AFOSR=TN-57-540 ASTIA Document No, AD 136 526 ENGINEERING RESEARCH INSTITUTE THE UNIVERSITY OF MICHIGAN ANN ARBOR AN INVESTIGATION OF REACTIONS IN FERROSPINELS BY DIFFERENTIAL THERMAL ANALYSIS Technical Report No. 4 Solid State Devices Laboratory Department of Electrical Engineering Co F Jefferson Project 2495 AIR FORCE OFFICE OF SCIENTIFIC RESEARCH AIR RESEARCH AND DEVELOPMENT COMMND SOLID STATE SCIENCE DIVISION CONTRACT NOo AF 18(603)-8, DIVISION FILE NO 40-24 July 1957

The University of Michigan * Engineering Research Institute ABSTRACT Differential thermal analysis has been used to follow the oxidation of magnetite in solid solution with a nickel-zinc ferrospinel, and the solid-state reaction leading to the formation of the ferrospinel from the oxides. Differential thermal curves for the series Ni.474 Zn.526 Fe204 - Fe304 are shown and explained. An attempt was made to follow the solid-state reaction which occurs during the formation of the ferrospinel from the oxides. ii

The University of Michigan ~ Engineering Research Institute THE OXIDATION OF THE SYSTEM Ni.474 Zn526 Fe24 - Fe304 Ferrospinels utilized as ferromagnetic elements in electric circuits have the advantages of high permeability combined with high resistivity. One detrimental factor which contributes to the lowering of the resistivity is ferrous iron in the structure. Ferrous iron also effects the anisotropy and magnetostruction which in turn effect the permeability. It is therefore necessary to understand the conditions under which it forms and methods of oxidizing it once it has formed, to control better the magnetic properties of the ferrospinels. The usual procedure for the preparation of the nickel-zinc ferrite is to mix intimately NiO, ZnO, and Fe203 in the desired proportions, press them into the required form, and fire at 1100~C to 1400~C until the solid-state reaction has been completed. Ferrous iron will form if Fe203 is present in excess of the amount required for the formation of NixZnl_xFe204. Jefferson1 has investigated the formation of ferrous iron in a nickel zinc ferrospinel under these conditions. When the oxides are mixed in stoichiometric proportions, there are two mechanisms by which ferrous iron can form. The formation below about 1300~C is caused by localized nonstoichiometric regions. Ferrous iron content due to this cause can be minimized by thorough mixing of the oxides and a prolonged firing time. The effect of firing time and extent of mixing on ferrous iron formation in a stoichiometric nickel-zinc ferrospinel was investigated and reported by Jefferson and Grimes.2 Above about 1300~C, the presence of ferrous iron in stoichiometric material is due to the decomposition of the nickel-zinc ferrospinel. Van Uitert3 has shown that volatilization of zinc occurs, which leaves the material with an excess of Fe203. Ferrous iron formation by this mechanism increases with firing, rather than decreases as occurs in the first case. It has been known for some time that two iron sequioxides exist. c6Fe203 is the common form and has the hexagonal structure. yFe20O is cubic and differs from magnetite in that cation vacancies exist in the structure. It is known that under certain conditions magnetite oxidizes to 7Fe203 followed by transformation to aFe203. Due to the similarity in crystal structure between yFe203 and magnetite, direct evidence for the, transformation has been difficult to obtain, and the subject remains controversial. Kojima4 studied the transformations in iron oxides by following the magnetic remanence and coercive force as a function of temperature in various atmospheres. He concluded that yFe203 is an intermediate product in both the oxidation of magnetite of aFe203 and in the reduction of CFe20O to magnetite. David and Welch5 investigated magnetite prepared by reduction of aFe203 and by precipitation. Only the magnetite prepared by precipitation showed the formation of yFe203 on oxidation. The authors concluded that 7Fe203 is stable only if water is present in the lattice. Behar6 conducted a micrographic study of. —---------._.. —-------------------— 1

The University of Michigan * Engineering Research Institute the oxidation of magnetite, and concluded that oxidation occurs first in the [111] direction. The oxidation of magnetite has been investigated by differential thermal analysis by several investigators. Magnetitie shows two characteristic exothermic peaks. The first peak occurs at about 300~C, while the second broader peak occurs at about 900~C. Kulp and Trite7 concluded that synthetic magnetite, prepared by a reduction of cFe203 at low temperatures, is transformed to yFe203, but they attributed the first peak in natural magnetite to recrystallization with no formation of yFe203 occurring. This conclusion is based on the facts that the size of the peak is proportional to particle size, and that the material heated beyond the first peak contains a considerable amount of ferrous iron. Schmidt and Vermaas concluded that the first peak in natural magnetite is due to surface oxidation and the second peak to volume oxidation instigated by recrystallization of the protective surface layer of cFe203. They agree with Kulp and Trite, that no yFe203 is formed. An investigation of the oxidation of the solid-solution series Ni474 Zn.526 Fe204 - Fe304 has been undertaken to attempt to understand the processes involved in oxidizing the ferrous iron in solid solution with the nickel-zinc ferrospinel. The samples used were prepared by ball-milling NiO, ZnO, and cFe203 in acetone for six hours. The oxides were dried, pressed into compacts, fired in air for 1425~C for one hour, and water-quenched. The samples were then crushed and passed through a 325-mesh sieve. The differential thermal analysis curves were obtained with a unit built according to the Department of Agriculture design. A platinum-vs-platinum-lOo-rhodium differential thermocouple was used at a sensitivity setting of 2.9 microvolts per centimeter. A chromelalumel thermocouple was used to determine the temperature of the block. This thermocouple was calibrated against the inversion temperature of quartz using the procedure given by Faust.9 The heating rate was 12~C per minute. The differential thermal curves are shown in Fig. 1. The curves show a slight exothermic peak at 360~C from the Curie temperature of the nickel sample holder. The curve for magnetite shows the two characteristic exothermic peaks along with a break at the Curie temperature of magnetite, 585~C. To identify the cause of the exothermic peaks, the oxidation of magnetite was followed by determining the change in weight, and the change in ferrous iron content of a sample heated in air for 15 minutes at 500C temperature intervals. Figure 2 shows the percent change in weight. There is a break in the curve between 250~ and 300~C. This same break was obtained in the ferrous iron content of the samples. The change in rate of oxidation at this temperature interval would appear to explain the low-temperature exothermic peak in the differential thermal curve of magnetite. It is concluded that this peak in magnetite is caused by surface oxidation, as proposed by Schmidt and Vermaas.8 Schmidt and Vermaas state that the oxidation of magnetite is accelerated at the Curie temperature. Figure 2 shows no indication that this is the case. It appears that the second peak is not due to any sudden initiation of oxidation 2

The University of Michigan * Engineering Research Institute but rather makes its appearance when the continuously increasing rate of oxidation becomes sufficiently rapid. Curves 1 through 9 of Fig. 1 are the differential thermal curves for the solid-solution series Ni474 Zn 526 Fe204 - Fe304. Curve 1 shows a broad exothermic peak, due to the oxidation of the ferrous iron formed in the stoichiometric ferrospinel at the temperature of formation of 1425~C. As the magnetite content is increased in Curves 2 to 9, this peak becomes less broad and moves down in temperature and corresponds with the low temperature peak in magnetite. Curves 8 and 9 have what appears to be two superimposed peaks. There is the possibility that the first of these two peaks increases in amplitude to become the low-temperature peak in magnetite. This does not seem as.likely as the above explanation. In Curve 4 a second exothermic peak makes its appearance. This peak moves down in temperature with increasing magnetite content and decreases in amplitude, until it has almost disappeared in Curve 10. A third higher temperature peak makes its appearance in Curve 6. (The second peak in Curve 5 might well be a summation of two superimposed peaks.) This third peak moves up in temperature and corresponds with the high-temperature peak in the magnetite curve. By identifying the exothermic peaks in the solid-solution series with those in magnetite, it can be concluded that the surface oxidation of the magnitite in solid solution occurs at a higher temperature than in magnetite itself, and the oxidation in the interior occurs at a lower temperature. The peak occurring at the intermediate temperature is thought to be due to the transformationr of yFe203 to ceFe203. Since Michel and Lensen10 and others have shown that certain foreign ions stabilize the yFe203 structure, it seems reasonable to assume that in the solid solutions investigated some stabilization of yFe203 has occurred. THE SOLID-STATE REACTION The use of differential thermal analysis to study the formation of the ferrospinel from the oxides was attempted. The sintering of the material to the thermocouple proved to be a problem, and only one or two runs were possible before it became necessary to replace the differential thermocouple. The differential thermal curves of Fig. 3 were obtained from oxide samples ball-milled to insure proper mixing. The equipment used was similar in design to that described above, but the recording was made with an X-Y Recorder. The sensitivity was 200 microvolts per cm and the heating rate was 250 per minute. A stainless steel sample holder replaced the nickel one. All the curves in Fig. 3 show an exothermic peak between 300 and 325~Ce Curve 1 is from magnetite prepared by firing aFe203 at 1425~C while Curve 2 is from aFe203 ball-milled for six hours in acetone. Curve 3 is from the same 5

The University of Michigan * Engineering Research Institute CoFe203 before it was ball-milled. From the resemblance between the low-temperature peaks in Curves 1 and 2, it is concluded that a surface layer of magnetite is formed on the tFe203 particles. Curves 4 through 8 are for the following oxide mixtures~ NiO + Fe203; ZnO + Fe203; o4 NiO +.6 ZnO + Fe203;.474 NiO + o526 ZnO + 1.294 Fe203; and.4 NiO + o6 ZnO + Fe203 +.0075 V205, respectively. The break at 680~ in Curves 2 through 8 corresponds to the temperature of transformation of cFe203 from an antiferromagnetic state to a paramagnetic state. According to Willis and Rooksky,11 there is a sudden expansion along the triad axis and a change in the rhombohedral angle on cooling through this temperature. There is an indication that NiO + Fe203 (Curve 4) reacts at a higher temperature than ZnO + Fe203 (Curve 5). The curves for samples of.4 NiO +.6 ZnO + Fe203 and.474 NiO +.526 ZnO +.1294 Fe203 are difficult to interpret. Curve 6 resembles Curve 5, while Curve 7 appears to have a double peak at the higher temperatures. The addition of V205 in small amounts has been shown by Grimes, Thomassen, Jefferson, and Kothary12 to lower the temperature of ferrospinel formation. Curve 8 from material containing a small amount of V205 shows a decided exothermic peak at about 820~C. It is interesting to note that the melting point of V205 is given as 800~C in the Handbook of Chemistry.13 CONCLUSION The oxidation of magnetite occurs in two stages. The surface oxidation at -about 250-300~C is marked by a sharp exothermic peak in the differential thermal curve. The rate of volume oxidation increases with temperature and becomes sufficiently rapid at the higher temperatures to give rise to a second broad exothermic peak at 900~C. The surface oxidation of magnetite in solid solution with a nickel-zinc ferrospinel occurs at a higher temperature than in magnetite itself, while the volume oxidation occurs at a lower temperature. A third exothermic peak occurs in the intermediate compositions which has tentatively been identified with the transformation of Fe203 to CFe203. The use of differential thermal analysis to study the solid-state reaction occurring during the formation of the ferrospinel from the oxides did not prove very informative. There is an indication that NiO + Fe203 reacts at a higher temperature than does ZnO + Fe203. The presence of an exothermic peak in all material ball-milled in steel ball mills was traced to C4Fe203. The similarity in the differential thermal peak of this material to that of magnetite leads to the conclusion that magnetite has formed on the surface of the particles. The addition of small amounts of V205 to the oxide mixture is known to lower the ferrospinel formation. Differential thermal curves of the material shows an exothermic peak near the melting point of V205. 4

The University of Michigan ~ Engineering Research Institute REFERENCES 1. C. F. Jefferson, An Investigation of the Composition on an Iron-Rich Nickel-Zinc Ferrite. Technical Report No. 66, Electronic Defense Group, Department of Electrical Engineering, The University of Michigan, June, 1956. 2. C. F. Jefferson, and D. M. Grimes, A Study of the Preparation of NickelZinc Ferrites, Technical Report No. 58, Electronic Defense Group, Department of Electrical Engineering, The University of Michigan, January, 1956. 3. L. G. Van Uitert, "dc Resistivity in the Nickel and Nickel-Zinc Ferrite System," J. of Chem. Physics, 23, 1883-1887 (1955). 4o Ho Kojima, On the Magnetic Property of Iron Oxides, The Science Reports of the Research Institutes, Tohokei University, Series A 6, 178-183 (1954). 5. I. David, and A. J. E. Welch, "The Oxidation of Magnetite and Related Spinels," Trans. of the Faraday Society, 52, 1642-50 (1956). 6. I. Behar, "Micrographic Study of the Oxidation of Magnetite," Comptes Rendus, 243, 1877-1880 (Dec., 1956). 7. S. L. Kulp,and A, F. Trite, "Differential Thermal Analysis of Natural Ferric Oxide," American Mineralogist, 36, 23-44 (1951). 8. E. R. Schmidt, and F. H. S. Vermaas, "Differential Thermal Analysis and Cell Dimensions of Some Natural Magnetites," American Mineralogist, 40, 422-431 (1955). 9. G. T. Faust, "Thermal Analysis of Quartz and Its Use in Calibration of Thermal Studies," American Mineralogist, 33, 337-345 (1948). 10. A. Michel, and M. Lensen, "On the Stabilization of the Cubic Sequioxide of Iron," Comptes Rendus, 243, 1422-1423 (1956). 11. B. T M. Willis, and H. P. Rooksky, "Crystal Structure and Antiferromagnetism on Haematite," Phys. Society of London, B, 65, 950-954 (1952). 12. D. M. Grimes, L. Thomassen, C. F. Jefferson, and N. C. Kothary, "Effect of V205 on Nickel-Zinc Ferrite Formation," J. Chem. Phys., 23, 2205, 1955. 13. Handbook of Chemistry, Ed. N. A. Lange, Handbook Pub., Inc., Sandusky, 1944. 5 ---

The University of Michigan * Engineering Research Institute TEMPERATURE, ~C 200 400 600 800 1000 I Ni 474 Zn r53p Fep04 ~2 | 1~- _Ni 474 Zn.526 e204 +.090 Fe304 Ni.474 Zn526 Fe204 +.196 Fe304 4 N.474 Zn.526 Fe204 +.380 Fe004h 5 Ni.474 Zn.526 Fe204 +.626 Fe304 z 6 7 / \/ \ Ni 474 znt526 Fe204 + 1.33 Pe.304 w7 h74 Zn 526 Fe-204 + 2.00 Fe304 8 Ni.4(4 Zn.526 Fe204 + 2.66 Fe304 9 \ Z; \N.474 Zn.526 Fe204 + 5.78 Fe304 |. / \ ^Fe 304 10 I I. I I I I, I 200 400 600 800 1000 TEMPERATURE, ~C Fig. 1. Differential thermal curves for the solid solution series Ni 474 Zn 526 Fe204 - Fe304. 6

The University of Michigan * Engineering Research Institute co.rt Ur II 10 (0.r-C h0 8 -p -c Q) O (0CX ~ ~ ~ ~ ~ ~ ~( 0J OD c 0 LHSI3M NI 3SV3HONI IN308Jd nr-'H -7 —-- 8 oJ ~.~~~r ~mm ~~~~~~~~~~~~~~~~~~ (P CI~~~~~~~~~~~~~~~~~~~~ Q Q) ~~~~~~~~~~~~~~~~~~~~~~ C~~~~~~~~~~ O~~~~?d d d dho~~~~ ~. ~ L~1mm 8I3'~3Rl L 3 t~ ~ -t-) ~.rm "''" — ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~TC)

The University of Michigan * Engineering Research Institute TEMPERATURE, ~C 200 400 600 800 1000 II I I I I I I I cFe20 3'ball milled in steel ball mill 2 cfFe203 before ball milling 3 O 4 NiO + Fe20g H4 LJ in~ | ~~~ZnO + Fe203 -J O 6 1.4 NiO +.6 ZnO + Fe20.474 NiO +.526 ZnO + 1.294 Fe2O3 7 o.4 NiO +.6 ZnO + Fe203 -.0075 V205 8 I I I I I I I I 200 400 600 800 1000 TEMPERATURE, ~C Fig. 3. Differential thermal curves of the formation of the ferrospinel from the oxides. 8

The University of Michigan * Engineering Research Institute DISTRIBUT ION LIST No. of No. of Agency Copies Agency Copies Commander 5 Director of Research and 1 Hq,, AF Office of Scientific Development Research, ARDC Hq., U. S. Air Force Attn: SRQB Attn: AFDRD-RE- 3 Washington 25, D. C. Washington 25, D. C. Commander 4 Department of the Navy 2 Wright Air Development Center Office of Naval Research Attn: WCRRH Attn: Code 423 Attn: WCRRL Attn: Code 421 Attno WCRTL Washington 25, D. C. Attn: WCRTM-i Wright-Patterson Air Force Officer-in-Charge Base, Ohio Office of Naval Research Navy iNo. 100 Commander 2 Fleet Post Office Air Force Cambridge Research New York, New'York Center Attn: Technical Library Commanding Officer 1 Attno CRRF Naval Radiological Defense L. G. Hanscom Field Laboratory Bedford, Massachusetts San Francisco Naval Shipyard San Francisco 24, California Commander 1 Rome Air Development Center Director, Research and Develop- 1 Attn: Technical Library ment Division, General Staff Griffiss Air Force Base Department of the Army Rome, New York Washington 25, D. Co Director, Office for Advanced 1 Division of Research 1 Studies Uo S. Atomic Energy Commission Air Force Office of Scientific 1901 Constitution Avenue, N.Wo Research Washington 25, D. C, P. 0. Box 2035 Pasadena 2, California U. So Atomic Energy Commission 1 Library Branch Document Service Center 10 Technical Information Divi.sion, Armed Services Technical ORE Information Agency P.o O Box E Knott Building Oak Ridge, Tennessee Dayton 2, Ohio 9

The University of Michigan * Engineering Research Institute DISTRIBUTION LIST (concluded) No. of No. of AoAgency Copies Agency Copies Oak Ridge National Laboratory 1 Commander 1 Attn: Central Files Western Development Division P. 0. Box P (ARDC) Oak Ridge, Tennessee Attn: WDSIT, P. 0. Box 262 Inglewood, California Brookhaven National Laboratory 1 Attn: Research Library Document Custodian 1 Upton, Long Island, New York Los Alamos Scientific Laboratory P. O0 Box 1663 Argonne National Laboratory 1 Los Alamos, New Mexico Attn: Librarian P. O. Box 299 Arnold Engineering Development Lemont, Illinois Center, Attn: Tech. Library P. Oo Box 162 Ames Laboratory 1 Tullahoma, Tennessee Iowa State College P. 0. Box 14A, Station A Commanding Officer Ames, Iowa Ordnance Materials Research Office Watertown Arsenal Knolls Atomic Power Laboratory 1 Watertown 72, Massachusetts Attn~ Document Librarian P. 0. Box 1072 Commanding Officer 1 Schenectady, New York Watertown Arsenal Watertown 72. Massachusetts National Bureau of Standards 1 Attn: Watertown Arsenal Labs. Library, Room 203 Tech. Reports Section Northwest Building Washington 25, D, C. Commander, Hq., AF Office of 2 Scientific Research, ARDC National Science Foundation 1 Attn~ SREC, Tech. Library 1520 H. Street, N.W. Washington 25, D. C. Washington 25, D. C. C nd Commander, WADC 1 Air Technical: Intelligence Center Director, Office of Ordnance 1 Director,_1~ Attn: Deputy for Documentation Research, Box CM, Duke WPAFB Ohio Station Durham, North Carolina (Inner envelope-no postage) Commander, European Office 1 Office of Technical Services 1 c/o American Embassy, ARDC Department of Commerce Brussels, Belgium Washington 25, D. C. (Outer envelope-postage) Superintendent, Diplomatic Pouch Rooms, Department of State 1512 H. Street, N.W. Washington 25, D. C. Washington 25, D. C. 10

UNIVERSITY OF MICHIGAN 3 90111 1 111 11 111111111111111 3 9015 03025 4984