F-\. m: Thin mmmmm. 5' 1.131411 3’ Michiga n Sm, 93 Unvfrbity THESIS 0‘1 9 —- . ~ g ABSTRACT ELECTRON MICROSCOPIC STUDY OF LNWELLAR EUTECTOID IN A HYPO-EUTECTOID ALUMINUM BRONZE By C. P. Kamath This work is an investigation of the interface between the i“ or fill-phase and the eutectoid lamellae in a hypo-eutectoid aluminum bronze. It has been found under the electron microscope that both C! and )22 lamellae touch the figl-phase unlike the small gap of ferrite present between the cementite lamellae and the austenite interface in steels as found by Darken and Fisher. The ftl-rosettes seen under the optical microscope are found under the electron microscope to be clusters of oval shaped masses abutting each other. Again, the dark field, seen under the optical microsc0pe, in between the E3l-rosettes is found under the electron microsc0pe to be miniature rosettes, perhaps of fi%l, just after nucleation and growth. ELECTRON MICROSCOPIC STUDY OF LAMELLAR EUTECTOID IN A HYPO-EUI CTOID ALUMINUM BRONZE By C. P. Kamath A THESIS Submitted To Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Metallurgy, Mechanics and Materials Science 196A ACKNOWLEDGEMENT The author is indeed very much indebted to Dr. A. J. Smith for his valuable suggestions, guidance and also for affording all the faci- lities of the Metallurgy Department for the completion of this work. He is also grateful to Dr. D. E. Scherpereel for his valuable guidance in the electron-micrograph work and having taken the electron micro- photographs. Finally he wishes to express his gratitude to the United States Agency for International DevelOpment (U.S., A.I.D.) and the Government of Maharashtra for having deputed him for studies and re- search at Michigan State University. ii II. III. IV. VI. TABLE OF CONTENTS Introduction . Previous Work and Background . . Scope of Investigation . Materials and Methods . Results and Observations Discussion and Conclusions References . . . . . . iii Page 11 1h 17 3O LIST OF FIGURES Copper-rich portion of the copper-aluminum equilibrium diagram . Fully annealed sample. 250 X. Fully annealed sample. 1200 X. . Sample cooled at 2OC/min. between and quenched at 5000C. 250 X . . Sample cooled at ECO/min. between and quenched at 50000. 1200 X . Electron micrograph of the sample Electron micrograph of the sample Electron micrograph of the sample another spot. 15000 X. . . . . . Electron micrograph of the sample another spot. 70000 X. . . . . . iv 6500 and 50000 in Fig. A. in Fig. A. in Fig. A at in Fig. A at 6500 and 50000 6000 X 28000 X Page . 21 22 23 2h 25 26 27 28 29 I. INTRODUCTION Aluminum Bronzes i.e. the copper-rich alloys of the Cu-Al system have been the subject of study ever since H. C. H. Carpenter and C. A. Edwards(l) brought their useful properties before the engineering world. These alloys have tensile strengths ranging from 60,000 to 110,000 psi after different heat treatments. Elongation may vary from 5 to 30% and Brinell hardness number from 120 to 250. These alloys can be rolled or forged and, with some difficulty, are machinable. Their most important and useful property is resistance to corrosion, especially marine corro- sion. Though many investigators carried out work mainly on mechanical pro- perties after quenching and tempering heat treatments, very few have worked on the morphology of the structures of these alloys. The c0pper-rich portion of the Cu-Al system is shown in Fig. 1‘2) The eutectoid decomposition of the [3 - phase of these alloys is in many respects similar to those of steels both in equilibrium cooled, and quenched conditions, and this fact was recognized at an early date even though the kinetics of decomposition in both the systems have yet to be fully understood. The decomposition of'fl3- phase is somewhat simpler to experiment with than in steels because of lower temperatures involved and also the more sluggish diffusion of the substitutional aluminum atoms than the interstitial carbon atoms in steels. These ‘alloys are amongst some of the non-ferrous alloys used for studying the martensitic-type and other transformations, and carrying the results to experiment with and understand martensitic and other transformations in steels, since most of the non-ferrous alloys can be studied at lower temperatures and hence it is easier to carry out experimental work. -2- There are some differences in the crystal lattices of these alloys and those of steels. For example, Austenite in steels is F.C.C. where- as, the corresponding phase, [3 in.aluminum bronzes is B.C.C. In the decomposition of this 3 - phase by quenching, we get meta-stable transi- tion constituent or constituents, {31 and/or [3" * and hence some com- plexity of transformation is met with; whereas in the case of steels no such meta-stable transitional constituents have been found as yet and, perhaps, may not be existing. In the decomposition of austenite of eutectoid composition to its constituents ferrite plus cementite (pearlite), there is controversy as to which of the above two constituents, ferrite or cementite is the initial nucleating agent. Most of the investigators are in favour of cementite to be the nucleating agent of pearlite, but there are others who support ferrite to be the one nucleating first, at least in hypo- eutectoid steels. The compromising formulas have also been propounded. Recently L. S. Darken and R. M. Fisher(u) have used the electron microscope to get much higher magnifications than it is possible in Optical microscopes, to get better resolution of the tri-junction of austenite, ferrite and cementite. Their finding of a small gap of, perhaps, ferrite between the austenite interface and the cementite lamellae is an important contribution towards the understanding of pear— lite formation. Since aluminum bronzes also show a lamellar eutectoid and other ** types of eutectoids, which are called by some authors pearlite , con- H There is controversy regarding the existence ofl3 (3). ** Now-a-days, the term pearlite is used in systems other than iron- carbon system to denote eutectoid transformation products in metal- lographically analogous structures. -3- * sisting of Czcu and 7 2 constituents in alternate layers as a result of the decomposition of (3- phase, the electron microscope can be used to study the morphology of it. The study in this work has been carried out by this method. Before l9h8 7/2 was designated by the symbol <5 . II. PREVIOUS WORK AND BACKGROUND (1) Early investigators like H. C. H. Carpenter and C. A. Edwards , (5) (6) (7) (8), B. E. Curry , J. M. Greenwood , P. Braesco , J. Bouldoires (9’10) , C. Grard and that too it forms in smaller quantities. Slow cooling is necessary for their proper formation. Large amounts of acicular and cellular eutec- toids along with thick alternate irregular layers of <1 and 7'2 are formed. Outside this composition range the eutectoid is mostly cellu- lar. It is difficult to get lamellar eutectoid from a phase having considerable concentration gradients<20). The farther the over-all composition of the alloy is from the eutectoid, the slower should be the cooling rate to get a lamellar structure. Therefore, the slower cooling method worked by C. S. Smith and.W. (19) E. Lindlief has been used to get lamellar eutectoid. Isothermal transformation technique has not been used. Different cooling rates -9- -10- were tried so as to get an interface between.,3 and [31 and the lamellar eutectoid. Both optical microscopic and electron microscopic structures have been studied for comparison. IV. MATERIALS AND METHODS The hypo-eutectoid aluminum bronze used in this study was prepared by melting high purity copper and aluminum in a plumbago crucible under graphite in an electric resistance furnace. It was cast into an ingot, 7" x 1%" x l", in a metal mould, the pouring temperature being about 125000. To get homogenous alloy, the case ingot was remelted at 12500C under similar conditions as given above and chill cast. fHuecast ingot was homogenized at 90000 for 15 hours and next rolled from 1" to 5" thickness with intermediate reheatings to 9000C. After the rolling Operation was over, the ingot was furnace cooled from 9000C. It analyzed: Element .;fi_ Copper 88.10 Aluminum 11.70 Iron 0.10 *Silicon 0411 *Manganese 0.005 *Nickel 0.005 *Tin 0.005 *Zinc 0.005 *Lead 0.005 * Semi-qualitative Sample pieces, 1" x 5/8" x 1/2" were cut from the rolled ingot for heat treatment at different rates of cooling. Sample I was heated to 8000C for two hours, after which it was slowly cooled in the furnace to 6500C in about four hours. The cool- ing, through the critical range, was done at a uniform rate of about (19) 2OC/min. as suggested by C. s. Smith and w. E. Lindlief in order to get rosettes of‘{3l and the lamellar eutectoid growing in them. -11- -12- Therefore, from 6500 to 5000C the cooling was done more or less uniform- ly in about one hour and fifteen minutes. At 5000C, the sample was quenched in water. The quenched specimen was cut into two equal halves and the newly cut surface of one of the pieces was filed, first polished with emery and finally with fine alumina on the polishing wheel. Etching was done with Buhler electrolytic etching apparatus for three seconds, using 1% aqueous chromic acid solution as etching reagent as suggested by W. C. Coons and D. J. B1ickwede<3l). Three more samples were subjected to heat treatments similar to. that given to Sample I, the difference being the cooling rates through the critical range were 100, %0C and %OC per min. respectively. These samples also, after heat-treatment, were polished and etched like Sam- ple I. Since, Sample I gave the desired results, it was taken up for study. There were a few good junctions between {3, fi3l and a few patches of lamellar eutectoid. Other samples did not show the presence of {3 or £31 and hence were not taken for further investigation. Sample I was repolished, lightly etched (this etching was not satisfactory for optical microscope work) and a negative plastic replica was taken using collodin solution supplied by Hitachi Co. along with their 11A electron microscope. The replica was taken on 200 mesh nickel grids and covered with a drOp of polystyrene latex solution in water for shadow contrast. The latex solution was dried in a vacuum desiccator. The dried replica was platinum shadowed in a high vacuum unit. The replica was next studied under the electron-microscope and photographs were taken at 6,000 and 15,000 magnifications. The nega- tives were enlarged about h.75 times to get high magnification prints. -13- Optical microscope photOgraphs of Sample I were taken. Sample V is a piece of the rolled ingot which was heated to 80000 and slowly furnace cooled in order to have a fully annealed structure for comparison. This sample was also polished and etched with 1% aqueous chromic acid solution as etching reagent. Optical photograph of the structure obtained was taken. V. RESULTS AND OBSERVATIONS The aluminum bronze used contained about 11.70% aluminum and hence, the composition is fairly close to eutectoid composition on the hypo- eutectoid side. The homogenized and fully annealed sample showed little pro-eutectoid (2 areas dispersed throughout the sample and their sizes varied. Those which were found in cellular eutectoid regions were smaller in size and more numerous, whereas, acicular eutectoid regions showed slightly larger in size but fewer in numbers. The matrix was mostly consisted of acicular eutectoid, cellular eutectoid and regions of thick alternate layers of (1 andy2 (see Figs. 2 and 3). Sample I gave the desired results, i.e., (3 or B1 interface and lamellar eutectoid junctions. Most of the eutectoid was either cellu- lar or acicular. Only in a few regions, lamellar eutectoid could be found either abutting acicular eutectoid or within the regions of it (see Figs. h and 5). The locations of the lamellar eutectoid were fewer and far between. Out of these fewer locations also, the clear junctions between the {3 or [31 and lamellar eutectoid were meagre. No Widmanstatten structure was found. Samples II to IV did not show any [3 or {31 and hence, were not investigated further. The electron micrographs of Sample I are shown in Figures 6 to 9. Even at 70,000 X, the C9. and 7 2 lamellae are touching the (31 interface and look continuous with it, unlike the result found by L. S. Darken and R. M. Fisher in steels. So. the kinetics of formation of lamellar eutectoid, is perhaps, different from that of lamellar pearlite in steels. E. P. Kliers and S. M. Grymko(17) have stated on the basis of finding inter—lamellar space first decreasing with -lh- _15- decrease in temperature of formation and later on increasing with further decrease in formation temperature, that the kinetics of formation of lamellar eutectoid in aluminum bronze is different from that of pearlite formation in steels. The electron-micrograph finding in the present study is a further proof of their statement. It is probable that both <2 and.‘7é are forming at the same time. Their growth in £31 is more or less well defined. (20) R. Haynes' work on ll.h% aluminum alloy does not show any for- mation of rosettes of (31. In the present work beautiful rosettes of {31 have been obtained (see Figs. 2 and 3). In the optical microscope, at higher magnifications, in a deeply etched structure Of Sample I, some faint irregular structure was visible in the etched portion between the rosettes of 81' In the electron micrographs, the rosettes are seen to be, actually, roughly oval shaped masses abutting each other in clusters (see Fig. A and 6) and this fact could not be seen clearly in the optical micro- graphs. The region in between the rosettes (mentioned earlier about the seeing of faint structure under the optical microscope) showed clearly in the electron micrograph to be miniature rosette—like structures (see Fig. 6) which at 70,000 X (See Fig. 9) is clearly indicated to consist of irregular shaped curved masses, perhaps, consisting of oval shaped masses closely abutting-each other (not very clear in Fig. 9). It is possible that these miniature masses are actually £31 just after mucleation and growth. There is indication in Fig. 6 that these minia- ture masses coalesce and form quite large oval shaped masses which cluster together and are seen as rosettes under the Optical microsc0pe. -16_ The large difference in sizes between the big £31 rosettes and the minia- ture rosettes and the numerous indistinct etching lines amongst the latter makes them look darker as a whole under the Optical micrOSCOpe when compared.with the bigger F31 rosettes. This is not the case in the electron micrograph and the etching lines are clearly seen because of better resolution. The mass behind the interface is seen to be a continuous one, like a band, and is B 1 (see Fig. 9). But in Fig. u this fact cannot be realized and it looks as if lamellar pearlite is forming indiscriminately both in rosette E31 and the so-called fig(19) (dark etched areas) which is actually £31 under formation. On close examination, it is found that the interface band does not have etching lines crossing it. The minia- ture rosettes have all coalesced just before the interface. The complete envelOping of proeutectoid C1 by rosettes of E51 has been observed. The very dark round spots on the electron micrographs are not a part of the structure but they are globules (about 0.26 microns in dia.) of polystyrene latex, dispersed over the replica to give shadow contrast, magnified. VI. DISCUSSION AND CONCLUSIONS A. Discussion: The similarity between the /3 - phase decomposition in aluminum bronzes and the austenite in Fe-C system has been referred to in Sec- tions II and III. Both 8 -phase and austenite are solid solutions having eutectoid phase transformation in them. Austenites of hypo-eutectoid compositions, under equilibrium conditions, decompose into pro—eutectoid ferrite and pearlite consis- ting of eutectoid ferrite and eutectoid cementite. The same is the case with.[3-phases of hypo-eutectoid compositions under equilibrium conditions. They decompose into proeutectoid <1 and eutectoid consis- ting of eutectoid cn and eutectoid.y 2. Under non-equilibrium condi- tions, i.e., under different rates of cooling, in Fe-C system, marten- site, bainite and pearlite are met with, whereas in aluminum bronzes, we come across 8' and.y" marten-ites Egl-rosettes, acicular, cellular and lamellar eutectoids. The most recent divergent views on pearlite formation in Fe-C system are those of M. Hillert (29) on one side and L. S. Darken and R. M. Fisher<3o> on the other and this has already been referred to at the end of Section II. Hillert argued that the ferrite gap found by Darken and Fisher is due to the rate of quenching which affects the s/so ratio (where‘sflis the actual spacing and s0 is the minimum spacing between the lamellae) giving rise to gap; and he gave in support of his argument certain micrographs, wherin with progressive increased rate of cooling, the development of the gap between the cementite lamellae and the austenite interface was shown. Fisher was not satisfied with Hillert's arguments and proposed further experiments to clarify at -17- -18- least some of the points which Hillert had raised against their paper. However, as Hillert said, the suggestion by C. Benedicks<32) that both ferrite and cementite could act as germs for formation of pearlite, cannot be lightly brushed aside. Though, Hillert's theory of active partnership between ferrite and cementite in lamellar pearlite formation is logical, it has yet to be put to rigorous tests. In-the case of aluminum bronzes enormous work, as in the case of steels, has not been done. The nucleus for the formation of lamellar eutectoid is still to be understood, just like lamellar pearlite in the Fe-C system. The characterisoic intervening £3lwphase complicates a little the understanding of the transformation. The assumption by earlier workers of changes in the composition of a_?phase, mostly on the super-saturation side, to account for the decomposition mechanism of the fig-phase, under isothermal transformation, has been proved to be incorrect by R. Haynes(33) on the basis of finding, by x-ray method, that the lattice parameter of 0L did not change during transformation between 560 and hoooc and was in equilibrium with Y 2 (just as ferrite composition is almost uniform in the Fe-C system during transformation). 0n the whole both aluminum bronze and the Fe-C systems are close in their transformation mechanisms and morphology. The non-finding Of the gap between the Cl and 7’2 lamellae and the E>l interface in aluminum bronzes is more in line with that part of Hillert's theory in which he states that there is active partnership between ferrite and cementite in the lamellar pearlite formation, than those of Darken and Fisher, if the analogy between the two systems is accepted and the difference between the experimental methods followed by Darken and Fisher (iso- thermal transformation) and in the present study (predetermined cooling rate up to a particular temperature followed by quenching) are taken -19- into consideration. From a close examination of Fig. 7, it is clear that the [31 interface is continuous throughout both where the Cl and 7/2 lamellae are touching it and this means that the transformation is continuous at the interface. B. Conclusions: Just as in eutectoid and hyper-eutectoid aluminum bronzes, in hypo-eutectoid ones also rosette shaped £31 phase is met with. The dark matrix seen under the optical microscope between the rosettes of £31 is found under electron microsc0pe to be made up of miniature rosettes, which, perhaps, are also £31 phase. The difference seen under the optical microscope between the dark matrix and the bright colored rosettes of figl is due to the large difference in the sizes of the two types of rosettes and the numerous etched boundaries of the miniature rosettes. The big rosettes are clusters of roughly oval shaped masses abut- ting each other in groups, the abutting lines being not clear under the optical microscope. These have grown out of the miniature rosette masses coalescing together. The miniature rosettes also consist of roughly rounded masses. The phase next to the interface where the lamellar eutectoid is growing into, other than big rosettes of E31, is a continuous band of coalesced miniature rosettes. Very little lamellar eutectoid is met with and that too, they are found inside the regions of acicular eutectoid or at the junctions of acicular eutectoid and the untransformed fiBl. The lamellae of both CL and y’2 touch the interface of [31 (70,000 X). It is clear from Fig. 7 that the flgl interface is continuous -20- throughout, both where the Cl and 7’2 lamellae are touching it. 1 1050 1000 IEIPERATURE,°C 100 950 900 O 0' 0 § 760 700 650 600 550 Z 3 4 10 WEIGHT PER CENT ALUMINUH 5 6 7 8 9 1O 12 . 14 16 18 20 22 24 26 10370 160 1.0 ‘93 (7.5) (8.5) (9-5) I _-.—. 15 20 25 30 35 40 45 ATOMIC PER CENT ALUMINUM Copper—rich portion of the copperaaluminum equilibrium diagram. 28 -22- ig. 2 Fully annealed sample. 250 X. -23- Fully annealed sample. 1200 X. Fig. 3 -2h- 250 X. Sample cooled at 2OC/min. between 6500 and 5000C and quenched at 5000C. Fig. A Fig. 5 Sample cooled at 2OC/min. between 6500 and 50000 and quenched at 50000. 1200 x. _26- Fig. 6 Electron micrograph of the sample in Fig. A. 6000 X. (Plastic replica) -27- Fig. 7 Electron micrograph of the sample in Fig. a. 28,000 x (Plastic replica) -28- Fig. 8 Electron micrograph of the sample of Fig. 6 in a different spot. 15,000 X (Plastic replica) Fig. 9 Electron micrograph of the sample of Fig. 6 in a different spot. 70,000 X (Plastic replica) .00.---_,. ._, _ “WW—,- .-'A‘A l. O‘xm “'1 IO. ll. 12. 13. 1h. 15. 16. 17. 18. 19. 20. 21. VII. REFERENCES H. C. H. Carpenter and C. A. Edwards: Proc. Inst. Mech. Engr., (1907) p. 57. M. Hansen: ”Constitution of Binary Alloys" (1958), p. 86, McGraw- Hill Book Co., New York. C. W. Spencer and D. J. Mack: "Decomposition of Austenite by Dif- fusional Processes" (1962), AIME, p. 592, Inter Science Publishers, .a New York. E“? L. S. Darken and R. M. Fisher: Ibid, pp. 2A9-288. I A B. E. Curry: Journal Phys. Chem. (1907), Vol. 11, p. A23. J. M. Greenwood: Journal Inst. Metals (1918), Vol. 19, p. 55. ‘j P. Braesco: Ann. Phy. /9/ (1920), p. 5, Vol. 1h. J. Bouldoires: Rev. de. Met (1927), Vol. 2h, pp. 357 & A63. M. Matsuda: Sci. Repts. Tohoku Imp. Univ. (1922), Vol. 11, p. 23h. M. Matsuda: Journal Inst. Metals (1928), Vol. 39, p. 67. C. Grard: Aluminum and its alloys, Part V, New York (1922), D. Van Nostrand Co. I. Obinata: Mem. Ryojun College of Engineering (1928—30), Vol. 2, p. 205; (1929-31) Vol. 3, pp. 87, 285, 295. D. J. Mack: Trans. AIME (19u8), Vol. 175, p. 2A9. G. Wasserman: Metallwirtschaft (193A), Vol. 13, pp. 133-139. A. B. Greeninger: Trans. AIME (1939), Vol. 133, p. 20u-227. E. Kaminsky, G. Kurdjumow and w. Newmarx: Metallwirtschaft (193E), Vol. 13, p. 373. E. P. Klier and S. M. Grymko: Trans. AIME (19A9), Vol. 185, p. 611. E. S. Davenport and E. C. Bain: Trans. AIME (1930), Vol. 90, p. 117. C. S. Smith and W. E. Lindlief: Trans. AIME (1933), Vol. 10A, pp. 69-115. R. Haynes: Journal of the Inst. of Metals, (l953-5h) Vol. 82, p. A93. R. Haynes: Journal of the Inst. of Metals (l95h-55), Vol. 83, p. 105. -30- 22. 23. 2h. 25. 26. 27. 28. 29. 30. 31. 32. 33- -31- Metals Handbook (19A8), p. 1160. F. C. Hull and R. F. Mehl: Trans. ASM (1962), Vol. 30, pp. 381-A2u. C. S. Smith: Trans. ASM (1953), Vol. E5, p. 533. S. Modin: Jernkontorets Ann. (1951), Vol. 135, p. 169. M. E. Nicholson: Journal Metals (195A), Vol. - , p. 1071. A. Dube: Doctoral Thesis, Carnegie Inst. of Tech., Pittsburgh, (19A8). H. I. Aaronson: Doctoral Thesis, Carnegie Inst. of Tech., Pitts- burgh, 195A. M. Hillert: "Decomposition of Austenite by Diffusional Processes" AIME (1962), pp. 197-2A7, Inter Science Publishers. A. D. Hopkins: Journal Inst. Metals (195A), Vol. 82, p. 163. w. C. Coons and D. J. Blickwede: Trans. ASM (l9u5), Vol. 35, pp. 28h-297. C. Benedicks: Journal Iron Steel Inst. (london), Vol. 2 (1905), p. 352. R. Haynes: Trans. AIME (1958), Vol. 212, p. 5. MlCHiGAN STATE UNIVER IT I II I H III | 1|! mm... 1 225 116 3 1293 O3