Search Documents

By J. D. Mote, A. Rosen, J. E. Dorn

The effect of strain rate and temperature on the critical resolved shear stress for (1100) [1120] prismatic slip was determined for the intermediate hexagonal phase containing about 67 at. pct Ag and 33 at. pct Al. Whereas the flow stress increased only slightly as the test temperature was first decreased below room temperature, a rapid increase inflow stress was obtained with yet greater decreases in temperature from about 240" to 4°K. The effect of both temperature and strain rate on the flow stress over the low-temperature range could be completely rationalized in terms of the Peierls mechanism when the deformation is controlled by the rate of nucleation of pairs of kinks. A few years ago Mote, Tanaka, and Dorn1 observed that the critical resolved shear stress for prismatic slip of a 67 at. pct Ag plus 33 at. pct A1 hexagonal intermediate phase by the (1100)[1120] mode decreased precipitously as the test temperature was increased from 4" to about 160°K. Their results for both basal and prismatic slip are given in Fig. 1. It is obvious that only the flow stress for prismatic slip is strongly temperature-dependent, whereas the flow stress for basal slip is independent of temperature. If the temperature dependence of the flow stress were primarily due to interstitial impurities, it might be expected that they would influence the flow stress for basal slip as well. Since the activation volume at 115°K for prismatic slip was determined to be about 15b3, where b is the Burgers' vector, this thermally activated deformation process was tentatively ascribed as being due to the Peierls mechanism. But the various analyses of the Peierls mechanism then available as formulated by Seeger2 and also by Lothe and Hirth3 did not agree well with the experimental data. Since that time a more realistic theory for the operation of the Peierls mechanism under conditions where the strain rate is determined by the rate of nucleation of pairs of kinks has been presented by Dorn and Rajnak.4 The present investigation was, therefore, undertaken in order to provide the essential data for a careful check on the agreement between this theory and the low-temperature deformation by prismatid slip of the above-mentioned Ag2A1 intermediate phase. EXPERIMENTAL TECHNIQUE Single-crystal specimens so oriented that the [1120] direction and the normal to the (1100) plane were at 45 * 2 deg to the tensile axis were grown from the congruent melting composition of about Ag2Al for the intermediate hexagonal phase by the

Jan 1, 1964

By G. Derge, Ling Yang, G. M. Pound

The specific conductance and its temperature dependence were measured over the entire composition range of the molten Cu2S-CuCI system. At a typical temperature of 1200°C, 10 rnol pet of the ionically conducting CuCl reduced the specific conductance from about 77 ohm-lcm-l for pure Cu2S to about 32 ohm -1cm -1, and 50 mol pet CuCl reduced the conductance to that for pure CuCI—about 5 ohm 1cm1. The nature of electrical conduction in molten Cu2S, FeS, CuCI, and mixtures was studied by measuring the current efficiency of electrolysis at about 1100°C. The Cu2S, FeS, and mattes were found to conduct exclusively by electrons, but addition of 1 5 wt pet CUS to Cu2S produces a small amount of electrolysis. Addition of CuCl to Cu2S suppresses electronic conduction, and ionic conduction reaches almost 100 pet at a CuCl concentration of about 50 mol pet. These facts are interpreted in terms of electron energy level diagrams by analogy to the situation in solids. RESULTS of electrical conductivity studies on molten Cu-FeS mattes as a function of composition and temperature have been reported.' The specific conductances ranged from about 100 ohm-' cm-' for pure Cu2S to 1500 ohm-' cm-1 for pure FeS. This is in sharp contrast with the low specific conductance of molten ionic salts for which the transfer of electricity is by migration of ions in the field. In general, these ionically conducting molten salts, such as NaC1, KC1, CuC1, etc., have a specific conductance of the order of magnitude of 5 ohm-' cm-'. It was concluded on the basis of this evidence that molten FeS and Cu,S exhibit electronic conduction. Pure molten FeS has a small negative temperature coefficient of specific conductance, resembling metallic conduction, while pure molten Cu2S has a small positive temperature coefficient, resembling semi-conduction. The molten Cu2S-FeS mattes follow a roughly additive rule of mixtures, both with respect to specific conductance and temperature coefficient. Savelsberg2 has studied the electrolysis of molten Cu2S and Cu2S + FeS. He concluded that while molten Cu2S is an electronic conductor, there is some ionic conduction in molten Cu2S + FeS3 owing to the formation of the molecular compound 2Cu2S.FeS and its dissociation into Cu1 and FeS2-1 ions. The present work does not verify his results. Chipman, Inouye, and Tomlinson" have studied the specific conductance of molten FeO and report a high specific conductance, about 200 ohm-1 cm-1 of the same order of magnitude as that found for molten mattes, and a positive temperature coefficient. They interpret these results in terms of p-type semiconduction in the ionic liquid by analogy to the situation in solid FeO.1 imnad and Derne' detected appreciable ionization in molten FeO by means of electrolytic cell efficiency measurements. In order to verify the conclusion that electrical conduction in molten Cu2S and mattes is electronic, and to gain further insight into the structure of molten sulfides, the following investigations were carried out in the present work: 1) The specific conductance, s of the molten system Cu2S-CuC1 was measured as a function of temperature over the entire composition range. As discussed later, molten CuCl is an ionic substance. It was thought that if molten Cu2S were simply ionic in nature, addition of small amounts of CuCl might not have a catastrophic effect in lowering the high conductance of the Cu2S. On the other hand, if much electronic conduction occurs, addition of the ionic CuCl should have a large effect in destroying the electronic conduction. 2) The electrolytic cell efficiency of the following molten systems was measured at about 1100°C in specially designed cells: Cu3; Cu2S + FeS, 50:50 by wt; FeS; Cu2S + CuS, 15 wt pet; Cu2S + CuC1, 5.9 to 46.4 mol pet; and CuC1. This gives a direct measure of the fraction of current carried by ions in these melts. Further, the cell efficiency, extrapolated to zero ionic current, is given by cell efficiency = (s leasile + s elexstronic). [1] s lucile for molten CulS would be expected to be no greater than that for molten CuC1, whose s lonle is about 5 ohm-' cm-1, as will be seen in the following. u,.,,.,.,.......for molten Cu,S is of the order of 100 ohm-' cm-'.' Thus, a large increase in cell efficiency from 0 to values of 10 to 100 pet upon addition of CuCl to Cu2S would indicate destruction of the electronic conductance. Conductance Measurements Experimental Procedure—The apparatus and experimental method were the same as those described in detail in connection with the study of electrical conduction in molten Cu,S-FeS mattes.' A four terminal conductivity cell and an ac poten-

Jan 1, 1957

By W. A. Wood, W. H. Reimann, Maria Ronay

The basic mechanism of fatigue is studied in annealed a brass subjectecl to alternating torsion at room temperature, 100°, 200°, 300°, and 400°C, and in air. It is shown that the slip-zone micro-cracking which characterizes fatigue damage produced by small amplitudes at room temperature is progressively replaced by grain boundary cracking at elevated temperatures, replacement being complete at 400°C. Replacement occurs not because slip activity decreases hut because slip movements at elevated temperatures cease to concentrate in narrow zones and instead disperse. Decrease of amplitude at 400°C through permitting longer lives of specimens before fracture actually causes in- MOST studies of fatigue at elevated temperature have been concerned with data for design. Observations on the basic mechanism of fatigue have been largely incidental, though they have led to the interesting inference that this mechanism has features in common with that of creep.' This paper is concerned primarily with the mechanism itself. Advantage has been taken of mater- crease in pain boundary damage, an anomaly attributed to the greater difficulty at elevated tempwature of starting and propagating a crack. Surface corrosion was most pronounced along slip bands but since at elevated temperatures the slip hands showed no cracking the surface corrosion did not appear to influence onset and early spread of fatigue damage. Similarly the type of fatigue damage produced at room temperature by cyclic strain at large amplitudes, characterized by irregular mi-crocracks at zones between disoriented regions in a grain, was also replaced by mainly grain boundary cracking at elevated temperature, though not so completely. ials and procedures known to show readily how fatigue affects the microstructure of a metal at room temperature, so that direct comparison might be made with what the same procedures show when the material is subjected to fatigue deformation at higher temperatures. 1) MATERIAL This was 70/30 a brass, used extensively in previous studies at room temperature.' Specimens were turned from extruded rod to have parallel test portions 1-1/4 in. long and 7/32 in. diameter, annealed to a grain size -1/10 mm, and electro-

Jan 1, 1965

By L. F. Mondolfo, F. A. Crossley

The mechanism of grain refinement by the addition of small amounts of titanium, molybdenum, zirconium, tungsten, and chromium to aluminum was investigated. The results indicate that the grain refinement is caused by the peritectic reaction which, by transforming the intermetallic compound into crystals of aluminum solid solution, seeds the melt with nuclei above the freezing point of aluminum. THE feasibility of reducing the grain size of aluminum alloy castings by the addition of small amounts of titanium, zirconium, molybdenum, tungsten, columbium, boron, and chromium has been known for a long time, and the use of some of these elements as grain refiners is common practice in the aluminum foundry industry. The purpose of the present investigation was to determine the mechanism by which these additions produce the grain refinement. Asato et al., on the basis of extensive work on copper, antimony, and silver alloys, presented a theory of grain refinement based on the occurrence of the peritectic reaction. Briefly, this theory states that during cooling of the melt crystals of the primary phase form, which react peritectically with the liquid upon further cooling. The peritectic reaction transforms at least partially the primary crystals into crystals of the secondary phase, which then act as nuclei for solidification of the remaining melt. As the peritectic reaction takes place the primary crys- tals break in pieces, and each piece acts as a nucleus. Fine dendritic crystals which can easily break are the most effective for grain refinement. Sicha and Boehm," in investigating the refining effect of titanium on an AI-Cu alloy, concluded that TiA1, crystals serve as nuclei to initiate crystallization, but did not offer further information on the mechanism of refinement. Bonsack in the discussion of their paper suggested that refinement was due to clouds of some solid in the melt, probably TiO² which hindered grain growth. Eborall,' after studying the effect of titanium, zirconium, vanadium, tungsten, columbium, chromium, and boron on some alloys, concluded that the peritectic reaction was not an essential feature of grain refinement. Her conclusion was based mainly upon the observation that under her experimental conditions (sand casting) the Al-Cr alloys, which undergo a peritectic reaction, did not show appreciable refinement. Cibula" investigated the mechanism of grain refinement and reported that two types of grain refinement exist: 1—Grain refinement produced by restriction of grain growth by the concentration gradient mechanism proposed by Northcott.' 2— Grain refinement produced by the presence of particles in the melt, upon which the melt crystallizes easily. He further concluded that very marked grain refinement results from a combination of the two mechanisms. Since Tic was found present in aluminum alloys refined by the addition of titanium, it was identified as the nucleating agent. In the case

Jan 1, 1952

By L. F. Mondolfo, F. A. Crossley

SURFACE effects in the brittle fracture of materials such as glass and in the plastic slip of zinc and cadmium crystals are well known.' Recently, another surface effect has been found for zinc monocrystals. It has been observed that certain normally ductile zinc monocrystals become ex- tremely brittle when their surface is coated with a layer of oxide or copper plate.

Jan 1, 1952

By Nicholas J. Grant, H. C. Chang

Microscopic observations during creep tests were made on AI-20 pet Zn, 80 pet Ni-20 pet Cr, and 25 and 3S aluminum specimens. All these materials failed in an inter-crystalline manner under certain stress and temperature conditions. Particular atten-tion was given to the initiation and propagation of intercrystalline cracks in coarsegrained AI-20 pct Zn specimens. The geometry of the grains and grain boundaries in these specimens was known before the tests. On the basis of the observations and a knowledge of the interaction between grains and grain boundaries, a mechanism of intercrystalline fracture and propagation of the fracture has been proposed. METAL, and alloy fracture occurs essentially in two ways: transcrystalline and intemrystalline. In the past most of the work done in this field was centered on the determination of stress criteria for fracture. The specimens used for these studies were tested under varying conditions of stress state, temperature, and strain rate. Unfortunately, very little attention was given to the initiation and propagation of the fracture, whereas the nature of fracture was examined closely only after failure. Several extensive reviews of these studies have been published by Orowan, Smith, Barrett, and Petch.1, 4 It is well known that a material which exhibits a ductile and transcrystalline fracture can be made to exhibit a brittle and intercrystalline fracture by modifying its composition and by heat treatment. This modification may or may not result in microscopically observable precipitation in the grain boundaries." In the creep of any one alloy the transition of transcrystalline to intercrystalline fracture is a function of stress, strain rate, and temperature." The tendency for intercrystalline fracture to occur increases as the strain rate decreases and the testing temperature increases. Extensive work on high purity aluminum (99.995 pet) under creep conditions showed the importance of the interaction between grains and grain boundaries, and the different responses of grain and grain boundaries to the variables: stress, strain rate, and temperature.7, 8 It was also shown that grain boundary deformation necessarily has to be accommodated in the grains by various deformation means as. for example, fold formation. The type of fracture resulting from the creep of high purity aluminum was invariably found to be transcrystalline. On the other hand, 2S and 3s aluminum were, under certain conditions, found to fracture in an intercrystalline manner." One reason given for this was that the grains could accommodate to a lesser extent the deformation of the grain boundaries.19 It was decided to examine this line of reasoning in greater detail and to investigate the particular interactions between grains and grain boundaries that result in the incidence and progression of inter-ct,ystalline fracture. It is also the purpose of this paper to show where and how intercrystalline cracks start and how they propagate to result in fracture. Several alloys—Al-20 pct Zn, 80 pet Ni-20 pet Cr, 2S and 3s aluminum-— were used for this study, but attention was paid mainly to the behavior of an A1-20 pet Zn alloy, since its general creep behavior has been established." From the results of this study a mechanism of intercrystalline fracture is proposed. Experimental Procedure The experimental procedure was similar to that used in previous work and has been detailed elsewhere.' Two parallel flat surfaces were milled from an originally round specimen, giving a gage portion which was 1x0.09X0.17 in. Some round specimens, 1/4 in. round by 1 in., were also tested. The A1-20 pet Zn specimens. which were annealed at 1030°F for 24 hr, showed two to four grains across the width of the test bar, and most of the grains occupied the whole thickness of the specimenu. Before some of the specimens were subjected to creep, the grain arrangement was mapped out in a development drawing, Fig. 1. The initiation and

Jan 1, 1957

By J. J. Gilman

The dependence of ortho kink-band formation on crystal orientation, on temperature, and on the conditions at the ends of a specimen is described. Load-compression curves for crystals that kink are presented. The reversibility of ortho kinking is demonstrated, and motion of kink planes through crystals is shown. Finally, a phenomenological explanation of ortho kink-band formation is given. ORTHO compressive kink bands (Fig. la) were discovered for the case of metals by Orowan,1 and they have been discussed by Jillson2 and in some detail by Hess and Barrett." However, their history in nonmetallic crystals is considerably older. It is not known when they were first observed in mineral crystals, but descriptions of them may be found as early as 1885.* They were sometimes mistaken for twins.6 Mügge seems to have been first to realize their true nature and importance. He reviewed the knowledge of them for a wide variety of crystals: anhydrite, antimonite, kyanite, vivianite, gypsum, mica, graphite, molybdenite, barium bromide, cal-cite, sodium nitrate, and barite. Also, the "twins" observed by Bridgman7 in sapphire crystals resemble kink bands more than twins. The terms ortho and para kink bands have been coined in order to distinguish between two types of kink bands which seem to represent the same geometrical configurations, but which have different origins and sizes. Ortho kink bands are the usual type that form at low loads when zinc crystals are compressed parallel to the basal plane. Para kink bands form at much higher loads in highly deformed crystals. These terms refer to compression phenomena and are not to be confused with analogous tension phenomena such as tensile kink bands and deformation bands. Mügge described the geometry of ortho kink bands quite completely for kyanite crystals, Fig. lc. The geometry observed for cadmium by Orowan1 and for zinc by Hess and Barrett3 is very similar to that described by Mügge. (See Hess and Barrett for a complete discussion.) Mügge associated kinking with bending combined with basal slip. He pointed out that the kink angles

Jan 1, 1955

By F. D. Rosi, F. C. Perkins, L. L. Seigle

An investigation was made of the mechanism of plastic flow in coarse grained specimens of both sponge and iodide titanium at low (-196°C) and high (500° and 800°C) temperatures. Deformation by slip occurs predominantly on a {1011} plane and in a <1120> direction over this entire temperature range, and secondary slip on {101-1} planes becomes more prevalent with overthisentireincreasing temperature.range, The crystallographic habit of the predominant twin planes becomes more prevalent with type is temperature-dependent, and deformation by twinning increases as the temperature of testing is lowered. Kink bands were not observed at -196OC and rarely at the high temperatures. A LTHOUGH the deform a tional mechanisms in . -titanium at atmospheric temperature have been extensively investigated," there is little information regarding the influence of testing temperature on these mechanisms. The only available data come from the recent work by McHargue and Hammond, who extended single crystals of iodide titanium at a temperature of 815°C. No significant changes in the types of slip and twinning planes were observed, but increased activity on the type I, order 1 pyramidal slip planes and on the type 11, order 2 twinning planes was reported. These planes contribute very little to the deformation process at room temperature.&apos;&apos; &apos; Since there is interest in titanium as a material for use at both low and high temperatures, the present investigation was undertaken to examine the slip and twinning behavior in coarse grained specimens of titanium at -—196", 500°, and 800°C in order to determine the general effect of temperature on the mechanism of plastic deformation. Experimental Material Used and Specimen Preparation—Both sponge and iodide titanium were used in the present investigation; chemical analyses appear in Table I. These materials were arc-melted under argon into 25 g slugs, which were then cold-rolled approximately 90 pct to a sheet thickness of 0.05 in. Since the strain-anneal technique was used for obtaining coarse grains, the cold-rolled sheets were made into tensile specimens which were then annealed to produce a relatively coarse uniform grain structure prior to critical straining. The details of the strain-anneal treatment. size of grains, and the method for preparing the specimen surface for optical microscopy have been described in a previous report.&apos; In order to obtain a wider spread in crystal orientation for the present work, it was necessary to vary the rolling schedule to reduce the degree of preferred orientation in the annealed sheets prior to critical straining. Methods for Tensile Testing—The apparatus for experiments at the temperature of liquid nitrogen (—196°C) consisted of two concentric stainless steel

Jan 1, 1957

By F. D. Rosi, C. A. Dube, B. H. Alexander

The slip and twinning planes have been determined in deformed crystalsof titanium by an X-ray method of analysis. The slip planes are of the type {1010} and {1011}, while the twinning planes are of the type {1072}, {1121}, and (1122). In the case of the predominant (1010) slip, a type I digonal axis of indices <1120> was the effective slip direction. Microscopic observations on the appearance of the three twin types disclosed a distinct difference in their shapes, and an attempt was made to correlate this difference with the amount of twinning shear. The slip mechanism was discussed on the basis of the difference in c/a ratios for the common hexagonal close-packed metals. IT is a general rule that slip in metals occurs most easily on the planes of greatest atomic density and of the largest interplanar spacing, and that the direction of slip is a close-packed direction. Thus, of the hexagonal close-packed metals which have been investigated by the German school in the decade 1925 to 1935 (cadmium, zinc, and magnesium), it has been found that slip occurs on the basal plane (0001) and in a <1120> direction.&apos; In the case of magnesium above 225oC, slip occurs on a first-order pyramidal plane (1011) as well as on the basal plane.&apos; The hexagonal metals, moreover, deform in part by twinning, and, again with the exception of magnesium at elevated temperatures, the twinning plane has always been reported to be a second-order pyramidal plane {1012).&apos; In view of this consistency in the plastic behavior of these metals, it has been supposed that the same slip and twinning elements would be operative in the other important hexagonal close-packed metals (titanium, zirconium, and beryllium), although there are some reasons why this may not be so. Referring to the tabulation in Table I of the axial ratios of the important metals in this structural class, it may be seen that they can be categorized into three distinct groups depending on the value of their c/a ratios. In the first group are cadmium and zinc with large positive deviations from the ideal ratio for closest packing. This expansion in the direction of the c-axis would be such as to accentuate the basal plane as the slip plane, because it increases the interplanar spacing. The second group includes cobalt and magnesium, which closely approximate the ideal ratio. In the third group are zirconium, titanium, and beryllium whose axial ratios are decidedly less than the ideal. The lattice of the elements in this group is compressed along the c-axis which, in effect, tends to make the basal plane less favorable for slip inasmuch as it reduces the interplanar spacing and the atomic density of this plane. Thus, for metals in the third group, it may be said that there is a certain indefiniteness about the glide plane, in that planes other than the basal plane may be expected to operate. The fact that magnesium begins to exhibit planes other than the normal slip and twinning planes at elevated temperatures may indicate that this behavior will become more pronounced in metals of lower axial ratios. For this reason it might be expected that the slip and twinning behavior of zirconium, titanium, and beryllium would be different from that found in cadmium and zinc. One indication that there is a difference is reflected in the recent work which has been done on the deformation textures of these metals."&apos; As will be shown later, there is a significant difference between the rolling textures of cadmium, zinc, and magnesium on the one hand, and zirconium, titanium, and beryllium on the other, and this difference suggests that the mechanism of plastic flow is not the same in these two groups of metals. From these considerations, therefore, it may be seen that there are good reasons why it cannot be concluded that titanium will have the same crystal-lographic elements of slip and twinning as those exhibited by zinc and cadmium. For the complete determination of these elements it would be most advantageous to have single crystals. However, much valuable information can be obtained by using coarse grained polycrystalline specimens, provided the grains are sufficiently large to permit X-ray

Jan 1, 1954

By F. D. Rosi

The slip and twinning behavior in extended titanium crystals were studied in some detail. The formation and appearance of coarse kink bands are discussed. Their crystallographic geometry was determined by X-ray analysis. A phenomenological interpretation of the complexities in kink band development is also presented. The critical resolved shear stress, coefficient of shear hardening, and plane of fracture were determined for several crystals extended at room temperature. THE slip and twinning elements observed in the room-temperature deformation of titanium were enumerated in a previous paper1 in which considerations were advanced regarding the nature and selection of these elements and their effect on the known mechanical properties of this metal. The present study concerns the crystallographic, microscopic, and mechanical aspects of flow in relation to the slip and twinning elements, and includes a prediction of slip systems, nature of slip and twin markings, inhomogeneities of plastic flow, and stress-strain characteristics. Unless otherwise noted, arc-melted titanium sponge (99.77 pct) was used in these experiments. The method of production of crystals, their dimensions, and the surface preparation for micrographic examination have been reported.&apos; For obtaining stress-strain characteristics, only those crystals which traversed the entire width of the specimen and were at least 8 mm in length were used. Tensile deformation of the crystals was performed with conventional grips for sheet specimens and a constant-stress loading beam, designed after the method of Andrade and Chalmers.&apos; The specimens were loaded by allowing sand to flow from a reservoir into a bucket suspended from the longer end of a balanced 6:1 lever arm at a rate controlled to load the specimen approximately 2 kg per min. Strain measurements were made using the Baldwin SR-4, bonded, resistance-wire strain gage, Type A-8, which permitted a reading accuracy of 2 microinches per inch. The formulas used in the evaluation of shear stress and shear strain, as in deriving the coefficient of shear hardening, are given by Schmid and Boasv n terms of the original orientation and change in length of the crystal. For calculating the critical resolved shear stress, the standard equation was used. The crystallographic nature of the unpredictable slip observed in a number of specimens was determined by the single-surface X-ray method of analysis as described in ref. 1. Experimental Results Prediction of Slip System: It was reported&apos; that room temperature slip in titanium takes place predominantly on {10i0} and in a <1120> direction, giving three potential slip systems. Hence, it should be possible to predict the operative primary system in the manner used by Taylor and Elam&apos; for alu- minum (i.e., the one with the greatest component of shear stress in the direction of slip). The stereo-graphic construction in Fig. 1 shows that this is true. In all cases, slip was found to occur initially on the (0110) plane and, in a number of cases, in the [21f0] direction. (The direction of slip was not determined for all orientations.) It follows that duplex slip can be expected when two slip systems are geometrically equally favorable for slip, as demonstrated in Fig. 2. In both crystals slip took place simultaneously on the (1700) and (olio) prismatic planes, as indicated by Fig. 2a and b. In the case of crystal B the movement of the specimen axis with increasing extension was toward the (10i0) direction, which represents the stable end orientation resulting from alternate slip on the [2ii0] and [1120] directions. This is analogous to the duplex slip process in face-centered cubic metals." Unpredictable Slip: In a number of crystals whose orientations were well within the operation of a single slip system, secondary slip occurred on planes not predicted by the criterion of maximum shear stress. Examples are shown in Fig. 3. In addition to

Jan 1, 1955

By J. B. Newkirk

IN 1939 Bitter and Kaufmann1 suggested that iron, precipitating from a copper-rich, Cu-Fe solid solution, appears initially as coherent particles of r-Fe which transform to the body-centered-cubic form of iron when the alloy is cold worked. Since then there has been considerable discussion, supported only by indirect evidence, regarding the validity of the concept. For example, Reekie, Hutchison, and Hether-ington2 interpreted the changes in electrical resistance accompanying the aging of Cu-Fe specimens in terms of a coherent 7-Fe precipitate which transformed to the incoherent a-phase during long aging or cold working. Greninger- ound that Cu-Fe alloys which had been water quenched from 1050" and then held at 650°C for 3 hr changed from weakly to strongly ferromagnetic as the aged wire was subsequently cold worked. He also found diffraction lines corresponding to body-centered-cubic iron in Debye-Scherrer patterns given by aged and worked specimens but lines corresponding to only one face-centered-cubic phase in specimens which had been aged but not worked. Knoppwost,&apos; and later Cech and Turnbull,5 found that aged Cu-Fe specimens showed a slight increase in ferromagnet-ism after they had be? cooled in liquid nitrogen. Very recently Denney6 followed the kinetics of the precipitation reaction as indicated by the saturation magnetization, hardness, and diffuse X-ray effects. The experimental results reported by all of the authors mentioned above are readily explained in terms of the Bitter and Kaufmann concept. However, most of the evidence is indirect when used to define the structural changes involved. It was the aim of the present investigation to determine the mechanism of precipitation of iron from a Cu-Fe solid solution in the light of available published information and with additional new and more direct experimental data. Experimental Method At 1060°C copper will contain up to 3 atomic pct Fe in equilibrium solid solution.7 The solvus slopes rapidly toward lower iron concentration, so that at 600°C only about 0.5 pct Fe remains in solution at equilibrium. Below the eutectoid temperature (850°C), the iron-rich phase in equilibrium with the copper-rich solid solution consists of body-centered-cubic a-Fe. Above 850 °C the equilibrium iron-rich phase is face-centered-cubic r-Fe. The alloy to be studied was made by melting electrolytic iron and OFHC copper by induction under vacuum in an Alundum-lined crucible. The metal was held molten for about 5 min, after which time the power was turned off and the crucible was tipped about 70" to promote directional solidification. The resulting ingot, containing 2.5 wt pct Fe by analysis, was clean, large grained, and free of gross segregation. Slices % in. thick were cut from the ingot, rolled to 1/8 in. in thickness, homogenized in hydrogen at 1060°C for 24 hr and, finally, quenched in cold water. Specimens for all subsequent tests were made from these slices. Squares measuring about 1 cm on a side were cut from the slices to make specimens for studying hardness and microstructure. Wires (0.020 in.) for the Debye-Scherrer survey were drawn from 1/8 in. sq rods which had been cut from the homogenized slices. Large crystals for observing diffuse X-ray diffraction effects were prepared by mechanically

Jan 1, 1958

By G. H. Rowe, A. N. Stroh, D. P. Gregory

The magnitude and variation with strain of the parameters activation volume, V*; activation energy, H; and frequency factor, A, in the Arrhenius equation for strain rate are determined for colunlbium single crystals and polycrgstals. Comparison of dislocation structures in deformed columbium crystals as determined by transmission electron microscopy with the rate IF deformation of a metal crystal is a thermally activated process, rate theory can be used as a powerful tool in the study of dislocation mechanisms. seeger1 develops the following expression for plastic deformation governed by a single thermally-activated theory results suggests thai formation of lattice vacancies at jogs in screw dislocations is responsible for thermally activated deformation of poly crystalline columbium. Single crystals appear to deform by a different mechanism. Work hardening in fully recrystallized fine grain columbium is found to result from simultaneous decrease in activation volume and in the number of mobile dislocations with increasing strain. Work hardening in recovered polycrystals is found to result from decrease in activation volume with strain only. dislocation mechanism. His equation for strain rate is where N is the number of mobile dislocations per unit volume held up at energy barriers, v is the frequency with which the dislocations "try" the energy barriers, b is the Burgers vector, a is the area a dislocation loop sweeps out after overcoming a barrier, V* is the activation volume which will be defined more completely later, and t* is the thermal component of stress equal to (Ta-Ti), where Ta is applied shear stress and Ti is internal stress of the dislocation network17&apos; given by

Jan 1, 1963

By B. Chalmers

THE practical importance of the phenomena of melting and freezing must have been recognized for a very long time. The difference between ice and water, for example, has had a profound influence on the history of mankind and the evolution of society. The possibility of melting a metal and allowing it to freeze in a mold of chosen shape has been an essential ingredient in our mastery of the art of shaping metals, and therefore in the evolution of the: machine age in which we find ourselves. The importance of melting and freezing, as applied to metals and alloys, has been so great, in fact, that empirical solutions have been found for the multitude of practical problems that have arisen. This approach has been so successful that relatively little attention has been directed to arriving at an understanding of the fundamentals of the processes. But metallurgy has come to a stage at which we may expect that some, at least, of the more complex problems that have not yet been solved (or perhaps even recognized) may be handled more effectively by scientific study, theoretical understanding, and logical experimentation than by trial and error. In this lecture, therefore, I propose to describe in outline what I think really happens when a metal freezes. In doing so I hope to explain many of the phenomena which have been observed, and in particular to account for the structures that are obtained in actual ingots and castings. The basic problem, to which this lecture represents a tentative partial answer, is this: a mass of metal, containing known proportions of various elements, is melted, heated to a given temperature, and then allowed to freeze under specified conditions. What will be the "structure" of the resulting metal? The term structure includes: 1—crystal size, shape, and orientations, 2—distribution of chemical elements, and 3-—shape, including cracks, cavities, pores, etc. The Solid-Liquid Interface We will first consider what takes place if a single crystal of a metal in the form of a rod is heated, not uniformly, but so that one end is hotter than the other. If this heating process is continued long enough, the hotter end will eventually melt; we will suppose that the rod is in a containing vessel so that the molten metal does not run away, Fig. 1. When some of the metal has melted, we have some solid, some liquid, and an interface or surface of contact between them. If the source of heat is now removed, the interface will move so that some of the liquid freezes, and if the supply of heat is suitably adjusted the interface will remain at rest. This very simple arrangement allows us to study the basic processes of melting and freezing, and if we fully understand this simple case, we may be able to account for what takes place under practical conditions where the heat does not all flow in the same direction, and where the heat flow is determined not by a controllable source of heat but by the heat capacity and temperature of metal and mold, and by the heat loss from the mold surface. The solid-liquid interface is evidently the region of the greatest interest to us; on one side of it there is crystalline solid, and on the other, liquid. In the solid, each atom has a well defined position, around which it vibrates as a result of thermal agitation. It only leaves this position in the relatively rare event of a "diffusion jump." The liquid is much less systematically organized. The atoms are about as far from their neighbors as in the solid, but the arrangement is much less systematic and is continuously changing. The solid and the liquid are represented diagrammatically in Fig. 2. The average energy of the atoms in the liquid is greater than in the solid by an amount that corresponds to the latent heat of fusion, i.e., the amount of heat that has to be supplied to convert unit mass of solid into liquid at the same temperature. The Two Processes As has recently been shown by Jackson and Chalmers,3 many of the features of the processes of freezing and melting can be understood if it is assumed that a continuous and rapid interchange of atoms between solid and liquid always takes place at a solid-liquid interface." It is necessary to con- sider two distinct processes, that of melting, in which atoms leave the surface of the solid and become part of the liquid, and the converse process,

Jan 1, 1955

By Bernhard Blumenthal

A melting process was developed by which high purity electrolytic uranium crystals can be converted into sound ingots without serious contamination. Careful preparation of the crystals, melting in a high vacuum, and directional solidification led to a metal of better than 99.993 wt pct purity. The metallographic examination showed a substantially clean metal with only residual amounts of a second phase which responded to heat treatment. The density of high purity uranium is 19.05 g per cu cm. FOR the study of the fundamental properties of uranium and its alloys, a base metal of high purity is needed, and an investigation was started at Argonne several years ago to examine various methods of preparing high purity uranium. Early success was achieved by the fused salt electrolysis process in preparing high purity metal in crystal form,&apos; and the problem became that of melting the crystals into ingots without contamination. The electrolytic crystals contained a few ppm of heavy metals, less than 15 ppm C, and large quantities of potassium and lithium in the form of trapped fused salt electrolyte. On melting in vacuo in magnesia or urania crucibles, the metal picked up considerable quantities of carbon and some iron, copper, and silicon but lost most of the potassium and lithium. It was not metallographically clean, and much research work was necessary to determine the optimum melting conditions for minimum contamination. This work has resulted in the development of a melting procedure by which sound ingots can be prepared with only a limited contamination by carbon. In the section entitled "Melting of the Electrolytic Metal," the equipment and processes that have been developed for melting the electrolytic crystals are described in detail. The "Metallography" section is devoted to the metallography of high purity uranium, and the section "Density of High Purity Uranium" contains the first property measurement on high purity uranium—a density determination. Melting of the Electrolytic Metal Apparatus—Melting and Vacuum Equipment: Uranium may be melted in either resistance or induction-type electric furnaces. Resistance heating was found to be preferable to induction heating for melting down the electrolytic crystals, for several reasons: Although induction heating permits fast heating, this advantage cannot be utilized when crucibles with limited thermal shock resistance are used. Temperature control of induction-heated melts is poor, and the stirring effect increases the rate of contamination of the melt from the crucible, slag, or the furnace atmosphere; liquation and controlled solidification are difficult to carry out. Because of these objections, resistance heating was chosen. A 10 kw Globar resistance furnace of low heat capacity permitted uniform heating of a uranium charge to its melting point in 45 min. The furnace was insulated with a removable aluminum shield so that, after melting and removal of the shield, the furnace could be quickly cooled to room temperature. The furnace was equipped with an apparatus for directional solidification; see Fig. la. The crucible was suspended in the center of the furnace; and when the melting was complete, the crucible was slowly lowered into the cold end of the furnace tube. Solidification took place upward from the bottom; and when the rate of lowering was small, a sound ingot free from sink holes and shrinkage cavities was produced. To remove the large quantities of gas contained in the electrolytic crystals, melting in a high vacuum is mandatory. While, from the standpoint of contamination, melting in a highly purified argon atmosphere appeared feasible, lithium and potassium removal proved less efficient under such conditions. Also, a vacuum-melted ingot has a cleaner surface than one melted in an inert atmosphere. The vacuum system must be of very high pumping speed and leak tight so that a vacuum of less than 2x10-7 mm Hg is attainable when the system is pumped down cold. The system must have a high pumping speed, so that the pressure maximum which may rise to 1x10-4 mm during heating can be quickly reduced

Jan 1, 1956

By D. K. Deardorff, Earl T. Hayes

An improved technique is described for the accurate determination of melting points of metals in the temperature range 1500&apos; to 2500°C. The improvements consist of gradient heating and refinements in cavity preparation to obtain true black-body conditions. The melting point of hafnium, as determined by this method, is set at 2222°±30°C. This is over 200°C higher than one of the values reported earlier. Discrepancies in the previously reported values for the melting point of hafnium are explained. The melting points of zirconium and titanium are found to be 1855O±15OC and 1668°±10°C, respectively. Reproducibility of results was ±3OC. PRIMARILY, this report is being made to correct a 1975o±25oC1 value for the melting point of hafnium, and to present an improved method for melting point determination. As a matter of interest, the technique was also applied to determining the melting point of zirconium and titanium. There is a wide difference in the melting point of hafnium as reported by competent investigators. DeBoer and Fast&apos; reported a value of 2230°±500C, and later work of McPherson, as reported by Aden-stedt, placed the melting point at 1975o±25oC. Zwikker" calculated that the melting point was 2430°C. In resolving this discrepancy, improvements were made in the procedure for determining the melting points of refractory metals in the range 1500" to 2500°C. Special techniques are necessary to determine the melting point of these metals accurately. Since the metals oxidize in air above 900oC, the determinations must be carried out in an inert gas or vacuum. Also, every known refractory contaminates them to some extent. In addition, the attainment of true black-body conditions is always difficult. DeBoer and Fast actually determined the melting point of Hf-Zr as-deposited iodide process alloys and extrapolated to 100 pet Hf. McPherson employed a tungsten wire suspension similar to that described by Schramm, Gordon, and Kaufmann,1 and used hafnium metal produced by the iodide process. There was no apparent reason for the wide spread in melting points, since the quality of the metal was very good and the work was under the direction of highly competent investigators. Retracing some of this work revealed one understandable error in the case of hafnium, and led to development of an improved technique for determining melting points of reactive metals in the range 1500" to 2500°C. Material Crystal bar hafnium, zirconium, and titanium were used for this work, with analyses as shown in Table I. The hafnium was of the highest purity that has been prepared, though not as pure as the zirconium and titanium. At the instigation of the Pittsburgh Area Office of the AEC, a special lot of Hfo2 was purified, converted to the tetraflouride, and bomb-reduced with calcium, all at the Iowa State College Laboratories. This crude hafnium metal was then refined by the iodide process at the Foote Mineral Co. All specimens were cut with a Sic wheel, ground, pickled, and dried carefully before use. Procedure General Tungsten Suspension Method—In this method, a small specimen is suspended by a fine tungsten wire in a slender, induction-heated tungsten or tantalum cylinder, all under vacuum. The melting point is determined by heating the specimen slowly while observing it with an optical pyrometer focused upon a portion of the specimen that radiates as a black body. This means that there should be no reflected radiation, either from colder objects near the pyrometer or from hotter surfaces. The induction heater would be the only hotter object and, if it is slender enough, the specimen will attain the same temperature. Often a slender hole is drilled into the specimen as a source of black-body radiation. This was the general method used by McPherson in obtaining his value of 1975°C for the melting point of hafnium. In trying to repeat this work, it was found that the points of contact between the

Jan 1, 1957

By W. Rostoker, H. Nichols

It has been demonstrated in previous work1&apos;2 that wetting of aluminum alloys by liquid mercury can cause fracture to occur with substantial suppression of prior plastic flow. This has been interpreted as the action of a critical liquid species in reducing the cohesive strength of a stressed solid at sites of potential crack nucleation. In this way the critical transverse stress for fracture nucleation is reached at much lower stresses than is normally the case. The stresses for both fracture nucleation and fracture propagation are depressed by the same mechanism. In so doing the plastic-energy component of fracture energy is reduced to a level which is of the same order of magnitude as the surface, interfacial, or cleavage energy. Recent work2 has illustrated the influence of dispersed precipitates, cold work, and combinations of these on brittle fracture associated with plastic strains of the microyield order of magnitude. This has been rationalized in terms of the contribution of internal microstress fields to the stress magnifications at points of obstruction of active slip bands in a stressed metal. The presently reported studies on an Al-4-1/2 pct Mg alloy amplify this subject by introducing thermal-mechanical conditions which can produce strain aging. As amply demonstrated by Phillips, Swain, and borall, the tensile behavior of aluminum alloys containing magnesium reveals a yield-point elongation and successive discontinuous elongations over the whole plastic-strain range. This gives a stepped or serrated character to the stress-strain diagram. The yield-point elongation is not temperature-dependent. On the other hand, the discontinuous extensions in the plastic-strain range are gradually suppressed by lower temperatures of testing. At -78"C, only the yield-point elongation persists. The yield-point elongation itself can be suppressed by rapid quenching to room temperature from an annealing temperature of 500°C. The argument presented by Phillips et 1. and Cottre114 is that the yield-point elongation derives from surmounting at a critical stress the restraint to slip by grain boundaries which have been reinforced by appreciable magnesium-atom segregation. The serrated character of the plastic-strain range is the result of rapid strain aging during the test. This is made possible by the inevitable proximity of magnesium atoms to any dislocation line. In the Al-4-1/2 pct Mg alloy there is one magnesium atom for every twenty-two aluminum atoms and, with a coordination number of 12, a magnesium atom is on the average only about two atomic distances from any given lattice site. Rapid strain aging is therefore a reasonable expectation. A study of mercury embrittlement of the Al-4-1/2 pct Mg alloy (commercial designation 5083) provides an opportunity to relate the interplay of grain size, grain boundary restraint to slip and strain aging in the brittle-fracture process. One lot of commercial sheet material of 0.125-in. thickness was used for all of the tests. The specimen shape and test procedures used were the same as those described in a previous publication.2 A range of grain sizes between 0.0057 and 0.0125 mm was produced by various combinations of plastic prestrain and annealing. The coarsest grain was somewhat elongated in shape. In this case the measurement was taken in the direction transverse to the axis of tension as more meaningful to a brittle-fracture process. In the annealed condition, specimens wetted with mercury failed with zero permanent elongation at stresses slightly below the normal yield point, i.e., stresses at 0.05 to 0.2 pct strain. Since this occurs before yield-point elongation can begin, fracture derives from slip contained within individual grains and restrained from propagating across grain boundaries. No significant difference in wetted fracture stress was observed between specimens which had been water-quenched or furnace-cooled from the annealing temperatures. This signifies that the degree of magnesium segregation at grain boundaries

Jan 1, 1964

By H. P. Leighly

WORNER1 briefly studied the embrittlement of titanium by mercury. He found that mercury will wet the titanium surface at 400°C in vacuo, if the specimen had been heated previously to 700°C to dissolve the oxide coating. Subsequent exposure to the atmosphere causes the mercury film to recede rapidly. Bending of the titanium specimen even after the mercury film has receded results in a brittle fracture. For the study of the embrittling effect of mercury on titanium, specimens 2 by 12 by 0.051 in. were cut from RC-130-A sheet, an alloy having a nominal 8 pct Mn content and a mixed a-ß structure. The specimens were stressed horizontally at a predetermined value after which a plastic ring having a cavity 1/2 in. diam by 1/2 in. deep was placed on the specimen. The cavity was filled with mercury and a hole was drilled in the specimen through the pool of mercury. If the stress level was high enough, fracture occurred instantaneously with the drilling. By repeating the test at different stress levels, the threshold stress can be determined. Mercury embrittlement causes a drastic reduction in the ultimate tensile strength of this alloy from 132,000 psi to about 15,700 psi. The drilling of the hole, in the absence of mercury, does not change the tensile properties. Fig. 1 shows the typical fractures observed as a result of the embrittlement of this alloy by mercury. The central portion of the fracture was brittle represented by the jagged region which appears to have occurred on the shear planes at about 45 deg to the stress axis. Ductile fracture, which is perpendicular to the stress axis, occurred near the specimen edges accompanied by a reduction of the specimen thickness. Examination of the fracture reveals that only in the brittle region does mercury wet the

Jan 1, 1962

By Irving B. Cadoff, Ernest Levine

The crack formation and propagation in the single -phase Cu-4 pct Ag alloys were studied. The alloys were loaded in mercury to various stress levels, the mercury was removed, and the specimen examined for cracks. Cracks were found to develop below the fracture stress; the frequency of such cracks increased with increasing stress level. Some cracks were nmpropagative. Fracture in mercury was found to occur by the link-up of cracks formed at various stress levels rather than by the growth and propagation of a single crack. If the mercury environment is removed prior to a critical amount of crack formation, then continued loading results in ductile fracture. The appearance of the cracks at selected grain boundaries is related to the relative orientation of the boundaries, as are the propaga-tive characteristics of the crack. The mercury interaction appears to be one of lowering the strength of the metal-metal bonds in the high-stress area of the grain boundary. GRIFFITH&apos;S microcrack theory1 proposed a critical crack size above which a crack in an elastic material grows with decreasing energy at a stress of From his theory it was proposed that the presence of a liquid tends to lower the surface energy of the microcrack faces2 leading to a decrease in the critical crack size necessary for spontaneous fracture propagation. stroh3 proposed that the stress concentration at a grain boundary due to pile-up may initiate a microcrack at the grain boundary. petch4 and Stroh5 evaluated the stress distribution at the head of a pile-up in a polycrystal-line material and deduced that the critical crack size and hence of is dependent on the grain size. Experimental verification of this dependence was found by petch6 for hydrogen embrittlement of steel. Studies in stress-corrosion cracking7 have provided a picture of fracture which shows that initial separations occur in a scattered, independent fashion in regions of high tensile stress. A minimum or threshold stress is necessary to produce a sufficient stress concentration to initiate frac- ture. These separations join up to form a crack. The extension of fracture is largely discontinuous and consists of a joining up of cracks. In recent worka evidence of this scattered crack network was found in a Cu-Ag alloy embrittled by mercury. For the Cu alloy-Hg couple, the crack path has also been found to be dependent on the orientation of adjacent grains, and with the addition of zinc to mercury a reduction in embrittlement along with a change in fracture morphology was found.9 In this present study, a mercury-dewetting method was used to observe crack initiation and fracture morphology when a Cu-4 pct Ag alloy is deformed in mercury and Hg-Zn solutions. PROCEDURE Specimens of Cu-4 pct Ag were prepared as in previous crack-path studies.&apos; The specimens were heated at 770°C for 24 hr and water-quenched Tension tests using a table-model Instron were carried out in mercury and in various concentrations of Hg-Zn. Loading was in steps up to the fracture stress, with the load being removed and the specimen examined for surface cracks at each step. The specimens were dewetted after each load to permit examination of the surface structure and rewetted prior to continued loading. The specimens were wetted by electro polish ing in phosphoric acid, rinsing in alcohol, and then immersing in a pool of mercury. Dewetting was accomplished by flame heating the specimen for 30 sec in a vacuum. Some surface contamination was found, but not enough to obscure crack configurations and grain boundaries. RESULTS Fracture Characteristics in Mercury. Fig. 1 is a stress-strain curve showing the progressive step-wise loading of the specimen. As may be seen from the graph, the first position stopped at a is at a stress 5000 psi below the expected fracture stress of 25,000 psi. Examination of the specimen after removal of mercury showed only one crack. The appearance of this crack at a stress far below the fracture stress of this alloy in mercury did not affect the stress-strain curve in any manner. The specimen was then recoated with mercury and deformation was continued (curve b, Fig. 1) raising the stress by 4000 psi, and the same procedure re~eated. The initial crack was located and appeared as in Fig. 2 (crack lb). In this figure the crack is

Jan 1, 1964

By H. D. Mellom

The direction of the optic or "C" axis of a uniaxial metal crystal can be found with the metallurgical polarizing microscope by examining two planes of section on the crystal. Complete orientation of the crystal can be determined if a twin crystal is common to both surfaces of section. Examination of Uniaxial Metals in Polarized Light-Rotation of a uniaxial crystal surface, properly prepared,&apos; shows four extinction positions 90 deg apart, provided that the surface is not close to a basal section. One of these two extinction directions corresponds with the projection of the optic axis on the surface of section. This direction can be found if the &apos;sign of the rotationr2 for the metal is known. To find the direction of the optic axis, the crystal is rotated on the stage 45 deg from an extinction position. With a crystal of negative sign, to produce extinction the analyser must be rotated towards the optic axial direction and vice versa. All sections, other than basal, of a uniaxial crystal have the same sign of rotation and, since there is no change in sign with wave length, white light can be used for these measurements. U the sign of the rotation is not known the presumed direction of the optic axis can be determined assuming a positive sign. A third section cut normal to the presumed optic axis will show basal extinction if the assumption is correct. It is convenient to use the edge between the two planes of section as a fiducial direction common to both surfaces. The optic axis can be related to this direction. If the above measurements are repeated on the other plane of section and the angle between the two section planes measured, stereographic projection can be used to determine the orientation of the crystal optic axis. Determination of Complete Orientation—The complete orientation of a uniaxial crystal can be determined provided that a twin crystal is common to both surfaces of section, and the twin law is known. The orientation of the twin crystal optic axis can be determined in the same way as for the main crystal

Jan 1, 1962

By G. H. Kesler, J. E. Oberele, C. E. Dryden, J. H. Oxley

Heat-transfer rates were measured in a model of a multifilament vapor-deposition bulb fo the preparation of high-purzty metals. Local transfer coefficients for heat transfer frow the filaments to the circuates process stream were determined as a functiol of gas flow rates and bulb, inlet, and filament, geowzetvies. These results a-e then convertted to n mass-t7,ansfer basis to provide a method of correlating vapor-deposition rates and to predict the performance of commercial urzits. The firm1 cor?,elating equation for mass transfer was found to be- wheve k, is the mass-transfer coefficient; D, is bze gas difficsivity : P is the system pressure; R is the gas constant; T is the gas temperature; Df, DB, and Di ave tlze diameters of the filament, bulb, and inlet, respectively; k is the gas conductivity; C, is the specific heat of the gas, M is the wzolecular zleight of the gas; Npr . N,,, NRe a7&apos;e the Prandtl, Schmidt, and Reynolds numbem of the system, respectively; LB is the height of the bulb; and SB and Sf are the total suyiace area of the bulb and filaments, respectively. ONE of the more conventional methods of preparing very high-purity metals is the vapor deposition of the metal upon heated filaments. The reactions which can be employed in such a process generally fall into two classes: 1) Pyrolysis of a volatile metal compound. Generally because of their instabilities, metal iodides are used; however, other metal halides and hydrides have been employed. These processes are frequently carried out under vacuum conditions, and the filaments are maintained at high temperatures. 2) Hydrogen reduction of a volatile metal compound. A metal chloride is normally the preferred feed for these processes, but other halides are sometimes used under special conditions. These hydro- gen-reduction processes are usually operated at atmospheric pressure. Many of the current processes for the production of semiconductor materials are actually vapor-deposition processes. However, the vapor-deposition process itself is relatively old, and almost all metals can be prepared in a state of very high purity by the use of vapor-deposition techniques. Loonam has recently given an excellent review of the iodide process for metal deposition,&apos; and Owen has started some detailed studies with a carbonyl system.&apos; A more general treatment of all vapor-deposition processes has been outlined by Powell, Campbell, and Gonser.3 Despite the fact that vapor-deposition techniques have been employed since 1890, there is a surprising lack of fundamental information regarding the transport characteristics of even the simplest deposition system, the heated filament. The purpose of the investigation described in this paper was to extend some of the earlier data obtained at Battelle to predict deposition rates and the uniformity of deposition in large-diameter, multifilament deposition bulbs under forced convection conditions.4&apos;5 A number of other investigators have already presented a simplified treatment of the deposition process under non-flow conditions, i.e., pure diffusion.6"9 To obtain information on the transport characteristics of a deposition bulb, a model of a typical plant deposition unit was constructed, and local heat-transfer coefficients from the filaments to a circulating air stream were determined under various conditions of air-flow rates and bulb, inlet, and filament geometries. These results were then converted to a mass-transfer basis, and consequently provided a method to correlate experimental vapor-deposition rates and to design commercial deposition units. EXPERIMENTAL WORK Description of Model. The apparatus which was used in this work was a scaled model of one type of deposition bulb. It was constructed of Lucite and was easily altered to give several combinations of diameter of bulb liner, inlet size and location, number of filaments, and outlet geometry. The model is shown in Figs. 1 and 2. The bulb shell was made octagonal for convenience in construction and observation, and the bulb liners were cylindrical. Air, introduced through one of several different jet inlets, was used as the test fluid in these model studies. Filaments for deposition of metal were simulated by metal rods 1/4-in. in diam. Two of the rods could be heated by passing electrical current through them ¦ while the remaining rods could not be heated. This procedure minimized the amount of heat which was

Jan 1, 1962