IL25480A - Shaped articles and their manufacture from low viscosity melts - Google Patents

Description

im»»X 9¾»?3 £3»3¾1?JI3 D"rtX*M c nxw D» S1B 8HAPED ARTICLES AUD THEIR MANUFACTURE FROM LOW VISCOSITY MELTS This invention relates generally to shaped articles produced from melts and their method of manufacture and, more particularly, to the formation of shaped articles directly from free-streaming materials of low melt viscosity, •Sho prooont ο ^¾:ί-β·β>β.·€>?ι-·»θΏβ 4. ) oo-^-c*>?ytj.-y»*a--^4eH-»¾--p¾y^-^-e>¾ir-eqp¾adifig app-1ίο¾ 4-βη-8τ τ-·<· ¾·9β37 Though the following discussion may have particular reference to the production of ilamentary-like articles of an aspect ratio (length/diameter or L/D) characteristically larger than unity, it is to be understood that articles of other shapes, such as tubes, rods, films, ribbons, and wires of any desired size and cross-sectional or superficial configur tion, are as well contemplated where their formation from free-streaming low viscosity melts is attended by problems analogous to those of filamentary formation, insofar as such problems are overcome according to the concepts and practices herein disclosed, or their equivalents .

As here employed, the term low melt viscosity materials generally connotes those materials which, in their molten state, are of such low viscosity as to be considered, according to conventional melt-spinning techniques, excessively susceptible to a surface-tension-driven tendency towards stream breakup and consequent shot formation; i.e., the viscosity is too low compared to surface tension forces to constrain the free-streaming liquid against breakup until it can be caused to solidify in the lengths desired for many applications. directly from a melt has required that the molten material have a rather substantial viscosity, rarely, if ever, less than 500 poises. The present invention is primarily directed towards, and most beneficially applicable to, the many important materials of considerably less, even negligible, viscosity; many, such as the metals in general, possess a viscosity in the molten state approaching the range of only a few hundredths if a poise to several poises.

The minimum viscosity essential to successful melt-spinning by conventional modes is a function of, inter alia, stream size and the physical properties of the melt (particularly surface tension and density) and is, therefore, difficult of any precise definitions; viewed pragmatically, the term "low viscosity" has reference to any material whose interrelationship of viscosity, surface tension and density is such as to thwart filament-like formation by conventional melt-spinning techniques, especially when attempting small diameter production. Thus, the viscosity levels to which the present invention will find beneficial, if not critical, application may range from that of virtually inviscid melts to levels as high as 1,000 poises in the case of melts of low density and/or high surface tension. Taken from another aspect, low viscosity connotes those melts whose free-stream breakup time, as later defined herein, is less than 1,000 times the breakup time of a stream of theoretically zero viscosity, as given by the quantity where p - melt density, D = stream diameter and Y = surface tension of the melt. As a matter of practicality, the term "low viscosity" is here employed as referring to those materials in which the time necessary to effect stream solidification exceeds stream breakup time.

Many workers in the field have expended considerable effort in attempts to answer the long recognized need for methods amenable to large-scale production of filament-like and other shaped articles (many of which would be of novel existence) from materials of such low melt viscosities as to defy any practical application of the conventional melt spinning techniques commonly employed in the spinning of high molecular weight and/or high viscosity glasses and organic polymers. Typical of the materials possessing low viscosity in the molten state are the metals, their alloys and intermetallic compounds; most oxides, sulfides, nitrides, and other salts, and most other inorganic substances and mixtures thereof; also, many organic substances and mixtures thereof; and mixtures of these substances. Materials having significant viscosity in the molten state are glasses, high polymers, materials of large molecular size, and other materials which interact in the liquid state to form liquids similar to glasses.

It will readily be appreciated that many of the low melt viscosity materials, were they economically available in filamentary and other desired forms, would find widespread application, particularly where considerations of strength, modulus, toughness, temperature resistance, electrical and thermal conductivity, chemical resistance, reflectivity and opacity to radiation, are important. Typical of such applications, some unique and some of greatly increased feasibility, would be used in - composites generally, heat resistant fabrics, reinforced metals, plastics and ceramics, brake linings, sintered assemblies, laminates, etc.

Among present day practices, particularly as regards the production of metallic filaments, is that of drawing a relatively large diameter wire or rod through successively smaller dies until the desired diameter is achieved. However, the cost of making wire rises rapidly as the diameter decreases and there is a limit below which such die-shaping procedures cannot be employed with any practical success. Continuous lengths of wire have also been made by ram-extruding metal through a die. However, this method is limited to very ductile materials and diameters too large for many desired applica ions. A third method of making continuous metallic filaments or the like, is that of drawing the wire while embedded in a dissimilar metal matrix to thereby confine the wire under-going drawing against disrupting influences. This technique, as presently known, is likewise expensive and difficult to adapt to large-scale production, particularly when applied to the formation of ultra-fine filaments.

Still another method employed in the production of continuous, small diameter filaments or wire involves vitreousJ the drawing of the molten meta while encased in a v?¾er©trs sheath, otherwise known as the sheath technique.

However, in many applications, the presence of the vitreous sheath is undesired and must be removed, a problem often difficult and expensive of solution. Also, for each metal so processed, there is the problem of selecting a t with the molten metal and which has sufficient viscosity above the melting point of the metal as will protect it from disruption during filament formation. Again, all of these problems combine to render this a costly process.

Still others have attempted a melt-extrusion approach, but, apparently through failure to discern the mechanism of low viscosity free-stream breakup, as well as any practical mode of its suppression or control, have suppositioned, for example, that extrusion must be into a chamber maintained at a pressure exceeding the vapor pressure of the melt, or that the chamber must be maintained at temperatures considerably below the melting point of the extrudant, or that chamber atmospheres reactive to the extrudant are to be avoided. As will more clearly appear hereinafter, it has been ascertained that no such criteria are essential to successful free-stream melt-extrusion of low viscosity melts when carried out according to the teachings of the present invention.

Although the above referred to practices are directed primarily to the production of filamentary-like articles from metallic materials, fully analogous problems and proposed solutions are encountered in the handling of other low melt viscosity materials, such as those already mentioned.

In light of the limitations attending such present-day practices, the need becomes quite apparent for a new approach to the production of filamentary and other shaped articles from materials of low melt viscosity which would be amenable to economical, large-scale production. An ap roach which would seem to fulfill these requirements would be that of extruding filaments directly from the melt as free streams. This approach, as already indicated, has until now, been thwarted by virtue of the fact that small-diameter jets issuing from low-viscosity melts tend to experience severe disruption and disintegration within a short distance so that material traveling from the orifice reaches the end of the continuous portion of the jet prior to the time in which it is possible to effect such solidification of the jet stream that it may be continuously self-sustaining. In that it has been observed that the unbroken jet length is a function of jet diameter, it is possible that certain of these materials could be melt-extruded in relatively large diameter, rod-like forms wherein the sheer size of the extruded body affords a sufficient resistance to disruptive forces and, therefore, a sufficient length of time for the material, during its traverse through the length of the jet, to undergo at least partial solidification prior to stream breakup. However, in the smaller diameters necessary to many desired applications, the jet disintegrates so rapidly that the traverse time of the material through the length of the jet is insufficient to allow freezing into forms of useful length. The period of time elapsing between initial issuance of the stream and its breakup will, of course, depend upon, inter alia, the physical properties of the liquid, particularly its viscosity, density and surface tension. The physical properties of many desirable materials are such that they are incapable of being solidified prior to traversal of the jet length, especially when the diameter does not exceed at least several millimeters, or even centimeters. It is unsurprising, therefore, that the process of free-stream melt spinning by the formation of a liquid jet in which the material is solidified or frozen to form an article of the desired shape and dimension has been successfully applied only with regard to relatively viscous liquids such as polymers and glasses. With liquids of low viscosity, however, the material issues from the end of the jet stream in the form of drops or shot before it can be frozen in its extruded form.

The understanding of liquid jet breakup essentially originates from the fundamental studies published by Lord Rayleigh in 1878 (see "Theory of Sound", Vol. 2, p. 3 lffj First Amer. Ed., Dover Publ., New York, and has been developed to the point of modern applications in such diverse fields as metallurgy, emulsification, fuel injection, and rocket engine behavior.

In aid of a more complete understanding of our invention, it is deemed appropriate at this point to discuss briefly the theory believed to account for the breakup of free-streaming liquids of low viscosity, but it is to be understood that the successful operation and novel results wrought in the practice of our invention, both as regards its process manipulations and the products thereby obtained, are in no manner intended to depend upon the validity of the theoretical discussion here set forth or upon its manner of application.

Liquid jets are subject to breakup by the action of forces which result initially from normally unavoidable stream disturbances in the form of turbulence, limitation that any liquid jet is unstable with respect to its surface energy. Breakup of the jet is enhanced by such surface energy or surface tension and is resisted by the inertia of the material and its viscosity. In spinning fibers from low-viscosity melts, stream instability due to "varicose breakup", as below defined, must be overcome. At higher spinning velocities, stream instability in the form of "sinuous breakup" will occur. Also, even where the varicose and sinuous breakup mechanisms are suppressed to the point that the stream becomes self-sustaining through at least partial solidification, aerodynamic deceleration of the partially or completely solidified stream may result in undeslred distortion or disruption of the extruded mass; the lower the density of the extruded melt and/or the higher the density of the atmosphere into which the melt is extruded, the greater the risk to distortion by such deceleration. It is the varicose breakup mechanism, however, which constitutes the first limit upon jet length and it is this mechanism which must, at any given spinning velocity, be overcome if the jet length productive of articles exhibiting high aspect ratios is to be established. As will be more fully related to, sinuous breakup and aerodynamic deceleration become more important factors at the higher spinning velocities.

Varicose breakup is believed to result from the surface-tension-driven tendency of slightly attenuated portions of a liquid cylinder to further attenuate. The surface tension gives rise to localized pressures within the cylinder, which pressures are, quite significantly, a roximatel ro ortio to t c re o h surface of the cylinder. A liquid cylinder of perfectly cylindrical conf guration would, theoretically, be stable; unavoidable variations in diameter along the cylinder, however, give rise to pressure differentials that result in transfer of material from regions of smaller diameter to adjacent regions of larger diameter. Once established, this transfer of material progresses with increasing rapidity until the cylinder is severed at the nodes thus generated. This degeneration of the cylinder is not limited to the point of initial diameter variation, but radiates in both directions, each enlarging region receiving material from both directions and each attenuating region expelling it in both directions. In a jet stream, it will be apparent that the time which has been available for disruption of the stream will increase with increasing distance from the point of jet origin. Thus the jet will initially appear smooth, some variation in the jet diameter becoming apparent further along the stream, and, at the end of the jet, the liquid appears as a series of drops or shot.

The causes of the initial disturbance which generate the incipient and, ultimately, disruptive variations in stream diameter can be traced to numerous minor turbulences within the stream, action of turbulent gases around the stream, external vibrations, interaction between the jet liquid and the face of the extrusion orifice and the like. The time from emergence of material at the stream origin until its varicose breakup at the end of the stream is thus limited. The effect of viscosity is to hinder the oca rowth of i t ba ce a well as its wave propagation. With relatively low melt viscosity materials, such as molten metals, however, the degenerative effects occur substantially unretarded and with extreme rapidity.

Further development of Lord Rayleigh's analysis has demonstrated that the length (L) of the continuous portion of a free stream of low viscosity, before serious varicose degeneration occurs, is approximated by the relation - L = KV^D3/y Where L = continuous portion of the stream (cm) V = stream velocity (cm/sec) D = stream diameter (cm) γ = surface tension (dynes/cm) f = density (gms/cm3) K = proportionality constant.

Thus, it becomes evident that the length, L, of the liquid stream decreases with increasing surface tension. Because the length of the stream is approximately proportional to the velocity, no appreciable increase in the time required for material to traverse the jet length is gained by higher velocity. That is, the "breakup time" (that time elapsing between the entry of the material into the jet source or origin and its passage through the length of the jet to the point of breakup and consequent shot formation) is substantially constant over the optimum range of extrusion velocity and decreases on either side of this range; an increase in extrusion velocity, at least up to a point, therefore results in an increase in the jet length, but has little effect on breakup time and, therefore, little effect upon the brief interval of time which is available to effect solidification of the jet. In the case of low viscosity melts, particularly those having a relatively high surface tension and/or low density, this time interval becomes so brief as to render jet solidification prior to breakup virtually impossible at the heat transfer rates conventionally attainable. Thus, even at high velocities of extrusion of low viscosity materials, the usual result is shot formation rather than filamentary lengths of high aspect ratios.

It is, of course, necessary to either freeze the material in the stream, or in some i^ray stabilize it within this continuous region in order to produce filamentlike shapes having aspect ratios sufficiently high to be of interest for many possible applications. As can readily be understood, because varicose breakup in low viscosity streams occurs so soon after emergence of the material into the stream, the amount of heat it is possible to transfer in this short period of time from relatively small diameter streams renders the possibility of their being stabilized in continuous form by freezing remote, if not impossible.

In summary, then, if the stream velocity of a low viscosity material is too low, surface-tension-driven amplification and propagation of normally unavoidable (though initially minor) stream disturbances prevents the formation of an efficient jet. At intermediate velocities the jet is disrupted by varicose breakup. With increasing velocity, sinuous breakup and aerodynamic deceleration, in which the stream becomes contorted by interaction with the atmos here become im ortant se d turbances are resisted by stream inertia and viscosity, but the viscosity of many desired materials is negligible to the point that undue breakup of the stream normally occurs before it can be solidified or frozen in the lengths desired.

With the foregoing problems and prior art limitations in mind, it therefore becomes an object of our invention to provide a method for the manufacture of shaped articles by free-stream extrusion of materials having such low viscosities as to normally be incapable of sustaining high aspect ratio configurations pending their solidification.

A further object is the production of novel articles of extrusion from materials having a low melt viscosity.

A further object is the manufacture of articles having a high aspect ratio from materials of low melt viscosity by the efficient retardation and suppression of those forces normally disruptive of the free-streaming melt prior to solidification.

Yet another object is the provision of shaped metallic articles exhibiting a reduced dendrite spacing as compared to conventionally shaped articles.

Still another object is the production of metallic articles which may be rapidly homogenized by conventional heat-treating techniques.

According to the present invention, the foregoing and still other objects are attained in the provision of a method for the manufacture of shaped articles by the free-stream-melt-extrusion of materials having such low melt viscosities as to normally experience disruption accomplished by the discovery that, by causing such materials to be jet-extruded as free streams into controlled atmospheres chosen according to the present invention, there may be effected a rapid, almost instantaneous formation of a rigid or viscous jet-stabilizing film about the liquid jet material which serves to prevent breakup pending solidification. Heretofore, there has been an apparent failure to appreciate what is believed to be the mechanism by which low melt viscosity streams suffer undue breakup. As applied to melt extrusion operations, any practical solution has, therefore, until now escaped detection. Having properly characterized the mechanism of jet breakup, it has now been discovered that it may successfully be suppressed by the generation of a stabilizing film of minute thickness about the nascent stream prior to its breakup and pending its solidification by normal heat transfer phenomena.

Further, the stabilizing film may be generated by one or a combination of modes, which may be generally characterized as (1) reactive film formation, wherein the surface of the jet material enters into a chemical reaction with the atmosphere of the spinning chamber, (2) decomposition, wherein either or both the surface of the stream and the spin chamber atmosphere undergo controlled decomposition resulting in the formation of a thin film along the surface of the jet (as by pyrolysis) and (3) deposition, as by evaporation or sputtering.

In the deposition mode of film formation, a layer is deposited on the surface from the vapor. Two techniques of film deposition are well known. The first of these is eva oration. In this rocess the material to be coated and the coating material are both placed in a vacuum. Conventionally, the coating material or source is then heated to a temperature such that its vapor pressure is at least 10 Torr (1 Torr=l mm. Hg). Atoms or molecules vaporizing from the source surface then proceed in a straight line. If they strike a relatively cooler surface, they condense upon it to form a coating or layer. The spinning of metal fiber involving the formation of a stabilizing layer in this manner may require several evaporation sources in order that all sides of the stream be coated evenly. This technique is obviously most suitable to spinning of high melting materials where sufficient heat loss by radiation could be accomplished in a spinning chamber of reasonable size.

Sputtering, the second commonly known means of depositional coating, consists of placing the object to be coated and the coating material in a partial vacuum, usually on the order of one to a few Torr. An electrode is placed in the system and a connection is made to the source materials so that it also acts as an electrode.

A high potential is then applied between the two electrodes. Gaseous ions created by the high potential strike the source surface giving sufficient energy to atoms or molecules so they escape into the vapor phase, thus supersaturating the vapor with the molecules of the coating material.

The supersaturated source material in the gas phase then coats fairly uniformly on all objects inside the enclosure. Thus a single source is suitable for sputtering. In this case, since there will be a pressure of one to several heat loss to occur by convection and the process is extendable to materials of lower melting points. The sputtering technique also provides alternative possibilities. For example, sputtering in an oxygen atmosphere could result in oxidation of the sputtered material. Nickel sputtered onto a stream in an oxygen atmosphere, for example, forms a layer of nickel oxide to stabilize the stream. Other examples are obvious.

By whatever mode formed, it has been discovered that the film material should possess certain attributes relative to the jet material if optimum results are to be achieved.

Preferably, the solubility of the stabilizing film or layer in the molten stream material should not exceed 10 by weight of the stream material at temperatures between the melting point of the stream material and the desired extrusion temperature. The present invention comprehends as well, however, those film-stream combinations wherein the interrelationship between film solubility, film formation rate and diffusion rate is such that film solubilities x^ell in excess of 10% can be accommodated to excellent advantage j this follows from the recognition that the film-formation event is necessarily of such rapid occurrence that a relatively high rate of film solution in the stream being stabilized may easily be offset by a rapid rate of film formation. In those cases where the rate of film formation is not sufficiently rapid compared to the rate of solution of the film to counteract unduly high levels of film solubility, successful operation may still be achieved by partially or completely saturating the melt prior to, or concomitant with extrusion, with the material of the film or such other material as will reduce film solution in the melt to operative levels.

Also, some melts may best be film-stabilized by the addition of a small amount of a constituent whose decomposition or reaction product possesses a lower diffusion rate and/or solubility in the melt material, as typified in Examples 27 and 29.

In the case of rigid, as opposed to viscous, film formations, the film material must possess a melting point higher than that of the stream material if stabilization is to occur. Again, as out of considerations of film solubility, there may be added to the melt a small amount of a constituent whose decomposition or reaction product possesses a higher melting point than the film material that may otherwise be formed. The same purpose may as well be served by melt modif cations resulting in streams of lower melting points, as for example, the formation of various alloys and eutectics. Where the supporting film composition is non-rigid, but viscous, the viscosity of the film material should be greater than 1,000 poises at the melting point of the stream material.

Further, it has been found desirable to choose the film generating constituents so as to avoid the formation of excessive amounts of products of any reaction or decomposition whose subsequent movement within the stream may cause disruption of an otherwise successfully formed stabilizing film.

Another most important aspect to the practice of our tabili ation roces lies the rovi ion that the velocity of the free-streaming melt, ereina ter referred to as the Rayleigh parameter, should be controlled to lie within such limits that the dimensionless quantity (which quantity is the square root of the well known Weber Number and in which V is stream velocity, D is stream diameter and p and γ are density and surface tension, respectively, of the melt) lies within the range of 1 to 50 and preferably within the rage of 2 to 2 . It has been discovered that, where the velocity of extrusion fails to satisfy this condition, the breakup time of the jet becomes so shortened that effective film stabilization is not established. For a given melt composition of a known surface tension and density extruded as a free stream at a given diameter, the optimum velocity lying within the Rayleigh parameter range of 1 to 50 will normally be determined experimentally, primary consideration being given to the relative density of the melt to that of the atmosphere into which extrusion takes place and the temperature of extrusion relative to the temperature of the spin chamber atmosphere. In general, varicose breakup determines the lower limit of the stated Rayleigh parameter range, while either or both sinuous breakup and/or aerodynamic deceleration (wherein the partially or fully solidified stream is caused to jam upon itself to create a jointed appearance) determines its upper limit. The upper limit of the range is approached as the relative density of the melt to the spin atmosphere increases; i.e. the greater the density of the melt and/or the lesser the density of the spin atmosphere, the higher the Rayleigh though optimum performance may dictate a somewhat lower level .

As a further illustration of the importance of extrusion velocity, two series of runs are set out in Tables I and II below, wherein only the pressure of extrusion, and therefore extrusion velocity, was varied.

In the runs reported in Table I, tin having a density of about 6.8 g/cm^ in the melt and a surface tension of about 573 dynes/cm was heated to 260°C. The melt was then extruded into an atmosphere consisting of 7S helium and 33^ oxygen having a density of 0.6 gms./l. through an orifice 100 microns in diameter, with the following results: TABLE I Spin Spin Rayleigh Pres. Velocity Parameter (Psig) (Cm/sec ) Remarks 316 3.1 Poor jet; no fiber Fiber with many nodules. i 528 5.3 Fiber with no nodules. 631 6.3 Wo joints; no nodules. 25 706 7.1 Very long fibers. (30 ft. ) 65 II38 11.k Fiber lengths up to ft. Fiber had a number of joints. 100 11+10 15 Fiber lengths were up to one foot; number of joints greatly increased.

In the runs reported Table II, liquid lead at a temperature at about 360°C, having a surface tension of about I4.50 dynes/cm and a density of about 10.5 /cm^ consisting of Q0% inert gas and 20% oxygen having a density of 1.28 gms./l., with the following results: TABLE II Spin Spin Rayleigh Pres . Velocity Parameter (Psig. ) (Cm/sec ) Remarks ) 1 - 3.69 No streaming 2.5 I80 5.8 Fiber with nodules 255 8.3 Fiber with nodules 361 11.7 Good quality fiber 510 16.5 Good quality fiber ho 721 23 Good quality fiberj lengths up to 1.5 meters This series of runs show that a minimum pressure, or equivalent velocity in terms of the Rayleigh parameter are required before streaming occurs.

At a relatively low Rayleigh parameter, it is seen that the fiber does not have good quality and exhibits many nodules. In the intermediate range, good quality fiber is obtained in both cases and, finally in the case of tin (Table I), as the upper limit specified is approached, one begins to encounter joints in the wire or filament due to aerodynamic deceleration. In the case of higher density lead (Table II), it is seen that fiber.,quality is on the increase at a Rayleigh parameter exceeding 23 that is, the optimum spinning velocity (or Rayleigh parameter level) increases with increasing melt density and it has been calculated that fiber formation from very high density melts would optimize at a Rayleigh parameter Though it may be theoretically possible to stream a low viscosity liquid without some form of stabilization, it could only be accomplished at stream diameters so large as to be of little practical value. This can best be appreciated by the observation that, from theoretical considerations, it may be estimated that a non- iscous liquid stream might become sufficiently rigid to hold its shape when the solids content becomes higher than about %. In this state, no stabilizing film would be required to maintain the shape of the semi-liquid stream against the varicose breakup normally encountered due to surface tension-surface free energy considerations, i.e., the Rayleig waves previously discussed. From conventional heat transfer considerations, such as latent heat of fusion, stream size, density, melting point, surface tension and heat transfer coefficients, one may estimate the heat loss per unit length which would be required for any given molten material in order to extrude a liquid stream which would freeze sufficiently to become sel -stabilizing against breakup. That is, one may calculate the minimum diameter of such a liquid stream which would be sufficiently stable for a sufficient time prior to varicose breakup for the material to solidify to the extent of 30 , at which point the semi-solid stream becomes self-stabilizing.

The particular value for the minimum diameter liquid stream which may possess this self-stabilizing behavior may vary slightly with variations in; 1. The heat transfer coefficient, which will vary slightly with the densit and com osition of the fluid into which the liquid stream is emerging: 2. The surface tension of the stream, which is obviously dependent on the composition and temperature of the material, though it may also vary slightly depending on the fluid into which extrusion occurs.

Thus, no absolute limit can be drawn except all variables involved be carefully specified, that is, the system itself be completely described. None of the variables involved have a sufficiently large effect, however, to affect the minimum diameter in a gross manner, and given the general spinning conditions and the composition for the extruding stream, close estimates of this minimum size may be calculated which will vary only slightly with relatively sizeable changes in the aforementioned variables. For example, it can be calculated that in order to successfully "freeze-spin" into an inert gas maintained at a temperature of 2$°C. without a stabilizing surface film being required, at least the following minimum stream diameters would be required when employing nitrogen as the spinning atmosphere; TABLE III Latent Heat Surface Melting Minimum of Fusion Tension Density Point Diameters Material (cal/gm) (dynes/cm) (gm c¾^) ( °c · ) (microns) Al 96 860 2. 3 625 1+. 8 X 1< Cu 9 1300 7. 9 1058 2. 7 X Pe 65 I835 7.2 1 10 2.8 X i< Ni 72 1920 7.8 1*1-30 X KC1 7k 90 I.98 751 2.0 X Thus, it is seen that the minimum "self-stabilization diameter" is several orders of magnitude larger than the filament dimensions which can be obtained using a stabilizing atmosphere .

According to another aspect of the present invention, it has been discovered that, when extruding metallic melts in accordance with our novel stabilization process, there are obtained shaped articles possessing novel and unexpected internal structural characteristics of most beneficial significance, particularly with regard to heat treating practices. This is readily understood by considering the fact that, when any multicomponent or less than pure system freezes, the material first to solidify will be richer in one or several of the components or impurities, while the remaining components or impurities will concentrate in the liquid phase. Such solute distribution is inherent in all solidification processes artificially (except those in which solidification is a-rfefcPie-a-Mty forced into a single phase region of the phase diagram of the substructures within individual crystals or grains which take the shape of dendritic arms of the "pure" component separating the minor component or impure regions.

Metal properties, including fracture strength, susceptibility to corrosion, deformability and surface character depend signi icantly on the degree of homogeneity of the alloy. It is therefore desirable in metal fabrication to minimize the inhomogeneity brought on by the micro-segregation of materials during the freezing process. This is usually done by post-heating, or annealing. The annealing process brings about increased thermal diffusion and causes the regions of discontinuity, i.e., the regions of impurity concentration, to diffuse and disperse, yielding a material of improved strength and reduced brittleness.

The efficacy of annealing treatments is dependent on the initial solidification structure of the metal or alloy, i.e., on the size of the internal structures, or dendrites, and on the solute distribution in the structures. That is, homogenization or dispersal of impurities will depend on the extent of microsegregation in the original material, and on the relative distances over which they must be redistributed to achieve a uniform structure. Since redistribution of impurities by thermal diffusion depends on the square of the distances through which redistribution or homogenization must occur, it is an obvious advantage to have a minimum distance between these dendrites or regions of inhomogeneity.

In normal macroscale castings, it is common to find dendrite arm spacings of about 100-1000 microns. It containing minor impurities are shaped by the techniques of this invention, a highly unusual internal structure is obtained wherein the dendrite spacings, which is the visible evidence of the microsegregation of materials, are much reduced over spacings normally encountered in other methods of fabrication, such as casting, and are usually in the range of a few microns. Although the spacings are dependent somewhat on composition and on extrusion conditions, they are usually noted to be about 5-25 microns apart. That is, about 5-25 microns separates the minor component or relatively higher impurity regions from one another. In typical castings, these spacings are normally found to be up to 200 times as large as the equivalent structures of this invention.

A major practical advantage accrues from such novel structures in that after-treatments of the spun metals can be carried out very rapidly compared to the expensive and extremely lengthy heat treatments required by structures in which dendrite spacings are larger.

Since redistribution of these solid impurities by thermal diffusion (annealing) depends on the square of the diffusion distance, dendrite spacings of the order of 5 microns allow practical annealing treatments in l/l+OOth of the time required for a material in which the spacings are 100 microns.

This advantage persists in the metal filaments of this invention compared with metal wire of identical size drawn from larger rods. In the initial preparation of large rods to be drawn down to wires, the original dendrite structure introduced b the ori inal billetin persists to some extent in the drawn wire, although micro-segregation may be somewhat reduced by drawing or other mechanical disruption processes. Thus, dendrite spacings in conventional wires are still several times larger than those of the filaments of this invention, though it is possible by rolling squeezing, or working, to mechanically disrupt dendritic structure and obtain partial homogenization. Since working is less effective than annealing and also requires an extra processing step, it would be desirable to eliminate it.

Heat-treating of the conventionally drawn, unworked wire would still require, therefore, relatively longer times to achieve comparable homogenization of the microsegregated impurities or minor components than would the heat treating of the filaments of this invention.

Standard metallographic etching techniques were used in observing the novel internal structures of the filaments obtained in several of the examples.

Metal filaments obtained from Example 27, composed of lj.06 stainless steel, and having filament diameters of approximately 75 microns were mounted in Trans-optic (a cold-curing epoxy resin mounting material) and successively dry ground and polished, wet ground and polished with 600 mesh silicon carbide, 0.3 micron alpha alumina and 0.05 micron gamma alumina, the latter three steps using conventional polishing cloths. Care was exercised throughout to avoid any substantial heating or excessive working of the samples. The samples were then etched with a standard netallographic etching solution composed of -..O g. CuSO 20 ml. cone. HC1 and 20 ml H for 0 eco d to reveal the internal structure. Typical dendrite spacings ranged from 5 to 20 microns as measured on micrographs of the etched samples, with the majority being separated by -10 microns.

A typical sample of filament was then heated to 970°C. for 1 minutes. Mounted, polished, etched samples showed complete absence of dendrites indicating that homogenization of the microsegregations was complete in this time. Comparable annealing results on large castings are known to take several hours.

Examination of the filaments of Example 31 by mounting, grinding, polishing and etching for 10 seconds with a Keller's solution of 1.0 ml. cone. HP, 1.5 m * cone. HC1 , 2.5 ml. cone. HNO and 95· 0 ∞1 . water revealed average dendrite spacings of approximately 10 microns.

Chromel R filaments of Example 22 were mounted, prepared and etched with solution of ferric chloride in hydrochloric acid/nitric acid solution. Dendrite spacings in the filaments were observed to be approximately I . - 5 microns.

Gold filaments of Example ij.2 were ductile and difficult to grind and polish. Mounted samples etched with an aqueous solution containing 10 potassium cyanide and 10$ ammonium persulfate showed no evidence of dendrites. This was as expected since the gold was extremely pure; i.e. the high purity level resulted in'no discernible microsegregation. Similarly, the high purity zinc fibers from Example 33, when prepared and etched with Palmerton's reagent ( 200 g. CrO^, 15 g. Na2S0^ and 1000 ml. H20 ) , showed no dendritic tructure but on ew rain boundaries in what appeared to be predominately a single crystal structure.

In summary, then, the present invention relates to methods for and products resulting from free-stream melt extrusion of materials having low melt viscosities into selected atmospheres productive of a rapid, jet-stabilizing film formation, whether by reaction, decomposition or deposition, which film is of a composition so chosen or modified as to have a melting point above that of the stream material or, in the case of non-rigid films, a viscosity above 1,000 poises at the melting point of the stream material. Further, for optimum process controllability and product uniformity, the velocity of extrusion should be maintained within the range given by the relationship 1 < V J p D/y *\ 50, preferably escription of an illustrative apparatus which may be employed in the practice of the present invention, reference is now had to the drawing, in which there is depicted a simplified, partially sectionalized vertical view of an induction-heated spinning aparatus.

As there shown, such an apparatus may comprise an induction-heated spinning assembly, generally indicated by arrowed numeral 10, mounted upon an elongated cylindrical catch chamber 12. The spinning assembly comprises a melt crucible which, for the examples which follow, was fabricated from boron nitride or alumina (A^O^), but may be formed from any suitable refractory material which is found com atible i.e. non-reactive with the melt it is desired to process. The bottom surface of the crucible is carefully drilled to provide either a small diameter orifice 16, or a carefully machined hole in which a watch-sized jewel, such as a sapphire, having a small diameter orifice formed therein is seated. Although the crucible illustrated is provided with only a single spinning orifice, production versions would, of course, be provided with a plurality of similar such orifices.

The crucible rests upon a supporting and insulating cylinder l8 of quartz construction, which, in turn, rests upon support plate 20. The crucible thusly mounted may be enclosed by a conventional susceptor 22 which encircles the crucible when it is desired to process non-conductive/non-coupling materials; i.e. materials which cannot be directly heated inductively. The susceptor may be held in place by suitable refractory cording, not shown. The crucible and susceptor are enclosed within a heavy-walled housing 2l of Pyrex or quartz construction. Mounted at the upper end of the housing i is a fitting 26 suitable for connection to source of pressure, which fitting is clamped in gas-tight relationship between upper plate 28 and support plate 20 by means of suitable stud bolts 30· By this arrangement, an inert, pressurized gas may be imposed upon the material being melted within the crucible to effect its extrusion through the orifice. As is well known in the art of induction heating, an induction coil y≥ of suitable electrical characteristics encircles the crucible to supply heat to the contents thereof according to the principles of magnetic induction. nte o d tw and the catch chamber 12 is a support cylinder 3 °f quartz or Pyrex construction which stands upon base plate 36 and bears against the crucible support plate 20. Suitable flexible gaskets 38 are provided between the base plate and support plate connections to assure a gas-tight assembly. Formed in the base plate 36 is a relatively large catch chamber entry port i+0 which is located to be in substantial alignment with the central aperture \Z formed in support plate 20. The catch chamber 12 is provided with connection through which suitable spinning atmospheres may be introduced or withdrawn. Provision is also made for an observation port ij.6 arranged to give a diametrical view across the catch chamber and an access plate ij.8 of relatively generous dimensions to facilitate installation and removal of suitable take-up equipment in the bottom of the catch chamber. At the very bottom of the catch chamber, a capped collection port 0 is provided for product removal.

It is to be understood that the just-described spinning apparatus merely represents a typical assembly which may be employed in the practice of the present invention, which is in no way limited to the details of construction of the apparatus. For example, a resistance-heated spinning assembly could as well be employed in conducting many of the experiments.

Reference shall now be had to the following examples, which are to be taken as illustrative, but not limitative of the principles and practice of the present invention and in which the apparatus represented in the drawing was employed. Unless otherwise specified, extrusion ressures i.e. ressure over the melt are au e and percentages are by weight.

Examples 1-35 demonstrate the suppression of stream breakup by the formation of a stabilizing film via a chemical combination of the extruded molten stream and reactive atmosphere.

EXAMPLE 1 The entire system was evacuated to 0.5 mm Hg pressure and an aluminum alloy having a melt viscosity of 0.03 poise and containing i+.O percent Cu, 0.5 percent Mn, 0. percent Mg and 95 percent Al was melted by inductive heating to 700°C. After melting was complete, argon at 10 psig. pressure was applied to eject the molten alloy through a l80 micron orifice into the vacuum within the chamber below the orifice maintained at a temperature of °C. The stream of aluminum, in the absence of a film-forming atmosphere, rapidly disintegrated and failed to solidify within a fall of 8 feet to the bottom of the spin chamber.

EXAMPLE 2 The experiment described in Example 1 was repeated with the exception that a 100 micron orifice was employed, the vacuum was replaced with nitrogen at 1 atmosphere pressure and the pressure on the melt was increased to 16.5 psig., resulting in an extrusion velocity i6 cm/sec. As in Example 1, the nitrogen did not perform as a film-forming atmosphere and only aluminum shot was formed.

Molten aluminum will react with nitrogen to form aluminum nitride which might then function as a stabilizing film.

This reaction, however, does not take place rapidly enough to stabilize the molten stream; consequently, no fiber formation.

EXAMPLE 3 The experiment described in Example 1 was repeated with the exception that the vacuum was replaced with an atmosphere which consisted of argon at 1 atmosphere pressure. Again, as was the case in the previous experiments where film-forming atmospheres were absent, no fibers were detected and only shot was formed.

EXAMPLE Example 1 was repeated with the exception that the vacuum was replaced with an atmosphere which contained O.OOlj. atmosphere oxygen and 0.996 atmosphere nitrogen. In this case, the major portion of the spun charge was in the form of shot. Approximately 1% of the total charge exhibited incipient fiber formation, indicating that the oxygen concentration was insufficient to establish an effective stabilizing film.

EXAMPLE 5 The experiment described in Example 1 was repeated with the exception that the extrusion pressure was increased to 20 psig and the vacuum in the spin chamber was replaced with pure oxygen at 1 atmosphere pressure; extrusion velocity was approximately 5 cm./sec. The oxygen was introduced after the molten alloy stream had commenced streaming smoothly. It was observed that if oxygen was present below the orifice during melting, the aluminum exposed in the orifice formed an oxide film which plugged the orifice and prevented spinning. The introduction of oxygen into the chamber below the orifice after streamin was initiated caused an aluminum oxide stabilizing film to be formed on the surface of the aluminum alloy stream which prevented breakup of the stream and resulted in the formation of uniform aluminum fibers with a diameter of approximately 100 microns, a tensile strength of 17,000 psi, and an elongation at break of 26.6% The fibers were lustrous in appearance and, when analyzed for aluminum, were found to contain 9i|-.8$ by weight aluminum, thus giving a clear indication that the film formed was extremely thin.

EXAMPLE 6 The experiment described in Example $ was repeated with the exception that the oxygen pressure was reduced to 0.033 atmosphere in the area beneath the orifice. At this reduced spin chamber pressure, only a very small amount of extremely short fiber was obtained, indicating that the oxygen concentration was too low.

EXAMPLE 7 The experiment described in Example 6 was repeated with the exception that the oxygen pressure were increased to 0.067 atmosphere. The extruded molten alloy stream formed fibers similar to those described in Example >. It is thus apparent, in comparing Example 6, that the concentration of the reactive spinning atmosphere must be maintained at levels adequate to assure a sufficiently high rate of film formation.

EXAMPLE 8 The experiment described in Example 1, was repeated with the exception that the extrusion pressure was increased to 15 psig. and the vacuum was replaced with at 1 atmosphere pressure, resulting in the formation of an aluminum nitride film which effectively stabilized stream. The fibers obtained had a diameter of approximately 100 microns and an analysis of the fibers indicated the composition of the alloy was essentially unchanged, indicating that only a thin surface film is involved in the stabilizing process. It is also demonstrated that any inert diluent, in this case argon, can be used in the presence of a sufficient concentration of the reacting atmosphere, in this case ammonia.

EXAMPLE 9 Example 8, was repeated with the exception that the spin chamber atmosphere consisted of 0. 9 atmosphere nitrogen and 0.1 atmosphere hydrogen sulfide and the extrusion pressure was increased to 30 psig. Reaction with the spin chamber gas resulted in the formation of an aluminum sulfide film which effectively stabilized the stream until fiber formation was established. The fibers averaged 85 microns in diameter.

EXAMPLE 10 Example 1 , was repeated with the exception that pure electrical conductivity grade aluminum (99»k-5/° aluminum) was used instead of aluminum alloy, a spin chamber gas consisting of 0. 97 atmosphere nitrogen, 0.015 atmosphere hydrogen and a 0.015 atmosphere hydrogen sulfide was employed and extrusion was carried out under a pressure of 30 psig. Spherical and elongated shot and strings of beads resulted, demonstrating that the hydrogen sulfide concentration was not sufficient to permit successful fiber formation was in the incipient stages, i.e. stream disruption was occurring at a very late stage.

EXAMPLE 11 Example 10 was repeated with the exception that the spin gas was replaced by one consisting of 0.90 atmosphere nitrogen, 0.033 atmosphere hydrogen, and 0.067 atmosphere hydrogen sulfide, resulting in good fiber formation, thus demonstrating that a concentration of 0.067 atmosphere hydrogen sulfide is sufficiently rich to obtain good fiber formation, whereas, in Example 10, 0.015 atmos. hydrogen sulfide was inadequate.

The above series of Example 1-11 demonstrate that aluminum fibers can be successfully spun in an atmosphere that results in the formation of a stabilizing film on the molten aluminum streams. These experiments further demonstrate that although a reaction is known to occur at the spinning temperature employed (e.g. aluminum and nitrogen to give aluminum nitride) it does not always insure successful production of fiber. Example 2 is a good illustration; here, the reaction between molten aluminum and nitrogen to form aluminum nitride is well known, but the reaction rate is obviously too slow to provide a film of sufficient thickness to permit a degree of stabilization requisite to good fiber formation. In contrast, however, the reaction between aluminum and ammonia (see Example 8) is sufficiently fast to provide good fiber formation. In addition, these experiments (see Examples ij., 6, and 10) also show that a minimum concentration of the film-forming atmosphere is necessary and that the quality EXAMPLE 12 Examples 12-16 demonstrates the utility of the sheath-stabilized spinning process in preparing filaments from metalloids. Such materials are extremely difficult to shape by any known means. Example 12 further demonstrates that a boron nitride film is a very effective stabilizing agent.

The equipment of Example 1 was charged with zone-refined boron having a purity of 99.9995$· The apparatus was evacuated to a pressure of 0 microns and the boron charge was melted (m.p. approx. 2300° C). A 50 psig. argon pressure was then applied to the melt and the boron extruded through a 150 micron orifice into a gaseous mixture composed of 90 nitrogen and 10% ammonia maintained at 1 atmosphere pressure. Very long lengths of boron filament were obtained. The filaments were rather lustrous, smooth and uniform, averaging about 115 microns in diameter and having a tensile strength of greater than 100, 000 psi.

Examples 13-16 further demonstrate the necessity for observing the various limitations discussed with respect to reactivity of the atmosphere, solubility limitations of the stabilizing sheath, and the like.

EXAMPLE 13 Utilizing the apparatus of the previous examples equipped with a 125 micron extrusion orifice, there was placed in the crucible a charge of pure silicon in the form of small pellets. The system was evacuated to less than 0.1 Torr pressure and heated to melting at approximately 1-.50°C., at which point the vacuum below the orifice was and 90 ammonia at 1 atmos. pressure. Concurrently, the vacuum above the orifice was replaced with argon at 3 psig. Although the molten silicon began streaming, the chamber atmosphere was ineffective to form a suitable sheath and only silicon shot was obtained.

EXAMPLE IJ4.

As before, silicon was charged to the apparatus and the same procedure followed except the atmosphere into which extrusion took place was pure ammonia and an argon pressure of 80 psig. was used for extrusion.

Somewhat nodular fibers were obtained which had diameters of approximately 120 to 1 0 microns.

It is apparent from this example that a spinning atmosphere of 100 ammonia is just sufficient to give the reaction 3 Si + I. NH3— si3¾_ + to an adequate extent to stabilize.

EXAMPLE 15 The experiment described in Example 13 was repeated with the exception that the molten silicon was extruded into a mixture composed of 20 oxygen and 80$ nitrogen. Instead of fibers, a white, fluffy cottonlike mass was obtained which showed no structural details when examined under the optical microscope at 3 magnifications. The evidence indicated that the reaction of silicon with oxygen took place very rapidly to form oxides of silicon which were, however, ineffective in establishing a stabilizing film due to their relatively high degree of solubility within the molten silicon and/or the high volitility of the oxides. molten silicon into a fluffy, white powder of silicon oxides .

EXAMPLE 16 To further demonstrate the effect of excessive solubility of the film material in the molten stream material, an experiment was conducted in which it was attempted to form a stabilizing film of carbon and/or silicon carbide.

Carbon and silicon will readily form a homogeneous melt at lower carbon concentrations. Thus, neither a carbon film nor one of silicon carbide will normally form to a sufficient extent to perform a stream stabilizing function.

Silicon was once again charged into the apparatus and the procedure of Example 15 was followed, except that the gas below the spinning orifice was changed to propane at 1 atmosphere pressure. On extrusion, only fine spherical shot was produced, indicating that neither carbon from the decomposition of the propane nor silicon carbide, if such was ever formed, was effective in stabilizing the stream.

EXAMPLE 17 Following the general procedure of Example 1 , charged/ beryllium was fSbe&igoit to the crucible, melted and extruded through a 220 micron orifice under an argon pressure of 50 psig. into a spin chamber gas containing argon at a pressure of 0.87 atmosphere. The temperature of the molten beryllium was maintained between 1 , 300 - 1,3$0°C. during spinning. No fiber was formed and only beryllium shot was collected.

EXAMPLE 18 ex er ment d crib d i am wa repeated with the exception that the argon atmosphere was replaced with an atmosphere containing 17.3 oxygen and 82.7 argon. A beryllium oxide film was formed about and effectively stabilized the stream to give good fiber formation. The fiber surface was delustered and gray in color.

EXAMPLE 19 An alloy was prepared which consisted of 9% aluminum alloy and 91$ 1030 steel. The final composition of the alloy was: 89.6 , Fe, 8.6' Al, 0.36$ Cu, 0.77$ Mh, 0.31$ C, 0.23$ Si, 0.0 Mg, 0.026 S, and 0.013$ P.

The alloy was melted in a vacuum at a temperature of 1,500°C, held in the molten state for five minutes, cooled and the surface machined smooth. The alloy was then placed in the illustrated apparatus and melted. The vacuum was replaced in the spin chamber below the orifice with argon at one atmosphere and the argon pressure over the melt was raised to 15 psig. to extrude the melt through a 100 micron orifice. Wo fibers were formed in the argon atmosphere.

EXAMPLE 20 Example 19 was repeated with the exception that the argon below the orifice was replaced with the oxygen-argon atmosphere of Example l8 at one atmosphere pressure. This atmosphere, which was introduced after streaming of the melt had occurred, formed a stabilizing film of aluminum oxide on the molten stream; fiber formation occurred as the vacuum was replaced by the film-forming atmosphere. The fibers formed had diameters ranging from 90 to 100 microns.

EXAMPLE 21 The ex eriment described in Exam le 1 was repeated with the exception that an alloy of 73· 5 £ nickel, 20%> Chromium, 3.5% aluminum and 3% iron was employed. The alloy was prepared by melting metals whose purity was greater than 99.5$ at 1, 560°C, cooling the melt and machining the surface of the alloy to remove any surface scale. The alloy was placed in the spinning apparatus, the system evacuated, the alloy melted and the vacuum below the orifice replaced with argon at one atmosphere pressure and the melt extruded by applying argon at 100 psig. No fibers were formed.

EXAMPLE 22 The experiment described in Example 21 was repeated with the exception that the argon below the orifice was replaced with an atmosphere containing 0.2 atmosphere oxygen and 0.8 atmosphere nitrogen. The fibers obtained averaged 37 microns in diameter, possessed a tensile strength of 70, 000 psi and an elongation of 9.6%.

Here again, the stabilizing layer was aluminum oxide.

EXAMPLE 23 This example demonstrates that a stabilizing film must be applied to the streaming melt rapidly in order to stabilize against varicose breakup and that the reaction of either aluminum or iron with nitrogen takes place too slowly to stabilize the stream.

An alloy was prepared which consisted of National Bureau of Standards 90% steel (containing 0.6% carbon) and 10 electrical conductivity grade aluminum.

The alloy was melted in vacuo, held for 10 minutes, cooled, and the surface of the slug machined smooth. The alloy ° in vacuo and the pressure on the melt raised to one atmosphere argon while melting took place. The alloy was completely molten at l, .2$°G, The argon pressure above the melt was increased to 100 psig. and the vacuum below the orifice replaced with one atmosphere of nitrogen. Filaments were extruded through 100 micron diameter sapphire orifices. Due to the low rate of reactivity of aluminum with nitrogen, the aluminum nitride sheath did not form rapidly enough to completely stabilize the stream and a mixture of spherical shot and bead-like chains were obtained.

EXAMPLE 2k This example demonstrates that the reaction between aluminum and ammonia proceeds rapidly enough to provide a stabilizing film of aluminum nitride which then protects the stream against breakup.

An alloy was prepared consisting of 90 steel (containing 0.1 carbon) and 10$ electrical conductivity grade aluminum. The alloy was melted in vacuo, held in the molten state and mixed with a gentle stream of argon for about 5 minutes, cooled and the surface machined smooth.

The alloy was placed in the spinning apparatus and melted, extruded under 100 psig. argon pressure into a chamber atmosphere containing 0.9 atmosphere pressure of ΙίΗ^ and 0.1 atmosphere of nitrogen. The fibers obtained were 75 -100 microns in diameter, up to 80 cm. in length and showed ultimate tensile strengths of 56, 000-58, 000 psi. and an elongation of 1 · 0-^.6$.

EXAMPLE 25 This example demonstrates that thorium and ox n r c stabilizing sheath to allow filament formation.

An alloy was prepared by melting in vacuo 2. 25 grams of thorium metal and 20.19 grams of nickel, each of which had purities greater than 99.5$. When the melt had reached 1300°C, a gentle stream of argon was bubbled through to insure complete mixing. After 10 minutes at l-.00-l500°C. , the melt was cooled and polished. The alloy was then introduced into the spinning apparatus and melted in vacuum. When the temperature of the melt reached approximately ll4.00°C, the chamber was pressurized with 100 psi. argon and the alloy extruded into a mixture of 80$ nitrogen and 0 oxygen at 1 atmosphere pressure. Short fibers were obtained which were smooth, iridescent, and coated with a thin film of the greyish oxide of thorium. Since the surface tension of molten thorium is greater than that of molten aluminum, this experiment, upon comparison with that of Example 2l+, demonstrates that the surface energy relationship between the constituents of a freely streaming melt does not, as regards stream disruption by varicose action, override the importance of film stabilization.

EXAMPLE 26 This example, as distinguished from the results obtained in Example 23 demonstrates that with more highly reactive metals, the reaction with nitrogen is sufficiently rapid to yield a stabilizing nitride film.

Uranium metal having an analysis of 0.10$ chromium, 0.10$ iron, 0. 00j?$ manganese, 0. 010$ nickel, 0. 002$ vanadium and greater than 99.75$ uranium was intr duc d c d at about 1200°C. The melt was then pressurized with 80 psig. argon and the melt extruded into nitrogen at 1 atmosphere pressure through multiple orifices 100 microns in diameter.

Fibers ranging in length up to about 1 foot were obtained. The fiber was slightly rough surfaced, shiny and grey in color. The stabilizing sheath was presumed to be uranium nitride.

EXAMPLE 27 This example demonstrates that, with utilization of the proper stabilizing atmosphere, such desirable materials as stainless steel can be successfully spun into filaments.

Type I.O6 stainless steel in the form of 25 gauge strip was premelted in vacuo, cooled and machined. The alloy had an analysis of 12.6 chromium, 3.50$ aluminum, 2.10$ nickel, 0.25$ manganese, 0.15$ carbon, 0.10$ copper and 81. 30$ iron. The alloy was charged to the spinning apparatus, melted in vacuo and pressure extruded under 65 psig. argon into 0.8 atmosphere of nitrogen pressure and 0.2 atmosphere oxygen at a melt temperature of about 1-+80°C. Fibers with average diameters of 90-100 microns were obtained in lengths up to several feet.

EXAMPLE 28 The experiment described in Example 19 was repeated with the exception that an alloy containing approximately 9i+$ copper and 6$ aluminum was employed. The alloy was prepared by melting electrolytic grade (99.9$) copper and the aluminum alloy of Example 1 in vacuum. The cooled melt had the following approximate analysis: Metal Percent by Weight Copper 3.

Aluminum 5.83 Magnesium Ο. 36 The alloy was exposed to aqua regia, rinsed and placed in the apparatus for spinning. The alloy was melted and a pressure of 15 psig. nitrogen applied to extrude the melt into an argon atmosphere at 1 atmosphere pressure.

No fiber was obtained.

EXAMPLE 29 The experiment described in Example 28 was repeated with the exception that the atmosphere below the orifice was replaced with the oxygen-argon atmosphere of Example 18 at 1 atmosphere pressure, resulting in the formation of an aluminum oxide film on the molten stream to give successful fiber formation. The fibers exhibited an average diameter of 70 microns.

EXAMPLE 30 An alloy containing 3 electrolytic grade zinc (99.99 Zn) and 66 high purity aluminum (99.99 A1) was extruded. The procedure adopted consisted of charging the material to be extruded into the crucible and placing the crucible in the apparatus shown in the drawing. The system was closed and evacuated to a vacuum of approximately 0.1 Torr. The vacuum in the section containing the crucible was broken with argon to 5 psig. pressure. The sample was heated to 760°C. to provide a uniform melt and the pressure over the melt was increased to 25 psig. The molten stream of the zinc-aluminum alloy failed to form fiber when extruded into the vacuum.

EXAMPLE 31 The experiment described in Example 30 was repeated with the exception that the vacuum below the orifice was replaced ..with the oxygen-argon atmosphere of of either an aluminum oxide or a zinc oxide stabilising film. Fibers with an average diameter of 8 microns were obtained. The tensile strength of the fibers average -4.0 , 000 psi. with an elongation of 3>.$%. The surface of the fibers was observed to be smooth.

EXAMPLE 32 Example 30 was repeated with the exceptions that the vacuum below the orifice was replaced with argon at 1 atmosphere pressure and that electrolytic grade zinc ( 99. 99/» Zn) was used instead of the zinc-aluminum alloy.

Ho fiber was formed when the metal was heated to L|.700C. and extruded through a $0 micron orifice into the inert atmosphere .

EXAMPLE 33 Example 32 was repeated with the exception that the argon atmosphere below the orifice was replaced with an atmosphere consisting of 17.5$ anhydrous ammonia and 82.5 argon at 1 atmosphere pressure. Fibers with diameters ranging from 160- 300 microns were obtained. The fibers were observed to be very ductile and had a lustrous appearance, indicating that the stabilizing film (presumably zinc nitride was quite thin.

EXAMPLE 3I4.

Example 30 was repeated with the exception that high purity grade tin ( 99. 99^ Sn, m.p. 232°C) was employed in place of the zinc-aluminum alloy. An orifice with a diameter of l80 microns was employed. No fiber was formed in the inert argon atmosphere and only tin shot was collected.

EXAMPLE 3$ Example 3k- was repeated with the exception that the atmosphere below the orifice was replaced with one containing 2.0%, oxygen and Q0% argon. Tin fibers with an average diameter of 120 microns were obtained. The as-spun fibers were ductile and highly lustrous, indicating that the tin oxide (SnO or Sn02) stabilizing film was of a minute thickness.

The next series of experiments demonstrate the feasibility of the concept that streams of low viscosity liquids may be stabilized by decomposition of the spinning atmosphere to form stabilizing films. This technique of stabilizing non-viscous streams long enough to permit solidification is to be contrasted with the concept demonstrated by the previous examples, where the stabilizing film was formed by a reaction between the molten stream and various reactive atmospheres.

EXAMPLE 36 The illustrated apparatus was employed to melt high purity manganese (99.96% Mn). The space above the melt was charged with argon gas at 100 psig. The melt was extruded into 1 atmosphere of argon. The' molten manganese, when spun through a 100-micron diameter orifice, failed to form fiber.

EXAMPLE 37 Example 36 was repeated with the exception that carbon disulfide was introduced into the area below the orifice to maintain a pressure of 0.39 atmosphere.

Black-coated manganese fibers with an average diameter of 100 microns were obtained. The coatin was easil removed to give a lustrous surface. The stabilizing film in this was either manganous sulfide or carbon.

EXAMPLE 38 High purity copper ( 99. 99 Cu) was melted in a vacuum. The space above the melt was charged with argon gas at 100 psig. while argon was the space below the orifice to bring the pressure to 1 atmosphere. No fiber was formed in the inert argon atmosphere.

EXAMPLE 39 Example 38 was repeated with the exception that carbon disulfide was vaporized into the area beneath the orifice to maintain a pressure of 0.27 atmosphere in place of the argon atmosphere employed in Example 38.

Copper fibers with an average diameter of 80 microns were formed. The fibers exhibited a black-coated surface which was easily removed by rubbing with a cotton cloth to reveal a shiny copper surface. It is postulated that the black coating (presumably the stabilizing layer) was carbon. The tensile strength of the copper fibers was 13, 500 psi. and their elongation was 22.0 .

EXAMPLE 1+0 Example 38 was repeated with the exception that gold ( 99. 97 Au) was employed instead of copper and the temperature of the melt was maintained at approximately 1100°C. Wo fiber formed after ejection of the molten gold into the inert argon atmosphere.

EXAMPLE l Example JL4.O was repeated with the exception that the argon was replaced with Ο. 33 atmosphere of bo o f ber w formed on extrudin the molten gold into the carbon disulfide atmosphere. It is speculated that the potentially stabilizing film of carbon formed by decomposition of the carbon disulfide was disrupted by volatilization of sulfur (which has a boiling point of only kl+5°C.), since the temperature of the molten streams was approximately 1100°C.

EXAMPLE 1+2 Example I4.O was repeated with the exception that the argon atmosphere was replaced with propane gas diameter at 1 atmosphere pressure. Gold fibers with the(¾¾«fl-*e*> of 50 microns were obtained.

Examples 1+3 and demonstrate that the process of spinning materials with low viscosities by suppressing Rayleigh wave formation is not restricted to metals and may be applied to organic substances. The material selected for the substrate in these examples was unsubstituted adipamide. Adipamide when melted has a viscosity, when measured at 236°C. with a conventional Ostwald viscometer, of approximately 1 poise. This viscosity is obviously unsuitable for melt spinning by conventional techniques where polymeric organic substances with viscosities in excess of £00 poises are normally required. The objective in these examples was to deposit a boron oxide stabilizing film on the surface of the streaming liquid adipamide stream by the reaction of a gaseous boron trichloride with the moisture in the system to stabilize the stream.

EXAMPLE 1+3 Commercial grade unsubstituted adipamide, which had been exposed in a humid atmosphere to provide suspended within the melt indicated that the melt temperature varied from 230-260°C. during the course of the experiment. The pressure above the melt was adjusted to approximately 1 psig. argon to retard the loss of water from the melt.

The area below the orifice was charged with air at 0.5 atmosphere and the molten adipamide extruded through a 100 micron orifice by the application of argon gas at 60 psig. No fiber was obtained on extruding the melt and only adipamide shot approximately 500 microns in diameter was obtained.

EXAMPLE U Example .4.3 was repeated with the exception that after the molten adipamide began to flow from the orifice, boron trichloride gas was introduced beneath the orifice at a pressure of 1 atmosphere. A slightly tacky, amber fiber was obtained which melted at 192-210°C. The fiber gave a positive flame test for boron.

Examples k-S and ij.6 demonstrate that the process of spinning materials with low viscosities by suppressing varicose instabibity with a stabilizing film can be extended to include, for example, the refractory inorganic oxides. In this particular case, the material selected for the inorganic oxide was a mixture of calcium oxide and aluminum oxide.

EXAMPLE k$ The illustrated apparatus was employed to melt the refractory inorganic oxide sample, which was prepared by fusing a thoroughly mixed powder consisting of calcium oxide and aluminum oxide. The cooled melt was machined and subsequently fused at 1600°C. by inductively heating in a graphite crucible. The area above the melt was charged with argon gas at 50 psig. and the molten mixture extruded through a 225 micron orifice into an atmosphere of argon. No fiber was produced, only shot being formed.

EXAMPLE i+6 Example £ was repeated with the exception that propane was introduced into the area below the orifice to maintain a pressure of 1 atmosphere. Fibers from the aluminum oxide-calcium oxide mixture with an average diameter of 200 microns were obtained. The stabilizing film of carbon was readily removed to give almost transparent fibers which had a tensile strength of 105*000 psi.

EXAMPLE 1+7 The following example demonstrates that the presence of the stabilizing film is capable of effecting successful spinning even when the emerging stream remains well above its melting point for a considerable distance beyond the spinning orifice, thus emphasizing its stabilizing influence .

The system of Example 1 was evacuated to 0.5 m . Hg pressure and a charge composed of high purity tin, as employed in Example 3 > was used. The molten tin was extruded through a 100 micron orifice at a temperature of 260-325°C into a film-forming atmosphere of 7% helium and 33ί^ oxygen at 1 atmosphere pressure.

Although this entire temperature range is above the melting point of tin, whereby solidification was delayed, good fibers were obtained over the entire range utilized via formation of a tin oxide layer resulting from the reaction of the molte tin stream with the ox en-containin atmosphere.

EXAMPLE 1+8 The following example further demonstrates the efficacy of stabilizing the molten stream with a surface film at temperatures wherein the stream remains molten for a considerable distance beyond the orifice. In addition, the adverse effect of using a higher density atmosphere is also demonstrated.

Utilizing the apparatus of Example 1+7 and again employing highly pure tin ( 99. 9 Sn), the system was evacuated to less than 5 mm* Hg pressure and the tin melted. The vacuum above the orifice was replaced with 1+0 psig. argon and the vacuum below the orifice with one atmosphere pressure of Q0% nitrogen-20^ oxygen mixture.

Again, although the melt temperature was varied from 2l+5-1+60°C, fiber was formed throughout the temperature range.

Because the l^/Og mixture was more dense than the Ο2/ΗΘ mixture of Example 1+7, the extruded filaments exhibited a number of joints or enlarged segments along its length.

Examples 1+9 and $0 illustrate the capability of forming small diameter filaments. In addition, Example 1+9 shows the effect of excessively high spinning velocity.

EXAMPLE 1+9 Wood's metal at 190°C. was extruded through a 35 micron orifice into an atmosphere consisting of 67% helium and 33% oxygen, resulting in the formation of filaments 20 microns in diameter. The spinning velocity was maintained within a Rayleigh parameter range of 17-22+» Good fiber was obtained over this entire range. However, velocity increased.

EXAMPLE 50 Tin metal at 3 0°C. was extruded through a 35 micron orifice into an atmosphere of 67 helium and 33 oxygen and the spin velocity was maintained within a Rayleigh parameter range from 6 to 7. Tin fiber 25 microns in diameter was obtained in lengths exceeding feet.

As previously mentioned, for successful spinning the stabilizing film must have a rather low solubility in the molten metal, in addition to a relatively higher melting point. Table IV shows the relatively low solubility of stabilizing films in several representative metals .

TABLE IV Film Solubility Temp, of Film-Generating Film in Melt ¾■ Melt Metal Atmosphere Composition ( t. %) (T°C. ) Tin °2 SnO or S 02 0.1 250° Lead °2 PbO 1 X 10" 3 360 ° Aluminum °2 A1203 X io"2 670° Beryllium °2 BeO 0.2 1270° Copper cs2 C 1.0 X io"1 1100° Manganese cs2 MnS 0. 3 1238° Gold C 3H8 C 0. 3 1063° Aluminum H2S A12S3 8 X 10~2 670° Aluminum NH3 A1N X IO"3 670° w ■55- Hansen, M. , Constitution of Binary Alloys, McGra^ - Hill Book Co. , Inc. (1958 ) .

The sulfide of manganese, besides having the is also relatively insoluble in the molten metal (see above Table IV). An attempt to spin manganese into a third of an atmosphere of carbon disulfide, where the reaction 2 Mn + CS2 C + 2MnS is possible, resulted in a copious yield of fiber having a thin black coating. See Example 37. Presumably, manganese sulfide was mainly responsible for the coating since carbon is soluble in molten manganese.

In the spinning of alloys, the problems of sufficiently high melting point and solubility of the film material can be resolved if the alloy contains even a small amount of a constituent which readily forms a satisfactory stabilizing film, as heretofore characterized. Small additions of aluminum to melts of otherwise difficult spinnability resulted in satisfactory filament formation under otherwise unchanged spinning conditions.

A moderate addition of aluminum to 1030 steel (a carbon steel) resulted in filaments when spinning was carried out in an oxygen-containing atmosphere. See Example 20. This is particularly significant, for in the case of iron, the oxide is not only soluble in the molten metal, but also has a lower melting point. Similarly, no difficulty was encountered in producing fiber from Chromel R (A superalloy of nickel containing yfo aluminum).

It will readily be appreciated that, though the preceding discussion has dealt in large measure with vertical spinning under pressure, other modes for generating the low viscosity stream are as well contemplated. For example, the stream may be generated by centrifugal cone rotated at high speed and supplied centrally by the melt would form a uniform film over the face of the cone, which film would depart the periphery of the cone in the form of ligaments that may be gas-stream attenuated pending solidification. Also, it is as well contemplated that conjugated filamentary structures can be extruded according to the present film-stabilization concept, wherein two or more discrete sources of melt are brought into confluence in the vicinity of the extrusion orifice. It should also be emphasized that the formation of the film at the proper time has been found to be far more important than the exact film thickness, though the necessary film thickness has been found in virtually all ins tances to be quite minute. For this reason, it is frequently more desirable to spin considerably above the melting point in order virtue/ to speed the film formation process by vihpu e of the increased rate of reaction normally attained.

It may now be appreciated that there has been herewith disclosed a unique and most beneficial process by the practice of which there may economically be obtained useful and often unique articles. By the concepts here set forth, one is now enabled to obtain shaped articles of extrusion directly from low viscosity melts, a practice heretofore fraught with great dif culties.

By virtue of having ascertained the mechanism by which low viscosity streams are caused to break up, it has now been discovered that such streams may be efficiently stabilized by a process in which a suitable film is caused to envelope the stream as it is formed and to thereby maintain substantial continuit endin solidification of the stream by normal heat transfer phenomena. Thus enlightened, many obvious variations, modifications and substitutions will readily occur to those skilled in the art. It is to be understood, therefore, that the invention here set forth, both as regards the many possible process manipulations and product modificat ons, is limited only by a proper construction of the language of the appended claims.

Claims (3)

1. HAVING NOW particularly desorlbed and ascertained the nature of our said invention and in what manner the same is to be performed, we declare that what we claim is : 1. The method of forming shaped articles directly from low viscosity melts characterized by generating the melt as a free stream, simultaneously contacting said stream with a film-forming atmosphere to thereby generate a film about said stream as it is given issue, whereby said stream is substantially stabilized against breakup pending solidification.

2. The method of Claim 1, characterized in that the free-streaming melt has a breakup time of less than where = density of the melt D = stream diameter Ύ= surface tension of the melt.

3. The method of Claim 1, characterized in that the free-streaming melt has a viscosity of less than 1,000 poises. l . The method of Claim 1, characterized in that the free-streaming melt is caused to issue at a velocity within the range given by the relationship 1 < V /yD y ^ , where = density of the melt D = stream diameter y = surface tension of the melt V = stream velocity. 5· The method of Claim 1, characterized in that the free-streaming melt is maintained at a velocity within the range given by the relationship l^ y>D/y <(30, where Y3 = density of the melt D = stream diameter = surface tension of the melt. V = stream velocity. 6. The method of Claim 1, characterized in that the free-streaming melt is maintained at a velocity within the range given by the relationship 2 ^ V yD/ ^2 > where = density of the melt D = stream diameter = surface tension of the melt V = stream velocity. 7. The method of Claim 1, characterized in thatthe free-streaming melt is maintained at a velocity within the range given by the relationship V D/y 10, where density of the melt D = stream diameter y = surface tension of the melt. V = stream velocity. 8. The method of Claim 1, characterized in that said film-forming atmosphere is chemically reactive with said melt to form said stabilizing film. 9. The method of Claims 1 to Ij., characterized in that said film-forming atmosphere is decomposed to form said stabilizing film. 10. The method of Claims 1 to Ij., characterized in that said film-forming atmosphere is precipitated upon the stream surface to form said stabilizing film. 11. The method of Claim l , characterized 25 80/2 in that the stabilising film is characterized by a solubility in said melt of less than 10¾ by weight of the melt at the melting point of said melt* 12· The method of claim 1+, characterized in that the stabilizing film has a viscosity of at least 1 ,000 poises at the melting point of said stream* 13* The process of forming fibers and filaments according to the method olaimed in claims 1 -12 wherein the melt is a molten metal or metal alloy which is extruded as a free stream into a film-forming atmosphere to thereby form a stabilizing film on the molten free stream prior to solidi ication* 1-4.. The process of forming fibers and filaments of molten metal or metal alloy as claimed in claim 1 , which comprises extruding the molten metal or metal alloy as a free stream into a film forming atmosphere through an orifice in sapphire or single crystal aluminum oxide* 5· A process for forming metal fibers and filaments as claimed in claim 13 which comprises extruding into air an alloy comprising the first metal the oxide of which is soluble in the molten alloy and the second metal the oxide of which is insoluble In the molten alloy. 16* The process of forming fibers and filaments as claimed in claim ill the alloy being an iron alloy. 17* The process of forming fibers and ilaments as olaimed in claim 16 which comprises extruding the molten alloy as a free stream into an oxygen containing atmosphere to thereby form an oxide film on said free stream prior to solidi ication, said iron alloy containing at least a second metal the oxide of which is substantially insoluble in said free stream* 252*80/2/3 18. The process of claim 15 wherein the second metal is aluminum. 19· The process of claim J+ wherein said atmosphere contains elemental oxygen. 20. The process of claims 16 or 17 wherein the stainless iron alloy is/steel. 21 · The process of forming fibers and filaments as claimed in claim 15 of a molten metal the oxides of which are soluble in the molten metal which comprises extruding the molten metal into a gaseous hydrocarbon atmosphere which gaseous hydrocarbon decomposes to provide a stream stabilizing carbon film on the molten free stream upon contact therewith. 22. The process of claims , 1k, 1 and 21 wherein the molten metal is copper. 23· The process of claim 20 wherein the gaseous hydrocarbon is propane. 2 » A shaped metallic article having an internal micro-structure characterized by an average dendrite spacing of less than about 25 microns. 25· The article of claim 21+, characterized in that said average dendrite spacing is less than about 5 microns. 26· The article of claim 2 » further characterized by a core and a thin film substantially enveloping said core, said film having a solubility of less than 0$ by weight of said core at the melting point of said core* 27· The article of claim 26, characterized in that said film has a melting point higher than that of said core. 25U80/23 28. The product of* claim 26, characterized in that said film has a viscosity of at least 1 ,000 poises at the melting point of said core. 29· The product of claim 26, characterized by as hereinbefore being in the form of a ilament having an aspect ratio^" defined-) freater than unity. 30. The product of claim 29» characterized by being in the form of a filamentous network. 31 · A method of forming shaped articles directly from low viscosity melts substantially as described in the herein Examples. 32· A shaped article whenever produced by the method elalmed in any one of claims 1 to 12 and 31 · 33· The articl of elalm 32 In the form of a fiber or a filament. 3Ι -· Fibers and filaments of metal or metal alloy whenever produced by the process claimed in any one of claims 13 to 22· Attorney for Applicants