ENGINEERING RESEARCH INSTITUTE UNIVERSITY OF MICHIGAN ANN ARBOR DEVELOPMENT OF PROCEDURES FOR TEE IDENTIFICATION OF MINOR PHASES IN BEAT-RESISTANT ALLOYS BY ELECTRON DIFFRACTION j, By 5 L. 0. BROCKWAY Professor ef Chemistry and W. C. BIGELOW Research Associate'PROGRESS REPORT NO. 7 for the period 15 July to 15 October 1953 PROJECT NO. 2020 CONTRACT NO. AF-33(616)-23 EXPENDITURE ORDER NO. R463 Br-1 To AERONAUTICAL RESEARCH LABORATORY (WCRRL) RESEARCH DIVISION WRIGHT AIR DEVELOPMENT CENTER WRIGHT-PATTERSON AIR FORCE BASE, OHIO

-C. f \ r Ij1 v

ACKNOWLEDGEMENTS This project is coordinated with the program on Heat-Resistant Alloys which is in progress under the direction of Professor J. W. Freeman of the Department of Chemical and Metallurgical Engineering of this University. Professor Freeman and his associates have supplied all the alloys studied here and have also provided extensive metallurgical data on these alloys to aid in the interpretation of the diffraction results. The electron-microscope studies of this report were carried out with the electron microscope of the School of Public Health through the courtesy of Professor Thomas Francis and Dr. R. E. Hartman. The experimental work on the N-155 and Inconel-X alloys reported here was performed by Mr. J. A. Amy and Miss Rosemary Jacobson. ii

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ii SUMMARY iv INTRODUCTION 1 RESULTS AND CONCLUSIONS 2 Replica Techniques for Electron Microscopy 2 Inconel-X Alloy 9 N-155 Alloy 14 Mechanical Polishing Procedures 19 X-ray Diffraction Studies of Carbides of 16-25-6 Alloy 21 DISCUSSION 23 BIBLIOGRAPHY 25 iii

SUMMARY Using the electron diffraction method, minor phases have been identified in samples of Inconel-X alloy representing aging treatments of 100 and 1000 hours at 1400~F and 10, 100, and 1000 hours at 1600~F. Columbium carbonitride and M23C6-type carbide have been identified in each of the samples aged at 1400~F; however, only the columbium carbonitride has been identified in the samples aged at 16000F. These results suggest a strong influence of aging temperature on minor-phase formation; the investigation is therefore being extended to include samples aged at 12000F. Studies have also been made of the minor phases of low-carbon N-155 alloy using samples aged 10, 100, and 1000 hours at 1400~F and 10 and 1000 hours at 16000F. Patterns corresponding closely to an intermetallic compound, Fe2W, have been obtained from the samples aged 1000 hours at 1400~F and 16000F. Columbium carbonitride and an M23C6 carbide have been identified in the samples aged 10 and 100 hours at 1400oF. These results are discussed in terms of the prior heat treatment of the samples. Mechanical polishing procedures were used instead of the usual electrolytic polishing in preparing the N-155 alloy samples for the electron diffraction examination. The results indicate that satisfactory surface preparations may be obtained in this manner and are discussed in terms of the possible advantages and disadvantages of using mechanical polishing. In addition, positive- and negative-plastic-replica techniques have been compared to determine the most convenient replica methods for obtaining electron micrographs for use in connection with the electron diffraction studies. Carbides have been separated from several samples of 16-25-6 alloy and examined by x-ray diffraction to compare the results obtained by the electron diffraction and x-ray diffraction methods of minor-phase identification. iv

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN DEVELOPMENT OF PROCEDURES FOR THE IDENTIFICATION OF MINOR PHASES IN HEAT-RESISTANT ALLOYS BY ELECTRON DIFFRACTION INTRODUCTION Recent metallurgical research has shown that the high-temperature properties of heat-resistant alloys are closely related to the minor phases which form in these alloys during thermal and mechanical processes. The lack of suitable methods for detecting and identifying these phases has proven to be a major obstacle in determining their influence on the metallurgical properties of the alloys. This project was therefore undertaken to apply the highly sensitive electron diffraction method developed by Heidenreich, Sturkey, and Woodsl to this problem. There are two principal objectives of this project: first, the development of procedures for adapting the electron diffraction method to the heat-resistant alloys and, second, the application of the method to the study of the influence of high-temperature aging on the development of minor phases in some alloys typical of those currently used in the construction of jet aircraft engines. The adaptation of the electron diffraction method to the hea-tresistant alloys is largely a problem of finding suitable polishing, etching, and rinsing procedures for preparing the alloy specimen for the electron diffraction examination. These problems have been discussed in general terms in a previous report2 and a paper describing etching techniques for this particular application has been prepared for publication in a technical journal. In the application of the electron diffraction method to the heatresistant alloys, an extensive study has been made of the influence of hightemperature aging treatments on the development of carbides in 16-25-6 alloy. This work has been sumnarized in the annual report for the previous year2 and is now in preparation for publication. In addition, the method was used to corroborate results of x-ray diffraction studies of the minor phase of a sintered aluminum product3. Currently, studies of the minor pha.ses of N-155 and Inconel-X alloys are in progress.

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN The paper on etching techniques was submitted in lieu of a regular progress report for the preceding quarter; therefore this report describes the work carried out since the last progress report was submitted on 15 April 1953. Included in this work are: (1) further studies of the variation of the lattice parameter of the M6C phase of 16-25-6 alloy, using x-ray diffraction methods, (2) discussion of the relative merits of positiveand negative-replica methods for the electron microscope studies in conjunction with the electron diffraction method, (3) preliminary consideration of the use of mechanical polishing methods in the preparation of the alloy samples, and (4) further studies of the minor phases of N-155 and InconelX alloys. The details of the experimental procedures used in the electron diffraction identification of minor phases have been described previouslyl,2,4)5. Briefly, however, the alloy samples are polished and then etched, using reagents and conditions which preferentially attack the matrix metal and leave the minor phases protruding in relief from the surfaces. Electron diffraction patterns are then obtained from the protruding phases by passing an electron beam across the surfaces at a grazing angle. These patterns are analogous to x-ray diffraction patterns and may be used similarly to identify the phases. It is also advantageous to make concurrent optical and electron microscope studies of the etched surfaces to evaluate the effectiveness of the methods of surface preparation and to correlate the microstructures of the samples with the diffraction results. RESULTS AND CONCLUTSIONS Replica Teclmniques for Electron Microscpy There has recently been considerable discussion on the selection of replication procedures for electron microscope studies of alloy microstructures6)7 and particular attention has been directed to the relative merits of positive and negative plastic replicas8)9. These discussions have been concerned with studies in which the electron microscope is used primarily as a suibstitute for th'-e lighlt microscope to provide information on the form and distribution of microconstituents. Considerable emphasis:Las therefore been placed on th' similarity in the appearanrce of th;.e resulting micrographs and conventrional optical micrographs. In th.e present investigation, the electron microscope stuldies are particularly valuable for detecting traces of insoluble products formed b-y the et-ching procedures and for obtaining information on the effectiveness of t.he etching procedures in exposing the minor-phase particles in relief

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN with respect to the surrounding matrix metal. While a number of replication techniques have been developed for the study of metal surfaces4,5110oll, including the polystyrene-silica technique, various oxide-replica methods, and the positive- and negative-plastic-replica techniques referred to above, the two plastic-replica techniques are most convenient to use and are best suited to the requirements of the combined electron diffraction- electron microscope studies. In previous work on this project, only the negativeplastic-replica technique has been used; however, the positive-replica method offers some decided advantages for certain purposes. It is therefore important to examine these two methods in connection with the requirements of the electron diffraction studies. The procedures by which negative plastic replicas are prepared have been described in detail in the literature4'9 and in a previous report2 of this project. Briefly, a few drops of a dilute solution of Formvar (polyvinyl formal) in ethylene dichloride, or of collodion in amyl acetate, are placed on the etched surface and the solvent is allowed to evaporate, leaving a' thin plastic film on the surface. The film is separated from the surface, shadow-cast, and examined in the microscope. The separation of the replica from the etched surface may be accomplished by several methods 9v12v13, however the "dry-stripping' method of Schaefer 12 has proved adequate in this project, provided some care is exercised to avoid distortion of the replicas when surfaces are heavily etched. In this method a l/8-inch-diameter specimen screen is placed on the coated metal surface; water vapor is condensed on the surface to loosen the plastic; cellophane tape is pressed over the plastic and the screen and is carefully lifted to remove the plastic film from the metal. The screen is then separated from the tape and the portion of the film which it carries is shadow-cast and examined in the electron microscope. Replicas prepared in this manner reproduce the etched metal surface in negative relief; i.e., protrusions on the surface are reproduced as indentations on the replica and vice versao The preparation of positive plastic replicas by the method of Schwartz, Austin, and Weberl is a two-stage procedure involving a thick intermnediate replica from which the final thin replica is made. A few drops of a 15% aqueous solution of polyvinyl alcohol are evaporated on the surface to produce a thick, pliable negative replica of polyvinyl alcohol which can easily be removed by means of tweezers and a razor blade. The surface of this replica which was in contact with the metal specimen is then flooded with a dilute solution of Formvar in ethylene dichloride. When the solvent evaporates a thin Formvar replica is produced which is a direct reproduction of the original metal surface. The composite film is cut into pieces slightly larger than the specimen mounting screens and these are floated, with the polyvinyl alcohol side down, on clean water to dissolve the polyvinyl alcohol. The pieces of the positive Formvar replica

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN remain floating on the water. Specimen screens are then carefully placed on top of these floating pieces of replica, so that they can be lifted from the water with wet newsprint in the manner described by Wyckoffl4. The pieces of replica are thus mounted on specimen screens in a manner suitable for shadow-casting and subsequent examination in the electron microscope. It is possible to substitute Zapon lacquer for the polyvinyl alcohol for preparing the intermediate negative replica in the first stage of this procedurel0o in which case the Formvar positive replica is separated by one of the methods described by Fullaml5. Shadow-casting, which is used with both types of replicas, consists of evaporating a very thin layer of a heavy metal, such as palladium, on the surface of the replica at an angle of about 300 from a hot tungsten filament. Certain areas of the replica are shaded from the filament by irregularities in the surface of the replica and receive no metal. These areas are less opaque to electrons than surrounding areas where the metal was deposited; consequently the contrast of the replica is considerably enhanced. Furthermore, the shape, direction, and length of the shadows make it possible to obtain information concerning the relief of various features on the original metal surface. It is convenient to place polystyrene latex spheresl6 on the replicas just prior to shadow-casting, as these spheres are easily identified in the micrographs and aid in determining the angle and direction of shadowing. In addition they are highly uniform in diameterl7 (2580+30 A) and are useful in calibrating the magnifications of the micrographbs. Fig;ures 1 and 2 show schematically the essential features of these replica methods, including the shadow-casting operation and the appearance of the resulting micrographs, for typical features such as a protruding carbide particle, an etch pit, and a polystyrene latex sphere, and also illustrates the basic points which must be kept in mind in interpreting the micrographs. When negative replicas are used, protruding particles are reproduced as indentations in the replica surface and appear in the micrographs with shadows which fall inside the outlines of the particles. Conversely, etch pits are reproduced as projections from the replica surface and have external shadows. In many cases the thickness of the collodion in these projections is sufficient to render them opaque to electrons, so that they appear lighter than surrounding areas in the micrographs. With positive replicas the projections on the metal surface are reproduced as protrusions on the replicas and appear with external shadows which may be used to determine the height and shape of the projections if the shadow-casting angle is known. Indentations or pits in the metal are reproduced as pits in the replica and have internal shadows. With either type of replica, the polystrene latex spheres cast external shadows and, since they are opaque to electrons, appear white in the micrographs.

PARTICLE OF DIRT ON SURFACE- - / ETCHED METAL PROTRUDING CARBIDE SURFACE PARTICLE -J - - ETCH PIT- -------- DIRECTION OF METAL SHADOWING- -- \ METAL FILM —- THIN COLLODION METAL FILM -- - - —...;:. _.:.::.:REPLICA POLYSTYRENE / LATEX SPHERE - -. ~ APPEARANCE OF FEATURES IN ELECTRON MICROGRAPH / / / / / / SHADOWS - - -_ _-/_ _/ FIGURE I. SCHEMATIC REPRESENTATION OF NEGATIVE- REPLICA METHOD

PARTICLE OF DIRT ON SURFACE ETCHED METAL PROTRUDING CARBIDE SURFACE PARTICLE - ETCH PIT J.. THICK, INTERMEDIATE,.,,"'-' POLYVINYL ALCOHOL REPUCA DIRECTION OF METAL SHADOWING —. METAL FILM - THIN FORMVAR REPLICA *~~... *.* POLYSTYRENE LATEX / SPHERE. APPEARANCE OF FEATURES IN ELECTRON MICROGRAPH SHADOWS -- - J FIGURE 2. SCHEMATIC REPRESENTATION OF POSITIVE- REPLICA METHOD

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN In the evaluation of surface preparations where it is desired to determine the type of etching action obtained and the extent to which the minor-phase particles are in relief, consideration of the relation of the shadows to the outlines of the features in the micrographs as described above generally is adequate to determine whether the various features project or are recessed on the metal surface. However, the positive-replica method is particularly suited for determining the extent of the relief of the minor-phase particles, for with this method the protruding particles are reproduced by identical protrusions on the replica and the length of their shadows may be used to calculate the height of the particles on the surface. The only information required is the angle of shadow-casting, and this may be obtained by reference to the polystyrene latex spheres. In general, micrographs from negative replicas cannot be used in this manner, for the protruding particles are reproduced as irndentations in these replicas and the shadows fall inside the indentations. Since the slope of the sides of the indentations are generally not known, the length of the shadows cannot be related to the height of the particles. These considerations are illustrated in Figure 3 by the micrographs taken from positive and negative replicas of the etched surface of a sample of S816 alloy. In many cases, however, it is desirable to study the results obtained when etching procedures are used which produce deep pitting of the surfaces; here the negative-replica method provides more extensive information, as shown in Figure 4. The micrographs in this figure were taken from a sample of 16-25-6 alloy which was etched electrolytically with 10% nitric acid. This treatment dislodged the minor-phase particles and severely recessed the grain boundaries, giving a "knife-edge" attack of the type described by Kinzel. These features are most clearly shown by the micrograph from the negative replica. In connection with the electron diffraction studies it is very useful to be able to examine the etched surfaces for small amounts of foreign materials such as insoluble etching products, since traces of such materials may obscure the minor-phase particles or produce diffraction patterns which confuse their identification. Such an examination can often be effectively accomplished by electron microscopy using the negativereplica method. The foreign materials become embedded in the thin plastic replicas and are lifted from the surfaces and easily detected in the replicas by the electron microscope. Positive-replica techniques cannot be used for this purpose, however, for although the contaminants are picked up from the surfaces they remain embedded in the thick polyvinyl alcohol replicas and are not transferred to the thin Formvar replicas and therefore do not appear in the micrographs. The negative-replica method is simpler and considerably more rapid than the positive-replica method; usually a negative replica can be obtained in less than five minutes, while in the positive-replica procedure

ENGINEERING RESEARCH INSTITUTE ~ U*V'Ri'S i'r',, O[f MiCHIGAN a. Micrograph from Negative Replica b. Micrograph from Positive Replica Figure 3. Comparison of Electron Micrographs fromh Positive saz.d Negative Replicas of Alloy Surface with Protruding Carbj.de PaiL —icles. X8000)O a. Micrograph from Negative Replica b. Micrograph from Positive Replica Figure 4. Comparison of Electron Micrograph fron Posi-tivie a-nd Negati-ve Replicas of Alloy Surface Deeply Pitted by Etching. X 8010

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN several hours may be required to dry the polyvinyl alcohol replica and then to separate the polyvinyl alcohol and Formvar replicas. The positivereplica method is definitely superior for use with rough or heavily etched surfaces, however, for there is considerable danger of introducing strains or artifacts into the thin replicas of the negative-replica method when separating them from such surfaces. This is often an important consideration in the electron diffraction studies where heavy etching treatments are required to provide adequate exposure of the minor-phase particles. On the basis of these considerations it appears that the simplicity and convenience of the negative-replica method makes it well suited for preliminary evaluation of etching procedures and for routine use in connection with the electron diffraction studies; however, the more complete information provided by the positive-replica method makes it highly desirable for final correlations of the diffraction results with the microstructures of the alloys. All the micrographs in the following sections of this report are from negative collodion replicas shadowed with palladium. Inconel-X Allo Considerable difficulty has been encountered in the preparation of this alloy for the electron diffraction studies due to the formation of adherent films of insoluble products on the samples during the polishing and etching procedures. In previous work on this alloy,3 satisfactory electrolytic polishing procedures were developed, using a solution of 1/3 perchloric acid and 2/3 acetic anhydride; however, investigation of a number of etching reagents failed to reveal any that were satisfactory. The recent work on this alloy has therefore been concerned primarily with the development of etching techniques. Consideration has been given to the use of various electrolytic polishing solutions as electrolytic etchants, and a number of solutions which give satisfactory results with other alloys have been tried. These include solutions of the following compositions: (1) 60 ml phosphoric acid, 15 ml sulfuric acid, 12 ml water (2) 32 ml phosphoric, 10 ml sulfuric acid, 10 ml water (3) 12 ml phosphoric acid, 47 ml sulfuric acid, 41 ml nitric acid (4) 66 ml sulfuric acid, 17 ml hydrofluoric acid, 17 ml hydrogen peroxide (3%) (5) 33 ml hydrochloric acid, 67 ml glycerine

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN (6) 14 ml phosphoric acid, 36 ml nitric acid, 50 ml sulfuric acid (7) 30 ml phosphoric, 60 ml sulfuric acid., 10 ml water (8) 50 ml nitric acid, 50 ml glacial acetic acid The concentrations of the acids used in the solutions are as follows: hydrochloric, 35%; hydrofluoric, 48%; nitric, 70%; sulfuric, 966; phosphoric, 85%. With the exception of No. 3, these solutions all produced adherent films of insoluble products on the alloy samples and. were therefore considered unsuited for etching purposes. Solution 3 stains samples aged at 16000F but appears to be suitable for use with samples aged at 1400~F and 1200~F. In an attempt to avoid the formation of the insoluble etching products, a number of nonaqueous reagents have been investigated. The compositions of these reagents are as follows: (9) 2% bromine in anhydrous methanol (10) ethylene glycol saturated with HC1 gas (11) liquid ammonia saturated with ammonium iodide (12) 10% fuming nitric acid (s.g. 1.60) in anhydrous methanol (13) 10 g of citric acid dissolved in 90 g of ethylene glycol The solution of bromine in methanol is an adaptation of the reagent of Mahla and Nielsen19 and is used as an immersion etchant. The attack of this reagent is rapid but varies from one crystal grain to another apparently depending on the crystallographic faces which are exposed, which results in a slightly roughened and pitted surface. A typical preparation is shown in the electron micrograph of Figure 5a, which was taken from a sample, aged 1000 hours at 16000F, after etching for 15 seconds. This micrograph was from a palladium-shadowed negative collodion replica and shows that the minor-phase particles are in relief in the surface. Good electron diffraction patterns of columbium carbonitride, the principal minor phase of this sample, were obtained from this surface (Figure 5b). The ethylene glycol- HC1 and the liquid-ammonia reagents are both used as electrolytic etchants. They have not been thoroughly investigated, but they have given satisfactory results in several cases. The liquidammonia reagent is unpleasant to work with, but it etches rapidly and without formation of insoluble products. Figure 6 shows an electron O10

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN a. Electron Micrograph X8000 b. Electron Diffraction Pattern Figure 5. Electron Micrograph and Diffraction Pattern Obtained from Inconel-X Alloy Using Bromine Etch. a. Electron Micrograph. X8000 b. Electron Diffraction Pattern Figure 6. Electron Micrograph and Diffraction Pattern Obtained from Inconel-X Alloy Using Liquid Ammonia Etch. 111

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN micrograph and an electron diffraction pattern obtained from a sample of the alloy, aged 1000 hours at 1400~F, prepared by etching for 3 minutes at 0.2 ampere per square inch in this reagent. Although the surface of the sample is somewhat roughened, the minor-phase particles are in good relief. The diffraction pattern contains rings of columbium carbonitride and M2 C6 carbide. Equally favorable results have been obtained with the HC1- etylene glycol reagent. Reagent 12 consisting of fuming nitric acid in methanol, produces heavy films of insoluble products when used electrolytically and is very slow to act as an immersion etchant. It is considered to be of little value in preparing this alloy. rThe citric acid reagent is ineffective as an electrolytic etch because of its low conductivity; it is also inactive as an immersion etchant. Very satisfactory surface preparations have also been obtained by immersion etching with aqua regia. The action of this reagent is similar to, though somewhat slower than,the action of the bromine-methanol reagent, and the etching varies for the different matrix grains depending on their orientation in the surface. It has also been found necessary to allow the aqua regia to stand for several days before using; otherwise it produces adherent films of insoluble products on the samples. With these various etching procedures, electron diffraction patterns have been obtained from five different samples of the alloy representing aging treatments of 10, 100, and 1000 hours at 16000F and 100 and 1000 hours at 1400~F. The patterns from the samples aged at 16000F show only rings of the columbium carbonitride, regardless of the etching procedures used. A typical pattern from the 1000-hour sample was shown in Figure 5 and the interplanar spacings and relative intensities corresponding to the diffraction rings are listed in Table I. The patterns from the samples aged at 1400~F show rings belonging to a complex carbide of the M23C6 type in addition to those of the columbium carbonitride. A. pattern of this type was reproduced in Figure 6 and typical values of the interplanar spacings and relative intensities are included in Table I. These diffraction results are summarized in Table II and show an interesting relationship between precipitate formation and aging treatment for this alloy. It appears that both the M23C6 and columbium carbonitride form when the aging is carried out at temperatures below 1600~F, but that the M23C6 phase does not form when the aging occurs at 16000F. There is an alternative possibility; namely, that the M23C6 carbide is present in samples aged for very short periods at 16000F but dissolves when the period is increased. This is considered to be an important point and will be investigated by studying samples aged for periods of less than 10 hours at 1600~F. The diffraction data obtained to date show no 12

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN evidence for a transition from M23C6 carbide to the M6C carbide such as was found in the 16-25-6 alloy. Furthermore, there is no evidence for the presence of TiN, a phase which was identified in this alloy by Ro,enbaum0 using x-ray diffraction methods. These matters are currently under i.nvestigation. In addition it is proposed to study a series of'iamp:l.; aged at 1200~F to provide more complete information on the influence of aging temperature. TABLE I INTERPLANAR SPACINGS AND RELATIVE INTENSITIES* FOR ELECTRON DIFFRACTION PATTERNS FROM SAMPLES OF INCONEL-X ALLOY Sample aged Sample aged 1000 hours at 14000~F 1000 hours at 1600~F dhkl I dhkl I 6.3 w 5.t w 3.%8 w 3.2 w 3.1 vw 2.55 w 2.,7 m 2.54 s 2. 12') ms 2. 06 s 2.19 s 1. 38 ms 1.'30 ms 1.69 vw 1.60 w 1.55 vw 4.843 w 1.56 s 1.)42 w 1.38 w 1.33 w 1.28 wm 1.34 s 1.26 wm 1.23 wm 1.26 ms *Xw = weak; m = medium; s = strong; v very. See Table III, page 15, for standard comparison patterns. 13

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN TABLE II MINOR PHASES IDENTIFIED IN SAMPLES OF INCONEL-X ALLOY BY ELECTRON DIFFRACTION Time of Temperature of Aging__ Aging, hours 14000F 1600~F 10 Cb(C,N) 100 M23C6 + Cb(C,N) Cb(C,N) 1000 M23C6 + Cb(C,N) Cb(C,N) N-155 Alloy An extensive investigation of the influence of high-temperature aging on the development of minor phases in low-carbon N-155 alloy is in progress. In carrying out this work it was decided to make an additional study of the suitability of mechanical polishing procedures for use in preparing alloy specimens for the electron diffraction. The mechanical polishing procedure which has been employed is an adaptation of one frequently used in optical metallographic studies. Samples were prepared by abrasion on metallographic emery papers through 3/0 or 4/0 grade and were polished on a canvas-covered polishing wheel impregnated with "Linde A" polishing powder. To avoid excessive cold-working of the surfaces, light pressure was used in holding the samples on the wheel and the canvas was liberally wetted with a dilute aqueous suspension of the polishing powder. With care suitably polished surfaces may be obtained in this way in a few minutes. More refined polishing may be obtained by finishing the surfaces on a silk-covered wheel using "Linde B" polishing powder; however, this has not been found necessary for satisfactory electron diffraction results. Following the polishing operation the surfaces were thoroughly cleaned to remove the polishing powder by rubbing them with a clean, wet cellulose sponge or cotton swab. An alternative procedure consists of treating the polished surfaces on a polishing wheel covered with a soft cloth which is moistened with clean water but which is not charged with a polishing powder. Satisfactory etching of surfaces polished in this way has been obtained through the use of an electrolytic polishing solution consisting of 300 ml of sulfuric acid (96%), 150 ml of phosphoric acid (85%), and 50 ml water. Current densities ranging from 0.1 to 6.0 ampere per square inch are used and the period of etching varies from 2 to 30 seconds, depending on the response of the various samples. Following the etching treatment the samples are thoroughly rinsed by methods previously described2 to remove etching products and reagents.

ENGINEERING RESEARCH INSTITUTE' UNIVERSITY OF MICHIGAN Using these procedures of sample preparation, good electron diffraction patterns of minor phases have been obtained from five different samples of the alloy representing aging treatments of 10, 100, and 1000 hours at 1400~F and 10 and 1000 hours at 16000F. The pattern from the sample aged 10 hours at 1600"F contains only a few relatively strong diffraction rings which correspond closely to those of columbium carbonitride. The patterns from the samples aged for 10 and 100 hours at 14000F contain the columbium carbonitride rings and additonal weaker rings which correspond closely to those of a complex carbide of the M23C6 type. A typical pattern from the 100-hour sample is reproduced in Figure 7 and the corresponding interplanar spacings and relative intensities are listed in Table III tog;ether with similar data for the M23C6 and Cb(C,N.) phases. The diffraction patterns obtained from the samples aged 1000 hours at 14000F and 1600~F differ considerably from those from the other samples. A typical pattern is reproduced in Figure 8 and the interplanar spacings and intensities are listed in Table IV. TABLE III INTERPIANAR SPACINGS AND RELATIVE INTENSITIES* FOR ELECTRON DIFFRACTION PATTERNS OBTAINED FROM N-155 ALLOY: SAMPLE AGED 100 HOURS AT 1400~F E. D. Pattern M23C6 Pattern** Cb(C,N) Pattern** dhkl I dhkl I dhkl I 2.7 A w 2.66 w 2.54 s 2.54 s 2.39 w 2.38 s 2.19 s 2.17 s 2.18 s 2.05 m 2.04 s 1.87 w 1.88 m 1.80 w 1.79 s 1.69 vw 1.68 w 1.61 w 1.62 w 1.54 s 1.55 ms 1.47 vw 1.48 vw 1.41 vw 1.42 vw 1,38 w 1.38 Vw 1.32 m 1.33 w 1.33 ms 1.26 w 1.26 s 1.26 m *w = weak; m = medium; s = strong; v = very **See Reference 20. 15

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN a. Electron Micrograph. X8000 b. Electron Diffraction Pattern Figure 7. Electron Micrograph and Diffraction Pattern Obtained from N-155 Specimens, Aged 100 hours at 14000F, after Mechanical Polishing and Light Etching. a. Electron Micrograph. X8000 b. Electron Micrograph Figure 8. Electron Micrograph and Diffraction Pattern Obtained from N-155 Specimen, Aged 1000 hours at 1600~F, After Mechanical Polishing and Heavy Etching. 16

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN TABLE IV INTERPLANAR SPACINGS AND RELATIVE INTENSITIES* FOR ELECTRON DIFFRACTION PATTERN FROM N-155 ALLOY: SAMPLE AGED 1000 HOURS AT 1600~F E. D. Pattern Fe2W Pattern** dhkl I dhkl I 4.1 w 4.07 3.83 3.6 w 3.60 2.8 m 2.80 ms 2.43 vw 2.35 s 2.36 s 2.18 s 2.17 s 2.08 w 2.05 m 2.01 s 2.02 s 1.98 s 1.98 s 1.92 m 1.94 m 1.80 w 1.82 w 1.74 m 1.75 wm 1.55 wm 1.51 w 1.53 w 1.43 w 1.44 w 1.36 w 1.37 m 1.32 m 1.33 s 1.28 m 1.29 s 1.22 m 1.24 s 1.20 w 1.21 w *w = weak; m = medium; s = strong; v = very **ASTM Card 3-0920 an entirely satisfactory identification of this pattern has not been obtained, although all the likely conpounds listed in the ASTM File of xray data and a number of phases recently reported in metallurgical diffraction studies have been considered. The best correlation so far reached is with an intermnetallic phase, Fe2W, which is listed in the ASUM File. As shown in Table IV the agreement with the diffraction pattern of this material is good with respect to both the values of the interplanar spacings and the intensities of the diffraction ring. These electron diffraction results are summarized in Table V and suggest some interesting relationships between the heat-treatments and 17

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN TABLE V MINOR PHASES IDENTIFIED IN SAMPLES OF N-155 ALLOY BY ELECTRON DIFFRACTION Time of Temperature of Aging aging, hours 1400~F 1600~F 10 Cb(C,N)+M C6 Cb(C,N) 100 Cb(C,N)+M 23C6 1000 "Fe2W" 26 "Fe2W"* *These samples were solution-treated 10 hours at 2200~F prior to aging. All others were solutiontreated for 1 hour at 22000F. the minor phases which form in this alloy. The most interesting point concerns the results which have been obtained from the samples aged for 1000 hours at 1400~F and 16000F. These samples differ from the others in two important respects: first, they were aged for periods much greater than the others and, second, they were solution-treated for 10 hours at 2200~F prior to aging while the others were solution-treated for only 1 hour at 22000F. It is difficult to determine at present which of these factors is most closely related to the very different electron diffraction results which have been obtained from these samples. To clarify this point additional samples of the alloy which have been solution-treated for the 10-hour period and aged for less than 1000 hours at 1400 and 16000F are being studied. Further consideration must also be given to the identification of the patterns from these two samples. While the agreement with the Fe2W pattern is good, this identification is considered to be tentative particularly since this material is not commonly encountered as a minor phase of this type of alloy. More detailed consideration will be given to this matter in future work. The results obtained from the samples aged for periods less than 1000 hours suggest that the formation of the M23C6 phase in this alloy follows a trend similar to that observed for the Inconel-X alloy and occurs only when the aging temperature is below 1600~F. Further studies of the samples aged at 16000F are in progress and a series of samples aged at 12000F is being examined to provide additional information on this matter. 18

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN Mechanical Polishing Procedures In all work to date where the electron diffraction method has been used in the study of the minor phases of alloy systems, the polishing of the alloy samples prior to etching has been carried out electrolytically. There is, however, the alternative possibility of using the mechanical polishing methods commonly used in optical metallography. In keeping with the objectives of this project it was of interest to consider the mechanical polishing methods and their possible advantages in the study of the heat-resistant alloys. While a thorough investigation of this matter has not been completed, the recent work with the N-155 alloy provides a considerable background of information and permits the more general aspects of the problem to be presented at this time. The principal objection to the use of mechanical polishing methods has been based on the fact that these methods produce severe distortion and working of the surface metal. It has been shown by various workers21'22 that normal polishing procedures, consisting of treatment of the surfaces on soft cloth wheels impregnated with fine polishing powders, destroys o the crystalline structure of the metal to depths as great as 50 to 100 A, producing a surface layer which is severely cold-worked and essentially amorphous in structure. Disturbances of lesser degree may extend to depths considerably greater than this. Recent x-ray diffraction studies23,24 of surfaces produced by surface grinders and cutoff wheels showed severe strains to exist to depths as great as 0.025 inch. These disturbances are sufficiently severe to introduce the possibility of an effect on the structure and composition of minor phases near the surface. Since the electron diffraction method is confined to the study of minor-phase particles which occur in this region, the use of mechanical polishing could lead to results which are not representative of the sample as a whole. The work with N-155 alloy reported here suggests that this objection may not hold for many of the heat-resistant alloys in which the minor phases consist of well developed carbides, for these phases appear to be highly resistant to both chemcial and mechanical influences. The electron diffraction patterns of Cb(C,N) and M23C6 phases which have been obtained from surfaces polished mechanically show no evidence of disturbance of the structure of the minor phases even in cases where the subsequent etching treatments were extremely light and did not even remove all the worked surface metal. The diffraction rings are as sharp and well defined as those obtained in other cases where electrolytic polishing was used. An example is shown by the micrograph and diffraction pattern reproduced in Figure 7. This diffraction pattern contains rings of both Cb(C,N) and M3C6 and compares favorably with similar patterns obtained from the Inconel-X alloy using electrolytic polishing (see Figures 5 and 6). 19

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN The patterns which have been tentatively identified as produced by the intermetallic compound, Fe2W, are of equally good quality as shown in Figure 8. The micrograph in this figure shows the type of surface from which the pattern was obtained. Here the etching treatment was fairly heavy and removed virtually all evidence of the mechanical polishing. There appear to be certain advantages which may result from the use of mechanical polishing methods. The resulting surfaces are considerably smoother than those produced electrolytically and permit the electron beam to strike a greater number of exposed minor-phase particles. This would undoubtedly be advantageous in obtaining patterns of phases which are present in small amounts or which have very small particles. It should also produce patterns having nearly continuous diffraction rings which are much easier to measure than the patterns of very spotty rings so frequently obtained with rougher surfaces. Occasionally, different minor phases of an alloy have particles which are greatly different in size. Such a case is shown in the micrographs of Figures 5 and 7, where the large round particles are the Cb(C)N) phase and the smaller particles are the M23C6 carbide. When electrolytic polishing is used, the large particles protrude very high above the surface and may prevent the electron beam from striking the smaller particles. Mechanical polishing levels the large particles along with the matrix metal and permits a better exposure of the smaller particles. This same effect may make it very difficult to obtain electron diffraction patterns from large particles, however. In general, the electron beam will not pass through more than about 1000-2000 A of solid matter without being absorbed. When particles are larger in diameter than this, patterns are obtained mainly from corners or edges which are thin enough to transmit the electrons. With mechanical polishing these corners and edges are polished off; particles are left very flat and smooth, as shown in Figure 7, and electron diffraction patterns may not be obtained. The mechanical polishing procedures are also advantageous in eliminating the difficulties so frequently encountered in electrolytic polishing due to the formation of films of adherent, insoluble products; however there is the corresponding problem of removing the polishing powders from the surfaces. This may be quite difficult, particularly with soft metals such as copper and aluminum, as shown by previous studies25. In the work with N-155 alloy, however, relatively simple washing procedures have proven adequate. Finally, it is evident that rather heavier etching treatments may be required to expose the minor-phase particles when mechanical polishing is used, since the metal is essentially flowed around and over the particles in the polishing. With electrolytic methods, the polishing procedures can usually be chosen to leave the particles partially exposed. 20

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN X-ray Diffraction Studies of Carbides of 16-25-6 Alloy In a previous report3, comparisons were made of electron diffraction patterns of the M6C carbide in samples of 16-25-6 alloy aged 1000 hours at 1200~F and 1400~F which indicated a difference of about 1h in the lattice parameter of the carbide of the two samples; however, limitations of the method made it impossible to determine the absolute values of the parameter with sufficient accuracy to establish this difference. Since such parameter variations may be closely related to variations in the compcsition of the carbide, it was considered important to obtain more accurate data. X-ray diffraction studies have therefore been made of carbides separated from each of the above samples and in addition from samples aged for 1 and 500 hours at 1600~F. The separation of the carbide from the alloy samples was based on the procedure described by Mahla and Nielsen19, which employs a 10% solution of bromine in anhydrous methanol to digest the matrix metal so that the carbide residues can be collected for the x-ray examination. The digestion was carried out in 40-ml centrifuge tubes containing 25-30 ml of the bromine solution. Samples were immersed in the solution overnight to assure complete reaction of the bromine, and on removal from the solution were carefully washed with a stream of methanol to dislodge adhering carbide particles. The tubes were then centrifuged to settle the residues and the liquid was carefully pipetted off. The residues were washed several times with clean methanol and, after drying, were mixed with small amounts of Duco cement and rolled into rods approximately 0.3 mm in diameter for use as diffraction specimens. The x-ray diffraction patterns were taken with a General Electric XRD-1 diffraction instrument using nickelfiltered copper radiation and a powder camera 14.32 cm in diameter. The diffraction data obtained from the residues from the four samples are listed in Table VI. The residues from the samples aged 1000 hours at 1400 and 12000F and 500 hours at 16000F gave strong patterns of an M6C-type carbide. Values for the unit cell edge were calculated from each of the lines of these patterns and the average value for each pattern is listed at the bottom of the respective columns in Table VI. Further comparison was made for the carbides of the samples aged 1000 hours at 12000F and 14000F by careful examination of a diffraction pattern from a mixture of approximately equal quantities of the residues from these samples. A detectable "splitting" of the strong lines of the M6C pattern would be expected if there was an appreciable difference in the unit cell size, but no such splitting was observed. These x-ray diffraction results do not support the previous electron diffraction evidence for a variation in the lattice parameter of the 21

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN TABLE VI INTERPLANAR SPACINGS AND RELATIVE INTENSITIES* FOR X-RAY DIFFRACTION PATTERNS OF RESIDUES FROM SAMPLES OF 16-25-6 ALLOY Heat Treatment of Samples 1000 hours, 1200~F 1000 hours, 14000F 500 hours, 1600~F 1 hour, 1600~F dhkl I d kl I dhkl I dhkl I 3.26 w 3.12 w 2.70 w 2o69 w 2.70 m 2.47 w 2.47 w 2.48 m 2.49 w 2.37 vw 2.39 s 2.20 s 2.21 s 2.21 s 2.17 vw 2.17 s 2.07 s 2.07 vs 2.08 vs 2.09 m 2.05 vw 2.05 vs 1.97 vw 1.92 m 1.91 s 1.92 ms 1.88 ms 1.80 w 1.80 1.81 m 1.79 s 1.64 w 1.51 w 1.52 ms 1.52 m 1.41 w 1.41 m 1.42 m 1.35 w 1.32 m 1.32 m 1.27 ms 1.27 s 1.28 s 1.18 m 1.19 wm ao = 10.80 A ao = 10.80 A ao = 10.82 A ao = 10.6 A * s = strong; m = medium; w = weak; v = very M6C carbide in the samples of 16-25-6 alloy. The lattice parameters calculated here are considered to be reliable to better than 0.5% and within these limits show no difference in the lattice parameter of the M6C carbide of the three samples studied here in which this carbide is the principal minor phase. Further evidence of this kind is provided for the 1200~F and 1400~F samples by the pattern from the mixture of the residues from these samples. If the lattice parameter differed by as much as 0.5% for these samnples, the difference in the radii of the 2.21 and 1.91 lines of the pattern would amount to approximately 0.2 and 0.4 mm respectively and would 22

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN have been easily detected as a "doubling" or "splitting" of the lines on the film, The lattice parameter variations which were proposed on the basis of the electron diffraction results were well within the limits of accuracy of the method. Furthermore, the electron diffraction results are sensitive to other factors such as the condition of the surfaces of the samples, so that the present x-ray results are considered to be more reliable. The residue from the sample aged 1 hour at 1600~F gave a diffraction pattern consisting of several strong lines of the M23C6 carbide and several weaker lines of the M6C carbide. The lines in this pattern are rather broad and not well suited for accurate measurement; however, the lattice parameter of the M23C6 carbide appears to be about 10.63 A. It will be noted that several lines of the M23C6 carbide also appear in the pattern from the sample aged 1000 hours at 14000F. These results are in agreement with the previous electron diffraction studies which showed the principal minor phase in the sample aged for 1 hour at 16000F to be the M23C6 carbide, rather than the M6C carbide, and which showed traces of this carbide in several samples which contained primarily the M6C carbide. DISCUSSION Considerable progress has been made in the studies of the minor phases of both the Inconel-X and the N-155 alloys, and the electron diffraction results which have been obtained indicate that heat treatments strongly affect the minor phases which form in these alloys. For the Inconel-X alloy it appears that the principal minor phase is columbium carbonitride when the aging temperature is 1600~F, while both the columbium carbonitride and a complex M23C6-type carbide form at the lower temperature of 1400~F. A similar variation is indicated for the N-155 alloy by results obtained from samples aged less than 1000 hours, while apparently neither of these phases appears in samples aged for 1000 hours. The best agreement for the patterns obtained from the 1000-hour samples is with an intermetallic compound, Fe2W. This identification is considered to be only tentative, particularly since this material is not usually found as a minor phase in the heat-resistant alloys. The interpretation of the results is further complicated by the fact that both the 1000-hour samples were solution-treated, prior to aging, for a longer period than the samples aged for lesser periods. Further studies are being made to clarify these points and the work on both alloys is being extended to include series of samples aged at 1200~F. It is anticipated that the completion of this work will provide considerable new information on the minor phases of these alloys. 25

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN While the x-ray diffraction studies of the carbides of the 16-25-6 alloy samples were undertaken primarily to investigate possible variations in the lattice parameter of the M6C carbide, the results permit an important comparison to be made of the results of the x-ray and electron diffraction methods. The x-ray studies were carried out with well established procedures which are widely used for identification of complex carbides in high-alloy steelsl81l9. It is considered significant that these studies did not detect any minor phases not previously identified by electron diffraction, while the electron diffraction studies indicated traces of M23C6 carbide in the samples aged 1000 hours at 1200~F and 500 hours at 1600~F where the x-ray studies indicated only the M6C caxbide. These results indicate that the electron diffraction method gives results comparable to the x-ray methods with the advantage of being more sensitive to small traces of minor phases. This is one of the first direct comparisons which have been made of the two methods and is important in establishing the relatively new electron diffraction method. 24

BIBLIOGRAPHY 1. Heidenreich, R. D., Sturkey, L., and Woods, EH. W., J. Applied Phys., l_ 127 (1946). 2. Brockwayy L. 0., and Bigelow, W. C., Annual Sumnmary for 1952 on Project 2020, Eng. Res. Inst., Univ. of Mich., to Flight Research Laboratory (WCRRL), WADC, 15 Jan. 1953. 3. Brockway, L. 0., and Bigelow, W. C., Progress Report No. 5 on Project 2020, Eng. Res. Inst., Univ. of Mich., to Aeronautical Research Laboratory (WCRRL), WADC, 15 April 1953. 4. Heidenreich, R. D., "Electron Microscopy" in Physical Methods in Chemical Analysis, edited by W. G. Berl, Academic Press, Inc., New York, 1950, Vol I, p. 535. 5. Heidenreich, R. D., "Electron Diffraction and Microscopy" in Modern Research Techniques in Physical Metallurgy, American Society for Metals, Cleveland, 1953. 6. American Society for Testing Materials, Committee E-4, Proceedings, American Society for Testing Materials, 0, 444 (1950) and 52, 543 (1952). 7. Report of Subcommittee XI of Committee E-4, American Society for Testing Materials Preprint No 57, 1953. 8. Schwartz, C. M., American Society for Testing Materials, Preprint No. 94c, 1953. 9. Forgeng, W. D., and Lamont, J. L., American Society for Testing Materials, Preprint No. 94d, 1953. 10. Austin, A. E., and Schwartz, C. M.,.J. Applied Phys., 22, 487 (1951). 11. Schwartz, C. M., Austin, A. E., and Weber, P. M., J. Applied Phys., 20, 202 (1949). 12. Schaefer, V. J., Science, 97, 188(1943). 25

13. Schaefer, V. J., Phys. Reviews, 62, 495 (1942). 14. Wyckoff, R. W. G., Electron Microscopy, Interscience Publishers, Inc., New York, 1949. 15. Fullam, E. F., American Society for Testing Materials Preprint No. 94a, 1953. 16. Backus, R. C., and Williams, R. C., J. Applied Phys., 20, 224 (1949). 17. Gerould, C. H., J. Applied Phys., 21, 183 (1950). 18. Kinzel, A. B., J. Metals, May 1952, p. 469. 19. Mahla, E. M., and Nielsen, N. A., Transactions, Amer. Soc. Metals, 43, 290 (1951). 20. Rosenbaum, B. M., National Advisory Committee for Aeronautics, Technical Note No. 1580, 1948. 21. Thomson, G. P., and Cochrane, W., Theory and Practice of Electron Diffraction, MacMillian and Co., London, 1939, p. 185. 22. Vacher, H. C., J. Research Nat. Bu. Standards, 29, 177 (1942). 23. Frey, D. N., Freeman, J. W., and White, A. E., National Advisory Committee for Aeronautics, Technical Note No. 2385, 1951. 24. Frey, D. N., Freeman, J. W., and White, A. E., National Advisory Committee for Aeronautics Report 1001, 1950. 25. Brockway, L. 0., and Karle, J., J. Colloid Sci., 2, 277 (1947). 26

i3 9015/ 02526317