ll--l0hhhhelQ^M -l-;^O3*

0 g S3 mD0; t o0,NMDC z

C CONlO QlCO OOCCCCVNCCCC~fI OC) a5 C)eE5 VV''Qs^OCTHO*C,-C NNNN COCOmm VVVVNNNNNN

pq 1 C a I O 'hb'i. m co t 1 ->01>0 'g M.0 m

.a gf C) OrOiQr-lmC^~mMMCO'}<^-^KS>0CHOc1ul3l-GOmOmC V V C

0 o I 'C-S '^^CVV'e COCrOCO> at^COOCV10 OCt(NCTPC O-'

C)) O C > 'CC i COC C SN)CeOC l.O 35 S Vj CC V *-0 . VOCCCCO Cr^M 'ÔVVCVCCV-OO,. ^Ô^tiM'C'CCCCOOCVVO,-. OC)CCCC 0 Cf)C . CO CCOVOVOm-OCOO'VCOOVCOVVV-lCOVCO VC3 C-, OCCVOGO a 5 CC COCCVVVNN)CmmcÔdbtdmOVQO)QOC

M.Q 0 C) C); N .CCCVC)C)CCCCC)CCCCO-^ ->'0i£)C£tbttOQOCCVNNVVOdoOCOC QE~z0 , c, c', N mmm mm mmogo -C^z

Ci C) -5-

V NVa C V V V * 1^C) CS O'' V2 't O )VV't^OOCnrtCfO -QOCO VVC)CCCCN~CCC)CC)C)CC)-tM~C'Oi~ -i NV^CC C)SC 1VV C~CO CCC CV VV NNN VQ V

Vi C

S C NfCIVlCrw'*^0CO V ViMrI*~'C VNV,-VVN ,-^-S-lC VVtQC^^O OCCC)CC)CC VNCC)'-OC) ^^£

S CP.O OCVV^ OVVVCVNVOi-l~VOV ilCNV^OCVCO VV~lCONO tC)iVrOl-CON~^.- C),a*r NC&COCO VVVNNNV VOCVO CO C ILLINOIS ENGINEERING EXPERIMENT STATION

o o

0a , 0 9 k 0 ,,,,,,, 3~n=gg ~ h~ ZO" ~ r-al~

Sa Cl3hD om~~md~~0~0h S *3~0.0 N ~ S 5E " Bd0

0 ~L ' o "~~! l oNoo C·~h~mO~Td(Odr) 5~- 0. ~cN~-o.-0o l I ~~0~

S ______

S ~ E ~NNNmmNNN M ""S$3 ý S ~ c~ ;a~~0h;S:Im~O- ~lR 0 : o~ m~~~~~mo~o a

u 0 - No E 00als0~

mb$

~d mmOd~~ m~d -- RECENT PUBLICATIONS OF THE ENGINEERING EXPERIMENT STATIONt Bulletin No. 198. Results of Tests on Sewage Treatment, by Harold E. Babbitt and Harry E. Schlenz. 1929. Fifty-five cents. Bulletin No. 199. The Measurement of Air Quantities and Energy Losses in Mine Entries. Part IV, by Cloyde M. Smith. 1929. Thirty cents. Bulletin No. 200. Investigation of Endurance of Bond Strength of Various Clays in Molding Sand, by Carl H. Casberg and William H. Spencer. 1929. Fifteen cents. Circular No. 19. Equipment for Gas-Liquid Reactions, by Donald B. Keyes. 1929. Ten cents. Bulletin No. 201. Acid Resisting Cover Enamels for Sheet Iron, by Andrew I. Andrews. 1929. Twenty-five cents. Bulletin No. 202. Laboratory Tests of Reinforced Concrete Arch Ribs, by Wilbur M. Wilson. 1929. Fifty-five cents. Bulletin No. 203. Dependability of the Theory of Concrete Arches, by Hardy Cross. 1929. Twenty cents. Bulletin No. 204. The Hydroxylation of Double Bonds, by Sherlock Swann, Jr. 1930. Ten cents. Bulletin No. 205. A Study of the Ikeda (Electrical Resistance) Short-Time Test for Fatigue Strength of Metals, by Herbert F. Moore and Seichi Konzo. 1930. Twenty cents. *Bulletin No. 206. Studies in the Electrodeposition of Metals, by Donald B. Keyes and Sherlock Swann, Jr. 1930. Ten cents. *Bulletin No. 207. The Flow of Air through Circular Orifices with Rounded Approach, by Joseph A. Poison, Joseph G. Lowther, and Benjamin J. Wilson. 1930. Thirty cents. *CircularNo. 20. An Electrical Method for the Determination of the Dew-Point of Flue Gases, by Henry Fraser Johnstone. 1929. Fifteen cents. Bulletin No. 208. A Study of Slip Lines, Strain Lines, and Cracks in Metals under Repeated Stress, by Herbert F. Moore and Tibor Ver. 1930. Thirty-five cents. Bulletin No. 209. Heat Transfer in Ammonia Condensers. Part III, by Alonzo P. Kratz, Horace J. Macintire, and Richard E. Gould. 1930. Thirty-five cents. Bulletin No. 210. Tension Tests of Rivets, by Wilbur M. Wilson and William A. Oliver. 1930. Twenty-five cents. Bulletin No. 211. The Torsional Effect of Transverse Bending Loads on Channel Beams, by Fred B. Seely, William J. Putnam, and William L. Schwalbe. 1930. Thirty-five cents. Bulletin No. 212. Stresses Due to the Pressure of One Elastic Solid upon Another, by Howard R. Thomas and Victor A. Hoersch. 1930. Thirty cents. Bulletin No. 213. Combustion Tests with Illinois Coals, by Alonzo P. Kratz and Wilbur J. Woodruff. 1930. Thirty cents. *Bulletin No. 214. The Effect of Furnace Gases on the Quality of Enamels for Sheet Steel, by Andrew I. Andrews and Emanuel A. Hertzell. 1930. Twenty cents. Bulletin No. 215. The Column Analogy, by Hardy Cross. 1930. Forty cents. Bulletin No. 216. Embrittlement in Boilers, by Frederick G. Straub. 1930. None available. Bulletin No. 217. Washability Tests of Illinois Coals, by Alfred C. Callen and David R. Mitchell. 1930. Sixty cents. Bulletin No. 218. The Friability of Illinois Coals, by Cloyde M. Smith. 1930. Fifteen cents. Bulletin No. 219. Treatment of Water for Ice Manufacture, by Dana Burks, Jr. 1930. Sixty cents. *Bulletin No. 220. Tests of a Mikado-Type Locomotive Equipped with Nicholson Thermic Syphons, by Edward C. Schmidt, Everett G. Young, and Herman J. Schrader. 1930. Fifty-five cents.

tCopies of the complete list of publications can be obtained without charge by addressing the Engineering Experiment Station, Urbana, Ill. *A limited number of copies of bulletins starred are available for free distribution. ILLINOIS ENGINEERING EXPERIMENT STATION

*BulletinNo. 221. An Investigation of Core Oils, by Carl H. Casberg and Carl E. Schubert. 1931. Fifteen cents. *Bulletin No. 222. Flow of Liquids in Pipes of Circular and Annular Cross- Sections, by Alonzo P. Kratz, Horace J. Macintire, and Richard E. Gould. 1931. Fifteen cents. Bulletin No. 223. Investigation of Various Factors Affecting the Heating of Rooms with Direct Steam Radiators, by Arthur C. Willard, Alonzo P. Kratz, Maurice K. Fahnestock, and Seichi Konzo. 1931. Fifty-five cents. *BulletinNo. 224. The Effect of Smelter Atmospheres on the Quality of Enamels for Sheet Steel, by Andrew I. Andrews and Emanuel A. Hertzell. 1931. Ten cents. *Bulletin No. 225. The Microstructure of Some Porcelain Glazes, by Clyde L. Thompson. 1931. Fifteen cents. Bulletin No. 226. Laboratory Tests of Reinforced Concrete Arches with Decks, by Wilbur M. Wilson. 1931. Fifty cents. *Bulletin No. 227. The Effect of Smelter Atmospheres on the Quality of Dry Process Enamels for Cast Iron, by A. I. Andrews and H. W. Alexander. 1931. Ten cents. *CircularNo. 21. Tests of Welds, by Wilbur M. Wilson. 1931. Twenty cents. Bulletin No. 228. The Corrosion of Power Plant Equipment by Flue Gases, by Henry Fraser Johnstone. 1931. Sixty-five cents. *Bulletin No. 229. The Effect of Thermal Shock on Clay Bodies, by William R. Morgan. 1931. Twenty cents. Bulletin No. 230. Humidification for Residences, by Alonzo P. Kratz. 1931. Twenty cents. *BulletinNo. 231. Accidents from Hand and Mechanical Loading in Some Illinois Coal Mines, by Alfred C. Callen and Cloyde M. Smith. 1931. Twenty-five cents. *Bulletin No. 232. Run-Off Investigations in Central Illinois, by George W. Pickels. 1931. Seventy cents. *BulletinNo. 233. An Investigation of the Properties of Feldspars, by Cullen W. Parmelee and Thomas N. McVay. 1931. Thirty cents. *BulletinNo. 234. Movement of Piers During the Construction of Multiple-Span Reinforced Concrete Arch Bridges, by Wilbur M. Wilson. 1931. Twenty cents. Reprint No. 1. Steam Condensation an Inverse Index of Heating Effect, by Alonzo P. Kratz and Maurice K. Fahnestock. 1931. Ten cents. *Bulletin No. 235. An Investigation of the Suitability of Soy Bean Oil for Core Oil, by Carl H. Casberg and Carl E. Schubert. 1931. Fifteen cents. *Bulletin No. 236. The Electrolytic Reduction of Ketones, by Sherlock Swann, Jr. 1931. Ten cents. *Bulletin No. 237. Tests of Plain and Reinforced Concrete Made with Haydite Aggregates, by Frank E. Richart and Vernon P. Jensen. 1931. Forty-five cents. *Bulletin No. 238. The Catalytic Partial Oxidation of Ethyl Alcohol, by Donald B. Keyes and Robert D. Snow. 1931. Twenty cents. *Bulletin No. 239. Tests of Joints in Wide Plates, by Wilbur M. Wilson, James Mather, and Charles O. Harris. 1931. Forty cents. *Bulletin No. 240. The Flow of Air through Circular Orifices in Thin Plates, by Joseph A. Polson and Joseph G. Lowther. 1932. Twenty-five cents. *Bulletin No. 241. Strength of Light I Beams, by Milo S. Ketchum and Jasper 0. Draffin. 1932. Twenty-five cents. *Bulletin No. 242. Bearing Value of Pivots for Scales, by Wilbur M. Wilson, Roy L. Moore, and Frank P. Thomas. 1932. Thirty cents. *Bulletin No. 243. The Creep of Lead and Lead Alloys Used for Cable Sheathing, by Herbert F. Moore and Norville J. Alleman. 1932. Fifteen cents. *Bulletin No. 244. A Study of Stresses in Car Axles Under Service Conditions, by Herbert F. Moore, Nereus H. Roy, and Bernard B. Betty. 1932. Forty cents. *Bulletin No. 245. Determination of Stress Concentration in Screw Threads by the Photo-Elastic Method, by Stanley G. Hall. 1932. Ten cents. *Bulletin No. 246. Investigation of Warm-Air Furnaces and Heating Systems, Part V, by Arthur C. Willard, Alonzo P. Kratz, and Seichi Konzo. 1932. Eighty cents. *Bulletin No. 247. An Experimental Investigation of the Friction of Screw Threads, by Clarence W. Ham and David G. Ryan. 1932. (In press.) *Bulletin No. 248. A Study of a Group of Typical Spinels, by Cullen W. Parmelee, Alfred E. Badger, and George A. Ballam. 1932. Thirty cents. UNIVERSITY OF ILLINOIS HARRY WOODBURN CHASE, Ph.D., LL.D., L.H.D., President

The University includes the following departments: The Graduate School The College of Liberal Arts and Sciences (Curricula: General with majors, in the Humanities and the Sciences; Chemistry and Chemical Engi- neering; Pre-legal; Pre-medical; Pre-dental; Pre-journalism; Applied Optics) The College of Commerce and Business Administration (Curricula: Gen- eral Business, Banking and Finance, Insurance, Accountancy, Trans- portation, Industrial Administration, Foreign Commerce,' Commercial Teaching, Trade and Civic Secretarial Service, Public Utilities, Com- merce and Law) The College of Engineering (Curricula: Ceramics; Ceramic, Civil, Electri- cal, Gas, General, Mechanical, Mining, and Railway Engineering; En- gineering Physics) The College of Agriculture (Curricula: General Agriculture; Floriculture; Home Economics; Smith-Hughes-in conjunction with the College of Education) The College of Education (Curricula: Two year, prescribing junior stand- ing for admission-General Education, Smith-Hughes Agriculture, Smith-Hughes Home Economics, Public School Music; Four year, ad- mitting from the high school-Industrial Education, Athletic Coaching, Physical Education. The University High School is the practice school of the College of Education) The College of Law (three-year curriculum based on a college degree, or three years of college work at the University of Illinois) The College of Fine and Applied Arts (Curricula: Music, Architecture, Ar- chitectural Engineering, Landscape Architecture, and Painting) The Library School (two-year curriculum for college graduates) The School of Journalism (two-year- curriculum -based on two years of college work) The College of Medicine (in Chicago) The College of Dentistry (in Chicago) The College of Pharmacy (in Chicago) The Summer Session (eight weeks) Experiment Stations and Scientific Bureaus: U. S. Agricultural Experiment Station; Engineering Experiment Station; State Natural History Sur- vey; State Water Survey; State Geological Survey; Bureau of Educa- tional Research; Bureau of Business Research. The Library Collections contain (July 1, 1931) 832,643 volumes' and 221,000 pamphlets (in Urbana) and 45,241 volumes and 7,875 pamphlets ,(in Chicago) For catalogs and information address TmE RiwcSTRAR, Urbana, Illinois.

- ^ .11 xi 6 s· ;'-'~ i: