US 20010046597 A1
In accordance with the invention, a reactive multilayer structure comprises alternating layers of materials that exothermically react by a self-propagating reduction/oxidation reaction or by a self-propagating reduction/formation reaction. This combination of a reduction reaction and either an oxidation or formation reaction can lead to ductile reaction products and is frequently accompanied by the generation of large amounts of heat. As compared with conventional multilayer foils, the new multilayer structures are easier to fabricate, easier to handle, and produce more reliable bonds.
1. In a reactive multilayer structure comprising alternating layers of materials that react exothermically to produce one or more reaction products, each of the layers having a thickness in the range 1-10,000 nanometers and the multilayer structure having a total thickness in the range 10 micrometer to 1 centimeter, the improvement wherein:
the materials of respective alternating layers react by a self-propagating reduction/oxidation reaction or a self-propagating reduction/formation reaction.
2. The improved reactive structure of
3. The improved reactive structure of
4. The improved reactive structure of
5. The improved reactive structure of
6. The improved reactive structure of
7. The improved reactive structure of
8. The improved reactive structure of
9. The improved reactive structure of
10. The improved reactive structure of
11. The improved reactive structure of
12. The improved reactive structure of
13. The process of bonding two bodies comprising the steps of:
disposing between the two bodies an improved reactive structure in accordance with
pressing the bodies against the reactive structure; and
igniting the reactive structure.
14. The method of
15. The method of
 This application claims the benefit of U.S. Provisional Application Serial No. 60/201,292 filed by T. P. Weihs et al. on May 2, 2000 and entitled “Reactive Multilayer Foils”. It is related to U.S. patent application Ser. No. ______ filed by M. E. Reiss et al. concurrently herewith and entitled “Method of Making Reactive Multilayer Foil and Resulting Product” and U.S. patent application Ser. No. ______ filed by T. P. Weihs et al. concurrently herewith and entitled “Freestanding Reactive Multilayer Foils”. These three related applications are incorporated herein by reference.
 This invention was made with government support under NSF Grant Nos. DMR-9702546 and DMR-9632526, and The Army Research Lab/Advanced Materials Characterization Program through Award No. 019620047. The government has certain rights in the invention.
 This invention relates to reactive multilayer structures, and, in particular, to reactive multilayer structures that can be easily processed to produce ductile reaction products.
 Reactive multilayer coatings are useful in a wide variety of applications requiring the generation of intense, controlled amounts of heat in a planar region. Such structures conventionally comprise a succession of substrate-supported coatings that, upon appropriate excitation, undergo a self-propagating exothermic chemical reaction that spreads across the area covered by the layers. While we will describe these reactive coatings primarily as sources of heat for welding, soldering or brazing, they can also be used in other applications requiring controlled local generation of heat such as propulsion and ignition.
 Many methods of bonding require a heat source. The heat source may be external or internal to the structure to be joined. An external source is typically a furnace that heats the entire unit to be bonded, including the bodies (bulk materials) to be joined and the joining material. An external heat source often presents problems because the bulk materials can be sensitive to the high temperatures required for joining. The bulk materials can also be damaged in cooling from high temperatures due to mismatches in thermal contraction.
 Internal heat sources often take the form of reactive powder. Reactive powders are typically mixtures of metals or compounds that react exothermically. Such powders, developed in the early 1960s, fostered bonding by Self-Propagating, High-Temperature Synthesis (SHS). However, the energy released and its diffusion is often difficult to control in SHS reactions. As a result, bonding by powders may be unreliable or insufficient.
 Reactive multilayer structures, which were subsequently developed, reduced the problems associated with reactive powder bonding. These structures are typically comprised of thin coatings that undergo exothermic reactions. See, for example, T. P. Weihs, Handbook of Thin Film Process Technology, Part B, Section F.7, edited by D. A. Glocker and S. I. Shah (IOP Publishing, 1998); U.S. Pat. No. 5,538,795 issued to Barbee, Jr. et al. on Jul. 23, 1996; and U.S. Pat. No. 5,381,944 to D. M. Makowiecki et al. on Jan. 17, 1995. As compared to reactive powders, reactive multilayer structures permit exothermic reactions with more controllable and consistent heat generation. The basic driving force behind such reactions is a reduction in atomic bond energy. When the series of reactive layers is ignited, the distinct layers mix atomically, generating heat locally. This heat ignites adjacent regions of the structure, thereby permitting the reaction to travel the entire length of the structure until all the material is reacted.
 The individual layer thickness in the foils defines the average diffusion distance that is required for materials to mix in these exothermic reactions. An exothermic reaction in a multilayer foil can self-propagate far more easily and far faster at room temperature than the same reaction in a powder compact because the layers are many orders of magnitude smaller than the powders. Individual layer thicknesses typically range from 1-1000 nm while typical powder diameters range from 10 to 100 μm. Consequently, reaction velocities in foils typically range from 1-30 m/s while reaction in powders range from 0.01 to 0.1 m/s. An additional advantage for multilayer foils is that the thicknesses of their individual layers are far more uniform, consistent, and controllable than diameters of corresponding powders. Thus, reaction properties are more easily controlled and modified. Lastly, while reactive foils are fully dense and free of contaminants at interfaces between its reactants, reactive powder compacts are rarely fully dense and often contain many contaminants at reactant/reactant interfaces due, for example, to oxide coatings on the particles. Both the lack of full density and the presence of contaminants can limit reaction kinetics and velocities compared to reactive foils.
 While a clear improvement over powders, reactive multilayer structures encountered their own difficulties. For example, when attached to a substrate, the reactive foils often debond or delaminate from their substrates upon reaction. This debonding is caused by inherent reactive foil densification during reaction and by non-uniform thermal expansion on heating and contraction during cooling. In the case of joining, it significantly weakens the bonding joint. More significantly, most reactive coatings react to produce a brittle intermetallic compound, which can be detrimental at the center of a joint, lowering its fracture toughness and causing it to behave in a brittle fashion when deformed. Consequently, internal or external stresses can cause catastrophic mechanical failure of the joint.
 To date, most research and development of self-propagating reactions in foils has been directed primarily to formation reactions wherein two or more elements (A/B) mix and react to form a compound product (ABx). While such reactions may produce large heats of reaction, many difficulties are encountered in fabricating and using the requisite foils. The reactants are typically expensive or are brittle, hard to deposit and difficult to use. Many of the foils are subject to unwanted ignition. Moreover, many of the reactions produce brittle final products.
 These difficulties can be illustrated by the problems with foils having B, C or Si layers. Table 1, which lists pertinent conventional formation reactions, shows that many of these reactions combine a transition metal such as Ti, Zr, Hf, V, Nb, Ta, Ni, Pd, or Pt with a light element such as B, C, Si, or Al. It also shows that the borides, carbides, and silicides generally have higher heats than the aluminides. (The two exceptions are very expensive due to the use of Pd and Pt, and therefore have very limited commercial potential.) Thus reactive foils with high heats of reaction generally employ reactions that form borides, carbides, or silicides.
 Unfortunately, reactive foils with B, C or Si are difficult to fabricate and use. As compared with aluminum, for example, foils with B, C or Si are more likely to delaminate or fracture during vapor deposition. When deposited at the relatively low temperatures required for making reactive multilayer foils, B, C and Si deposit in an amorphous state. Consequently the deposited layers are subject to considerable growth stresses. Thus, multilayer foils with amorphous layers of B, C and Si have a higher driving force to delaminate. In addition, multilayer foils with amorphous layers of B, C or Si are more susceptible to fracture, cracking, and delaminating than foils with Al layers.
 An additional difficulty in fabricating foils with B, C, or Si is that these materials sputter deposit at very slow rates, far slower than Al. Since sputter deposition is a preferred method of fabricating reactive multilayer foils, slow sputter rates are a severe limitation on the eventual commercialization of these foils.
 Reactive foils that contain B, C, or Si, also tend to be brittle and unstable. The amorphous layers of B, C or Si in these foils have a lower fracture toughness than the alternative Al layers, so the multilayer foils will be more susceptible to fracture and cracking during handling. This susceptibility makes cutting and patterning the foils difficult and raises the likelihood of unwanted ignition during handling due to fracture or cracking. In addition, since transition elements diffuse rapidly into amorphous layers, faster than into Al at a similar temperature, foils based on B, C, or Si will also have lower thresholds for ignition, which also makes them more susceptible to unwanted ignition.
 Lastly, when any of the above formation reactions are completed, the final reaction product is brittle at room temperature. Thus, future handling or use of this product, whether in a joint, a propellant, or a combustion reaction, can be degraded. In the particular case of joining, the presence of a brittle boride, carbide, silicide, or aluminide at the interface between the two components is bound to lower the fracture strength, fracture resistance, and fatigue resistance of the joint. Accordingly, there is a need for new reactive multilayer foils that can be easily processed and handled and can be easily used to produce ductile, reliable bonding.
 In accordance with the invention, a reactive multilayer structure comprises alternating layers of materials that exothermically react by a self-propagating reduction/oxidation reaction or by a self-propagating reduction/formation reaction. This combination of a reduction reaction and either an oxidation or formation reaction can lead to ductile reaction products and is frequently accompanied by the generation of large amounts of heat. As compared with conventional multilayer foils, the new multilayer structures are easier to fabricate, easier to handle, and produce more reliable bonds.
 The nature, advantages, and various additional features of the invention can be seen by consideration of the various illustrative embodiments now to be described in detail in connection with the accompanying drawings. In the drawings:
FIG. 1 is a schematic diagram of a reactive multilayer structure in accordance with the invention;
FIG. 2 is a schematic diagram of a reactive foil with particle composite geometry in accordance with the invention; and
FIG. 3 illustrates bonding using the ductile metal reaction product of a multilayer structure as a joining material.
 It is to be understood that these drawings are for purposes of illustrating the concepts of the invention and, except for graphical illustrations, are not to scale.
 In accordance with a preferred embodiment of the invention, a reactive multilayer structure (generically referred to herein as a “foil”) is provided as a local heat source in a variety of applications such as a process for joining materials together. As illustrated in FIG. 1, the reactive foil (14) with a layered structure is made up of alternating layers 16 and 18. The foil contains two materials, which in their simplest form consist of an element α and an oxide or compound βΓx, where α, β and Γ can designate any element and x can be an integer or a fraction. The foil will react by the element (α) reducing the initial oxide or compound (βΓx) and forming a more stable oxide or compound αΓy and the element β. This combination of a reduction and either an oxidation or a formation reaction leads to the release of heat. Examples include reactions wherein a reactive element like Al (or Si, Ti, Zr, or Hf) reduces an oxide with a low heat of formation (e.g., Fe2O3, CuO, or ZnO) and forms a metal (Fe, Cu, or Zn) plus an oxide, Al2O3 (or SiO2, TiO2, ZrO2, or HfO2) that has a very high heat of formation. Examples also include a reactive element(s) such as Ti, (or Zr and Hf) that reduces a compound with a low heat of formation, such as NiB, and then subsequently forms a metal (Ni) plus a compound (TiB2, or ZrB2 or HfB2 which has a high heat of formation.
FIG. 2 illustrates an alternate form of a reactive multilayer structure which we will call a composite particle foil 50 wherein one of the reactive materials is in the form of particles 52 (e.g. spheres, disks or fibers). The other reactive material can be in the form of layers establishing a matrix 54 for the particles. The structure 50 can use the same materials and the same reactions as the foil 14 of FIG. 1.
 The materials (α/βΓx) used in the reactive structures (14, 50) are preferably chemically distinct. In one preferred embodiment, layers 16, 18 (52/54) alternate between Al and an oxide with a low heat of formation (e.g., Fe2O3, CuO, or ZnO). In another embodiment, layers 16, 18 (52,54) alternate between a reactive early transition element such as Ti, Zr, or Hf and a boride, silicide, or carbide compound with a low heat of formation such as (NiB, FeB, FeSi, Cu3Si, Ni3C, Fe3C). In yet another embodiment, layers 16, 18 (52,54) alternate between Pt, Pd or alloys of these elements and an aluminide compound with a low heat of formation such as (CuAl2, TiAl3) In another preferred embodiment, the initial compounds comprise metallic elements that are ductile such as Fe, Cu, or Ni, so that the final product consists of this ductile metal and a hard compound. Preferably, the pairs of materials α, βΓx, are chosen so that the reactions form stable compounds with large negative heats of reaction and high adiabatic reaction temperatures. These reactions will self-propagate in a manner similar to the formation reactions described in T. P. Weihs, “Self-Propagating Reactions in Multilayer Materials,” Handbook of Thin Film Process Technology (1997), which is incorporated herein by reference in its entirety.
 When a multilayer structure 14, 50 is exposed to a stimulus (e.g., a spark or energy pulse) neighboring atoms from the two materials mix. The change in chemical bonding caused by this mixing results in a reduction in atomic bond energy, thus generating heat in an exothermic chemical reaction. This chemical bonding occurs as layers with α-αbonds (i.e., layer 16, 52) and layers with β-Γ bonds (i.e., layer 18, 54) are exchanged for α-Γ and β-β bonds, thereby reducing the chemical energy stored in the foil, and generating heat. As FIGS. 1 and 2 further illustrate, the heat that is generated diffuses through foil 14, 50 (in a direction from reacted section 30 through reaction zone 32 to unreacted section 34) and initiates additional mixing of the reactants. As a result, a self-sustaining/self propagating reaction is produced through the structure 14, 50. With sufficiently large and rapid heat generation, the reaction propagates across the entire structure 14, 50 at velocities that can exceed 10 m/s. As the reaction does not require additional atoms from the surrounding environment (as would, for example, oxygen in the case of combustion), the reaction makes foils 14 or 50 self-contained sources of energy capable of emitting bursts of heat and light, rapidly reaching temperatures above 1400 K and a local heating rate reaching as high as 109 K/s. This energy is particularly useful in applications such as propulsion, joining, and ignition requiring production of heat rapidly and locally.
 When a reaction propagates across a multilayer structure 14, 50 as illustrated in FIGS. 1 and 2, the maximum temperature of the reaction is typically located at the trailing edge of the reaction zone 32. This may be considered the final temperature of reaction, which can be determined using the heat of reaction (ΔHrx), the heat lost to the environment (ΔHenv), the average heat capacity of the sample (Cp), and the mass of the product (M). Another factor is whether or the not reaction temperature exceeds the melting point of the final product. If the melting point is exceeded, then some heat is absorbed in the state transformation from solid to liquid of at least part of the product. With reduction/oxidation and reduction/formation reactions very often the metallic component in the product can melt due to the high reaction temperatures, while the stable compound may not. The final temperature of reaction may be determined using the following formulas (where To is the initial temperature, ΔHmm is the enthalpy of melting of the final metallic phase, Tm is the melting temperature of the final metallic phase in the product, ΔHmc is the enthalpy of melting of the final compound phase, and Tmc is the melting temperature of the final compound phase in the product),
T f =T o−(ΔH rx +ΔH env)/(C p M)
 If no melting of final product occurs;
 If the metallic phase in the product melts only partially;
T f =T o(ΔH rx +ΔH env +ΔH mm)/(C p M)
 If the metallic phase in the product completely melts.
 If the compound phase in the product melts only partially; and
T f =T o(ΔH rx +ΔH env +ΔH mm +ΔH mc)/(C p M)
 If the metallic and compound phases in the product melt completely.
 Intricately related to the heat of the foil reaction is the velocity of the propagation of the reaction along the length of foil 14, 50. The speed at which the reaction propagates depends, in particular, on how rapidly the atoms diffuse normal to their layering or particles (FIG. 1 or 2) and how rapidly heat is conducted along the length of foil 14, 50. However, now at least three elements are involved in the reaction α, β, and Γ compared to simple formation reactions that can involve only two elements. But, only one of the elements must diffuse to complete the reduction/oxidation or reduction/formation reactions. In most cases the diffusion of O, Si, B, or C between the layers (or particles and matrix) will control the rate of the reaction. The propagation velocity is a strong function of the foil's multilayer thickness or average particle thickness. As the thickness of individual layers 16, 18 (or particles 54) decreases, the diffusion distances are smaller and atoms can mix more rapidly. Heat is released at a higher rate, and, therefore, the reaction travels faster through the foil structure. Reactive foils typically have diffusion distances that range from 1-1000 nm while reactive powder compacts typically have diffusion distances that range from 10-100 μm. Hence, reaction rates and reaction velocities are many times faster in foils than in powder compacts.
 In accordance with a preferred embodiment of the invention, reactive multilayer foils 14, 50 may be fabricated by physical vapor deposition (PVD), chemical vapor deposition, electrochemical methods, electroless methods, mechanical methods, or some combination of these methods. A magnetron sputtering technique, for example, may be used to deposit the materials α/βΓx on a substrate (shown in FIG. 1 in dashed outline form as layer 35) as alternating layers 16, 18. Substrate 35 may be rotated over two isolated sputter guns in a manner well known in the art to effectuate the layering of materials α/βΓx into alternating layers 16, 18.
 The vapor streams from the two sputter guns or the two electron beam hearths are isolated from one another during deposition of a reactive multilayer foils 14. This isolation reduces intermixing and unwanted reaction of the elements being deposited. It is important to isolate the two vapor streams from one another to prevent loss of the energy of the reaction during deposition.
 Substrate 35 is shown in dashed outline form to indicate that it is a removable layer that facilitates fabrication of the reactive foil 14 as a freestanding foil. Substrate 35 may be any substrate (e.g., Si, glass, or other underlayer) having the characteristics of providing sufficient adhesion so as to keep the foil layers on the substrate during deposition, but not too adhesive to prevent the foil from being removed from the substrate following deposition.
 In accordance with a preferred embodiment, an additional wetting layer (e.g., tin) may be used as an interface layer between the first layer of foil (16 or 18) and the substrate 35 to provide the necessary adhesive. When no wetting layer is employed, selection of the appropriate material αor βΓx as the first layer deposited on the substrate will ensure that the necessary adhesive requirements are met. When a reactive foil using Al/Cu2O as materials α/βΓx, is to be fabricated, for example, without a wetting layer, the exemplary reactive foil would be deposited on a substrate such as Si with the first layer being Al deposited on the substrate. Al is preferably selected as the first layer in such case because Al will sufficiently adhere to Si during depositing, but will allow peeling off of the substrate after the foil is formed.
 A fabricated foil 14 may have hundreds to thousands of alternating layers 16 and 18 stacked on one another. Individual layers 16 and 18 preferably have a thickness ranging from 1-1000 nm. In a preferred embodiment, the total thickness of foil 14 may range from 10 μm to 1 m.
 Another preferred method of fabricating is to deposit material in a codeposition geometry. Using this method, both material sources are directed onto one substrate and the atomic fluxes from each material source are shuttered to deposit the alternate layers 16 and 18. Again, care must be taken to isolate the two physically distinct atomic fluxes from each other.
 In accordance with a preferred embodiment, the degree of atomic intermixing of materials α/βΓx that may occur during deposition should be minimized. This may be accomplished by depositing the multilayers onto cooled substrates, particularly when multilayers 16 and 18 are sputter deposited. To the extent that some degree of intermixing is unavoidable, a relatively thin (as compared to the alternating unreacted layers) region of pre-reacted material 20 will be formed. Such a pre-reacted region 20, nevertheless, is helpful in that it serves to prevent further and spontaneous reaction in foil 14.
 As illustrated in FIGS. 1 and 2, the reactive foil 14 or 50 can have a layered or particle composite geometry. While a layered geometry will typically result from vapor deposition of a reactive foil, another preferred embodiment of this invention is the mechanical formation of α/βΓx reactive foils. In this method, sheets of α and βΓx are stacked, inserted in a removable protective jacket, and then deformed into a multilayer sheet, as by swaging and rolling. The jacket is then removed. This mechanical processing can result in either a layered or particle composite geometry and generally is less expensive than vapor deposition. For further detail concerning mechanical processing, see U.S. application Ser. No. ______, filed by M. Reiss et al. concurrently herewith and entitled “Method of Making Reactive Multilayer Foil and Resulting Product” which is incorporated here by reference.
 Reactive foils in accordance with the invention may be adapted for use in a variety of applications. In one preferred application, the foils may be used to ignite another reaction that releases a signal, more heat, or a gas, as in a combustion or propulsion application. In this case, a freestanding foil can be inserted into a metastable material to be ignited, with all sides of the foil being covered by the metastable material.
FIG. 3 illustrates another preferred application of the invention wherein the reactive structures react to form a ductile composite product 40 that contains particles 42 or layers of a hard oxide or compound in a matrix 41 of a ductile metal. The ductile metal 41 can be a simple element such as Cu, Ni, or Fe or it could be a ductile alloy of two or more elements. In a preferred embodiment, these reactive structures can be used in the joining of two bodies or components (43,44). In these applications, the reactive multilayer is positioned between the two bodies (43, 44) to be joined, the bodies are pressed against the reactive multilayer, and the latter is ignited. The ductile metallic product 41 resulting from the self-propagating reaction can serve as a solder or braze that flows and wets the surfaces of the bodies, and consequently forms a strong joint. Thus, reactive multilayers described in the present invention essentially enable a braze-free room-temperature joining process. This process provides significant advantages over known reactive joining methods which require braze of solder material to be deposited on, or positioned next to the free-standing foil and/or on the surfaces of the components.
 In another embodiment of the invention, the foils are fabricated using inexpensive materials such as Al and CuO, Fe2O3 or ZnO. These materials are less expensive than many of the elements used to fabricate reactive foils with very exothermic formation reactions (as opposed to reduction/oxidation or reduction/formation reactions) such as Ti, Zr, Hf, or Nb.
 Preferred embodiments of the invention are useable as freestanding reactive foils 14 with increased total thickness. The total thickness of such a reactive foil depends upon the thickness and number of the elemental layers (e.g., 16 and 18) utilized to form the foils. Foils that are less than 10 μm are very hard to handle as they tend to curl up on themselves. Foils on the order of 100 μm are stiff, and thus, easily handled. Thicker foils also minimize quenching. In joining applications, for example, using reactive foils, there is a critical balance between the rate at which the foil generates heat and the rate at which that heat is conducted into the surrounding braze layers and the joint to be formed. If heat is conducted away faster than it is generated, the reaction will be quenched and the joint cannot be formed. The thicker foils make it harder to quench the reaction because there is a larger volume generating heat and the same surface area through which heat is lost.
 Thicker foils can be utilized with reaction temperatures that are lower, generally leading to more stable foils. Foils with high formation reaction temperatures (as opposed to reduction/oxidation or reduction/formation temperature) are generally unstable and brittle and therefore are dangerous and difficult to use. Brittle foils, for example, will crack easily leading to local hot spots (through the release of elastic strain energy and friction) that ignite the foil. Cutting such brittle foils (e.g., for specific joint sizes) is very difficult to do as they are more likely to crack into unusable pieces or ignite during the cutting process.
 In accordance with a preferred embodiment, the thicker reactive foils are on the order of 10 μm to 1 cm. Although a number of different systems may be employed to create the thick freestanding reactive foils, a unique process in selecting the fabrication conditions is advantageous. In accordance with a preferred embodiment, for example, deposition conditions such as sputter gas and substrate temperature are advantageously chosen so that stresses remain sufficiently low in the films of the foil as they are grown in the system. Since the stress in the film times its thickness determines with the driving force for delamination, the product of stress and thickness should be kept below 1000 N/m. Stresses often arise in the films during the fabrication process. As the films grow thicker, they are more likely to peel off their substrates or crack their substrates than thinner films, thereby ruining the final foil production. By characterizing the stresses on the films and selecting conditions to minimize the stresses, the fabrication process can be completed without the peeling off (or cracking) of the substrate.
 Utilizing one or more embodiments of the invention, a number of different applications can now be performed more effectively and efficiently. For example, freestanding reactive foils can be incorporated directly into solid propellants, enabling the uniform and complete combustion of components within the foil with extremely large releases of heat. Alternatively, a number of materials can now be joined more efficiently. Semiconductor devices may be bonded to circuit boards or other structures, using reduction/oxidation reactions where the final metallic product serves as a braze or solder material to join the components.
 The invention may now be more clearly understood by consideration of the following examples:
 Al/CuO/Cu Composites
 Electrodeposition of Cu from standard cupric sulphate solution is alternated, with electrodeposition of CuO (or Cu2O) from a 3 M copper lactate, 0.4 M cupric sulphate solution (pH>10 by addition of NaOH). The alternating layers can be fabricated either by moving the substrate from one bath to another or by draining and refilling the same bath with different solution. Rotation of the substrate during deposition and the suspension of a large volume percent of aluminum particles in the either or both solutions enables the aluminum particles to be incorporated in the electrodeposited matrix.
 Instead of an electrodeposited multilayer structure, the pH of the copper lactate solution can be modified such that Cu and CuO are deposited simultaneously. The oxygen content can be controlled by optimizing the pH, current density, and electrolyte concentrations.
 The aluminum particles can be 100 micron diameter down to 100 nm diameter. Cold rolling of the electrodeposited foils will create pancake structures with much smaller average diffusion lengths. Swaging will create oblong particles which can then be roll flattened for even further reduction in diffusion distances.
 It is to be understood that the above-described embodiments are illustrative of only some of the many possible specific embodiments, which can represent applications of the principles of the invention. Numerous and varied other arrangements can be made by those skilled in the art without departing from the spirit and scope of the invention.