Creator:W.F. Claussen, R.C. DeVries, and J.D. Mackenzie Date Created:March 1960 Place Created:Schenectady, New York Keywords:crystalline and vitreous ceramic materials Context:document from General Electric research offices ************************************************** PROGRESS REPORT NO. 3 THE INFLUENCE OF HIGH PRESSURES AND HIGH TEMPERATURES ON TRANSFORMATIONS AND COORDINATION IN CRYSTALLINE AND VITREOUS CERAMIC MATERIALS W.F. Claussen, R.C. DeVries, andJ.D. Mackenzie Contract No. AF-33(616)-6295 March 1960 Wright Air Development Center SCHENECTADY, NEW YORK 60-GC-84 GENERAL @ ELECTRIC When Government drawings, specifications, or other data are used for any purpose other than in connection with a definitely related Government procurement operation, the United States Government thereby incurs no responsibility nor any obligation whatsoever; and the fact that the Government may have formulated, furnished, or in any way supplied the said drawings, specifications, or other data, is not to be regarded by implication or otherwise as in any manner licensing the holder or any other person or corporation, or conveying any rights or permission to manufacture, use, or sell any patented invention that may in any way be related thereto. This document may not be reproduced or published in any form in whole or in part without prior approval of the Government. Since this is a progress report, the information herein is tentative and subject to changes, corrections and modifications. PROGRESS REPORT NO. 3 TO WRIGHT AIR DEVELOPMENT CENTER THE INFLUENCE OF HIGH PRESSURES AND HIGH TEMPERATURES ON TRANSFORMATIONS AND COORDINATION IN CRYSTALLINE AND VITREOUS CERAMIC MATERIALS W.F. Claussen, R.C. DeVries, and J.D. Mackenzie Contract No. AF-33(616)-6295 Period: October 1 - December 31, 1959 Research Laboratory GENERAL ELECTRIC COMPANY Schenectady, New York March 1960 60-GC -84 THE INFLUENCE OF HIGH PRESSURES AND HIGH TEMPERATURES ON TRANSFORMATIONS AND COORDINATION IN CRYSTALLINE AND VITREOUS CERAMIC MATERIALS W.F. Claussen, R.C. DeVries, and J.D. Mackenzie INTRODUCTION The broad objective of this study is to gain an understanding of the crystal chemistry, phase equilibria, and structural relationships of ceramic materials formed under conditions of high pressure and temperature. Experimentally this will be accomplished by: (1) subjecting selected materials (which are known or expected to exhibit coordination changes leading to denser phases or forms) to various combinations of pressure and temperature (up to approximately 2000°C and 100,000 atmospheres); and (2) subsequent examination with various techniques, to define the phases and structures produced. The work includes corroboration of previous work at high pressures and temperatures, extension of the investigations to more extreme conditions, and the study of new systems. Emphasis will be placed on the examination of the structural and stability aspects of the resulting phases. The study can be divided into two parts on the basis of the structural types of the two ceramic systems to be investigated. Part A covers the investigation of silica and other inorganic compounds of similar structure. The structural changes in both the crystalline and the glassy forms will be investigated. This part is under the supervision of J.D. Mackenzie and W.F. Claussen. Part B is a study of the two-component Si02-Al203 system and some isotypic systems formed by the separate or combined substitution of Ge02 for Si02 and Ga203 for AI2O3. This part is supervised by R. C. DeVries and W. F. Claussen. PART A Compared with work on crystalline systems, few high-pressure studies have been made on vitreous phases. It is generally considered that the density of a glass is necessarily less than that of the corresponding crystalline phase. Bridgman and Simon (1953), however, have demonstrated that vitreous silica may be compacted until its density is greater than that of cristobalite. In crystalline silica, the different -1- arrangements of Si04 tetrahedra in the continuous network can give rise to polymorphs having grossly different properties. Thus, cristobalite of density 2.32 is fairly rapidly attacked by hydrofluoric acid, whereas coesite of density 3. 00 is completely inert. It is interesting to subject vitreous silica to high pressures and temperatures and to investigate (a) if the preparation of a glass of density greater than that of quartz (p = 2. 65) is possible, and (b) its different physical and chemical properties. During this period, nine more high pressure experiments have been carried out, three on vitreous germania and six on vitreous silica. A summary of the experimental conditions is presented in Table I. Only platinum capsules were used and the cell employed was similar to that shown in the previous report. Pressure was applied at a rate of approximately 2000 atm/sec and the sample maintained at the maximum pressure for 2 minutes. Both x-ray and microscopic examinations were made immediately and also four weeks after the experiment to ensure no devitrification of the glass had taken place. TABLE I Summary of High-Pressure Experiments on _Vitreous Silica and Germania_ Pressure Temp Refractive Density Run No. Substance (atm) (°C) Index Density Increase 275 Si02 90,000 150 1.510 2.470 11,2 340 Si02 80, 000 400 1. 522 2.524 13.7 341 Si02 80,000 500 1.528 2.558 15.2 423 Si02 100,000 300 424 Si02 100,000 500 450 Si02 100,000 150 451 Si02 100, 000 300 2.550 444 Si02 100,000 600 2.595 420 Ge02 100,000 250 1.694 4.16 13.3 421 GeO 2 100,000 400 1.720 4.25 15.8 422 Ge02 100,000 RT 1.668 4.00 9.0 The observed refractive indices of the compacted glass are recorded in column 5 of Table I while the density data are given in column 6. In the last report, the density data were evaluated from a plot of refractive index and density of crystalline silica, it being assumed that values of the compacted glass will fall on this smooth curve. Over this period, calibration of the liquid density columns have been completed for the density range 2.0 to 3.0 g/cc. For silica, therefore, density could be determined directly. In Fig. 1, the density-refractive index relationship for the compacted glass is seen to follow the smooth curve as assumed previously. No density column has yet been prepared to cover the range 4.0 to 5.0 g/cc, and so the density results for germania were still evaluated from the measured refractive index. It is observed in Table I that, for germania, glasses of density equal to that of the crystalline phase have now been prepared. For silica, however, the highest density observed was 2. 56 as compared to the density of 2. 65 for crystalline quartz. 2.0 2.2 2.4 2.6 2.8 3.0 DENSITY GMS/ml. Fig. 1 Density-refractive index relationship for crystalline and compacted silica. o COMPACTED Si02 GLASS □ OTHER FORMS OF Si02 - Fig. 2 Hardness-density relationship for some compacted silica. HARDNESS MEASUREMENTS In view of the large increases observed of the density and refractive index of the glasses, it was thought that the hardness of the samples might show a similar variation. The hardness of compacted silica of various densities were accordingly measured with a Kentron microhardness tester. Samples of suitable size and surface were uncommon because of the small dimensions of the pressure cell. The results obtained on nine samples are shown in Fig. 2. It is seen that both the deviations and scatter were too large to warrant any conclusion. Hardness measurements of this type will therefore not be carried out again in the coming period. SOLUBILITY IN HF SOLUTIONS Coes (1953) had demonstrated that the solubility of ordinary vitreous silica and all the known modifications of crystalline silica in dilute hydrofluoric acid is a direct function of density. This is not known - for compacted glass. Over this period, an empirical technique has been developed for studying this property of the dense glass. Selected pieces weighing about 10 mg were put into 5 per cent HF solutions for 24 hours. The loss of weight was then determined. The method is necessarily approximate because of the variation in shapes and sizes of the samples. Only a few measurements have been made so far and the results are inconclusive. These experiments will be continued in the next period. Germania will also be examined. In the coming period, density columns covering the range of 4. 0 to 5.0 will be prepared for the germania samples. Infrared studies will be made on the compacted glass to determine if any significant structural changes have occurred. Further high-pressure experiments will be carried out in an attempt to prepare vitreous silica and germania of even higher density. More accurate measurements of temperature will be made by the use of thermocouples in the high-pressure cells in order to obtain PVT relationships for both vitreous silica and germania. PART B INTRODUCTION During the third quarter of the contract period, emphasis has been given to the study of the decomposition and transformation of the naturally occurring polymorphs of A^SiOs. This departure from the expressed intention (second quarterly report) to focus on the substitution of Ga203 for AI2O3 in Al2Si05 was made because of the consistent failure to produce any of the desired polymorphs from gel or mechanical mixtures in the Ga203-Si02 system and the only limited success in the Al203-Si02 system. Either reaction rates in both systems are too slow to reach what has been considered to be equilibrium at high pressures and temperatures in these systems or the assemblages found (Ga203 plus glass and corundum plus glass, respectively) do indeed represent equilibrium in the P-T region studied. The P-T conditions for the runs that have been made were based on the work of Clark et al. (1) (or an extrapolation of the results of that study), who concluded that the frequently found crystalline assemblage of AI2O3 and Si02 was metastable. By studying the decomposition of natural kyanite, andalusite, and sillimanite, the stability regions can be determined from a different direction than that offered by the synthetic approach, and such information 1. S,P. Clark, Jr., E.C. Robertson, and F. Birch, Am. J. Sci., 255, 628-640 (1957). - can then be used as a guide for future work with synthetic materials. As a result, a tentative curve for the decomposition of kyanite as a function of temperature and pressure has been determined, and some new observations on reaction paths and relative reaction rates for the decomposition of the polymorphs have been made. The data for kyanite are not in agreement with the previous workU) in which kyanite and sillimanite were synthesized from kaolinite, andalusite, and sillimanite. EXPERIMENTAL High-Pressure Cell The same cell type previously described in the second quarterly report was used. The cell assembly was the same as described recently in regard to diamond synthesis. (2) The internal heater in most of the runs was a graphite cylinder separated from the sample by a cylinder of alumina. The sample was enclosed in Pt or 80 Pt-20 Rh foil. Temperature and Pressure Measurement Several runs have been made with the above cell with a thermocouple of which the Pt capsule containing the sample is a part. In these runs a plot of watts vs temperature was made and the reproducibility of the plot was found to be ± 50°C. With this information it is possible to make the runs without inserting a thermocouple in every run, and considerable time is saved without appreciable sacrifice of precision. Most of the runs are made this way with an occasional use of a thermocouple to check the watts vs temperature curve. No correction has been applied for the change of emf as a function of pressure. The problem of an absolute pressure calibration is a difficult one, and we are in the unsatisfactory position of being able to state with certainty only the precision of the pressure measurement (± 2000 atm). The pressure calibration is made by measuring the change in resistance of Bi, Tl, or Ba wires with pressure at room temperature. The pressure transmitting medium adjacent to the wire is silver chloride. It obviously would be more satisfactory to have fixed points that could be determined under the same conditions as the runs. 2. Chemical and Engineering News, 24-25 (November 29, 1959). - Starting Materials Naturally occurring kyanite and andalusite from Minas Gerais, Brazil, and sillimanite from Dillon, Montana, were used. The specimens were examined optically and found to be free of other phases. A chemical analysis for the kyanite sample is currently being made. A few runs using mechanical mixtures of kaolinite and a synthetic aluminum trihydrate and some additional runs using previously described gels were also made. Identification of Samples In addition to the usual x-ray techniques and the transmitted polarized light methods of identification, all runs are also being examined by reflected light in polished section in order to see the microstructural relationships of the phases. This has proved to be a very sensitive technique for observing the beginning stages of the changes taking place. Identification still requires the combination of all the techniques. RESULTS AND DISCUSSION The data are summarized in Tables II through IV and are shown graphically in Figs. 3 and 4. More data are needed to firmly establish the kyanite decomposition curve shown as the heavy dashed line in Fig. 3, but the general trend is clear. At one atmosphere, decomposition was observed in 48 hours at about 1100°C. This is in agreement with the work of Greig. (3) Clark et al. (1) report no decomposition of kyanite at 1000°C in 4 days at 1 atm. At 1500°C and 45,000 atm, kyanite is stable while decomposition is observed at the same temperature at 35,000 atm. Other data consistent with the curve determined from kyanite are the partial conversion of sillimanite to kyanite at 1700°C and 50, 000 atm and the synthesis of kyanite from a gel at 50, 000 atm in the temperature range 1500° to 1800°C. Kyanite has also been synthesized from a gel at 20, 000 atm and about 1200°C, but not reproducibly, and there is some doubt as to the temperature of the successful run. 3. J.W. Greig, J. Am. Ceram. Soc., 8 (8), 465-484 (1925). - TABLE II Results of Experiments with Kyanite Pressure Run Starting Temp (atm) Time No. Material (°C) x 103 (min) 1624 Ky* 1850 30 4 1634 Ky 1830 50 10 1641 Ky 1820 20 10 1642 Ky 1790 40 5 1648 Ky 1530 25 45 1649 Ky 1500 30 35 1652 Ky 1850 45 10 1653 Ky 1500 35 13 1654 Ky 1950 50 10 1655 Ky 1720 35 10 1656 Ky 1350 30 15 1657 Ky 1320 20 15 1659 Ky 1910 45 10 1674 Ky 1500 45 30 1675 Ky 1300 15 30 *Ky = Kyanite + G = Graphite * Glass or material amorphous to x-rays. Heater Phases Present and Remarks Gt Mullite, corundum, glass* G Kyanite beginning to decompose, some glass and corundum at grain boundaries G Mullite, glass G Kyanite decomposing at grain boundaries, some glass and corundum G Mullite, glass, corundum at grain boundaries of original kyanite G Mullite, undecomposed kyanite, glass G Kyanite decomposition evident, glass and corundum at grain boundaries G Kyanite decomposition uncertain G Kyanite, considerable decomposition, mullite, corundum, glass G Kyanite decomposition, corundum and glass at grain boundaries G Kyanite decomposition uncertain G Mullite, some undecomposed kyanite, glass, corundum (?) G Considerable decomposition of kyanite; mullite, glass, corundum at grain boundaries G Kyanite G Mullite, corundum, glass - TABLE III Results of Experiments with Andalusite and Sillimanite Run No. Starting Material Temp (°C) Pressure (atm) x 103 Time (min) Heater Phases Present 1609 Andalusite 1000 15 80 G* Andal, corundum 1629 Andalusite 1820 35 7 G Mullite, glass, trac corundum 1630 Sillimanite 1790 35 5 G Glass, corundum 1672 Sillimanite 1700 50 13 G Kyanite, corundum, coesite (?), glass 1673 Sillimanite 1670 25 20 G Glass, corundum *G = Graphite TABLE IV Results of Experiments with Gel Mixtures and Others Run No. Starting Material Temp (°C) Pressure (atm) Time (min) Heater Phases Present 1613 Gel 1000 20 120 G* Quartz, corundum 1614 Gel >1000 <1300 20 180 Ni Kyanite, corundum 1620 Gel 1025 25 180 Ni Quartz, corundum 1623 Gel 1900 30 2 G Corundum, glass 1650 Gel + H20 1300 35 60 G Corundum, glass 1658 Gel 1800 1500 50 50 5 30 G Kyanite, corundum, glass 1660 Gel + H20 1900 50 15 G Corundum, glass 1664 Gel 1300 25 40 Ni Corundum, glass 1669 Gel 1000 20 300 Ni Corundum, glass 1667 Kaolin and Gibbt 1700 35 15 G Corundum, glass 1668 Kaolin and Gibbt 1700 20 15 G Corundum, glass *G = Graphite tGibb = Gibbsite - Fig. 3 Decomposition of kyanite as function of temperature and pressure (heavy dashed line). The solid line and the light dashed line are taken from Clark et al. (1) To the right of the curve above about 20, 000 atm, there is evidence that the equilibrium assemblage may be corundum plus glass. Kyanite is stable to the left of the curve. The rectangle represents the precision of the temperature and pressure measurement. K = kyanite; S = sillimanite; A = andalusite; C = corundum; M = mullite; G1 = glass; Cs = coesite. -10- KYANITE Fig. 4 Plot of runs listed in Table III. Key as in Fig. 1: G = gel, Q = quartz, KG = kaolinite plus gibbsite. >11- When kyanite decomposed, the first evidence usually was the formation of glass and small crystals of AI2O3 from the fine-grained material at the grain boundaries [Fig. 5(a)]. The larger crystals of kyanite break down to form mullite* and glass [Figs. 5(a), (b)] and then at longer times or at higher temperatures the increased formation of AI2O3 is seen [Fig. 5(c)]. Although more data are needed for a firm conclusion, the tentative interpretation is that the final products from the decomposition of kyanite at high temperatures and pressures (above about 20,000 atm) are A1203 (corundum) and Si02-rich glass, but the reaction path may involve the formation of mullite before the final stage is reached. (a) (b) Fig. 5 Polished and etched sections as seen in reflected light, bright field. HF etchants; C = corundum; G1 = glass; M = mullite; K = kyanite. 500X (a) Run No. 1648. Corundum and glass at original kyanite grain boundaries; mullite and glass within original kyanite grains. (b) Run No. 1649. Remnants of kyanite surrounded by mullite plus glass. (c) Run No. 1624. Large mullite crystals (light grey) in glass (darker grey). Corundum crystals (white) in mullite. (d) Run No. 1630. Corundum and glass formed from sillimanite. *Below about 25, 000 atm both indices of refraction and the x-ray patterns are consistent with those reported for mullite. There may be some increase in refractive indices at higher pressures. - (c) (d) Fig. 5 Andalusite appears to decompose in the same way as kyanite, but sillimanite transforms very quickly to AI2O3 plus glass and the intermediate mullite stage has not been seen [Fig. 5(d)]. The behavior of sillimanite under these conditions seems qualitatively consistent with the results with kyanite after the mullite stage is reached in large crystals. It would be expected that mullite and sillimanite would decompose in a similar way because of their structural similarity. From the point of view of the mechanism of the reactions, it would be interesting to see if the results can be explained in terms of the respective structures and the differences in the aluminum ion coordination characteristic of each. From the point of view of phase equilibria, it is possible that the formation of corundum plus glass from the nominal Al2C>3:SiC)2 composition can be considered in terms of a shift of the corundum plus liquid region of the Al203-Si02 system toward the silica end of the system with increasing pressure. An extrapolation of - the preliminary results suggests that the Al203-SiC>2 phase diagram at high pressure would show a large region of corundum plus liquid with kyanite melting incongruently into this region. The decomposition curve of kyanite as presented here explained in part the predominance of corundum plus glass in most of the runs that have been made with gels and other starting materials in the present study. The P-T curves of Clark et al. Cl) are included in Fig. 1 to illustrate the obvious disagreement between their results and the present study insofar as the stability of kyanite is concerned. If one disregards the nature of the reactions taking place and considers only the stability region of kyanite, it would be necessary to lower the curve up to 25, 000 atm in order to be in agreement with the previous work or an extrapolation of it. KYANITE X-RAY PATTERN The differences between peak intensities in the x-ray spectrometer patterns of synthetic and natural kyanite (mentioned in second quarterly report) have been shown to be due to preferred orientation of the latter. If finely ground natural kyanite is sprinkled onto a slide covered with a thin layer of grease, the resulting x-ray spectrometer pattern is essentially identical in all respects to that of synthetic kyanite. This technique reduces the amount of preferred orientation when mounting the powder of a mineral with well-developed cleavage on a flat surface via a liquid medium. Random orientation is apparently achieved in the synthetic material because of the fine grain size and the inter grown arrangement of the crystals. PROGRAM FOR THE NEXT QUARTER During the final quarter it is proposed to (1) complete the determination of the P-T curves of the A^SiOs polymorphs from both the synthetic and decomposition directions; (2) to try to obtain some data on the kinetics of the transformations; and (3) to see if the mechanisms of the transformations can be treated from the available structural data. ACKNOWLEDGMENT Mrs. M. Houle prepared the polished sections and photomicrographs. - DISTRIBUTION LIST Contract No. AF-33(616)-6295 Progress Report No. 3 1. Director Naval Research Laboratory-Washington 25, D.C. 2. Commanding Officer Watertown Arsenal ATTN: Library Watertown, Massachusetts 3. Armour Research Foundation of Illinois Institute of Technology ATTN: D.J. MacPherson Technology Center Chicago 16, Illinois 4. Battelle Memorial Institute ATTN: Dr. R.I. Jaffee 505 King Avenue Columbus, Ohio 5. University of Denver Denver Research Institute ATTN: Mr. J. P. Blackledge Denver 10, Colorado 6. New York University College of Engineering ATTN: Dr. Harold Margolin Research Division University Heights New York 53, New York 7. Manufacturing Laboratories, Inc. ATTN: Mr. P. Kulin, President 249 Fifth Street Cambridge, Massachusetts 8. Office of Naval Research Building T-3 ATTN: Mr. J. Harwood (ONR-.423) Metallurgy Branch 17th and Constitution Avenue Washington 25, D.C. -2- 9. Nucor Research, Inc.„ ATTN: Dr. Bobrowski 2 Richwood Place Denville, New Jersey 10. Harvard University Department of Physics ATTN: Dr. Birch Cambridge, Massachusetts 11. Engineering Supervision Company 'ATTN: Mr. A. Zeitlin 787 United Nations Plaza New York 17, New York 12. University of California Institute of Geophysics ATTN: Dr. G. C. Kennedy Los Angeles 24, California 13. Linde - Division of Union Carbide ATTN: Dr. Ever sole Research Division Tonawanda, New York 14. Brigham Young University Department of Physics ATTN: Dr. H.T. Hall Provo, Utah 15. Director of Research and Development Headquarters, USAF ATTN: AFDR-RD Colonel J. V. Hearn, Jr. Washington 25, D.C. 16. Oak Ridge National Laboratory ATTN: Mr. W. Manly Post Office Box P Oak Ridge, Tennessee 17. Horizons, Incorporated ATTN: Dr„ Burwell 2891-2905 East 19th Street Cleveland 4, Ohio 18. Massachusetts Institute of Technology-Division of Industrial Cooperation ATTN: Professor N.J. Grant Cambridge 39, Massachusetts 19. University of Michigan Engineering Research Institute ATTN: Professor Sinnott or Dr. Freeman Ann Arbor, Michigan 20. Stanford Research Institute ATTN: R. Thielemann Metallurgical Department Menlo Park, California 21. The Wyman Gordon Company ATTN: Chief Metallurgist Worcester, Massachusetts 22. The Dow Chemical Company ATTN: Dr. T.E. Leontis Assistant to the Director Metallurgical Laboratory Midland, Michigan 23. Office of the Assistant Secretary of Defense (R and E) ATTN: Mr. Frank Huddle The Pentagon Washington 25, D.C. 24. California Institute of Technology Department of Mechanical Engineering ATTN: Professor P. Duwez Pasadena, California 25. Harvard University Division of Engineering and Applied Physics ATTN: Professor H. Brooks Cambridge, Massachusetts 26. University of Oklahoma Department of Physics ATTN: Professor S.E. Babb, Jr. Norman, Oklahoma -4- 27. Mallory-Sharon Metals Corporation ATTN: Mr. Stan Abkowitz Manager of Development Niles, Ohio 28. Boeing Airplane Company Seattle Division ATTN: Mr. J.W. Sweet Chief Metallurgist Seattle, Washington 29. Defense Metals Information Center Battelle Memorial Institute ATTN: C.B. Voldrich Columbus 1, Ohio 30. Monsanto Chemical Company ATTN: Dr. I. B. Johns Everett, Massachusetts 31. University of Utah Department of Physics ATTN: Dr. Peter Gibbs Salt Lake City, Utah 32. Iowa State University Department of Physics ATTN: Dr. C.A. Swenson Ames, Iowa 33. Air Reduction Company ATTN: Mr. Eric Gregory Central Research Laboratories Murray Hill, New Jersey 34. Owens-Illinois Technical Center ATTN: Mr. B. L. Steierman 1700 North Westwood Toledo, Ohio 35. Sandia Corporation Applied Research Division ATTN: Mr. J.D. Dudley Sandia Base Albuquerque, New Mexico