&gg “@5@?5 _'_‘ e _L‘-SIK? fifl@@@ N&?U@N&fl;. fl;&” - @@mw; RESEARGH LIB - DOCUMENT @@L& @@'TI’UDN ééfl@@w{m@m{? g@mc«fl fiw fiom:‘wfifilh d@@mmb @m@’l the flfie*ywfi @.0 @ fl ;o CENTRAL RESEARCH LIBRANKY OCUMENT COLLECTION RGY RESEARCH LIBRARIES O | s 3 445k 0515521 1 ORNL-4327 UC-80 — Reactor Technblogy MECHANICAL PROPERTIES OF ARTIFICIAL GRAPHITES — A SURVEY REPORT W. L. Greenstreet OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION Printed in the United States of America. 'Available from Clearinghouse for Federal Scientific ond Technical Information, National Bureau of Standards, U.S. Department of Commerce, Springfield, Virginia 22151 . Price: Printed Copy $3.00; Microfiche $0.65 LEGAL NOTICE This report was preparéd as an account of Government sponscred work. Neither the United States, _ndr the Commission, nor any petson acting on behalf of the Commission: _ A, Makes any warranfy or representation, expréssed or implied, with respect to the accuracy, completeness, or uvsefulness of the information ‘contained in this report, or that the use of - any information, apparatus, method, or process disclosed in this report may not infringe _privately owned rights; or , B. Assumes ony liabilities with respect to the use of, or for damages resulting from the use of any information, apparatus, method, or process disclosed in this report. ‘As used in the above, '‘person acting on behalf of the Commission'’ includes ony employee or confruc_for‘of the Commission, or employee of such contracter, to the extent that such employee ot contractor of the Commission, or employee of such contractor prepares, disseminates, or provides access to, any informafion pursuont to his employment or contract with the Commission, or his emplt:’)rr.nenfi with such contractor. ORNL-H527 Contract No. W-7405-eng-26 Reactor Division MECHANICAT, PROPERTIES OF ARTIFICIAL GRAPHITES — A SURVEY REPORT W. L. Greenstreet DECEMBER 1968 OAK RIDGE NATTONAL LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION - for the U.S. ATOMIC ENERGY COMMISSION LOCKHEED MARTIN ENERGY RESEARCH LIBRARIES i 3 4456 0515521 1 Abstract ....... .. 0., e . Cetene e Ceeen INtroduChtion sevesrsevranernonneeas e . ‘e Carbon and the Hiétgry of Graphite t.veveeivernernennn. structure e e ceeraan Manufacture ...ieeevciovicnentenenensnnnans ceseen Mechanical ‘Properties ..... . ceesaresas et en e Room~Temperature Properties .viiveeevevene. benseanars Derived btress-Strain Relationships .... v Temperature Effects ... . v, Créep S eee et eeces e aeat s act ettt s atananees ceana Combined Stress Behavior ..eieseeeeeas e tsecseeeesacan References ....00v0e covea veene . cresetenas Acknowledgements . . . . cenes ceeen iii CONTENTS * & s 8 . - * L . N . -------- O N 12 18 19 30 32 35 36 38 45 MECHANICAL PROPERTIES OF ARTIFICIAL GRAPHITES — A SURVEY REPORT W. L. Greenstreet Abstract A review of published mechanical properties data for artificial graphites is given in this report; high-tempera- ture as well as room-temperature data are included. The intent is to provide a unified description of the complex mechanical behavior associated with these materials. The report also contains a brief history on graphite and discus- sions of crystalline structure and of manufacturing methods. Introduction This report was written to provide a survey of mechanical properties data for artificial graphites, or electrographites. Although brief re- views of these data are to be found in the literature, there is a need for an overall description which unifies the many aspects of the complex mechanical behavior associated with these materials. This report was written with such unification as a major objective. We have chosen to limit our consideration, in the main, to so-called "nuclear-grade, or equivalent," electrographites, that is, either molded.or extruded graph- ites made from petroleum coke and coal-tar pitch. Included in this se- lected group are premium quality and-specialty graphites. As a.prelude to the discussion of mechanical properties, a brief his- tory on graphite is given. This is followed by descriptions of crystal- line structure and of the manufacture of artifiéial graphite. The history section traces developments leading to the manufacture of graphite With- brief mention of the use of this material in the nuclear industry. Although the data considered are of a phenomenological nature on the macroscopic scale, a knowledge of crystalline structure and manufacturing methods is necessary to the understanding of graphite behavior. Hence, rudimentary discussions of these aspects are included as essential parts of this report. Both room- and high-temperature data are reviewed with the greatest emphasis being placed on the former. 1In the case of room-temperature be- havior, simple tensile and compressive stress-strain curves for monotonic and for cyclic loading are considered. Lateral as well as longitudinal strain data are reviewed. The discussion of behavior under combined stresses is necessarily brief due to the paucity of reported data on this subject. Carbon and the History of Graphite Aside from its use as a fuel, carbon plays a very important part in our everyday existence. Its great importance in industry is emphasized by the volume and dollar amounts of carbon and graphite products produced. Liggett [196L4] tells us that the estimated world production in 1962 was in excess of 1.4 billion pounds with a domestic value of about $400 mil- lion. The United States, France, Japan, West Germany, and England are the major producers and exporters of carbon and graphite. These five countries account for approximately T75% of the world production. Of these, the United States is the leader, producing in 1962 an estimated 410.2 million pounds with a value of about $151.7 million. It would be superfluous to enter into a lengthy discussion of car- bon and its uses here since our interest centers only on artificial graph- ites. Uses of carbon in its elemental or allotropic and mahufactured or fabricated forms are given by Mantell [1946].. The introduction in this book contains an interesting summary showing the scope of applications. A recent account of the uses is given in the Kirk-Othmer, Encyclopedia of Chemical Technology, Vol. 4, 196L. Diamond and graphite are allotropic crystalline forms of carbon. Nightingale [1962c] describes "amorphous carbons," the third form of car- bon found in nature, as those carbons in whioh the graphite structure is not completely developed or in which the graphitic structure is limited to volumes on the order of a few thousand angstroms. Examples of "amor- phous carbons” are coal and lignite. Hence we see that all natural forms of carbon are crystaliine. Both graphite and diamond can be synthesized. Diamond is completely converted to graphite when heated to 2000°C (Black- ‘man [1960b]) . At atmospheric bréssure, graphite sublimes directly into the gaseous state. Blackman [1960b] reports the sublimation temperature as around 3500 to 3700°C, while Walker [1962] quotes a value of around 3%350°C. These figures indicate the potential of graphite as a refractory. Graph- ite is characterized by high thermal conductivities, low coefficients of thermal expansion, low elastic moduli, and strengths which increase with temperature to about 2500°C. Hence, it is an excellent high temperature material. While diamond is one of the hardest substances known, graphite is one of the softest. The firsf discoveries of graphite are lost in antig- uity. Mantell [1946] tells us that it was known in early times since it was used for decorative purposes in prehistoric burial places in Europe and in ancient graves. It was long confused with other minerals includ- ing ores of lead, molybdenum, antimony, and manganese. These were be- lieved to be one and the same substance, or at least members of the same family. Acheson [1899] attributes this confusion to their outward resem- blance and to the fact that they produce mérks on paper. Because of this graphite was called molybdaena, plumbago, graphite, and black-lead. The true identify of graphite, or plumbago, was not recognized until late in the 18th century when the Swedish apothecary, Carl Wilhelm Scheele, demonstrated its carbon content (Thorpe [1894]).. Mantell credits the German mineralogist (Abraham Gottldb) Werner with givinglthe mineral the name graphite (from the Greek word graphein, "to write") in 1789. Aikin and Aikin [1814] concluded from the results of three independent experi- ments on combustion, which at that time were recently completed, that no decided chemical difference can be detected between diamond, pure plum- bégo, and charcoal. The earliest use of graphite was undoubtedly as an instrument for drawing and writing. The first mention of it is found in the Middle Ages . where it is described as a substance used.for this purpose (Mantell [1946]) . In 1564 the Barrowdale graphite deposit in Cumberland, England, was diécovered with the result that the manufacture of lead pencils on a commercial scale was originated. Accbrding to Seeley and Emendorfer [1949], authentic records show the use of clay-graphite crucibles in Bavaria in 1400. Thus, Nicholson [1795] gives the use of plumbago as a material for making pencils and crucibles. He stated that powder of plumbago, with three times its weight in clay and some hair, makes an excellent coating for retorts, and, ffirther, that the Hessian crucibles were composed of the same materials. Aikin and Aikin [1807] list, in addition to the above uses, the use of graphite as a lubricant for machining instead of o0il and as a sub- stance for the protection of iron from rust. ¥Finally, natural graphite is still used today in the manufacture of crucibles, certain molds and other equipment for foundry work and metal smelting, and for pencils. It is a major constituent in lubricéting mixtures and in the rubber tire industry for hardening and improving wear resistance. These are a few of the many uses; for more information, see Seeley [1964]. Manufactured; or artificial, graphite is used in many applications made possible by the existence of this product, and it can now be sub- stituted for natural graphite in almost all uses of the latter material. Artificial graphite products can be made in a variety of shapes and sizes, and the properties can be adjusted during manufacture to tailor the ma- terial to the specific requirements of a given application. Artificial graphite is actually a crystalline graphite, with the only artificial attribute being the method of production. While natural graphite practi- cally always contains admixtured impurities, artificial graphites with purities that may exceed 99.999% carbon (Eatherly and Piper [1962]) can e manufactured. The history of the artificial graphite industry is closely related to the development of carbon electrodes. It is generally believed that Sir Humphry Davy was the first to use carbon electrodes with the electric arc. In a letter to Nicholson, which was written in 1800 (Davy, J. [1839]), Davy tells of finding that well-burned charcoal produces shock and sparks when connected to the ends of a voltaic pile, or battery. He also tells of the use of two long, thin sli@s of dry charcoal (carbon electrodes), connected to a voltaic pile, in the decomposition of water. Later, in 1802, Davy described a spark "of vivid whiteness" which was ob- tained when pieces of well-burned charcoal were connected to a battery (Davy, J. [1839]). During succeeding years, batteries with increased power were made, and in his writings published in 1812 (Davy, J. [1840]) Davy talks of points of charcoal connected to a battery producing "a light so vivid, that even the sunshine compared with it seemed feeble." In this same reference, we find that Davy used a powerful 2000 double-plate battery of the Royal Institution (sometimes referred to as the "great battery" of the Royal Institution) with two pieces of charcocal, each about 1l-in. long and 1/6 in. in diameter, to produce a constant discharge of at least 4 in. in length fihich produced a brilliant arch of lightfl In addition, - he was able to producé a discharge 6 or 7 in. in length in a partial vac- uum. This was the forerunner of great developments. During most of the 19th century, the majority of the development work in the carbon industry was directed toward giving improved electrodes. Hinckley [1921] points out that impetus for this was provided by the in- vention of the dynamo and the development of great quantities of hydro- | electric power at Niagara Falls. _(Details of the carbon electrode devel- opments during this period are given by Mantell [1928, 1946].) The Frenchman, M. F. Carré, 1s called the founder of the arc-carbon industry. His electrodes were made of pulverized pure coke, calcined lampblack, and sugar syrup; the mixture was patented in 1876. He pounded these in- gredients together and.kneaded them into a hard paste, then pressed the paste hydraulically and baked it at high temperature. The superiority of his product was described by him in Carré [1877]. ' Carré therefore established the crude beginnings of the industrial operations of célcining, grinding, mixing, shaping, and baking. These general operations in the manufacture of electrodes are largely followed today, although improvements have been made in every one of these steps. Carré's mixture also indicates the raw materials used in making manu- factured carbons. These are (1) the carbonaceous particles making up the bulk of the product, termed "body" materials, and (2) "binders," which serve to hold together the finely ground particles of the "body" materi- als. Miscellaneous substances_are added to the mixture to give certain specific physical or mechanical properties. Hinckley [1921] reports that before 1850 small quantities of graph- ite were made in an-electric furnace and the tips of carbon'electrodes after use for some time were found to be converted to graphite. However, the comercial production of artificial graphite did not begin until near the end of the 19th century. Castner was granted a United Statés patent for electrical baking of carbon electrodes in 1896 (Castner [1896]), and in the same Year, Acheson, then of the Carborundum Company, obtained a patent (Acheson [1896]) for the manufacture of graphite in an electric furnace. Acheson discovered that graphite can be produced in an electric fur- nace while studying the effect of very high temperature on carborundum, ‘or silicon carbide. He found that the material decomposed with the sili- con being vaporized and carbon being left behind not in the amorphous but in a graphitic form. Fitzgerald [1897] in describing the manufacture of carborundum referred to the formation of pure gréphite next to the core of the furnace. He told of Acheson's conclusion that graphite is not formed simply by the process of subjecting amorphous carbon to very high temperature. Instead, the carbon enters into chemical combination with some oOther element and this compound then decomposes, leaving carbon in graphitic form. Additional discussion about the formation of graphite in the electric furnace is given by Acheson [1899]. Concerning his method for the production of graphite he states: "This method of manufacturing graphite, I would define, as con- sisting in heating carbon, in association with one or more oxides, to a temperature sufficiently high to cause a chemical reaction between the constituents, and then continuing the heating until the combined carbon separates in the free state. It is not, - however, limited to the use of oxides, as pure metals, their sulfides, and other salts may be used; but for various reasons the oxides are to be preferred.” Although many metal carbides do decompose to yield well crystallized graphite, Acheson's conclusion is fallacious. During the 1920's and 1930's, investigators at the National Carbon Company disproved his con- clusion in many controlled experiments. It is, for example, possible to show that a pure petroleum coke undergoes at least as high a degree of graphitization as one with added iron, iron oxide, or other similar im- purities (MacPherson [19681). The stages of manufacture of carbon and of graphite electrodes are very similar. True electrographites, or artificial graphites, are made up as "amorphous" carbons or graphite carbons, but are given a final very high heat treatment to transform microecrystalline carbon into crystailine graphite (Kingswood [1953c]). This last step is called the graphitization process. The differences between carbon and graphite and the temperatures involved in graphite production will be discussed presently. Mantell reports that in June 1897, Acheson produced the first graphite electrodes. These were made at the reguest of Castner for use in electro- chemical processes. Three years after the patent for graphite manufacture was granted, the Acheson Graphite Company was incorporated and began to build a plant at Niagara Falls, which became the center of the industry. By that time [1899] Acheson had used the furnaces of the Carborundum Com- pany for a year or more and had produced over 200,000 carbon electrodes 15-in. long and having a cross-sectional area of 1 in.® for use in the Castner alkali process, both in the United States and in Furope (Acheson [1899]) . The life of these electrodes was many times that of the same electrodes ungraphitized. Electric resistance furnaces of the type invented by Achescn [1895] are used today for the graphitization process. His was the first electric furnace in which temperatures approaching 3000°C, as needed for graphit- ization, could be attained. | As stated by Walker [1962]: "This marked the beginning of a new era in which the carbon in- dustry expanded and developed improved baked and graphitized carbon for use primarily in (1) electrolytic manufacture — e.g., of alkalies, chlorine, aluminum, and manganese; (2) electro- thermic production — e.g., of calcium carbide and silicon car- bide; and (5) electric furnaces — e.g., for steel, copper, ferro- alloys, and phosphorous. Then in 1942, when Fermi and a group of scientists produced a self-sustaining nuclear chain reaction, they used graphite as a moderator in their reactor. This opened up a whole new outlet for graphite and at the same time produced materials problems on which much research and development studies are still in progress." We must add to this by saying that nuclear reactor applications and the more recent uses in.the aerospace fields have demanded a thorough exami- nation of the mechanical properties of artificial graphites. These prop- erties received little attention, as evidenced by published works, prior to their use in these advanced technological applications. Currie, Hamister, and MacPherson [1956] point out that graphite was selected as a neutron moderating material in the first nuclear reactors because it was the most readily available material that had reasonably good moderating properties and a low neutron capture cross section. In addition, its use was made widespread due to its low cost and the ease with which it can be precisely machined. Smyth [1948] gives a historical account of the first uses of graphite in the nuclear industry. Night- ingale [1962b] lists 75 graphite-moderated reactors which were in opera- tion, under construction, or definitely planned at the time of writing. The number aiso includes reactors which were taken ofit of operation but omits a few low-power training and educatiohal reactors. Gwinn [1949] states that the use of graphite as a moderator in the- plutonium piles in the atomic energy plants at the University of Chicago and Hanford, Wash- . ington, is perhaps the most dramatic and widely publicized application of this material. This application was announced in 1945. In addition to the appellation "artificial graphite," the terms "syn- mn f1 n 1 thetic graphite,'" "graphitized products," "electrographite,' or "graphite" réfer to those products in which "baked carbon" is further heat treated, generally in an electric furnace, at temperatures of 2200°C or higher (Liggett [1964]). Artificial grarvhite can be made from almost any or- gariic material that leaves a high carbon residue on heating. However, since the beginnifig of the industry, the principal material_used for making up the bulk of the finished article (called "pody material" or "filler material") was probably petroleum ccke (MacPherson'[l968]). The growth of the petroleum industry resulted in increased availability of petroleum coke, which was found to be the purest form of carbon available in large quantities. Thus, petroleum coke has remained the principal ma- terial for graphite manufacture, and this preeminence is held throughout the world today. In the first edition of his book on the carbon industries, Mantell [1928] lamented the paucity of information by writing: "While most industries have a more or less abundant literature, with numerous books, pamphlets, and articles, the manufacture of carbon electrodes is a notable exception. Were it not to be considered sarcasm, the industry might even be referred to as a 'black art!, first because of the secrecy usually surrounding its processes, and second, because of the absolute physical dirtiness of the usual electrode plant. Carbon works have generally been regarded as fortifications, through whose gates only the initiated might enter. Most of the manufacturers have been of the opinion that the less said on the subject the better for them. One reason for this, perhaps, is that in Europe, the art of manufacturing high-grade electrodes was always regarded as a secret, of which only a few had definite knowledge." An echo to this state of affairs is perhaps represented by the choice of title, "Black Magic," by Speer Carbon Company [1949], for their booklet issued on the occasion of the 50th énniversary celebratioh of the company. However, the situation changed drastically after that time, particu- larly since 1950, and dearth has been réplaced by deluge. The literature now abounds with published works on carbons and graphites, with a notable exception being the subject of mechanical behavior of graphite under com- bined stress states. Survey articles have been written by Kingswood [195%a, 1953b, 1953%c, 1953d], by Currie, Hamister, and MacPherson [1956], by Blackman [1960a, 196Cb, 1960c], by Walker [1962], and by Shobert [1964L]. The article by Currie et al. details the influences of raw materials and processing methods upon variations in properties of artificial graphites. Books on this subject, in addition to those by Mantell [1928, 1946], are Ubbelohode and Lewis [1960], Nightingale [1962a], and Union Carbide Corporation.[l96h]. The book by Simmons [1965] is of less general inter- est since it is devoted to nuclear-irradiation effects on graphite. Structure The hexagonal lattice structure proposed by Bernal [192k] is now accepted as the ideal structure for graphite. This is a layered struc- ture and is composed of a syétem of infinite layers of fused hexagons, that is, the atoms of‘carbon in graphite lie in planes in which they form sets of hexagons. The nets are in successive parallel planes superposed 10 so that half the atoms in one net lie normally above half the atoms in the net beneath, while the other half lie normally above the centers of the hexagons of this net; see Fig. 1. Since alternate nets lie atom forr atom normally above each other, the stacking sequence is ABAB .... . ‘ The carbon atom spacing within a plane is 1.415 3, while the separa- tion between parallel layers is 3.3538 E (15°C) (Walker [1962]). The theo- retical density for this structure is 2.267 g/cm® (Blackmen [1960a]). The carbon atoms in the layer planes are held by strong valence forces, where- as the interplanar binding forces are very weak. Joined atoms between layers are pinned by weak forces allowing adjacent planés to be easily displaced parallel to each other or rotated around an axis perpehdicular to the planes. In addition, this feature probably accounts for the aniso- tropic properties of graphite crystals such as shown by electrical con- ductivity, thermal conductivity, and linear expansion measurements. The electrical and thermal conductivities are greater in directions parallel to the layer pilanes than normal to them, while the thermal expansion is greater in the direction normal to the layer planes. Although the most common stacking in graphite is hexdgonal; in some cases a small percentage has a lattice in which the carbon hexagons have shared edges, but the stacking of hexagonal plates in layers 1is such that every fou:th layer is in juxtaposition with respect to the c-axis (the axis normal to the planes). Thus, the stacking sequence is ABC ABC ...; ufie} 'L'/\ Co < INTERPLANAR DlSTANCOJE 3.354 A Fig. 1. Structure of the Hexagonal Form of Graphite (Seeley [1964]). 11 see Pig. 2. This modified lattice, called rhombohedral, was proposed by Lipson and Stokes [1942]. It is not usually present in artificial graph- ite, but is found in natural graphite. This form is metastable with re- speét to the hexagonal form from which it may be produced by mechanical deformation, as in fine grinding (Bacon [1952]). It reverts to the hex- agonal form on heating to 1300°C. Deviations from the ideal graphite structure among most carbons of cormercial interest are prevalent (Walker [1962]). The most important deviation is the presence of stacking disorders between layer planes in carbon. Since it allows us to bring out an important distinction between carbon and graphite, this deviation will be described here.* The simple analogy of Seeley [1964] in terms of a deck of playing cards provides a clear illustration. Suppose each card represents a sin- gle plane of hexagons in which the carbon atoms are ordered in two dimen- sions with the proper spacings. When the deck is evened at the sides and *Chemically, there is a ready means for differentiating between amor- phous carbon and graphite. One part of the substance to be tested is treated with three parts of potassium chlorate and sufficient concentrated nitric acid to render the mass liquid. The mixture is then heated on a water bath for several days. GCraphite is converted into golden yellow flakes of graphitic acid, while amorphous carbon is altered to a brown substance soluble in water (Mantell [1946]). d. SPACII\OJG 3.354 A Fig. 2. Structure of the Rhombohedral Form of Graphite (Seeley-[l964])f 12 at the ends, ready for dealing, it has three-dimensional ordering and may be considered as representing graphite structure. After the cards are dealt, played, and bunched without evening the ends and sides nor rotating the cards for redealing, the deck represents turbostratic structure. This is the structure of so-called amorphous carbon, that is, the cards, though parallel, are without order in the third dimension.® Refinements are required to complete the analogy. Hexagonal graph- itic structure requires that every other card in the ordered deck be moved laterally the same distance and that the cards be the equivalent of 3.3538 R apart. The vertical distance between cards represents the interlayer, or d, spacing in the crystalline structure. Franklin [1951] concluded from an analysis of many nongraphitic and graphitic carbons that two distinct and well-defined interlayer spacings exist for a system of parallel carbon macromolecules. These are 3.4L4 A between adjacent misoriented layers and 3%.354 K.between correctly oriented layers. Hence, the turbostratic structure of carbon requires that each card in the bunched deck be separated by a minimum of 3.4LL A. Through the work of Franklin [1951] and Bacon [1951] the d spacing has been related to the proportion of diéoriented layers, p. With in- creagsed disorder of the layer stacking from ideal graphite, p varies from O to 1. Thus, using X-ray diffraction patterns to determine p, the ratio of graphitic carbon to nongraphitic carbon in a specimen can be estimated from its mean d spacing. Manufacture To complete the general description of artificial graphite, we will give a brief-summary of the manufacturing sequence. This will aid in understanding the bulk behavior of the material and help to relate micro- scopic to macroscopic aspects. Since we are interested in nuclear-grade, or equivalent, graphites, the primary sources of information used are Eatherly and Piper [1962] and Union Carbide Corporation [1964]. % The term "turbostratic" was coined by Biscoe and Warren [1942] to describe this unordered material. 13 As'élready mentioned, in most cases, artificial graphites are pro- duced from petroleum-coke filler material, which is true of nuclear graphites. The binder is coal-tar pitch. The petroleum coke is a by- product in the refining of petroleum crude, and today it is largely ob- tained by the cracking of a heavy refinery oil. When heated to a tem- perature of 2800 to 3000°C, the carbon from petroleum coke achieves a high degree of crystallinity (or-degree of graphitization), that is, the crystalline properties (lattice dimensions) approach those of a perfect erystal. A highly crystalline graphite has high thermal and electrical conductivities and a large crystallite size. The raw cokes from the refineries have textured, partially aligned structures. The crystallites show no three-dimensional order, but the lamellar order is sufficient to cause alignment of adjacent crystals to varying degrees. The fixed carbon (the carbon remaining after heating to 1000°C) in practically all of the petroleum cokes used ranges from 85 to 90%. The volatile content ranges from 7 to 16%, with a typical value of 11%. Other constituents are ash and various impurities. Ccal-tar pitch is the heavy residue derived from the distillation of coal-tar from by-product coke ovens used in preparing metallurgical coke. This coke is derived from the destructive distillation of bitumi- nous coal, and it is the chief reducing agent employed by the steel in- dustries in the blast-furnace reduction of iron ore. Coal-tar pitch is an excellent material for graphite manufacture because it is solid at room temperature and fluid at higher temperatures; in addition, it has a high carbon content. The thermoplastic property allofis for thorough mix- ing of the filler with the binder, facilitates the forming of the filler- binder mixtures, and permits storage and handling of the formed articles at room temperature without adversely affecting the shape of the product. The carbon content of coai-tar pitch is approximately 93%, and after heating to 1000°C about 55% of the pitch remains as binder carbon. Be- ~cause of its high carbon content it has been described as a form of "lig- 1 ~ uid carbon,"” which, through the addition of 7% alloying constituent, has a softening point at 100°C. The processing steps in the manufacture of a conventional, extruded, nuclear graphite are summarized in Fig. 3. The raw petroleum ccke is 14 RAW PETROLEUM COKE () CALCINED AT 1300°C CALCINED COKE MILLED AND SIZED FLOUR . PARTICLES MIXED (D COOLED EXTRUSION OIL EXTRUDED | GREEN ARTICLE I BAKED TO BOO°C BAKED ARTICLE IMPREGNATED WITH PITCH GRAPHITIZED TC ~3000°%C NUCLEAR GRAPHITE Fig. 3. Flow Diagram for Manufacture of Nuclear Graphite (Eatherly and Piper [1962]). ‘ first calcined at temperatures up to 1400°C, usually in a large rotary gas- or oil-fired kiln. The purpose is to remove volatile hydfocarbons and to affect a shrinkage of the filler material before it is incorpo- rated in the formed article. About 25% of the weight of the raw coke is | lost during this process. During calcination the aligned structure of the raw coke is preserved, but the layer planes of carbon atoms increase 15 in dimension over that present in the raw coke. The calcined petroleum coke is, as yet, a turbostratic carbon. After calcination, the material is broken down by crushers and mills and sized, through screens, into a series of carefully controlled frac- tions. The finest fraction, termed coke "flour," has a maximum particle size of 0.015 in. for most electrode and specialty graphites, while the coarsest fraction has particles as large as 0.5 in. Selected size frac- tions are recombined to produce a dry aggregate wherein the proportion of fractions and fraction size are varied, within limits, to control the properties of the end product. The maximum particle size for a coarsé— grained nuclear graphite is 1/32 in. When the calcined coke is crushed or milled, the individual parti- cles, although irregular in shape, tend to have one dimension longer than the other two; - The shapes of the particles and the alignment of the rudimentary crystallites in these particles depend on the éoke source, that is, they depend on the refinefy practice and the charge stocks em- ployed. However, the predominant orientation of crystallite layer planes is paréllel to the longer particle dimension. The next step in the manufacturing sequence is that of mixing the dry coke aggregafe.with the coal-tar pitch to make a formable plastic mix. The mix is heated to a temperature normally in the range from 165 to lTO°C, where the binder is quite fluid. This allows for a good dis- tribution of the pitch in the petroleum-coke filler materials. In the ideal situation, each coke particle is coated with a film of pitch. When forming is done by extrusion, about 30 parts by weight of binder are added to 100 parts of filler. The proportions may differ from these when the product is to be molded. » Furnace blacks or extremely fine (<10u) coke particles may be added t0o the coke mixture to increase the bulk density of the artificial graph- ite. These additives fill voids that would otherwise exist between the large particles. Also, when extrusion is the forming method to be em- ployed, the addition of lubricating oil to the mix is common practice. The oil serves to reduce the friction between the surface of the die and the mix. Thus, the quantity of pitbh otherwise necessary for reasonable 16 rates of extrusion is reduced, which is desirable because the evolution of additional volatiles during the baking operation gives rise to a struc- turally poor product w1th inferior properties. As noted above, the coke-pitch le is formed either by extru51on or by molding. In forming, the long axis of the coke particles take a pre- Aferred oriéntation either in the direction of extrusion or perpendicular to the direction of molding. The final graphite product retainé the same pattern of .grain orientation. The with-the-grain direction is parallel to the extruéion axis in extruded graphite and perpéndicular to the di- rection of molding pressure in a molded piece. The against-the-grain di- rection is perpendicular to this. Hence, there are marked differences in properties between the two directions with.the anisotropy in molded graphites being generally less than that found in extruded materials. Mrozowski [1956] points out that the anisotropy of physical prop- erties is due to two causes. One is the crystaliite alignment.in the particles, as discussed above, and the other is purely geometrical in nature. In the latter, the particle alignment creates an anisotropy due to the relative frequency of binder-bridges per unit path in different ‘directions. The rélativé contributions of the two sources of anisotropy depend on the type of physical property 1nvest1gated | ‘The artlcle, after forming, is called a "green'" carbon body and con- sists of calcined petroleum-coke filler bonded by coal-tar pitch. This .body.is then baked in a gas-fired furnace to convert the pitch from a thermoplastic material to an infusible solid while at the same time main- taining the shape imparted.by forming. This operation is critical‘and must be carefully controlled. | The green carbon loses its strength during the first part of the cycle, and the volatiles in the pitch, which yield a sizeable volume of gases, must escape through a relatively impermeable mass without disrupt- ing the structure. Polymerizétibn and cross linking proceed within the | binder and between binder and filler materials so that the platelets of the pitch increase in size from their original’dimensions of a few ang- .stroms. When the tempefature reaches 800 to 1000°C (the final baking temperature), the cross-linking process causes the carbon to become ex- tremely hard and brittle. At the same time, the binder shrinks about 17 5% by volume, creating high stresses that can crack the carbon body. Mrozowski [1956] states that because of the calcining operation in which the filler has been preshrunk, the only way in which the shrinking binder surrounding a coke particle can decrease its volume is by way of bpening crackslperpendicular to the particle surface. The entire baking cycle may take from a few weeks to two months, depending on the furnace size and the méthods used. | The baked carbon has a porosity of about 25%. 1In drder to reduce the porosity and thus increase the bulk density, an impregnation opera.- tion in which coal-tar pitch is forced into the pores or voids in the body is used. The fimpregnating pitch differs somewhat from' the binder .pitch in that some of the heavier fractions normally present in the binder pitch are missing. The impregnation is performed by preheating the car- bon to a temperature above 200°C, immérsing it in molten pitch, and pres- surizing it in an autoclave to a pressure of the order of 100 psi for several hours. | The ultimate crystal orientation and size are inherent in the raw materials, but, even after the gas—baking process, the carbon possesses little true crystal structure. Both-long-range order and internal per- fection,'aithough latent, are not yet developed. It i1s the purpose of the last step in the manufacture of graphite stock to affect crystal | growth and to ?erfect the internal order; that is, to convert carbon to graphite. This is the graphitization procesé mentioned earlier, and re- gquires a temperature in the 2600 to 3000°C range. Upon heating the material during graphitization the dominant pfocess between lSOO°C and about 2500°C is crystal growth with internal crystal structure still imperfect. Above'2500°C, continued minor crystal growth occurs, but the major éffects are diffusion and annealing. The stacks of parallel planes are ordered according to the layered stacking of graph- itic structure in this last stage of heating. The graphitization opera- tion takes about two weeks. | | Tests have shown that the properties of any piéce of graphite are. directly dependent upon the highest temperature reached in graphitization. However, very little chahge occurs in the room-temperature mechanical properties on heating above 2500°C. 18 Although the theoretical density for a graphite crystal is 2.26 g/cm?, a typical value of the bulk density for artificial graphite is 1.70 g/cm? at room temperature. This is due to0 porosity between filler particles and between crystallites within both the filler and binder carbon. New methods in fabrication such as forming and baking a carbon article under pressure and the use of a hot-working process, which produces a reorien- " tation and recrystallization of the carbon, have led to graphites with bulk densities in excess of 2.1 g/cm® (Eatherly and Piper [1962]). These materials have higher tensile strength than the more conventional lower density graphites. | From the preceding discussion it is seen that the anisotropy associ- ated with artificial graphite should have an axis of symmetry with prop- erties in a plane normal to this axis being independent of direction, that is, the material may be classified as transversely isotropic. The axis of symmetry is in the direction of molding pressure (against—the— grain) or in the direction of the extrusion axis (with-the-grain), de- pending on the forming method used in the manufacture. This hypothesis is supported by stress-strain measurements made by Kennedy [1961], who examined the stress-strain behavior of an extruded graphite in the di- rection parallel to the extrusion axis -and in two orthogonal directions normal to this axis. The material exhibited deformatiofi resistance in- dicative of rotational symmetry, but the data also show that the ultimate strengths and elongations do not necessarily follow this pattern. Mechanical Properties Artificiallgraphites are brittle materials with low strengths at ordinary tenmperatures in comparison with metals. ThéSe graphites have large variations in properties, as might be inferred from the descrip- tion of their manufacture. A piece of this material may be described as a mixture of graphite and poorly graphitized binder carbon. The varia- tions in properties stem not only from the raw materials used but also fram the size and shape of the finished article. Differences are found from piece to piece in a given lot and grade, with some variation in properties within each piece. Indications of variations in density, 19 electrical resistivity, modulus of elasticity, and flexural strength to be expected within formed pieces, molded and extruded, are given by Wright [1956]. These data show that differences of 15 to 30% across the diameter of a large piece may be expected. In general, both the modulus and strength decrease with distance from the outer edge of a transverse section. Despite these variations, which the manufacturers are striving to reduce, definite characteristics are associated with the'prOPertieé of graphite as a class of materials, and it is the purpose here to focus attention upon these. The discussion will be limited to mechanical prop- erties. The review articles and bocks already cited contain summaries on electrical and physical properties. Room-Temperature Properties Typical room-temperature properties for a nucléar graphite are given in Table 1. The values are for a fine-grained extruded pilece about 4L in. by 4 in. in cross section. TFrom this table, it may be seen that the strengths and elasticlmodulus are greater in the with-the-grain (par- allel) direction, which is generally true for artificial graphiteé as | would be expected from the discussions on crystal structure and manu- facture. Complete stress-strain diagrams for simple tension and compres- sion are plotted together in Fig. 4. The material is a nuclear-grade graphite and the data are for the'with-the-grain direction. As indicated in Fig. L4, the stress-strain curves for graphite are nonlinear even at low stress levels, and there are pronounced differences between both stress and strain at fracture in tension and in compression. Typically, fracture strains on the order of 0.1 to 0.2% and 1.0 to 2.0% are found in tension and in compression, respectively. Arragon and Berthier [1958] performed compression test studies on 216 specimens made from extruded, petroleum-coke, industrial graphite.v Three types of test were used: (1) simple compression, (2) cyclic tests (loading-unloading-reloading) in which the cycles were made from in- creasing stress levels spaced at equal intervals, and (5) cyclic tests between zero stress and a fixed maximum. The simple compression test curves did not show any abrupt changes ih slope corresponding to the incipience of plasticity. This lack of observable breaks in the curves 20 Table 1. Room-Temperature Properties of a Typical Nuclear Graphite (Eatherly and Piper [1962]) l to grain Value Deviasion Bulk denéity, g/cm> 1.70 0.02 Electrical resistivity, uohm-cm: || to grain T3k 59 l to gréin 9HQ 111 Thermal conductivity, cal/(sec)(cm)(°C): || to grain 0.5k43 | to grain 0.330 Tensile strength, psi: || to grain 1440 25k l to grain 1260 308 Compressive Strength, psi: || to grain 5990 658 | to grain 5960 918 Flexural strength, psi: || to grain 2400 506 | to grain 1970 509 Young's modulus, psi: | || to grain 1,19 x 10° 0.14 x 10° | to grain 1.11 x 10° 0.09 x 10° Coefficient of thermal expansion, per °C: o || to grain 2.22 X 107 0.39 X 1076 3.77 x 107® 0.hk2 x 107 21 2000 TENSILE 0 2 GOMPRESSIVE w @ - 2000 _ - o - ® / -4000 -6000 -1.4 -{.2 -1.0 -0.8 -0.6 -0.4 -0.2 o 0.2 0.4 ’ STRAIN (%) Fig. 4. Uniaxial Stress-Strain Curves, Parallel to Extrusion Axis (Greenstreet et al. [1965]). is in accordance with concomitant occurrence of both elastic, or revers- ible, and plastic straining throughout the stress range. The slopes of the stress-straifi curves at the origins were of the same order of magni- tude as given by sonic measurements. Seldin [1966a] alsoc found good comparisons between sonic moduli and the slopes at zero stress as mea- sured from tensile and compressive stress-strain curves for several - grades of molded graphite. | | Arragon and Berthier'demonstrated.that hysteresis loops are pro- duced on unloéding and reloading. For illustration; a stress-strain dia- gram for test type b is shown in Fig. 5. The curves are concave toward the load axis on unloading and slight concavities toward this axis are found on reloading. But, in first approximation, the reloading curves - ~are straight lines. A "compression limit" (~700 psi) was also reported. Above this 1limit, there is residual permanent deformation upon unloading which increases with unloading stress up to rupture. The authors report . that below this limit, cyclic ldading curves reclose at the origih, leaving no permanent deformation. Tt is interesting to note that neither concavity toward the load axis for reloading curves nor a limit corresponding to the'compression limit.deséribed above have been reported by Andrew, Okada, and Wobschall [1960], Losty and Orchard [1962], and Seldin [1966a]. The first authors investigated hysteresis effects in bending and torsion, while the others . 22 200 ) HQ _ —————————————————————— ;77‘* £ S 150 = gé | %./ ’ ICI) - b 8p e o © A 7 / / 2 100 o e P © = o o? o g G S o~ S o g o A8 S g P s o S LA e o7 s s - O ./do | e e o7 Fog © vy e T L } 3;6, X ZooT o O Cr d = —— o 1 2 3 4 DEIFORMATION (Fo) Fig. 5. Stress-Strain Diagram for Compressive Tests of Type b (Arragon and Berthier [1958]). ' performed cyclic loading tests in simple tension and compression. The residual strains on unloading were observed to increase always with in- creasing maximum stress. Jenkins [1962a] speculated that curious fea- tures of the curves obtained by Arragon and Berthier may be explained by end effects in the Specimens. _ . Arragon and Berthier found that cyclic compressive tests of type b cause the apparent density to increase. The hysteresis lobps increase in size with increased maximum stress, and the slopes of straight lines connecting the unloading and reloading points decrease with increased stress. (The sldpe of a line connecting the two points of one cycle is termed the "paraelastic modulus.") The envelope curve corresponds to that for simple compression. | | The results from tensile and compressive tests which were conducted by Losty and Orchard [1962] and Seldin [1966a] corroborate these find- ings concerning the hysteresis loops and the character of the envelope curve. The studies made by these investigators will be discussed shortly. In the case of test type ¢, each specimen was subjected to 12 cycles. During the first cycles the total deformations at the unloading and re- loading points increased with increased cycle number, but after the sixth cycle these deformations were essentially constant. The form of the hys- teresis loop remained constant as did the paraelastic modulus. There was a slight increase in the apparent density during cycling. 23 Using a stress-strain diagram corresponding to test type b, Arragon and Berthier discovered that straight lines drawn through the unloading point and the point at zero stress for each cycle converge at a single point. (Using the curves from compression tests given by Losty and Orchard* and by Seldin, one finds additional support for this finding.) The coordinates of this point are both negative (taking compressiofi as positive), and it was reasoned that the existence of this point is a manifestation of the history of the virgin specimen. ( During manufac- ture, differential thermal expansion causes internal stresses that aré probably quite high, as explained later.} The envelope stress-strain curve was identified as a segment of a hyperbola with asymptotes paral- lel to the stress and strain axes. , For the tests-of type b, conducted by Losty and Orchard, the speci- mens were machined from extruded graphite stock material. The curves obtained are all convex toward the stress axis on loading and concave toward this axis on unloading. The relocading curves in both tension and compression asymptotically approach the envelope stress-stfain curve after each cycle of loading, unloading, and reloading in the same manner as shown by Arragon and Berthier (see Figs. 6 and T7).T Seldin [1966b] found from his tests on molded graphites [1966a] that, in general, the reloading curve passed through the unloading point. At stresses greater than those at unloading, the reloading curves were coincident with ex- trapolations of the corresponding initial loading curves. Thus, in com- parison, the point of tangency to the envelope curve 1is shifted more toward.the stress axis for these molded graphites than for the extruded graphites tested by the other investigators. ' Losty and-Orchard found the slopes of the curves at low stresses in extruded graphites to be about the same in tension and compression, that: * . Additional discussion will be given shortly. TThe slopes of the straight lines drawn through the hysteresis loops in Fig. T correspond to the paraelastic moduli mentioned above. The ' slopes of the three shorter lines in the loop obtained after prestressing to about 3000 psi and reloading indicate the paraelastic moduli corre- sponding to the stresses of 700, 1500, and 2250 psi on reloading. By extending the straight lines for the four loops, one finds that they es- sentially converge at a single point. 24, 1250 4000 —===LCADING — UNLOADING 750 STRESS (psi) 500 20% 102 0 0.400 STRAIN (%) Fig. 6. Tensile Stress-Strain Diagram for Nuclear Graphite, Parallel to Extrusion Axis (Losty and Orchard [1962]). 3000 me== LOADING — UNLOADING % 2000 e w w ] o }._ ) 1000 59 e o4 . 02 0.3 0.4 05 06 STRAIN (%) Fig. 7. Compressive Stress-Strain Diagram for Nuclear Graphite, Parallel to Extrusion Axis (Losty and Orchard [1962]). : 25 is, Young's modulus was approximateiy_the same'for the two loadings. This was true for both with-the-grain and against-the-grain directions. These authors also measured lateral strains during testing. The strain ratios reported are of the order of 0.10, and the ratio of strain in- duced in the paraliel direction for a perpendiculér (against~the—grain) specifien is a little less than one-half that induced in the perpendiculér direction for all stress levels. The strain ratios for compressive specimens loaded both with-the-grain and against-the-grain were essen- tially independent of stress. Seldin [1966a] conducted a definitive study using eight different grades of molded graphites and employing various uniaxial tests. These include cyclic tests in tension and in compression and cycling between given tensile and compressive stresses. In each case, lateral as well - as longitudinal strains were measured. The transverse, or lateral, stress-strain diagrams obtained have different shapes in tension and compreséion; the diagrams for tension are concave tdward the stress axis while those for campression are‘convex. The transverse-to-longi- tudinal strain ratios are independent of stress, yielding constant vélues, in compression. However, the ratioé in tension are functions of stress, decreasing as the stress is increased. These results for compression are in good agreement with those of Losty and Orchard. Greenstreet et al. [1965] also found from tests on an extruded graphite that ‘the ratios in compression are approximately constant, and the ratios tend to decrease continually with increasing stress in tension. Snyder [1966] also studied the stress-styain behaviors of molded graphites. He used compression and flexure tests, and, in each case," the.program was that of loading to a given stress level, finloading, and reloading to failure, giving a two-cycle test. He obtained elastic mod- "uli, Poisson's ratios, and failure data; his elastic data for ATJI* (the * ATJ graphite is a fine-grained (maximum particle size of 0.006 in.) premium-quality molded graphite made by Union Carbide Corporation, Car- bon Products Division. It is similar to nuclear-grade graphites and is often used as a "standard" in making property measurements and compari- sons. ' 2 26 graphite grade commofi to both studies) are in good agreement with those of Seldin [1966a]. | The transverse, residual strain was positive for all graphites studied by Seldin regardless of whether the load was a tensile or a com- pressive one. This residual strain was slightly greater for a given tensile stress than an equal and opposite compressive stress. Thus, the volume of a specimen pulled in tension and released 1is increased since all linear dimensions are increased. In addition; the longitudinal re- sidual strain was greater in tension than compression. These results were borne out by cyclic tests as well as tests where- in the specimen was subjected to either simple tension or compression. Cyclic tests between equal stresses in tension and compression produced strains in the against-the-grain direction that increased slightly Wifh each . cycle regardless of whether therlongitudinal axis of the specimen was in the with-the-grain or against-the-grain direction. Thus, fatigue failure is possibly associated with this slight increase in strain. The occurrence of cumulative damage with repeated reversed loadings was dem- onstrated by Dally and Hjelm [1965]. Seldin further shows that a test specimen can be made to reproduce its original stress-strain responses in the longitudinal and transverse directions by amnealing it at its graphitization temperature. Provided it is generally true, this is a revolutionary discovery in graphite test- ing, especially in the potential it provides for removing uncertainties in data analyses and interpretations that arise because of the inefiitable variations in graphite properties. - In first approximation, the stress-strain curves are identical in tension and compression. Exceptions to this may be found by comparing‘ the against-the-grain data for some graphites. Seldin found through close inspection of the results from his tests that in each case there is a tendency toward gréater deformation reéistance in compression than in tension. It is often stated that prestressing a graphite specimen in compres- sion reduces the elastic modulus, E, in tension, and, likewise, pre- stressing in tension reduces the modulus in campression. ©See Losty and Orchard [1962] and Seldin [1966a], for example. TIn the case of these 27 writers, the modulus referred to is found from the stress-strain dia- gram, and the changes observed may be explained, in part, as foliows. In view of the demonstrated nonlinearity of the diagram, the slope of the curve obviously depends upon the position along the curve. Considef the schematic stress-strain diagram in Fig. 8 which corresponds to pre- stressing a specimen in tension to the point A, unloading to pcint B, and loading in compression. The slope at point O is fhe same as that at the instant of unloading from A and represents the elastic modulus. Now, the curve from A to B is an unloading curve, and the segment from B to C is clearly a continuation of this nonlinear curve. Therefore, strictly speaking, the only slope along the unloading curve which can be expected to have meaning in terms of an elastic modulus is that at the unloading point, A, and not the one at B. Comparisons for deter- mining changes in modulus due to prestreséing are probably best made by using dynamic modfilus measurements at low strain amplitudes. In regard to dynamic moduli, Jenkins [1962b] has shown through res- onant frequency measurements that prestressing in compression does de- crease the apparent elastic modulus. This investigation was concelved as a result of his fracture studies in which he observed isclated crack- ing at stresses below those required for major fractures. He feasoned ORNL-DOWG 67-7659 m tan™! Eo A tan™! Eo Pig. 8. ©Schematic Stress-Strain Diagram for Explaining Change in Modulus Due to Prestressing. _ 28 that beéause of such isolated cracking the bulk physical and mechanical -properties of the material should change. In particular, the dynamic ‘ modulus of elasticity for very low strains should change. To test this hypothesis, dynamic modulus measurements were made on extruded nuclear graphite specimens after prestressing in compression to progressively higher levels. It is reported that the best fit to the data up to fracture is given by £ = expl-¥0/(0_ ~ o)1 , (1) o where EO = agpparent modulus at zero stress, E = apparent modulus after prestressing to a given stress, k = nondimensional constant, o = stress, o_ = a critical stress. c The data do not exhibit any significant directional dependence for the directions normal and parallel to the extrusion axis. From the data given, E/EO'E 0.70 near fracture, and the fracture strengths ranged from ~3700 to ~L4LT700 psi. It is reported that k was found to have a constant value of about 0.08, and g, varied from 4500 to 5000 psi. . Because of the nature of graphite, that is, the material consists of discrete filler particles embedded in the graphitized binder which forms a more or less continuous matrix with random structure, relation- ships between specimen size and mechanical properties appear likely. Thus, strength versus specimen diameter studies have been made. -Losty and Orchard [1962] investigated specimené ranging in diameter from 0.282 to 0.798 in. Their material was extruded nuclear-grade graphite, and both directions with respect.to-the extrusion axis were examined. The strengths were independent of diameter for the three larger sizes with those for the smaller diameter being Jjust significantly lower. In a similar study using flexural specimens ranging in cross section from a square, 1/4 in. on a side, to a 3/L4 in. squaré, no significant effect of specimen size on flexural strength was found. 29 Digesu and Pears [1965] used ATJ graphite specimens of four dif- ferent diameters ranging from 0.063 to 1.0 in. The ultimate tensile strengths for the three smaller sizes were about the same, but the strength fbr the 1.0-in.-diam specimen was about ll% less. These in- vestigators also found small decreases in strength for a given speéimen with increase in strain rate and with increase in surface réughness} Greenstreet et al. [1965] found that for specimen diameters ranging from 0.128 to 0.750 in. size effects in terms of fracture stresg and strain‘ and elastic modulus were small or nonexistent. The maximum pérticler size for their material was 1/32 in. Flexural fests are commonly used for control purposes in manufac- | ture {Currie, Hamister, and MacPherson [1956]) . Two types are employed. The most widely used 1s that of applying the locad at the center of the beam supported near its ends. The second is that of applying the load at two points so a greatef length of beam is under maximum stress. This is the recommended test (Currie et al.) and is designated as the "third- point loading" testrby ASTM [1964]. 1In either case, thé.strengths are calculated using simple beam theory based on linear elasticity. Accord- ing to the above authors, the second test gives strength values 20 to 35% lower than the single-point loading test. In addition, they report the ratio of tensile strength to flexural strength from single-point loading as about 0.53% with a range of 0.47 to 0.68. Through tests and data analysis in which nonlinear behavior as well as differences in stress-strain properties in tension and compression were accounted for in bend specimens, Greenstreet et al. [1965] made comparisons of bend and uniaxial test results. As expected, the results from bend tests were shown to agree with those from uniaxial tests in regard to stress-strain diagrams and to fracture stress and strain. 'Andrew, Okada, and Woschall [1960] observed creep in end-loaded graphité cantilever beams at room temperature. The creep strain had a logarithmic time dependence indicating the absence of viscous creep, and, over the range of time observed (~1200 min), the creep deformation was small in comparison to the total. Kennedy [1961] conducted uniaxial creep tests on nuclear-grade graphite specimens at temperatures of 78, 750, and 1100°F. At each temperature, there was slight initial creep f 30 which saturated in a few hours and no creep occurred after that for to- tal time durations up to 2000 hours. Similar results were found using flexural tests. However, using more sensitive measuring techniques, .Kennedy [1963] later found that low level creep does persist over time intervals of 2000 hours. Kennedy [1960, 1961] also examined the influence of rate of loading on tensile specimens at temperatures of 78, 750, and 1100°F. He used short-time and incremental loading tests in which the loading rates were approximately 600,000 psi/hr and 2 psi/hr, respectively. There appeared to be no significant differences between the stress-strain curves from the short-time tests at 78°F and those from incremental ldading tests at 78 and T750°F. 1In addition, the results indicated that the fracture stresses and strains are essentially time and temperature independent. Derived Stress-Strain Relationships Starting from microstructural considerations and relating micro- structural aspécts to bulk material behavior, Jenkins [1962a] deduced a méchanical analogy for use in predicting stress-strain behavior in ten- sion and compression; This mechanical analog is a series of equal fric- tion blocks connected by springs with equal stiffnesses. On the basis of this model, initial loading is described by e = Ao + Bo® » (2) where ™ | total strain, stress, 1/E = elastic compliance, W = Q 1l a material constant which characterizes the plastic deformation. Thus, the first term on the right of Eq. (2) represents the elastic strain, and the second gives the plastic strain. The equation for un- loading from a maximum stress, Gm, is 1 e —e=Alo -0 + 280 ~0)?, (5 31 where € is the strain at S Finally,* for relocading from G with g = Gm, the equation is c—¢c =Mo-0) +3Blo-0)% . () Calculated results were compared with loading-unloading-reloading ‘curves in compression, and the relationships wefe found to hold for stresses to 1200 psi. The calculated results were found to duplicate the envelope curve, the residual strain on unloading, and the hysteresis loops to a good degree of accuracy within the range of\applicability established. ' | | Other stress-strain relationships were derived by Hesketh [196L4] and by Woolley [1965]. The equation due to Hesketh is o gs(-ggy), (5 where E is the elastic modulus for a single crystal and ey is the yield strain for a éingle crystal. In inverted form, the expression is 1,3 .36 0 (3 .36\ 16 - 25 Ee 16 5 Eey 3 (% 2_6-_‘2_) +] (6) Y In this form it is evident that the equation of Jenkins, Egq. (2), corre- WO =iq sponds to the first two terms. Woolley based his derivation on a model for dislocation movement and reasoned that‘subgrains become plastic progressively. His relation- ship is exponential in form and is given by *In his paper, Jenkins took the specific case o, = O. In the paper of Jenkins and Williamson [1963], illustrations of stress-strain curves deduced from the model are given for stresses both in tension and com- pression. ' 32 | —¢/ | o = Ee (1 — e eo) . | (7 where E is the elastic modulus of bulk material and €5 is a constant. This equatibn, with appropriately determined eohstants, was shown to fit all points of both tensile and compressive curves for a nuclear-grade graphite. ' _ : Because all three equations, Egs. (2), (5), and (7), give good fits to stress-strain data at low strains, Jenkins [1965] compared them on the basis of slope, dc/de, versus stress. Equation (2) gives a curve convex toward the origin which agrees well witfi the experimental data up to a compressive stress of 2000 psi. On the other hand, the equation of Hesketh gives a cufve éoncave toward the origin, while that of Woolley gives a'straight line, neither of which fit the data. Seldin [1966a] examined the model of Jenkins by comparing-predicted and experimental results for several molded graphites. The lower and intermediate stress level data were used, and the conc¢lusions are as follows. Equation (2) usually gives a good fit to the entire tensile stress-strain curve and to fhe compressive curve up to approximately 60% of the breaking strength. Equations (2) and (3) predict a residual strain of %-Boi for loading to the stress, Gm’ followed by unloading to zero stress. The experimental results show that for tensile tests these strains are proportional to 02, but théy are greater than predicted. In compression; they are proportional to Gn, where n ranges from 1.6 to 1.9, and they are less than predicted. The shapes of the stress-strain curves and widths of the hysteresis loops on release and reapplication of stress as given by the model agree well with observations. Temperature Effects - Pioneering work regarding the influence of témperature upon mechan- ‘ical properties was done by'Malfistrom, Keen, and Green [1951], who sys- tematically studied changes in short-time tensile strength and elastic modulus. They also studied tensile creep behavior at temperatures in the approximate range of 2100 to 2900°C. Subsequently, the temperature dependence of strength has been investigated by Green [1953%], by Wagner, 33 Driesner, and Kmetko [1958], by Martens, Jaffe, and Jepson [1959], by Lund and Bortz [1960], by Martens, Button, Fischbach, and Jaffe [19603, by Digesu and Pears [1963], and by Smith [196ka]. In these tests, com- mercial as well as laboratory-made graphites were used. The résults for the various types are somewhat different, but there are general trends as follows: 1. The short-time strength increases from room temperature tb ap- proximatély 2500°C where the value may range to twice that at room tem- perature. Above this tempéfature the strength decreases rapidly. | 2. The strength increases with bulk density. 3. The elongatiofi at fracture is small at room temperature (on the order of 0.1 to 0.2%) and remains so up to about 2000°C, above which it begins to increasé with a large-increase ocecurring between 2500 and 2750°C. Elongations at rup£ure up to about 4LO% at the higher tefipera- ture were reported by Martens et al. [1960], although the values are usually much less than this. | | 4. The ductility decreases with increasé in strain rate. 5. Graphites are stronger and less ductile when 1baded in tension " in the with-the-grain direction than when loaded in the against-the- grain direction. | | 6. The denéity decreases during tensile testing and increases during comfiresSion testing. | ' | 'Aé an illustration of strain-rate effects, we will summarize the work of Smith [1964a], who made a careful investigation of these effects in terms of tensile strength and rupture elongation on a molded, nuc- lear-grade graphite.‘ The elongations measured by Smith are low in com- parison to values.for other graphites. His tests were made ifi'a helium vatmosphere and in vacuum at strain rates of 0.005, 0.5, and 2.0 (min) 7%, The influence of ameSphere was small, so this factor will be ignored in summarizing the results. At the lowest strain rate, the strength increased by a factor of almost three between room temperature and | 2500°C. The intermediate stréin'rate gave results which differed signi- ‘ficahtly from those at 0.005 (min)~' only at temperatures above 2000°C, where the slope of the strength versus temperature curve was reduced. At 2.0 (min)'l,'the increase in strength over the range from room 34 temperatufe to 2500°C was about 50%. The elongations at rupture were independent of stirain fate up to about 2200°C. There was a distinct in- crease in elongation at 2500°C (1% elongation at 2500°C) at the lower strain rate, but the elongation remained at the 0.1 to 0.2% level for the higher rates. Dynamic elastic modulus measurements (using vibrating beams) were made by Malmstrom, Keen, and Green [1951], by Faris, Green, ahd Smith [1952], by Davidson, Losty, and Ross [1958], and by Lund and Bortz [1960]. Moduli were obtained from the low stress portions of stress- strain curves by Digesu and Pears [1965].- These studies also covered a range of graphite grades. . Although differences exist between various graphites and betfieen dynamic values and those obtained from stress- strain curves, there is again general agreement between the results. The overall trend is an increase with increasing temperature to about 2000°C followed by a decrease. As an example, Faris et al. [1952] found an increase over the-rénge from 25 to 2000°C of about 50% in the with- 'the-grain direction and about.90% in the against-the-grain direction for one grade of extruded material. -The moduli also increase with increase in density. Mrozowski [1956] pointed out that "frozen-in" stresses are respon- sible for changes in short-time strength with temperature. These stresses arise from the highly anisotropic nature of the graphite crys- tal for which he took thermal exXpansion cbefficients of 1 x 107® and S35 X 107® (°¢c) L as average values over the range from O to 2500°C in the direction of the basal plane and perpendicular to this direction, respectively. This anisotropy leads to differential shrinkages as graph- ite cools from thé graphitization to room temperature. He reasoned that on cooling from the graphitization temperature to 2300-2500°C the stresses induced are relieved by creep, and he did not expect any appre- ciable plasticity below 2000°C. (These views are supported by tests conducted by Smith [196Lb]). Thus, large stresses build up on cooling Ato room temperature. The increase in short-time strength with tempera- ture according to this hypothesis is then due to the gradual release of "frozen-in" stresses on reheating. The drop in strength above 2500°C is due to the onset of rapid creep. 35 A commbn'explanation for the increase in elastic modulus with tem- perature is expansion of the crystallites into voids between them and a consequent tighténing of the structure. However, Davidson and Losty [1958a] suggest. instead that an increase in crystal shear fesistance with increased temperature is the important factor. . Creep The creep of graphite in flexure and torsion was studied by David- son and Losty [1958a, 1958b]. Measurements were made over the tempera- ture range 1000 to 2000°C. The typical behafior was an initial defor- mation followed by a slow, time-dependent éreep.deformation and over the range 1200 to 2000°C was best described by an equation of the form A, =A+k4nt+Bt, (8) where At = total deformation at time t, A = initial deformation, B,k = temperature and stress-dependent parameters. Thus, both transient and steady creep are exhibited in this range. The parameters B and k were found to be linear functions of stress. When the load is removed, the transient component of the creep is recovered according to ' | \At =mddn t . ' (9) Short-time tensile creep was measured in the range 2000 to 3000°C on a number of different commercial graphites by Malmstrom, Keén, and Green [1951] (as mentioned above), by Wagner; Driesner, and Kmetko [1958], by Martens, Button, Fischbach, and Jaffe [1960], and by Seldin [1964]. Wagner et al. also studied compressive creep; creep in flexure was measfired by Seldin and Draper [1961] and by Seldin [1962]. There is not general agreement on the form of the expression for describing the transient creep. The "steady, or secondary, creep rate depends on Gn, whefe values of n ranging from ~2 to ~4 have been reported. The difference ih,reported values may be attributable to differences in 36 materials. The exponent appears to be independent of direction with re- . spect. to grain crientation. For a given stress, the sfieady creep rate increases with increased temperature, and there does not appear to be a minimum corresponding to the peak in short-fime strength. 1In concert with the general tendency - for greater short-time strength in the with-the-grain direction, the resistance to creep is greater in this direction than against-the-grain. Recovery occurs when the load is remcved, but it amouhts to only a frac- tion of the transient creep deformation. | | The creep tests were mainly of fixed-time duration, with the lon- gest time being about 2 hours. Thus, rupture-elongation data usually were not obtained. For those specimens which did fracture, maximum rup- ture elongations of roughly 20 to 50% at the higher temperatufes were observed. Wagner et al. found two types of behavior in compression. In the first, failure occurs during the secondary creep stage, as is the case in tensile creep tests, while in the second there is accelerated creep followed by fracture. Combined Stress Behavior Although a great deal of information is available concerning uni- axial behavior, the literature contains very iittle concerning_combined " stresses. Theoretical analyses haVe been made treating the material as linearly elastic but taking transverse isotropy into account. Two such analyses are given by Weng [1965) and by Witt and Greenstreet [1966]. Ely [1965] tested tubular specimens of graphite under combined in- ternal pressure and axial loads. The specimens were machined from two grades of extruded tube stock to give uniform gage sections 3 1/2-in. leng, 1.0 in. in inside diameter, and 0.060 in. in wall thickhess.‘ Only strengths are reported; these correspond to ten biaxial stress condi- tions. | The biaxial fracture stresses were best correlated by a combination of failure theories. In the tension-tension quadrant for the two-dimen- sional stress space, failure appears to be governed by the maximum nor- mal stress, while in the hoop-tension, axial-compréssion quadrant the - results seem to correlate with an expression due to Stassi-D'Alia [1959]. 37 The latter expression is given in terms of principal stresses by (o —0,)% + (0p — 03)% + (053 — 6,)% . | 2 +2{p = 1o, (o, + 0p + 03) =2p0 , (10) where o, = strength in simple tension, o’ o p = 'OT' 2 o Gé = strength in simple compression. This equation represents a paraboloid of revolution in principal stress space. Note that for p = 1 this reduces to a von Mises surface as used in the mathematical theory of plasticity, that is, an infinite circular cylinder. 38 References Acheson, E. G. [1895]. "Producing pure electric-light carbon,” U. S. Patent No. 542,982. Acheson, E. G. [1896]. "Manufacture of graphite," U. S. Patent No. 568,525._ Acheson, E. G. [1899]. | "Graphite: its formation and manufacture,” Journal of the Franklin Institute, 147, 475-L36. Aikin, A. and C. R. Aikin. [1807]. Dictionary of chemistry and miner- alogy, Vol. 1, London, 300-303. Aikin, A. and C. R. Aikin. [1814]. An account of the most important recent discoveries and improvements in chemistry and mineralogy (appendix to Dictionary of chemistry and mineralogy), London, Oi— 88. American Society for Testing and Materials. [196L4]. "Standard method of test for flexural strength of concrete (using simple beam with ' third-point loading)," 1964 Book of ASTM Standards, Part 10, Phila- delphia, 37—39. Andrew, J. F., J. Okada, and D. C. Wobschall. [1960]. '"Elastic constants and permanent set in carbons and graphite at room temperature, " Proceedings of the Fourth Conference on Carbon, held at the Uni- versity of Buffalo, Pergamon Press, New York, 559-575. Arragon, P. P. and R. M. Berthier. [1958]. "Caractérisation mécanique du graphite artificiel," Industrial Carbon and Graphite, Society of Chemical Industry, London, 565—578. Bacon, G. E. [1951]. "The interlayer spacing of graphite," Acta Crystallographica, 4, 558-561. Bacon, G. E. [1952]. "The reduction of the crystalline perfection of graphite by grinding," Acta Crystallographica, 5, 392. Bernal, J. D. [1924]. "The structure of graphite,” Proceedings of the Royal Society of London, 106, Series A, TLo-T7T73. Biscoe, J. and B. E. Warren. [1942]. "An X-ray study of carbon black " Journal of Applied Physics, 13, 364—3T1. Blackman, L. C. F. [1960]. "Properties of graphite," Research, Applied in Chemistry, London, 13. [1960a]. Part I. "Preparation, structure, and mechanical prop- erties," 13(10), 390-397. 39 [1960b]. Part II. "Thermal and electronic properties," 13(11), Yy1-LL8, [1960c]. Part IITI. "Crystal compounds and irradiated graphite," 13(12), Lo2-502. Carré, M. F. [1877]. "Sur la fabrication de conducteurs en charbon, pour la lumiére électrique," Comptes Rendus hebdomadaires des séances de 1'Académie des Sciences, 8L, 346—3L8. Castner, H. Y. [1896]. "Anode for electrolytic processes," U. S. Patent No. 572,L72. | Currie, L. M., V. C. Hamister, and H. G. MacPherson. [1956]. "The pro- duction and properties of graphite for reactors," Proceedings of the International Conference on the Peaceful Uses of Atomic Energy, Geneva, 1955, 8, United Nations, New York, 451473, Dally, J. W. and L. N. Hjelm. [1965]. "Proof-testing graphite and its design implications,” Journal of the American Ceramic Society, U8, 338—-3)1 ., Davidson, H. W. and H. H. W. Losty. [1958a]. "Elastic and plastic prop- erties of carbon and graphite,'" Mechanical Properties of Non-Metallic Brittle Materials, edited by W. H. Walton, Interscience Publishers, New York, 219—238. ' Davidson, H. W. and H. H. W. Losty. [1958b]. "Plasticity of graphite," Nature, 181, 1057-1059. Davidson, H. W., H. H. W. Losty, and A. M. Ross. [1958]. "The mechani- cal properties of graphite at elevated temperatures," Industrial Carbon and Graphite, Society of Chemical Industry, London, 551-559. Davy, J. [1839]. "Vol. II — Early miscellaneous papers, 1799-1805," The Collected Works of Sir Humphry Davy, Smith, Elder and Co., London, 150—-15k, 211-213. Davy, J. [1840]. "Vol. IV — Elements of chemical philosophy," The Col- lected Works of Sir Humphry Davy, Smith, Elder and Co., London, 109-111. Digesu, F. J. and G. D. Pears. [1963]. "Some parameters of the mechani- cal behavior of graphites" (paper presented at the American Ceramic Society Meeting, Los Angeles, California, October 1963), Southern Research Institute Report. Batherly, W. P. and E. L. Piper. [1962]. '"Manufacture," Nuclear Graph- ite, edited by R. E. Nightingale, Academic Press, New York, 21-51. Bly, R. E. [1965]. "Strength of graphite tube specimens under combined stresses," Journal of the American Ceramic Society, 48, 505-508. 40 Faris, F. E., L. Green, Jr., and C. A. Smith. [1952]. "The thermal de- pendence of the elastic moduli of polycrystalline graphite," Journal of Applied Physics, 23, 89-95. Fitzgerald, F. A. [1897]. "The manufacture and dévelopment of carborun- dum at Niagara Falls," Journal of the Franklin Institute, 143(2), 81-96. . Franklin, R. E. [1951]. "The structure of graphite carbons," Acta Crystallographica, 4, 253P61. . T Green, L., Jr. [1953]. "High-temperature compression testing of graph- ite," Journal of Applied Mechanics, 20, 289-29L. Greenstreet, B. L., J. E. Smith, F. J. Witt, G. T. Yahr, S. E. Moore, and J. G. Merkle. [1965]. "Room-temperature mechanical properties of EGCR-type AGOT graphite,” USAEC Report ORNL-3728, Oak Ridge National Laboratory . Gwinn, G. R. [1949]. '"Graphite," Industrial Minerals and Rocks {Non- metals Other Than Fuels), 2nd ed., American Institute of Mining and Metallurgical Engineers, Committee on the Industrial Minerals Volume, The Institute, New York, 417-U435, Hesketh, R. V. [1964k]. "Crystal anisotropy in graphite and uranium," Journal of Applied Physics, 35(12), 3604—3606. Hinckley, A. T. [1921]. "The development of the modern furnace elec- trode," Electrical World, T78(26), 1263—1265. Jenkins, G. M. [1962a]. "Analysis of the stress-strain relationships in reactor grade graphites,” British Journal of Applied Physics, 13, 30-32. Jenkins, [1962v]. "Fracture in reactor graphlte, Journal of Nuc- lear Materlals, 5(3), 280-286. Jenkins, G. M. and G. K. Williamson. [1963%]. '"Deformation of graphite by thermal cycling," Journal of Applied Physics, 14(9), 2837—28L1. Jenkins, G. M. [1965]. "The stress-strain relationships of polycrystal- line graphite under compression up to fracture,” Carbon, 3(1), 93— oL, Kennedy, C. R. [1960]. "Mechanical properties of graphite," Gas-Cooled Reactor Program Quarterly Progress Report for Period Ending March 31, 1962, USAEC Report ORNL- 295h Qak Ridge National Labo- ratory, 99-101. Kennedy, C. R. [1961]. "Mechanical properties of AGOT graphite," Gas- Cooled Reactor Program Quarterly Progress Report for Period Ending March 31 1961 USAEC Report ORNL-3102, Oak Ridge National Labo- ratory, 116—124 41 Kennedy, C. R. [1963]. "In-pile creep of graphite," Gas-Cooled Reactor Program Semiannual Progress Report for Period Ending March 31, 1963, USAEC Report ORNL-3445, Oak Ridge National Laboratory, 221-225. Kingswood, V. S. [1953]. "Carbon in the engineering and metallurgical industries," Metallurgia, U48. [195%a]. Part I. "Historical introduction and constitution of carbon allotropes," Metallurgia, 48(286), 55-62. [1953b]. Part II. "Industrial carbons,"” Metallurgia, L8(287), 133—-138. | [195%3c]. Part III. "Industrial graphites, diamonds and special mixtures," Metallurgia, 48(288), 169-17k. [1953d]. Part IV. "Refractories, electrodes and other metallurgi- cal uses," Metallurgia, 48(289), 221027, Kirk-Othmer [196L]. Encyclopedia of Chemical Technology, 2nd ed., Inter- sclence Publishers, New York. Liggett, L. M. [196h]. "Carbon (baked and graphitized products, manu- facture) " Kirk-Othmer, Encyclopedia of Chemical Technology, 2nd ed., Interscience Publishers, New York. ~Lipson, H. and A. R. Stokes. [1942]. "The structure of graphlte," Pro- ceedlngs of the Royal Society of London, 181, Series A, lOl—lOS Losty, H. H. W. and J. S. Orchard. [1962]. "The strength of graphite," Proceedings of the Fifth Conference on Carbon, held at Pennsylvania State University, 1, MacMillan, New York, 519-532. Lund, H. H. and S. A. Bortz. [1960]. "High temperature physical prop- erties of molded graphites," Proceedings of the Fourth Conference on Carbon, held at the University of Buffalo, Pergamon, New York, 537546 MacPherson, H. G. [1968]. Oak Ridge National Laboratory, personal commu- nication to W. L. Greenstreet. : Malmstrom, C., R. Keen, and L. Green, Jr. [1951]. "Some mechanical prop- erties of graphite at elevated temperatures," Journal of Applied Physics, 22, 593600. Mantell, C. L. [1928]. Industrial Carbon, Van Nostrand, New York. Mantell, C. L. [1946]. Industrial Carbon, Its Elemental, Adsorptive, and Manufactured Forms, 2nd ed., Van Nostrand, New York. 42 Martens, H. E., L. D. Jaffe, and J. E. Jepson. [1959]. "High-tempera- ture tensile properties of graphites,” Proceedings of the Third Conference on Carbon, held at University of Buffalo, Pergamon, New York, 520542, / Martens, H. E., D. D. Button, D. B. Fischbach, and L. D. Jaffe. [1960]. "Tensile and creep behavior of graphites above 3000°F," Proceedings of the Fourth Conference on Carbon, held at the Unlver51ty of Buffalo, Pergamon, New York, 511-53%0. Mrozowski, S. [1956]. '"Mechanical strength, thermal expansion and struc- ture of cokes and carbons,'" Proceedings of the First and Second Conferences on Carbon, held at the University of Buffalo, Waverly Press, Baltimore, 31-45. Nicholson, W. [1795]. A Dictionary of Chemistry, 2, London. Nightingale, R. E. (editor). [1962a]. Nuclear Graphite, Academic, New York. Nightingale, R. E. [1962b]. "Graphite in the nuclear industry,” Nuclear Graphite, edited by R. E. Nightingale, Academic, New York, 1-20. Nightingale, R. E. [1962c]. "“Structure," Nuclear Graphite, edited by R. E. Nightingale, Academic, New York 37-115. Seeley, S. B. and E. Emendorfer. [1949]. "Carbon (natural graphite),” Encyclopedia of Chemical Technology, edited by R. E. Kirk and D. F. Othmer, Interscience Encyclopedia, Inc., New York. Seeley, S. B. [1964]. "Carbon (natural graphite)," Kirk-Othmer, Encyélo- pedia of Chemical Technology, 2nd ed., Interscience, New York. Seldin, E. J. [1962]. "Research and development on advanced graphite ma- terials, Vol. VI — Creep of carbons and graphites in flexure at high temperatures,”" WADD-TR-61-72, Vol. VI. Seldin, E. J. [1964]. "Research and development on advanced graphite ma- terials, Vol. XVIII — High temperature tensile creep of graphlte, WADD-TR-61-72, Vol. XVIII. Seldin, E. J. [1966a]. "Stress-strain properties of polycrystalline . graphites in tension and compression at room temperature," Carbon, W2), 177-191. Seldin, E. J. [1966b]. Union Carbide Corporation, Carbon Products Divi- sion, personal communication to W. L. Greenstreet. Seildin, E. J. and R. N. Draper. [1961]. "Research and development on advanced graphite materials, Vol. V — Analysis of creep and recov- ery curves for ATJ graphite,'" WADD-TR-61-T72, Vol. V. 43 Shobert, E. I., IT. [1964]. '"Carbon and graphite," Modern Materials: Advances in Development and Applications, edited by B. W. Gonser and H. H. Hausner, 4, Academic, New York, 1-99. Simmons, J. H. W. [1965]. Radiation Damage in Graphite, Pergamon, New York. Smith, M. C. [196ka]. "Effects of temperature and strain rate on trans- verse tensile properties of HLIM graphite tested in helium and in vacuum," Carbon, 1(2), 147-153. Smith, M. C. [1964b]. "Effects of prestrain on the tensile properties of graphite," Carbon, 2(3), 269-27k. Smyth, H. D. [1948]. Atomic Energy for Military Purposes; The Official Report on the Development of the Atomic Bomb Under the Auspices of - the United States Government, 1940-1945, Princeton University Press, Princeton. Snyder, R. D. [1966]. Development of constitutive equations for nuclear grade graphite for space applications 1, NASA-6-65. Speer Carbon Company. [1949]. Black Magic; The Story of Manufactured Carbon, 1899-1949, Speer Carbon Company, St. Marys, Pennsylvania. Stassi-D'Alia, F. [1959]. "Limiting conditions of yielding of thick- walled cylinders and spherical shells," U. S. Army European Re- search Office, Frankfurt, Germany, No. 9851 DU; Contract No. DA-91- 591-EUC-1351, 01-4263-60. Thorpe, Sir Thomas E. [1894]. Essays in Historical Chemistry, MacMillan, London. R ‘ ~ Ubbelohode, A. R. and F. A. Lewis. [1960]. Graphite and Its Crystal Compounds, Clarendon, Oxford. Union Carbide Corporation, Carbon Products Division. [1964]. The Indus- trial Graphite Engineering Handbook, Union Carbide Corporation, New York. Wagner, P., A. R. Driesner, and E. A. Kmetko. [1958]. "Some mechanical properties of graphite in the temperature range 20 to 3000°C," Proceedings of the Second United Nations International Conference on the Peaceful Uses of Atomic Energy, Geneva, 1958, 7, United Nations, New York. Walker, P. L., Jr. [1962]. "Carbon — an old but new material," American Scientist, 50, 259-293. Weng, T. [1965]. "Thermal stresses in anisotropic hollow cylinders," Journal of Basic Engineering, Transactions of the American Society of Mechanical Engineers, 87, Series D, 391-397. bty Witt, F. J. and B. L. Greenstreet. [1966]. "Influence of cross-sectional shape on irradiation-induced stresses in graphite colums,"” Nuclear Science and Engineering, 25, 141-151. Woolley, R. L. [1965]. "The yield curve and the compressive strength of polycrystalline graphite,” Philosophical Magazine, 11(112), 799-807. Wright, M. S. [1956]. "Properties of carbon and graphite at high tem- peratures,"”" High Temperature Technology, edited by I. E. Campbell, Wiley, New York, 92-113. 45. Acknowledgements It is a pleasure to thank Dr. H. G. MacPherson for his generosity in criticizing the Carbon and the History of Graphite section -of the re- port. I also want to express my thanks to the following organizations and individuals for permission to use the material indicated below. Professor P. L. Walker, Jr., and the Society of the Sigma Xi, quotation on page 7 : John Wiley and Sons, Inc., Figs. 1 and 2 Academic Press, Inc., Fig. 3 and Table 1 Society of Chemical Industry, Fig. 5 H. H. W. Losty and Pergamon Press, Inc., Figs. 6 and 7 10. O DO W Ao ;ug:;»su.mgv_o;uguokomb*u<:z&4w_z:w:>namozqumqqobwoewcowmo L ORNL-4327 UC-80 — Reactor Technology Internal Distribution . Baumann . Beall . Beatty . Bettis . Billington lumberg . Butler . Cagle . Carlsmith Chakraborty . Collins Coobs Corum Culler Derby Donnelly Epperson . Fontana Gabbard Gehlbach Godfrey oldman . Gray . Greenstreet . Gwaltney . Hammond . Harms . Haynes . Herndon . Hewette . Jones, Jr. . Kasten . Keilholtz - . Kennedy . Kerr . Larson (Y-12) . Longest . MacPherson . MacPherson . Malinauskas . Marrow (Y-12) . MeClung ooy EHEQEsHt S OEHEZQOCO Oy OEHOEHEHEQESHE OSSR nrdnEHE HY 52. 53, 5. 55 . 56. 57. 58. 59. 60. 61. 62. 136. 137-138. 139-141. 1h2-162. - 163. HOEHO IO PO USUNGOPLURTUNQEPEHNEPGQNE S GHEDOT . McCoy . McHargue . McLain . McLean . McWherter . Merkie . Meyer . Moore . Morgan . Napier (Y-12) . Narten . Neill . Ohr . Overholser Perry . Pratt Ragan . Reed Robbins . Robertson Rosenthal . Sanders oavolainen lsman Stradley Strehiow Sundberg Tallackson Trauger Unger . Valachovice . Weinberg . Weir . Werner . Wichner . Witt . Yahr . Zapp : Y-12 Document Reference Section Central Research Library Laboratory Records Department Laboratory Records, R.C. OPBQ*UQOZU)WW?UPPQCD.EWSOEbbwgmgmmzm*flmfi‘.fibbfilt{i 16k, 165. 166. 167. 168. 169. 170. 171. 172. 173. 17h. 175. 176. 177. 178. 179. 180. 181. 1862. 183. 164 . 185. 186. 187. 188. 189. 190. 191. 192. 19%. 194 . 195. 196. 197. 198. 199. 200. 201, 202. 203. o0k . 205. Z:!D'JUZCDEPF_I‘;E!W'I. ' 3 A %8 L,;_‘ External Distribution S. J. Acquaviva, Army Materials and Mechanics Research Center, Watertown, Mass. C. A. Anderson, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. . W. Babel, Douglas Aircraft Company, Santa Monica, Calif. . L. Bailey, Great Lakes Carbon Corporation, Niagara Falls, N.Y. . W. Behrman, RDT O. Bergman, National Engineering Science Company, Pasadena, Calif. Boltax, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. A. Bortz, IIT Research Institute, Chicago, Ill. Brussalis, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. K. Carlson, POCO Graphite, Inc., Garland, Texas I. Chalfant, Pratt & Whitney Aircraft, Hartford, Conn. E. Crowe, Los Alamos Scientific Laboratory Jack Cully, SNPO-A, c/o USAEC, P. O. Box 5400, Albuguerque, N.M. R. J. Dietz, Los Alamos SC1ent1flc Laboratory D. M. Forney, Jr., Air Force Materials Laboratory, Wright-Patterson Air Force Base, Ohio L. L. France, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. C. W. Funk, Aerojet-General Corporation, Sacramento, Calif. E. A. Henne, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. Harold Hessing, SNPO-A, c/o Los Alamos Scientific Laboratory, CMB Division V. Kachur, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. B. T. Kelley, UKAEA, Culcheth, Warrington, Lancashire, England "Fred Kirkhart, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. J. J. Krochmal, Air Force Materials Laboratory, Wright-Patterson Air Force Base, Ohio B. F. Langer, Westinghouse Electric Corporation, Pittsburgh, Pa. W. J. Larkin, AEC, ORO ‘ R. G. Lawton, Los Alamos Scientific Laboratory J. J. Lombardo, NASA Lewis Research Center, Cleveland, Ohio H. H. W. Losty, The General Electric Company, Ltd., HIRST Research Centre, Wenmbley, England L. L. Lyon, Los Alamos Scientific Laboratory D. P. MacMillan, Los Alamos Scientific Laboratory - M. Manjoine, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. W. D. Manly, Union Carbide Corporation, New York, N.Y. H. E. Martens, Jet Propulsion Laboratory, Pasadena, Calif. R. L. Maxwell, University of Tennessee, Knoxville, Tenn. J. T. Meers, Parma Research Center, Cleveland, Ohio Capt. W. E. Mercer, III, SYMSE, U.S. Air Force, Norton Air Force Base, Calif. J. L. Mershon, USAEC, Washington R. A. Meyer, Douglas Aircraft Company, Santa Monica, Calif. W. C. Morgan, Pacific-Northwest Laboratory, Richland, Wash. J. E. Morrissey, SNPO, USAEC, Washington R. E. Nightingale, Pacific-Northwest Laboratory, Richland, Wash. C. D. Pears, Southern Research Institute, Birmingham, Ala. 206. 207. 208. 209. 210. 211. 212. 213. 21k, 215. 216. 217. 218. 219. 220. 221. 222, 223, 22k, 225, 226-24L7. 2U8. 249-468, o'v'mfz:mimiz:@ 49/ Pis A -t C. A. Pratt, Air Force Materials Laboratory, Wright-Patterson - Air Force Base, Ohio W. G. Ramke, Air Force Materlals Laboratory, Wright- Patterson Air Force Base, Ohio C. Rowley, Los Alamos Scientific Laboratory 5. Scheib, SNPO, USAEC, Washington C. -Schwenk, SNPO, USAEC, Washington . J. Seldin, Parma Research Center, Cleveland, Ohio A. Shaw, Auburn University, Auburn, Ala. H. Singleton, Westinghouse Astronuclear Laboratory, ittsburgh, Pa. D. Slagle, Pacific-Northwest Laboratory, Richland, Wash. M. C. Smith, Los Alamos Scientific Laboratory H. L. Snider, Lockheed Aircraft Corporation, Marietta, Ga. G. B. Spence, Parma Research Center, Cleveland, Ohio Milt Swope, Aerojet-General, Sacramento, Calif. J. S. Theilacker, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. Norman R. Thielke, SNPO, NASA Lewis Research Center, Cleveland, Ohio Donald Earl Thomas, Materials and Fuel Facility, Westinghouse Astronuclear Laboratory, Pittsburgh, Pa. P. L. Walker, Pennsylvania State University, University Park, Pa. E. 0. Waters, Yale University, New Haven, Conn. Tu-Lung Weng, Parma Research Center, Cleveland, Ohio E. M. Woodruff, Pacific-Northwest Laboratory, Richland, Wash. Gulf General Atomic Library (Attn: Wayne Bell, J. C. Bokros, D. B. Coburn, E. Creutz, A. J. Goodjohn, R. J. Price, D. V. Ragonne, C. Rickard, M. T. Simnad, Librarian — 13 copies) Laboratory and University Division, AEC, ORC Given distribution as shown in TID-4500 under Reactor Technology category (25 copies — CFSTI)